13;}: 2. 4.. . L .7. r...

! .yav. . x... ..J#. .

12;: t
s 73 j. .
.. .J :1

 

 

 

 

 

 

THE MS

3 Illllllllllllllllllllllllllllllllllllllllllllllll

3 1293 014172

This is to certify that the

dissertation entitled

Effect of Ion-Sputtered Thermoelastic NiTi
Coatings on Fatigue Crack Initiation in
Copper

presented by

Seungho Nam

has been accepted towards fulﬁllment
of the requirements for

 

 

Ph.D. degree in Materials Science

D. S. Grummon

 

Major professor

Date August 1, 1995

 

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

Lt wR—ARY
Michigan State
University

 

 

 

PLACE It RETURN BOX to remove We checkout from your record
To AVOID FINES return on or bdore due due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

-i§
]

MSU leAn Afﬁrmative Action/Emu om Inetttulon

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.7; _

 

 

EFFECT OF ION-SPUTTERED THERMOELASTIC NiTi COATINGS ON
FATIGUE CRACK INITIATION IN COPPER

By

Seungho Nam

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Materials Science and Mechanics

1995

ABSTRACT

EFFECT OF ION-SPU'ITERED TI-IERMOELASTIC NiTi COATINGS ON FATIGUE
CRACK INITIATION IN COPPER

By

Seungho Nam

Surface effects are particularly important in fatigue, since the surface is often the
preferred site for fatigue crack initiation, and the fatigue life of materials may often be
dominated by the crack initiation stage. NiTi thin films, which give rise to shape memory
and superelasticity effects were used as a new surface coating approach to improve fatigue
lifetime. The objective of the present work was to study the effect of NiTi surface coatings
on fatigue crack initiation in polycrystalline copper, chosen as a model f cc metal. NiTi thin
ﬁlms were deposited on copper fatigue specimens by ion beam sputtering from a NiTi alloy
target. By a crystallization annealing process, superelasticity properties were provided to
the NiTi films. Fatigue specimens were subjected to carefully controlled fatigue cycling at
constant plastic strain amplitude at various temperatures with respect to the transformation
temperature of the films. The development of fatigue cracks on specimens was studied by
optical microscopy and scanning electron microscopy. The crystal structure changes on
surface films were investigated by X-ray dif f ractometery.

Adhesion between the film and an ion sputter cleaned substrate was very strong.
X-ray diffraction study demonstrated that stress induced martensites were produced in NiTi
thin films on a plastically deforming metallic substrate. The hardening rate and the
saturation stresses were not affected by NiTi thin film surface modification. Amorphous
NiTi films demonstrated their brittleness and the fatigue crack initiation life of amorphous
film coated specimens was shorter than the unmodified copper specimens. On the other

hand, all crystalline film coated specimens demonstrated improved fatigue crack initiation

lives compared to bare copper control specimens. The greatest increase in fatigue crack
initiation life was obtained from Ni49_6Ti50,4 film coated specimens, which showed around
100 % improvement over bare copper specimens. These films had austenite finish
temperatures approximately 35 °C above the test temperature, a condition associated with a
minimum in the residual superelastic strain.

The characteristics of notch-peak topography resulting from plastic strain
localization and incidence of cracks produced by cyclic strain were delayed by the presence
of thermoelastic surface coatings. In addition, the austenite phase of NiTi coatings, which
has higher yield strength and elastic compliance than the copper substrate, blocked the
egress dislocations and inhibited penetration of slip bands. Thus, the ability of crystalline
NiTi films to block (suppress) PSB extrusion from the substrate is attributed not only to
superelastic deformation of the film, but also to its strength, ductility, and high elastic

compliance.

To my wonderful parents and family:

for their endless love, encouragement and patience.

iv

ACKNOWLEDGMENT

I am deeply indebted to Dr. David S. Grummon, my major professor, for his
guidance, interest and continuous support throughout my Ph.D effort. Additionally,
sincere appreciation is extended to my guidance committee members: Prof. Martin A.

Crimp, Prof. Thomas R. Bieler, and Prof. William P. Pratt Jr.

Special thanks go to all my colleagues for their support and good friendship:
Richard Schallek, Zhiwei Cai, Zhibo Zhao, Li Hou, Heyun Yin, Kevin Durning, and
Bejeir Brooks. Thanks are given to Mr. Peter Scheuer for proof-reading drafts of my
thesis. I would also like to thank the entire MSM Dept. staff for their kindness during the
whole of my life in the U.S.A. as a student, especially Mr. Kurt Niemeyer for his help in

making the specimens for my research.

My sincere appreciation is also expressed to the National Science Foundation for

providing financial support of this research under Grant #MSS8821755.

Finally, I would like to express my appreciation to my wife, Jean—Young, and my
lovely daughter, Angie (Youn-Joo), for their never-ending love and patience throughout

my school life.

TABLE OF CONTENTS

LIST OF TABLES

LIST OF FIGURES

1. INTRODUCI‘ ION ................................................................................... 1

2. LITERATURE REVIEW ......................................................................... 5
2.1. Fatigue Behavior in FCC Metals ......................................... 6

2.1.1. Cyclic Stress-Strain Response in FCC Metals
2.1.2. Dislocation Structure in Faigued FCC Matals
2.1.2.1. Evolution of the Dislocation Structure during Cyclic Hardening
2.1.2.2. Formation of Persistent Slip Bands (PSBs)
2.1.3. Fatigue Crack Initiation in FCC Metals
2.1.4. Secondary Cyclic Hardening in Fatigued FCC Metals
2.2. Effect of Surface Modification on Fatigue Life in FCC Metals ............ 16
2.3. Transformational Superelasticity ................................. 20
2.3.1. General Aspect of Martensitic Transformation and Superelasticity
2.3.2. Displacive Transformation in NiTi Alloys
2.3.2.1. Transformation Behavior in Near-Equiatomic NiTi Alloys
2.3.2.2. Effect of Composition on Transformation Temperatures
2.3.2.3. Effect of Heat Treatment on Transformation Temperatures
2.3.2.4. Effect of Third Element Additions and Impurities
2.3.2.5. Self -Accommodation and Surface Relief
2.4. Physical and Mechanical Properties of NiTi Alloys ......................... 33
2.4.1. Deformation of NiTi Alloys
2.4.1.1. Sress Strain Behaviotr in NiTi Alloys
2.4.1.2. Two Stage Yielding in NiTi Alloys
2.4.1.3. Plastic Deformation of NiTi Alloys
2.4.1.4. Grain Size Effect on Pseudoelasticity in NiTi Alloys

vi

2.4.2. Fatigue Properties of of Near-Equiatomic NiTi Alloys
2.5. Thin Film Processing ........................................................................... 38
2.5.1. General Features of Thin Film Processing
2.5.2. Thin Film Processing by Sputter Deposition
2.5.3. Sputter of Alloys
2.5.4. Stress in Thin Films
2.5.5. NiTi Thin Film by Sputter Deposition
2.5.5.1. Crystallization of NiTi Thin Films
2.5.5.2. Transformation Behaviors of NiTi Thin Film

3. EXPERIMETAL METHODS .................................................................. 49
3.1. Fabrication of Fatigue Specimen .......................................................... 50
3.2. NiTi Film Deposition by Ion Sputtering ............................................... 52

3.2.1. Ion Beam Sputter Deposition System
3.2.2. NiTi Thin Film Deposition by Ion Sputtering
3.2.3. Crystallization of NiTi Thin Films
3.3. Chracterization of NiTi Film ................................................................. 55
3.3.1. Thickness Measurement
3.3.2. Composition Analysis
3.3.3. Transformation Temperature Measurement in NiTi Thin Films
3.3.3.1. Electrical Resistivity Measurement
3.3.3.2. Differential Scanning Caloriometry (DSC)
3.3.4. Crystal Structure Evaluation
3.4. Low Cycling Fatigue Test ..................................................................... 57
3.4.1. General Features of the Fatigue Testing System
3.4.2. Calibration of Testing System
3.4.3. Specimen Mounting
3.4.4. Plastic Strain Controlled Low Cycling Fatigue Tests
3.5. Observation of Sputtered Film Structure and Surface Damage

Accumulation ......................................................................................... 62
4. RESULT AND DISCUSSION .............................................................. 64
4.1. NiTi Thin film Fabrication by Ion Beam Sputter Deposition ................. 64

4.1.1. Composition
4.1.2. Microstructure
4.1.3. Crystallization

vii

4.1.4. Surface Relief
4.1.5. Residual Stress
4.1.6. Transformation Behavior
4.2. Fatigue Experiments ....................................................................... 74
4.2.1. Specimen Characteristics
4.2.2. Cyclic Stress Strain Hysteresis
4.2.3. Cyclic Hardening Curve and Saturation Stress

4.3. Stress-Induced Martensitic Transformation in NiTi Film .................... 76
4.4. Surface Damage Accumulation ............................................................ 79
4.4.1. Surface Crack Formation on Flat Surface Gauge Area of Surface
Modified Specimens

4.4.1.1. Specimen F 1
4.4.1.2. Specimens F2 and F3
4.4.1.3. Specimen F4
4.4.2. Fatigue Damage Accumulation in Shallow Round Notched
Surface Modified Specimens
4.4.2.1. Bare Polycrystalline Copper
4.4.2.2. Amorphous Film Coated Specimens
4.4.2.3. A/SIM Film Coated Specimens
4.4.2.4. A/T E Film Coated Specimens
4.4.2.5. A/R Film Coated Specimens
4.5. Fatigue Crack Initiation Life ............................................................. 89
4.5.1. Bare Copper specimens
4.5.2. Amorphous Film Coated Specimens
4.5.3. A/SIM Film Coated Specimens
4.5.4. A/T E Film Coated Specimens
4.5.5. A/R type Film Coated Specimens
4.5.6. Effect of Post Annealing on Fatigue Crack Initiation Life

5. CONCLUSIONS 98
APPENDIX 199

BIBLIOGRAPHY 21 1

viii

LIST OF TABLES

Table Page

2-1: Cyclic Stress-Strain Characterictics of FCC Crystals ......................................... 101
2-2: Sputtering Yield Data for Various Metals (atoms/ions) ..................................... 102
3-1: Nominal Composition of OFHC Polycrysatlline Copper Used for Fabriaction

4—1:

42:

4—3:

4—5:

4-7:

4-9:

of Fatigue Specimen ............................................................................................ 51
Measured Concentrations of Ni and Ti in Deposited NiTi Thin Film for

Different Vertical Emission Angles ..................................................................... 103
Transformation Data, Compositions, and Fatigue Crack Initiation Life for

NiTi Thin Films on Flat Gauge Sectioned Copper Fatigue Specimens ............... 104

Fatigue Tests on Bare Copper Specimens and Amorhpous Film Coated

Specimens ........................................................................................................... 105
Transformation Temperatures and Composition of A/R Film ............................ 106
Fatigue Tests on AIR-type Specimens ................................................................ 106
Transformation Temperatures and Compositions of A/TE-type Films ............... 107
Fatigue Tests on A/TE-type Specimens ............................................................. 107
Transformation Temperatures and Compositions of A/SIM-type Films ............. 108
Fatigue Tests on A/SIM-type Specimens ............................................................ 109

ix

Figure

2-1:

2-2.

2-3

2-4:

2-5:

2-9:
2- 10:

LIST OF FIGURES
Page

Schematic representation of pesistent slip band (PSB) formation during

cyclic deformation. PSBs form at sites of original slip planes which are

softest. Extrusions and intrusions at the free surface cause the stress
concentration resulting in promoting additional slip and fatigue crack
nucleation .......................................................................................................... 110

Cyclic stress—strain curve for copper single crystal, which is obtained
by plotting saturation stress, 1:3, as a fuction of plastic strain amplitude, ypl.

1:3“ stands for the saturation stress. The volume percent of P838
increases with strain amplitude in region B ...................................................... 111

(a) Hysteresis stress-strain loop for the first 10 cycles of a copper crystal

subjected to a constant strain amplitude of Ay = 0.02 at 200 K and
(b) schematic of a cyclic strain hardening curve. 112

Dislocation structure in (a) matrix and (b) persistent slip bands of fee metals

at cyclic saturation. The steady state deformation occurs by a to and fro

glide of the same dislocations in the matrix,whereas a dynamic equibrium
between multiplication and annihilation exists in the wall structure of

persistent slip bands .......................................................................................... 113

Persistent slip band formation by non-promary slip model: (a) the ideal

infinite Tylor lattice in veins prior to saturation and (b) when cyclic

saturation is reached slip occurs on a secondary slip plane and thus extensive
annihilation of opposite edge dislocations can occur ....................................... 114

(a) the relaxed position at zero applied stress (b) the relaxed position at
the possible maximum stress before disintegration into dipole wall ................ 114

Stress-tempreature phase diagram illustrating region of shape memory
effect and transformational pseudoelasticity. The required stress to induce
martensite shows the linear variation which obeys the Clasius—Clapeyron

relationship ....................................................................................................... 1 15
Electical resistance vs.temperature curve for a Ni46,5Ti47,5CU5 alloy film

annealed at 923 K for one hour ......................................................................... 116
Equilibrium phase diagram for Ni-Ti alloys ..................................................... 117

The dependence of the transformation temperature (M3) on nickel
conentration in NiTi alloys ............................................................................... 118

2-11:

2-12:

2-13:

2-14:

2-15:

2-16:
2-17:

A schematic of surface relief morphologies showing the typical
(a) triangular self -accomodating morphology of martensite phase and
(b) saw-tooth self -accomodating morphology of R-phase in NiTi alloy .......... 119

Five types of Stress—Strain curves depending on the deformation
temperature relative to the transformation temperatures .................................. 120

Stress-Strain Behavior of NiTi thin ﬁlm as a Function of Deformation
Temperatures .................................................................................................... 121

Seven types of Stress-Strain curves showing dependence on the deformation
temperature relative to the transformation temperatures associated
with the R-phase and martensitic transformations ........................................... 122

Three types of yield stresses, the stresses for slip in martensite and austenite

phases, the stress for reverse transformationm (OR), and the elongation as a
function of test temperature for NiTi alloy; the yield stress represents for
twinning deformation below M3, for inducing martensite between Ms and

Md (0M), and for slip deformation above Md (0A) .......................................... 123
Grain size effect on psudoelasticity in NiTi alloy ............................................ 124
The total strain per cycle plotted against number of cylces. The longest life

was obtained at about 15 °C above Ms temperature for total strain of 0.01

which is a typical strain of PSB ........................................................................ 125
Effect of test temperature on fatigue life of Ti - 50.8 at% Ni alloy with

austenite finish temperature (Af) of 300 K .................................................... 126
Variation of sputtering yield with angle of incidence for 1 keV Ar+

incident on Ag, Ta, Ti, and Al .......................................................................... 127
Sputtering yields of (a) nickel and (b) titanium as a fuction of ion energy

and ion mass ..................................................................................................... 128
Schematic representation showing the superposition of physical

processes that established structure zone ......................................................... 129
Angular Distribution of Spittered Atoms for Ion Bombardment at Normal
Incidence ........................................................................................................... 130
Flow Chart of the Experimental Methods ........................................................ 131
Shape of surface modified specimens: (a) films were partially, (b) fully

covered in gauge section on flat area to observe surface morphology

and crystal structure, and (c) on shallow round notched area to

investigate fatigue life ....................................................................................... 132
Polycrystalline OFHC copper fatigue specimen microstructure after

annealing at 1173 K for 1 hour. The average grain size is 300 mm.

(Electrolyte was used 1 part of I-INO3 and 9 parts of water solution) .............. 133

xi

3-4:

3-6:

3-8:
3-9:
3—10:

3-11:

3-12:

3-13:

4—1:

42.

4—3:

Schematic Diagram of Electropolishing Apparatus: Electropolishing
was performed in a solution of 33% of nitric acid and 67% of methanol
at around 228 K ................................................................................................ 134

Schematic Diagram of Sputtering System: After sputter-cleaning by
bombarding to remove dirt and the oxide layer, films were deposited
on the specimens (side view) ............................................................................ 135

Steel ﬁxture used to position fatigue specimens in vauum chamber during
sputter-cleaning and ion beam deposition. This is designed to acilitate
sputter-cleaning and film deposition on four specimens in one batch .............. 136

Time-temperature plot for the crystallization annealing treatment of
as-sputtered NiTi thin ﬁlms .............................................................................. 137

Schematic Drawing of The Electrical Resistivity Measurement Apparatus... 138
Block Diagram of Servohydraulic System ..................................................... 139

A Schematic Diagram of the Cooling System for Low Temperature Fatigue

Tests and Details of an Aluminum Cooling Block. The specimen was
conductively cooled by aluminum blocks through coolant was made to

circulate ............................................................................................................ 140

Hysteresis stress-strain loop for constant strain cycling. Low cycling
fatigue test in present study was carried out at a constant plastic strain

amplitude (Asp) ................................................................................................ 141

Cyclic hardening curve from the annealed polycrystalline copper. grain

size was 300 um and the plastic strain amplitude was 0.002. The curve
was constructed by connecting the peak stresses for each cycle and was
terminated 1 % peak stress drop as a criterion for fatigue crack intiation ........ 142

Detail of cyclic hardening curve near fatigue crack initiation life (N1)
deﬁned by 1% stress-drop ................................................................................ 143

Schematic Diagram of Angular Distribution of Ion Beam Sputtering
Deposition: The incident angle to taget is 45° against the target plane normal.

The vertical emission angle (rp) is defined as an angle from the target plane

normal and the horizontal emission angle (0) is deﬁned as an divergent
angle on the same vertical emission angle ........................................................ 144

Measured concentrations of Ni and Ti in deposited NiTi thin ﬁlms for
various vertical emission angles ....................................................................... 145

EDS spectrum of crystallization annealed NiTi thin ﬁlm: The argon Kor
peak was detected from both as-sputtered and annealed ﬁlms ......................... 146

xii

4-5:

4-7:

4-8:

4—9:

4—10:

4-11:

4—12:

4—13:

4—14:

4-15:

Structure of N iTi thin films formed by ion sputter deposition on glass

subtrates. Fractured surface (bottom) and surface morphology (top) of

(a) the as-sputtered ﬁlm (amorphous phase) which shows more brittle

nature than (b) the crystallization annealed ﬁlm .............................................. 147

DSC trace obtained by heating the 3 mm-thick amorphous NiTi thin film

with a heating rate of 10 K/min. The exothermic peak appeared

approximately 773 K and was taken to be the crysatliization temperature

for the thin ﬁlm ................................................................................................. 148

X-ray diffraction sans of (a) as-sputtered and (b) crystallization annealed
NiTi thin ﬁlms ................................................................................... 149

Electrical Resistivity as a Function of Temprature for N iTi Amorphous
Thin Film .......................................................................................................... 150

SEM micrograph for the NiTi thin film annealed at 873 K for 10 minutes
showing grain size and grain morphology ........................................................ 151

Optical micrographs showing the surface relief morphologies in Ni43,3Ti51,2
thin ﬁlm (Af = 322 K and Ms = 286 K) deposited on (a) glass and (b) copper
substrates at 298 K followed by quenching into liquid nitrogen ...................... 152

Optical micrographs showing the surface relief morphologies in Ni43,3Ti51.2

thin film (Af = 322 K and M3 = 286 K) deposited on copper substrates at

298 K (a) cooled from above Af, (b) followed by quenching into ice water

and then warmed back to 298 K, and (c) followed by quenching into liquid
nitrogen and then warmed back to 298 K ......................................................... 153

Electrical resistivity versus temprature curves for various compositions of
NiTi thin ﬁlm alloys annealed at 873 K for 10 mimutes ................................. 154

Characteristic X-ray line scanning at the interface between Ni43,3Ti51,2 film
and substrate showing concentration proﬁles for (a) Cu before annealing,
(b) Cu after annealing, (c) Ni, and ((1) Ti after annealing ................................. 156

The hysteresis loops for the ﬁrst 10 cycles (top) and after reaching
saturation for polystalline copper specimen annealed for 1 hour at 1173 K.

(Asp = 0.2%) ..................................................................................................... 158

Cyclic hardening curve from the polycrystalline copper annealed at 1173 K
under plastic strain amplitudes of 0.002 and 0.003. The fatgiue life is shorter

and the saturation stress is higher at Asp = 0.3 % than at Asp = 0.2 % ........... 159

Cu-Ka X-ray diffraction scans for Ni50_3Ti49,7 ﬁlm coated specimen at

298 K (Af= 345 K and M8 = 273 K): showing stress-induced martensite
peaks on a monotonic loaded (a), crystallization annealed (b), and quenched
ﬁlms followed by warm back to 298 K (c), and as-sputtered films ((1) ........... 160

xiii

416:

417:

418:

419:

420:

421:

422:

423:

424:

425:

426:

427:

428:

X-ray diffraction results from Ni50.3Ti49_7 ﬁlm on fatigue specimen

(Af= 345 K and Ms = 273 K). Definite (002), (ll-1) and (111) martensite

peaks are present after fatigue which are believed to have formed in regions

of severe local straining above underlying PSBs ............................................. 161

DSC heating scans from Ni50.3Ti49.7 film deposited on copper foils.
The sharp endotherm at 273 K in the scanned trace indicates the melting
point of ice ........................................................................................................ 162

(a) X-ray diffraction scans of Ni 50,3Ti49,7 ﬁlm on glass substrate on cooling
from 348 K to 173 K and (b) X-ray diffraction spectra of Ni 50,3Ti49_7 film
on glass substrate on heating from 173 K to 348 K ................................. 163

Peak shif t resulted from a slight change in the height of exposed surface;

both (110)32 and (111)Cu peak were shifted to a higher angle at X-ray

diffraction scan from fatigued specimen. ................................................ 165
Optical micrograph of the Specime Fl; No prominent cracks in the film

coated area, while severe surface damages were observed in uncoated area

(21“..Aep = 70) ..................................................................................................... 166

(a) SEM micrographs of specimen Flat the boundary between modified (top)
and unmodified area (bottom): Significant number of cracks and severe
extensive PSB extrusions distributed in unmodiﬁed area, whereas no surface
damage was observed in the ﬁlm modiﬁed area at a low magnification and
(b) SEM micrographs of specimen Flat the boundary between modified (top)
and unmodiﬁed area (bottom) at a high magnification : The extrusion-
intrusion bands in a substrate grain were totally suppressed in tne coated

area (22Aep = 70) ............................................................................................. 167

SEM micrograph from specimen F1: the film delamination is observed at the
triple points of grain boundaries (arrow) ......................................................... 169

SEM micrograph from specimen F lzuncovered local matrix area showed
PSBs but film adhered ﬁrmly and the slip bands were supressed by the ﬁlm.. 170

Optical micrograph showing film cracks on the Specimen F2 after 1% of
peak stress drop (222111»:p = 77) ........................................................................... 171

SEM micrographs from specimen F2: (a) cracks developed perpendicular
to the loading direction and (b) few 'wrinkles' were observed at a high
magniﬁcation. .................................................................................................. 172

Optical micrograph showing film cracks on the Specimen F3 after 1% of
peak stress drop (2213ap = 67) ........................................................................... 173

SEM micrographs from specimen F3 showing a high density of cracks
developed (a), severe 'wrinkles' were observed (b), and a crack penetrated
from the film to the substrate (0). .................................................................... 174

Cu K-a X-ray diffraction scans of the specimen F3. TizNi peaks were
observed with copper and BZ austenite peaks .................................................. 176

xiv

429:

430:

431:

432:

433:
434:

435:

436:

437:

438:

439:

(a) A schematic illustration of the gauge area of specimen and applied

stress (on), (b) when parallel cracks developed prependicular to the loading
axis on the ﬁlm, the stress effect in x-direction on the ﬁlm between the

cracks decreases and the stress (ac) along y-direction, will intensify with

an aide of residual compressive stress on the film, and (c) the resultant

strain along y-direction goes more compressive side continuously, called

cyclic compressive creep' ................................................................................. 177

Optical micrograph showing no cracks in the ﬂat gauge section area on the
Specimen F4 after 1% of peak stress drop (22Aep = 90) ................................. 178

SEM micrograph showing a large crack in the opposite side of the ﬂat gauge
section area on the Specimen F4 after 1% of peak stress drop (22Aep = 90)... 179

Persistent slip band f romation in unmodified copper after 500 cycles

(22A2p = 2). The slip bands were formed in only favored crystallographic
grains (a) and were pareIIel with same spacing in an individual grain (b) ....... 180

Persistent slip band structure in bare copper after 500 cycles (22Aep = 2) ...... 181

SEM photographs of unmodified copper specimen at various cummulated

strains (22118;) = 4, 10, 20, 40, and after failure) at A8p = 0.2%. The spacing
of the slip lines was narrowed and the intencity of the extrusions increased... 182

SEM micrographs showing the presence of secondary slip trace near grain
boundary after fatigue crack initiation occurring (22Aep = 60) ....................... 183

SEM micrographs of the specimen Amor.-2 showing fatigue cracks
propagated in a straight character (zerp = 44) ............................................... 184

SEM micrographs from amorphous ﬁlm coated specimen after 1 % peak
stress drop (a) cracks were nucleated from amorphous ﬁlm and

(b) the typical brittle nature was observed. (22133p = 44) ................................. 185
SEM micrographs of the specimen A/SIMl showing a fatigue crack

propagated perpendicular to the loaidng axis (ZZAep = 100) ........................... 186
SEM micrographs of (a) near a large crack and (b) crack tip area of the

specimen A/SIMS, and (c) crack tip area of the specimen A/SIM 1 ................ 187

SEM micrographs of the specimen A/SIM5 showing a large crack with ﬁne
cracks independent of substrate grain orientation (22133;, = 90) ...................... 189

SEM micrograph of the specimen A/TEl—l showing a large wavy crack
with surface rumples (ZZAep = 120) ................................................................ 190

SEM micrographs of crack tip area of the specimen A/TEl-l showing
a 'zig-zag' of crack propagation (22Acp = 120) ..................................... 191

XV

A-l:
A-2:
A-3:
A-4:
A-S:
A-6:
A-7:
A-8:
A9
A-10:
A-l 1:
A-12:

SEM photographs of the specimen A/TE3-1 at various cummulated strains

(22Aep = 4, 10, 20, 40, 80, and after failure) ................................................... 192
SEM micrograph showed the crack tip area of the specimen A/R- 3

(221121,: 88) ..................................................................................................... 193
Effect of the Difference between Test Temperature (Td) and Af on Fatigue

Crack Intiation Life (N1) of A/SIM-type Film Coated Specimens ................... 194
Effect of the difference betweenTest Temperature (Td) and Af on Fatigue

Crack Initiation life of AME andA/SIM-type Film Coated Specimens ........... 195

Figure 447. Fatigue crack initiation life for various types of ﬁlm coated
specimens. (Test Temperature was 298 K) ..................................................... 196

Hysteresis loops representing the assumed cyclic stress srtain behavior of
the overall (inner, narrow loop), PSB (outer,wide loop) and presumed
superelastic film lying above the PSB. The critical stress for the formation

of stress-induced martensite is higher at Td>>Af than at Td~Af ..................... 197

Fatigue crack initiation life for A/TEB type ﬁlm coated specimens, one
kinds specimens fatigued until fatigue crack initiation occurred and the
other annealed at 673 K for 20 min. after 10,000 cycles, and resumed

the fatigue test until fatigue crack initiation ..................................................... 198
Electrical Resistivity as a Function of Temperature for F2 Film ..................... 199
Electrical Resistivity as a Function of Temperature for F3 Film ..................... 200
Electrical Resistivity as a Function of Temperature for F4 Film ..................... 201
Electrical Resistivity as a Function of Temperature for A/R type Film ........... 202

Electrical Resistivity as a Function of Temperature for A/T E1 type Film ....... 203
Electrical Resistivity as a Function of Temperature for Afl‘ E2 type Film ....... 204
Electrical Resistivity as a Function of Temperature for A/TE3 type Film ....... 205
Electrical Resistivity as a Function of Temperature for A/SIMl type Film ..... 206
Electrical Resistivity as a Function of Temperature for A/SIM2 type Film ..... 207
Electrical Resistivity as a Function of Temperature for A/SIM3 type Film ..... 208
Electrical Resistivity as a Function of Temperature for A/SIM4 type Film ..... 209
Electrical Resistivity as a Function of Temperature for A/SIMS type Film ..... 210

xvi

1. INTRODUCTION

Fatigue is the phenomenon of failure of materials under cyclic loading, and it is
strongly influenced by both microstructure and environment. It is known that failure can
occur under cyclic loading at stress levels much lower than those which cause fracture
under monotonic loading. The fatigue life of materials is determined by the sum of the
number of cycles required to initiate a crack and the number required to propagate the
cracks to a critical length for fracture. Although the propagation process is of crucial
importance in the engineering of many alloy systems, it is well known that the initiation
stage of fatigue may often dominate the total fatigue life of materials. Surface effects in
fatigue are particularly important, since in most cases the surface is the preferred site for
fatigue crack initiation. In face centered cubic (fcc) metals, it is widely accepted that
fatigue crack initiation is a result of cyclic strain localization in persistent slip bands
(PSBs), which behave as if they are softer than the matrix, causing all cyclic plastic strain
to become concentrated in the bands. As a consequence, local plastic strains associated
with P883 are very high (~ 1 %). These PSBs produce extrusions and intrusions at the
free surfaces, leading to stress concentrations at the notch roots and ﬁnally fatigue crack
nucleation. The important role of the surface on fatigue properties was dramatically
demonstrated by Basinski et al. in copper (1983). When a copper specimen was fatigued
to a large fraction of its fatigue life, and its surface then removed by electropolishing, the
specimen exhibited a fatigue life as long as that of a virgin specimen.

Several techniques have been developed to prolong fatigue crack initiation life by
modification of stress state, microstructure, and topography of the surface or near-surf ace

region. Since surface tensile stresses have a deleterious influence on fatigue life, it is a

1

2
generally useful practice to create residual compressive stresses at the surface by

mechanical operations such as surface rolling or shot peening (Almen and Black, 1963).
Also, surface hardening by nitriding and carburizing strengthens the surface by inhibiting
dislocation movement (Chen and Starke, 1976). Plated coatings using different materials
to give a surface strengthening effect, which resulted in lowering the stacking fault
energy at the surface, have also been studied as possible methods for improving fatigue
life of materials (Greenﬁeld and Purohit, 1980). Although plating often degrades fatigue
life, ion implantation and ion beam mixing techniques have been investigated as a means
of improving the fatigue life by shallow surface modification with controlled thickness,
composition and structure of the modified region (Morrison et al., 1989; Grummon et al.,
1988). Unfortunately, no surface modification is able to significantly modify the chief
driving force for crack initiation, namely the localization of bulk plastic strain, which
results from either grain boundary incompatibility, or from dislocation dynamics which
lead to the formation of some form of persistent slip band structure.

If the bulk plastic shear strains were perfectly homogeneous, a strengthened
surface layer might remain intact indefinitely during fatigue. Since, however, plastic
strain in the bulk rapidly concentrates into PSBs (with characteristic plastic shears of
~1%), a strong ﬁlm will be subjected to intense localized cyclic straining (Meyers and
Grummon, 1990) and can be expected to eventually rupture, and/or suffer slip band
penetration, forming a potent crack nucleus. Thus, the problem of crack initiation in the
presence of surface films is a problem of fatigue of the surface film itself, in localized
regions overlying subsurface PSBs.

In the present work, a new surface coating approach proposed by Grummon
(1990), which is significantly different from previous surface modiﬁcation techniques, is
evaluated with respect to its ability to improve fatigue crack initiation life. This
technique is based on unique fatigue properties associated with the thermoelastic

martensitic transformations that give rise to shape memory and superelasticity effects

3
(Melton and Mercier, 1978). In particular, the superelasticity effect results in unique

stress-strain behavior in which anelastic shear strains of several percent occur when a
martensite phase is induced by applied stress and reverted back to the original austenite
phase by removing that stress. Deformation by this mechanism is able to confer
exceptionally high fatigue life under strain controlled conditions. Therefore, surface
modification with superelastically deforming materials, which can exhibit large strains
without dislocation activity, may resist the concentrated strains which are the result of
bulk plastic strain localization during fatigue. Such a modified surface, it is hoped, will
inhibit persistent slip band (PSB) emergence to the free surface by blocking dislocation
egress, or by promoting slip reversibility through the shielding of slip-steps from
environmental interaction, and therefore delay the onset of the kind of severe, localized
surface damage that normally leads to fatigue crack initiation in many metals.
Essentially, the idea is to impart the high fatigue resistance of superelastic NiTi alloy to
other metals by applying the former material as a coating.

Among the many known shape memory alloys, NiTi alloy has high ductility and
extraordinarily long fatigue life under high strain amplitude fatigue. This is associated
with the superelasticity effect, which occurs not only in austenite-to-martensite
transformations, but also in austenite-to-R-phase transformations (Miyazaki and Otsuka,
1984; 1986). These materials, in the form of thin films, have been produced by dc
magnetron sputtering (Chang, 1993; Busch et al., 1990) or r.f. magnetron sputtering
(Ishda et al., 1993), and the optimum conditions for superelasticity and shape memory
effects have been investigated. Recently, NiTi alloys in the form of thin films produced
by dc. magnetron sputtering have been shown to posses full transformational
superelasticity (Hou, 1995). The stress for inducing martensite obeyed the Clausius-

Clapeyron relationl, as with a bulk NiT i alloy, and the fully superelastic phenomenon

 

1. This relation explains the realtionships between variation of temperature and stress which are necessary
for the maintenance of equilibrium between two phases. For the formation of thermoelastic martensite,
there is an equivalence between temperature and stress: a decrease in temperature is equivalent to an

4
was also observed. The superelasticity effect may allow the ﬁlm to be deformed up to a

plastic strain of 5 ~ 10 % without accumulating severe slip damage that enhances the
nucleation of fatigue cracks on the surface.

The first objective of the present study was to fabricate NiTi thin films by ion
beam sputter deposition which could be shown to undergo the thermoelastic martensitic
transformations, and to understand transformation behavior of NiTi thin films. The
second objective was to study the effect on fatigue crack initiation of thermoelastic NiTi
coatings.

This thesis is organized as follows: in Chapter 2, a detailed review of the literature
on fatigue crack initiation in FCC metals is provided, and fatigue life studies on materials
with several conventional surface modification types are introduced. In addition, general
concepts of transformational superelasticity and mechanical (especially fatigue)
properties of near equiatomic NiTi alloys, with respect to their transformational behavior
are reviewed, and recent research on NiTi thin films are discussed in this chapter. An
overall description of NiTi thin film fabrication by ion beam sputtering, characterization
of the deposited ﬁlms, low cycle fatigue tests, and other experimental details are given in
Chapter 3. In Chapter 4, results of the study are presented and discussed. The first part
of this chapter presents characteristics of the thin films such as their microstructure,
crystallization behavior, and their transformational behavior. The second part describes
the stress strain behavior of the thin films with respect to their transformation
temperatures, and provides comparisons of fatigue crack initiation life influenced by NiTi
thin surface film coatings. Finally, the thesis is summarized in Chapter 5, where

conclusions of the present study are drawn and suggestions for future work are provided.

 

increase in stress, both stabilizing martensite. According to this relation, the stress needed to produce
stress-induced martensite increases linearly with temperature.

2. REVIEW OF THE LITERATURE

Introduction

The present study was undertaken as part of a project entitled "Surface
Superelastic Microalloying for Improved Fatigue Performance in Ni, Ti, Fe, and Cu
Based Alloys” supported by National Science Foundation under grant # MSS-8821755.
The main objective of this work was investigation of the thermoelastic NiTi surface
coating on fatigue properties of polycrystalline copper. In ductile metals, fatigue crack
initiation at moderate strain amplitudes involves two principal processes: the localization
of plastic strain in the interior of the material, in the form of persistent slip bands (P888),
and the progressive development of stress concentrating geometries at the surface where
these slip bands emerge. The irreversibility of shear displacements along the slip bands
in the near-surface region results in a roughening of the surface. This parasurface region
can strongly influence the homogeneity and reversibility of cyclic slip, by mediating
interactions with the external environment, controlling the egress of mobile dislocations,
and altering the operation of nearby dislocation sources. Modification of the chemistry,
stress state, or atomic structure of the surface region has therefore been a logical approach
to the improvement of fatigue performance in metals.

In the following chapter, mechanisms of fatigue in ductile metals, including cyclic
plasticity, strain localization, and crack initiation, will be treated briefly in section 2.1. In
section 2.2, different approaches to improve fatigue properties by surface modification
will be discussed. In section 2.3 and 2.4, general features of thermoelastic martensitic
transformation, including superelasticity, shape memory effect, and stress-induced

martensite formation, will be discussed, and the deformation behavior and fatigue

5

6
characteristics of superelastic NiTi alloys will be reviewed. In section 2.5, fundamental

aspects of ion beam sputter deposition of NiTi thin films, along with the effect of various
process parameters that have an effect on the film, will be outlined. Basic studies related
to microstructure and thermoelastic transformation characteristics in sputter deposited
NiTi film, which were carried out by Chang (1993), will be discussed in this section.
Finally, the inﬂuence of the crystalline nature of the deposited film in relation to its

desired properties will also be discussed in this review.

2.1. Fatigue Behavior in FCC Metals

In face centered cubic (fcc) metals, it is widely accepted that fatigue crack
initiation is a result of cyclic strain localization in PSBs, which behave as they are much
softer than the matrix (Neumann, 1983). These PSBs produce extrusions and intrusions
at the free surfaces where subsequent stress concentrations occur and promote additional
slip, and finally, fatigue crack nucleation; as shown schematically in Figure 2-1. The
population of fatigue cracks increased linearly with the number of cycles and the applied
strain amplitude after the crystals reached saturation (B. -T. Ma and C. Laird, 1989),
because the volume fraction of PSBs is proportional to applied strain and with
accumulating cycles. Details of the cyclic stress-strain response, PSB formation, and

fatigue crack initiation in FCC metals will be described in the following sections.

2.1.1. Cyclic Stress-Strain Response in FCC Metals

Mechanical testing methods for studying cyclic stress strain response may utilize
one of essentially three kinds of control modes: (1) constant stress amplitude mode
(Am), (2) constant total strain mode (AeT), and (3) constant plastic strain amplitude
mode (Asp). In these approaches, the fatigue life of a component is defined as the total
number of cycles needed to induce fatigue damage and to initiate a dominant fatigue flaw

which is propagated to final failure.

7
Among these control modes, constant stress amplitude-fatigue life approach was

ﬁrst introduced in the 18603. Out of this work evolved the concept of an 'endurance
limit', which characterizes the applied stress amplitude below which the material is
expected to have infinite fatigue life. Although this method has found widespread use in
fatigue analysis, it does not provide a realistic account of fatigue life when considerable
plastic def orrnation occurs during cyclic loading.

Strain controlled fatigue testing is commonly used as a basis for structural design
in components where fatigue crack initiation ahead of stress concentration is of primary
concern. Under typical cyclic loading conditions involving low values of imposed strain,
the plastic strain (8p) is only a fraction of the total strain amplitude (er) and it decreases
with cyclic work hardening. Thus, fatigue tests conducted with total strain amplitude
control do not provide a clear description of the progression of cyclic deformation.
Instead, the most common method of performing strain-controlled tests involves constant
plastic strain amplitude cycling.

Annealed fcc metals generally exhibit a rapid increase in yield stress with
increasing number of cycles at a constant plastic strain amplitude during the initial period
of cyclic deformation. This process is known as cyclic strain hardening. With further
increases in the number of cycles, the hardening rate progressively diminishes and the
variation of stress with plastic strain reaches a steady-state condition where the stress
amplitude reaches a 'saturation' value. Once the saturation state occurs, there are no
further changes in the stress-strain hysteresis loop shape and area with increasing number
of cycles, until fatigue cracks form.

A plot can be constructed from the peak shear stress (1:3) at saturation as a
function of the plastic strain amplitude (ypl) used in the test, which is called the cyclic
stress-strain curve (CSS curve). A typical CSS curve for a copper single crystal is
illustrated in Figure 2-2 (Mughrabi, 1978). Based on the magnitude of the plastic strain

amplitude (Ypl), the plot has been divided into three distinct regions. In regions A and C

8
which represent very low and high plastic strain amplitudes respectively, the saturation

stress increases with increasing plastic strain amplitude, but in the intermediate region, B,
the CSS curve shows nearly constant saturation stress independent of plastic strain
amplitude. The typical shear stress at the plateau, observed for a copper single crystals at
room temperature, is approximately 28-30 MPa, and the plastic strain amplitudes
associated with the plateau ranges from 6 x 10 ‘5 to 7.7 x 103. 1

At a very low plastic strain amplitude, range A in Figure 2.2., 1:3 increases with
increasing ypl, and the deformation is reasonably homogeneous. Under these conditions,
cyclic deformation is essentially reversible and thus formation of PSBs and strain
localization does not occur, thereby extending the fatigue life of any given material to
inﬁnity. Thus, the upper limit of this region will correspond to the fatigue limit. With
increasing plastic strain amplitude, range B, the shear stress amplitude will reach a steady
state condition, where the saturation stress (13*) is independent of the plastic strain
amplitude. This region is related to inhomogeneous deformation localized in PSBs. In
this region, the volume fraction of P883 increases linearly from 0 % to 100 % as the
plastic strain amplitude is raised, until region C appears. Further increasees in the plastic
strain amplitude (i.e. high strain fatigue) results in an increased saturation stress, range C,
and homogeneous deformation. As a manifestation of this, new dislocation patterns will
be formed and materials will exhibit a different failure mode. Cyclic stress-strain data for
various FCC metals and alloys, fatigued at different temperatures, are given in Table 2-1.
Saturation stresses depend on the testing temperature. The saturation stress for copper
single crystals fatigued at 4.2 K was almost twice as high as one fatigued at 295 K.

However, the plastic strain amplitude range (ypl) are almost independent of the materials.

 

1. The ratio of saturation stress to shear modulus is remarkably constant for copper, silver, nickel, and or-
iron (Mughrabi et al., 1979).

A'D

.it " I

9
2.1.2. Dislocation Structures in F atigued FCC Metals

Annealed FCC metals develop dislocation structures over the range of plastic
strain amplitude and number of cycles. At a low number of cycles, the dislocation
structure of fatigued pure metals consists of dense bundles (veins) of primary edge
dislocations in multipole configurations separated by dislocation depleted regions
(channels). With increasing the number of cycles, at saturation stress state, the
dislocation structure evolves continuously. The matrix structure, which was created
during cyclic hardening, breaks down and is converted into the 'wall' or 'ladder' structure
of the PSB's which are softer, and in which the strain concentrates subsequently. The
P883 are described as thin lamellae (~ lum) which experience high slip activity, lying
parallel to the primary slip plane, embedded in a matrix with a significantly different
dislocation arrangement (Mughrabi, 1978). The PSB is composed of regularly spaced
edge dislocation walls of high dislocation density, while the matrix veins are irregularly
arranged. The last stage in the evolution of dislocation structures in fatigue is that of cell
formation. Dislocations collect into a cellular arrangement upon further increases in the
number of cycles. However, the cell formation may or may not occur and is apparently
not necessary for fracture (Kuhlmann-Wilsdorf and Laird, 1977). In the following
section, evolution of the dislocation structure during rapid hardening and formation of

persistent slip bands (PSBs) in FCC metals will be discussed in detail.

2.1.2.1. Evolution of the Dislocation structure during Cyclic Hardening A
typical set of hysteresis loops obtained from the first 10 cycles of fatigue of a copper
specimen, and a schematic illustration of a cyclic hardening curve for fee metals are
shown in Figure 2-3. A fatigue hardening curve gives the peak stress as a function of the
cumulative strain (22Aep) or the number of cycles. The rapid hardening range is
comparatively short. One description of the dislocation process which may lead to rapid

hardening is as follows: In the early stage of cyclic hardening, initial plastic straining

 

 

a?

10
occurs on some favored slip planes. As a result of non-conservative dislocation motion,

debris collects on this slip plane, and it hardens. Slip then becomes favored on another
plane which, in turn, accumulates dislocation debris and hardens. The glide planes,
where slips are in both negative and positive sense, are mostly separated and the glide
dislocations trap one another to form dipoles or multipoles with increase in number of
cycles. In this process, screw dislocations are mutually annihilated through cross slip
while edge dislocations form very stable dipole groupings with preferential orientation of
the dipoles perpendicular to the primary slip direction. The process continues until the
entire volume of the material has undergone some cyclic slip, by which time the material
has hardened to a level approaching the saturation stress. As a result, the dislocation
substructure during cyclic hardening of fatigue hardened pure metals consists of parallel
veins of metastable edge dislocation dipoles (multipoles) and channels, which are almost
dislocation free areas, as shown schematically in Figure 2-4(a). With increasing
cumulative plastic strain, deformation begins with the bowing out of edge dislocation
from the veins. When this bowing segment encounters an adjacent vein, annihilation of
edge segments takes place. Two screw segments of opposite sign are then liberated and
my shuttle back and forth with screws of opposite sign. Thus, most dislocations between

the veins have predominantly screw character.

2.1.2.2. Formation of Persistent Slip Bands (PSBs) The transition
from rapid hardening to the saturation state is associated with PSB formation.
Dislocation models regarding PSB formation were developed mainly by two groups: (1)
Essmann, Gosele, and Mughrabi and (2) Kuhlmann-Wilsdorf and Laird. Essmann,
Gosele, and Mughrabi (1981) proposed a dislocation model based on experimental
observations of point defect generation by the annihilation of edge dislocations and the
majority of these point defects have vacancy character. According to this model, the

saturation state is characterized by the existence of a dynamic equilibrium between

 

—\.

11
dislocation multiplication and annihilation. Dislocations can multiply when edge

segments bow out and cross the channels, increasing both edge and screw dislocation
densities, or when screw dislocations glide along channels, laying down edges along the
walls, which is shown in Figure 2—4(b). This is balanced by mutual annihilation of
dislocations of opposite signs approaching on glide planes less separated by than a critical
distance. Screw dislocations annihilate by cross slip if the distance between glide planes
is less than ~50 nm, but the corresponding distance for edge dislocations on the glide
plane is less than ~l.6 nm apart in the wall of the PSBs (Mughrabi, 1979). Elimination of
edge dipoles is envisaged as giving rise to vacancies. Vacancy production by dipole
collapse is then balanced by vacancy capture by dislocations passing within critical
distance. The sequence of microscopic slip processes, accompanied by dipole
annihilation, has the effect of depositing edge dislocations at opposite PSB-matrix
interfaces such as that their extra planes are directed into the PSB.

Kuhlmann-Wilsdorf and Laird (1980) proposed a mechanism for continuous
reﬁnement and densification of veins, which were formed during rapid cyclic hardening,
with increasing stress amplitude. The veins became hard shelled because dipoles at their
surfaces were maintained at a small spacing by the arrival and departure of dislocations
gliding under stress. The softer interiors contained wider dipoles, which were protected
from moving dislocations. The dislocations were arranged in such a way that there were
an equal number of parallel edge dislocations of positive and negative orientation (inﬁnite
Taylor lattice) in the veins before the onset of PSB formation. As the stress amplitude
rises toward the saturation level, secondary glide occurs, and produces relative shifts in
the lattice and thus extensive annihilation of edge dislocations, leading to collapse of the
veins to form a PSB. The formation of the dipolar walls in the P888 begins with a local
breakdown of a vein as shown in Figure 2—5. They described a series of events in which a
flood of glide dislocations move through the matrix channels annihilating their screw

components and trapping their edge components into the wall segments. The wall

12
segments blocked each other's progress and became stacked up in groups which in turn

stopped more walls, resulting in the formation of persistent slip band ladder structures.
Neumann (1986) calculated a theoretical stable configuration of vein and
considered the finite Taylor lattices2 as starting configurations. He concluded that the
relaxed and tetragonal Taylor lattices were stable and the changes in polarization between
positive and negative dislocations were directly connected to the plastic strain. Figure 2-
6(a) shows such a relaxed configuration of a diamond-shaped section of a Taylor lattice
in which all the dislocations are at their equilibrium positions at zero applied stress.
According to his calculations, when the maximum stress appropriate for a given dipole
configuration was exceeded, the veins disintegrated into dipole walls of the type observed

in PSBs, which is shown in Figure 2—6(b).

2.1.3. Fatigue Crack Initiation in FCC metals

Since the initiation of cracks in fatigued metals is closely connected with the
existence of PSBs, many detailed observations on PSBs profiles and crack initiation
behavior have been undertaken. Several methods of obtaining accurate cross sections
have been introduced to study the fatigue crack initiation in copper single crystals.
Hunsche and Neumann (1986) sectioned fatigue crystals without distortion of the edge
using a micro milling technique. Basinski and Basinski (1984) used replication and
sectioning technique, and Ma and Laird (1989) used the sharp corner polishing technique.
Each of these techniques provides high resolution observations of PSB morphology and
initial crack formation. Hunsche and Neumann recorded a number of cracks at various

depths across a PSB profile and found that the protrusion3 geometry was important in

 

2. This lattice is composed of a ﬁnite regular planar array of positive and negative edge dislocations and
the dislocations in the conﬁgurations form a face-centered tetragonal lattice. The tetragonality of the lattice
ranges from 1.8 for small (36 dislocations) to 2.2 for large (324 dislocations) sections (Neumann, 1983).

3. A protrusion is a surface uplift (a large extrusion), many micrometers in height (about 7 um), where a
macro-PSB, tens of micrometers wide (about 30 um) and containing tens of matrix and/or PSB lamellae,
emerges at the free surface (Ma and Laird, 1989). It is connected eith excess vacancy production within the
PSB.

13
determining where the most damaging cracks were formed. The height of a protrusion

was linearly proportional to the width of the PSB. A crack could occur anywhere in a
PSB, but PSB-matrix interfaces were the preferred sites for continued crack growth.
Recently Ma and Laird (1989) have made a systematic study of crack growth for fatigue
of copper crystals using the above technique. They found that PSBs consist of extrusions
and intrusions, that all cracks originated at intrusions, and that the preferred crack sites
were isolated intrusions at PSB-matrix interfaces.

Unlike a single crystal, polycrystals contain numerous boundaries and grains with
relatively different crystallographic orientation. Thus, cracks can nucleate in a particular
grain, while adjacent grains appear to suffer no damage. This may be due to the
difference in crystal orientations and the constraint effects of surrounding grains. Also,
the initiation of cracks in polycrystalline copper revealed a strong dependence on the
strain amplitude used in the test (Figueroa and Laird, 1983). At a high level of strain (Asp
> 0.2%), the majority of grains were plastically deformed homogeneously, likewise
multiple slip conditions developed in most grains. Intergranular crack nuclei were
formed at the intersection of the slip bands with grain boundaries, but were not
significantly developed at twin boundaries. Similar observations reported that grain
boundaries were the preferred site for both crack initiation and propagation in fatigued
copper polycrystals under high plastic strain amplitude conditions (Kim and Laird, 1978).

At a low level of strain (Asp < 0.02 %), the distribution of PSBs varied from grain
to grain and the density of the bands was non-uniform across the surface of individual
grains. At this level of cyclic plastic strain, PSBs were formed in the most favorably
oriented grains and the plastic strain was localized in them. Nevertheless, even for cracks
nucleated in PSBs, the preferred crack initiation sites were grain boundaries. This crack
nucleation tendency has been explained by the PSB impingement mechanism (Mughrabi

et al., 1982). In the model, cracks were expected to nucleate along the grain boundaries

14
due to the stress concentration between P883 and grain boundaries caused by piling-up of

edge dislocations (PSB-GB crack).

Neumann and Tonnessen (1987) reported that the relative frequency of cracks
nucleating at twin boundaries increased with decreasing plastic strain amplitude for
polycrystalline copper, and that intergranular crack initiation was insignificant. They
rationalized the observation of fatigue crack formation at every other twin boundary
utilizing a mechanism which relied on the elastic anisotropy of materials containing
twins. For example, the shear modulus of copper varies with direction by a factor of 3.2.
In order to ensure strain compatibility at twin boundaries in elastically anisotropic
materials, internal stresses must be generated in the vicinity of the boundaries. They
considered a stack of lamella twins, where the crystallographic orientation of lamellae
changes back and forth from that of the matrix to the twin as one traverses across the
boundaries. With the change in orientation, there is also a change in the direction of the
internal stresses. The internal stresses act in concert with the resolved stresses arising
from the applied loads at every twin boundary. When the resultant stress is of
sufficiently high magnitude, a PSB is formed near the twin—matrix interface and
eventually develops into a crack. However, the PSB-GB type crack was recently
encountered even from low to intermediate amplitudes at room temperature (Liu et al.,
1992).

Studies of dislocation structures in polycrystalline copper (Ramussen and
Pedersen, 1980; Winter et al., 1981) revealed that PSBs crossed subgrain (twin)
boundaries only with very local modifications to the wall structure, but large angle grain
boundaries were usually impenetrable. The sudden disappearance of a PSB inside the a
polycrystal must be accompanied by a rather large compatibility stress, but these seemed
to be accommodated at a grain boundary. Those researchers explained that the
incompatibilty of deformation across grain boundaries would be elastically

accommodated in bulk polycrystals. Polak and his colleagues (1984; 1992) concluded

so
$-

\

51,

t1.

15
that most grains exhibited very similar spatial dislocation structures to that of a single

crystal. In their successive studies, they observed regular cell and wall structures within
the high amplitude regions and vein structures with a high density of P883 at the low
amplitude region. However, they proposed that PSBs could be interlinked, which means
that they were nucleated within one grain and propagated into the suitable slip system of
a favorably oriented neighboring grain.

In general, fatigue cracks are nucleated only at PSB boundaries in a single crystal,
whereas cracks in polycrystalline materials are initiated either at PSB boundaries
(transgranular crack) or at the boundary intersecting P888 and grain boundaries (PSB-GB
crack). In polycrystals, however, PSBs formed within interior grains produce slip that is
confined to the individual grain. The large transfer of material causing the creation of a
rough topography is possible at surface grains, but not at the interior ones, because of
constraints from the surrounding matrix. Therefore, the role of the free surface in crack
initiation is very important for both single crystals and polycrystals with regard to their

fatigue properties.

2.1.4. Secondary Cyclic Hardening in Fatigued FCC Metals

Secondary cyclic hardening was reported for strain amplitudes above 7 x 10'4
(Polak et al., 1992) and was related to a considerable fraction of secondary slip within a
majority of grains in polycrystalline copper.4 The source for the secondary cyclic
hardening of copper specimen was attributed to: (1) build up of the dislocation forest
density due to the action of a secondary slip (multiple) system within individual grains, or
(2) an activation of slip systems with a lower Schmid factor with continuing strain (Polak

and Klesnil, 1984).

 

4. According to the study by Wang and Mughrabi (1984), the secondary cyclic hardening was prominent in
the range of around 10'3 < Aep< 2 x 103 and became vague with increasing As The distinction between
primary and secondary hardening was disappeared when Asp was above 5 x 10' .

16
According to the dislocation study relating the secondary cyclic hardening in

copper by Wang and Mughrabi (1984), the gradual formation of a misoriented dislocation
cell at the PSB-matrix interface, and their lateral expansion, resulted in a local hardening
in single crystals. The secondary cyclic hardening in polycrystals was observed at lower
plastic strain amplitudes and slightly earlier than in single crystals. It was claimed that
this difference resulted from the variety of grain orientations and the constraint stresses
enforced multiple slip at lower amplitudes. The relative amount of secondary hardening
with respect to the saturation stress increased with an increase in the plastic strain
amplitude.

In another study of polycrystalline copper by Mendez et al. (1988), secondary
hardening was observed to occur more easily in vacuum than in air, when fatigue tests
were conducted under the same constant plastic strain amplitude (Asp = 6 x 104). The
maximum peak stress in vacuum was slightly higher, and the fatigue life was longer than
those in air, whereas the saturation stresses were same whether the tests were carried out
in vacuum or in air. This occurrence of secondary cyclic hardening in vacuum was
associated with a prolonged fatigue life. Thus, it is deduced that the saturation stress

depends only on the constant plastic strain amplitude.

2.2. Effect of Surface Modification on Fatigue Life in FCC metals

The surface of the material is exposed directly to the environment, and
mechanical properties, chemical stabilities, as well as electrical properties are affected by
the surface physics and chemistry. There is considerable evidence to indicate that the
mechanical properties and deformation behavior of crystals can be altered by modifying
the surface conditions. The influence of solid surface films on the mechanical behavior
of crystal substrates is likely to depend on the elastic properties of both the film and the
substrate components, the degree of atomic misfit at the interface, and the residual

stresses in the film. Various types of surface modification techniques have been studied

17
in the context of fatigue, including formation of thin oxide layers, deposition of metal

films, ion beam surface modification, and shot peening. Generally, these surface
treatments result in surface-strengthening, i.e., increase in the critical resolved shear
stress. Patterson and Greenfield (1971) investigated surface alloying effects on the
plastic deformation behavior in copper single crystals and explained the nature of the
surface-strengthening by nickel and gold thin films. They concluded that both nickel and
gold films on copper substrates showed a remarkable surface strengthening effect, which
was attributed to solid solution hardening, misfit dislocation networks, and image forces.

The surface condition of the material is a very important factor in fatigue since the
surface is often the preferred site for fatigue crack initiation. The important role of the
surface in fatigue was demonstrated by the observation that when a specimen was
fatigued for a large fraction of its life, and its surface then removed by electropolishing,
the specimen showed a remaining fatigue life as long as that of a virgin specimen.
Fatigue life is also generally higher in a vacuum. In general, the increase in fatigue crack
initiation resistance by surface modification is due to homogenization of cyclic plastic
deformation, which leads to lower surface damage accumulation rates, or to effective
suppression of slip band penetration to the free surface. Thus, a variety of surface
modifications have been developed, such as surface rolling or shot peening, to supply a
compressive stress on the surface or surface hardening.

Chen and Starke (1976) investigated the effect of stacking fault energy (SFE) of
ion plated surface layers on fatigue life. In experiments on copper substrates, nickel was
used as the high SFE coating, silver as the low SFE coating. Copper itself was used to
investigate the effect of interface imperfections. They concluded that fatigue crack
initiation was sensitive to the topography of the slip bands on the surface, which varied
depending on differences in the stacking fault energies of the surface layer and bulk
material. Nickel layers produced coarser slip steps and reduced fatigue life, whereas the

opposite effect was observed in silver films. Copper self-coated specimen did not show

 

l
{I

(A .1

18
any appreciable change as compared to the uncoated ones. Lowering SFE at the surface,

it was argued, reduced the incidence of cross slip and resulted in more homogeneous and
reversible slip, causing a slower accumulation of surface damage.

Greenfield and Purohit (1980) studied the influence of both discrete and diffused
surface layers on the initiation and movement of dislocations, in fatigue tests performed
on copper specimens with nickel, zinc, silver, gold, and aluminum coatings. In this
study, they considered the relative shear modulus of the surface with respect to the
substrate, along with the stress and strain for yielding at the surface. The regions of
initiation of plastic deformation depended on the elastic stiffness of the surface with
respect to the substrates and the magnitudes of the stress and strain. This analysis
showed that the increased yield strength of the surface layer is more important in
enhancing fatigue life than the modulus effect. If the surface film is harder than the
substrate, then the substrate deforms plastically while the surface layer behaves still
elastically, then the fatigue damage accumulation rate will slow during cyclic
deformation.

Grummon et al. (1988) studied the effect of surface microalloying by ion beam
modiﬁcation on fatigue crack initiation, using direct ion implantation by krypton or ion
beam mixing of aluminum, in polycrystalline nickel. This study pointed out that a
surface ﬁlm which is strong and more elastically compliant than the substrate, might be
expected to both block the egress dislocations and inhibit penetration of slip bands, and
also resist plastic deformation itself, and therefore suffer a low rate of fatigue damage
accumulation. Grummon et al. attributed the increased elastic limit of the material in the
modified surface layer to solid solution hardening, the high defect population induced by
radiation, and extremely ﬁne grain size of the surface modiﬁed area.

It has been shown that direct ion implantation and ion beam mixing techniques
can be effectively used for improving fatigue life of materials, by shallow surface

modification. The approach has the capability to control thickness, composition, and

{if
~ I

lift

19
structure of the modified volume of the material. In studies on fatigue behavior of boron

and nitrogen ion implanted into nickel polycrystals (Hohmuth et al., 1985) and copper
polycrystals (Mendez et al., 1987), it was found that increasing the implantation dose
reduced the density of slip bands at the surface, and increased the fatigue lifetime.
Additional experiments, however, demonstrated that the hardening of the surface layer by
ion implantation of interstitial species did not improve fatigue resistance. Favorable
effects did result, however, from implanting with the low stacking fault energy
substitutional species, which promoted better slip reversibility during cyclic loading.

Carbon ion implantation in a nickel chromium alloy improved fatigue life by solid
solution strengthening and by the development of high compressive stress due to the
presence of the interstitial carbon. However, carbon implantation in copper and stainless
steel shortened the fatigue life because of slip line coarsening, which was observed on the
surface by SEM, that causes stress concentration and hence decreased fatigue life. But
they did not mention any plausible reasons such as a residual stress effect on the surface
or stacking fault energy change related to the surface modification (Sleeswyk et al.,
1980).

Fatigue properties were improved in polycrystalline nickel with microalloyed Ni-
Al surface layer created using ion beam mixing and direct aluminum ion implantation
(Grummon et al., 1989). Self-implanted nickel demonstrated that the microcracks were
formed along P883 and that fatigue features were not altered. However, microalloying
the surface with aluminum increased surface hardness and inhibited the formation of
surface PSB features. The crack initiation mode was changed from transgranular to
intergranular, and the change was usually accompanied by suppressing PSB penetration
to the surface. Also, surface residual compressive stress, which may be caused by direct
implantation with larger atom species, produced longer fatigue crack initiation life. The
increased fatigue life may result from lowering of stacking fault energy on the surface,

causing finer and planar slip, resulting in a delay of crack initiation.

20
In summary, surface effects in fatigued materials are very important since, in most

cases, fatigue failure originates from fatigue cracks at the surface. These cracks initiate at
the sites of surface roughness which has developed a notch-peak geometry as a result of
cyclic slip irreversibility. Particularly for fcc case strain localization in PSBs can exceed
the surrounding matrix strain value by a factor of 100. Thus, various approaches to
surface modifications have been carried out for the purpose of inhibiting or delaying the
fatigue crack initiation on their surface regions. According to the finite element study by
Meyers and Grummon (1990), however, the surface layer above underlying active PSBs
should possess an elastic limit on the order of 1%-1.5% of shear modulus in order to
withstand localized strain during cyclic deformation. Therefore, it is believed that PSB
shear displacement will force the surface ﬁlm in its plastic range so that the improvement
of fatigue life by such conventional surface films, which were used in various surface

modiﬁcation techniques in this section, are limited.

2.3. Transformational Superelasticity

Considering plastic strain localization during fatigue, a promising surface coating
system to withstand such a large strain is superelastic alloys. Superelasticity results in
unique stress-strain behavior, which exhibits large anelastic strain associated with a
reversible stress-induced martensitic transformation, and thus superelastic alloys may
show extraordinarily long fatigue life under high strain amplitude loading. Especially,
fatigue life for NiTi alloy can be extrapolated to over 108 cycles at strain amplitudes
approaching 1% (Melton and Mercier, 1978; 1979), which is the strain associated with
PSB shear in fee metals. Therefore, surface coatings of NiTi alloys, which are able to
accommodate large strains without accumulating severe slip damage, are suggested to be
an effective way of delaying fatigue crack nucleation in metals.

In the following section, general features of thermoelastic martensitic

transformation, including superelasticity, the shape memory effect, and stress-induced

x.”

21
martensite, will be discussed. Following this, phase transformation behavior in NiTi

alloys and factors which affect these transformations, will be presented in detail.

Physical and mechanical properties of NiTi alloys will also be reviewed.

2.3.1. General Aspect of Martensitic Transformation and Superelasticity

Martensitic transformations are diffusionless phase transformations in which
atoms are cooperatively rearranged into a new and more stable crystal structure without
changing the composition of the matrix. Since no long-range atomic movement is
involved, the progress of the martensitic transformation does not depend on time, but on
temperature. On cooling, the high temperature phase (austenite) starts to transform to the
low temperature phase (martensite) at a temperature designated as Ms, and finishes at a
lower temperature, Mf. On heating, martensite starts to transform to austenite at a
temperature designated as A3 and finishes at a higher temperature, Af. In the temperature
range between M8 and NI“ on cooling (and between A3 and Af on heating), austenite and
martensite phases coexist. The primary driving force for the transformation is the free
energy difference between the two phases. However, a strain energy is accumulated in
the lattice during the transformation which resists further progress of the transformation
unless a higher driving force is supplied by a subsequent decrease in temperature.

Martensitic transformations can be divided into two categories, thermoelastic and
non-thermoelastic transformations. For materials which follow the non-thermoelastic
transformation the martensite-parent interfaces are relatively immobile. Therefore, when
the martensite is formed once, the interface does not return back to the parent phase after
heating and the austenite phase must be renucleated. However, in thermoelastic
transformation the interface of the martensite-parent undergoes reversible movement
during heating or cooling due to the balance between thermal and elastic effects. Thus,
the occurrence of a thermoelastic martensitic transformation has certain requirements: (I)

ordered cubic parent phase, (2) small volume change during martensitic transformation,

 

It

22
and (3) most of its deformation occurs by intervariant boundary motion before slip. Such

thermoelastic martensitic transformations give rise to the shape memory effect. In NiTi
alloys, martensite can be deformed up to 5 to 10%, by reorientation of martensite variants
without activating dislocation slip. The plastic strain resulting from the rearrangement of
martensite variants at low temperature is recoverable on heating above Af, because each
martensite variant reverts to austenite in the original orientation.

Shape memory alloys can have another unique mechanical property called the
superelastic effect (SE). Phenomenologically, within a speciﬁc temperature range (As <
T < Md5) in the superelastic effect, material suffering large anelastic strain on loading can
revert to its original shape upon unloading. Figure 2-7 describes the relationship of these
effects with temperature and stress. Superelastic effect requires a reversible stress-
induced martensite transformation. In other words, the application of stress has a similar
effect on the transformation as a decrease in temperature (Swalin, 1972). The stress
needed to produce stress-induced martensite (SIM) increases linearly with temperature.
The relationship between stress and the Ms can be speciﬁed as follow. When the state of
a material which has been elongated or shortened, is considered, the conventional
enthalpy (H) and free energy (G) have to include the mechanical work (i.e.

thermodynamical effect of stress). Therefore,

H=U+PV-Fl (2.1)

G: U+PV-TS-Fl

where F is the force applied to the specimen, 1 is specimen's length, and the other

symbols have the usual meanings. If the free energies of the austenite and martensite

 

5. The temperature at which Ms equals the critical stress for slip in the martensite is designated as Md.

3th5 “I"

311315011 ‘

r
h

.oran 11

{.‘gl . i .
.‘Glitl‘vllfl!
t-

itum Eq

2pm rt;

23

phases under stress are GA and GM, then when both phases are in equilibrium at the

transition temperature, AG for the reaction is zero.
GA = GM or dGA = dGM (2.2)

For an infinitesimal change in temperature and force under constant pressure, the

following relations can be obtained:

dGA = - SA dT - l A dF (2.3)
dGMz-SMdT- 1 MdF.

FromEq.(2.2)
-SAdT-lAdF=—SMdT-lMdF (2.4)

upon rearrangement of the above equation:

31E — ASA—’M AHA-’M

at W a ' W (2'5)
where ASA—’M and AHA-’M are the entropy and enthalpy changes of the transformation
(A-rM) respectively, and it was assumed that the changes in entropy (ASA-’M) and
enthalpy (AHA—’M) during the transformation are constant over a fairly wide rage of
temperatures. To is the transformation temperature without the applied stress. If the
cross sectional area of the crystal is A and the volume of the crystal is V, then F/A = 0;

let All I = e and T = M3. Thus, the linear variation in stress to induce martensite can be

expressed by the Clausius-Clapeyron equation, which is written as

24

 

 

 

 

dF _ AHA—’M _ AHA—’M-l
dMs-A _ TONA‘M-A ' TOAIATM-V
d0 _ _ AHA—’M (2'6)
dMs V°T0sA—’M

where o is the applied stress and eA"M is the strain associated with the transformation
and AHA-’M is now the volume-based enthalpy (J/m3 = Pa). However, superelastic

effects cannot be expected if the temperature of the material is above Md, in which case
slip occurs in the austenite before induction of the martensite phase. If the temperature of
the material is below As, SME, rather than superelasticity, can be expected due to the

stabilization of the martensitic phase at that temperature.

2.3.2. Displacive Transformation in NiTi Alloys

2.3.2.1. Transformation Behavior in Near-Equiatomic NiTi Alloys Near-
equiatomic NiTi alloys exhibit unique properties, the shape memory and superelasticity
effects associated with the thermoelastic martensitic transformation. For NiTi alloys,
three kinds of phases can occur during the transformations, the B2 (CsCl structure) parent
phase, rhombohedral R-phase, and monoclinic (819') martensite phase. Fully annealed
binary NiTi alloys transform directly from the BZ parent phase to the monoclinic Bl9'
phase. However, thermally cycled or thermo-mechanically treated NiTi alloys transform

in two-steps, namely from the B2 to R-phase and then to 819' phase. Upon cooling, the

transformation occurs in the following sequence: B2 parent phase -’ Incommensurate
phase at T,‘ -* Commensurate phase (R) at T, -’ monoclinic (B 19') martensite phase.

The R-phase transformation was ﬁrst treated as a precursor phenomenon to the
martensitic transformation. Later investigation showed however, that the R-phase
transformation does not necessarily occur prior to the martensitic transformation.

Hwang et al. (1983) interpreted R-phase transformations in Ni50Ti47Fe3 alloy as an

25
incommensurate-to-commensurate charge density wave transition. They explained that

the B2 parent phase first underwent a second order transition to an incommensurate state
associated with a rhombohedral distortion that produces 1/3 (110) and (111) reflections.
The onset temperature of the B2-to-incommensurate phase transition was designated as
T,'. The extra superlattice reﬂections in fact were not located in the exact 1/3 positions
(thus the phase 'incommensurate'), and their intensities kept increasing on cooling. A
subsequent first order transition occurred involving a structural change from a distorted
incommensurate B2 phase to a rhombohedral one causing the extra reflections to lock
into precise 1/3 positions. The onset temperature of the incommensurate-to-
commensurate phase transformation was defined as T,. The R-phase transforms to the
martensite phase at M3 on further cooling. The shape memory effect and superelasticity
arise in both 820R and ReBl9' transformations (Miyazaki and Otsuka, 1986).

Physical properties, such as electrical resistivity, elastic modulus, damping
capacity, and magnetic susceptibility, change during phase transformations in NiTi
alloys, and transformation temperatures in this alloy may be determined by measuring
any of these characteristics. The change in electrical resistance with temperature and the
measurement of the latent heat during phase transformations are widely used to monitor
the transition temperatures. Differential scanning calorimetry (DSC) curves for the
transformation show the exothermic peaks on cooling and the endothermic peaks on
heating during the phase transformations.

The change in resistivity of NiTiCu alloys, which undergo a distinct two step
transformations both on cooling and heating, is illustrated in Figure 2-8. On cooling from
high temperature where the parent (austenite) phase is stable, the electrical resistance
slightly decreases first and then starts to increase, where the R-phase formation occurs
from the 32 phase, due to the negative temperature coefﬁcient of electric resistivity in the
R-phase. A sharp increase in resistance occurs at T,‘ and increases further beyond Tr until

Ms is reached. At the M3 temperature, the resistance starts decreasing until Mr is reached,

26
after which the rate of decrease becomes gradual on further cooling. On heating from a

very low temperature, where only the martensite phase exists, the resistance increases
gradually up to the As temperature where the curve changes its slope and then increases
until Af is reached, where B 19'-R transformation will finish. The resistance decreases
until the R—rBZ transformation occurs, and then slight increases is followed on further

heating. Hereinafter B29R and ReBl9' transformations are named R-transformation

and martensitic transformation respectively.

2.3.2.2. Eﬂ’ect of Composition on Transformation Temperatures The
nickel-titanium phase diagram, which is illustrated in Figure 2-9, shows stoichiometric B-
NiTi phase at high temperature with a steep solvus line on Ti-rich side. An eutectoid
decomposition from B-NiTi phase to TizNi + Ni3Ti at 903 K has been proposed.
However, Wasilewski et al. (1971) proposed that the B-NiTi phase remained a stable
phase down to room temperature, with a very narrow stoichiometric range. On the Ni—
rich side, Nishida et al. (1986) and Kim et al. (1986) reported the existence of metastable
phases other than Ni3Ti, such as TizNi3 and Ti3Ni4. But the equilibrium phase diagram
on the Ni-rich side for the low temperature range is not well established and controversy
on the composition and structure of those phases exists. Most researchers accept that the
microstructures of NiTi alloys are primarily single phase, with small amounts of other
phases distributed throughout the matrix. These phases can change the composition of
the matrix and thus affect the transformation temperatures in this alloy system.

Transformation temperature changes are extremely sensitive to the alloy
composition, especially on the Ni rich side of NiTi alloy systems, which is shown in
Figure 2-10. For example, the Ms temperature decreased linearly with increasing Ni
concentration, at about 100 K per 1 at% of Ni, for alloys with composition of 49.6 to
52.0 at % of Ni after annealing between at 773 K and 923 K (Nishida et al., 1984; Shugo

et al., 1985). However, the change in Ms temperature was relatively insensitive on the Ti

27
rich side due to the low solubility of Ti and TizNi second phase distributed around the B2

grain boundary. The quantity of TizNi particles did not affect the transformation
temperature (Lin and Wu, 1992). The compositional accuracy required for Ni-Ti in fact
exceeds uncertainty levels commonly found in routine chemical analysis. Therefore, in
most cases for quality control purposes the transformation temperatures themselves are
measured, rather than the chemical composition (Melton, 1990).

Besides the dependence of transformation temperatures on Ni concentration, there
are other factors which influence the transformation behaviors in NiTi alloy. These will

be reviewed briefly in the following sections.

2.3.2.3. Effect of Heat Treatment on Transformation Temperatures As
mentioned in the previous section, stable TizNi and TiNi3 precipitates with metastable
phases, such as TizNi3 and Ti3Ni4, can be formed in NiTi alloys. Formation of these
precipitates changes the composition of the matrix resulting in a shift of the
transformation temperature.

Recent study by Xie et al. (1990) stated that the Ti3Ni4 precipitate in aged
specimens had good coherence with the matrix and an internal stress field was built
around it Ti3Ni4 precipitate formation caused a widening of the R-phase range and a
significant decrease of Ms temperature by the effect of the internal stress fields at the
coherent boundary between precipitates and parent phases. Such dispersed ﬁne Ti3N i4
precipitates affected the crystallographic direction of the martensite and thus a single (or
strongly preferred) variant martensite could be formed in a parent grain. It was also
reported that most TizNi precipitates were segregated at the grain boundaries and ﬁne
Ti3Ni4 precipitates were distributed within the grains in annealed N iTi thin films (Busch
et al., 1990; Chang, 1993; Ishida et al., 1993).

In addition, the dispersed precipitates raise the strength of the austenite phase by

precipitation hardening resulting in an increase in the critical stress for slip in NiTi alloys.

28
It is important to raise the critical stress for slip in order to stabilize the superelasticity

characteristics against cyclic deformation.

In NiTi alloys, the transformation temperatures, recovery strain, and yield stress,
which are associated with the f orrnation of precipitates, will be affected by heat treatment
parameters, such as aging temperature and time, cooling rate, and post annealing after
cold working. Details of these effects will be reviewed briefly.

(1) Aging effect: Dependence of the transformation temperature on the aging
temperatures was studied by Miyazaki et al. (1982) and extensively by Nishida and his
coworkers (1984, 1988). In these studies, all Ti-51.0 at.% Ni alloys were solution treated
at 1273 K and aged at various temperatures for different times and then quenched into ice
water. The highest transformation temperatures were obtained from the alloys aged
between 673 K and 773 K for 1 hour, and transformation temperatures were lowered with
increasing aging temperatures. The effect of aging time on transformation was compared
between 1 hour and 50 hours for aged alloys with aging temperatures of 673 K and 773
K. The longer aging time displayed a slightly higher transformation temperatures for
each aging temperature. Both the 1 hour-773 K and 50 hour-673 K specimens revealed
similar behavior due to the similarity in dispersion density and size of the precipitates.
Thus, the transformation temperatures are severely affected by precipitate formation
which are related to aging time and temperature.

(2) Cooling rate: NiTi alloys used for transformation temperature studies by
Nishida et al. (1986) were solution treated at high temperatures (1273 K) and aged at
various temperatures before finally quenching to a low temperature. Systematic work
was done for Ti-50 at.% Ni and Ti-52 at.% Ni. The Ms temperatures was 327 K for
quenched Ti-50 at.% Ni and below liquid nitrogen temperature for quenched Ti-52 at.%
Ni. Time-Temperature-Transformation (TIT) curves were constructed for Ti-52 at.% Ni
from specimens that experienced different cooling rates: as-quenched, which was fast,

and fumace-cooled, which was slow. The precipitation sequence for various aging

Pl‘lm
said \1
train

..
D. I ‘

.rL

temper
Ihlth

Him
31 «'00!
“min
We

the ma

29
processes was investigated. TiNi3 and TizNi3 precipitates formed in slow cooled Ti-

52at.% Ni alloy but, unfortunately, this work did not treat the influence on transformation
temperatures. They also concluded that neither eutectoid nor peritectoid reactions took
place in non-stoichiometric Ti-52 at.% Ni. The precipitation sequence in Ti-52at.% Ni
alloy was expressed as TiNi -* Ti3Ni46+ TiNi -+Ti2Ni3 + TiNi -* TiNi3 + TiNi, where
Ti3Ni4 and TizNi3 are metastable phases. The formation of such precipitates causes
compositional changes in the matrix. For example, higher Ti concentration can be
expected due to preferential Ni consumption.

(3) Post annealing effect after cold working: The mechanical and shape memory
properties critically vary with heat treatment after cold working. One prominent effect of
cold working combined with heat treatment is the separation of the R-phase
transformation and martensitic transformation in near equiatomic NiTi alloys. Depending
on the post annealing temperature and the amount of cold working, the transformation
temperatures can change as a consequence of the dislocation or precipitate structure
which might strongly impede the movement of martensite interfaces and, therefore,
depress the M8 temperature signiﬁcantly. The effect of the amount of cold work on the
transformation temperatures without subsequent annealing was studied by Lin et al.
(1991) using several methods such as internal friction, shear modulus measurement,
hardness, and TEM observation. The BZ—rB 19' transformation temperature on cooling
was decreased but reverse transformation (Bl9'—rB2) temperature on heating was
increased with increasing cold work. They concluded that martensite phase stabilization
at room temperature could be provided by cold working. With higher amount of cold
working the martensite phase became more effectively stabilized. The deformed
martensite structure, containing induced dislocations was considered to be responsible for
the martensite stabilization. Legresy et al. (1991) provided more detailed information for

g

6. Ti3Ni4 precipitate phase was expressed as Ti11Ni14 in several other reports but based on the crystal
structure Ti3Ni4 is more approprate and this phase denotation was used in later report by Nishida

mi’ormat

“Jana

.‘J
1",

tguum
ummtl
shuns
amnhd

ﬁﬁSlCdsC

tmmnu
Ltmxr
bf dune
character
tummy
“is {is 1".
SUbSlllUl
[Ensitin
bl Nam
tic”

lhtSe r

3O
transformation behavior, and showed that Af temperature increased, whereas As, Mr, and

Ms decreased with increasing cold work.

2.3.2.4. Effect of Third Element Additions and Impurities There
is great interest in the phase transformation behavior of NiTi alloys when a third metallic
element, especially a 3-d transition metal, such as Co, Fe, Cu or Cr, is partially
substituted for nickel. These substitutions affect the transformation temperatures,
particularly Ms temperature, by different degrees depending on the substituted species. In
most cases of such substitution with Co, Fe, or Cr, the intermediate phase7 region is
increased by suppressing Ms temperature more rapidly than the starting temperature of
the intermediate phase. For example, Fe substitution for Ni shifted both Ms and T,
temperatures lower; but the drop of Ms temperature was three times greater than that of
T, temperature (Hwang et al., 1983). However, up to 30 at% Cu substitution for Ni could
be done without changing the Ms temperature. In addition, the shape memory
characteristics were still exhibited in NiTiCu ternary alloys (Melton and Mercier, 1978).
According to their investigation, the transformation sequence for Ti 50(Ni40Cu10) alloy
was as follows: cubic B2 -* orthorhombic B19 -> monoclinic B19'. However, if the
substituted Cu for Ni exceeded about 10 at.% , the alloy showed only one stage
transformation: cubic B2 -> orthorhombic B 19. Recently, more studies were undertaken
by Nam and his colleagues (1990) of Cu substitution in NiTiCu alloys, especially for 10
at.% Cu, were compared with binary NiTi alloys under constant thermal cycling test.
These researchers observed that BZ-> orthorhombic Bl9 transformation starting
temperature increased slightly, whereas Bl9—t monoclinic B19' transformation starting
temperature decreased. Thus, the temperature region where 819 was stable became

wider with increasing Cu content. But, for NiTiCu alloy in which Cu concentration was

‘

7. It has been reported that the transformations were clearly produced by the two-stage processes: high
temperature phase -’ intermediate phase —’ low temperature phase due to the substitution of Co, Fe, and Cr
(Fmakubo, 1984)

attic 13.5
animal

The
317118111. 0
sprorimr‘
rise 70"!

“it”.

M,

11.618 It,
ttialirc in
it pitttr
bntt‘acnct,
lilrr‘duccl
Mg tempt
d‘lﬁng pr
Hamil. rm
bl hut;
lit shap
ﬂu 199'";

3Wis \1
the Ship
Ph‘dSe, I
8. Hldro

Whig
1th icing.

31
above 12.5 at.%, only the BZ—rBl9 transformation was faVOrable while the Bl9-tB19'

transformation was suppressed.

The transformation temperatures in NiTi alloys are also sensitive to impurity
content. Oxygen causes a decrease in Ms temperature, which changes monotonically by
approximately 100 K/at.% with increasing oxygen concentration from an alloy annealed
above 700 K (Shugo et al., 1985). The experimental equation reported from that study

was:
Ms/K = (351.4 :1: 0.5) - (92.63 :1: 0.93) x (xo/ at.%) (2.7)

where X0 = 0.13 - 1.06 at.%. Most oxygen was bound in Ti4Ni20, and resulted in a
relative increase in nickel concentration. In addition, the TizNi phase was stabilized by
the presence of oxygen in this alloy system, and the contamination by oxygen increased
brittleness (Nevitt, 1960). Other interstitial impurities, such as carbon, which is easily
introduced during melting from a graphite crucible, and nitrogen, have a similar effect on
Ms temperature, but to a lesser degree, due to the preferential consumption of titanium
during precipitation of TiC, TizNi, and Ti4Ni2N. However, hydrogen has little effect on
transformation temperatures8 or mechanical properties in NiTi alloy even though hydride
by hydrogen absorption may fornr (Shorshorov et al., 1985; Asaoko et al., 1992). Also,
the shape recovery was decreased with presence of small amount of hydrogen (Wade et

al., 1990).

2.3.2.5. Self-Accommodation and Surface Relief Shape change
occurs when one martensite plate is formed within a single crystal of the parent phase and
the shape change is manifested directly as changes in the external shape of the parent

phase. Even though the parent phase consists of a single crystal, a number of martensites

h

8. Hydrogen-induced shape memory effect, which was believed to be the result of surface hardening,
caused higher tranformation temperatures by hydrogen charging on NiTi alloy but the change disappeared
after removing the surface layer (Wade et al., 1990).

mlh Lilli-t
it calls
displace
331.115 1
strain er
aroma;
taunts.
1113.17.11
1 13pm
timed
Slflln m
the 3hr;
such a
resultn
accomr

littlt‘iru

studied
fundrn
Ianan:
here (1
315 C;

Ollf am.

free gu

32
with different habit plane indices. These martensites with different habit plane indices

are called 'variant'. During a martensitic transformation, each martensite plate causes a
displacement resulting in an increase in the total strain energy. However, martensite
crystals in thermoelastic alloys, since martensite variants are created to minimize the total
strain energy through a twinning process, mutually reduce the transformation strain
accompanying a multi-variant combination during the formation of the martensite
variants. This effect is called 'self-accommodation'. According to the observation by
Miyizaki et al. (1989), the martensitic transformation in NiTi alloys occurred by forming
a typical triangular morphology consisting of three variants and these three variants
formed plate groups clustered about one of the {100}132 poles. The calculated shape-
strain matrix for the overall triangular morphology showed that the shear components of
the shape-strain were very small. They found that there were 16 groupings which formed
such a triangular self-accommodation morphology about each of the {100}132 poles,
resulting in 48 possible combinations in all to form such triangles. The ability of self-
accommodated martensite to be deformed by means of detwinning is a fundamental
factor underlying shape memory properties.

Self-accommodation mechanisms of the R-phase in NiTi alloys have been
studied by optical and electron microscope observations (Fukuda et al., 1992). The
fundamental unit of the self-accommodation in R-phase of NiTi alloys consisted of four
variants. The four variants self-accommodated around one of the <001>132 axes, and
there were 6 groupings self -accommodation morphology around each of the <001>132
axes canceling their shape deformation each other. These variants were twin related to
one another and there were two types of twin relations.

A martensitic transformation is accompanied by various characteristics like
diffusionless character, surface relief, and the presence of many lattice imperfections. It
is an evident feature of a martensitic microstructure that a surface relief is produced on a

free surface of bulk materials. The surface relief is not irregular, and the angle of incline

tithe uphe;
hal.l97t
amnionna
hehteen V
1331131311011 1
ashiltn.
Phase Ms

1].. I993)

thingii

It Phys
3.4.1. De}
Al
thse 01
tritium
that 01 m;
thetsi
high mob
WI be dc

dlihjcam

33
in the upheaval has a definite value which depends on the crystal orientation (Nishiyama

et al., 1978). Optical micrographic observations have been performed to observe the
transformation morphology and to determine the habit planes and junction planes
between variants in self-accommodating morphology. The surface relief due to the
formation of B 19' martensite has been described as a dominance of three variants forming
a small triangular grouping consisting of three variants (Miyazaki et al., 1989) and the R-
phase was characterized by saw-tooth like and long straight band morphology (Fukuda et
al., 1992) in bulk NiTi alloys, which is illustrated in Figure 2—11. These are typical

features of self -accommodation of B 19' martensite and R-phase in NiTi alloys.

2.4. Physical and Mechanical Properties of NiTi Alloys
2.4.1. Deformation of NiTi Alloys

Almost all physical properties of austenite, except the density, are different from
those of martensite, due to a significant change in crystal structure during the
transformation. The yield strength of TiNi austenite is about two to six times higher than
that of martensite, and its elastic modulus has the lowest value near Ms temperature. It is
believed that the low yield strength of thermoelastic martensite is due to the relatively
high mobility of the boundaries between neighboring variants. In NiTi alloys, martensite
can be deformed up to 5 to 10% by reorientation of martensite variants without activating

dislocation slip.

2.4.1.1 Stress Strain Behavior in NiTi Alloys The stress strain
behavior of NiT i alloy can be generally divided into five different types depending on the
test temperatures (as normalized to the canonical transformation temperatures: Ms, Mr,
A s, and Af). The five types of the stress-strain curves for NiTi alloy are shown in Figure
2-12. If the deformation temperature is below Mf (Type I), the alloy is entirely in a

martensite phase so that deformation proceeds according to the migration of the

ittfla
[JCTEIIS

:nitlt‘t

has

it‘ll

34
intervariant boundaries within the martensite phase. When the test temperature is

increased to above Mf, but is below Ms (Type II), the alloy deforms because of stress-
induced growth of martensite phase fraction, and continues with the same mechanism as
Type I. In the temperature range between Ms and Af (Type III), the def orrnation is only
by formation of stress-induced martensite. In Type IV, which temperature range is
between Af and Md, superelastic characteristics are revealed. Here, the martensite phase
forms under stress, and disappears when the stress is released. Above Md (Type V), the
plastic deformation in the austenite phase occurs before stress-induced martensite is
formed, because the critical stress for slip in the austenite is lower than the stress to
induce martensite.

The stress strain behavior of a NiTi alloy as function of temperature has been
extensively studied by Miyazaki et al. (1982). In the study of bulk NiTi alloy,
superelastic strain recovery (without supplying external heat) was found at test
temperature about 15 K above Ar. Recently, the stress strain behavior of NiTi thin film
alloy produced by do magnetron sputtering were obtained at various test temperatures by
Hon and Grummon (1995). Results are shown in Figure 2—13. Most of characteristics of
the thin ﬁlm were similar to those of bulk NiTi alloys. The results showed that the stress
for inducing martensite obeyed Clausius-Clapeyron relations, and the full superelasticity
was found at about 30 K above Af. It is worth noting that the slope of the loading curve
before stress induced martensite has formed is higher than that of the unloading curve.
The formation of stress-induced martensite results in the differences of Young's modulus
(EA > BM) in the films between the loading and unloading stages. The amount of
unrecovered strain, which was designated as an in Figure 2-13, increased when the test
temperature deviated from 333 K, which is 28 K above Af, in both sides: due to
occurrence of slip at higher temperature and stabilization of stress induced martensite at
lower temperature. This higher temperature for full recovery in the ﬁlm may be

attributed to the internal defects during deposition process.

shall all
5:1th11 .'
supcrt

R5310]

hair
19%).

that:
turd t
scum
rearnn
appha
harm
tensile
cnhcr
screen
165‘. le-
conch

Exhiht

35
2.4.1.2. Two Stage Yielding in NiTi Alloys NiTi alloys usually

show an R-phase transformation prior to the martensitic transformation, as described in
section 2.3.2. R-phase transformation is accompanied by rhombohedral distortion so that
a superelasticity effect associated with R-phase can be observed. The mechanical
behavior related to R—phase formation by an external applied stress was studied for NiTi
binary alloy and NisoTi47Fe3 ternary alloy which exhibited a well-defined R-
transforrnation and a subsequent martensitic transformation (Miyazaki and Otsuka, 1984;
1986). Miyazaki and Otsuka (1984; 1986) carried out tensile tests over a wide
temperature range to study stress-induce transformation associated with the R-phase and
found that the first yielding occurred at low stress (~50 MP3), with subsequent yielding
occuning at higher stress (above 200 MPa). The first yielding was attributed to the
rearrangement of R-phase variant (or variants), which was more favorable under the
applied stress, whereas the second yielding was attributed to the stress induced
martensitic transformation. The first stage yielding was confirmed by well-controlled
tensile tests, where the stress was sufficiently low to avoid the second stage yielding. The
critical stress to form R-phase was minimal at T, temperature and increased with
increasing temperature above T, temperature. The stress-strain curves associated with the
test temperature for the alloy were classified as shown Figure 2—14. These researchers
concluded that both the R-phase transformation and the martensitic transformation

exhibited the shape memory effect and the transformational superelasticity.9

2.4.1.3. Plastic Deformation of NiTi Alloys Recent study of plastic
deformation of NiTi alloys (Miyazaki et al., 1991) showed that the stress for slip in
austenite phase was almost independent of the temperature, whereas the stress in
martensite phase decreased with increasing temperature. The elongation of the martensite

was always larger than that of austenite phase. Thus, one can combine the relation of the

 

9. These phenomena were more clearly observed in NiTi binary alloys than in NiTi Fe ternary alloys due
to the high critical stress for slip in binary alloys (Miyazaki and Otsuka, 1986).

{365585 I01

n
P11156111

2.4.
grain size
141.51:
he polycr
innumnt
discard
it “relax
tine grin
sequent:
nancns:
5th its
grain bu
hate 3h
subsist
line gm:
[1.03 .0

neatly c

9156,9110:
maﬂénsit
'0 nld be

Sﬂ‘eral N

36
stresses for slip in martensite and austenite phases and for inducing martensite and R-

phases with elongation as a function of temperature, which is illustrated in Figure 2- 15.

2.4.1.5. Grain Size Effect on Superelasticity in NiTi Alloys The
grain size of NiTi alloy affects the superelasticity effect in NiTi alloys (Saburi et al.,
1984). Single crystal NiTi alloys did not display superelasticity at any temperature but
the polycrystalline alloys exhibited complete superelasticity behavior above Af. It is
important to note that the superelasticity becomes more pronounced as the grain size is
decreased, indicating that the reduction of the grain size is very effective in improving
superelasticity, which is shown in Figure 2-16. In a later work (Motohashi et al., 1991),
ﬁne grained NiTi alloy was produced by therrnomechanical treatment which involved
sequential stages of annealing and cold rolling. Motohashi et al. found that the
martensitic transformation temperatures, particularly Ms temperature, increased as grain
size decreased. This shift in Ms temperature to higher temperature was attributed to the
grain boundaries as nucleation sites for the martensitic transformation. Recent studies
have shown that when depositing NiTi with d.c. magnetron sputtering, heating the
substrate to temperature between 598 K and 723 K produced a deposit with extremely
ﬁne grain size (Hou and Grummon, 1995). They have observed grain size on the order of
0.02 - 0.05 um and the films have shown classical transformational superelasticity, and

nearly complete strain recovery at temperatures above Af.

2.4.2. Fatigue Properties of Near-Equiatomic NiTi Alloy

Early studies of the fatigue of NiTi alloys by Melton and Mercier (1979) showed
exceptionally large accumulated strains to failure when the alloy was tested either in the
martensitic condition, or in the austenitic condition at temperatures where martensite
could be stress induced during low cycle fatigue measurement. After investigating

several NiTi alloys which had different Ms temperatures, fatigue life was found to

37
increase with decreasing Ms temperature. A significant finding in this work was that

superelastic deformation lead to high fatigue life for NiTi alloys in both the austenitic and
fully martensitic conditions. When fatigue life was plotted against the strain amplitude
for different alloys, the longest fatigue life was obtained at about 15 °C above Ms
temperature, but the fatigue life decreased at higher temperatures, as shown in Figure 2-
17.

In a later work by Miyazaki et al. (1989), instead of the tension-compression
cycling used by Melton and Mercier, loading-unloading (tension-tension) cycling at
various temperature was adopted. The critical stress for inducing martensite increased
with the test temperature and the fatigue life of the alloys was improved with increasing
test temperature, which is shown in Figure 2-18. Here, all test temperatures were above
A f and below Md, which was presumed to be 130 K above Af. In another study
(Miyazaki et al., 1986), it was found that the critical stress for inducing martensite
gradually decreased, the superelastic strain hysteresis loops became smaller, and residual
strain increased with increasing cycles during cyclic loading-unloading of NiTi alloys.
But all these values became insensitive to cycling with increasing cycle number and
maintained a steady state with respect to superelasticity characteristics. Superelasticity
characteristics can also be influenced by heat treatment (different aging temperature,
time, and annealing histories) and mechanical treatment (cold working), all of which can
raise the critical stress for slip in NiTi alloy (Tadaki and Wayman, 1980). The combined
effects of precipitation hardening and hardening due to high density of dislocations can
be effective in stabilizing superelasticity characteristics against cyclic deformation.

In summary, NiTi alloys have been found to possess exceptionally long fatigue
life under conditions of high strain amplitude fatigue. The phenomenal performance of
these alloys is attributable to their ability to undergo stress-induced, reversible,
martensitic transformations (up to 5 - 10% of strain) and reversible twinning and

detwinning reactions in the martensite phase.

38
2.5. Thin Film Processing

2.5.1. General Features of Thin Film Processing

Thin films can be fabricated broadly by either physical or chemical deposition.
The former includes vacuum evaporation and sputtering, and the latter consists of
chemical vapor deposition (CVD), thermal oxidation, and electroplating. Thin film
properties are strongly dependent on the method of deposition, the substrate materials,
substrate temperature, and the rate of deposition. For instance, CVD is a process
whereby a solid material is deposited from vapor by a chemical reaction occurring on or
in the vicinity of the heated substrates. Although the CVD technique is able to produce
coatings with uniform thickness and low porosity, even on substrates of complicated
shape, the properties of CVD materials are affected by the interacting processes
associated with chemical reaction near or on a heated substrate surface. To date, no
reports have appeared which use CVD process for NiTi ﬁlms.

Electroplating, which is the process of depositing a substance by the passage of
electric current through an electrolyte, has been used in both decorative and engineering
application as an economical technology. However, the process is problematic in the
case of titanium,and will not be discussed further.

In the vacuum evaporation process, vapors are produced from a material located
in a source which is heated by direct resistance, radiation, eddy current, electron beam, or
an arc discharge. Deposition rates are relatively high compared to sputter deposition,
especially with the electron beam evaporation technique. However, it is difficult to
produce alloy ﬁlms in the stringent stoichiometry limits needed for thermoelastic NiTi
films. Also, this technique is limited if conformable coverage and strong adhesion to the
substrate are needed, since there is a relatively low kinetic energy in the evaporant atoms.

On the other hand, sputter deposition permits a wide selection of ﬁlm materials
and a good control of deposit thickness. It also produces ﬁlms with high purity and better

controlled composition, while providing films with good adhesion and homogeneity. The

39
transformation behavior of NiTi alloys is highly dependent on the composition.

Therefore, sputter deposition is the most promising method for the surface coating by
NiTi alloys where careful composition adjustment is needed to provide desired surface
film properties, such as superelasticity. In this section, general aspects of thin film
structure produced by sputter deposition and ion beam sputter deposition will be
described and, in particular, the characteristics of NiTi thin films produced by sputtering

will be discussed.

2.5.2. Thin Film Processing by Sputter Deposition

Sputtering is described as the removal of atoms from a surface or near surface of a
solid by energetic particle bombardment. This phenomenon has many applications,
ranging from surface cleaning and etching, to thin film deposition. The sputtering yield,
1’, is defined as the average number of ejected atoms per incident ion, and is a function of
many parameters, including the energy, the mass, and the incident angle of the projectile
(bombarding energetic particle), as well as the mass and the surface binding energy of the
ejected atom. The dependence on sputtering yields for the ion and ion energy for nickel
and titanium are shown in Figure 2-19. Within the energy range of practical interest in
sputtering processes (50 to 5000 eV), the sputtering yield increases with ion energy, but
the variation of the sputtering rate in the low energy region is more sensitive. The
sputtering yield data and threshold energies for various pure metals are listed in Table 2-
2. The incident angle of the projectile also inﬂuences the sputtering yields, usually the
maximum yield is obtained near 65 °- 75 °.10 Figure 2-20 shows the variation of
sputtering yield with incident angle for Art“ on different targets. Also, the emission angle
(0) of the sputtered particle affects the yield and the angular distribution, from the target

surface-normal, of the sputtered atom roughly follows a (cos0)'1 relation.

 

10. The maximum sputtering yield for light ion sputtering was obtained from a higher incident angle than

the heavy ion sputtering. The maximum was not reached before 80 ° by H+ for Ni, Mo, and Mo target
(Anderson and Ray, 1981).

40
Among the several sputtering processes, ion beam sputtering, which has been

used in the present study, and magnetron sputtering will be discussed briefly. In ion
beam sputter deposition, an inert ion beam generated at an ion source is extracted into a
high vacuum chamber and directed to a target of the desired material, which is sputtered
and deposited onto a nearby substrate (George, 1992). Ion beam sputtering permits
independent control over the energy and current density of the bombarding ions. Low
background pressures (s 0.1 mTorr) are used, resulting in minimized gas inclusions and
reduced scattering of the sputtered particles during transit. Also, the directionality of the
beam allows the investigator to vary the angle of incidence of the beam against the targets
as well as the angle of deposition on the substrate. Adhesion of the ion beam sputtered
ﬁlms is superior to evaporated films, the improvement is attributed to the higher energy
of arrival of the sputtered particles. The arriving particle energy helps to clean the
substrate by sputtering away absorbed gases, it gives greater mobility to the condensing
atoms, and it provides implantation into the substrate, enhancing chemical bonding and
localized rearrangement

On the other hand, in magnetron sputtering deposition, ions are extracted from the
plasma and accelerated to strike the target, resulting in the sputtering of the target
material. Most magnetron sources operate in the pressure range from 1 to 20 mTorr and a
cathode potential of 300-700V. In this process, the magnetic field does not directly affect
the ion motion; however, because of electrostatic attraction the ions move with the
electrons, keeping the plasma neutral. In a conventional magnetron, most of the
discharge is confined near the cathode surface and, therefore, bombardment of the
growing film by electrons and ions is minimized.

Thin films can be formed on the desired substrates by sputter deposition. The
nucleation of the film on the substrate can be described in several steps. Several modes
of thin film growth have been observed depending on the ﬁlm-substrate interactions: (1)

island (Volmer-Webber) mode (2) layer (Frank-van der Merwe) mode, and (3) Stranski-

41
Krastanov mode (Oring, 1992). These film growth modes depend on the film—substrate

interaction, producing island growth when the condensing atoms are bound more strongly
to each other than to the substrate and layer by layer growth when the interaction with the
substrate is stronger. In the latter case, the interaction will monotonically decrease as
each layer is added.

The microstructure which results from sputter deposition is sensitive to deposition
conditions such as the pressure of working gas, and the substrate temperature. A
relationship between microstructure development and processing conditions was
proposed by Thornton (1978) in the form of a structure zone diagram, as shown in Figure
2-21. According to his diagram, the microstructure of the sputter-deposited film is
affected by the gas pressure and the substrate temperature expressed as a fraction of the
absolute melting point (T/Tm). At low film temperatures, shadowing effects (a simple
geometric interaction between the roughness of the growing surface and the incident
angle of the arriving atoms) result in poor-quality films with open boundaries, which is
called Zone 1 structure. Increasing ambient pressure extends the Zone 1 structure to the
higher temperature region. At very high homologous temperatures (T/Tm 2 0.5), bulk
diffusion enhances the crystallization process, which produces large columnar or
equiaxed grains (Zone 3 structure). Zone 3 structures are recognized by dense grain and
twin boundaries, and by grain shapes that do not coincide with the substrate and coating
surface topologies. Thus simple metals deposited in high TIT m regions are similar to
fully annealed bulk materials. With limited atom mobility to overcome the shadowing
effect, dense fibrous structures (Zone T) and columnar structures (Zone 2) are formed at

intermediate temperatures (0.3 < T/Tm < 0.5).

2.5.3. Sputtering ofAlloys
When a virgin multicomponent alloy target is subjected to energetic particle

bombardment, the elemental composition of the sputtered flux is usually different from

42
that of the target surface. This phenomenon is called preferential sputtering (Betz and

Wehner, 1983). The preferential removal of one component from the sputter-target
surface leads to the formation of the so-called 'altered-layer', which is a near-surface
region with a composition different from the bulk. At sufficiently low target
temperatures, where thermal diffusion of the constituents is negligible, the thickness of
the altered-layer is limited to the depth in which the collision cascade occurs. Steady-
state conditions can be attained in which case the sputtered ﬂux and the bulk alloy are
constrained to have the same composition. When steady-state has been achieved, the film
composition will depend on the interaction of the sputtered ﬂux with the working gas,
and by differential sticking coefficients and preferential re-sputtering of film component
atoms.

It has been reported that both ion incident and emission angles can affect the
composition of deposited thin film in ion beam sputtering deposition (Theirich and
Engemann, 1991). In their study, WTi thin films were produced from a single WTi alloy
target. For a normal incidence beam the composition change with emission angle was not
prominent, but for an oblique incidence angle the titanium content decreased with
increasing emission angle. The concentration of titanium in the ﬁlm was always lower
than that of target. Such lower contents of titanium in the deposited ﬁlm was attributed to
the low sticking coefficient of titanium on the growing film. The sticking coefficient for
sputter particles in most cases can be expected to be 1. But, recent studies have indicated
that the sticking coefficient may be less than 1 for a low atomic weight element, while the
sticking coefficient may be equal to 1 for a high atomic weight element (Emmoth and
Bergsaker, 1988). This seems to be due to reﬂections of the light sputtered atoms from
heavier atoms of the substrate. In their studies, the targets were bombarded by a 10 keV
Ar beam with a current density of 10-20 rim/cm2 for the production of sputtered metal
atoms and relative sticking probabilities for sputtered atoms on various substrates were

investigated. For nickel, stainless steel and vanadium the calculated relative sticking

43
probabilities on lighter materials, such as beryllium, graphite and silicon, differed by only

1-2%. However, the sticking coefﬁcient of sputtered vanadium (M = 50.94) on silver (M
= 107.87) substrates has been determined to be around 0.82.

Concerning the angular distribution of sputtered atoms for the normal incidence, it
is known that the distribution is occasionally ”under cosine” for heavy ions whereas the
angular distribution is "over cosine" for light ions, which is illustrated in Figure 2-22. In
addition, the angular distribution gradually changes to a more symmetrical one as the ion
beam energy increases (Okutani et al., 1980). However, it appears very difficult to
predict what the angular distribution of the components will be for alloys. The studies for
Cu/Ni and Fe/Ni alloys by Olson et al. (1979) showed ambiguous results for Ar ion
bombardment under 1 keV. In their work, even though Fe and Cu were located on the
opposite side of nickel in the periodic table, i.e. M1:e (55.84) < MN, < MCu (63.55), nickel
was enriched in the normal direction to the surface in both cases. Therefore, the atomic
weight effect is not enough to interpret the enrichment of nickel in the direction normal to
the surface. No logical explanation for preferential emission for these alloy systems was
offered. However, the binding energy of each element was set equal to its heat of
sublimation, and there the energies for Fe and Cu are 4.28 eV and 3.49 eV respectively.
The result indicated that the binding energy effect for an Fe/Ni alloy and the atomic
weight effect for a Cu/Ni alloy were dominant.

In addition, the directionality of the beam during ion beam sputter deposition will
give rise to a deviation from the cosine distribution of sputtered atoms. There is usually a

small lobe opposite the incident ion beam, which results from the enhanced sputtering
due to glancing collisions of ions. Under bombardment conditions, where random

collision cascade theoryll applies, both experiment and linear transport theory give cosine

 

1 1 - . The impact of energetic ions with a solid produces a varity of effects. The incoming ion loses energy
by Inelastic collisions which produce electronic excitations and by elastic collisions which displace atoms
from their lattice positions. The displaced atoms produce additional collisions in a short period of time and

e Series of events is called collision cascade. For the collision cascades started by the primary knockon

a , these are classiﬁed into three different regimes: (l) the single-knockon regime, (2) the linear cascade

 

4.4
distributions. The angular distributions are independent of the incident angle of the beam

in the regime of cascade sputtering. The ejection distribution is rotationally symmetric
with respect to the surface normal, whatever the angle of incidence. However, in the low
energy regime of sputtering, the emission distribution depends on the incidence angle and
deviates from the symmetry. Very little information is available on angular distribution
in the regime of nonlinear cascades. The angular distribution of sputtered Si atoms was
studied by Okutani et al. (1980) at angle of incidence of 60° by Ar+ for 3 keV and 10
keV. It has been also reported that the angular distribution of sputtered polycrystalline W
was measured at an angle of incidence of 60° by Kr‘l’ from 5 keV to 19 keV by Oechsner
(1975). Both results revealed similar trends in the angular distribution by an oblique
incident beam in this regime. With decreasing bombarding energy the direction of
preferential emission was continuously shifted from the target normal into the region of

specular reﬂection of the incoming ions for moderate bombarding angles.

2.5.4. Stress in Thin Films

There may be either intrinsic and extrinsic stresses in thin films. Intrinsic stresses
come from defects such as dislocations and precipitates in the ﬁlm and extrinsic stresses
from the differences of the thermal expansion between the ﬁlm and the substrate. Thin
ﬁlms deposited by physical vapor deposition, such as sputtering and evaporation, possess
a residual stress which can result in film buckling, cracking, or delamination. Generally,
sputter deposited films prepared at high pressures have tensile stresses, but at low
pressure, energetic particle bombardment of the growing film may give rise to high
compressive stresses due to the recoil implantation of the surface atoms, in what is called
the atomic peening effect (Mattox 1994).

Most importantly, the existence of residual stresses in NiTi films can shift the

transformation temperatures, according to the Clausius-Clapeyron equation described in

 

regime, and (3) the spike regime (Sigmund, 1981). The detailed characteristics of each regime will be
provided in the reference.

45
section 2.3.1. Residual stresses in both as-deposited and annealed (873 K, 15 min.) NiTi

thin ﬁlms on silicon wafers bombarded by Art during deposition were studied by Walles
(1993). Films deposited by ion sputtering without assist beam bombardment showed
tensile residual stresses of about 50 MPa. After annealing, the tensile residual stress
increased to about 380 MPa due to large differences of thermal expansion coefficients

between silicon and NiTi thin film.

2.5.5. NiTi Thin Film by Sputter Deposition

2.5.5.1 Crystallization of NiTi Thin Films NiTi alloys in the form
of thin films have been produced by d.c. magnetron (Busch, 1990; Kim 1986; Chang,
1993) or r.f. magnetron sputtering (Ishida, 1993; Rai and Bhattacharya, 1987), and the
crystal structures and shape memory effects in the films have also been investigated. As-
sputtered thin ﬁlms of NiTi alloys deposited at substrate temperature below 598 K were
amorphous in structure (Hou, 1995), with no phase change or shape memory effect. The
stability of the amorphous phase was attributed to the small free energy gap between the
amorphous and crystalline phases, the sluggish diffusion of titanium, and a large negative
enthalpy of mixing (Chang, 1993). Although such as-deposited amorphous ﬁlm can be
used as corrosion resistant and wear resistant coating, the films exhibit shape memory
effects only in the crystalline structure, and hence it is necessary to crystallize them.

The crystallization temperature has been reported to be about from 753 K to 853
K (Busch, 1990; Ishida, 1993; Kim, 1986) and the transformation is exothermic. But,
Moberly et al. (1992) reported the crystallization temperature changes depending on the
thickness of the films. It was concluded that the complete crystallization of thicker film
occurs at lower temperature than the thinner ones, but that nucleation occurs
homogeneously and results in random orientations. No nucleation occurred until the
temperature was greater than 853 K for a less than 100 nm thick film. It was also

reported that the crystallization progressed much faster in thicker films (Johnson, 1991).

46
The crystallization temperature increased monotonically from 748 K to 778 K with

decreasing Ti contents from 51.0 at.% to 47.4 at.%. An increase in the heating rate raised
the crystallization temperature, but depressed the associated enthalpy (Chang, 1993).

According to in—situ TEM observations by Moberly et al. (1992), the
crystallization of the amorphous NiTi thin ﬁlms occurred homogeneously and the grains
were grown spherically until they were impeded by the grain boundaries. The grain size
of the annealed thin NiTi film was almost the same or greater than the local film
thickness. Thus, the regions which possessed more than one grain through the film's
thickness were rare (Moberly et al., 1992; Chang, 1993). Such large grain formation was
attributed to the difficulty of the nucleation due to strain build up in the amorphous
matrix which hindered further nucleation after the beginning of two dimensional growth
(Moberly et al., 1992).

However, sputter deposition of NiTi alloy on hot substrates produces crystalline
deposits, and the post-crystallization annealing process is not necessary. Recently, the
thin ﬁlms grown by d.c. magnetron sputter deposition of NiTi alloy on hot substrates,
with temperatures between 598 K and 723 K, by Hou and Grummon (1995) revealed
ultra-ﬁne grain size (20-50 nm). Furthermore, when the substrate temperature was in the
range of 598-673 K the produced film exhibited a strong texture of a (110) 32, but this
texture was disappeared when the film deposited at higher substrate temperatures (above
673 K).

2.5.5.2. Transformation Behaviors of MD Thin Films Transformation
behavior of NiTi thin ﬁlm can be inﬂuenced by its chemical composition, impurities in
the ﬁlm, and heat treatment. The inﬂuences of heat treatment and precipitation were
studied by Busch et al. (1992). Film annealed isothermally at 813 K showed a decrease
in the transformation temperatures and a more pronouncing R-phase with increasing

annealing time. When the ﬁlms were annealed in the range of 813 K to 1073 K for 30

47
minutes, the transformation temperatures were decreased and R-phase range was

increased with increasing annealing temperature. Busch et al. observed two types of
precipitates which were formed at the grain boundaries and within the grain interiors.
The precipitates inside grains were identified as TizNi, and those formed along grain
boundaries as Ti3Ni4. TizNi precipitates in grain interior were attributed to a decrease of
M3 temperature due to the preferential use of titanium in the matrix. Also, they explained
that Ti3Ni4 precipitate formation caused a widening of the R-phase range and a
significant decrease in the Ms temperature by the effect of the internal stress fields at the
coherent boundary between precipitates and parent phases. The finely dispersed Ti3Ni4
precipitates did not provided the nucleation sites for martensite, however, the coarsened
Ti3Ni4 precipitates provided the nucleation sites for martensite and constrained it from
further development (Xie et al., 1990). This effect might be more dominant than an
increase in the relative titanium concentration in the matrix. However, annealing at 813
K for 2 hours did not produce any precipitates at grain boundaries in their studies.
Another study showed that the thickness of the crystallized NiTi thin film, which also
was deposited by d.c. magnetron sputtering, affected the Ms temperature (Xie et al.,
1990; Moberly et al., 1992). The Ms temperature of those films increased with thickness.
They observed that the thicker film (> 7 pm) showed a MS temperature above room
temperature, whereas the thinner film (< 1 pm) showeda Ms temperature below room
temperature.

In summary, NiTi alloys in the form of thin ﬁlms, which display the shape
memory effect and superelasticity, have been studied recently to aide in the development
of the several microdevices, microactuators, microvalves and deformable membrances.
For practical engineering applications, the proper control of transformation temperatures
in the ﬁlms is important. The transformation temperatures are inﬂuenced by several
factors such as grain size, heat treatment, precipitate formation, and chemical

composition. Until now, most techniques have used cold substrate deposition by

48
sputtering. In this case, as-deposited ﬁlm has an amorphous structure which has no shape

memory effect or superelastic properties, and thus crystallization annealing at a certain
temperature (~773 K) is required. Recently, however, sputter deposition of NiTi alloy on
hot substrates has been introduced, where the post-crystallization annealing process is not
necessary. In this case, ultra-fine crystalline NiTi films have been produced, which is

effective in improving the superelastic effect.

3. EXPERIMENTAL METHODS

Introduction

The experiments performed in this study are divided into three major parts: (1)
NiTi thin film deposition by ion sputtering and film crystallization procedures, (2) the
characterization of the films, and (3) the detailed study on fatigue life and surface damage
accumulation during and/or after low cycle fatigue. In this chapter, specimen
microstructure and detailed sample geometries will be discussed, along with material
sources and sample preparation procedures. The section on thin film characterization
includes compositional analysis using ZAF correction, the martensitic transformation
behavior of crystallized films by electrical resistivity measurement, and crystal structure
determination by X-ray diffraction analysis. Since many fundamental fatigue studies of
fee metals have been reported for single crystals and polycrystals of copper, and fatigue
phenomena, such as persistent slip band (PSB) formation and fatigue crack initiation, are
by now well established for this material. Thus, pure polycrystalline copper was used as
a substrate metal in the present work. The fatigue lives of surface ﬁlm coated specimens,
both with amorphous and crystallized films, and bare copper specimens, were
investigated by low cycle fatigue tests conducted under constant plastic strain amplitude.
Fatigue tests were carried out at various temperatures with respect to the transformation
temperatures of coated films, which were martensitic, austenitic, or dual-phase mixtures.

In addition, the effect of initial phase1 of the ﬁlm on fatigue properties was studied.

 

1. Since NiTi alloys exhibit a transformational hysteresis, the initial phase of the ﬁlms at the same test
temperature will be different depending on thermal history.

49

50
3.1. Fabrication of Fatigue Specimens

The various steps involved in the fabrication of low cycle fatigue specimens are
illustrated in Figure 3-1. Dog-bone shaped fatigue specimens of 127 mm length were
machined from cylindrical oxygen free high conductivity (OFHC) copper rod. The detail
sample geometries are depicted in Figure 3-2. The nominal composition of this material
is tabulated in Table 3- 1.

Table 3-1
Nominal Composition of OFHC Polycrystalline
Copper Used for Fabrication of Fatigue Specimen 2

 

 

 

 

 

Element Amount (%)
Cu 99.99 min.
0 0.0010 max.
Hg 0.0001 max.
Cd . 0.0001 max.
Zn 0.0001 max.
S 0.0018 max.
Pb 0.0010 max.
Bi 0.0010 max.
Se 0.0010 max.
Bi 0.0010 max.
P 0.0003 max.
Te 0.0010 max.
Se 40 ppm max
Bi 40 ppm max
As 40 ppm max
Sn 40 ppm max

 

Two different kinds of specimens were fabricated depending on the purpose of the
study. One had 2 mm ﬂat surface region in the gauge length area, to facilitate direct

observation of crack development on the surface, and investigation of changes in crystal

 

2. The Young's moulus is 120 GPa and the coefﬁcient of thermal expansion is 17x10 '6/ °C.

51
structure by X-ray diffractometry. These specimens were rotated during deposition to

allow for coverage of the entire gauge section. Because of the difficulty in deposition and
crystallization of the films on the entire surface in the gauge section, a second specimen
geometry (for evaluating the fatigue life) was provided with a shallow round notch on one
side, to increase local stress and cause crack to initiate at the notch, with films applied
carefully to the pre-dertermined crack initiation zone. This specimen was coated without
rotation. Both specimens had smooth ends for use with hydraulic collet grips.

After machining, the surfaces of the specimens were mechanically polished (in a
specially designed fixture) using silicon carbide papers down to 1200 grit; further
polishing was carried out using 600, 5, and 1 pm alumina abrasives. Since the final
surface finish was produced by electropolishing after an annealing treatment, a mirror
ﬁnish was not required at this stage. However, the surfaces were free of deep scratches,
which would be difficult to remove during electropolishing.

The specimens were annealed for 1 hour in vacuum at 1173 K and cooled slowly
to room temperature to relieve stresses which may have developed during machining.
The vacuum was kept below lxlO‘4 Pa during heating and cooling to minimize surface
oxidation3. The ﬁnal grain size after annealing was determined by optical metallography.
Samples, cut from the specimens, were mechanically polished with silicon carbide and
alumina down to 0.3 um, and etched in a solution of 10% HNO3 and 90% H20 for 10
sec. The microstructure of the annealed copper is shown in Figure 3-3. The average ﬁnal
grain size of these samples was approximately 300 um (annealing twin boundaries were
excluded in this measurement).

The gauge sections of the heat treated specimens were electropolished to provide
a smooth surface free of residual stress. The finish also facilitated investigation of
surface morphology changes. A masking lacquer was applied to the grip end to shield it

from the electrolyte, and the area exposed in the gauge section was kept constant to

 

3. Even though slight oxidation may occur during the annealing process, the oxide layer was removed
during the electropolishing step.

52
maintain identical electropolishing current density each time. Electropolishing was

conducted in a moderately agitated bath of 33% nitric acid and 67% methanol maintained
at approximalety 228 K. The applied current density was 2.0 mA/mmz. The apparatus
for electropolishing is shown in Figure 3-4. The specimen was also rotated at 60 rpm to
prevent bubble adhesion on the surface, and the total electropolishing time for each
specimen was approximately 30 sec.4 The masking lacquer was removed from the
specimen surface and cleaned by acetone, methanol, and deionized water successively in

an ultrasonic bath.

3.2. NiTi Film Deposition by Ion Sputtering
3.2.1. Ion Beam Sputter Deposition System

A schematic drawing of the ion sputter deposition system is illustrated in Figure
3-5. The sputtering chamber, a 14inch diameter metal bell jar, was evacuated by a
rotary mechanical pump and by an Alcatel 5400 turbomolecular pump (the pumping
speed is 400 liters/sec. of air), and by an APD cryopump—SS, which pumping speed is
12001iters/sec. of air, in parallel. The chamber was baked for over 12 hours at 353 K.
After baking, the chamber was cooled and pumped for 8 to 10 hours. After these
operations, the final base pressure was obtained around 7.7x106 Pa. To monitor the
partial pressure of the residual gases in the chamber a Dycor M102 quadrupole mass

spectrometer was used.

3.2.2. NiTi Thin Film Deposition by Ian Sputtering
The sputtering target alloy, with composition of Ni5o,17Ti49,g35 ,was

commercially provided in the form of a 62 mm diameter extruded rod. The alloy was cut

 

4. Because of the presence of oxide layer on the suface of the specimen, the current and voltage increased
in the early stage of electropolishing. After removal of the oxide layer, the current and voltage stabilize and
this particular stage was believed to be the optimal electorpolishing condition. It took about 15 sec. to
electropolish the specimen after achieving the optimal condition

5. No further information of the trace element contents was available from the company.

53
and hot rolled at about 1173 K into a thin plate. The ﬁnal dimensions of the target were

approximately 85 mm diameter and 3 mm thick. The target surface was mechanically
polished up to 600 grit silicon carbide paper and cleaned in successive bathes of acetone,
methanol, and deionized water.

The ion sputter deposition system was equipped with two different ion guns. One
was used for sputter-cleaning and the other was used for deposition of NiTi films on
specimens. High purity argon gas (99.999% min.) was supplied to both ion guns through
a gas purifying cartridge (OxiClear DGP-250 R1) to remove trace oxygen and moisture.
A special steel fixture was designed to facilitate successive sputter-cleaning and
depositing on four fatigue specimens in one batch (Fig. 3-6). Initially, the specimens
were placed in the center of the sputtering chamber, normal to the ion gun which was
used for cleaning. Prior to deposition of nickel-titanium films, the specimens were
sputter-cleaned by bombarding the substrate surface with 500 eV argon ions for 10
minutes by a Anatech 5-cm cold cathode gun, to remove dirt and the oxide layer. The
operating conditions were as follows: a beam current of 9.6 mA, acceleration current of
1.6 mA, and 150 mA of emission current. The working argon gas pressure in the
chamber was maintained approximately at 2.9x10‘2 Pa during sputter-cleaning.

After sputter-cleaning, the specimens were rotated 90° by a linear-rotary
feedthrough, to face the target, without breaking the vacuum. Ion-sputter deposition was
accomplished by irradiating the NiTi alloy target with 800 eV argon ions generated from
a 3-cm Kaufman-type gun (Ion Tech Model 3.0-1500-100), equipped with converging
optics. Even though a convergent grid was used for the ion sputter deposition, some
amount of the ion beam diverged. Therefore, a titanium aperture was used to reduce
beam coverage, and to avoid sputtering undesired elements, such as Fe and Cr, from the
chamber wall and other fixtures in the chamber. The incident angle of the ion beam on
the target was 45° and the gun was operated with a discharge voltage of 55 V, a beam

current of 40-55 mA, and an acceleration current of 1-3 mA, supplied by an MPS-3000

54
FC power supply. During deposition the working chamber pressure was maintained at

3.9x 10‘3 Pa. To compensate for the preferential resputtering of titanium at the substrate,
which was described in section 2.5.3, 0.25 mm diameter titanium wires (99.8 % purity)
were placed on the target to control the composition of the ﬁlms. The average effective
composition of the target surface was changed to around Ni47Ti53. The distance between
target and substrate was 130 mm and the distance from the ion source to the center of the
target was 70 mm. Total deposition time to produce the desired 3 pm film thickness was
approximately 5 hours.

For direct comparison of crack development between the coated and uncoated
surface areas on the fatigued specimens, the specimens were partially masked with
copper foil in the gauge area. An additional set of specimens, which were used for
evaluating fatigue life, were made with films in a shallow notched area. Simultaneously,
ﬁlms were deposited on slide glass substrates for thickness measurements, resistivity

measurement, and composition analysis.

3.2.3. Crystallization of NiTi Thin Films

Crystallization temperatures of as-sputtered ﬁlms were studied using a DuPont
910 differential scanning caloriometer (DSC). Specimens, which had been deposited on
glass, were loaded in sealed aluminum pans and heated at a constant heating rate of 10 K
per minute from ambient temperature to 873 K. The crystallization temperature was
determined by the peak temperature from the exotherrn in the DSC trace.

Crystallization of as-sputtered films on fatigue specimens was performed in
vacuum with two 625 watt quartz-halogen lamps at a distance of 25 mm from the surface
of the specimens for 10 minute at 873 K with a heating rate of around 20 K per minute.
The oxygen partial pressure was kept below 5.0x10‘7 Pa and the nitrogen partial pressure
was maintained near 7.0)(10‘6 Pa during the crystallization procedure. The specimens

were cooled at an initial rate of 50 K per min. The time-temperature curve for the

55
crystallization annealing is shown in Figure 3-7. Crystallization of the films were

conﬁrmed by X-ray diffraction analysis on a Scintag 2000 diffractometer.

3.3. Characterization of NiTi Films
3.3.1. Thickness Measurement

Films were deposited simultaneously on microscope slide glass substrates, which
were placed next to the fatigue specimens, for thickness measurements. A profile step
was produced by masking with an annealed copper foil during deposition. The thickness
of the as-deposited film was measured by Dektak II profilometer with a scan rate of

1mm/min. The final thickness of deposited ﬁlms was in the range of 2.9 to 3.2 um, and

the deposition rate was approximately 100A per minute.

3.3.2. Composition Analysis

Elemental compositions of the ﬁlms were determined by energy-dispersive X-ray
(EDS) microanalysis by Link analtical AN-1000 with pentafet detector on a Hitachi S—
2500 scanning electron microscope (SEM). Standards of pure nickel, pure titanium, and
Ni 50, 1Ti49,9, which was provided by a vendor, were chosen to determine the composition
of the films. All standard samples were polished using up to a 0.3 pm abrasive alumina
powder to obtain a mirror-like surface and then cleaned by acetone and methanol. The
ﬁlms deposited on the glass surveillance substrate were mounted on a graphite stage,
together the with standards, and were illuminated by the electron beam with an energy of
20 keV. After stabilization of the SEM LaB6 ﬁlament6, the spectra from the films were
checked for any existence of characteristic Si X-ray peaks from the glass substrate.
Absence of these peaks conﬁrmed that no interaction occuned between the electron beam

and the glass substrate. All the spectra were collected with a working distance of 15 mm

 

6. The stablization of the ﬁlament (electron source) took approximately 30 min. and could be decided by
comparing the dead time percentage and time interval without changing any other conditions with the
previous spectrum. If the ﬁlament reaches a stable condition, those values become exactly the same as the
previous ones.

56

and magnification of 1,000x, without tilting of the SEM stage. The composition was

calculated by using the ZAF-4 program supplied by Link Analytic Inc. to analyze the all

spectra.

3.3.3. Transformation Temperature Measurement

3.3.3.1. Eletrical Resistivity Measurements The transformation
temperatures of the annealed films were studied by recording electrical resistivity as a
fuction of temperature for the surveillance films deposited on glass substrates. The
resistivity measurement apparatus is shown in Figure 3-8. Two independent power
sources were used for the measurement: one (12 V d.c. battery) for supplying a constant
current to the film and the other (HP 6010A d.c. power supply) for controlling the
heating rate. Thin-gauge copper wires and a T-type thermocouple were attached by
pressing pure indium onto the film. A constant direct current (4 mA) was passed
through the circuit, while the voltage drop (AV) across a 4 mm long film strip was
recorded on an XY-recorder. Both the cooling and heating rates for the resistivity scans

were held at approximately 10 K/min.

3.3.3.2. Differential Scanning Caloriometry (DSC) Phase
transformation temperatures and enthalpies of the annealed ﬁlms were determined on a
DuPont 910 DSC with heating rate of 10 K/min. The baseline curve of DSC is the
continuous change in the heat ﬂow that a material undergoes without transformations.
The actual curves are the deviation in the heat ﬂow that a material undegoes with
transformations. The transformation start and finish temperatures are determined by
superimposing the actual curves upon the baseline curve, and using the points of
deviation that the actual curves exhibit from the baseline curve. The transformation start
temperature is where the peak begins to deviate form the baseline, while the finish

temperature is where the peak return to the baseline.

57
3.3.4. Crystal Structure Evaluation

A Scintag 2000 diffractometer with copper X-ray tube was used to investigate the
crystal structure of the deposited ﬁlms before and after annealing. A sample holder for
the 5 inch dog bone fatigue specimens was designed for mounting on the X-ray sample
stage properly. An operating voltage of 40 kV and tube current of 30 mA were chosen,
and both step and continuous scanning modes were employed. For step scans, the
samples were irradiated by Cu-Kor radiation with a step size of 003° and a pretime of 5
sec. For continuous scans, the samples were irradiated with a step size of 003° and a
scan rate of 3°lmin. Structure changes resulting from stress-induced transformations
were studied on fatigued specimens after plastic deformation and fatigue cycling.

A Rigaku RU-200B X-ray diffractometer equipped with a Rigaku CN 235181
cold stage was used to study the crystal structure of the isothermally annealed NiTi films
at various temperature from 348 K to 173 K. An operating voltage of 45 kV and tube
current of 100 mA were chosen, and a continuous scanning mode was employed.
Samples were irradiated by Cu-Kor radiation, with a step size of 003° and a scan rate of
2°lmin. from 2-theta angles of 35° to 48°. The whole cold stage resided in a sealed
vacuum camber. Sample temperature was controlled by a built-in heater mounted on a
cold finger, which was connected to a liquid nitrogen dewar. The temperature ﬂuctuation

during scanning was maintained within :1: l K.

3.4. Low cycling Fatigue Test

All low cyclic fatigue tests were undertaken at a constant plastic strain amplitude
using sinusoidal fully reversed uniaxial loading, in a computer-controlled servohydraulic
testing machine. The computerized control and data acquisition system for the low cycle
fatigue testing program (CPS-405) written in the BASIC language were developed by Dr.

Grummon. The schematic block diagram of servohydraulic test and cooling system is

58
shown in Figure 3-9. Detailed procedures for fatigue testing will be described in the

following section.

3.4.1. General Features of the Fatigue Testing System.

The low cycle fatigue testing was conducted in a closed looped servohydraulic
system (Instron 2150 series) with a 20,000 pound load cell (Instron 6467-107) and a 10
mm gauge length extensometer of :1: 1.0 mm full scale (Instron 2620-530). The gripping
system consisted of a set of collet grips (Instron 2742 series) for round specimens and on
an electric pump pressurization system.

A cooling system was employed for low temperature fatigue tests. A detailed
schematic is illustrated in Figure 3-10. The coolant used was 50 % ethylene glycol with
water and was cooled by Neslab refrigerated circulating bath aided by a cryocool
immersion cooler (Neslab CC-lOOII). The lowest temperature obtained in the bath was
around 203 K. The specimen was conductively cooled by aluminum cooling blocks
through which coolant was made to circulate. The hoses, which carried the coolant, were
insulated to minimize thermal losses during circulation and the aluminum blocks were
also covered by insulation to prevent heat dissipation to the grips. The specimen and
cooling blocks were isolated from the external atmosphere by wrapping the entire
assembly with insulating materials. Liquid nitrogen, if needed, was blown to the gauge
area to lower the temperature for providing fully martensitic NiTi films.

The gauge section area was calculated according to the formula given below.

r2
A=Jv2 ——(9—sin 0)
2
where (3-1)

_ Jrz -(t-r)2
fi==Tan1 (t-r)

 

59
where A is the cross section area, r is the radius of the cross section if there is no ﬂat face,

and t is the distance between the ﬂat face and the opposite curved part of the cross
section. The program, CPS 405, provided a constant plastic strain amplitude control and
converted the measured load from the load feedback to stress based on the calculated
gauge section area. The measured stress and strain were displayed graphically on the
monitor during the test, and the peak stresses for each block were stored continuously.
Testing could be terminated or interrupted by various criteria such as a specified amount

of the peak stress drop, a predetermined total plastic strain target, or number of cycles.

3.4.2. Specimen Mounting

Since the annealed copper specimens, irrespective of the surface condition (coated
or uncoated) were extremely soft, special caution was needed for mounting the
specimens. Perfect alignment of the grips had to be maintained to prevent imparting any
plastic deformation during this procedure. Initially, the specimen was inserted into the
lower grip. The extensometer with the installed gauge length pin was mounted on the
specimen by rubber bands. The hydraulic pump and the actuator were turned on under
load control mode and then the ram was moved up very slowly. After the specimen was
positioned in the upper grip, the upper grip was clamped. Prior to testing, the initial load
was maintained at zero when the specimen was clamped, which minimized loading on
the specimen before the beginning of the test. While the system was still in load control,
the gauge length pin was removed from the extensometer, and both the strain feedback
and the strain command were zeroed under zero load. After finishing the above
procedures, the testing system was transferred from load control to strain control. All
low cycle fatigue tests in these studies were undertaken under strain control mode, with
drive amplitudes continuously adjusted by the computer to achieve constant plastic strain

amplitude.

60
3.4.4. Plastic Strain Controlled Low Cycling Fatigue Tests

Figure 3-11 illustrates the cyclic response of a material subjected to repeated
loading in the form of a stress-strain hysteresis loop. The total width of the loop is Aer

(total strain range) and the total height is A0 (total stress range). The total strain is the

sum of the elastic and plastic strains as shown below.

Aer - A8, + Asp

so, Asp - Aer - A8, (3.2)
_ MT _ A1:
A: E

where, AP is the load amplitude corresponding to elastic region, A is the gauge section
area, and E is the Young's modulus of specimen. The plastic strain amplitude can be
determined from the hysteresis loop by measuring the distance of the two parallel straight
lines corresponding to the elastic responses between the compressive and tensile side. In
the present work, the plastic strain amplitude was determined from the strain differences

between the two sides at zero stress7. This is expressed int he following,

Ar: = 8 (3.3)

p 0-0+ " 0:0-
The plastic strain amplitude of interest of every cycle was compared with the pre-
determined plastic strain amplitude and was controlled via computer. Real time stress
strain hysteresis loops were displayed during the fatigue test.

To minimize the overshooting or undershooting from a predetermined plastic

strain target (e.g. A8,, = 0.2 %) in the early stage of the fatigue test the first 10 cycles

 

7. The actual plastic strain amplitude might be little higher than the target due to Baushinger effect, but the
deviation was negligible.

61
were completed manually. The peak stresses and the cycle numbers were recorded for

every cycle. After the ﬁrst 10 cycles the specimens were fatigued under a plastic strain
control at A8,, = 0.2 % using a constant frequency of 2 Hz which was controlled by a
computer using the 12 bit analog-digital converters. The peak stress was recorded at the
end of every block, which consisted of approximately 25 cycles and 1,000 data points,
and was compared with that of the previous block. This comparison of the peak stresses
was used for the criterion of fatigue crack initiation.

The cyclic hardening curve was constructed from the peak stresses as a function
of cycle number. Figure 3-12 illustrates an example of the cyclic hardening curve. The
saturation stresses in the present study were determined from the cyclic stress strain data.
The fatigue crack initiation life was established as the point where the peak stress
dropped by 1%, as shown in Figure 3-13, as detected by the computer controlled system.

The accumulated plastic strain (eAcc) is deﬁned as follows,

II
8,406 = 22118;, (3.4)

The fatigue test was interrupted, if necessary, to monitor surface morphology
changes whenever the accumulated plastic strain reached the predetermined value, such
as SAcc = 4, 10, 20, and 40. Damage accumulation at increasing cumulative strains were
compared between bare copper and NiTi ﬁlm coated specimens.

Other sets of the NiTi film coated specimens were removed from the load frame
after a specified number of cycles (10,000 cycles). Then, the specimens were annealed
by heating rapidly up to 673 K, by quartz lamps, and held for 20 min. in vacuum, and

were then fatigued again at 298 K. Effects of post annealing8 of films which had

 

8. It is important to raise the critical stress for slip in NiTi alloys in order to stabilize the superelasticity
characteristics against cyclic deformation. The post annealing may cause a thermal rearrangement of the
dislocations, and precipitation hardening in coated ﬁlm raising the critical stress for slip.

62
experienced cyclic deformation less than that needed for fatigue crack initiation were

studied.

Stress-strain behavior in NiTi alloy, as described in section 2.4, is different
depending on its transformation temperature. Most of the fatigue tests were carried out at
298 K, but the relative temperature with respect to the martensitic transformation
temperature of coated ﬁlms was different. Therefore, the ﬁlm on fatigue specimens may
exhibit fully austenite, R-phase, martensite, or dual phase structures (mixture of austenite
and R-phase or mixture of R-phase and martensite). However, since the Mf temperatures
of the produced NiTi films were always lower than 298 K, fully martensitic ﬁlms could
be obtained only at low temperatures. In this case, the films were cooled down below the
Mf temperature by blowing liquid nitrogen on the gauge section area, while circulating
the coolant into the cooling blocks, which is shown in Figure 3-10. Once the desired
temperature had been reached, liquid nitrogen cooling was stopped, but the temperature
was maintained low enough so as not to transform the film back to austenite by

continuously circulating the coolant during the fatigue test.

3.5. Observation of Sputtered Film Structure and Surface Damage Accumulation

Surface characterization was carried out by optical microscopy and scanning
electron microscopy (SEM). The cross sectional areas of both as-sputtered and
crystallization annealed ﬁlms grown on glass were observed by SEM (Hitachi S-2500C).
The surface morphological differences between the B2 phase and martensite phase on the
crystallized thin ﬁlms having different thermal histories were observed through an optical
microscope at 298 K.

Direct comparison of fatigue damage on the surface region in the gauge section
was performed between the area covered with the ﬁlm and the uncoated (masked) copper
area with the same cyclic history. The progress of surface damage accumulation was

recorded by interrupting the tests, recording surface features, and applying more cyclic

63
deformation. Macroscopic characteristics such as crack populations and surface

rumpling were studied by optical microscopy and SEM. Fatigue crack development on
the specimens and the morphological differences between surface film coated specimens
and bare copper specimens were also studied at higher magnification. Furthermore,
careful observations were canied out on the sputter deposited specimens to understand
the crack initiation modes and crack propagation corresponding to the crystal structures

of the ﬁlms and the associated testing temperatures.

4. RESULTS AND DISCUSSION

Introduction

This chapter is composed of two parts: (1) a discussion of NiTi thin films
processed by ion beam sputter deposition and their characterization, and (2) a treatment
of the effect of the thermoelastic NiTi thin film coatings on surface damage accumulation
and fatigue crack initiation in polycrystalline copper. The microstructures of as-sputtered
and crystallized NiTi thin films will be described brieﬂy. The transformation behavior of
the annealed films, which were used to determine test temperatures, will be reported. The
second part will be concemed with the fatigue properties in relation to the transformation
temperatures of the films. The fatigue crack initiation life of NiTi ﬁlm-coated specimens
cycled, at various temperatures, will be contrasted to that of a bare copper specimen.
Also, a direct comparison of surface damage near the boundary between NiTi thin film

coated and uncoated regions of the specimen will be presented and interpreted.

4.1. NiTi Thin Films Fabricated by Ion Beam Sputter Deposition
4.1.1. Composition

The compositions of the thin films fabricated by ion sputter deposition are
tabulated in Table 41. The chemical compositional data were collected four times from
the same vertical and horizontal emission angles for each run and variations were
monitored for every run. The composition of all of the ﬁlms was different from that of
the target, whose average effective composition after placing Ti wires on the target
surface was Ni47Ti53. The titanium content of the deposited films, with a range of 48.1-

52.7at. Ti%, was generally lower than that of the composition of the target surface. The

64

65
depletion of Ti was attributed to the preferential resputtering of Ti from the growing film

(Lee, 1994). In NiTi amorphous thin film studies using ion beam assisted deposition
(IBAD) (Chang, 1993), the sputtering yield results showed that YT, > YNi and the yield
ratio of titanium (YT, / YNi) was increased with decreasing ion energy in amorphous
ﬁlms (YT, / YNi = 1.75 at 500 eV). In addition, considering the sticking coefficient
related to relative atomic weight (MT, 2 47.90 and MN, = 58.71), which was reviewed in
section 2.5.3, the sticking coefficient for titanium will be lower due to the low atomic
weight of titanium. Thus, the low sticking coefficient for titanium may lead to lower
titanium content in the deposited film compared to that of the target NiTi alloy.
Composition changes depending on vertical and horizontal emission angles were
studied. The schematic diagram of the experimental setup used in this study is shown in
Figure 4-1. The angle of ion incidence was always 45° to the target normal and the
substrates were located at different emission angles, as described in section 3.2. The
composition of the films was apparently affected by only the vertical emission angles.
No composition changes were observed in the difference of horizontal emission angle.
The variations of composition with emission angle are illustrated in Figure 42. Ni
concentration decreased monotonically with increases in the emission angle in ranges of
25° to 60°. When alloys or compounds are sputtered, the angular distribution of the
sputtered atoms can be different for each component. In contrast to other sputter
deposition processes, ion beam sputter deposition has the directionality of the beam, so
one can change the incident angle of the beam on the target as well as the angle of
deposition to the substrate. Therefore, when films are sputter deposited from the
multicomponent alloy target, the composition can vary depending with where the target
and substrate are positioned. The change of the titanium concentration on emission angle
can be explained by differences in atomic weight and surface binding energy (ET, : 4.85
eV and EN, : 4.44 eV) between titanium and nickel, which was discussed in section 2.5.3

(Okutani et al., 1980; Oechsner, 1975). Since titanium has a lower atomic weight and

66
higher binding energy than nickel, the atomic weight effect leads to a preferred emission

of nickel from an ion bombarded NiTi alloy in the direction normal to the target. Also,
the effect of binding energy on sputtering yield causes same preferred emission of nickel
from an ion bombarded NiTi alloy in the direction normal to the target.

EDS study of the films showed a small amount of residual argon in both as-
sputtered and annealed films , as shown in Figure 43. No change in the intensity of the
characteristic peak for argon was observed after the annealing process. The effect of

trapped argon gas in the ﬁlm is not yet understood.

4.1.2. Microstructure

The cross sections and surface morphologies of both as-sputtered and annealed
NiTi ﬁlms on a glass substrate, as observed by SEM after fracture by bending, are shown
in Figure 44. The thickness of deposited ﬁlms was approximately 3 pm. Unlike the r.f.
magnetron sputter deposited NiTi film by Ishida et al. (1992)1, there is no evidence to
indicate formation of zone structures in the ﬁlms, which was described in section 2.5.2.
This may be attributed to the very low argon gas pressure (3.9 x 10'3 Pa) used during ion
beam sputter deposition, compared to typical r.f. or d.c. magnetron sputter deposition. A
very low argon working gas pressure increases the sputtered atom mean free path,
resulting in less collisional scattering, and thus the smoothness of the coating surface
increases (Thornton and Hoffman, 1989). In addition, the low operating pressure results
in an increased energetic particle bombardment of the substrate, and leads to increased
surface mobilities of adatoms on the surface (Thornton, 1974), and hence the density and
smoothness of the growing ﬁlm increase. Also, the fractured surface of the as-sputtered

amorphous film exhibits a more brittle nature than that of the annealed film.

 

1. In their study, Ar working gas pressures were from 0.67 Pa to 13.3 Pa and the resultant ﬁlms revealed
columnar structure with pores.

67
4.1 .3. Crystallization

Since nickel-titanium is an easily amorphized alloy systemz, the thin ﬁlms formed
on cold substrates by ion sputtering deposition were amorphous. The driving force for
crystallization is the free energy difference between the amorphous and crystalline
phases.

The crystallization behavior of the ion sputter deposited NiTi thin films was
studied by differential scanning calorimetry (DSC). In the present work, a heating rate
was chosen to quantify a reference annealing crystallization temperature that would
ensure full crystallization for the deposited film on copper fatigue specimens. No further
study relating to the heating rate has been done as a part of this work. Figure 45 shows a
DSC scanning trace for an as-sputtered film which was heated from room temperature to
823 K with a heating rate of 10 °K/min. An exothermic peak appeared at around 773 K,
which is identiﬁed as the crystallization temperature of the film.

Two crystallization peaks were observed in the DSC scans. This is comparable to
other observations of the crystallization of NiTi films (Busch et al., 1990) and
Tix(NiCu)1-x ﬁlms (Chang, 1993), where a small side peak was observed along with a
strong crystallization peak. It was reported that the disordered BCC structure was formed
during crystallization (Busch et al., 1989). Such double crystallization peaks have been
commonly observed in crystallization of metallic glasses, such as Ni 532r42 and
Fe40Ni4oBzo (Cahn, 1983). It has also been reported that small amounts of dissolved
oxygen in Cu-Zr glasses caused a transition from single stage to double stage
crystallization (Chevalier et al., 1983). Thus, the role of impurities on the side peak may

not be excluded. However, the nature of the second peak has not been clearly identiﬁed.

 

2. The stability of the amorphous phase of NIT i alloy has been attibuted to the small free energy gap
between amorphous and crystalline phases, a large negative heat of mixing (- 30 kJ/mol) in the amorphous
phase, and abnormally fast diffusion of nickel in Ti (Zoltzer and Bormann, 1988). Schwarz et al. (1985)
prepared amorphous NiTi alloy powders by mechanical alloying the elemental powders in the concentration

range between 28 and 72 at.% Ti.

68
X-ray diffraction studies were carried out on as-deposited and post-annealed films

to investigate the crystal structure changes. The as-sputtered films exhibited a broad
diffraction peak at approximately 20 = 42° which was located near the strongest peak of
(110)}32 of NiTi alloy, as shown in Figure 46. This broadened peak is in agreement with
previous reports for the amorphous phase of NiTi alloy (Kim et al., 1986). The
amorphous structure was marked by the absence of long range order and characterized by
a diffuse X-ray diffraction peak. X-ray diffraction scans of crystallization annealed NiTi
films are given in Figure 46(b). The largest peak at approximately 20 = 42.5° appeared
along with two smaller peaks at 20 = 62.0° and 20 = 78.2°, which correspond to (110)132,
(200)132 , and (211)32 respectively, in NiTi.3

The resistivity measurement of an as-sputtered film in the range of 123 K to 373
K is shown in Figure 4-7. The resistivty curves did not exhibit any transformational
characteristics, as described in section 2.3.2.

The grain size and grain morphology of annealed NiTi films are shown in Figure
48. This SEM micrograph exhibits artifacts apparently associated with radial growth on
both copper and glass substrates, which is believed to result from the growth of pancake-
shaped microcrystallites. Therefore, such crystallization occurs homogeneously and
proceeds until grain impingement. Similar large grain crystalline structures were reported
for CuTi liquid quenched amorphous alloys (Marshall et al., 1984). The application of
energy dispersive X-ray (EDS) analysis allowed the measurement of the chemical
composition of the center of the grain and the radially grown arms. No differences in
chemical composition between the two areas were detected. In addition, X-ray
diffraction results also showed only a single phase of B2 structure with the same

composition, which indicates the crystallization process is a polymorphic reaction4.

 

3. Even though TizNi and Ti3Ni4 precipitates were observed in the crytsalline ﬁlm due to a narrow solid
solution region in this alloy system, a supersaturated single phase solid solution can be formed.

4. There is no change in composition during crystallization and the crystal shape depends only on the
crystallographic anisotropy of growth rate (Cahn, 1983).

69

Planar-equiaxed pancake grains, 3 to 5 pm in diameter, were observed both on
glass and copper substrates after the crystallization process. The average grain size was
larger than the thickness of the f ilms., and was larger than those observed in other studies
(Chang, 1993; Busch et al. 1991). According to their results, crystals of NiTi were
produced from free standing films annealed at 773 K and the average grain sizes were 1-2
um, which was independent of film thickness. For eutectic or polymorphic
crystallization, once a stable nucleus has formed, the crystal growth will be linear with
time during isothermal annealing (Koster et al., 1984). Therefore, the large grained
structure can be predicted from an Arrhenius dependence of growth rate on time and

temperature for polymorphic crystallization. The simple expression of this equation is:

u = a() v0 exp (-AEa/kT) (4.1)

where u is crystal growth rate, a0 is the distance across the crystallization front, v0 is the
atomic jump frequency, and AEa is the activation energy for an atom to leave the
amorphous to crystal interface. In the present work, the crystallization temperature was
100 K higher than the previous investigations for free standing films (773 K). Also, the
large grains were attributed to slow cooling after the crystallization process, due to a
relatively large thermal mass of the substrate. In addition, it has been reported that in situ
observation of crystallization form the amorphous NiTi thin ﬁlm involves dramatic grain
growth, with some grains growing in excess of 10 um (Moberly et al., 1992). Moberly et
al. attributed such rapid grain growth to the difﬁculty of nucleation, such that the grains
could undergo extensive growth in the two dimensional ﬁlm before impinging other
grains.

Other explanations for the origins of the large grained ﬁlm can be attributed to the
defects in the deposited ﬁlm. The presence of point defects generated during ion beam

deposition, which is a well known consequence of the bombardment of the solid with

 

 

70
energetic particles (Smidt, 1990), can lead to abnormally large grains in the film. Since

ion beam sputter deposition operates at higher energy conditions than the usual
magnetron sputtering deposition, the induced point defect density will be higher. Thus,

such a high point defect density can enhance grain growth by facilitating grain boundary
mobilities in the film (Thompson, 1990).

4, 1 -4- Surface Relief

Surface relief from the retained martensite could be observed in the Ni43,3Ti51,2
film (Af = 322 K and Ms = 286 K) at room temperature, which is shown in Figure 49.
The films, deposited on both glass and copper substrates, were quenched in liquid
nitrogen and then warmed back to room temperature. This surface relief phenomenon
was conﬁrmed by observing the same area of the film on the copper substrate from
different thermal conditions, as shown in Figure 4-10. This is the first direct observation
0f the surface relief effect due to the formation of the martensite phase in the thin film
form of the NiTi alloy. The surface upheaval on the film due to the martensite phase
f0"nation disappeared when the film was heated above its Af temperature. Successive
°bServations showed that exactly the same shape of martensite variants reappeared and
disappeared repeatedly. The surface features of thin films produced in the present study
Show ed neither the triangular morphology of B 19' martensite nor the saw tooth
morphology of R-phase reported by others for bulk alloys, as described in section 2.3.2.
Instead, the martensite developed in random directions on the thin ﬁlm. These random
mar"—ensite variants might be the result of thin ﬁlm constraint effects and residual stresses,
Whi Ch would form during cooling due to the difference in thermal expansion coefficient
bet"Veen the ﬁlm and substrate. These combined effects will inﬂuence the formation of

favotable martensite variants locally on the film, but the overall distribution of the

martensite variants seems to be random.

71
4.1.5. Residual Stress

The existence of large intrinsic and extrinsic stresses in thin films have been the
subject of numerous investigations because of their importance in the application of thin
ﬁlms. Unlike thermally evaporated metal films, which are often in a state of tension, thin
films grown by ion beam sputter deposition tend to possess high compressive intrinsic
stresses due to the atomic peening effect. In other words, these compressive intrinsic
stresses can be introduced by the bombardment of energetic reﬂected neutrals which
cause forward sputtering and recoil damage to the growing film. Thus, the residual stress
Of an as-sputtered, unannealed, film in the present study is expected to be compressive.5
In the as-sputtered condition, the films had negligible extrinsic thermal stresses
because of a relatively low substrate temperature used during the sputter deposition
Process. For crystallized films, the intrinsic stresses would be relaxed after crystallization
annealing at 873 K, and the extrinsic stresses would form during cooling due to the
difference in thermal expansion between the film and substrate. For a thin film with a
thickness less than 104 times the substrate thickness, plastic ﬂow in the substrate can
generally be neglected (Thornton and Hoffman, 1989) and the thermal stress induced in

the film by the substrate is given in a one dimensional approximation (neglecting the

poisson ratio) by:
Orh=Ef(Of'Gs)(ra'an) (42)

where Ef is the Young's modulus for the film, (If and as are the average coefficients of
therl‘nal expansion for the film and the substrate. Ta and Tm represent the annealing
tenrlI>erature and the measuring temperature (room temperature) respectively. In the

pl‘esent work, the crystallization annealing temperature of the films was 873 K. The
\

5‘ The residual stress calculations were not carried out in the present study, but it was reported that
compressive stress (approximately -60 MPa) was developed in as-sputtered NiTi ﬁlm produced by ion
beam sputtering deposition (W alles, 1993). In this study, the deposition conditions was same as the present
Work and 0.3Sum thick ﬁlm deposited on Si single crystal was used.

72
themal expansion coefficient of bulk c0pper is 17x10'6/K (Metals Handbook, 1979).
The thermal expansion coefficients of austenite NiTi is 11x1045/K, and that of martensite
NiTi is 6.6x10‘6/K (Jackson et al., 1972). The Young's moduli of the austenite and
martensite phase in NiTi thin film alloys were measured to be 42.5 GPa and 22.5 GPa
respectively (Hou, 1995). From the calculation using Eq. 4.3, if the annealed film
possesses only 32 austenite or only B 19' martensite phasees, the residual stresses will be
approximately -147 MPa and -78 MPa respectively at room temperature (298 K).

Therefore, the residual stress resulting from the annealing process leads to compressive

stresses on the surface film (B2 or 819' phase). However, the actual residual

compressive stress was expected to be slightly lower than this due to stress relaxation

during cooling.

4 - I - 6- Transformation Behavior

Transformation temperatures varied with the compositions of the films. Figure 4-
1 1 provides the idea of transformation sequences relating to composition of the films.
The Ni50.9Ti49.1 film showed only R-phase transformation. The martensitic
trans f ormation showed an increase in titanium concentration. The highest Ms and Mf
temperature values were observed for Ni43,3Ti 51.2 film. These transformation
temperatures decreased when the compositional deviation from 51 at.% Ti was increased,
and Similar results were obtained from continued runs. This behavior is different from
bulk materials, where near-stoichiometric NiTi alloys showed the highest transition
temperatures. In addition, the transformation temperatures of thin films were generally
much lower than those expected of bulk alloys with similar composition. It is believed
that oxygen contamination during film deposition may have affected the transformation
beflavor signiﬁcantly. Most oxygen is bound in Ti4Ni20, which was reviewed in section

2'3 ~2, and results in a decrease in the Ti concentration of the film matrix.

73

Transformation temperatures for various films used for surface modification of

fatigue specimens are presented in Table 4-2 through 4—8. All data in this table were

obtained from contemporaneously deposited surveillance specimens having identical
deposition and vacuum annealing conditions given to the actual fatigue specimen.

Diffusion between the films and copper substrates was investigated, since
transformation temperatures of NiTi alloys are sensitive to composition variation. Figure

4- 12 shows the concentration profiles of nickel, titanium, and copper at the interface
between Ni48.1Ti51,9 thin films and copper substrate before and after annealing. For as-
Sputtered film on a copper substrate, the concentration profile across the interface was
Very sharp, indicating that no interdiffusion occurred during deposition. For annealed
films on copper substrates, the concentration profiles for each element became smoother
at the interface. However, the diffusion layers of both the copper substrate and NiTi
films were so limited that the effect of composition change on the transformation
behavior of the overall surface film due to diffusion is negligible.

In summary, composition analysis indicated that titanium was depleted in the
sputter-deposited films. This was produced primarily by preferential resputtering of Ti
from the growing ﬁlm and a low sticking coefficient of Ti. The deposited films showed a
dense structure, which was attributed to a very low working gas pressure. As-deposited
films were amorphous in structure and the crystallization process which followed
provided their thermoelastic transformation properties. Diffusion between the films and
copper substrates, during crystallization, was negligible. Most crystallized films
underwent two well separated austenite 0 R-phase 0 monoclinic martensite

“31'le orinations, and surface relief was observed due to the formation of martensite on the
ﬁlms

74
4.2. Fatigue Experiments

4.2.1. Specimen Characteristics

As was indicated in Chapter 3, two different kinds of fatigue specimens were
fabricated. Specimens with a flat surface on the gauge section were used for direct
observation of crack development on the surface, and investigation of changes in crystal
structure by X-ray diffractometry. Four different kinds of film-coated specimens were
made on the flat surface in the gauge area of copper fatigue specimens. The specimens
were designated as Fl through F4 depending on their transformation behavior and
composition of the coated films. The transformation temperatures and compositions of
the films are given in Table 4—2, and the resistivity curves are provided in Appendix.
Shallow notched specimens with various surface coatings and transformation
tern Peratures were studied with respect to fatigue crack initiation life and surface damage
accumulation rate. The transformation temperatures of the films were determined by
ele(Strical resistivity measurement, which are given in Appendix. The film types are
claSSified into three different categories depending on the characteristics of the coated
filmS- A/R series films revealed only 32 (austenite) 9 R-phase transformation during
themed cycling between 373 K and liquid nitrogen temperature. In A/TE type films the
Af temperature was lower than 295 K and thus only the austenite phase existed at room
temperature (298 K). A/SIM type films represent the film with an Af temperature was
higher 295 K. Both A/TE type film and A/SIM type film exhibited 32 (austenite) 9 R-

phase 0 319' (martensite) transformations in resistivity measurements. The number

neXt to the film types indicates their relative Af temperature, namely, the higher number
means a higher Af for that film type. The fatigue crack initiation life and transformation
temperatures for the various films are listed in Table 43 through 4-9, including bare
copper specimens and amorphous film coated specimens. In the present study, the test
‘emperature (Td) was generally fixed at 298 K and the Af temperature of the films

determined the expected transformational behavior during fatigue.

75
4.2.2. Cyclic Stress Strain Hysteresis

No significant difference in the stress strain hysteresis loops between surface

modified and the unmodified specimens was observed. Figure 4—13 illustrates a typical
example of the hysteresis loops recorded in the first 10 cycles of annealed polycrystalline
copper. The yield stress of the specimen in the first cycle was very low (approximately
20 ma) and increased with further cycles. During cyclic hardening, the stress amplitude
and strain amplitude increased continuously. After reaching the saturation condition, the
hysteresis loops maintained an essentially constant shape until the secondary cyclic
hardening stage occurred. On occurrence of secondary cyclic hardening, the peak

Stresses and corresponding strain amplitudes slowly increased again up to fatigue failure.

4~2 -3 - Cyclic Hardening Curve and Saturation Stress
The cyclic hardening curve was obtained from the peak stresses on each cycle and
the Corresponding accumulated plastic strain. Figure 414 shows a typical example of the
Cyclic hardening of polycrystalline copper6 under constant plastic strain amplitudes of 0.1
% (Asp = 0.2 %) and 0.15 % (Asp = 0.3 %). These experiments demonstrated that the
fati gue life was slightly shorter and the saturation stress was higher at Asp = 0.3 % than at
A“3p = 0.2 %. Fatigue crack initiation life studies were carried out only for the specimens
cycled at Asp = 0.2 % for various surface conditions. Cyclic hardening occurred rapidly
in tl‘le ﬁrst few cycles and the rate decreased gradually, then reached the saturation stage.
The behavior depicted in these results are characteristics of cyclic hardening curves
ac"'(lllired from both bare copper specimens and NiTi film coated specimens. The
Satllration condition was obtained after approximately 250 cycles and the saturation stress
was approximawa 123 MPa regardless of the surface condition of the specimens. No
Si gnificant differences were observed in the hardening rates and saturation stresses

betVVeen bare copper and NiTi film coated specimens. These results indicated that the
\

6' All specimens were annealed at 1173 K for 1 hour.

76
ﬁlm coatings, whether amorphous or crystalline, did not significantly affect the saturation

stresses or hardening rates.
Cyclic hardening curves for unmodified specimens in the present investigation
showed an initial saturation stage followed by a secondary cyclic hardening stage in
Figure 4-14. The saturation stage lasted up to approximately 7,000 cycles (ZZAep = 28)
and then secondary cyclic hardening occurred. Identical behavior for secondary cyclic

hardening was also observed for the surface film coated specimens.

4.3- Stress-Induced Martensitic Transformation in NiTi Film
The superelasticity effect results in unique stress-strain behavior in which
anelastic shear strains of several percent occur when a martensite phase is induced by
applied stress and reverted back to the original austenite phase by removing that stress.
Deformation by this mechanism is able to endow exceptionally high fatigue life under
strain controlled conditions. Therefore, surface modification with superelastically
def 01ming materials, which can provide large strains without dislocation activity, may
resist the concentrated strains which are the result of bulk plastic strain localization
dul‘ing fatigue. Evidence of formation of stress-induce martensite in the films on
cyclically deforming substrate thus is significant in the present work. The finding of
StresS-—induced martensite in the films suggests that the surface coatings, for which Af is
lower than the test temperature, are presumed to be superelastically deformed during
cyclic deformation, resulting in delaying the onset of the kind of severe, localized surface
Cla‘rrlé‘tge that normally leads to fatigue crack initiation.

Stress-induced martensitic transformation was observed in Ni 50.3Ti49.7 thin ﬁlm
alloy on plastically deformed copper specimens. Figure 4-15 illustrates the X-ray
diffraction scans, taken at 298 K, of the Ni50.3Ti49.7 thin film coated specimen.
Dirfraction scan (d) showed a broad amorphous peak from the as-deposited film.

DitTraction scan (c) was from the same specimen which had been annealed, furnace

77
cooled to 298 K, then quenched in liquid nitrogen, and allowed to warm back to room

temperature. In this scan, the largest peak was from (110)132 but the other three peaks
were from (002), (MT) and (111) of monoclinic martensite respectively. Reheating the
specimen for a few minutes to 400 K caused the martensite peaks to disappear, as shown
in the scan from (b). Diffraction scan (a) displays diffraction peaks from the stress-
induced martensite phase after the specimen was monotonically overloaded". Since the
furnace cooled specimen had only austenite peaks before deformation, the peak should be
the same as that of the diffraction scan from (b). Therefore, this observation
demonstrates the presence of substantial amount of stress-induced martensite in a NiTi
thin film on a plastically deforming metallic substrate.
Similar results were obtained from cyclically loaded specimens which were
Coated with the same film as the specimen discussed above. The specimens were
exatrlined before and after cyclic loading to determine whether martensite had been
Strass-induced by the modest strain levels associated with the low cycle fatigue test. A
SPeeirnen with an Ni50.3Ti49.7 film, fatigued for 11,755 cycles at a plastic strain
amplitude of 0.0015 (Asp = 0.3%), showed clearly the presence of (002), (11 T) and (111)
monoclinic martensite peaks along with a (110)132 and (111)Cu peaks, as shown in Figure
4‘ 1 6- The relatively weak peaks compared to the peak from the austenite phase indicates
that the martensite formation occurred over only limited areas of the specimen surface. It
is Possible, then, that it was induced only in regions which were subjected to large local
Snail”! excursions caused by underlying PSBs. Thus, it may be concluded that stress-
in<1llced martensite can form in NiTi thin films lying atop inhomogeneously deforming
metallic substrates and thus can exhibit the superelastic effect during fatigue.
DSC data for a surveillance f ilm8 is shown in Figure 417. The Ni 50_3Ti49_7 film

ChsIDlays distinct double endothems in the DSC scan on heating. The first endotherm
\

7d}, 'I‘he specimen was smashed accidently and then the gauge section area suffered a high plastic
8 fOl'mation whose precise magintude is not shown.

' '1‘.he ﬁlm was prepared by deposition on annealed copper foils placed near the gauge area of the fatigue
speclmw.

78
resulted from the martensitic transformation and the second one from the R-phase

transformation. This is consistent with the enthalpy of the R-phase transformation which

is very small (about 2—5 J/g) compared to the martensitic transformation (approximately

20 J/ g). The austenite start (As) and finish temperatures (Af) were 307 K and 345 K

respectively. The martensite start temperature (M3) was near 273 K and the martensite

f i ni sh temperatures (Mf) was below 173 K. These were determined by X-ray
di ff ractometry at various temperatures on cooling, and are given in Figure 4—18.

The critical stress for stress-induced martensite at room temperature (298 K) was
calculated using Eq. (2.6), Clausius-Clapeyron equation. To, which is the equilibrium
trans formation temperature without an external applied stress, was approximated as To =
(M3 + As)/2 = 290 K, and used a material density of 6.5 g/cm3 and a theoretical
trans formation strain of 0.1299 (Matsumoto el al., 1987). AI-I for the transformation was
estimated to be 25 J/g, where AH = AHM—oR + AHR—SA. The critical stress for stress-
indllced martensite of the film will thus be approximately 255 MPa at room temperature.
This estimated value is believed to be low enough to form stress induced martensite in
coated ﬁlm in the regions of severe local straining above underlying PSBs9.

On the other hand, it was reported that the pseudoelasticity effects were observed
in both R-phase and martensitic transformations (Miyazaki et al., 1989). R-phase
transformation exhibited smaller enthalpy, a narrower hysteresis, and lower critical stress
Val Des than the martensitic transformation (Fukuda et al., 1992). In other words, the R-

phase can be formed with a small driving force. Therefore, the stress-induced
transformation for the R-phase can occur more easily than that for martensite. Since the
R‘Phase appears prior to the martensite, the initial R-phase will be formed in plastically
def Qrmed NiTi thin film alloys by applying an external stress at a certain temperature

range, which was described in section 2.4.1. For example, if the T, temperature of the

\
%‘igonsidering that the localized strain at PSB is approximately 1% and the Young's modulus of austenite
1

film, which was used in the present study, is 42.5 GPa (Hou, 1995), the stress of NiTi ﬁlm on a PSB is
exPected to be slightly higher than 400 MPa.

79
film is near the test temperature, the shape memory and superelasticity effects of the R-

phase can be expected. However, no distinguishable R—phase peak10 was found in the X-
ray diffraction scan of a A/TEl type film coated specimen, with a T, = 284 K, after
fatigue at 298 K. Even though the stress induced R-phase can be formed in the films on
deforming copper substrate at this condition, it is very difficult to distinguish whether the
peak, which is near 20 = 42.8°, is from the B2 or R-phase, especially when the (111)Cu
peak is from the substrate“. Also, since the maximum transformational strain of the R-
phase, subjected to an external applied stress, is very small, with the maximum shear
strain of less than 0.3 % (Miyazaki and Otsuka, 1986), the stress induced R-phase would
transform to martensite again at a higher strain (~1%) during cyclic deformation. Thus,

the small peak of the R-phase may not be detected on the X-ray diffraction scan.

4.4. Surface Damage Accumulation
4.4.1 . Surface Crack Formation on Flat Surface Gauge Area of Surface Modiﬁed
Specimens

Four different kinds of ﬁlm-coated specimens were made with a flat surface in the
gauge area of copper fatigue specimens, which are give in Table 4—2. For the F1
specimen, a fully annealed copper foil was attached to the specimen which remained in
place for the deposition process, resulting in a boundary between the thin film modified
area and the completely uncoated pure copper. The specimen was examined for crack
development between the coated and uncoated areas. The other specimens had fully
covered gauge areas and were examined for crack density and crack modes. The
specimens were cycled in strain control at a constant plastic strain amplitude of 0.015

(Asp = 0.3 %) and the saturation stress was approximately 140 MPa, regardless of the

 

10. Peak splitting below Tr, associated with R-phase formation from BZ austenite phase, was reported as
(01 DR and (0-11)R at 425° and 43° respectively (Ling and Kaplow, I981).

11. Reinstallation of the 5 inch long specimen on a X-ray stage after mechanical test also caused a slight
change in the height of exposed surface resulting in the peak shift, as shown in Figure 4—19.

80
surface condition. Tests were terminated when the maximum peak stress dropped by 1

%.

4.4.1.1. Specimen F1 Viewing through an optical micrograph
showed no prominent cracks in the film coated area, while cracks and slip bands were
developed in the uncoated area, as shown in Figure 4-20. A detailed SEM micrograph of
the surface between a film coated and uncoated area of the specimen is shown in Figure
4-21. The strain amplitude and number of cycles in both areas of this specimen are
identical. Direct comparison of the fatigue damage accumulation in the surface region
was performed in the vicinity between the gauge section covering the film and the
uncoated copper area.

It is obvious that the uncoated portion of the specimen behaves much differently
than the coated part. In the unmodiﬁed region of the specimen (lower half), a significant
number of cracks and severely extensive PSB extrusions were formed throughout the
entire area, whereas the area coated with the film apparently has suffered no surface
damage, as shown in Figure 4-21 (a). With the exception of the area at the edge of the
ﬁlm-copper boundary, the morphologies shown in these micrographs are typical of the
behavior of the specimen as a whole. A detailed study of the coated portion of the
specimen revealed no detectable signs of PSB extrusion or microcrack formation. As a
whole, when the P838 encountered the coated portion of the specimen, they became fully
indistinct. It is noted that in the uncoated portion, the extrusion—intrusion bands in a
substrate grain were clearly developed but those bands were totally suppressed when they
projected into the coated zone, as shown in Figure 4—2l(b). The result of fatigue tests,
shown in this ﬁgure, are quite different from the results from the surface modiﬁcations by
ion implantation. Unlike the surface modifications by ion implantation, distinguishable
microscopic feature changes, such as the crack initiation mode changes (Morrison et al.,

1989) or slip line spatial density changes (Grummon, 1988), were not observed in the

81
coated area. The surface morphology changes with accumulated strain will be described

in section 4.4.2.

Some PSB surface activity was present in the immediate vicinity of the film
boundary, but this is considered to be an artifact associated with the masking procedures
which slightly shadowed the ﬁlm in this region, causing it to be much thinner, and also
probably contamination of the film near the boundary. Occasionally, cracks were
observed at a few areas such as near the film boundary (cracks from unmodified region)
and at some delaminated areas of the film especially at the triple points of grain
boundaries, which is marked by an arrow in Figure 4-22. But, in general, the film
adhered very ﬁrmly and no cracking occurred from those locally uncovered areas (or at
defected areas which were formed during deposition) as shown in Figure 423, where
PSBs were developed in the copper substrate. This observation showed that these ﬁlms
strongly adhered to the matrix, suppressed PSB penetration to the free surface, and

inhibited fatigue crack initiation.

4.4.1.2. Specimens F2 and F3 As was provided in Table 42,
the ﬁlms on specimens F2 and F3 exhibited a higher titanium concentration than
stoichiometric NiTi alloy. In addition, the film on specimen F3 did not show
transformation behavior in electrical resistivity measurement, as shown in Figure A-2.
The crack modes and density of the films of specimens F2 and F3 were compared and
analyzed.

An optical micrograph showed several cracks on specimen F2, as shown in Figure
4-24. Detailed crack mode and density on the specimen were studied on the SEM
micrographs in Figure 4-25. The average distance between cracks, developed
perpendicular to the loading direction, was 0.2 mm and the average crack length was
approximately 0.4 mm. Most cracks developed perpendicular to the loading axis, while a

few 'wrinkles' were observed parallel to the loading axis.

82
On the other hand, a high density of cracks was observed by an optical

micrograph on specimen F3, as shown in Figure 4—26. A detailed SEM micrograph of the
crack shapes is shown in Figure 4-27. The average distance between cracks, which
developed perpendicular to the loading direction was approximately 0.06 mm and the
average crack length was 0.6 mm. All cracks were produced in the transverse direction
of the loading axis, and the films near those cracks were severely wrinkled parallel to the
loading axis.

Almost identical features were observed on specimens F2 and F3, with only the
crack populations and modes being slightly different. All cracks were developed
perpendicular to the loading axis on both specimens F2 and F3. This result indicats that
these cracks penetrated into the copper substrate from the surface ﬁlm and propagated in
the same direction as the lateral film cracking. This is in marked contrast to cracks in
fatigued bare copper specimens, which evolved along the PSBs or grain boundaries
associated with specific crystallographic directions, as was observed in the uncoated area
of specimen F1. The high density of cracks on specimens F2 and FB is believed to be the
result of a high Ti concentration on the films, which resulted in the formation of Ti2N i
precipitates. TizNi peaks were observed from X-ray diffraction scan of the specimen F3,
as shown in Figure 4—28. The TizNi particles usually precipitated out at grain boundaries
in NiTi alloys during the annealing process (Busch et al., 1990; Chang, 1993) and caused
deleterious effect on mechanical properties. The difference in crack population between
specimen F2 and F3 may be from higher Ti concetration on F3 film, causing an increase
in its brittleness. It can be attributed to the balance between the strength of the ﬁlm and
stress on the film. These relations result in more cracks on specimen F3 than on
specimen F2, because the reduction in stress caused by cracking occurs only over a finite
distance away from the crack plane.

On the other hands, the ‘wrinkles’ were developed parallel to the loading axis on
both specimen F2 and F3. The ‘wrinkles’ which developed on the specimen F3 were

83
more frequently observed than on the specimen F‘2. This result occurred because of the

relative distance between the cracks and the crack lengths. For the specimen F3, the
distance between cracks was shorter and the average crack length was longer than that of
the specimen F2. Careful observations by SEM showed that the ‘wrinkles’ occurred near
the center of each crack; it is believed that the fatigue crack originated there. Since the
thickness of the surface film is negligible compared to that of the substrate, the stress
distribution in film is taken in two dimensions. A situation of plain stress typically exists
in a thin film loaded in the plane of the ﬁlm where there is no stress normal to the free
surface. The ‘wrinkles’ in the film might be explained by 'compressive cyclic creep' in
the ﬁlms during fatigue testing. Consider alternating compressive and tensile stresses on
the surface film along the loading axis due to cyclic deformation. While the tensile stress
in the loading axis produces an extension along that axis on the film, it also produces a
compressive strain perpendicular to that axis. The transverse strain would be balanced
with the applied stress along the loading axis before cracking. However, the balance will
be broken after developing cracks on the films, which were perpendicular to the loading
axis. Let the loading direction be the x-axis and the transverse direction be the y axis, as
shown in Figure 429. After cracks formed in the ﬁlms, the tensile stress along the x-axis
would act for crack opening and closing by alternating stresses. As a result, the stress
effect in the x-direction might be nulliﬁed between cracks and only the stress along the y-
axis would be left between them. Thus, it can be assumed that the crack propagated in
the y-direction from the origin of the crack, which is believed to be the weakest area,
during cyclic deformation. But the effect of the stress along the y-axis increases with
decrease in the distance between cracks and it also increases with an increase in the crack
length. It was considered that the residual compressive stress in the surface film, which
was described in section 4.1.5, intensified the stress effect between the cracks. With an
increasing the number of cycles, the crack length will increase, and thus the stress along

the y—direction tends to go more to the compressive side continuously, which was named

84
, . . , . . . . .
compressrve cyclic creep as shown in a schematic stress—straln response in Figure 4-29.

The wrinkles are presumed to occur from the weakest area when the compressive stress
was high enough to form. In this study, other stress effects, such as stress near crack tips,

were ignored.

4.4.1.3. Specimen F4 An optical micrograph did not show any
crack formation on the F4 specimen, as shown in Figure 4-30. The crack was found on
the opposite side of the flat surface in gauge section area. Only a large crack was
observed after faigue crack initiation, shown in Figure 4-31. It is different from the
surface coated region of the specimen F l, in which the crack was formed on the uncoated
area proir to the crack formed on the coated area. The crack on specimen F4 formed on
the coating region.

In summary, the crack population and crack size of each specimen were
signiﬁcantly different depending on the film properties related to its comnposition, and
the transformation temperatures with respect to the testing temperature (298 K). The
coated area of specimen F1 and specimen F4 did not show any prominent cracks, while
specimens F2 and F3 showed a relatively high cracking density. The relationship

between the films and crack mode will be dicussed in following section.

4.4.2. Fatigue Damage Accumulation in Shallow Round Notched Surface Modiﬁed
Specimens

Shallow round notched specimens with NiTi films were also studied after fatigue
crack initiation occurred. All specimens were cycled in strain controlled at a constant
plastic strain amplitude of Asp = 0.2 %. All tests were terminated when the maximum

peak stress was dropped by 1%, which was described in section 3.4.

85
4.4.2.1. Bare Polycrystalline Copper - Figure 432 shows several

extrusions of persistent slip bands in a polycrystalline copper specimen fatigued after 500
cycles (22Aep = 2). It is quite evident that the slip bands were formed only in specifically
favored crystallographically oriented grains dependent upon the loading direction and
compatibility with neighbor grains, especially at low cumulative strains during cyclic
deformation. This observation showed that the slip lines indicated parallel primary single
slips. The spacing of slip bands was almost equal in individual grains, but the intensities
and shapes of extrusions were different from grain to grain. It is also apparent that the
shape of PSB extrusions and the spacing of slip bands varied with orientation of the
grains in a polycrystalline copper. Figure 43 illustrates a tongue-shape extrusion
emerging at the crystal surface due to PSBs which exhibited a typical width of
approximately 1 pm.

Successive observations of the same area of a copper specimen were carried out at
various cumulative plastic strains. The specimen was reinserted in the testing frame and
fatigued for a predetermined number of cycles and further observations were carried out
using SEM. The specimen was observed after five interruptions of the fatigue test
(22Aep = 4, 10, 20, and 40) and after fatigue crack initiation. An example of such
observations is presented in Figure 4-34. The area, as a reference, was chosen to be
easily identified in continued observations. Unlike the extrusion shown in Figure 433
which exhibited generally small and discontinuous tongue-shapes, the spacing of slip
lines has become narrower and the extrusions have been altered to a continuous ribbon-
shape upon increasing the number of cycles. These observations indicate that the
extrusions with tongue-shapes were developed in only a small number of cycles and then
grew and coalesced into ribbon-shapes with an increased amount of accumulated plastic
strain. Continued observations showed that slip bands were developed in more grains
and the extrusions were intensified continuously with increasing the number of cycles.

The heights of extrusions and intrusions were larger, but the spacing was not changed any

86
further continued cycling. However, there was little change of surface morphologies

between N = 10,000 (22Aep = 40) and N1 = ~15,000, where fatigue crack initiation
occurred. Secondary slips occurred near the grain boundary in the reference area. The
above observations demonstrated that the intrusions made notches where stress
concentrations occurred and eventually fatigue cracks would initiate. It was also
observed that crack nuclei formed at the intersection of the slip bands and grain
boundaries (PSB-GB crack).

Another observation in SEM of fatigued polycrystalline copper after crack
initiation was accomplished near grain boundaries, as shown in Figure 4-35. This
observation shows the presence of secondary slip traces in the vicinity of grain
boundaries. It is believed that such secondary slip systems near the grain boundary are
activated to accommodate the strain compatibility between neighboring grains consisting
of different crystallographic orientations. This is the major difference between single
crystal and polycrystalline copper, where the secondary hardening related to secondary

(or multiple) slip occurs due to a wide variety of orientations.

4.4.2.2. Amorphous Film Coated Specimens The surface
morphology of the specimens were observed at low magniﬁcation after fatigue failure, as
shown in Figure 4-36. Several major cracks developed in these specimens, unlike
crystalline film coated specimens, which will be described in the following sections.
Fatigue cracks propagated in a straight manner perpendicular to the loading-axis and the
remaining surface area did not show any distigiushable features. Figure 437 (a) shows
that amorphous NiTi film coated specimens demonstrated a typical brittle nature, with
spalling of the ﬁlms on the substrate. It is also observed that some part of the ﬁlm was
delaminated from the copper substrate, as shown in Figure 4-37 (b). Detailed

examination of the crack shapes and directions on the film and substrate demonstrated

87
that the cracks penetrated from the surface film to the copper substrate regardless of the

crystallographic orientation.

4.4.2.3. A/SIM Film Coated Specimens The surface
morphologies, at low magnification, of the specimens A/SIMl and A/SIMS displayed a
large crack which developed perpendicular to the loading axis after fatigue failure, as
shown in Figure 4-38. The cracks developed small wavy morphologies, unlike the
amorphous film coated specimens, but the crack displayed more of a straight character in
the specimen A/SIMS than that in the specimen A/SIMl. Observations near the crack tip
area and near a large crack on the specimens were carried out at high magnification. It
was found that all specimens had developed only one major crack in the high stress
region, and the crack tip area of the specimen A/SIMl is shown in Figure 439.
However, it was revealed that very fine microcracks formed near the major crack and
crack tip area of the specimen A/SIMS, which is shown in Figure 4-39. These
microcracks also developed perpendicular to the loading axis and independent of
substrate grain orientations, as shown in Figure 440.12 But, these microcracks were not

found in other A/SIM type specimens, which exhibited longer fatigue life.

4.4.2.4. A/TE Film Coated Specimens Observation of the
surface morphology for the specimen A/TEl at low magnification, also showed that a
crack developed in the specimen. Surface rumples, also, were observed, as shown in
Figure 4—41. It was found that the crack propagated in a wavy manner, as was observed
crack tip area of the specimen A/T E1, shown in Figure 442. All crystalline ﬁlm coated
specimens showed wavy crack propagation, while the amorphous film coated specimens

displyed a straight crack propagation. However, the high degree of 'zig—zag' crack

 

12. In this micrograph, the grain boundaries, including twin boundaries, were revealed underneath of the
ﬁlm as an artifact of electropolishing.

88
propagation was observed near the crack tip area, which indicated that crack propagation

would be slower than in other specimens.

Since the servo-hydraulic grip allowed for specimen removal and reinsertion, it
was possible to monitor a specific region of the specimen surface using SEM. It was
checked whether the fatigue crack initiation life was affected by several interruptions of
the test or not. The resultant fatigue lives for both cases did not show any effect due to
the interruption of the test; the data is presented in Table 4—7. In the previous section, it
was found that cyclic hardening of specimens occurred only at the very early stage of the
fatigue tests and then reached the saturation stage, which was approximately after 250
cycles under present fatigue conditions. Thus, the evolution of surface damage
accumulation was not monitored from the cyclic hardening stage but from the saturation
stage until the fatigue crack initiated. Successive fatigue damage accumulations of
A/TE2-4 and A/TEB-l specimens were also investigated at various accumulated strains,
that is, 22Aep = 4, 10, 20, and 40, 80 and after fatigue crack initiation, as shown in Figure
4-43 for the A/TE3-1 specimen. The observations of surface damage accumulation in
A/TE2-4 and A/TE3-1 specimens disclosed very similar results. No sign of surface
cracks was found until 22Acp = 80. Unlike the unmodified copper specimens in which
damage accumulation occurred continuously, as shown in Figure 4-33, the crack
formation was substantially delayed in these specimens and then the crack developed

abruptly.

4.4.2.5. A/R Film Coated Specimens The observation of
surface morphology for specimen AIR did not show any significant difference from other
crystalline film coated specimens, which revealed a crack perpendicular to the loading

axis. SEM micrograph near crack tip area was given in Figure 4-44.

89
In summary, for both shallow round notched specimens and ﬂat gauge sectioned

specimens, all cracks were produced in the transverse direction against its loading
direction whether the surface films were crystalline or amorphous in structure. It is noted
that cracks of both amorphous and crystalline film coated specimens initiated from the
surface ﬁlm first and propagated to the substrate. The surface observation results showed
that the adhesion between substrate and film was very strong, and thus once a microcrack
was formed at the surface this cracking led to failure of the substrate. If the adhesion is
not strong enough to transfer the stress, the film will be delaminated from the substrate.
Therefore, once cracks are formed on the surface film, it is possible that the fracture of
the film generates a notch effect and facilitates the propagation of fatigue crack into the
matrix, and also that growth of the crack within the interior is due to the loss of the
protective influence of coating, thus markedly impairing fatigue life.

The crack shape was related to fatigue crack intiation life, which was given in
Table 4-3 thrrough Table 4-9. Amorphous film coated specimen showed several straight
cracks, which revealed the shortest fatigue life, while other film coated specimens
revealed a large crack at the shallow notched area. A/TE type film coated specimens
demonstrated a high degree of wavy crack propagation, which exhibited the longest
fatigue initiation life. A/SIM5 specimen dislplayed very fine 'hairy' cracks, which also

were deveolped perpendicular to the loading axis, associated with a large crack.

4.5. Fatigue Crack Initiation Life

Shallow round notched specimens were used for studying fatigue crack initiation
life. Fatigue crack initiation lives for the surface-modiﬁed specimens were evaluated
from cyclic hardening curves defined by a 1 % drop from the peak stress. The results of
fatigue crack initiation life for the surface films with various transformation behaviors,

including bare copper specimens, are given in Table 4-3 through Table 4-9.

90
4.5.1. Bare Copper specimens The annealed bare copper

specimens were used as a baseline comparing fatigue crack initiation lives for the
specimens given in Table 4—3. The average fatigue crack initiation life (N1) was 15,292

and its standard deviation was approximately 2,448.

4.5.2. Amorphous Film Coated Specimens The fatigue
initiation life of amorphous film coated specimens was reduced by approximately 30 %
compared to that of unmodified pure copper specimens. Degradation of fatigue
properties of copper in amorphous ﬁlm coated specimen resulted from the brittle nature
of the film, which was explained in previous section. It is normal to find that amorphous
materials exhibit brittle behavior before a well defined yielding in the stress strain
relation and strong inhomogeneous plastic flow (Kimura and Masumoto, 1983). The
short fatigue life of amorphous film coated specimens is attributed to these characteristics
of amorphous materials. Thus, fatigue failure occurs from the growth of cracks nucleated

at defects, and then facilitates crack propagation into the matrix.

4.5.3. A/SIM Film Coated Specimens The A/SIM type film
coated specimens, which exhibited relatively high transformation temperatures, showed
approximately 50% improvement over the bare copper specimen. Since the Af
temperature of A/SIM type films was slightly higher than the ambient temperature, the
stress induced martensite formed during cyclic loading was retained in the sample after
being removed from the load frame. The retained stress induced martensite peaks from
X-ray diffraction study were observed from A/SIM type ﬁlm coated specimens after
fatigue, which was given in Figure 4-16. The retained martensite is presumably due to
the stabilization of the stress-induced martensite formed during cyclic deformation of the
film. The results for A/SIM type film (AfZ Td) coated specimens showed that the

fatigue crack initiation life increased with decreasing Af temperatures, as shown in Figure

91
4-45. The A/SlMl type film coated specimens demonstrated the best fatigue strength

among A/SIM type film coated specimens. A/SIMS type film coated specimen showed
an improved fatigue strength, but the least among these type of film coated specimens.
Since the test temperature was well below As for A/SIMS type film, once the martensite
phase forms at highly localized strain areas in the ﬁlm on the deforming substrate, the
martensite will remain stable throughout the remainder of the fatigue test

A/SIM4 type film coated specimens were tested at different conditions from other
A/SIM type film coated specimens, which were tested at 298 K. A/SIM4 type film
coated specimens, with the Ar of 309 K and the As of 263 K, were tested at 276 K. 13 The
crystal structure of the films prior to the fatigue test was expected to have dual phases of
thermally-induced martensite and R-phase mixture under these conditions. A/SIM4 film
type coated specimens showed better fatigue crack initiation life than other A/SIM type
ﬁlm coated specimens, whose films were austenitic before the fatigue test. This is
probably a consequence of the reversible intervariant boundary motion of pre-existing
martensite in the A/SIM4 film. Additionally, the favorably oriented stress-induced
martensite variant formed during tension will be different from that formed in
compression in other A/SIM type films. In addition, these stress-induced martensites
formed during cyclic deformation will remain stable throughout the remainder of the
fatigue test, because the Af temperatures were higher than the test temperature (T d = 276
K). Thus, the superelasticity effect in those films may less than in A/SlM4 type film.
This contradicts the result of Melton and Mercier (1978), where a higher fatigue life was
observed for stress-induced martensite compared with thermal matensite. Melton and
Mercier claimed that no selection of favorable variants in a thermal martensite might
occur with respect to the applied stress. However, it is not clearly understood which

mechanism has been dominant in A/SIM4 film during cyclic deformation.

 

13. The specimens were quenched by liquid nitrogen, and then warmed back to 276 K, which temperature
was maintained during the fatigue tests.

92
4.5.4. A/TE Film Coated Specimens The A/TEl type ﬁlm

coated specimen, whose Af temperature was lower than the test temperature (Td = 298
K), showed the greatest improvement in fatigue crack initiation life. This ﬁlm was
believed to have a substantial amount of stress induced martensite formed during cyclic
deformation which, however, is presumed to have reverted back to austenite after
removal from the load frame. The unrecovered strainl4 depends on the test temperature
relative to the Af temperatures. A minimum in unrecovered strain is usually obtained
from at 10 K to 30 K above the Af, and the unrecovered strain increased at higher or
lower temperature (Hou, 1995; Miyazaki et al., 1981). The unrecovered strain was
attributed to residual martensite for Td > Af, which could be recovered after heating, and
slip of martensite phase for Td < Af, which would not be recovered. For A/TE type film
coated specimens, the test temperature was approximately 23 K to 35 K higher than Af
for A/TE type ﬁlms. If all A/TE ﬁlms are presumed to behave superelastically at present
test conditions, the critical stresses for inducing martensite will increase with decreasing
Af, according to the Clausius-Clapeyron relation. Therefore, the films are expected to
have a higher critical stress with a full recovery when the slip of austenite is not involved
among A/TE type films. A/T E1 type film coated specimens, which exhibited the longest
fatigue initiation life, were tested at approximately 35 K above Af, suggesting that
transformational superelasticity in coatings might be applicable to the enhancement of
fatigue properties of substrates. It can be connected with the stress-strain studies of NiTi
film produce by Hou (1995), where the perfect superelastic behavior was found at
approximately 30 K above its A{. The stress for slip in martensite and austenite phases
also may increase due to the internal defects which was formed during film deposition.
This result is comparable to the observations by other research groups, where the critical

stress for inducing martensite increases with increasing test temperatures and the fatigue

 

14. The unrecovered strain is deﬁned as the strain, when the stress is relesed to zero, from the origin of
stress-attain curve, which was shown in Figure 2-13.

93
life of NiTi alloys improves with increasing the test temperature (Td) above Af, but below

Md (Miyazaki, 1986).

Fatigue crack intiation life are plotted as a f uction of temperature difference (T d
~Af) for Afl’ E and A/SIM type ﬁlm coated specimens, whose Af temperatures of the ﬁlms
ranged from 263 K to 322 K, in Figure 4-46. The results indicate that the fatigue crack
initiation life increased gradually when the difference between the test temperature and
the Af temperature of the film increased. Some of the martensite induced during cyclic
deformation would be stabilized when Td s Af, whereas most of the induced martensite
presumed to revert back to austenite when Td > Af. Even though Td was slightly higher
than Af, unrecovered martensite was observed in Ni50.3Ti49_2 film, which is shown in
Figure 2-13. Therefore, the perfect superelasticity of NiTi thin films can be expected
when the test temperature is slightly higher than Af. In other words, the improvement in
fatigue life can be achieved by reducing the Af relative to the test temperature. However,
when the specimen is tested well above Af, that is Td >> Md, the film may not form stress
induced martensite and the amount of unrecovable strain due to slip of austenite will

increase”.

4.5.5. A/R type Film Coated Specimens A/R type ﬁlm coated
specimens, which showed only R-phase transformation, displayed better fatigue
properties than average A/SIM type film coated specimens, but worse than average A/T E
type film coated specimens. This prolonged fatigue strength of A/R type film coated
specimens can be attributed to the high ductility and yield strength of the austenite phase
in the ﬁlm, but not to the superelasticity of R-phase of the film. Even though the R-phase
can be stress induced and exhibit pseudoelasticity, the transformational strain may be
relatively small. In addition, the transformation strain associated with the rearrangement

of the R-phase variant decreases with increasing temperature difference (Td - Tr)

 

15. The yield stress of austenite is in ranges of 200 to 800 MPa at 473 K depending on Ni concentration
(Miyazaki et al., 1991), and the stress of austenite ﬁlm lying on PSB is enough to induce slip.

94
(Miyazaki and Ostuka, 1986) and thus the strain for R-phase is assumed to be diminished

or affected little at the present test temperature. Therefore, the austenite phase ﬁlm might
suppress the formation of extrusions and intrusions on the surface, thereby enhancing the
fatigue lifetime.

The results of fatigue crack initiation life indicates that all crystallized NiTi film
coated specimens demonstrated longer life and amorphous film coated specimens showed
shorter fatigue life than the bare copper specimens, as shown in Figure 447. The stress-
strain behavior for the surface film and substrate during cyclic deformation may occur
two ways. If the strain localization does not occur before fatigue crack initiating in a
material, the relations between the discrete surface ﬁlm and the substrate by Greenfield
and Purohit (1980) can be applied due to the same axial strain on the ﬁlm and substrate
near the interface. Considering this approach, the stresses on the film and substrate
depend only on their Young's modulus and yield strength. Since the Young's modulus is
lower for crystallized NiTi ﬁlm and the yield strength of crystallized NiTi film is higher
than those of copper”, the surface layer will experience less plastic (or only elastic)
deformation than the substrate during fatigue. Thus, it is desirable for fatigue that the
surface film hold higher compliance and yield strength than the substrate.

However, when the strain is localized into PSBs in a polycrystalline copper with
increasing the number of cycles and the plastic shear strain in the PSB is about 1%,
regardless of the applied plastic strain amplitude (Hunch and Neumann, 1986). As a
result, the strain of the surface film on PSBs and overall bulk are not identical any longer.
Thus, once the plastic shear strain is localized in the substrate during cyclic deformation,
the surface film above an underlying PSB will encounter a large strain. The computation
by finite element model showed if the surface film on a PSB responds only elastically,

the extreme tensile (or compressive) stress on the film could well approach 1 GPa

 

16. The Young' moduli of the austenite and martensite phase in NiTi alloy are 42.5 GPa and 22.5 GPa, and
the yield stress is 100 MPa near Ms, where the yield stress is minimum (Hou, 1995). The Young' modulus
of copper is 120 GPa and the yiels stress of annealed is ~20 MPa.

95
(Grummon et al., 1990).17 But, yielding of the surface film will occur before reaching

such a high stress level on the film. Therefore, the stress-strain behavior of the bulk,
PSB, and the presumed superelastic film lying atop active slip bands will be different.
Schematic stress strain curves by this model are depicted in Figure 448. The overall
specimen displayed a narrower hysteresis loop than PSB region. The slope of
superelasitc film on loading is steeper than unloading due to the formation of stress
induced martensite, but the slope for a copper substrate is much steeper (ECu > ENiTi(A) >
ENiTl(M))°

Based on the above model, the differences in fatigue crack initiation life for
crystalline film modiﬁed specimens are thought to be mainly due to two reasons. One is
due to the reversibility of film deformation during cyclic loading, which in turn relates to
transfromation temperatures of the films and test temperature, i.e., the amount of
unrecoverable strain. During cyclic deformation, the thermoelastic martensite in a
surface ﬁlm is presumed to exhibit twinning pseudoelasticity by the reversible motion of
martensite twin boundaries at below Ms temperature. The austenite phase is assumed to
show superelastic behavior by formation of stress-induced martensite below the Md
temperature. In addition, twinning pseudoelasticity in the austenite phase of NiTi alloys,
even well above Md (Dueli g and Zadno, 1990), might improve the fatigue life. However,
the unrecovable strain in martensite, if external heat is not supplied, and in austenite,
above Md, are larger than that in superelastically behaved austenite. Thus, it is expected
that the surface film, which has superelastic properties, is more effective in resisting large
strains induced during fatigue.

The other reason for prolonged fatigue life of NiTi film coated specimens may
result from the high ductility and strength of the surface film (Grummon et al., 1988).

The Young's modulus of copper is higher than that of NiTi alloy and the yield strength of

 

17. In their study, the shear modulus of the ion beam surface microalloyed film is 66.3 GPa. However,
since the shear modulus of NiTi alloy, which was used in present study, is around 27 GPa (Jacson et al.,
1972), the stress of NiTi ﬁlm on a PSB is expected to be slightly higher than 400 MPa.

96
the austenite phase of NiTi alloy is higher than that of copper. These results mean that

the ability of NiTi crystalline film to block PSB extrusion is attributed not only to
superelastic deformation of the ﬁlm, but also to its strength and ductility. It is important
to note that the increase in fatigue life is associated with the retardation or near

elimination of the formation of PSBs on the surface by NiTi thin ﬁlm coatings.

4.5.6. Eﬁect of Post Annealing on Fatigue Crack Initiation Life

The effect of post annealing, after cycling to a speciﬁed accumulated plastic strain
(22Aep = 40), on fatigue crack initiation was studied. It is important to raise the critical
stress for slip in order to stabilize superelasticity characteristics against cyclic
deformation (Miyazaki et al., 1986). Work hardening was expected due to the
introduction of dislocations in the film lying atop of a large locally strained substrate
during fatigue. Thus, the post annealing process might facilitate precipitation hardening
and/or the annihilation and rearrangement of dislocations introduced in the film during
cyclic deformation. The former process can be expected to increase the critical stress for
slip by precipitate hardening, resulting in enhanced superelastic effects and thus possibly
increased fatigue life. After cycling up to approximately 10,000 cycles, A/TE3-3 and
A/TE3-4 specimens were removed from the load frame, and then annealed for 20 minutes
at 673 K in vacuum18. After heat treatment, the specimens were cycled until fatigue
crack initiation occurred. The fatigue crack initiation life, for heat treated specimens, was
slightly longer (approximately 15 %) than that for the untreated ones, as shown Figure 4-
49. However, this slight change in fatigue lifetime can be affected by not only annealing
the film, but also annealing the copper substrate itself. The result of post annealing the
ﬁlms after cycling was not a very effective way to improve the fatigue life in the present

condition. It is believed that the limited post annealing time to prevent the formation of

 

18. According to precipiattion process in NiTi alloy, the formation of precipitate was insufﬁcient aged
temperature below 673 K (Nishida et al., 1986), and thus lowest possible temperature, expected to be
formed ﬁne precipitates in the ﬁlm, was chosen.

97
precipitates in NiTi film and the diffusion of Cu, Ni, and Ti atoms was not long (high)

enough to change the internal structure of the film. The effect of post annealing
associated with formation of precipitates and texture and diffusion between film and

substrate on fatigue lifetime of NiTi film coated specimen needs more study.

Shim

Amorphous NiTi ﬁlms demonstrated their brittleness resulting in the development
of straight fatigue cracks perpendicular to the loading axis on the surface. The fatigue
crack initiation life of amorphous film coated specimens was reduced compared with that
of bare copper specimens. However, the fatigue cracks in all crystalline ﬁlm coated
specimens formed wavy morphologies and their fatigue crack initiation lives were longer
than that of bare copper specimen. Fatigue cracks, in general, were developed
perpendicular to the stress axis in both amorphous and crystalline film coated specimens
and propagated from the surface film to the substrate. Fatigue crack initiation life for
crystalline NiTi film coated specimens was related to the transformational characteristics
of their surface film. The greatest improvement in fatigue crack initiation life was
observed from A/T E type ﬁlm coated specimens, which Af> Td, and was related to the
superelastic effect of surface films. A/R type film coated specimens also showed better
fatigue properties than bare copper specimens. The improvement of fatigue life in
crystalline NiTi film coated specimens suggests that the ability of NiTi crystalline ﬁlm to
block (or suppress) PSB extrusion from the copper substrate, resulting in the inhibition of
fatigue crack initiation, is attributed not only to superelastic deformation of the film, but

also to its strength and ductility.

5. CONCLUSIONS

NiTi thin films, which give rise to shape memory and superelastic effects, were
evaluated as a new surface coating approach to improve the fatigue lifetime in
polycrystalline copper. The films were produced by ion-beam sputter deposition on
copper substrates and the thickness of the deposited films on fatigue specimens was
maintained approximately 3 pm. All the as-sputtered films were deposited on cold
substrates and the resultant films were amorphous in structure. Subsequent
crystallization annealing was done, using a heating rate of 10 K/min. and maintained at
873 K for 10 minutes. The average grain size was 3 - 5 um, which was larger than the
thickness of ﬁlms.

Followings are the detailed conclusions of the present study:

(1) Strong adhesion between the film and substrate was obtained as a
consequence of ion-cleaning a substrate and highly energetic atom bombardment during
ion-beam sputter deposition.

(2) The concentration of titanium in the deposited films was always lower than
that of the effective target surface, Ni47Ti53, and increased with increases in vertical
emission angles. Depletion of titanium in the film is attributed to the preferential
resputtering of titanium from growing films and the lower sticking coefficient of titanium
relative to nickel. The directionality of the beam and the working conditions are
responsible for variation in concentration in the deposited ﬁlms.

(3) Annealed films exhibited lower transformation temperatures than near

stoichiometric NiTi bulk alloys. Short annealing time and oxygen contamination during

98

99
deposition is believed to be the main factors. Oxygen, especially, can be bound in

Ti4Ni20 resulting in a decrease in a Ti concentration of the films.

(4) Surface relief, due to the formation of thermally induced martensite, was
created on the free surface of the films. The martensite developed in random directions
on the ﬁlm, unlike triangular morphology of B19’ martensite in bulk alloys, behavior
which can be attributed to thin film constraint effects and residual stresses.

(5) Stress-induced martensites were produced in NiTi thin films, whose Af was
higher and M3 were lower than the test temperature (Td = 298 K), on a plastically
deforming copper substrates after cyclic deformation. Once stress-induced martensite
was formed from the initial austenite, the martensite could be retained but the area was
expected to be limited due to the strain localization during cyclic deformation.

(6) Amorphous NiTi films accelerate fatigue crack initiation, which is a result of
their brittle nature, and thus a shortened fatigue life of the copper specimens.

(7) Crystalline films improved the fatigue life by inhibiting PSB extrusions,
resulting in the suppression of fatigue crack initiation. This is attributed not only to
superelastic deformation of the films, but to their strength and ductility as well. The
greatest improvement in fatigue crack initiation life was achieved from A/T E1 type film
coated specimens (Ni49,5Ti50,4: Af (263 K) < Td (298 K)), which showed approximately
100 % (N;WEI mm / N,cu - 2.02) improvement over bare copper specimens.

(8) Thermoelastic NiTi coatings increase the fatigue life of a specimen, when the
difference between Td and Af becomes more positive in value.

(9) All cracks developed perpendicular to the loading axis, transfered from the
ﬁlm to the subsrate, whether the surface films were crystalline or amorphous in structure.
However, the fatigue lifetime of coated specimen was related to their crack modes, the
longer fatigue initiation life of specimen exhibited the higher degree of wavy crack

propagation.

100
Directions for Future Research

The study presented herein showed the possibility of a new surface alloying
system, which exhibits the thermoelastic martensite transformation, to improve the
fatigue life of structural metals. However, there are questions remaining and more areas
to study related to the transformation behaviors, especially in the thin film form of the
NiTi alloy.

First of all, the thin film form of the NiTi alloy produced by sputtering deposition
exhibits several behaviors that differ from the bulk material due to the non-equilibrium
state of the film. The transformation temperature varies with compositions, inclusions,
heat treatment, and the internal stresses induced by defects. Thus, the parameters to
decide its transformation temperatures should be adjusted so that the film can be applied
practically. In addition, when the films are deposited on metals such as Fe, Ni, and Cr,
expected diffusion between film and substrate would occur whether the cold substrate
deposition, which required a crystallization process, or the hot substrate deposition was
used. The diffusion of each element in film and substrate changes the relative
composition of the deposited films, which affects the physical and mechanical properties,
especially if the annealing temperature is high and the cooling rate is slow. Therefore,
the effects of diffusion, precipitates, and texture formations during deposition and/or
annealing processes on fatigue lifetime are left for future.

Even though physical properties of the coated'ﬁlms were normally identical, the
fatigue lifetimes may increase or decrease depending on the thickness of the film due to a
change of stress distribution. The boundary conditions between the surface film and
substrate can be constructed and related to the balance of stress and strain behaviors
during cyclic deformation. Detailed study of dislocation substructures near the interface
region between the film and the substrate using TEM should be carried out so that surface

hardening (softening) or secondary hardening phenomena could be interpreted.

101

Table 2-1. Cyclic Stress-Strain Characterictics of FCC Crystals. (Suresh, 1991)

 

 

 

 

 

 

Material Yw nghléc 123* (MPa)
Cu 1.0 x 10-4 3.0 x 10-3 73.0
(4.2 K)
Cu _ 8.0 x 10-3 48,0
(77.4 K)
Cu 6.0 x 10-5 7.5 x 10-3 27.5
(295 K)
Cu 1.0 x 104 1.0 x 10-3 14.0
(523 K)
Cu-2.0 at% Al 1.0 x 104 3.0 x 10-3 33.0
(295 K)
Cu-5.0 at% Al _ _ 32.0
(295 K)
Cu-2.0 at% Co 3.0 x 104 5.0 x 10-3 27,5
(295 K)
Ni 1.0 x 104 7.5 x 10'3 52.0
(295 K)
Ag 6.0 x 10-5 7.5 x 10-3 17.5
(295 K)
Al-l.6 at% Cu 1.5 x 104 1.5x 10-3 95.0
(295 K)

 

 

Ypl, AB is the plastic strain amplitude of starting the plateau stage in CSS curve.
7p], BC is the plastic strain amplitude of ending the plateau stage in CSS curve.

133* is the saturation stress in CSS curve.

102

 

 

mm
ON
ON

mm
5
VN
mm
ON
5
mm

3
ON
m—
mu

9% s_o>

2058.5.
e<

 

0.0

NM
94

m;
mwd

cam

04

w.m
coo;

54
Mad
wwd
Ned
mm.“
NNA
5.0
#04
8.—
mod
me.—
C .o
mmd
5.m
Nwd
mmd

0X

5.0
wed
Rwd
omd
mm.—
om;
Fwd
NH .—
8.—
mmN
we;
2 .o
wed
cod
00.0
5N4“

HM

Rd
Ed
Fwd
omd
0?.—
we;
owd
g.—
2.—
mmd
mm.—
N_ .0
Ed
OWN
mo.—
N~ .m

2

mad
med
wad
mod
86
2 ._
wed
mod
wad
of
8.0

Ned
mo;
mod
Ct;

02

5.0
8.0
5.0
Q .o
mod
9 .0
mod
mod
2 .o
VNd
me .o
bod
de
8.0
2 .0
end

 

«I

 

F
as

25363

Lee‘s”

00

cm
=<

m<

>8 saw
«no mucosa—m

 

Assamese was: son; .8 can so; wacoeaam .3 22¢

   

 

103

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8.8 «.8 8.8

-2 a. 1 1 1 1 1 1 1 1 .58 8.8 8.8
8.8 8.8 8.8 8.8 o8 Q8 _.8

-88 1 1 ea. 8.9V .me -_ .8 8.8 .3.. 1 1 .3...
88 Q8 8.8

88.9. 8.8 .me 1 1 1 1 1 1 1 1 as.
8.9. ea. 88 «.8 8.9. a8 _.8 now 8.9. ode now

-88 3.8 8.9. 8.8 8.8 8.9. .oa. 4.8 8.8 8.8 4.8 .88
3.9. now 88

.88 .88 so... 1 1 1 1 1 1 1 1 .88
8.8 be. as. 8.9. own an. «we

-88 1 1 -88 -88 8.8 -28 .88 e8 1 1 %8
8.9. New «.8

8.8 1 1 1 1 1 1 1 1 8.8 e8 .98
owao>< 2 o w a o 8 v m N a 283.
33a Ezv

.835. Beach 35> 22sec .8 SE 55. E2 osaoaoa 5 8 as _2 co aoagaooaoo 8.382 .2. 29¢

 

104

338.3 .8588 Mm .ancoao 8:58 35.35 Howe .

 

 

 

 

 

 

mm 8o§ M 28 M Sam M M 8 AM 82 58.8.? E
Mm. mm: .2 - - - - ammdﬁa mm
mm ovwd— M mom M mm M mom M :3 Wmmwcv mm
08 8a.: M 88 M Sm. M E M 82 8.988 8
ﬂ
8502
wages: ozv 3a a 52V
.95... £3 gangs 5 o< a: :2 8288550 586on
Swansea... M080 oswuam SE

 

 

.meogoeam cam new cameo eoeouoem omsec EE

8 2:8 5:. E2 to. a: 88:3 nose seem an .aaoeaoaaoo .58 88.5885. .3 22¢

105

Table 43. Fatigue Tests on Bare Copper Specimens and Amorhpous Film

 

 

 

Coated Specimens.
(Test Temperature: 298 K)
I Fllm Sat. Max. Peak Fatigue Crack
Specimen Type Stress Stress Initiation
(MPa) (MPa) Life (N1)
Amor.-l Amor- 122 126.6 9,792
Amor.-2 phous 123 128.4 1 1,473
Amor.-3 123 129.1 12,517
Amor.-4 123 127.9 9,942
Ave. 123 128 10,931
Bare Cu-l N/A 121 129.4 21,453
Bare Cu-2 121 128.9 15,199
Bare Cu-3 123 131.1 15,296
Bare Cu-4 122 129.1 12,117
Bare Cu-5* 125 134.1 14,415
Bare Cu-6 123 130.3 15,372
Bare Cu-7 123 129.1 14,245
Bare Cu-8 124 132.9 16,157
Bare Cu-9 122 128.4 12,384
Bare Cu-10** 123 131.7 17,568
Bare Cu-11** 123 130.3 14,008
Ave. 122.5 130.5 15,292

 

 

 

 

* Damage Accumulation was monitored (N = 1,01 1-2,518-5,025-10,032-
14,415).

** After cycling up to 10,000 cycles, the surface film of specimens were
annealed for 20 min. at 673 K then resumed cyclic deformation until fatigue

crack initiation occurred.

106

Table 44. Transformation Temperatures and Composition of A/R Film.

 

 

 

 

 

Film Ni-Ti at.% M; MS A, Af T, T,‘
T =
AIR 49.1-50.9 - - - - 283 K 297 K

 

Table 45. Fatigue Tests on AIR-type Specimens.
(Test Temperature: 298 K)

 

 

7 Film Sat. Max. Peak Fatigue Crack
Specimen Type Stress Stress Initiation Remark
(MPa) (MPa) Life (NIIL

AIR-1 NR 122 135.2 28,149

AIR-2 123 131.2 21,485

AIR-3 123 133.3 21,965

AIR-4 123 134.9 28,801

Ave. 123 133.4 25,100

 

 

 

 

107

Table 46. Transformation Temperatures and Compositions of A/TE-type Films.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fllm Ni-Ti at.% Mr M, A, Af Tr T,‘
T
F____ H”
A/TE 1 49.6-50.4 (<100 K) 228 K 210 K 263 K 284 K 290 K
A/TE 2 50.5-49.5 105 K 213 K 252 K 269 K 283 K 294 K
A/T E 3 47.8-52.2 134 K 224 K 254 K 275 K 284 K 299 K
Table 47. Fatigue Tests on A/TE—type Specimens.
(I est Temperature: 298 K)
Film Sat. Max. Peak Fatigue Crack
Specimen Type Stress Stress Initiation Remark
(MPa) (MPa) Life (N1)
A/I‘ E 1-1 A/T E 1 123 136.5 29,372
A/I‘E 1-2 124 138.5 31,316
A/I‘E 1-3 122 132.2 31,015
A/T E 1-4 125 137.8 30,001
Ave 123.5 136.3 30,426
* Damage
Afl‘ E 2-1 A/T E 2 123 134.5 27,317 Accumulation
Afl‘ E 2-2 125 132.8 23,297 was monitored
A/TE 2—3 123 136.4 29,299 (N _ 1 016-
* _ s
A/T E 2-4 124 135.9 26,551 25225940.
Ave 123.5 134.9 26,616 10,029-20,035)
* Damage
Afl’E3-1* A/TE3 123 133.0 24 861 .
’ Accumulatlon
A/I‘E 3-2 123 133.8 25,480 was monitored
Ave. 123 133.4 25,170 (N = 1,010-
2,517-5,023-
Afl‘ E 3-3* * 123 135.5 30,972 10,03420,039)
A/T E 3-4** 123 133.1 26,472
Ave. 123 134.3 28,772

 

** After cycling up to 10,000 cycles, the surface ﬁlm of specimens were annealed for 20

min. at 673 K then resumed cyclic deformation until fatigue crack initiation occurred.

 

108

Table 48. Transformation Temperatures and Compositions of A/SIM-type Films.

 

 

Film
T

A/SIM l
A/SIM 2
A/SIM 3
A/SIM 4
A/SIM 5

 

 

Ni-Ti at.% Mr Ms A, At T, T,‘
47.1-52.9 (<100 K) 245 K 277 K 296 K 305 K 326 K
48.1-51.9 195K 263 K 254K 299K 317K 321 K
47.3-52.7 174 K 243 K 256 K 300 K 295 K 323 K
47.7-52.3 230 K 288 K 263 K 309 K 31 1 K 316 K
48.8-51.2 250K 286K 311 K 322K 316K 323K

 

 

109

Table 49. Fatigue Tests on A/SIM-type Specimens.
(Test Temperature: 298 K)

 

 

 

 

 

 

 

 

Film Sat. Max. Peak Fatigue Crack
Specimen Type Stress Stress Initiation
(MPa) (MPa) Life (N1)
A/SIM 1 A/SIM 1 121 131.5 25,377
A/SIM 2-1 A/SIM 2 125 134.5 20,232
A/SIM 2-2 125 135.5 26,403
A/SIM 2-3 122 132.2 26,164
A/SIM 2—4 123 133.2 21,728
Ave. 123.7 133.8 23,631
A/SIM 3-1 A/SIM 3 125 133.2 22,241
A/SIM 3-2 124 132.4 25,335
A/SIM 3-3 122 130.5 21,064
A/SIM 3-4 123 131.8 25,804
Ave. 123.5 132.0 23,61 1
A/SIM 41* A/SIM 4 124 134.5 25,952
A/SIM 42* 125 133.8 25,646
A/SIM 43* 123 134.9 30,018
A/SIM 44* 125 135.1 27,102
Ave. 124.5 134.6 27,676
A/SIM 5 A/SIM 5 123 131.7 22,371

 

 

* The specimens were tested at 276 K, where the specimens were quenched by

liquid nitrogen, and then warmed back to the test temperature.

 

110

dozens—0:: M080 osweﬂ can 37. 1252200 wanes—ca E wee—=82 5:80:0050 888
05 08.80 008:8 00.: 05 8 8:08:55 new meoasbxm 80:8 08 £023 3.83 are REwto Mo 80:8 8 .ES
mmmm d23§20e 020.8 wage cones—08 8me econ A58 “Samoa Mo 2.0585858 ones—0:8 .TN 05mm

      

88:55 + a

 

 

 

 

111

$59 :3 E ignwsEv .m 8&2 E 003295 583 5.3 8.8205 mmmm ho 0:023
0823 of. .385 couﬂamm 05 5‘. mucﬁm *3. 40> dug—med“. 58% 0:003 00 .5503 a ma .3. .323

:oaﬁamm mango—Q mp 005030 m. £033 .1338 03.5 02500 .8 0E3 58508.8 0:9»0 .N-N Eswﬁ

a > dwazmsﬁx Embm

«-3 a? L: we

 

 

ﬂ _ _ _

 

m <

 

 

wmmm ho 600000 0823/ £0. 0

 

 

0..

8 3:6 mp

0m

112

1'
TENS'ON HALF CYCLE

7 00.02

COMPRESSION HALF CYCLE

 

A
m
v

 

Rapid Hardening Saturation

 

Peak Flow Stress, 1:

 

Cumulative Plastic Strain (or cycle)
(b)

Figure. 2-3. (a) Hysteresis stress-strain loops for the ﬁrst 10 cycles of a

copper crystal subjected to a constant strain amplitude of A7 =0.02 at 200 K
and (b) a schematic of a cyclic strain hardening curve. (Feltner, 1965)

113

._....4.._.._.._. .r4.._.4.._.4.._.
._.4.._.4.._.4.... .r4.._.4.._.4.._. .rd.4.._.4.._. 4.4.
k4.._.4.._.4.... .r4..r4.._.4..r .r4.._.4.._.4.._.
.r4..r4.._.4.._. .r4.._.4.._.4.._. .rd.4..r4..r
k4..r4.._.4..r ._.4.._.4.._.4.._. k4.._.d.4...r 4. 4. 4.4..
._.4.._.4..r4.... .r4..r4..r4.._. .r4..r4..r4..r .....L. .r LL.
4.4.4. 4.4.... 4.4.4. 4.4.4.4. 4.4.4.4.
PFF FF¢¥
\\ \ w 4. 4. 4.4.
g g \
\ l
a

834 .Bﬁnwsﬁv .853 BB E03500 .0 003036 :03 05 E $38 2240:5540 can 5500:9438
c0958 835500 282% a 30.0:3é508 05 E $538020 088 05 .0 03w o: 28 9 a .3 £0000 4.2380800 0480.
>30; 05. .couﬁaam 0:98 8 £508 000 .40 8:3 020 80430024 By 0.8 waE 3 5 0.30450 5.80205 .YN 05mm

 

E04500 E0Ew0m 0mg D.
320m wEE—w So wEBon

Figure 2-5. Persistent slip band formation by non-promary slip model:
(a) the ideal infinite Tylor lattice in veins prior to saturation and (b) when
cyclic saturation is reached slip occurs on a secondary slip plane and thus
extensive annihilation of opposite edge dislocations can occur.

(a)

(b)

114

TTTTTTTI’TTTT
.l..L.LL.I..l.,’.L.I-.I.J..l.
TTTTTT/ITTTTT
.l..L.l..L.L,’.Ll.L.l.LL
T'rT-rT/TTTTTT
.LJ..L.I.,’.L&L.LL&.L
'I'TT'I','T'I'TTTTT
.I..|..L,I.|..L.L.L.L.L.i..L

'I'TT'I'TT .’

.I.J..L.l.L 1'11-
Tffffrl'l-l-L-l-l
LLLL®TTTT

TTTTT’LLLLLL

LLL TTTTT
TTTT’LLLLLLL
LL 'I'T'I'TTT

’LLlLLLLL

(Kuhlmann-Wilsdorf and Laird, 1980)

Figure 2-6. (a) the relaxed position at zero applied stress (b) the relaxed
position at the possible maximum stress before disintegration into dipole

(a)

1'
=1-

1-
a...

(b)

 

wall. (Neumann, 1986)

115

9M9 dog—Say 9:25:88 5&098008020 05 3090 44053 45:88?
.4002: 05 $55 02240808 00:08 9 $05 60:500.. 05. 5405808095 3543822884
Ea 400:0 b04808 0&8? go 02w“: 9.5832: E8me 0083 Samoaﬁoamgm ANN 0844mm

T 083804483.

 

  

 

l a 2 a < m < m 2 :z
A \
\
x
\
\ \
_ \ \
\ \ .
\
\ \ 400tm 50802 2485
\
\ \ \r $9 woe I @2040
\ \ «o I
x \ 91. 365808095 IAN on
x x x \ [8538888388. ”coi0aﬂoéimzﬂo
/f”/////

        

 

 

 

    

 

 

      

 

$39115

 

116

.3004 .wﬁﬁov 50: 0:0 00.4 v— mmo 00 00400000 SE
3:0 0:003 .534 Z a .48 08:0 083800803.» 0008888 4000004...“ .w-~ 0.5me

0: 088800809

 

h F O. 0.8 we

...................

  

Emma 2 0,. a: .2

 

................................................................... .........................I... ..
“ ~ U u
. s t n
s \1 n
s \ V t
s H n
\ i .
“ n

..................... ........ .. ....,. ....... .............

 

 

(51pm .(eniqm)aoumsisa1 [1201110013

117

Weight Percent Nickel

 

 

 

 

 

 

 

 

 

o to so 40 so so 70 so 90 100
1.00 J ,1 Y 1 1 1 1 V l ‘ L Y 4 iﬁﬁ i Y
‘ m’c 67 f
1000; L Z»
1400-1 .
,0 : f
” 1 0°C 1 i
i... 3 ,s
:3 1 :
d 4
a 1200- l
h 1
a 1 ° ’
3 Tim E
5 1m” ' I i
‘— ‘ o .
1 ‘ : | E
ionic i t pi: ':
J __ . t _ :
coo-i I- ' 'l- t
I
q I:
: i: o ' . z
j l. .................. Lilo"? ........ t
600 v «v . v. . , '- . 4 A
o 10 20 no .0 so so 70 so 90 zoo
Ti Atomic Percent Nickel N

Figure. 2-9. Equilibrium phase diagram for Ni-Ti alloys. (Massalski, 1987)

118

 

 

 

 

200 1 r r I .
Ti-Rich <— : ——>Ni-Rich
-'
150 +- e i 4
I
O
l
n
100 F ' -
9 A i .
O Q. A
50- 'o —
I I i‘.‘ l
i A.‘
o ‘ 2:
I u "
I U .
I
_50 r_ 0 Wang et al. | i ,
AlHarnson et al.. :
OlHanlon et al.: § ;
-1OO _ uzPurdy et al. ' ‘ __l
l 3
I .
' . l
—150 i i l l i l
47 48 49 50 51 52 53

Nickel atomic (%)

Figure 2-10. The dependence of the transformation temperature (Ms) on
nickel conentration in NiTi alloys (Melton, 1990).

119

 

>0=m 4.52 E 00.29% .0 320.4308
w:0~0080000.:0m £000-38 3V 0:0 00840 8600808 .0 3200808 w=000080000uz0m
HEwSNE 03 0008.0 05 wEBoﬁ 023—00808 00:00 008:; .40 008004400 < 4 TN 803m

3v

3

 

 

 

 

 

 

 

 

 

 

 

 

 

m

 

Vb

 

 

 

 

Stress

I (Td <Mf)

 

120

Stress

 

. >
Stram

III (Ms < Td < Af)

 

‘ >

Stress

 

Stress

 

5
Strain

V (Af « Td)

Stress

 

 

Strain

 

II (M! < Td < Ms)

 

. >
Strain

IV (Af < Td)

/:/

 

 

 

Figure 2-12. Five types of Stress-Strain curves depending on the deformation
temperature relative to the transformation temperatures. (Funakubo, 1984)

ASL—\Lv mwﬂi 0W. ICE! 1 .

 

 

121

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A4000
é (a) 198K (b) 213K (c) 228K
v
gzmb /7 ﬂ
1: _
CD 0 L . . . 1 1 l l
0 1 2 3 0 1 2 3 0 1 2 3
“ 258K
@400 (d) 243K (6)
E b 273K
V2000 (O
a? _
1:! M
U) 0 . . J . . l l
o 1 2 3 0 1 2 3 0 1 2 3
600+
" (0 318K
76
EM” (h) 303K
V r (g) 288K
U)
8200-
1:
"’ ﬂ
0 l I l l I I I ' 1 r
0 1 2 3 0 1 2 3 b 1 2 3 4
Eu
(l)363K
800- (1:) 348K
' 333K
A600” 0)
g .
V400+
1‘3. _
1:.
“mn—
Strain (%)

Figure 2— 13. Stress-Strain Behavior of NiTi thin film as a Function of Deformation
Temperatures (Hou, 1995).

122

GwE .8330 28 Engv .m:o_.mE._o_m:§ oEmcotmE Ea 829m 05 £3» 38683 mBBEoQEQ comma—gouge“
05 9 038—8 238an2 :oumscohoc 05 co 8:00:38 wEBoﬁ 8E3 Ewbmémobm ‘5 we“; co>om .3 -N 2=wE

Ewbm
‘lll

 

\\

CC. Ab =>

Edam
‘IIIII

 

C<VHV£2V Z

 

 

 

 

 

 

 

Ewbm
‘lllll
F0
«0
cu m.
w a
S S
S Eb S
‘1
AvaHvrC 5
58am Ewbm
TI TI
kw ..
S S
w 3mm w
x w.

QEVFVHEV E

 

AWEVHXSC =

 

Edam
Alllll

 

 

GEN—1V5 <v >

 

swam

ssons

GENE _

ssang

 

ssons

123

.28 :2 26:: 5:88:80:

8:: :8 :8: A23 22 .8: :2 5953 0888.88 8:288 :8 52 383 5:88:80: 8883: :8 3:883:
:88: 22» 05 28:: wCZ :8 0:880:88 :8: :0 88088 a 8 888880 on: :8: .20V 8:888:888: 8:05:
:8 $86 on: $883 9888: 6:: 8:82.88 8 ER :8 830:6 on: $888: 20$ 80 8%: 8:5. .mTN Bswi

CC 0:380:80 H
:<

 

 

 

 

 

 

 

 

 

 

com cow 8m + 8N o2
o . o
_ A _ t _ 8 _ _ a
mo 0 I.
I 32 m2 0
) gauge—m o o l CON
% om I a J
O I O I 8V
m 1. 1
m: 0? I. 88:: 8:m:8m:< \ ..
O E “Em :8 325 k 0 . i cow
E I. a .q \ . no on: 0:,- oo coo:- uo\ \ \ cocoon-00000000.... on: II.
811 <0 \\ 1 cow
\ 888 8:88:82
\ 5 :8 5: agm
20
D

 

 

 

 

(Raw) 889118 mm

Tensile Stress (MPa)

300

200

100

300

200

100

300

200

100

 

124

 

 

 

 

 

 

At“ = 303 K

__ (a) Test Temp. = 313 K
"' grain size : 50 um

I l l l l
P (b)
r
_ grain size: 1 mm
A l l l I l
L (c)
i‘ single crystal

. l I l l

l - - -L ..... l ..... 3 4 5 6 7

Figure 2-16. Grain size effect on psudoelasticity in NiTi alloy (T. Saburi
et al., 1984).

125

.mm: :0 88:: 82:3 a m: :083 5.0 :0 88:m 8:0: 0
:8 0:880:80: m2 03:: 0.. m: 80:: :8 00:11:90 33 0:: :m0w:0_ 0E. .85: .:0:0:02 :80 5:02: 0. om :0 :8
8 :08:0::0: :03 0:880:80: :m0: 0:. @0030 :0 :388 580w: :38: 20:0 :0: 88:: 8:0: 2:. .::-N 083.:

@0396 :0 89:82: Z ‘i

we: be: 00: mm: VG: m0: no: :O: H
.::”:2 _ 1:32 1%. T 1.3:: ﬁ 45:. : 4:2: _ jadjlaniﬂl]_ ﬂood

 

  

...................... ".::: m #OAV H > .mmnm —8_Q%P

Alllll 1
.—1
O
O

  

l

[“111 1 1 1

UoONTHmE o #5
008..”me {
UomHHm—z o

Doorumwz a

 

(“19118 [9101.)13V <—

l

Seesaw 0:8 82 :28: a: 8:8:

 

 

1.1111111 1

 

Tensile Stress (MPa)

126

  

 

 

I
800 ~—
I
600 —-
400 — I T = 433 K ‘
0 T = 363 K

200 -

_ A T = 333 K

O l l I l
102 103 104 105 106
Fatigue Life (N)

Figure2-18. Effect of test temperature on fatigue life of Ti - 50.8 at% Ni
alloy with austenite finish temperature (A0 of 300 K (Miyazaki et al.,
1989).

127

V

 

.202 .5: 05 88.25
$08 :0: 0:0 >905 =0: :0 500:: a 00 82:8: 3: 0:0 .005: A3 :0 0203 w=t0==:m .::-N 0:03:

3: 3

 

 

          

 

 

 

:23: >08wa .>03. >Omw2u
.. ..I ~ 9 . ~ 3 . ~ 8 a 1:: a ~ 9 0 ~ :1. a ~ 3
114111411111. 4 Aq<<114 d uuqq<<J+ J1 44<1<dq a < 1 ﬂ +J| «111411 1 ll 11-1Td < 4 1 d «1.9
o c
guy-0.2 \ 1 ID.Z(—= 1..
. 1 o
. . “a ~
.8. tart-0.113. 0.... o: 1
88 233.50.... 0.... : .
3: 388188.68..- ..r. \ ..9
:8 .::-Ir... 0....
.3 3."?le ““03:th n.” \0 P u .42 «uggﬁbaiitozl u .9: r
...:..v..l...§8. ..t .0: 132...... ..o:
.8: 2i 0... 1 I: 3 03:55.3 . .5
S... .. it in. o... 1 :3 39.20.53... . .3 O.
..nn d..'.‘ 2 ..r .0. 1 ﬂ 0 and. ‘u 3- ..1d‘ 0 o.- O \ ~
.1: :.- 1.1.53. 1 .l .
.1 u :2.2...a§8 1... ,. x
I. 9 \ N
. ~. .2. .:.:-.0. u .1... . \
o 1 .
I \ \ 1.
~ 1 I o\ \B D 1
. \\ .\
. . L 7 / \ \‘ I“
,/ I 1 \
I .. I. > ' . .1 1 . 1 \q H >
1n H 1 1 01 OH
H 1 1 .9
1.9 1 1
o . I L ~
0 - u 1 n
1 :1 o 0 or 10 \o v. 1
1 1 Q... 0\ g
. ﬂ U
o 1 r L n
u H O 1
199 1| I \\ u
o I! o .109
o o. :.: .33... o... o \\o
..::-80.8. 0... 1~ 1 \o 1
Ir :9 e. .- tic 0.86.36.35.36... VA! 0‘.)
1111111111 11.011 \ 3: :53: o .3 1 Ill 1111111 .02 3 i .0518: 1.3.0.3... 1 ~
d 011 \\\\\ Entity-ml: 0.1.... 5.0... 1 1 D\I \\\ .1! 33.0-=08 .
Elb‘XiI 011111u .3....013652. 93...... 1.. \1 .08 a"; 0...... o... 1
.. . .1 W111 .8. ..atlpsfluusu o... . o .a 9.... 1 Ihclilllii11\\\\\\ \\\\1\\1 Sun 3 0235:... . . 1
o. 0. 111.?11?\\ .2 .: 2.01.1 ..l.o..-.o.....o.l.o.... u \\\\\ . . . .3 0.0.. 0.8
IIILYI.‘1H\WV\ .03.!33303’33: 0.3.0.5...3.O.l.h 9 \ . A! vugg 01. 0:! 1n
5. 1111.. .:.:-3.80.38 0......a.o.a.o.!.o.a: .. .£N033103.1.l.orﬁ
INN-H” p P . —11-P> h > D bePP- I p b bPP‘PD r 1 . DPFI FDPDP i F F (I . I

 

 

 

128

 

2.5 l l I
1
Ar+-Ag,'l'o,Ti,Al .1
1./
1
/ AI\\
'1 '\
U
I/cos 9 ’1 D/Ti\ ._.
2°” 1 0 \
1/
O x
I
01
a— /’I
O
z A-A-A
§ XA’ T° ‘0
g i 7’ O£>O\
: /%4’ / A0 0 D
“ 0w / \
Q / I O
1: of} / C \
8 ’89” /<>
-J . ’ .0 \
E jO.O
3; 1.0 Q -

0.5 1- ‘1
a

l l
O 30 60 90

ANGLE OF INCIDENCE (degrees)

 

 

 

Figure 2-20. Variation of sputtering yield with angle of incidence for 1 keV
Ar+ incident on Ag, Ta, Ti, and Al (Kelly, 1984).

129

ISOLATED ZONE 1

 

 

 

 

    

  

GROWTH DEFECTS
./"'
.": / .
_/, SUBSTRATE INDUCED . /
(a) .- ‘,;ZONE I STRUCTURE ' . I, -
”.::V’ __ . “g , , .
SHADOWING .7 == 1.1.. ,3; K. -'
PROCESSES ‘ " 1:? :31} h .. '
1 ‘ ,.
| ‘ I: .1 .J . . . v
I GA -. _~ 3:4:
(b) :1 20 ., .::,“
SURFACE I '“ . 3‘.“
DIFFUSION II “M ~
PROCESSES l “W 1
l . I; I I
i A‘ "Iv "I" ll A
: --:> /
(C) 1 RECRYSTALUZATION 11"" f,
BULK DIFFUSION GRA'N GROW”
PROCESSES |
I
I
I
I
I
(d)
COMPOSITE
STRUCTURE .
DIAGRAM 4 ‘

 

3
INE RT GAS _. , SUBSTRATE
PRESSURE IPaI 0.15 0.10-2 TEMPERATURE ITITmI

Figure 2-21. Schematic representation showing the superposition of
phySIcal processes that established structure zone (Thornton, 1977).

130

.300: S3: 0080205 88:02 :0 8080:0080m :0: :0: 0802 00:08am :0 80:80:35 8::m=< .mm-m 0:09:

 

 

/ _ \

. o
0582 8.3m

08000 :28:

08000 :0>O

 

131

 

Machining Fatigue Specimens (OFHC Cu)

l

[Mechanical polishing

l

lAnnealing (1173 K, lHr.) I

l

[Electropolishing (228 K, 15 Sec.) l

 

 

 

 

 

 

 

Glass Substrate

 

 

 

 

 

———>
‘—

 

 

Ion sputter Cleaning

 

 

‘— ——>

 

 

Ion sputter Deposition

 

 

——>

(873 K, 10 Min.)
1 * 1

Fatigue Test ( ASpl = 0.2%) I [Electrical Resistivity measurement

l

I X-ray Diffraction Work and SEM Study on Fatigue Specimens l

 

 

I Crystallization

 

 

 

 

 

Figure 3-1. Flow Chart of the Experimental Methods.

12.54mm

7 62

132

Masked

  

Modified Surface by Sputtering

 

 

l l1:0 2mm wide (ﬂat) )

I 5....
I

 

 

 

 

 

 

 

 

(1))

Shallow round notch

 

 

 

 

Figure 3-2. Shape of surface modiﬁed specimens: (a) films were partially, (b) fully
covered in gauge section on ﬂat area to observe surface morphology and crystal
structure, and (c) on shallow round notched area to investigate fatigue life.

133

 

Figure 3-3. Polycrystalline OFHC copper fatigue specimen microstructure

after anneailng at 1173 K for 1 hour. The average grain size is 300 um.
(Electrolyte was used 1 part of HNO3 and 9 parts of water solution).

134

.M wNN 085:0 :0 680508 :0 00:10 08.: E00 058 :0 0.0mm :0 808200
0 E 008:0:8: 003 wing—308005 ”038.0::< w:E0=0:0::00_m :0 Emma 0:880:0m Tm 08$”:

 

 

>4::Dm ”EEO:

 

 

 

1

 

 

 

0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0340:0200 80:0 0052 Q

m0:02_

 

_ _
ﬂ 0%... 05%..
J .. ..HHH.H.HHH.H,H.H.HHHHH.yum“. «<0 0.5205).
® ccccccc . .::.... ccecccc
. “HEM- .0535:
305:5: x wmﬁomﬁ .... 28.802 080
oz... 00. EB H. ......mm: + 002: :00:
025000 . . 0:50:53:
300:0 0032.50: 0002<
K Ew 00520200.:
m. 30:20 003.0

I $300500 810.0:

 

 

135

 

 

 

 

    

 

 

 

   

 

 

 

 

 

 

 

'Ll
"""""""""‘" I I I I
I I i I
Sputter-cleaning
Beam
| l I |
I I I |
I I I l
I I Specimen
I |
I I/
I ,/I
| I
I I I | | l
I v Heat E
J— ; Sink
V Target
|
I
E I

 

Cryogenic Pump

Figure 3-5. Schematic Diagram of Sputtering System: After sputter-cleaning by
bombarding to remove dirt and the oxide layer, ﬁlms were deposited on the
specimens (side view).

136

1/2" ROUND HOLE

  
 

SETSCREW

‘
\

ION BEAM
DEPOSITION

 

 

 

 

 

 

Figure 3-6. Steel ﬁxture used to position fatigue specimens in vauum chamber
during sputter-cleaning and ion beam deposition. This is designed to facilitate
sputter-cleaning and film deposition on four specimens in one batch.

Temperature (°C)

137

 

 

 

 

 

O TOO 200
Time (minutes)

Figure 3-7. Time-temperature plot for the crystallization annealing
treatment of as-sputtered NiTi thin films.

138

.030:0::< 808050002 b35200: 80.5001”: 05. :0 wEBED 0:080:0m .w1m 0::mE

2300083.? U
:00: 0w§0> U
00: 0w8_0> ”
8080800 U

8:80:80 .

V—INmﬁ'V‘I

 

 

IIIIII

 

.8300
/Ir

090:. 9502.7

 

.::::m :030:
OD <28 .::.:

 

 

 

 

 

 

 

     

502:2 050:

\

    

IIIIIIIIIIIIIIIIIIIIIIIIII

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

 

 

 

‘-I|III|I|||||||||||||||I|||||||I|||||I||II||||||||||||I||||||

V_iIIIII||||||||||||||||III|II|IIII|||IIIIIIIIIIIIIIIIIIIIIIIII

 

 

 

 

 

520$ 0::80m

 

::< 00000::800

 

50:000.: an

 

 

 

 

 

______“
0.0 >0:

<©XK n—I

 

 

 

820$ 0::0:0>:0>:0m :0 gamma 0:005 .:-m 0:03”:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   
 

 

 

 

 

   

 

 

 

 

 

 

+ _ 5:30 826: T , _ 00000000 82:0 :3:
:2: , 0:255»: . _ $1305.20“: 0
_ 1.08.200 0.3: _A . 2.533;: 0
....r. - . 220.0 40:50”; _
V ..\ .
\ \ _ I1. 000
m m 00.300028 02,328 3
u .... ...
\ \
W H. q ... 4_ w i + . UNUE
\ 0.55.9: w . “:1 .
\ m >
9 m t
B m -::.. ..... 133 000.080: 00.080
3:03:00
c0325:

 

  

 

¥U<m0mm: ZEN—Pm

 

§

0a 9:0wa

  

 

 

 

.. ...-.0018: .

x9305: o<o._

,H. . ..H.
\
&

 

 

8 ®_

 

.00:

140

0008080 0: 0008 003 80.000 @008: 00.003
808820 .3 00.80 303003800 003 008:00:0 05. 3005 $5000 80882.4 :0 :0 0:0:0Q 0:0
800:. 023:0”: 0:380:80 :. 33 :0: 880% 95000 05 :0 88w05 0:080:0m < .on 0::wE

  

 

 

 

    

 

 

 

 

 

 

 

    

          

 

 

 

 

 

 

 

 

              

    

 

 

 

 

 

 

        

 

 

    

 

 

  

 

 

 

 

 

 

 

 
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    

 

 

 

 

 

 

.\
8.58 20.00022. \
:20 023000 \
0000080 8280.802 “
:50 03:55»: \
. .\
1 \
0>._<> .8 Edoﬁzmvo “
x “
“ 20.2.8.0 \
85000 “ I “
. , I ..I... \
1. “‘.‘~. ‘
‘ 1 . I I I I S“ ..‘e..\\ “
3%: 0.5080285 x
“ V 5 § . \ 85200108:
I h I l.l| \
002.502. “ “
N I “ 2030 2:000 “
S n e _ “ 222.23.. m
x
x x
\
mr -“
.0580 A080:
2.0a: .3: .00 u e

 

 

 

141

 

 

 

«m— A86 +~n——— A8

 

 

 

 

 

 

 

 

 

Figure 3-11. Hysteresis stress-strain loop for constant strain
cycling. Low cycling fatigue test in present study was

carried out at a constant plastic strain amplitude (ASp).

142

800008 0.0000 003.0: .0: 00:00.00 0 00 0000 00000 0:000 00 0 000000800 003 0:0
0.000 0000 00: 0000000 00000 0.0 $500850 3 0000000000 003 0E00 0:0. .800 003 0005080 8000 0000.0

05 000 80 com 003 080 8080 000000 00500990300 00:00.80 05 800: 0t00 $080000: 0:000 .NTm 000$":

0:000 00 000802
we 02 moﬁ NS 3 o

0

0m

 

(9cm) ssens

 

OS

 

02

143

 

'IIIIIIIIIVYITTVrYUI'IVIIUUIj

 

 

O 0 '.
llllllljlll+lllll_Llll_lllli'lLl

 

3x10

I

2x104

' l
104
No. of cycles

l

T

l
5x103

I

 

 

O V3

135

F4 v—l

(em) ssans

O
M N N
F-l

Figure 3-13. Detail of cyclic hardening curve near fatigue crack initiation life (NI) defined

by 1% stress-drop.

144

 

.0:w00 00:00:80 :00E0> 0800 05 00 0:m00 0003020 00 m0 000: :00
m: 8: 0:w00 00:00:80 3000.000 05 000 08000 000:0 .0300 0.0 88: 0:w00 00 m0 000: :00 m: :00:

0:w00 00:00:80 000009» 000. .:08000 000:0 0030. 05 000:0w0omv m: 00w00 9 0:w00 0:00:00: 000.
”00060000 w000000m 800m 00: :0 00000505 00:0w0< :0 800w05 00080::0m .70 000mm

_ . 88 Om:

       

a: 205
00:00:8m:
3008003 .I

h

 

-l-_ ,
W4

   

30:00:00
:O:mm§
$00208 020%,:
000000: . 0:00:00: .

  

E00 2 0000000

 

 

_ a fl:
.2. w.

:08007:

 

 

 

145

 

600 eV —>Ni47Ti53

52-

N] at.%
is
l——o—-I
I—o—l

 

 

 

yet composition
l ' T

20 310 ' 40 50 ' 6O

 

Vertical Emission Angle (deg.)

Figure 4-2. Measured concentrations of Ni and Ti in deposited NiTi thin
films for various vertical emission angles.

146

 

 

 

 

 

 

 

 

 

 

 

-:::—.:::... -.:-

 

 

i

:r-1ém :

Figure 4-3. EDS spectrum of crystallization annealed NiTi thin film: The
argon Ka peak was detected from both as-sputtered and annealed films.

 

 

147

 

.80.: 00:00:00 00000580000 0.0 3: 00.0 8080 0:08:
808 032:0 000:3 00000 0000000800 80.: 020000000 05 A0: :0 :09: 32000008 80:00... 000 A8003:
80:00». 00000000 030050... 000:» 00 000000000 00000.». 00: :3 00800: 080.: 0:5 :.EZ :0 80.0005 .3 80$":

3: :0:

 

148

 

..........................................................................................................................................................................

b Exo

 

Heat Flow
Arbitrary Unit

 

 

 

........................................

Endo 4

 

 

 

 

 

450 °C 500 °C

Figure 4-5. DSC trace obtained by heating the 3 um-thick amorphous NiTi
thin film with a heating rate of 10 K/min. The exothermic peak appeared
approximately 773 K and was taken to be the crysatliization temperature for
the thin film.

stunoo jO JaqumN

smnoo JO JaqurnN

149

 

 

 

 

 

 

 

 

 

 

 

 

2.93 2.25 1.02 1.54 1.34 1.20
300.0 1 1 1 1 100
270.01 ~90
240.0< »so
210.01 +70
190.04 >50
150.01 >50
120 . 0 MM M i‘ o

90.0« “ﬂﬂm%w% ﬁMﬁﬁ .30
50.04 "I‘AWW ”l I. 2' } 4 p .
,0 , .. M will 'l‘llllll’l‘llillll’li‘ Ml'llllllll :

0.0 . . . . , T . . . 1 T . . . I . . . . , . . . . o

30 40 so so 70 so
26, Degree
(a)

5000:.93 2:25 1.82 1.54 1.34 1.130
450.0- -9o
400.0- »ao
aso.o< L1o
300.01 >60
250.0- ~50
200.01 »40
150.01 ~30
100 . o W W L20
5°-°* WW WWWM‘WWMWWWWWAM'1°

0.030 . 1 T ' 4‘0 7 ' I . s'o ' ' I ‘ 6‘0 ' ' . ' 7'0 1 Y 800

26, Degree
(13)

Figure 4-6. X-ray diffraction scans of (a) as-sputtered and (b) crystallization
annealed NiTi thin films.

150

8:": 00:0. 0000000814 502 00: 000080800. :0 000000": 0 00 00300600 :000000:m 6-0 000mm:

000000000800.

00 8w 0°C 00 8w- Uo CWT

 

n ...-us:.....a'..nl..........Il.u...a....la ..........................................................................................................................................................................................................

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(sum Kemcuv) MAIISISOH

151

 

Figure 4-8. SEM micrograph for the NiTi thin film annealed at
873 K for 10 minutes showing grain size and grain morphology.

152

.00m008 0:00: 08: m0E0000c 00 00328: M wow :0 00000000 000000 3: 000 000% :0: 00 00000000 3: 8m n 02
000 V: Nam n :3 8:: 0:5 «0050.002 0: 00:w0:0::0008 :0::00 000:000 0:0 w0:30::0 000003008: :00000 .010 000mm

3: :0:

 

(b)

(e)

 

I
IL.

" "it"‘u‘ .' . "(.11~5”‘: 5‘. I“

Figure 4—10. Optical micro phs showing the surface relief moxphologies in Ni483'l‘i512
thin ﬁlm (Ar = 322 K and Ms = 286 K) deposited on copper substrates at 298 K (a) cooled
from above At, (b) followed by quenching into ice water and then warmed back to 298 K,

and (c) followed by quenching into liquid nitrogen and then warmed back to 298 K.

123 K 173 K 273 K 373K

Ni50.9Ti49.1

Ni50.2Ti49.8

Ni49.2Ti50.8

Resistivity (Arbitray Units)

Ni48.8Ti51.2

 

-150 °c -100°C 0 °c 100 °C
Temperatures

Figure 4—1 1. Electrical resistivity versus temperature curves for various compositions
of NiTi thin film alloys annealed at 873 K for 10 minutes.

Resistivity (Arbitray Units)

 

155

123 K 173 K 273 K 373K

Ni47.1Ti52.9

-150 °c -100 °C 0 °c 100 °C
Temperatures

Continued.

$5385 Bun B. A3 use ._2 8v .wc:8=§ 5cm :0 3V .mE—wozcd 283 :0 c3 5.. 8an cougcoocoQ
mEBOﬁ 0:533 98 SE 339332 5953 83.85 05 3 9:288 0518* 2358830 .NTv oSwE

A3

(I...) i3<2l~<ef>§

03:35 :0

oﬁmbmnam :U

 

157

deans—50

 

Dag—355m SU

 

 

 

TEST
DATI
FILE;

EMC

TM
STE

ST]

«(mqmqm>mm@mm

TEST: BARE CU(10/7-2/93)
DATE: 10.07-1993
FILE: F1007

EMOD: 120 GPa

AREA: 43.92883 mm2
TARG: .002

STRS: 41MPa

ST RN: 0.001

NCYC: 10005

TEST: BARE CU(10/7-2/93)
DA’I‘E: 10-07-1993
TIME- 16:54:48

FILE: F1007

SPEC: F1008

EREM: -l.026lS4E-03
EPMX: 2.0553295-03
STMX: 124.6745 MPa
SCMX: -129.4277 MPa
EACC: 20.00103
EMOD: 120 GPa
AREA: 43.92883 mm2
FREQ: 2.202151 Hz
TARG: .002

EPLS: 1.998269E-03
TMAX: 60.01011
BLOK: 324
WRITING

F1008H

158

 

 

 

 

.001

.......................

.................

......u.....p...-....-...- ...............

......................

0......- ....................................

 

 

 

 

 

 

MONITOR

......-....‘.--o-c..nu..O.--uuo¢..u.20-.-.---.-vu ------------

.---I.--a.o..-.......-.-.--.....- ...........

 

 

 

 

-.n--u\--a..u---ug‘ ...........

 

 

 

 

 

 

 

......no-uuno...........- ...................................

o............-.................. ........,.,_....- ...........

 

.....ao-u— ..........

 

..p.........-.~. ............

..............................................................

 

 

 

 

 

.001

 

 

 

 

 

 

 

 

 

 

 

...-...:.--...-.—...............uuou-oo.-.---o~--- ...........

 

 

 

 

 

 

 

 

 

 

Figure 4— 13. The hysteresis‘loops for the ﬁrst 10 cycles (top) and after
reaching saturation for polystalline copper specimen annealed for 1 hour at

1173 K. (Asp = 0.2%)

159

.68 do u as 8 .85

6.8 md u .54 3 8:3: 2 $26 8:883 05 98 Steam E 01.: oawzd of. .806 new mood ..o $25.95
586 20me have: M Mb: 3 330:5 5&8 o:___5mbo>_oa 05 :5: 2:3 wsaogﬁ 0:0»0 .Eiv 0.5mm

298 m0 898: Z

S 42 mow New 32 o

m
________. F._P____O

_ -

 

O
In

§

 

 

 

on

 

 

(em) ssens

Number of counts, Arbitrary Scale

160

 

....................................................................................................................................................................................................

,.-.......4.........o.........a.........¢.........4.11..1...4~ .........................................................................................................................................

..................................................................
1 1 - c

 

 

 

.......................................................................................................................................................................................................
...........

 

 

.After Monotomc
overload,298Kg;ggzgg;gg3;;2é

1.1....—.1,..11.1....'. .................................................................................................................................

 
 

 
  

 

 

 

.--1 ....... ‘- .....
Reheated to 400K, '

Cooled .tO 298K

1 1 1 1 - 1 1 . . - 1 1 . 1
1 1 1 1 1 1 1 . . . . . 1 .
1 1 1 1 1 . 1 . 1 1 1 1 1 1 1
1 . 1 . 1 1 1 . 1 1 1 1 1 1
g 1 1 1 . 1 . 1 . . 1 . .
1 1 1 1 . 1 . . . 1 . 1 . . 1 .
............. ....a...11....‘... .1 ..a..... . .4....,....4..0 1. ..4.1....1........,..,‘.,.......L.....l..4,........4... .1...a...... ..4... , ...4...2 ....4-..12...,-.< ..... 4 ... ....a.. ......
n 1 1 1 1 1 1 1 1 . . . . 1 . . . .
1 1 1 - 1 1 1 1 1 . . 1 1 1 1 . 1 1 1
1 1 1 . . 1 1 1 1 . . . . . , , ,
t 1 1 . 1 1 1 . 1 1 1 . . 1 . 1 1 .
1 1 . . 1 1 1 . 1 . 1 1 1 1 1 . 1 1
1 1 1 1 1 1 1 1
v 1 1 1 1 1
t
1
a
1
......... .r........‘-..._..............,.
n
s
1
v

..........................................................................................................................

  

 

Annealed and
~LN2-Quenched,
wannedbackt023235§§%§?éi3i?
,298K . ,,,,, ; ........ -- ..... ......... -- .. -.....

Amorphous 1311111298 KW; _______ ......... ._ __________ . .........
1* i i i i i i i i i i i i i i i i i 1
36 38 40 42 44 46
20, Degrees

 

 

 

Figure 4-15. Cu-Ka X-ray dif fraction scans for Ni50.3Ti49.7 ﬁlm coated specimen
at 298 K (Ar = 345 K and Ms = 273 K): showing stress-induced martensite peaks
on a monotonic loaded (a), crystallization annealed (b), and quenched ﬁlms
followed by warm back to 298 K (c), and as-sputtered ﬁlms ((1).

Number of counts, Arbitrary Scale

161

 

 

After :Fatigue

 

 

X-ray diffraction Peaks from
Stress-I

 

   

nduced Manensite

 

 

 

 

 

 

 

Before Fatigue

 

r:'

 

 

i

i

 

i

.E,N.-
ECO'
IA.
iOé
EY—E
9F:

1'

   
 

 

 

 

..................

 

 

 

 

 

 

36

38

l
40

l
42

28, Degrees

i

 

 

46

Figure 4—16. X-ray diffraction results from Ni50.3Ti49.7 ﬁlm on fatigue specimen
(Af = 345 K and Ms = 273 K). Definite (002), (ll-1) and (111) martensite peaks
are present after fatigue which are believed to have formed in regions of severe

local straining above underlying PSBs.

162

 

 

 

 

 

Heat Flow (Arbitray Scale)

 

 

 

 

200 250 300 350

Temperature (K)

Figure 417. DSC heating scans from Ni50.3Ti49.7 film deposited on copper foils.
The sharp endotherm at 273 K in the scanned trace indicates the melting point of ice.

Number of counts

163

 

35 4O 45 48

29, Degrees

Figure 4-18 (a). X-ray diffraction scans of Ni50.3Ti49.7 film on
glass substrate on cooling from 348 K to 173 K.

Number of counts

 

164

i (Oll)R

26, Degrees

Figure 4-18 (b). X-ray diffraction scans of Ni50.3Ti49.7 film on
glass substrate on heating from 173 K to 348 K.

 

 

 
 
 
 
 

 
 
 
 

 
 
 

 
 
 
 
 
 
 

 
 

 
 

 
 

 

 

 
 

 

 

 

 
 
 
 
 
 

 

 

 

 
 
 

 
 

 

 

 

 
 

 

 

 

 

 

  

 

165

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 
 
 

 

 
 

 

 

 

 

 

 
  
 
 

 

 

 
 

 

 

 

 
 
 
 
 
 

 

 

 
  
 
 
 
 

 

 
 
 
 
 
 
 

 

 

  

 

 

 

 

 

 
 
 
 
 

 

 

 

 

 

 

 

 
   
 

 

 
 

 

 

 

 

 

 
 

 

 
 

 

 

 
 
 

26, Degrees
both (110)32 and (111)Cu peak were shifted to a higher angle at X-ray

n
’

 

 

 

 

. . . . . . . . . . . . . . . . . . .
1 _ . . . . . 1 . . . . , 1 . . . 1 .
_ . . . . . . . . . . . . . . . . . .
1 . . . . . . . . . . . . . . 1 .
. . . . . . . . 1 . . . . . . . . . .
. . . . . . . 1 . . . . . . . . . . .
11.11.11.111: 1.11.11.....1...11.11..1 11.1.. .11111111111111.2111....1111..1.1.. 11111111...\.11.1111.1.....11..1141... 111.111.1111...1‘.. .1 ...1....1..1.1. 1.1.1.. 11.1...111 .1 1. ..1.1.....11..1...1....1....11.11..11..11-11.11... 111111.11... 1 1
. . . . . 1 . . . . . . . _ . . 1
1 . . . . 1 . 1 . . . . . . . . 1
. . . . . . . . . . 1 . . . . . .
. . . . . . . . . 1 . . . . 1
. . . . . . . . . 1 .
1 1 . . . . . . . . . . . . .
v.1.11.1...m.....1..1.....1....1....1..1.1.11.1.1.1111113111711111... 11.....1. 11111111111. .1111111m.1...1111 11.11 1111*..11111.11v111111...~.1 1.1.1.4.. ...1. 1. ..w: . 1.. 1.... 111111 m..1..11111.1.111..1 11.—11.11111 ....1..11~. .. .
. 1 . . . . . . . . . 1 . . . . .
. . 1 1 . 1 1 1 1 1 1 1 . . . . .
. . . . . . . . . 1 . . . 1 . 1 . . .
. 1 , . 1 . . . . . . . . . . 1 . . .
. . . . . . . . . . . . . . . . . .
, . . . . . . 1 . . . . . . . . 1 .
11.111.111.1.. ....1....11....1.1.1.11... ........111.11..11.. .1....11111111..1.111 1..1.111111711111111... ..111111t11: 1111111111. ...1111. ..1. .1 ..1....1.111 ...1, .11 1 ..1 ..1 . 1..i1.1111.111111.1111.111|.........1111....... ..1....1.~ . ...
. . . . . . . . . 1 . . . . . . . 1 .
1 , 1 . . . . . . . . . . 1 . . . . .
. . . . . . . . . . . . , . . . . . 1 .
. . . 1 . . . . . 1 . . 1 . . . _ .
. . 1 1 . . 1 . . 1 . . . 1 1 . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . 1 . . .
1 11 11...... ......1..... .1 11.....1... .1....111..1:..11.11111 111.11.! 1.. 1.11. 11.1.11....\ .1.. ...1.. .. 11.\. 111.111 11.11.1113: . ....1 ..1.........1.1.1... ...1. . 1 1. .1.. .111 111.1..111.1..11.11.11..1....1.. 131...... 1 .1
. _ . . . . . 1 . . . . . 1 . . . . . .
. . . . . . . . . . . . 1 . . . . . . . .
. . . . . . . . . . . 1 . . . . . . . . 1
. 1 _ . 1 . _ 1 . . . . . . 1 . . 1 . . .
1 . . . . . 1 . . . . . . . . _ . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . 1 . . . . . 1 . . .
1..-..11.1..1... .. . .11....11.1....1111...1.1~. .. 11.1.1.1111.11«111.1.111.)111..11.... 1.11.11.11..1..1..1...1...11.11111..111. 11117111111114. .11 ..1a11 . 411.111.11.11... .1 ..v1 .....1.1.....1111111..1..1111 .1) 111.11 ...—111.1... 1......1...1.1 .1
1 . . . 1 . . . . . . . 1 . . . . . . .
. . . . . . . 1 . 1 . . . . . 1 . . . .
1 . . . . . . . . . . . . 1 . . . _ .
. , 1 . . . . . . . . . . . . . 1
. . . . 1 . . . . . . . . . . . . .
. . . . . . . . 1 . . . . . . . . . .
T1111... ...1 1.1.1.1... .1....1...1.11.11.1. 111111.1111.1..111 11111111\1.11..11.1 11.1.1..11.111111.111111.11...11.11..1 ..11.111.11 1.11.1.1 1.11.1.1. ..1....1..1..1.11 .11....1... 111. 11 .1..1.1..1..1..1 11111.11.1\11..1....11....11.1. 1111.....1.1 11.1
. . . 1 . . 1 . . . . . . . . . . . .
. . . 1 1 . . . . 1 . . . . . . . . .
. . 1 . . . . . . . . . . . . . 1 . .
. , . . . . . . . . . . . . . . . . .
1 . 1 . . . . 1 1 . . . . . . . . .
1 . . . . . . . . 1 . . _ . . . 1 . .
. . . . . . . . . . . . . _ . . . .
1.11.....1.1 ..... ...1.-...1..1..11........1..1...1..1..1n1.1...1.. 11.111.1‘111.11111. 11.111111.11.1111111111111...1-11111..1 11.111111......1.o11.1 1.1. ........1...1...1..1..11..11..1....1.111......11.111.111 ...1.... 1111.31.21.11. 11...... 11....
. . . . . . . . . . . . . 1 . . 1 .
. . . . . . . . 1 . . . . _ . . . .
. . 1 1 . . 1 . . . . 1 . . 1 1
. . . _ . . . 1 . 1 . . . . . .
. . . 1 . . 1 . . 1 . . . _ . . .
. . . . . . . . . . . . . . .
. . 1 . 1 . . . . . . .
v1 .1. .-.1‘.11..1111.11.1.1111..1~...1111..1a. 1V1. .11.. ..11 1110‘. 1.1.1..111~1..1... ...1....11s.1 4 .Q. 4 111.1111... 11.11.11.11111 11.111.111.111 ..1
. . 1 . . . . . . . . . . .
. . . 1 . . . 1 . . . . .
. 1 . . 1 . . . 1 . . . . .
. . . . . . . . . . . . .
. . . . 1 . . 1 . . 1 .
. . . . . 1 . . . . 1 1 .
1. 1.....r..111.111.111.11}?1111.1.1t1 1P 1.. . 1 1.1.1 . .....1.111v1111111 11 1 1111 r.1..111.11 11.
. . 1 . . . .
1 . . . . . _ . . .
1 . . . \ . . . . .
1 . 1 . . 1 . . 1
. . . I . 1 . 1 1 .
1 . _ . . . 1 1 .
. . . . . . . . 1
.1 . 1111111111 1111 1.1111111111111111. 1 1111.....11O.1.111111.1 1....
. . . 1 . . . .
. . . . . . .
. . . . . I'll . .
. . . . . . .
. 1r|ll . . 1 . . .
. . 1 . .
. . . . . . .
, 1 1. 111. 19.1111111111111111111 1111.11.11 111.1.1111. 111.11....1
. p . . 1 . .
. 1 . . 1 .
. . J . . ,
. . . 1 . .
. 1 . . . . .
. . . . . . . .
. . . . .
“.1.. v. 1 1.. « 21.11.12.141.
. . . . . .
. . . . . . . . _
. . . . . . . . 1 .
. 1 . . . . . . .
. . . . . . . . .
. . . . . . . . . .
111111111 211.11.11.11. .1....111\1111111.111 11111111111111111. 111.111.11.10....11.111 .1....11111111 1.111.... 1..1 1.11.. .11....1...11..1......1..11.11.11.1\..11..1.... 1.111111..\11 .11.. ..
. . . . 1 . . . . . .
1 . 1 1 . . . . 1 . . . . . .
. . _ . . . . . . . 1 . . 1 .
. . 1 . . . . . . . . . . . .
1 . . . . . 1 . . . . 1 . . 1
. . . 1 . . 1 . . . . . . . .
. . . . . . . . . 1 . . 1 . . 1
1111111111.....1..-..1.1111.111..11.11.11.11.11.1111111111111111 11.111.111.111}..111111.111..111111111... 11.11.111.121 111.11.1111. . .........11111111.11..1.. 1.11.1...111111111111.111.11 1.111.111.11. .11.. .1.
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. 1 . . . 1 . . . . . 1 . . .
. . . . . . . . . . 1 1 1 1 1
. . . . . 1 . . . . . . . . .
. . . . . 1 . . . . . . . .
. . . . . . . . 1 . . .
1......1..11.. ......1~1..1.....14..1..1...1....1111.11.111.11.11 11.....1111111111 11.11.1111\.11111..11 ..1.... 1\.. ..1111111 ‘11. 11....11111.1..11111....1.2.11.11.11.11......11111 .11111... .1 1.... 1..
. . . 1 . . . . . . . 1 . .
. . . . . . . . . . . . 1 .
_ 1 . 1 . . . . . . 1 . . .
. 1 1 . . . _ . . . . . . .
. . . . . . . 1 . 1 1 . . .
. . . . . . . . . . . . . .
1111.1.1 21.11-... .a. 1....1..11.1....1111.11111.1.1151.1...11111w111111111.\1.11.1111 ...111111111I 1.1.1.1.. 11.11.11.111: 111111. 11’... 1.11.1 1.11.11.15.1111111111. 111.1..1\1.111... 11.11.111.111. 1..11 .111
1 . . . 1 . . . . . . _ . .
. _ . . . . . . . . . . . . .
. . . . . _ . . . . . . 1 . .
. . . . . . . . . 1 . . . . .
. . . 1 . . . . . . . . . . .
1 . . . 1 . . . . . . . . . .
. . 1 . . . . . 1 . . . 1 . .
1111......11.1..11. ..11..........1..1..11....11111111111.....1111 1.111111.|1.11.11.. 11111111111112.1211: 1111111111101 ..1.... ..1. 1 .1111. 1.11.11.11.11..11..1111 .111 ...I1111...1 111.11.111.11 1.1.1 111.
. . . . . . 1 . . . . . . 1 1
. . . . . . . 1 . . . . _ . . .
. . 1 . . . 1 . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . 1 . . 1 . . . .
1 1 . . . . . 1 . . 1 . . . .
. 1 . . . . . . . . . .
11.111.111.111....1....111111111..11.11.1111..11. 1 1.11.11. ..11.111111\11 1.11... 1 .11 11.11. 1... 1......11111. 111.11.11.11. 1.11.11.11I. 1 ..1..
. . . . . . . . . . .
1 . . 1 . . . 1 . . .
. _ . . . . . . . . . .
. . . . . . . . . 1 . .
. . . 1 . . . . . . . .
. . . . 1 1 . . 1 . . .
#11111....1..1..1.....1r1111111111x..11111.1.11111111111511111111. . 11 .1 p 11111 111 . 1 11111}... .111 1111 111.1 1.11 1.111... 1111111111.». . .1 .111 .
. . . . . . . . 1 .
. . . _ . . . . . 1 . . 1
. . . 1 . . 1 . 1 . 1 . .
. . . . . . . . . 1 . 1 . 1
. . . . 1 . . . . . . 1 . .
. 1 1 . . 1 . . . . . . . .
. 1 . . . . . . . . . . . .
#1 1.1.11.11..1..1.1. 11.111.11.1111.111.11.11 t 1 111,.1 1 11. 111111.1 .11 1 ..1'111 11.... 11.11 11\ 1.11111. 11 111 \111111
. . . 1 . 1 . . . . . .
. . . . . . 1 . . . . .
. . . 1 . . 1 . 1 1 . . .
. . 1 . . . . . . . . .
. . . 1 1 v u 1 . . - .
. . . 1 . . . _ . .
. . . . . . . . .
11..1..11.1..1..11 111..111.11..11. 1.1111111111111 111.11.. 11 1. 111.111.11 1111111.. 1111111. 11.11 11111.11.1.1111...11.1..-.111.111111... 11111111111..11.111. 1 .111
. . . . . . 1 . . . . . .
. . 1 . . . . . 1 . . . . .
1 . . . . . . . . . . 1 . .
. . . . . . . . . 1 . 1 . .
. . . 1 . . . . . . . . . 1
1 . . . . . 1 1 1 . . . . .
. . . 1 . 1 . . a 1 . . 1 1 .
1..111.11....3..1.111.111.111.11.4.1111..11.11.11111111.a1..11111111 .111111.11111111111117-1111:.1 ...1—111111. 1 11111.11.1\1 1 11.111 11111....11111..11 14...... 1.11111... ..1..11..1J11. 1 1 .11.
. . . . . . . . . . _ . . .
_ . . . . . . . . . 1 . . .
1 . . . . . . . . . 1 . . .
1 . . . . . . . . . . . . cl
1 . 1 . . . . . . . . . _ v
11.11.11..1111..1..1..1:1.1111111111111.11.11111111111111L1 1111.111111.F 1 11111... (1.1 1.1.1. 11.111111 11.1.11.11v1 11.111.1\1111.1 11111111181 1 a ..1
1 _ . . . _ . . . . 1
. . . . . . . . . 1 .
. . 1 . . 1 . . 1 . .
1 . . , . . . . . . . . .
. 1 . . . 1 . m . . . . .
. . . . . . . . . . . .
1.1...111111........1..11....11...1...11..1... 111111111111.1111111. 11.1.1.11111111111. 11.11111... .111.111111. 1 11111111. 1111 1111.. 11......111..11.11...-.11111. 111.111... 111.111 111 1111.11
. . . . . . . O . . . . r u
. . . . . . 1 . . . .
. . . . . . . . . . . . e
. . . . . . . . . 1 .
. . _ . . . . . . f
1.111 11......1..1..11. 111 1:1...111111. 11111111141111.1111. 11.111111‘11111 11 1 1.. 11.111.111.11 111111.111. 11.111111... 11.11111
. . . . . 1 . . . . . .
. . . . . . . . . . . . 1
. . . 1 . . . 1 . . . . .
. . . . . . . 1 . . . . . .
. . . . . . . . . . . . . 1
. . . . . . . . . . . . . .

 
 

 

 

 

 

 

 

 

28m 9552. 18:38 Mo c3852

35
Figure 4—19. Peak shift resulted from a slight change in the height of exposed

surface
diffraction scan from fatigued specimen.

.65 n mqumv 88 838:: E 8111530 11:03 mama—can 885m 296m 2:?»
.83 @880 SE 05 E 8680 205an 02 ”E 250on 05 do amﬂwEBE 8250 .ON& oSwE

  

 

167

Film

Bare Cu

 

 

Figure 421. (a) SEM micrographs of specimen Flat the boundary between
modiﬁed (top) and unmodified area (bottom): Signiﬁcant number of cracks and
severe extensive PSB extrusions distributed in unmodiﬁed area, whereas no surface

damage was observed in the film modiﬁed area at a low magniﬁcation (22A8p = 70).

'lm

‘ Bare Cu

 

 

Figure 421. (b) SEM micrographs of specimen Flat the boundary between
modiﬁed (top) and unmodified area (bottom) at a high magnification : The
extrusion-intrusion bands in a substrate grain were totally suppressed in tne coated

area (ZEAsp = 70).

169

 

Figure 4-22. SEM micrograph from specimen F1: the film delamination is
observed at the triple points of grain boundaries (arrow).

170

 

matrix area -
ssed by the film

Figure 4-23. SEM micrograph from specimen Flzuncovered local
showed PSBs but ﬁlm adhered firmly and the slip bands were supre

171

.E n 1123 new
32% xwom .8 °on coca mm 586on 2: :0 3680 EE 9:305 :QEwEBE 7328 .VN& 059m

 

172

 

.:ozwocEwNE nwE a 3 ©3330 80>» hoist} 38 3v
new 8526 9:98. 05 9 536593 tome—«>3 8880 3 ”mm 5860mm Soc Emmi—web: 2mm .mmlv 05mm

3V 3

 

 

t

. .
~A" ‘V..

173

 

Figure 426. Optical micrograph shcwing film cracks on the Specimen F3 after 1% of peak stress

drop (22Aep

= 67).

174

 

.on 08:33 2: 9 EE 2: Soc @2885; x88 a can .Sv 33080 203 hoist}
898 AS coco—gov 8.080 ‘8 bicep :wE a wEBoﬁ mm 586an EC: 23838:: 2mm 5er oSwE

3v 3

 

175

 

62.5250

Number of counts, Arbitrary Scale

176

 

 

 

 

 

 

 

 

 

 

 

 

 

26, Degrees

Figure 4-28. Cu K-a X-ray diffraction scans of the specimen F3.
Ti2Ni peaks were observed with copper and 82 austenite peaks.

 

 

I:igurc
(0a), (
SerSs
21101] g
and (c
Called

177

 

 

 

 

cm

 

 

 

Film V Hf

 

 

 

S bs
u m 1H5
(W

 

i\_\

N=4 3 N=2 N=1
(c)
Figure 429. (a) A schematic illustration of the gauge area of specimen and applied stress

(0a), (b) when parallel cracks developed prependicular to the loading axis on the film, the

stress effect in x-direction on the film between the cracks decreases and the stress (0e)
along y-direction, will intensify with an aide of residual compressive stress on the film,
and (c) the resultant strain along y-direction goes more compressive side continuously
called 'cyclic compressive creep'.

178

.80 n saws new 32:. ﬁa .0 9% 3% E
2085on 05 :o 83 cocoom owzam an: 2: E 8680 o: wEBoﬁ :QEMEEE REED .omé Raw—m

 

179

 

Figure 431. SEM micrograph showing a large crack in the
opposite side of the flat gauge section area on the Specimen F4

after 1% of peak stress drop (22Aep = 90).

.SV ES» :32th an 5 9:0me 08am .2? 3:28 295 98 A3 28% oEmEmEEmbo 6293 3:0 5 cannot.
295 853 E7. 25. AN I nw<wmv mug—93 com coca 3&8 95:68:: E .5580: 923 mi Eamiom .Nmé Bswﬁ

3v 3

 

181

 

Figure 433. Persistent slip band structure in bare copper after
500 cycles (22Aep - 2).

182

 

Figure 4-34. SEM photographs of unmodiﬁed copper specimen at various cummulated strains (ZEN-‘4) s 4, 10,

20, 40,andafterfailme)atA£p=0.2%. 'l‘hespacingofthesliplineswasnarrowedandtheintencityofthe

extrusions increased.

183

Figure 4—35. SEM micrographs showing the presence of secondary
slip trace near grain boundary after fatigue crack initiation occurring

 

60).

(22A8p

184

 

Figure 4-36. SEM micrographs of the specimen Amor.-2 showing fatigue
cracks propagated in a straight character (22Asp = 44).

185

 

Gav H mm<NNv .uotomno 83 23a: 25.5 28E? 2: EV new EE mac—ESE“ So: @8863: 203
$6.80 23 coho 32% x89 oe H coca 5.50on 8800 EE 30:80:.“ Soc mnmﬁwouog 2mm 6le 2sz

E

 

186

 

Figure 4—38. SEM micrographs of the specimen A/SlMl showing a fatigue
crack propagated perpendicular to the loaidng axis (22Aep = 100).

187

A8 98

a

A .252 5862; 05 o 83 a: 8.5
35m? 586on 05 ‘8 88 a: v.88 A8 can xoﬂo owes a He: A8 do mnaﬁwouﬂﬁ ENE .omé. owamE

 

188

 

Continued.

 

. E1 U rn

I
.:|
L

Figure 4-40. SEM micrographs of the specimen A/SIMS showing a large
crack with fine cracks independent of substrate grain orientation (22Aep =
90).

190

 

Figure M1. SEM micrograph of the specimen A/TEl-l showing a large
wavy crack with surface rumples (22Aep = 120).

191

 

Figure M2. SEM micrographs of crack tip area of the specimen A/TEl-l
showing a 'zig-zag' crack propagation (22Asp = 120).

192

  

A833 Bag US. .8 .9. .on
.3 .v u gs «Ema 325883 Song 8 TBS 59.80% 05 .«o manganese 2mm .91. ”gamma

83 u £8 83?.“ “85‘

Ova—MSG

 

 

 

Figure 4-44. SEM micrograph showed the crack tip area of the specimen
AIR-3 (22Asp = 88).

194

 

Fatigue Crack Intiation Life (NI)

 

 

 

26000
i El
25000-
24000- a
a
23000.
1 El
22000-30 . 320 i -i0 ' 0 I 10

Td- Ar

Figure 445. Effect of the Difference between Test Temperature (Td) and Al"
on Fatigue Crack Intiation Life (N1) of A/SlM-type Film Coated Specimens.

Fatigue Initiation Life (N1)

195

32000

 

30000-

28000~

26000~

24000 .

 

22000 4

20000 . f
-30 -20

 

r

 

 

' I f I ' l ' I

-i0 0 10 20 30 . 40
Td-Af

Figure 446. Effect of the difference betweenTest Temperature (T d)
and Af on Fatigue Crack Initiation life of NT E andA/SIM-type
Film Coated Specimens.

30000

196

:2

10000
0

 

 

* : Single Test

r i .1

wig isnoqdrourv

-‘b

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

AAA

 

 

 

 

 

 

 

013622 ‘19)\
will SZI'\lIS/‘r7'\\ \

 

()1 96z= JV

 

E .
. : /*

wlu :IWIS/Vl %

 

()ISLZ= 1v).

 

win 93qu

 

 

 

” (>169z :JV)

 

wru zen/v. ,

 

 

  

will Iauviééaasgs

 

 

 

 

 

 

“3N/1N

Figure 4-47. Fatigue crack initiation life for various types of film coated specimens. (Test Temperature was 298 K)

 

 

Q
r:-
-=" s‘.’
3 y g
s 8: s
‘0 .
3 8' 5
_- ”‘0‘— :-
5 .:: £
5; s‘: f
.s
: \
és
y _
‘0
3:
0
~‘

Overall Specimen
Hysteresis
Superelastic Thin- L
Film Hysteresis (T d>>Af) —,‘1——H— -'
(lying above an active PSB) f :5 g

PSB Hysteresis ;
g .5...
Y F

Superelastic Thin-
)

 

 

 

 

 

 

Film Hysteresrs (Td~Ar)

(lying above an active PSB g.- g.
x 4: 5

‘ F 1 ‘ :3
§ 3

5 e 5

5 t .=

g s

5 s e 5'

—-' e 3 :

3 Q 3

:- ~' 3 5

~‘ : r

.3.: 3

s ‘

s e‘i

s c E

: 0 ¢ 3

' . ~ 0

3 s 5

3 e
5 ~ :
.5 gigs“ J

 

Figure 4-48. Hysteresis loops representing the assumed cyclic stress srtain behavior
of the overall (inner, narrow 100p), PSB (outer,wide loop) and presumed superelastic
film lying above the PSB. The critical stress for the formation of stress-induced

martensite is higher at Td>>Af than at Td~Af.

198

 

 

 

 

 

 

 

 

 

 

_ 40000
Annealed 673 K for 20 ‘
- min. after 10,000 cycling
2 H — 30000
Untreated s imen
s _
0
Z
Z -— 20000
Bare Cu
1 -
— 10000

Figure 4—49. Fatigue crack initiation life for A/TEB type ﬁlm coated
specimens, one kinds specimens fatigued until fatigue crack initiation
occurred and the other annealed at 673 K for 20 min. after 10,000 cycles,
and resumed the fatigue test until fatigue crack initiation.

IN

APPENDIX

Electrical Resistivity Curves

.::E E c2 EBEomEoH ho :05:an a an ﬁgamaom .8585 ._ -< Emmi
mogﬁomﬁor—r
0°02 00¢ 9.87 015.

v~ moN HF» MON

 

M mnm M MEN m5 M m9

(mun Kenwv) KIEAIISISGH

200

Do 03

M mom

.EE mm c8 258$th ho :ouoSE a mm @3553“ .8585 .N-< 853m
wouﬁﬁomﬁoh
Do 0 Do Sﬁn

M Mbm M m:

00 omﬁn

 

Ems

(sum Kanwv) KHADSISGH

373 K

273 K

173 K

1

1

 

 

 

 

   

 

 

 

 

  
 
    

   

 

 

 

 

 

         
 

 

 

 

   

 

   

 

       

         

 

100 °C

 

 

    
  

  

     

     

       

   
       

     
 

                   

   

 

 

 

       

Function of Temperature for F4 Film.

00

Temperatures

istmty as a

 

 

 

 

 

 

 
        

~100 °C
Figure A-3. Electrical Res

 

      

 

           

I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
I I I I I I I I
I I I I I I I
I I I I I I I
I I I I I I I
I I I I I I I
I I I I I I I
r I V r r 1 1 r
I I I I I I I I
. I I I I I I
I I I I . I I
I I I I I I - I
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I I
I I I I I I I I I I
I I I I I I I I I I
...AL..........5..I.......u.I....II-.I...I...I..I..........JII......I..\....I... .......A..........J-I.I.......o..........l......
I I I I I I . I I I
I I I I I I I I I I
I I I I I I I I I I I
I I I I I I I I I I
I . I I I I I I I I
I I I I I I I I I I
I I I I I l I I
I I I I I I I I
I I I I I I I I
I I . I I I I I
I I I I I I I .
......I..-J...I .... ..........s....-.....I...I..... “II”....I..I.....5II........I..I.... .7 ......I..J.. .....III.....I..I..s..”.... II.
I I I I I I I
I I I I I I I I
I I I I I I I
I I I I I I I
I I I I I I I I
I I I I I I I I
I I I I I I I I
I I I I I I I I
I I I I I I I I
I I I I I I I I
I I I I I I I I I
...-......-huv-naun-I-ﬁ-A-vI-IIv-t-I-An....al.oo.-..-..0.I.--QII-I ..IDQIQIC.‘IIIICIOVI 5 ...-I-ﬁ-vI—I-II.-§----1-.I ..-
I I I I I I I . I
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
t I I h D I l I I
I I I - I I I I I
I I I I I I I I .
I I I I I I I I .
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
......I..I....................II.........I..I_-.II.II...._I.II.... ............I................. II........I ..IIII......... ...
I I I I I I I I I
I I - I I I I I I
I I I I I I I I I I
I I I I I I I I .
~ I l O I u I I
I I I I I I I I
I I I I . I I I I
I I I I l I I I I
I I I I I I I I .
I I . I . . I I
I I I I I I I I
......I... ......I..........I..I ......I....... III.I.I..I...........U......II.......II I...-III.........I.II....I...I.III.... ...
I I I I I I I I I
. I I I I I I I I
I I I I I I I I .
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
I I I I I I I
......II. ......I. ......I.. ......n... I ....-o......... .....IHIA t ......IIIII....I...J..II....III~.II...I. II.
. I I I I I I
I I I I I I I I I I
I I I I I I I I I I I
I I I I l I I I I
. I I . . I I I I
I I I I I I I . I
I I I I I I I I I I
I I I . . I I . .
I I I . I I I I I
I I l I A v 1 I
I I I I I I I I
...II.... IIIIIII. ..II.II.II,......I...,...I...I. ....II..‘..II.II.I.I, . .I....I.II,II....I...~....I......,....I.I. III
I I : I I I I I I I
I I I I I I I I . I I
. . I . I I . I I I I
I I I I I I I I I -
. I I I I . I I I .
I I I I I I I I . I I
I I I I . I I I I I I
I I I I I I . I I I .
I . . . I . I . . I
I . I I I . I . . I I
”II—...“....-....I...I.........I..I.......I.I..I.....IIII........IIII..I..... .......I............I..I.I.....I...............I ...
I I I . I I I I I I I
. I . I I I I I I I
I I . I I . I I I .
I . I . - I I . - I
I I I I I I I .
I I I I I I I I
I . . I I I I I
I I I I I I I
I I I I I .
I I I I . .
......I... ...-IIIIIIISII—III-III}III-.II.-I I-I-{I-II-IIIIII........ .I...—I.»II.IIIIIIII-II'—I......II.),II..... .-.
I I I I I I I
I I I I I I I
I I I I I I I .
I I I I I I I I
I . I I I I I I
. I I I I I - I
. I I I I I . .
. . I I I I I
I I I I I I I I
I I I I I I I I
I I I I I I I I
I I I I I I I
. . I . v I I
I . I I I . I
I l I I I I I
. I I I I I I I
I I I I I .
I I I I c
...IIIII.I‘...I.I.I.I I... ...—u‘ I. ..,...I..... ‘ .I.................. .........,I,. ....I ...
I I I I I I I .
I I I I I I . I .
. I I I I I I I I
I I I . I I . I I
I I I I I I I . .
I - I I I I I I . .
. I I I I I I I I I .
. I I I I I I . . I .
n ‘ ’ I
.IIIIIIIII<.IIIIIIIIII.}IIII IIIIIIIIII‘I II- 2'..- ’IIII.II..I}- II 'IIIIII-IIIc: IIhIII... .....L.... III...
I I I I I I . I I I I
I I I I I I I . I I I
I I I I I I I I I I I
I I I I I I I I I I I
I I I I I I I I I I I
I I I I I I I I I I .
. . . I I I I . I
I I I I I I I . I
.-.-nuu-nulInvII-y-a-»\-~-- -II<-----Il.---------‘no. ...1.-— I .-.<-II.-l“lI‘ll.<'.—’.<i4.'lli. \I-Io->-«» ..1-
. I I I I I I
I I I I I I I I
I I I I I I I .
I I I I I I I I I
I I I I I I I I I
I I I I I . I I .
I I I I I I I I I
I I I I I I . I .
I I I I I I I I I
.-...I—I...I...........I .....-I...: III... 0 . ~..........4I-. ....4.......... II-..
I I I I I I I I I
I . I I I I I I . I I
I I I I I I I I I I I
I I I I I . I I - I I
I I I I I I . I . I I
I I I I - I I I I . I
I I I I I I I I I I I
I I I I I I I I I I I
I I I I I I I I I I I
...-...”.-...I.........IIII.I-.III O ...-.-.... ......I.......I.......,................I..II.I.... ......I.I.II.
I I I I I I I I I I I
I I I I I I I I I I I
I I I I I I I I I I I
l - I I I . I I I I I
I I I ﬂ I I I . . . . I
I I I I I I I I I I
I I I . . I I
..-...III ......III.s......... II. ﬂ ....J....II. ......I...t...IIII...AI..I..IIIIJI.....I.» I. ......s.....
I I I I I I I I
I I I I I I I I .
I - I . . I I .
I I I I I I I I
. I I I I I . I
I I I I I I I I
I I I I I I I ~ .
I I I I I . I I I I
.....-...I‘...........,..I.......,.- .‘ ,I.........,...II...I.,..........,..........‘IIIIHII...,I..I.....I,. ...-
I I I I I I I I I I I
. I I I I . I I I I
. I I I I I I I . I
I I . I I I I I I I
. I I I I I I I I -
I I I I I I I I I I
I I I I I I I I I I
......I.......-......... I..III..... -.......I....... ..II—IIIIIIII , -. I... ............I.......I..I............._........I........II....I
I I . I I I I I I I
I I I I I . I . I .
I I I I I I I . I I
. I I I I I I I I I
I I I I I . . I I I
I I I I I I I I . I
I I I I I I I I . I
‘ 1' I 0 a 1 I
-....-... I...III.II I.....I...,I-..I.II.. ......I... ....... v' ....'.......II.‘....... . ........,......I.I
. I I 1 I . .
I I I . I I I
I I I I . . . .
I I I . I . I .
I I I I I I I I
- I . I I I I I I .
A I I I I ‘ A A g A
v v v v V I I w I I
I I I I I . . .
I I I I . I I I .
I I I I I I I I .
I I I I . . I I I
I I I I I I . I I
I......III.IIII.II.I.I................I....I ...IIIIIII ..I. ....-I I I.........II.....I-... ......I.....I.II..I... I.
I I I I I I I I I
I I I I I I I . .
I I I I I I I . I
I I I l I . . I .
- - - . I I I .
..vv-uv-e-‘»-.II.II.I I.......I-..I-II.-II- ..Iu-II-‘I I I—III-Il-II-I-II-.d—II.-I..II-.I ......—.l. ..v‘
‘ ‘ I I C I n l
I I . I . I . I I
‘ ‘ ‘ 0 I l 1 l I I
' ‘ ‘ ' ' b I I I
I I I I I I I I I
III-...... ... I.”.}.I.....I.I'\.I.....II. ....II........ ......II ( ....IIIJIIUIIIIII.)_..I-.I.II .....I....5....I.
I . I I I I I I
I I . I I I I - I
I - - . I I I I I
I I I . I I I I I
I I I I I I I I
.......................,..........,.......... .........,..........I, ‘ .......‘.......... ..........,..........;......
' ' ‘ - - I I I I
' ' ' - - I I I I
- I I . I I I I I
' * ' t - I I I l
I I 1 I I I I x I
. I I . . , . .
I I . I 4 I I I
' ~ ' t v I I I
I I I I I I I I
I I I I I I
I..........‘I..III..I.., .I,..I-.....I ......I...... .,_.......I .‘I.........I‘,.........I;I.I.I.
I I l . I I I .
I I I . I I I I
I I I . I . I
I . I I I I I
I I I I I I
II...........I....... ......I........ I. ....v.. .....,.._...... “pH-”n“ .__ , n,
' ‘ v v I
I I I I I I I
I I I I I I
I - I I I I I I I

 

 

-150 °C

 

 

(sum Kentqw) mAnsrsaH

.::E 98“ M}. be 238388. “0 :ouocsm a mm @3363” .8585 .v.< Eswi
3:582”th
U0 OCH 0 Do OCH: Uo omﬁu

 

 
 
 
 

 

 
 
 
 
 
 

 

202

 

 

 

 
 

   
  
 
   
 
 
 
 

 

 
 

 

 

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 
 
 

 

 

 

 

 

 

 
 
 

 

 
 
 
 
 
 

 

 

 

 

I I I I I I I J I . i I‘ I I I ijl I
I I . . . . I I . . . . _ I . I . .
I I . I . . . . . . I I . . I .
I . I . . I . I . . . I . . I . .
I I I . . . . . . . I . . I . I . .
I . . . . . . . . . . . . . I . I I
I I . . . . I I . . . . I . . . I I .
I . . . . . . I I . . i I . . . . . I .
I . . I . . I . I I . . . . . I .
. I I . . . . . . . . . .
I IIIIIIIIII.IIIII...IIIIIII ..IIIIIIIII.IIIIIIIIv..I..III.I .IIIIIIII II .IIIIII..IIIIIIII.III.I.I .. .....II ..I. III...I IIIIIIIIIIIIIII II..I III-....IIIIIIII..IIII...II IIIIIIIII v. I
I . . I . I _ . I I . . . . . I . . I I
I I . . . . I . I I . . I . . . . I . c I
. I . . . I . . I . . . I I . . . . . . .
. I . . _ . . I . . . . . I . . . . . . .
I . . I . I I I . _ . . . I . . . I . . .
. . I I . I I . . . . . . . . I . I . I .
I . . I . . I . . . . I . . . . . . . . I
I I _ I . . I . . . . . . . . . . I . . .
. I . . . I I I . . . . . . . . . I . . .
. . I . . . I . . . . . I . . . . . . .
Inuit-I'Dn-A’IulllIllIG‘II00.I0Ilbl\lllunI-lIl-§l.I-Ill-Ith-Illn. (OI-hul-‘v-Itv-IIIIulfubaI.{OI-IIIDIJIIOIII I. III-IIIII'I'IIIIII\III1:IIII\II .III IhIIuII I IIIII§III IIII'III-nlohtII-IIIO CIIUIDKVIIIIIIOVn-vﬁlwIII\IIIIIO’AIIl\ l-II‘IIII 'x-n
. . . I . I . . . _ . . . . . . . . I .
I . . . I . I . . . I . . . . _ I . . . .
. . I I . I . . I . A _ . . . I . I . . I
. . I I I . . . I . . . . . . I . _ . I I
I . I . I I . . I l I . . I . . . . I I
I . . I . . I I . _ . . I . . . . . . I .
. . . . . . . . . . . . . I I I . I . I .
I I I . I . . . . . . . . . . I . . . . I
I I . . . . I . . . _ . . I . . . . . .
. I I . I . I , . . . . . I . . . . . I
. I I . . . . . . . . . . . _ .
I IIIII IIIIIIII.IIIIIIIIIIIIIIJIIIIIIIIIIIAIIIIIIIIIIImIIIIII IIIIIIIIJIIIIIIII.II\IIvIIIIIII.IIIIIIIIII II IIIIIIIIIwI.IIIIIIIIIIIIIIIIIJIIII.II IoI.IIIIIII~IIIIIIIAIII .IIIVIIIIIII IIIIII IIIIIIAIIIIIIoIIv IAIIIIIIIIIIIIIHIIIII I.III..IIIIII¥IIA
. I . I . . . I . . . . . I . . . . I .
I I _ I I I . . . . . . . . I . . . . I
. . . . . I . I I . I . I . . . I . .
. I . . . . . . I . . . . . . . I . _
I I . . . . I . I . . . . I . . . . .
. I . I I I I . . . . . . . . . I I
I . I . . I I . . . . . . I . I I .
I . . . I I . . . . . . I _ I I I I
. . . . I I . . . . . _ I I . I I .
. I . I I . . . . . . I I . . .
IIIIIIIIIII II.IIII..I.a.II.IIIIIII4III.IIIIIIIIIIIIIIIIIII-IIIIII IIII.IIIJIIIIIIIIIIIQIIIIIIIII \......I I. .IIII..IIIIIIIIII.IIIII.III.4II..IIIIIII. ..i. II.IIII1III.IIII~..I..I. ..IIII I.IIIIIIIII...IIIIII..IIIIIIIIIIs..IIJ IIIIJ
. . I I . . . . I I . I . . . . .
. . . . I . . . . . . I . . . .
I I I . . . . I I I . . . I . . .
_ I . . I . . . I . . I I I . . I
I I . . . . I . . . . . I I . .
. . I . . I. . . . _ . I . I c I .
I . . . I . . I . . . . . . . .
I . . . . . . I . I . . . . . .
. . I . , . I . . . . . . . .
. . _ . I . . .
I. I... IIII. IIIIIIIIIIIIIIIIIIIII II. IIIIIIIIJIII.IIIIIIIIIIIIIIIIII .III II. IIIIIIII III. .III . ..IIIIIIIIIAII ..I.... IIIIIIIIIII .III .....lIIII I... ..IIl.
. I I . I . . . . . I _ .
. I . . . I . . I . . . . I . . .
. . I . . . . . . . . . . . c
I . . . I . . . . . I I . .
. . I . . . . . . I . . . . . .
. I . I . . I . . . . I . I . _ I
. . . . . . _ I . . . . . . . .
I I . . I . . . I . . . . . I
. . . I I . . , . . . . I I ., . .
. . . . I . A I I . . I I I . .
llbI-IIODI- rIII.IIIIIltn.IIIDIA-in‘IIOIIDIIII-§!.IIIIIOIII I-II .IIIIOI I II-II‘IIIIII-II\IDIIIIIII VIIIIIII(III .I.III-IIIIDIIII-IIO'IDI'DILVDIIIODIIIIIIIIIID ' .I'IIA
I . I . I I . I I I I I I I .
. . I . . . . I I . I . I . .
I I . I I I I I I I . I I .
I I I I I I I I I I . . I I I
I I . I . . . I . . I I I I .
I I . . i . . I . I . . I . I .
I I I . I . _ . . . . .
. I I . . . . I . . . l .
. . I I . I . . . I . I
I . . I . . . I .
IIIIIIII.IIIrII I IIIIII I III ItIIIIIIIIIIIII.IIIIIIIIIIIIIII III I V II It...
I I . I I I
. I I .
“ . . . .
‘ I . . . . I .
I _ I I . . . I
I . I I I I I .
I I . . . I I I I I .
I I ' . _ . I . I I I I .
I . I I I I . I I I I . .
I I I I . I I . . . I . I .
. . I
I A h III IIIIII V « IIIII J V IIIIA. IIIII I— IIIIIAIIIIIIII‘I II III‘IIIII IIIIIIIIIII III-I IIIIIIIII\III
I I . I I . I . I I I . . I . I .
I I I . I I . . . . I . . I . . . . . .
I I . . . I I I I I I . . _ . . . . . .
I . . . I I I . I . . I . . . . . . . .
. . I . . . I . . . . . . . . I . . . I
I I I . . . I . . . . I . . . . . . . .
. I . I I . I I . . . . . I . I . I I .
I I . . I . I . . . . . . I . . I . . .
I . . . . . I . . I . . . I . I . . I .
I . . . . . . . . . . _ . . I I . . I I
IIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIIIIIIIIII.IIIII..III.IIII.II.IIIIII IIIIIIIIIIIIIIIII.III..III.IIIIIIIIII.IIIIIIIIIIIIIIII .II..I..III...IIIIlII.II..I.AI. ....III.IIIIIIII.III.II.IIIIIIII..I .III IIIIIIIIIII.IIIIIIIIIIIIIIIII..\IIIII IIIIJIIIIIIIIII
I . . . . . I . . I . . . . I . . . I I
I . . I I I . _ . . . . I . . . . . . .
I . I . . I . . I I . I I . . I . . I .
I . . . . . I . I . . I . I I I . . . .
I . . . . . I _ I . I I I I . I I . . .
I I . . . I . . I . I I I . I . I . . I
I . I . . . . . . I I . . . I . I I .
. . I _ . . I I . . I I . . . . . . .
. I . . . . . . . . I . . . . . . . . .
I . I I . I I . . . I . . . . . . .
IIIIIIIIII. IIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIAIIIIIIIIII IIIIII IIIIIIIIIII.IIII....IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII I.III..IIv..IIII. .I..... ..I.. .IIIIIIII .IIIII..III..II.III IIIIIIIIIIIIIIIIIIIIII. IIIIIIIIII. IIIIIIIIIIIIIIII
I I . I . . . I . I . .
I . . . I . I . . . . . . . I
I . . I . . . . . . _ . I . . .
I I I . . . . . . . . . . . . .
I I . . . I . I . . I . . . I .
I I . . . . . . . . . . I . .
I . . . . . . . . . . . . . .
I . I . . . . . I I . . I I . .
. . . . . . . . . . . . I .
I . . . . . . I I I . . I
IIIIIIIIIIIIII.IIIII.I.IIIIIIIII.IIIIIIIIIIIIIIIII.IIIIIIIII IIIIIII ..IIIIIIIIIIIIIIIIIIIIIIIIIIIII ..IIIIIIIIIIIIIIII II.IIII.I9..I.II.I|I..I.... ...I.. III... ..I..I. ..III...II.. III. IIIIIIIIIII ....II..I. II..III.II IIIIIIIIIIIIIII
. . . . I I . I . . I . . . .
I . I . . I . . _ . I I . . . . . . .
. I . . . . . . . I I . I I . . . . . .
I . . . I . . I . . . . . . I . . . . I
. I . I . . I _ _ I . . . . . I I . . .
. . . . . . . . . I I . . . . . I . I I
I I . . . . . . . . . . . _ I . . . . .
. . I . . . . . . . . . I . . . . . I .
. I . . . . I _ . . . . I . . . . _ . _
. I . . I . . . . . . I . . . . . . . . .
IIIIIIIIIII‘rIIIIIIIIIIs.IIIIIIIIIIuIIIIIIIIIIIsIIIIIIIIIIIuIIIIIIII IIIIIIIIIxIIIII.IIIII'IIIIIIIII IIIIIIIIIIIIIIIIIIII IIIIIIIIIrI.IIIIIIKIIIIIIII.IIIIIIIIIIIIIII. IIIII.LI.III.II.-III.III-IIIIIIIIIIIII.ILrIIIIIaIIIIIvaIIIAIIIIIIuIIIIII IIII\IIIIIIIII;
I . . . . . . . I . I . . . . . .
I . . . . . . . . . . . . . . . _ . . . I
. I . . . . . . I . . . . . . . . . . _ .
I I I . . . . . . . . . . . . . . I I . .
I _ . I . . . I I . . . . . . . . . . . .
. . . . . . I . . . I I . . . . . . . . .
I . . . . . _ I . I I . I . . . . . . I .
I . . . I . . _ I . . . I . . . . I . . .
. _ I I . . . . . . . . . . . . . I I . .
. . . . . I I I I . . . . I . . . . . . .
a u I I A I I u I u . - n - n I t
IIIIIIJIIIIIIIIIIIAIIII.IIIIIIaIIIIIIIIIIITIIIIIII IIIIIIIIIJIIIIIII. IIsIIIIIIII. IIIIIIIJIIIIIIIIII IIIIIIIII~IIIIII.II.IIIIII..II..IIII.II..I..I. I..IIIJIIII.III\I.IIII.‘I.IIIII.‘.III.IIIIIIIIaIIIIIh..IIIJIIIIIINIIIII III IIII\.II
. I I . . . . I _ I . _ I . . . . I . . .
I . I I . . I . . . . I . I I . . . . . .
I . . I I _ I . I . . . . . I I . I I I I
. I I I . I . I . . . . I . I I I . . . .
. I . . . . . . I . . . I I . . I . .
I I . . I . I . . . . . . I I _ . I I .
I I . . I . . . . . . . . . I . . I I I
I I . . . . . I I . . . . . I I . . I . .
I . . I . . . I I . . I I . I I . . . .
I I . . . . . I . . I . I . . I I . I I

 
 

 

 

 

 

 

 

(81mm Kemqlv) KJIADSISGH

 

203

Do o3

.EE 9?: EH}. 8.. BBEKEPH. ..o =08:an a ma bﬁaﬂmom .8585 .m-< Bswi
8.53.8;th
o o

 

 

 

 

1
u
.
~
A
A
.

 

 

A
.
.
.InIA-nutuAAr...I..I.».‘p.
.
A
A
A
.
1...... .....1-..uJ.A
.
A
A
.
A
.
A 4
A
A
.
na...s....- In..u..-.A.ln.
A
A

...:nn......

IAu-.I.-cou.

.::-....)

v......n...

rAvuo-ncnuA

 

I...ou.¢u.u

 

rouIO'no-nc

 

 

 

 

........~l....--u~--

  

.
.
A
A

s
.
A
A
A
.

J
A
.
o
u
.
.
.
.

...¢--..-Su‘..o.--.-Q.u.
.

 

I

 

 

A . 4 4 4 4 4 4 1 1‘ A 4 4 4 4 < q 1 4 . A
A . A A . A A A A A A A A . A A A A A .
A . A . A A . A A . A A A A . A . A
A A A A A A A A A A A A A A A A A . A
A A . A A A . A A . A . A . A . A A A
A A A A A A A A A . A A A A A A A A A
A . A A A A A A . A A A . A A A A A A
A . A A A A A . A A _ . A A . A . A A
A A . A A A A . . A . . . A . A A A
A A A A A A A A . A A A A A .
...... ..AAA...A.........A.A....:::...-..A...AAA.......I-..2....-.......A.....A...-|..........A. .....A...vA..-. A .....- .A ..-..A .....AA.............- ..:......v..... AAAA’AIIAA.-.
A A A A A A A A A A A A A A
A . A . A A . . A A A A . A A A A A A
A . A A A . A A A A A A A A A A A A A
A A A A A A A A A A A . A A . A . A A
. A . A A A A . A A . A A A A A A . A
A A A A A A A A . A . A A A . . A A .
A A A A A A . A A A A A A A A A . A A
A A . A . . A A A A A A A A A A A A A A
A . . A A A A . A A A A A A A A . .
A A A A A A . . . A A . . A A A A A A
pl.|I.l-p.un¢h-ltlvIII-utilnutua-a ..... In.|tnulun~¢l\¢n u'lnslnn'n»!.AI-D‘-hAnal-n.n|U)IC»-lclu-I Il-bocAJAI ........ \h~.l4.¢tn\ ......... tr ...... opnnul'.‘ .alu- In. 1‘71! 1 uti...»:u\ I’Iu:u-Allnl\ A
A . A A A A A A A A A A . A A A
. A A A A A . A A A A A A A .
A A . A A A A A A A A A A A
A A A A A A A A A . A A A A A .
A . A . A A A . A A A . A A A
A A A A . . . . A A A . A . A A
A A A A A A A A A A A A A A A
A A A . A A A A . A A A A A A A
A A A A . A A A A A . A A . . . A
A A A A A A A A A A . A A A A A
A A A A A A A A A
.:::...1: VA...A2.........-....I..-.....4......A....w-.A......A. A- .........~.A...A.A.J... A ......... A... ...... ..::... . .-.; ... .v..;
A A A A A A A . A A A A A A
A A A . A A A A A A A A A A A
. A A A A A A A A A . A A A
. . . A A A . A A . . A A A
. A A A A A . . . A A . . . A
A A A A A A A . A . A . . A .
. A A A A A A A A A . A A A A A
A A A A A . A . A . A A A A .
A A A A A A A A A A A A . . . A
A . A A A . . . A . . A A L
.....4....... .-.... 1AA... .. AAA: ...s... ...-... .AA AA... ...uAAAAAIAAAZI. 24.2. A... A ..::? ....-.‘.A..A...A...A. 3.... is:
A A A A . A A . A A . A A
A A A A . A A A . A A A . A
A A A A . A A A A A . A A
. A . A A A A . A A A A
. A A A A A A . . A . A .
A A A A A A . A A A A A A
. A A . A . A A A A . . A A
. . A A A A A A A A . A A
.. -...A A. A. ..............2-..............-..v..y....AAA. A- .........n....-...I......AA.........A. A 2.13.... ... A.-. I... . A :.:
A . . A A A A A A .
. A . A . . . A A A A .
A A A . A A A . A A A A A
. . . . a . A A A A A A o I.
A A . . A A A . A A A . .
. A A A . A A . A A A . A A
A . . A . A A . A A .
A A A A . A A A A
. A A A l A . A A
A A A A I . A A . A
. . A . A A . I A
A A A A A . . . A A
. . A . A A A A A
A . A A A A . A .
. A . A A A . A A . A
A A . A A A A A . A A A
A . A A . . A A A A A A
. u u u . A 1 u 0 ~ 1 a
A A A A A A A A A A . A
A A A A A A A . .
..I.-t.....o. p A - Ar ...: ... ...».A -A¢.Au.. A. .. ...... ..\auv-u ....n .A..\..-A..-r..;
A A . A A A A .
. A A . . A A A A
A A A A A A A A . A
A A . . A A A A A . A
A A A A A A A . A A
. . A A A A A A A A
t o u - v A . u - A -
~ g D c ‘ o . a o A o
A A A A A A A A A I.
. A . . A A A A
A A . A
a a v ’ ......A-..2.2.1.1..........A...-. ... .... .....A..
. A A . . A A A A A A A
. A A A A A A A A A . A . .
A A A . A . . . . A A A
A A A A A A . A A A A . A
A A A A A A . A A . A A
A A A . A A A A . A A .
. A A . A A A A A A A A A
A A A A A . A A A A . A
A A A . A A A A A A
A A A A A . A A A
..::.-. ...... ........-.... < 3.1:.-. A 1.311.... ........I..-...\..-.A ....1-..
A A . A A A . A A A A A .
A A A A A . A A A A A A A
A A A A . A A A A . A A A . A
. A A A . . A . A . A A A A A A
A A A A A A A A A A . . A A A A
A A . A A A . A A A . A A A A A
A . A A A . . . A A . A A A A A
A A A A A A A A . A A A A A A A
. A . A A . A . A A A . A A A
v u A o A u n . » . .
.....AAA......... ..-...A....-.A.......l... A A A.. . . A. . 2... AA... .-.. ....A.......... ....A ....I.
. A . A . A A . A A A
. A A A A . A A A A A A A A A . A A
A A A . A A A A A A A A A A A A A A A A O
. . A A A A A A A A A . A A . A A A A .
A A A A . A . A A A A . A . A A 1
A A . A A A A A . A A A A . A A
n n o . n u a . . l m A ~ . n . - - . S
A . A A A A A A . A A . A A A A
n n . v . ‘ . . I u — u w . 1 n
A A A A . A A . A A A A A A
......A. .-.. .......A...A u..-..A.....-.A....::::...-...A.-..A...:A\.A.. :.:-..::...Ar.A-.....\.A..... . .....a.r......Av.... ......-A.A\....A .... ..-.|.. (
. A . . A . A . A A A A . A A
. A . A A A A A A A A A A . . A .
. A A . A A A A A A A . . . . . A A A
A A A A . . . A A A A A A A A A A A A
. A A A . A A A A A A A A A A . A A A
A A A A A . A . A A A A . A A A A A A
A A A A A A A A . A A A A A A A A A .
A . A A . A A A . A . A A A A . A A .
. . A A . A A A A . . A A A A A A A A A
A A . A A A A . . A A . A A A A .
|.-‘n\.1 .I-A-nbl.AminoanI-v.«vnu- ..cvlnsI-nfoIIp-tin.n.luout:|!-¥l¢n1n.n.tn¢ osnt-Ainl u.-:AA.- ho!!-.:.I.~54-nol-AAl.u<4l|.nanun ul...‘clv I - 4b..A...”:ulIuLvt-iIobl...an-Iov uvl.lfcvtnn l..t\tI-.AI|I.QA
A A A A A A A A A A A A A A A A A A
A A A A A A A A A A A A . A A A A A
A . A A A A A A A A A A A . A A A A A
A A A A A . A A . A A . A A . . A A A
A A A A A . A A . A A A A A A A A . A A
A A A A A A . A . A A A A A A A A A A A
. . . A A A . A . A . A A A _ A A A A A A
A A . A A A A . A A A A A A A A . A A . .
A . . A A A A A A A A A . A . . A A A A .
A A A A A A . A A A A A A A .
A A A A A A A A
......AA. ...-....-...-......A -.....A..-~..AA.........-A.A...- ....A.-... .:::... ...........-..A.... .......AA. .....A.... .. .....t: .......L.. A... ......A..Au.s..A
A . A . A A A . A A A A A A .
. A A A . A A A A A A A A . A A
A A A A A A A A A A A A A A A A A
A A A A . A A A A A . A A . A A A A .
A A . A A A A A A A . A A A A A A . A .
A A A A A A . A . A . A A A A . A A A .
A A A A A A . . . A A A A A A A A A . .
A . A A A A . . A A A . A A A A A . A A
A . . A A A A A A . A A A . A . A A A
A A A A A A A A A . A . A A A A A A . .

 
 
   
 

 

   
   

 
 

 

     

  

 

 
 

 

 

 
   
 

 

 

 

 

           

 

 

 

 

 

 
 
 

 

 

 

 

 

 

  

 

 
   

 

 

 

 
 

 

 

 

 

 
   

 
 
 

 
 
 
 
 

 

 

 

 

 

M mnm

M MPN

2

.::E

98“ Nmth. go. 83an8».—. “o 835$ n ma bramaum 30585 .o-< 8:me

mouzuﬁumﬁorr

 

 

 

 

 

 

 

 

 

 

                  

 

 

 

 

 

 

 

 

 

     
 

  

 

 

 

 

 

 

 

 

 

 

 

.III....

 

 

 

 

 

 

 

 

 

   

 

 

   

 

 

 

 

  

         

n
O
-
I
I
I
.

 

 

 

I
l
I
,
h

     
  

      

 

 

 

 

  

 

 

 

 

 

 

I

   

4 i J1 I [W a . I 4 11 ‘ 4 4 4 1 4JT 1 1 [W 1
. . . . . . . . V . _ . . . . . . .
. _ V . . . . . . . . . . . . , . .
. . . . V . . . . . . . . . . . I .
. . . . . . . . _ . _ V . . . . . _
. . . . . . . . . . . . . . . . . .
. . . . . V . . V . V . _ V . . . .
. . . . . . . . . . . . . V . . _
. . . . . V . . . _ . . . . . .
. . . . . . . . . . V .
II . .I. ....I II.... .. . .I.I .I...I..I. . I..I.. ..II.. ...... ...t... .....I... ......I. . .- ......v......‘vII.II......u. .l.... . . .6....0.
. V . . V . . _ . . V V
. . . y . . . . V . . , . . .
. V V . . , . V . . V V . V V
. . . . . . . . . , , . . . V V .
. V . . _ . . . . . . . . . . V V
. . . . . . . . . . . . . . . V . . .
. . . . . . . V . . . V . . . . , . .
. . . . . . . . I . . . . . V V . .
. . . . . . . . I V . . . . . . . .
. . . . . . . . . . . . . V . . . . 1
It.. It.- ..s... I. In... ... I. .6.. I... If.....I. .-..I....\I....II.II. ....I . .\,.... I. .~. ......0.. u. ..I. .0. . I .....sI .\I.....l... .oI Is. .‘I..nr.
. I . . , . V . . . . . . . V V
. . V - . . . . . . . . V . V . . I
. . . . . . . . . . . . . . . . . . .
. . . . . . V . . . . . . . . . . . .
. . V . . . _ . . V V , . . . V . . . .
. . . . . . . . . . . _ . . _ . . . .
. . . . . . . . . . . . . . . . . . .
. . . o . . V . . V V . . . . . . . .
. V . . . . . . . V . . . . . . . .
. . I V . . . . V . . . . . . . . .
. . . . . h _ . V . .
.o A. ..JI . .9... ..I.. . .Jv. I... .‘II...-I.II<II.......o..oI.I..Ir.I. 5;. 4.. . If. . I. I.V.I,,.VI.~I ..I...I.\.. I. I I4 .I..I.. Iv
. . . . . . . . . . V V . V . .
. . . . . . . . . I . V . . . .
. . V . _ . , V . . . . .
. . V . . . V . . I . . I
V . | H . . . . . . . . . . . I
V . I . . . . . . . . . . . . V
. . . . . . . . . . . . . . .
. V . . . . . . . . . . . . .
. . . . . . . , _ . . . . .
. . . . V I . . V . V . .
..IIA1I.I«-I..I. ..1......I..I. .:.nv II... I. I ..I. A .....-..\..II. .1 I... u .. .o....anv.‘v. ..IIII ..‘.I... J... ‘.
. I . . . . . . . . I V
. V . V . . . . . . .
. . . V . . V . . . . .
. . V . . . . . . .
. . . I . I . _ . . .
. . . . . . . . .
. _ . . I . . . . .
. . . . . . . . . .
. . . . V . . . . . . . . .
I... IIIIIII~caIAa IunIaIIIIIIII. .IIoo In... I ...1 Iv I..I'.II.. Inna-.-.1‘f..a....v...:I.I. I .::..I....V|-..~I. ....I... ......II.IQII ..I.u-u II I! I I...I.
. . . . v . . . . .
. . . . . . . . .
. . . . . V .
. . . . . . .
. . . _ . V .
j . . . . . . .
. . . . . . .
. . . n . . .
v . . . . .
. . .
.
V

 

~>o—-----‘9.-¢-

 

1I-IIII..

1......I.

rIuI-

 

~.........VL.

rIv..

 

   

.
V
.

 

 

 

 
 

 

 

I
b

 

-
I
I
I
I

I
a
I
c
l
.
I

I
l
I

 
 

..........J..........Vy...VV.............-..J..........O....-.....A-.........n-...

I
u
I
I
a
C
l

I
a
-
I
I
-
.
a
u

I
I
I
I
I
.

I

..VI..VA........-.-......,...;.....................
.

a

 

.
s
.
.
.
.
.
.
A
.
.
.
.
.
.
.
.

.-..«......—....

 

 

.
V
.
.
.
.
.
.
.
.
V
.

 

 

 

 

 

ﬁ

I

.
s
.
.
.
.
.
J

 

I
t

 

QI.......-.‘..gunno-..J-c

 

 

..........J-..........\.

A
l

 

 

..........‘.-V.....-..,—V..V..V.V,.V.............--.....‘........V.‘...........,V.

 

,

~

I
u
u
I

I
.
l
I
I

 

.
o
.
.
.
.
.
.
.

tI-aI-IIIII .

 
 
 

 

 
 

.
.
.
.
.
..I..\III......I. .II...I.
.
.
.
.
.
.
V
.
.
.
u
.
.
.

 

I
I

 

V......,..........,......,._...........

I

---‘-~---~v---

v

 

.
.
.
.
.
.

.I..I-..

<5.

  

 

.
.

 

I
.-...V....J.....I..
I

I . . I

..\.....—..—_|.V........l’u.......Vl.--....-..J.-...-

    

 

  

 

 

 

I. lKl'n-l-b
.
.
.
.
.
IIu'n obbu. a
.
V
.
.
. I
. . I
. .
.
. .
..-.....~.........T.... J...
. .
. .
. .
. .
V .
. .
. . .
. . .
V
2:21.... ..

 

 

l
..........c-....V

 

{

,

V
IaIIIII...
.
_
V
_
II. .....c
V
.
_
.s.....

I
o
I

 

 

”......I,”.........--..-.

V

..........,..V..
I

     

~
I
5
D
.

 

..,-.........,V.-..v..

 

-..V....4......

.‘......-

 

 
 

 

. . l

I I ,

I . .

l I I o

. . l I

I l o I
I

....-.....‘...........,..........,.....-.V-..'...-..V...‘..V..

 

.
I

c
..ﬁ.......u

 

 

 

 

 

 

....J
I
I

 

 

  

 

I
...—...-.....,.......

u......a-—.’.._........\

>-.......-.

I
v
.‘....

.
. .
. .
V .
. .
II..IIII-.
. .
. .
. .
. .
. .
. .
. .
. .
. .
.III-.III

 

         

V
O
.
.
.
V
.
t
.

...V.....VI.V.V......
I

§v«n-.-I-~- a......’

.........IJ...........s........

 

  

O

I
I

....,...I....

u
........V...‘..I.V......,....VV

 

 

(sum Kenlqw) KHADSISQH

.EE 9:: mum—...? V8 Baﬁomﬁoh ho coaocam a mm @3563“ .8385 .b-< 2:me

00 OS

85222509

Uo

OOOMI

Uo OWMI

 

 

 

 

 

 

 

 
 

 

205

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

  

 

 

 

   

 

 

 

 

   

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

   

 

 

 

 

   

 

  

 

 

 

 

 

 
    

   

 

 

 

 

 

 

 

   

 

 

V V 4 I V 1 4 V 4 I 4 V 4 1 q 4 I
V . V V V V V V V V V V V V V V V
V V V V . V V . . V . V V V .
V V V V . V V V V V . V V V V V
V V V V V V V . V V V V V V _ V
V . V . V V V V V . V V V V V V
V V V V . V V V V V V . V V V V
V V V . V V V V V V V . . V V V
V V V . V V V V V . V . _ V V V
. . . V V V . V V V V V V
I IIVVVVIVVIIVIVVIVI IVIVIVVVI VVVIVVIVIVVVIIVVIIVV uV.IV..IVV ..V....VVII.II...VIII .VIVVIIV .VVV. . IVVIVIVIVVIVIVI VIVVIVVIV. ...... v. ..... VIVVVV V II ...... .VVVIV.o.. ......c ..I. ..1 IVVVIVII
. V . V V . V V . . _ V . .
V V . V V V V V V V V V . .
. V V V V V V V V . . V V V V
. V . V V V V V . . V V . _ .
. V V . V V . . . . . V V V
V V V V V V V . V V . . V V V
V V V V V V . . V V . V V _ V
. . V V V V V . . . . V V V V
V V V . V V V V V V V V V V V
V V . V V . _ V V V . V . V

IIIIVVVIIVV r. VIIVVI.V\.VIVI I.\II...VIV .I ... VVIVV-VVIVVVIVVVVIVVIV I V ..I. V I.\V .IIVI. V. IVVIVVVVI'VVII IIVS V V. .yVVVIV.sV V ... ..VVV.rIV.VV ...V ..r.. ..VV. ... . \ ..IVIV V
V V . V V V V V V . V V V V
V V . . V _ V . V V V V . V V V V V
V . . _ V V V V _ . V . . I . .
V V . V V V V . V V V . V . . V
V V . V V . V V V V V V . . V
. V . V V V V V _ . V V . V V V
V . . V . . V V V V . V V . V . V V .
V V V V V . V V V . V V V V . V . . V
V V V . . V . . V V V V . V V V V V . . V
V V V . V V . . V V . V . V V V V V . V V 5
. V . V . V . V V V . V

.-....VVVVI ........V...;IIVIVIV.1?...VVVVVVV ..I....VVLIV....:.:... ..... I.,.VV....V.... .....V. ....... ........-V....V.VVVV....V...)VVVVVVVIVVIVZVV....VVV.IVVVV.......V.... ..u. .VVIVVV ..- $.33... ......V...VV.VVV.
V V . V . . . V V V . V . V V I . V V
. V V . V V V V V V V . . V V V V V V V
V V V V V V . V V . V V _ V V V V V . V V
V V . V . . . V V V V V V . V V . . V V V
V V V . V V _ . V . V V . V V V V V V V V
V V V V V V V . V V V V V . V . . V V
V V . V V V V V V V V . V V . V . V V
. V V . V V V V V V V . V . V V . V
V V V . . V V . V V V V V V V V . V V
. V V V V V V V . V V . V V V V V V V

VIVaVVVIVVII. 4 VII... 4 ..IVIVVIVVII..IIIJVIVVIVVIVVI\.IV‘V.VIIVVVVII..V.V1..VV.V.V... VIVVVV.II‘IVV.III-l.IVV.IVIIaVIVVIIIIIoIVVVIV:VoIVIII.VaV..I....‘VVVVVVVVVV I...V\I .IVV a II
. V V V . V . . V V V V V V V V V V
V V V _ V . V . V . V V . . V . V V
V V V V V V _ V V V V V V . V V V
V . V V V V V V V V . . V V V V V
. V V . V V V V V V V V V . V .
V V V V . V V V V . . V . . V
V V V V V V . V V V V V V V V
V . V V V . . V V . V V . . V
V V V V . V V V . . V V . V
V V . V V V V
IIVVIIIIV. VVIVVIVVIV.III VVVIVVIV. .III.V...VV.VV. .VIIV V ..V.. VIVVVIVII II... V... I III... ..IIVVII
V V V V . V
_ V V V V V V
. V V V V .
. . V I H V V .
V V . l V V V V
V V V V V V
V V V V . V
V . V V V V V V . V
V V V V V V V V V
V V V . V V . V .
IlIllIl-II‘III114lll.Il.llll|.'llnl.|lllt'l'ntl llnt'hlllliullVI‘thtlnlhI‘U'Olol Dun. 0 ll . I I AID-llt-III'U'IQVIL l
V V . V V 'l n~ I VI V V V V V '
V V . V V V . . V V |
n n V n ' u o n V V V
. V V V V V V V V .
V V V . V V V V V V V
V V V V V V V V V V
. r . V V V V V V
V V V . V . V V
. V V V V . . . . V
. . V V V V V V V
IIIVVVIVtIVIVI VVIVVIVVIVIVIVIIVI§VVVV \IVVIVIII .IIVIVVI..IVVVIIVVI-IVIVVVIVVVIVVI VIVVIVI. VVV r (IVIVIVVVVIIVVI
V . V . V . V V V
V V V V V V V . V V
V V . V V V . V V V V
o o n o u n . V c v v
V V V . V V V . . V V
V V V V V V . V V V
- n n a u n o - v -
V V . . V V V V V V V
V . V V V V V V V V . V V
V V V V V V V V V . V V V
V V . . V V V

VIVIIVV VVIVIVV IVVIIVVIIVJVIVVIIVVIV. IVVIIV..V. VI... VIVVIJ IVIVIIVIkIIIVIIVII‘IIVIIIIIIQIVIVVIVJIVVVIVVVw.VVVVVI I.II.I\.VIVVIV ... ... VNVVII. I... .I
V V V . V V . V I V V V V . V
V . . V V V V . V . V V
V V . V V V V . . V . V V
V V . V V . V V V V V .
. V . V V V V . . . V V V
0 v V . - Ill - V V n A . . v
V V V V V V V . . V V . V
V V V . V V V V V V V V
V V V V V V V . V V . V
V V V V V V . V V . . V

.IVVVIVIIVI vVIVVI.VI.VoIVVIVVVIVIVVVIVVIVVVIVIVI VVVVIIVVIIVVIVVVI AVIIVVIVVVVVVVVI VIVVVIIIVV V I .VVVVVVVVVIV..IIIVVVVVVIIVIVVVVV I VIIVIVVVI\.IVII ..IVVIII
V V V V . V V V V . V V . V V
V V V . V V V V V . V V _ . V
V V s u - V n o u I n . n u . n
V V V V V . V V V V V . _ . V
V V V V . V V V V V V V V V V .
V V . . . V . V V V V V . V V .
V V V V V V V V V V V V V V V .
V . V V . . V V V V V . . V V V
V V V V V V V V V V . V V V V
V V V _ V V V . V V . .

IVIVVVIVVIVI VVIVVIVVAIIVIVVVVI VVIIVoI..IVVIVV..IIVV.V.VVVI0.VIVVVV IVVIVVIVV'VVIIVVV V I m I. V. V. II.. 0 V ..V VIIIVVVVI. .3.. IV.
V . V V . V V V I V V V V V V
V . V . _ V V V V . V V V V V
V V V V V . V V . V V V . . V
V V . V . V V . V V . V V V V . V V
V V V V V V V V V V V V V . V V V .
. V V V V V V V V V V V V . V . V V
V V V V V I . V V V V . V V V V V V
V V V . V V V V V V V . V V V . V V
V V V V . V V V V V . . V V V V V V
V . V V V . . V . V V V V V V V V V

I llVolIUIOLVQQ .I'tt‘nulll'OOlIn‘IQIICDIstIL Isn‘tb‘IVDVVOV-aVQV‘IIon‘v-C-IIAD‘IV'nVOIIoVVI bull-.IOIt-‘VUIHOIIJIII Io'tlolgnlt‘ InvllbII‘lnnltlo-Olt‘lﬁllllﬂ. IO-O-O‘IV-OIVII' llnI-IIICAQCgI‘thnrlO'I A t-D‘VOIIcI’V‘.1I Calla-II
V V V . . V V . V . . V V V V . V V V
. V . V . V . V V V V V V V . V . V V V V
. V V V V V . V V . V . V . . V V V V V V
V V V . . V V V . V V V V V V V V . . V
V V . V V . . V V V . V . _ V V . V _ V .
V . V V _ V V V V V V . V V V V V V V V V
V V . V . . V V V V V V . . V V V V V V V
V V V . V V V V V . . V V V V V V V V . V
V . V V . V . V V . V V . . . V V V V V .
V V V V V . . V V . V V . V V V . V V . V

IIVIIVVIVIV ..I.....VVVIsVVIVVVVIIVV~IIIVVVIVVIVsVVVIVIVVIVVuVIIVVVVV.V.oIV...VVVVVsVVIVVV..VVvIVVVVVVV.VVVVVVVVVVVV}.VVVVVVVIV VVVVIVVIVQVVVVIIVVtoIIVVVVVIs.VI.......VVVVVVIVVoI.VVVVVsVVVVVIVVo .VVVIIV'IVVVIVVVIVVVVVVV V VI.\IV..VIVIVIVVI .VVVIIVV
V V V . . V V V . V V V . V V V V V V V
. V V . . V . _ V V V . . V . V V V . V
V V V V . . V . V . V V . V V V V V V . V
V V . V V V . V V V V V V V V V V . V V V V
V V V V . V V V V . . . V V V V V V V V V V
V V V V . V V V V . . V V V . . . . . V V V
V V V . V . V V V V . V V V V V V . V V V V
V . V . V V . . . V V . V . V V V . V V V .
. . V V V V V V V V . V V V V V V V _ . V
V . . V . V V V V . V V V V V V V V V _ V

«VIIVIVVVIV v.IVIVIVVIIVTIIVIIVIIVVAIV.I..VVVVIgVVVVVVVVIVV.VVIV.VVVVVVTVIVVVIVIVVAVI.. VVVVVVV‘VVVVVVVVVI.V.I...V.....V..I.V.V.VV IV.VVVVVV.~VVIV.VVVIJVVV V.VIVIAIIV.VVV.V*.VVVIIVVI~VVVIVIIJVVVVVVVVVVw.VVVVVV.IVVIVIIVVIV..IVAVVI....: .. VVVVHIIVVIVV.
. V V _ V . _ V V . _ . V V V V V V V . . V
V . . V V V V V V . V V V V V V . V V V _ V
V V V . . V V V V V V V V V . . V V V V V V
V V V _ V V V I V V V V . V . V . V . .

V . V V V V V V V V V V . V V V V V V V
V V V V V V V V V V V . V V V _ V . V
V V V V V V V V . . . . V V V V V V V V
V . V V V . V V V V . . V V V . V V V V
V V V V V V V . V V V V . V V V V V V V
V V V V . V V V V V V V V V . V V V . V

 
 

 

        

 

 

 

 

 

 

 

 

M mum

M mm

M m5

(swirl Kemqlv) KIIAIISESQH

206

Do o8

.::E 93. 2252 .8 8383th ..o couoﬁi a ma 32563“ Reﬁoo—m .w-< Eswﬁ

mohdﬁomEoH

Uo CWT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    

   

 

 

  

 

 

 

 

 

 

  

 

 

 
              

 

J 4 V 4 4 4 JV 4‘ + [T < V 4 V 4 4 [1 I < V < 1 I. J
. V V V V V V V V V . V V V . V V V V V V V V
V . V V V . V V V V . . V V . V . V V V V V
V V V V V V V . V V V V _ . . V V V . V . V . .
V V V V V V V . . . . . V V . V . V V .

. . . V V . . . V V . V . . V . . . V V V

V V V V V . . V V V V . . V V . . V V

V V V V V V V . V V V . . V V . V . V

V V V V V V V V V V . . V . . . _ . .
.-VV .V- VVVVVVVVVVV u VVVVVVVVVVV V VVVVVVVVVV V ......... o VVVVVVVVVVV v .......... V .............................. u ........ I. . .....-...VV.V..V.V.V V. . I .1 . .............. .V ..... V ....................... U. VVIV .

V . . . . V . . V V . . . . V V . . V . V V

_ V V V . . V V V . V . . V . . V . V V V

V V V . V . . V V V V V V . V V V V V V .

V V V V V . V . V . V . V . V _ . V V V V

V V V V . V V V V V V V V V V V V V V V . .

V V V V . . . V . V V V . . . V . . V V

V V V V . V V V V V V V . V V V V V . . .

I IIII§ ttttttttttt I IIIIIIIIIII I VVVVVVVVVVV o VVVVVVVVV t VVVVVVVVVVV ' VVVVVVVVVV u VVVVVVVVVVV I. IIIIIIIIIII III I I'ln I ‘I ........ K --------- § nnnnnnn D VVVVVVV § VVVVVVVV I ...... ho VVVVVV 'V ‘7 v u. u \. VII I I I V \I It. A

V V V V . V V . V V V V V V . V . V V V V V V

V V V . . V . V . . V V . . V V V . . . V . V

V . V V . V . V V V V V V . V V V . V . V V

V V V . V . V V . V . V . V V V . . V V V . V

V V . V. V V _ V . . . V V . V V . . V . V V V

V V V V V V . V V V V V V . _ V V _ . V V V .

V V V . V V V . V V V . V V V V _ V V . V .

. V V V V V V V . . V . V . V . . V . . . V

. V V V V V V V V V . . V . V V V V V

V V . . V . . . V V V V . . . . V . V k

V V V V V V V V . V V V V .

VVVVVVVVVVVV V1....V.VVV.a......V..V.V...VV...V.V.V..V.V VVV.............LV J.VV_VV.. ......V..~..V......V....VV..VV7....V...V.......V._....-.......TZVVVVL ...... .4. VV.V ...1..) ..VV. ...s..
V V V . V V . . V V V V V V V V V V
V . I . V V . V V V V . . . V . . V
. V L V V . . V . . V . V . .

V . V V V . . . . V V . V V
V V V . V V . . V . V . _ .
V V . V . V V V V . . .
. V V V V . V
......::...a.V-......:.o..:.......-.-.V..VV.V .«.V VVVVVVVo. VVVVVVVVV r .......... V ....... ...VV‘VVVIV.-VVJVVVVVV...... ........ I. VVVVVVVVV :.:...V... VVVVVVVVV . ....... a VVVVVVVV V ....... V ....... I ..... .V. V. .V I.
V . V V V . . V . .
V V V V V V V V
V V V V . V .
u n I V V V V
V V V I— . V .
n V V I V
V V . .
V V . V
1‘ V V V .
VVVVVVVVVVVVVVVVVVVVVVVVVV ...VVVVVVVVIV..-VV..V. ...-9....IV.V..V..I-...V...IV..V-..V.VV.VVVV-V
V . V .
V V V
A‘l . V V
V V .
V n -
V V V
. V .
V V V V
V V V V
......... ...rVV..-.V...|..-V...VVVVI.VV....-.V. ..VVVVVVVV. ....V....V..VVVV...V.\V v.
V . V V V
V . . V V
V . . V V
V V . V V
V V V V V
V . . . V V
V V . V . .
V V V V . V V V
V . V V . V . .
V V V V V V V V . . V

IIIIIIIIIII VIII.IIIVIIKIIIII«II-II‘IIQIIII-II-OIVIIIVI IIIOIIIVO~IVIIIOIIIIO on ' IIVVIVIIOAII-IIIIGI .OOIOIOIIQV
V V V V V . V V . V V
V . V V V V V . . . V
V . . V . V V V . . V . ,

V V V . V V V V V V . . V V
. V V V . V V . . . . V
V . V . . V V V . . V V .

. . . . V V V V . V . . . , V
V . . . V V V V V . . . . V

VVVIVVVVVVVVVV VVVVVVVVVVV )V VVVVV VIVVVVIVV.I VVVVVVV «...-VVV.VVV.¢ V ...... V‘VIVIIVI II.IIIII V. V. VVV.VV...-VILVVVI-V..VVJVVV.....VImVIIVVVVVV-V . VLV V s VAv VVVVV m ..... o. VVVVVV V ...... u ........... ”......V‘VVA
V . . V _ V V V V V . V . V . . V V
. V . V V V V V _ . V . V V . .
V V . . . V V V V V V . . . V .
V V V V V V V V K V V V V V V V V
V V V V . V V V V V V . . V . V . . . V V
. V V V . V V V V V l V V . V V . . V _ V
V V V V . V . V . V V V V . V V V . . V V .
. V V V V . V V V _ V V V . V . V . V . . . V
. V V V . V . V V V V V V V . V V V V . V V V
. V V V . V V V V V V V V V . V V V . . . V .

--....VVVVV.‘ VVVVVVVVVV 0.-.. VVVVVVV I VVVVVVVVVVV V ........... V ........... . .......... a .......... V.‘ V ..1». 4 a. I V ...... V ........ ‘ ....... V ........ I ...... .V ..... V. .... ..... I ...... \ ..... -V .I. .VV‘..I
V . V V V . V V V V V V . . . . . . . V . V . .
V V . . V . V V V V V V V . V V V V . . . V V V
V . V V . V V V V . V V V V . . V V V . V V _ .
V V V . V . V V V . V . V . V V V V V V V . V V
V V . V V V V V V V V V V V V . V V V V V V
c n g n V — u u I n _ V . g a V V g - V -
V V . . V V I . V V V V V V . . V . V V . V
V V V V V V V V . V V _ V V V _ V . V . V
o c V . V V I u o I V . V . o I V V I . . .
V V V V V V V . V V V V V . V V . V V . V

I ..... ..I.. v .......... A VVVVVVVVVVV I ........... a ........... o ........... Iv ......... I .......... 1 ......... v VVVVV o..l-u.-IIV utoVa-IuIIVI VVVVVVVVV V ....... I ........ u ....... I ........ I ...... .V ..... V ..... V ..... o ..... V|V . 1 V. .l. ...!n..
V V V V V . V V V V V V . V V V V . V V V
V V V V . V V V V . V V V . . V . V V
V V V . . V V V V V V V V V V V V V . V
V V V . V V V V V . V . V . . V V . V
. . . V . . V V . V V V . V . V V V
V c - V . . o . a I V n V V V V v V .
. V V V . V V V V V V . V V V V V V
V V V V . V V V V V V . V V V V V V
V . V V V . . . V . . . V . . V V V
. . . V V . . V V . V . . V . V V V

IIIIIIIIIII VIICVQIIIII§IIIIOIIIrv.‘lIIllcalvulhlIIIDIIIlVVIuIIIaII.....I..VVI-IIIKIIIIVCVIV-I'm-VII-I-I.I.I..!II.II\-ltIul-lI-l .III-.I- JI' IIaQIII’IICIIOII‘QIIIID‘I!..IIIIIIIIID-s I!‘. I II' I III. alI VII .LVVVv..IV|II-Q1.tlt‘lI1I1-IVIVIQ I I‘d! I'IIA
. . V . V V . . V V V V V V V . . V V V
V V V V V . V V V V V . V V V . V . . . V V V
V V V . . . V V V V V V V . . V . V V . . V
. . . V . V . V V V V V V V . V . . V V . V
V V V V . V V _ V V V V V . V . V . . V V
V _ V . V V V . V V V . V . . . V V . . V
V . . . V . _ V V V V V V V V V V V V V V
V V V . . V . V V V V V V V V V V V . V V

I1tluI-Illl‘r lllllll 01I§ IIIIIIIIIII § IIIIIIIIIII s IIIIIIIIIII o VVVVVVVVVV o ......... u IIIIIIIIII I VVVVVVVVV > IIIIIIIIII ‘0 IIIIIIIIIIIIIIIIIIII f IIIIIIII V‘ IIIIIIIII \ IIIIIIIII ﬂ IIIIIIIII o IIIIIII 5 IIIIIIII I ....... O IIIIIIII 0 IIIII Av IIIII o VVVVV c ..... \ ssssss I. VVVVVVVVV K VVVVVV 0.1;
. . . . . V . . V . _ . V . . V V V . . V . . V
V . V V V V V . V V V . . . V V V V V V V _ V V
V _ V V V V V . V V V V _ V V . V . V V V V
V V V . V V V . V V V V . . V V . V V .

V V V V V V . . V V V V V V V V V V V V .

V V V V V V . V V V V V V V . . V V V . V

V V V V V V . V V V V V . V V V . . V V . V

V V V . V . . V V . V . . V V V . . V . .

. V V V . . V . V . . V . . V . V V V V
V V V V V V V . V V V V V V . .
J V VVVVV VVV.J.VVVV.V.V.VVL ........... . .......... ... VVVVVVVVV A .......... Vw VVVVVV VVVV_ VVVVVVVVVVV V. ......... V VVVVV ..VVsV VVVVVVVV V ......... 1V. ...... a ....... . ....... J ........ .... VVVVV g ........ ‘.VVV.VA+V .-. ..... TV...) ..... (V ..... ...-)VVVV.V~..A
V V V . V V . V . . V . V V V . V . V . V V V
V V V V . V . V V . . V V V . V V . . V V V V
V V V V . V V V V V V V V V V . V V V . . V . V
. V V V V . V V V V V V . V . V V V V V . V V
V V V V V V . V V V . . V V . V V V V V V V V V
V . V . V V V V V V V V V . . V V V . V V . V
V V . . . . . V V V V . V V V V V . V V V V
V V V . . V . V . V V V V V V V V V V V V V .
V V V V . V V V V V . V V V V V . . . V V V V
V V V V . . V V V V V V . . V V V V V V V V V

 

 

Mmbm

(sum wav) mAnszsaH

207

.::E PE“ 22%? .8 238888. .0 885$ a 3 3.335% 78585 .o-< oSwE

mEBEQQEoH

 

............

...................................................................................................................

..............................................................................................................

 

 

 

 

 

 

 

 

........ Vm ﬁNMN H LP» m m m
..... M «hm Mar—l ...........

....................................................................................................................

.................................................................................................................

 

 

 

Uoo

 

........................................................................................................................

...........................................................................................................................

 

l1

 

+ [1
. . .

Uo OO~u 00 gm:

 

 

 

(sum wav) K1!A!18!sau

208

00 OS

 

M mum

.::E 23“ $252 5h 9.388th ho couogm a ma mzzuﬂmom Horace—m .oT< Bzwi
$5389th
00 O 00 8m- 00 CWM-

M .8“;

m mom n F

Mg um: .

M MbN M m: M mg

(suun Kenlqw) KJIAHSISOH

209

.EE 9?: 35m? 5.. 2380ath go .8855 a ma §>smaom 32:35 A ~ -< oSwE

85:22“th

00 0 Do 8w: 00 Oman

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.........

 

 

 

1 J 1 xi \\4 J1‘ 4 1 1\\ I1 1 I||1 1 4 4
. . . _ .
. . . V .
. . '. .
. _ .
. . .
. . .
.o-ouuo.au. v.......u..\...........u...............‘........................nv........‘.......v..-.......u..........l....v...... ...;v....o........I:.,......A.......‘......‘. ..........n:.......v..... .,....I..v.......v...........-.:...I. ...........
. y .
. . .
. . .
. . .
‘ .
........................................................................................................................... .-
. . .
. . .
. .
. .
...................... J‘un..t..-.A..vcu.o..-.o.-..4n..a..m“‘ ‘..-p..o-..1.I-...L.|........-s ...-...nu— ...-.IV.nI..c.nuy‘u..
. .
. .
. .
. .
,. , .
...............................................................................................................................
. .
. .
. . .
. . .
. . .
. .
, .
...................................................................... HUI—H ‘mWAUVMmH
.
. . .
. . . . . . . .
. . . . . . . . .
. . . . . . . . . .
. ‘ . . . . . . . .
. . . . . . ‘
.................................................................................. I! ...
...... . . . . II. c - n ...-n . . . o t o. u u .
. . . q— . . .
. . ‘ 'I' . . . .
. . . . . . .
. . . . . . .
. . . . . . . . . .
. . . ‘ . . . . . .
_ . . . . . . . .
. . . . . . . . .
. . . . . . . . . .
Illl‘unl Ito-III'.AIh-I.-.ohaolnu..Au-
. a . . . . . .
. . . . . . .
. .
. .
. .
.
4
. .
. .
. .
. .
. .
.
. .
v .
. .
. 17 . .
....................................................................................................................................................................................
. . . .
. . .
. . . ‘
. . .
. . .
. . .
. . . .
. . .
. . .
. . . .
...-n...-.. 1......a..‘a...o.......5.................u.....a. ......-.¢-.-...u....a.........vs‘..uan o: ..I.-o..u-..1 .....u...‘........lu....-...l........‘o..... ..4-1‘.....a..A.....u......4-..‘o....lu..a...v.n..:o.....u...uol|....nl .....
. . . .
. . . . . .
. . . . , .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . - . . . .
. . . . . . .
. . . . . . .
A.........a r.1.-.....«u...u.......I..4....n...............-.............u..¢.....on..a...a.n....n1.to-cuouu nu...\4 ...............................................................................................................................
. . . . . . .
, . . . _ . .
, . . . .
. IIII _ . .
. I'll . . .
, . , .
s .. ‘n...4».luu..o‘...o~.... .\r~ .I .. .A....or - ‘ y.............o...\5 ...........
. . . . . . .
, . . . . , .
. . . . . .
. . . . . .
. . . . . .
. . . ‘ .
. . . . .
. . . ‘ .
. . . .
. . . , .
nnnnnnnnnnnn 1.......-..J.......o...J;-.........m....a.ou .............‘..........AaA..........s.....................J..u...c.... 1........§.. ............-..Jo.......‘b................-J........‘................v.v....‘v.............v..un......\v..n:
. . . . .
. . . . .
. . . . ‘
. . .
. . .
. . .
v . .
. . .
. . .
. . .

(sum quw) KIIAHSISGH

210

.EE 98“ 925?. M8 055380? Mo c2855 a ma brémaom Reﬁne—m .N_-< Bzwﬁ
mBBEoQEP—L
Do 8.— 009 Do 8a.. 00 OWM:

M Sm“ m<

MSM”;

Mowm H

 

M MR M E M m: M Q

(sum Kemcuv) KJIAIISISQH

WP

NQMP

9°

10.

ll.

12.
13.
14.
15.

16.

17.
18.
19.
20.
21.
22.

BIBLIOGRAPHY

J. O. Almen and P. H. Black, ”Residual stresses and fatigue in metals”, McGraw-
Hill Book Company, Inc. (1963) 65.

T. Asaoka, H. Saito, and Y. Ishida, ICOMAT 92 (1992) 1003.

H.H. Anderson and H. L. Bay in "Sputtering by Particle Bombardment I", eds. R.
Behrisch, Springer-Verlag, Berlin, (1981) 145.

Z. S.Basinski, R. Pascual, and S. J. Basinski, Acta Metall., 31 (1983) 591.
Z. S.Basinski and S. J. Basinski, Acta Metall., 37 (1989) 3263.
M.Bayerlein and H. Mughrabi, Acta Metall., (1992).

G. Betz and G. K. Wehner in "Sputtering by Particle Bombardment II”, eds. R.
Behrisch, Springer-Veﬂag, Berlin (1983) 11.

l. E. Bolotov, V. Y. Kolosov, and A. V. Kozhyn, Phys. Stat. Sol. A, 72 (1982) 645.

J. D. Busch, A. D. Johnson, D. E. Hodgson, C. H. Lee, and D. A. Stevenson
presented in ICOMAT (1989).

J. D. Busch, A. D. Johnson, C. H. Lee, and D. A. Stevenson, J. Appl. Phys., 68
(1990) 6224.

R. W. Cahn, "Physical Metallurgy”, R. W. Cahn and P. Haasen eds., North-
Holland Physics Publishing, Amsterdam (1983) 1779.

L. Chang, Ph. D. Thesis, Michigan State Univ. (1993).
E. Y. Chen and E. A. Starke, Jr., Mater. Sci. Eng., 24 (1976) 209.
H. S. Chen, Appl. Phys. Lett., 29 (1976) 12.

J. P. Chevalier, V. Calvaiac, A. Quivy, M. Hannelin, and J. Bigot, Acta Metall, 31
(1983) 465.

T. W. Duerig and R. Zadno in ”Engineering Aspects of Shape Memory Alloys”,
eds. T. W. Dueri g, K. N. Melton, D. Stockel, and C. M. Wayman, Butterworth-
Heinemann, London (1990) 369.

L. Diaz, J. Victoria, and R. Gonzalez, Mater. sic. Eng., 37 (1979) 103.

B. Emmoth and H. Bersaker, Nucl. Instr. and Meth., B33 (1988) 435.

U. Essmann, U. Gosele, and H. Mughrabi, Phil. Mag. A, 44, No. 2 (1981) 405.
D. Favier, Y. Liu, and P. G. McCormick, Scripta Metall., 28 (1993) 669.

C. E Feltner, Phil. Mag., 12 (1965) 1229.

C. E Feltner and C. Laird, Acta Metall., 15 (1967) 1621.

J. C. Figueroa and C. Laird, Mater. Sci. Eng, 60 (1983) 45.

211

25.
26.

27.

29.

30.
31.
32.
33.
34.

35.
36.

37.
38.
39.

41.

42.
43.

45.

47.

49.

51.

52.
53.

212
T. Fukuda, T. Saburi, K. Doi, and S. Nenno, Mater. Trans. JIM, 33 (1992) 271.
J. George, "Preparation of Thin Films”, Marcel Dekker, Inc., New York, (1992) 41.

K. Giesser, J. D. Busch, A. D. Johnson, and A. B. Ellis, Appl. Phys. Lett., 61 (1992)
1633.

I. Greenﬁled and A. Purohit, Mater, Sic. Eng., 46 (1980) 89.

D. S. Grummon, J. W. Jones, and G. S. Was, Metall. Trans. A, 19A (1988) 2775.

D. S. Grummon, D. J. Morrison, J. W. Jones, and G. S. Was, Mater. Sci. Eng., A115
(1989) 331.

I. G. Greenfield and A. Purohit, Mater. Sci. Eng., 46 (1980) 89.

L. Hou, unpublished data (1995).

K. Hohmuth, E, Richter, and B. Rauschenbach, Mater. Sci. Eng., 69 (1985) 191.

A. Hunsche and P. Neumann, Acta Metall., 34 (1986) 207.

C. M. Hwang, M. Meichle, M. B. Salamon, and C. M. Wayman, Plil. Mag.A, 47
(1983) 9.

C. M. Hwang, M. B. Salamon and C. M. Wayman, Phil. Mag. A, 47 (1983) 177.

C. M. Hwang, M. Meichle, M. B. Salamon, and C. M. Wayman, Plil. Mag. A, 47
(1983) 31.

A. Ishida, A. Takei and S. Miyazaki, Thin Solid Films, 228 (1993) 210.

K. Ikuta, H. Fujita, M. Ikeda, and S. Yamashita, IEEE CH2832-4 (1990) 38.

C. M. Jackson, H. J. Wagner, and R. J. Wasilewski, NASA-SP 5110 (1972) 49.
A. D. Johnson, J. Micromech. Microeng., 1 (1991) 34.

R. Kelly in ”Ion Bombardment Modiﬁcation of Surfaces”, eds. O. Auciello and R.
Kelly, Elsevier Science, Amsterdam, (1984) 101.

W. H. Kim and C. Laird, Acta Metall., 26 (1978) 777.
J. J. Kim, P. Moine, and D. A. Stevenson, Scripta Metall., 2A) (1986) 243.

H. Kimura and T. Masumoto in ”Amorphous Metallic Alloys", eds. F. E Luborsky,
Butterworth & Co. Ltd., Lodon, (1983) 187.

R. Kolodiziej and S. Soltys, J. Mater. Sci. Lett., 11 (1992) 349.

U. Koster, Z. Metallkd, 75 (1984) 691.

D. Kuhlmann-Wilsdorf and C. Laird, Mater. Sci. Eng., 46 (1980) 209.

V. T. Kuokkala and P. Kettunen, Acta Metall., 33 (1985) 2041.

J. Lee, Master Thesis, Michigan State University, (1994).

J. M. Legresy , B. Prandi, and G. M. Raynaud, J. De Phys. IV, 1 (1991) C4241.
H. C. Lin and S. K. Wu, Scripta Metall., 26 (1992) 56.

H. C. Lin, S. K. Wu, T. S. Chou, and H. P. Kao, Acta Metall., 39 (1991) 2006.
H. C. Ling and R. Kaplow, Metall. Trans., 11A (1980) 77.

W. Liu, M. Bayerlein, H. Mughrabi, A. Day, and N. Quested, Acta Metall., 40
(1992) 1763.

55.

57.

59.

61.

62.

63.

65.

67.

69.

70.

71.

72.

73.

74.

75.
76.

77.

78.

79.

213

B. -T. Ma and C. Laird, Acta Metall., 37 (1989) 325.
A. F. Marshall, Y. S. Lee, and D. A. Stevenson, Mater. Sci. Eng., 63 (1984) 215.
O. Matsumoto, S. Miyazaki, and H. Tamura, Acta Metall., 35 (1987) 2137.

D. M. Mattox in ”Handbbok of Deposition Technologies for Films annd Coatings",
ed. R. F. Bunhsah, Noyes Publications, Park Ridge, New Jersey, (1994) 352.

K. N. Melton in ”Engineering Aspects of Shape Memory Alloys", eds. T. W.
Duerig, K. N. Melton, D. Stockel, and C. M. Wayman, Butterworth-Heinemann,
London (1990) 21.

K. N. Melton and O. Mercier, Metall. Trans., 9A (1978) 1487.

J. Mendez, P. Violan and M. F. Denanot, Nucl. Instr. and Meth., B19/20 (1987)
232.

J. Mendez, P. Villechaise and P. Violan, "Strength of Metals and Alloys” (ICMSA
8) eds. P. O. Kettunen, T. K. Lepisto and M. E. Lehtonen, Pergamon Press, (1988)
737.

C. A. Meyers and D. S. Grummon, Mater. Sci. Eng., A130 (1990) 127.
S. Miyazaki, K Otsuka, and Y. Suzuki, Scripta Metall., 15 (1981) 287.

S. Miyazaki, Y. Omni, K. Ostuka, and Y. Suzuki, J. of De Phys., No.4 (ICOMAT
1982) (1982) C4-255.

S. Miyazaki and K Otsuka, Phil. Mag. A, 50 (1984) 393.
S. Miyazaki and K Otsuka, Metall. Trans., 17A (1986) 53.
S. Miyazaki and C. M. Wayman,/1cm Metall., 36 (1988) 181.

S. Miyazaki , Y Kohiyama, and K Otsuka, J. Jpn. Inst. Met. Int'l. Sysm. 6th. 1991,
Sendai, Japan (1991) 691.

S. Miyazaki, Y. Sugaya, and K Otsuka, MRS Int'l. Mtg. on Avd. Mats, 9 (1989)
263.

S. Miyizaki, K. Otsuka, and C. M. Wayman, Acta Metall., 37 (1989) 1873.

W. J. Moberly, J. D. Busch, A. D. Johnson, and H. Berkson in ”Phase
Transformation Kinetics in Thin Films", eds. M. Chen, M. Thompson, R.
Schwarz,and M. Libera, MRS lnt'l. Mtg. on Avd. Mats, 230, (1991) 85.

D. J. Morrison, J. W. Jones, D. E Alexander, C. Kovach, and G. S. Was, Mater.
Sci. Eng., A115 (1989) 315.

Y. Motohashi, K. Ohsawa, T. Hoshiya, K. Okamoto, and M. Ohmori, J. Jpn. Inst.
Met., 55 (1991)132 [Japanese].

H. Mughrabi, Mater. Sci. Eng., 33 (1978) 207.

H. Mughrabi, ”Strength of Metals and Alloys", P. Hassen, V. Gerold and G.
Kostorz eds., Pergamon Press, (1979) 1615.

H. Mughrabi, "Fatigue Mechanisms: Advances in Quantitative Measurement of
Physical Demage" AST M No.81] (1982) 5.

P. Neumann, ”Physical Metallurgy", R. W. Cahn and P. Haasen eds., North-
Holland Physics Publishing, Amsterdam (1983) 1553.

P. Neumann, Mater. Sci, Eng., 81 (1986) 465.

81.

$333833?

83.33

89.

88.339.88.538

.898

100.

101.
102.

103.

104.

105.
106.
107.

214

P. Neumann and Tonnnessen, Fatigue '87, Proc. 3rd Int. Conf. on Fatiguee and
Fatigue Thresholds, Charlottesville, VA, vol. I, eds. R. O. Ritchie and E. A. Starke,

(1987) 3.

M. V. Nevitt, Trans. AIME, 218 (1960) 327.

M. Nishida and T. Honma, Scripta Metall., 18 (1984) 1299.

M. Nishida, C. M. Wayman, and T. Honma, Metall. Trans., 17A (1986) 1505.
M. Nishida, C. M. Wayman, and A. Chiba, Metallography, 21 (1988) 275.

Z. Nishiyama, ”Martensitic Transformation", Academic Press, New York (1978)
276.

I. C. Noyan and J. B. Cohen, "Residual Stress Measurement by Diffraction and
Interpretation”, Springer-Verlag New York Inc., (1986) 232.

H. Oechsner, Appl. Phys.,8 (1975) 185.

M. Ohn'ng, ”The Materials Science of Thin Films”, Academic Press, Inc., San
Diego (1992) 197.

T. Okutani, M. Shikata, S. Ichimura, and R. Shimizu, J. Appl. Phys., 51 (1980)
2884.

R. R. Olson, M.E. King, and G. K. Wehner, J. Appl. Phys., 50 (1979) 3677

K. Otsuka, T. Sawamura, and K. Shimizu, Phys. Stat. $011. A, 5 (1979) 457.

W. R. Patterson and I. G. Greenfield, Acta Metall., 19 (1971) 123.

J. Polak and M. Klesnil, Mater. Sci. Eng., 63 (1984) 189.

J. Polak and K. Obrtlik, M. Hajek, and A. Vasek, Mater. Sci. Eng., A151 (1992) 19.
A. K. Rai and R. S. Bhattacharya, Mat. Sci. Eng., 85 (1987) 139.

K. V.Rasmussen and O. B. Pedersen, Acta Metall., 28 (1980) 1467.

K. V.Rao, R. Malmall, S. M. Bhagat, G. Backstrom, and H. S. Chen, IEEE Trans.
on Mag., MAG-l6 (1980) 896.

T Saburi, M. Yoshida, and S. Nenno, Scripta Metall., 18(1984) 363.

Y. Shugo, S. Hanada, and T. Honma, Bull. Res. Inst. Min. Met., Tohoku Univ., 41
(1985) 23.

P. Sigmund in ”Sputtering by Particle Bombardment I", eds. R. Behrisch, Springer-
Verlag, Berlin, (1981) 9.

A. W. Sleeswyk, H. J. G. Kok, and G. Boom, Scripta Metall., 14(1980) 919.

F. A. Smidt, Inter. Mater. Rev., 35 (1990) 61.

S. Suresh, "Fatgue of Materials”, Cambridge University Press, New York (1991)
40.

R. A. Swalin, "Thermodynamics of Solid”, John Wiley & Sons, Inc, New York
(1972) 33.

T. Tadaki and Wayman, Scripta Metall., 14 (1980) 914.
D. Theirich and J. Engemann, Nucl. Instr. and Meth., 859/60 (1991) 336.

P. E Thoma, M. Y Kao, S. Fan'abi, and D. N. AbuJudom, ICOMAT 1992 (1992)
917.

108.
109.
110.
111.
112.

113.
114.
115.
116.
117.

118.

119.
120.
121.
122.
123.
124.

215

C. V. Thompson, Ann. Rev. Mater. Sci., 20 (1990) 245.

J. A. Thornton, J. Vac. Sci. Tech., 11 (1974) 666.

J. A. Thornton, Ann. Rev. Mater. Sci., 7 (1978) 239.

J. A. Thornton and D. W. Hoffman, Thin Solid Films, 171 (1989) 5.

H. Tobushi, H. Iwanaga, K. Tanaka T. Hori, and T. Sawada, Continuum Mech.
7’hermodyn., 3 (1991) 79.

T. Todoroki and H. Tamura, Trans. Japan Ins. Metals, 28 (1987) 83.
N. Wade, Y. Adachi, and Y. Hosoi, Scripta Metall., 24 (1990) 1051.
B. J. Walles, Master Thesis, Michigan State University, (1993).

R. Wang and H. Mughrabi, Mater. Sci. Eng., 65 (1984) 235.

R. J. Wasileewski, S. R. Butler, J. E Hanlon, and D. Worden, Metall. Trans., 2
(1971) 229.

C. M. Wayman in "Engineering Aspects of Shape Memory Alloys", eds. T. W.
Dueri g, K. N. Melton, D. Stockel and C. M. Wayman, Butterworth-Heinemann,
London (1990) 3.

A. T. Winter, Phil. Mag., 30 (1974) 719.

A. T. Winter, 0. B. Pedersen, and K. V. Ramussen, Acta Metall., 29 (1981) 735.
C. Y. Xie, L. C. Zhao, and T. C. Lei, Scripta Metall., 24 (1990) 1753.

K. Zoltzer and R. Bormann, J. Less-Comm. Metals, 140 (1988) 335.

Metals Handbook, vol.2 (1979) 275.

Binary Alloy Phase Diagrams, vol.2, ed. T. B. Massalski, American Metal Society,
Metal Park (1987) 1768.

 

"‘llllllllHill'lllll“