Development, Characterization and Testing of Nickel ...

Development, Characterization and Testing of

Nickel Titanium Based High Temperature Shape

Memory Alloys

By

Saif ur Rahman

Supervised By

Dr. Mushtaq Khan

School of Mechanical and Manufacturing Engineering (SMME)

National University of Sciences and Technology (NUST)

(2016)

ii

Development, Characterization and Testing of

Nickel Titanium Based High Temperature Shape

Memory Alloys

Saif ur Rahman

2011-NUST-DirPhD-ME-45

This work is submitted as a PhD thesis in partial fulfillment of the

requirement for the degree of

PhD in Mechanical Engineering

Supervisor: Dr. Mushtaq Khan


National University of Sciences and Technology (NUST)

Sector H-12 Islamabad, Pakistan

(2016)

iii

Certificate

This is to certify that the work in this thesis has been carried out by Mr. Saif ur Rahman and

completed under my supervision in the School of Mechanical and Manufacturing

Engineering (SMME), National University of Sciences and Technology (NUST), H-12

Islamabad, Pakistan.

Supervisor: ______________

Assistant Prof. Dr. Mushtaq Khan

School of Mechanical and Manufacturing Engineering

Submitted through

Principal/Dean


National University of Sciences and Technology (NUST), Islamabad

iv

Dedication

To my respected and loving mother

Farosha

v

Acknowledgments

In the Name of Allah, the Most Beneficent, the Most Merciful. All the praises and

thanks be to Allah, the lord of the entire Universe.

I would like to extend my gratitude to my supervisor, Dr. Mushtaq Khan for his

guidance and encouragement throughout my research. I would also like to thank him

for pushing me to complete my thesis in time. Thank you Mushtaq Khan, It is great

to have my supervisor.

Dr. Aamir Nusair Khan, my co-supervisor is acknowledged for his contribution,

help and support in alloys development and characterization. Thanks Aamir Nusair

Khan for your guidance, unlimited help and support in the completion of this thesis.

Thanks to my Guidance and Evaluation Committee (GEC) members Dr. Liaqat Ali

and Dr. Syed Hussain Imran Jaffery for their guidance and fruitful discussion. I

would like to extend my gratitude especially to external GEC member Dr. Riaz

Mohammad for selection of the current research area. His continuous guidance and

encouragement enabled me to complete my thesis successfully.

I would like to extend my gratitude to my DG Syed Tahir Hassan Hashmi and

Director Syed Muzaffar Ali for their support and encouragement throughout my

study duration.

Thanks to technical staff Tahir Mehmood Khan, Liaqat Ali, Muhammad Israr

and Muhammad Waseem for their efforts in alloys development and

characterization.

I would also like to thank my PhD colleague Malik Mansoor Muhammad for his

valuable input and assistance.

Of course I cannot forget my parents, my loving parents who have brought me up to

this stage. Without their prayers, love and endless support, I would not have

accomplished this task. Thank you always for your love.

vi

At the end, thanks to my dearest wife Rashida and loving children Shumila Saif,

Somayya Saif, Moneeba Saif, Waqar Ahmad Saif and Fatima Saif. Words are not

enough to express my gratitude for your love and patience.

Saif ur Rahman

July 2016

vii

Abstract

TiNi-based shape memory alloys are well known for their excellent shape memory

and superelastic properties. TiNiPd alloys are considered as the better high

temperature shape memory alloys due to high transformation temperatures, small

hysteresis, reasonable strain recovery and comparable workability. However, by

further increasing the transformation temperatures i.e. by increasing the Pd content,

thermal hysteresis also increases. This has an adverse effect on the actuation

behavior of the alloy. At high temperature the critical stress for slip deformation of

TiNiPd alloys also decreases, which increases the permanent deformation and

reduces the strain recovery in the alloy. In order to prevent increase in thermal

hysteresis, reduce permanent deformation and increase strain recovery in TiNiPd

alloy, Ni has been replaced by 5 at%, 10 at% and 15 at% Cu. Four alloys;

Ti50Ni25Pd25, Ti50Ni20Pd25Cu5, Ti50Ni15Pd25Cu10 and Ti50Ni10Pd25Cu15 (all in atomic

%) have been developed and characterized for their microstructure, phase

transformation temperatures, mechanical and shape memory properties in solution

treated condition. By increasing the Cu content, the transformation temperature of

the TiNiPdCu alloys significantly increased, whereas thermal hysteresis decreased.

Similarly, the microhardness, yield and fracture strength also increased. Shape

memory properties like strain recovery and work output also improved. Therefore,

TiNiPdCu alloys showed improved transformation temperatures, strain recovery and

critical stress for slip deformation through solid solution strengthening mechanism.

The TiNiPdCu alloys were also aged at different aging temperatures i.e. 400°C,

500°C, 600°C and 700°C for 3 hours to investigate their transformation

temperatures, mechanical and shape memory properties and compared with the

solution treated samples. By aging the Ti50Ni25Pd25 and Ti50Ni20Pd25Cu5 alloys, the

transformation temperatures, mechanical and shape memory properties slightly

increased. After aging the Ti50Ni15Pd25Cu10 and Ti50Ni10Pd25Cu15 alloys, the

transformation temperatures and shape memory properties significantly decreased,

however the mechanical properties were improved.

viii

It can be concluded that aging of Ti50Ni25Pd25 and Ti50Ni20Pd25Cu5 alloys is

beneficial to increase their transformation temperatures and shape memory

properties. However it has an adverse effect in terms of transformation temperatures

and strain recovery by aging the Ti50Ni15Pd25Cu10 and Ti50Ni10Pd25Cu15 alloys.

ix

Publications

1. Saif ur Rehman, Mushtaq Khan, Aamer Nusair Khan, Syed Husain Imran

Jaffery, Liaqat Ali, and Aamir Mubashar (2015), Improvement in the Mechanical

Properties of High Temperature Shape Memory Alloy (Ti50Ni25Pd25) by Copper (Cu)

Addition, Advances in Materials Science and Engineering [Article in Press]

2. Saif ur Rehman, Mushtaq Khan, Syed Husain Imran Jaffery, Liaqat Ali,

(2015), Effect of Aging on Phase Transition Behavior of Ti50Ni15Pd25Cu10 High

Temperature Shape Memory Alloys, Advanced Materials Research, Vol. 1101, pp

177-180.

3. Saif ur Rehman, Mushtaq Khan, Aamer Nusair Khan, Liaqat Ali, Syed

Husain Imran Jaffery (2015), Two-step Martensitic Transformation in an Aged

Ti50Ni15Pd25Cu10 High Temperature Shape Memory Alloys, Acta Physica Polonica

A, Vol. 128, pp B-125 – 127.

4. Saif ur Rehman, Mushtaq Khan, A. Nusair Khan, Liaqat Ali, Sabah Zaman,

Muhammad Waseem, Liaqat Ali, Syed Husain Imran Jaffery (2014), Transformation

Behavior and Shape Memory Properties of High Temperature Shape Memory Alloy

Ti50Ni15Pd25Cu10 at Different Aging Temperatures, Materials Science and

Engineering - A, Vol. 619, pp. 171-179.

5. Saif ur Rehman, Mushtaq Khan, A. Nusair Khan, M. Imran Khan, Liaqat Ali,

Syed Husain Imran Jaffery (2014), Effect of Precipitation Hardening and

Thermomechanical Training on Microstructure and Shape Memory Properties of

Ti50Ni15Pd25Cu10 High Temperature Shape Memory Alloys, Journal of Alloys and

Compounds, Vol. 616, pp. 275-283.

x

Conference Presentations

1. Saif ur Rehman, Mushtaq Khan, Liaqat Ali Syed Husain Imran Jaffery, Aamir

Mubashar (2015), Effect of Cu Addition on Microstructure and Transformation

Temperatures of Ti25Ni25Pd25 High Temperature Shape Memory Alloys,

International Conference on Mechanical, Aeronautics and Production Engineering

(ICMAPE) London, 20th April, United Kingdom.

2. Saif ur Rehman, Mushtaq Khan, Syed Husain Imran Jaffery, Liaqat Ali, (2015),

Effect of aging on Phase Transition Behavior of Ti50Ni15Pd25Cu10 High Temperature

Shape Memory Alloys, International Conference on Nano and Materials Science, Jan

24-26, Zhuhai, China.

3. Saif ur Rehman, Mushtaq Khan, Aamer Nusair Khan, Liaqat Ali, Syed Husain

Imran Jaffery, (2014), Two-step Martensitic Transformation in an Aged

Ti50Ni15Pd25Cu10 High Temperature Shape Memory Alloys, International Conference

on Computational and Experimental Science and Engineering (ICCESEN-2014), Oct

25-29, Antalya, Turkey.

xi

Table of Contents

Dedication ------------------------------------------------------------------------------------ iv

Acknowledgments --------------------------------------------------------------------------- v

Abstract ------------------------------------------------------------------------------------- vii

Publications ---------------------------------------------------------------------------------- ix

Conference Presentations ------------------------------------------------------------------ x

Table of Contents --------------------------------------------------------------------------- xi

List of Figures -------------------------------------------------------------------------- xviii

List of Tables ---------------------------------------------------------------------------- xxix

List of Abbreviations ------------------------------------------------------------------- xxx

Chapter – 1 Introduction and Overview .................................................................. 1

1.1 Shape memory alloys ............................................................................................. 1

1.2 Phase transformation temperatures ........................................................................ 3

1.3 Effect on transformation temperatures due to ternary alloying with NiTi ............. 4

1.4 High temperature shape memory alloys ................................................................. 5

1.4.1 TiNiPd based high temperature shape memory alloys ................................... 5

1.5 Problem statement .................................................................................................. 7

1.6 Research aim .......................................................................................................... 7

1.7 Research objectives ................................................................................................ 8

1.8 Research methodology ........................................................................................... 8

1.8.1 Material selection ........................................................................................... 8

1.8.2 Material processing and heat treatment .......................................................... 8

xii

1.8.3 Sample preparation and characterization ....................................................... 9

1.9 Thesis outline ......................................................................................................... 9

Chapter – 2 Background ......................................................................................... 11

2.1 Introduction .......................................................................................................... 11

2.2 Discovery of shape memory alloys ...................................................................... 11

2.3 Mechanism of shape memory effect and superelasticity ..................................... 12

2.4 Crystallography of NiTi-based shape memory alloys .......................................... 14

2.5 High temperature shape memory alloys ............................................................... 15

2.5.1 Nickel (Ni) based high temperature shape memory alloys .......................... 17

2.5.1.1 Nickel Aluminum (NiAl) high temperature shape memory alloys ...... 18

2.5.1.2 Nickel Manganese (NiMn) high temperature shape memory alloys .... 19

2.5.2 Copper (Cu) based high temperature shape memory alloys ......................... 22

2.5.3 Nickel Titanium (NiTi) based high temperature shape memory alloys ....... 25

2.5.3.1 Nickel Titanium Hafnium (NiTiHf) and Nickel Titanium Zirconium

(NiTiZr) high temperature shape memory alloys ............................................. 25

2.5.3.2 Titanium Nickel Platinum (TiNiPt) high temperature shape memory

alloys ................................................................................................................. 29

2.5.3.3 Titanium Nickel Paladium (TiNiPd) high temperature shape memory

alloys ................................................................................................................. 31

Chapter – 3 Equipment Setup and Material Processing ...................................... 41

3.1 Introduction .......................................................................................................... 41

3.2 Development of shape memory alloys ................................................................. 41

3.2.1 Cleaning of constituent elements ................................................................. 43

3.2.2 Melting of constituent elements ................................................................... 43

3.2.3 Homogenization ........................................................................................... 44

3.2.4 Chemical analysis ......................................................................................... 45

xiii

3.3 Sample preparation............................................................................................... 45

3.3.1 Sample dimensions ....................................................................................... 45

3.3.2 Solution treatment and aging ........................................................................ 46

3.4 Materials characterization .................................................................................... 47

3.4.1 Optical microscopy ....................................................................................... 47

3.4.2 Scanning Electron Microscopy (SEM) ......................................................... 48

3.4.3 X-Ray Diffractometry (XRD) ...................................................................... 48

3.4.4 Differential Scanning Calorimetry (DSC) .................................................... 49

3.5 Mechanical Testing .............................................................................................. 50

3.5.1 Microhardness testing ................................................................................... 50

3.5.2 Isothermal tensile testing .............................................................................. 50

3.6 Measurement of shape memory properties .......................................................... 52

3.6.1 Equipment setup ........................................................................................... 52

3.6.2 Constant stress thermal cycling tests ............................................................ 53

3.7 Summary .............................................................................................................. 56

Chapter – 4 Effect of Copper Addition and Aging on Microstructure of TiNiPd

Alloys ......................................................................................................................... 57

4.1 Introduction .......................................................................................................... 57

4.2 Microstructure analysis of solution treated TiNiPdCu alloys with varying Cu

percentage .................................................................................................................. 57

4.2.1 Second phase precipitates ............................................................................. 57

4.2.2 Grain size ...................................................................................................... 60

4.3 Effect of aging temperature on microstructure of TiNiPdCu alloys with varying

Cu percentage ............................................................................................................. 62

4.4 Phase analysis of solution treated TiNiPdCu alloys with varying Copper

percentage .................................................................................................................. 68

4.5 Phase analysis of aged TiNiPdCu alloys with varying Copper percentage ......... 70

xiv

4.5.1 Phase analysis of aged 0Cu alloy ................................................................. 70

4.5.2 Phase analysis of aged 5Cu alloy ................................................................. 71

4.5.3 Phase analysis of aged 10Cu alloy ............................................................... 72

4.5.4 Phase analysis of aged 15Cu alloy ............................................................... 74

4.6 Summary .............................................................................................................. 76

Chapter – 5 Effect of Copper Addition and Aging on Transformation

Temperatures of TiNiPd Alloys .............................................................................. 78

5.1 Introduction .......................................................................................................... 78

5.2 Effect of Cu addition on phase transformation temperatures............................... 78

5.3 Effect of aging on phase transformation temperatures ........................................ 80

5.3.1 Effect of aging on phase transformation temperatures of 0Cu alloy ............ 81

5.3.2 Effect of aging on phase transformation temperatures of 5Cu alloy ............ 82

5.3.3 Effect of aging on phase transformation temperatures of 10Cu alloy .......... 84

5.3.4 Effect of aging on phase transformation temperatures of 15Cu alloy .......... 87

5.4 Effect of thermal cycling on phase transformation temperatures ........................ 89

5.4.1 Effect of thermal cycling on phase transformation temperatures of 0Cu alloy

............................................................................................................................... 89

5.4.2 Effect of thermal cycling on phase transformation temperatures of 5Cu alloy

............................................................................................................................... 92

5.4.3 Effect of thermal cycling on phase transformation temperatures of 10Cu

alloy ....................................................................................................................... 94


alloy ....................................................................................................................... 96

5.5 Summary .............................................................................................................. 98

Chapter – 6 Effect of Copper Addition and Aging on Mechanical Properties of

TiNiPd Alloys .......................................................................................................... 100

xv

6.1 Introduction ........................................................................................................ 100

6.2 Effect of Cu addition on hardness ...................................................................... 100

6.3 Effect of aging on hardness ................................................................................ 101

6.3.1 Effect of aging on hardness of 0Cu alloy ................................................... 101

6.3.2 Effect of aging on hardness of 5Cu alloy ................................................... 102

6.3.3 Effect of aging on hardness of 10Cu alloy ................................................. 103

6.3.4 Effect of aging on hardness of 15Cu alloy ................................................. 105

6.4 Effect of Cu addition on mechanical strength .................................................... 107

6.4.1 Effect of Cu addition on mechanical strength in martensite phase ............ 107

6.4.2 Effect of Cu addition on mechanical strength in austenite phase ............... 109

6.5 Comparison between the mechanical properties of martensite and austenite

phases in solution treated condition ......................................................................... 111

6.6 Effect of aging on mechanical strength .............................................................. 114

6.6.1 Effect of aging on mechanical strength in martensite phase ...................... 114

6.6.2 Effect of aging on mechanical strength in austenite phase ........................ 116

6.7 Comparison between the mechanical properties of martensite and austenite

phases in 600°C-aged condition ............................................................................... 118

6.8 Summary ............................................................................................................ 121

Chapter – 7 Effect of Copper Addition on Shape Memory Properties of TiNiPd

Alloys ....................................................................................................................... 123

7.1 Introduction ........................................................................................................ 123

7.2 Shape memory properties of TiNiPd alloys with varying Cu percentage .......... 123

7.2.1 Shape memory properties of 0Cu alloy ...................................................... 123

7.2.2 Shape memory properties of 5Cu alloy ...................................................... 127

7.2.3 Shape memory properties of 10Cu alloy .................................................... 130

7.2.4 Shape memory properties of 15Cu alloy .................................................... 133

7.3 Effect of Cu addition on transformation temperatures ....................................... 136

xvi

7.4 Effect of Cu addition on transformation strains ................................................. 137

7.5 Effect of Cu addition on recovery ratio and work output .................................. 139

7.6 Summary ............................................................................................................ 141

Chapter – 8 Effect of Aging on Shape Memory Properties of TiNiPd Alloys .. 142

8.1 Introduction ........................................................................................................ 142

8.2 Shape memory properties of 600°C-aged TiNiPd alloys with varying Cu

percentage ................................................................................................................ 142

8.2.1 Shape memory properties of 600°C-aged 0Cu alloy .................................. 142

8.2.2 Shape memory properties of 600°C-aged 5Cu alloy .................................. 145

8.2.3 Shape memory properties of 600°C-aged 10Cu alloy ................................ 148

8.2.4 Shape memory properties of 600°C-aged 15Cu alloy ................................ 151

8.3 Comparison of shape memory properties between solution treated and 600°C-

aged 0Cu alloys ........................................................................................................ 154

8.3.1 Comparison of transformation temperatures .............................................. 154

8.3.2 Comparison of transformation strains ........................................................ 155

8.3.3 Comparison of recovery ratio and work output .......................................... 156


aged 5Cu alloys ........................................................................................................ 157





aged 10Cu alloys ...................................................................................................... 160




xvii


aged 15Cu alloys ...................................................................................................... 163




8.7 Summary ............................................................................................................ 167

Chapter – 9 Summary of Results and Discussion ............................................... 168

9.1 Introduction ........................................................................................................ 168

9.2 Effect of Cu addition on microstructure............................................................. 168

9.3 Effect of Cu addition on transformation temperatures ....................................... 169

9.4 Effect of Cu addition on mechanical and shape memory properties ................. 169

9.5 Effect of aging on microstructure and transformation temperatures ................. 170

9.6 Effect of aging on mechanical and shape memory properties ........................... 171

Chapter – 10 Conclusions and Recommendations for Future Work ............... 173

10.1 Summary of experimentation ........................................................................... 173

10.2 Conclusions ...................................................................................................... 174

10.2.1 Effect of Cu addition ................................................................................ 174

10.2.2 Effect of aging .......................................................................................... 175

10.3 Recommendations for future work .................................................................. 176

References: .............................................................................................................. 177

xviii

List of Figures

Fig. 1.1 Mechanism of martensitic transformation ----------------------- 2

Fig. 1.2 Dependence of martensite start temperature, Ms on Ni-content

in binary NiTi alloys ---------------------------------------------------

4

Fig. 1.3 Dependence of Ms temperature on the addition of Pd-content by

replacing Ni, in an equi-atomic TiNiPd system --------------------

6

Fig. 2.1 Mechanism of shape memory effect can be observed by

following the path a-b-c-a, while superelasticity can be realized

by following the path c´-c- c´ -----------------------------------------

13

Fig. 2.2 Different transformation paths in TiNi based alloys --------------- 15

Fig. 2.3 Dependence of Ms temperature on aluminum content in NiAl

alloy ----------------------------------------------------------------------

18

Fig. 2.4 Dependence of martensite peak temperature, Mp on either Ni or

Mn contents in NiMnAl alloy ----------------------------------------

20

Fig. 2.5 Composition dependence of the martensite start temperature, Ms

on aluminum content in CuAlNi alloys -----------------------

24

Fig. 2.6 Composition dependence of the martensite peak temperature,

Mp as a function of hafnium content in NiTiHf alloys --------

26

Fig. 2.7 Composition dependence of the martensite peak temperature,

Mp as a function of zirconium content in NiTiZr alloys -------

28

Fig. 2.8 Composition dependence of the transformation temperature as a

function of platinum content in TiNiPt alloys ---------------

30

Fig. 2.9 Change in martensite start temperature, Ms with respect to

palladium content in equi-atomic TiNi50-xPdx alloys ----------

33

Fig. 2.10 Martensite start temperature, Ms as a function of Ti/(Ni,Pd)

ratio ---------------------------------------------------------------------

35

xix

Fig. 2.11 (a) Specific work output for a Ti50.5Ni19.5Pd30 alloy as a function

of applied stress loaded in both tension and compression and (b)

the corresponding transformation strain versus applied stress

38

Fig. 2.12 Work output for a series of TiNiPd and TiNiPt alloys as

function of the transformation temperature range (Mf to Af)

39

Fig. 3.1 Process flow chart presenting the sequence of operations for

materials processing and characterization ----------------------

42

Fig. 3.2 Schematic representation of homogenization and solution

treatment processes -------------------------------------------------

46

Fig. 3.3 Schematic representation the aging process at various aging

temperatures ------------------------------------------------------------

47

Fig. 3.4 Measurement scheme of transformation temperatures from DSC

heating and cooling cycles -------------------------------------------

50

Fig. 3.5 Special gripping arrangement for holding of 0.3 mm thick

samples ------------------------------------------------------------------

51

Fig. 3.6 Internal view of lever arm creep and stress rupture tensile

testing system ----------------------------------------------------------

53

Fig. 3.7 Measurement scheme of transformation temperatures,

recoverable and irrecoverable strains from typical strain-

temperature curve ------------------------------------------------------

55

Fig. 4.1 SEM images showing the second phase precipitates formed

along the grain boundaries in solution treated samples of (a)

0Cu, (b) 5Cu, (c) 10Cu and (d) 15 Cu alloys --------------------

58

Fig. 4.2 EDS spectrums shown for solution treated samples of 0Cu,

5Cu, 10Cu and 15Cu alloys------------------------------------------

59

xx

Fig. 4.3 Optical micrographs (at 200X) of (a) 0Cu (b) 5Cu (c) 10Cu and

(d) 15Cu alloys solution treated at 900°C for 1 hour ----------

60

Fig. 4.4 Optical micrographs (at 500X) of (a) 0Cu (b) 5Cu (c) 10Cu and

(d) 15Cu alloys solution treated at 900°C for 1 hour ----------

61

Fig. 4.5 Effect of increasing Cu-content on grain size -------------------- 61

Fig. 4.6 Back-scattered SEM images presenting the microstructure and

grain boundaries in 0Cu alloys after aging for 3 hours at

temperature of (a) 400°C, (b) 500°C, (c) 600°C and (d) 700°C

63




64




65




66

Fig. 4.10 XRD profiles at room temperature for the samples solution

treated of 0Cu, 5Cu, 10Cu and 15Cu alloys ---------------------

69


treated and aged at 400°C, 500°C, 600°C and 700°C of 0Cu

alloys --------------------------------------------------------------------

70



alloys ---------------------------------------------------------------------

72

xxi



alloys ---------------------------------------------------------------------

73



alloys ---------------------------------------------------------------------

75

Fig. 5.1 DSC heating and cooling cycles of solution treated 0Cu, 5Cu,

10Cu and 15Cu alloys -------------------------------------------------

79

Fig. 5.2 Effect of Cu addition on transformation temperatures ------- 80

Fig. 5.3 DSC heating and cooling cycles of the samples solution treated,

aged at 400°C, 500°C, 600°C and 700°C of 0Cu alloy -------

81

Fig. 5.4 Effect of aging temperatures on transformation temperatures as

a function of increasing aging temperatures of 0Cu alloy ----

82


aged at 400°C, 500°C, 600°C and 700°C of 5Cu alloy ------

83



84


aged at 400°C, 500°C, 600°C and 700°C of 10Cu alloy -----

85



86


aged at 400°C, 500°C, 600°C and 700°C of 15Cu alloy -------

87



88

xxii

Fig. 5.11 DSC curves representing the transformation temperatures

during five thermal cycles of 0Cu alloy ------------------------

90

Fig. 5.12 Effect of thermal cycles on transformation temperatures of 0Cu

alloy ----------------------------------------------------------------------

91


during five thermal cycles of 5Cu alloy ---------------------------

93

Fig. 5.14 Effect of thermal cycle on transformation temperatures of 5Cu

alloy ---------------------------------------------------------------------

93


during five thermal cycles of 10Cu alloy -------------------------

95


alloy ----------------------------------------------------------------------

95


during five thermal cycles of 15Cu alloy -------------------------

97


alloy ---------------------------------------------------------------------

98

Fig. 6.1 Microhardness of solution treated samples of 0Cu, 5Cu, 10Cu

and 15Cu alloys ------------------------------------------------------

101

Fig. 6.2 Microhardness of 0Cu alloy aged at different temperatures 102




Fig. 6.6 Tensile stress-strain curves of solution treated 0Cu, 5Cu, 10Cu

and 15Cu alloys, tested in martensite phase (Mf – 50°C) ---

108

xxiii

Fig. 6.7 Tensile stress-strain curves of solution treated 0Cu, 5Cu, 10Cu

and 15Cu alloys, tested in austenite phase (Af + 50°C) ------

110

Fig. 6.8 Effect of partial substitution of Ni by Cu in solution treated

0Cu alloy on martensite yield stress, tested at (Mf – 50°C) and

austenite yield stress, tested at (Af + 50°C) ----------------------

111

Fig. 6.9 Effect of partial substitution of Ni by Cu in solution treated 0Cu

alloy on martensite fracture stress, tested at (Mf – 50°C) and

austenite fracture stress, tested at (Af + 50°C) ------------------

112

Fig. 6.10 Effect of partial substitution of Ni by Cu in solution treated 0Cu

alloy on martensite fracture strain, tested at (Mf – 50°C) and

austenite fracture strain, tested at (Af + 50°C) ------------------

113

Fig. 6.11 Tensile stress-strain curves of 600°C-aged 0Cu, 5Cu, 10Cu and

15Cu alloys, tested in martensite phase (Mf – 50°C) ----------

115

Fig. 6.12 Tensile stress-strain curves of 600°C-aged 0Cu, 5Cu, 10Cu and

15Cu alloys, tested in austenite phase (Af + 50°C) ------------

117

Fig. 6.13 Effect of partial substitution of Ni by Cu in 600°C-aged 0Cu

alloy on martensite yield stress, tested at (Mf – 50°C) and


119


alloy on martensite fracture stress, tested at (Mf – 50°C) and

austenite fracture stress, tested at (Af + 50°C) -----------------

119


alloy on martensite fracture strain, tested at (Mf – 50°C) and


120

Fig. 7.1 Strain-temperature curves representing the shape memory

properties of solution treated 0Cu alloy at stress levels of 100 –

500 MPa ---------------------------------------------------------------

124

xxiv

Fig. 7.2 Change in transformation temperatures of solution treated 0Cu

alloy at stress levels of 100 – 500 MPa --------------------------

125

Fig. 7.3 Recovered and irrecoverable strains of solution treated 0Cu

alloy at stress levels of 100 – 500 MPa ---------------------------

125

Fig. 7.4 Recovery ratio and work output of solution treated 0Cu alloy at

stress levels of 100 – 500 MPa --------------------------------------

126


properties of solution treated 5Cu alloy at stress levels of 100 –

500 MPa ---------------------------------------------------------------

127



128



129

Fig. 7.8 Recovery ratio and work output of solution treated 5Cu alloy at

stress levels of 100 – 500 MPa -------------------------------------

129


properties of solution treated 10Cu alloy at stress levels of 100

– 500 MPa ---------------------------------------------------------------

130



131



131

Fig. 7.12 Recovery ratio and work output of solution treated 10Cu alloy

at stress levels of 100 – 500 MPa -----------------------------------

132


properties of solution treated 15Cu alloy at stress levels of 100

– 500 MPa --------------------------------------------------------------

133

xxv


alloy at stress levels of 100 – 500 MPa -------------------------

134


alloy at stress levels of 100 – 500 MPa -------------------------

135


at stress levels of 100 – 500 MPa ----------------------------------

135

Fig. 7.17 Comparison of martensite start temperatures of solution treated

0Cu, 5Cu, 10Cu and 15Cu alloys under stress level of 100 – 500

MPa ----------------------------------------------------------------

136

Fig. 7.18 Comparison of austenite finish temperatures of solution treated

0Cu, 5Cu, 10Cu and 15Cu alloys under stress level of 100 – 500

MPa -----------------------------------------------------------------

137

Fig. 7.19 Comparison of recovered strain of solution treated 0Cu, 5Cu,

10Cu and 15Cu alloys under stress level of 100 – 500 MPa

138

Fig. 7.20 Comparison of irrecoverable strain of solution treated 0Cu,

5Cu, 10Cu and 15Cu alloys under stress level of 100 – 500 MPa

138

Fig. 7.21 Comparison of recovery ratio of solution treated 0Cu, 5Cu,


140

Fig. 7.22 Comparison of work output of solution treated 0Cu, 5Cu, 10Cu

and 15Cu alloys under stress level of 100 – 500 MPa --------

140


properties of 600°C-aged 0Cu alloy at stress levels of 100 – 500

MPa -------------------------------------------------------------------

143

Fig. 8.2 Change in transformation temperatures of 600°C-aged 0Cu


144

Fig. 8.3 Recovered and irrecoverable strains of 600°C-aged 0Cu alloy at


144

xxvi

Fig. 8.4 Recovery ratio and work output of 600°C-aged 0Cu alloy at


145


properties of 600°C-aged 5Cu alloy at stress levels of 100 – 500

MPa --------------------------------------------------------------------

146



146

Fig. 8.7 Recovered and irrecoverable strains of 600°C-aged 5Cu alloy at


147



148


properties of 600°C-aged 10Cu alloy at stress levels of 100 –

500 MPa ----------------------------------------------------------------

149



150

Fig. 8.11 Recovered and irrecoverable strains of 600°C-aged 10Cu alloy

at stress levels of 100 – 500 MPa ----------------------------------

150



151


properties of 600°C-aged 15Cu alloy at stress levels of 100 –

500 MPa ----------------------------------------------------------------

152



152


at stress levels of 100 – 500 MPa -----------------------------------

153

xxvii



153

Fig. 8.17 Comparison of transformation temperatures of solution treated

and 600°C-aged 0Cu alloy under stress level of 100 – 500 MPa

155

Fig. 8.18 Comparison of transformation strains of solution treated and

600°C-aged 0Cu alloy under stress level of 100 – 500 MPa

156

Fig. 8.19 Comparison of recovery ratio and work output of solution

treated and 600°C-aged 0Cu alloy under stress level of 100 –

500 MPa ----------------------------------------------------------------

157



158



159



500 MPa ----------------------------------------------------------------

160


and 600°C-aged 10Cu alloy under stress level of 100 – 500

MPa ---------------------------------------------------------------------

161



162



500 MPa ----------------------------------------------------------------

163

xxviii


and 600°C-aged 15Cu alloy under stress level of 100 – 500

MPa ---------------------------------------------------------------------

164



165



500 MPa ----------------------------------------------------------------

166

xxix

List of Tables

Table 3.1 Chemical compositions of the four alloys given in weight

percent -----------------------------------------------------------------

44

Table 3.2 Chemical composition of homogenized alloys ------------------- 45

Table 3.3 Dimensions of samples for various characterizations ----------- 46

Table 4.1 Compositional analysis of the overall alloy and second phase

precipitate in solution treated condition for 0Cu, 5Cu, 10Cu

and 15Cu alloys ------------------------------------------------------

59

Table 4.2 Compositional analysis of black and white precipitates formed

in 10Cu and 15Cu alloys after aging for 3 hours at various

aging temperatures ---------------------------------------------------

67

Table 6.1 Yield stress, fracture stress and fracture strain calculated from

stress-strain curves of Fig. 6.6 for solution treated alloys

tested in martensite phase (Mf – 50°C) --------------------------

109


stress-strain curves of Fig. 6.7 for solution treated alloys

tested in austenite phase (Af + 50°C) -----------------------------

111


stress-strain curves of Fig. 6.11 for 600°C-aged alloys tested

in martensite phase (Mf – 50°C) -----------------------------------

116


stress-strain curves of Fig. 6.12 for 600°C-aged alloys tested

in austenite phase (Af + 50°C) -------------------------------------

118

xxx

List of Abbreviations

SMA Shape Memory Alloy

SME Shape Memory Effect

TWSME Two Way Shape Memory Effect

PE Pseudo elasticity

TT Transformation Temperature

ΔT Thermal Hysteresis

εrec Recovered Strain

εirr Irrecoverable Strain

εf Fracture Strain

σy Yield Stress

σf Fracture Stress

σDT Detwin Stress

WQ Water Quenched

Ms Martensite Start Temperature

Mp Martensite Peak Temperature

Mf Martensite Finish Temperature

As Austenite Start Temperature

Ap Austenite Peak Temperature

Af Austenite Finish Temperature

1

Chapter – 1

Introduction and Overview

1.1 Shape memory alloys

Shape Memory Alloys (SMAs) are the materials which are capable to undergo

martensitic transformation and exhibit two unique properties, i.e. Shape Memory

Effect (SME) and Super-Elasticity (SE). In SME, the material is deformed in the

low temperature martensite phase, it recovers its original shape by the mechanism of

reverse transformation upon heating to a specific temperature called the reverse

transformation temperature. In SE, the material recovers a significant amount of

strain above its reverse transformation temperature. SE is an isothermal phenomenon

which is associated with a large nonlinear recoverable strain upon the removal of

externally applied load. For example NiTi wires recover about 8% strain which is

about 40 times larger as compared to conventional metals like steel having only 0.2%

strain [1, 2]. Shape recovery takes place at a particular temperature and thus the

SMA can function both as a sensor and as an actuator. Therefore SMAs are often

called smart or intelligent materials. Due to both Shape Memory and Super-Elasticity

properties of the SMAs, they are being used in various applications such as

automobile industries, high value consumer electronics, pipe couplings, medical

implants and guide wires and antennae for cellular phones. Apart from that, SMAs

are also used extensively in Micro Electro-Mechanical Systems (MEMS).

The properties of SME and SE are the result of a reversible martensitic

transformation which is a diffusionless solid state phase transformation process that

can be activated by temperature change or magnetic field. In such type of

transformation, the atoms of the material move cooperatively or by shear-like

mechanism. The high temperature phase (parent phase) is usually cubic and the

lower temperature phase (martensite) has a lower symmetry. The schematic of

martensitic transformation is shown in Fig. 1.1 [3].

2

Fig. 1.1 Mechanism of martensitic transformation [3]

By cooling the SMAs below some specific temperature, the process of martensitic

transformation exhibits by shear-like mechanism, as shown in the Fig. 1.1. From the

same schematic it can be seen that the structure of martensites present in regions A

and B are the same, however they are oriented in different directions. This difference

of the direction is called variants of the martensite. Due to having a lower symmetry,

considerable number of martensite variants can be formed from the same parent

phase. Similarly by increasing the temperature above the critical value, the available

martensites become unstable and the reverse transformation takes place. In this way

the martensites transform back to the original orientation as was present in the parent

phase and recovers its original shape. Thus the formation of martensite due to

temperature change is called the temperature-induced martensite which provides the

origin for shape memory effect. The phenomenon of SE takes place when the SMA

is deformed above the reverse transformation temperature. At high temperature, the

parent phase is stable in the absence of external load. However by applying load

(induced stress), the parent phase becomes unstable and consequently the martensites

are formed which are oriented in the direction of applied load. Upon removal of

external loading, the martensites become unstable and transform back to the parent

phase in original orientation. Thus the formation of martensite due to induced stress

within the material is called the stress-induced martensite, which is responsible for

the superelastic behavior.

3

1.2 Phase transformation temperatures

From the application point of view, it is important for SMAs to know the

temperatures at which the forward and reverse transformation starts and ends. The

temperatures at which the alloy starts and completes transformation from austenite

phase to martensite phase during forward transformation (cooling from high

temperature to low temperature) are called martensite start (Ms) and martensite finish

(Mf) temperature respectively. Conversely, during reverse transformation (heating

from low temperature to high temperature), the temperatures at which the alloy

begins and ends the transformation from martensite phase to austenite phase are

called austenite start (As) and austenite finish (Af) temperature respectively.

In the NiTi-based shape memory alloys, the phase transformation temperatures are

very sensitive to the Ni content. By increasing the Ni content above 50% (or

decreasing Ti-content below 50%), the transformation temperatures decrease very

rapidly [4]. However, decreasing Ni content below 50% (i.e. Ti content above 50%),

the transformation temperatures remain almost constant as shown in Fig. 1.2. This

shows the importance of the relationship between the transformation temperatures

and the alloy composition in case of NiTi-based shape memory alloys. The Ni-rich

NiTi shape memory alloys normally show superelasticity at room temperature since

the transformation temperatures are below room temperature. These alloys are

mainly used in applications where superelastic effect is required to function i.e.

orthodontic wires, stents, springs and guide wires etc. The Ti-rich NiTi shape

memory alloys have relatively higher transformation temperatures but have less

sensitivity towards the composition. These alloys normally show shape memory

effect at room temperature. The Ti-rich NiTi shape memory alloys are mainly used

as actuator type of applications where the shape memory effect is required to

function. The binary NiTi alloys can only be used at temperatures below 100°C,

since the highest Ms that can be achieved is about 77°C [4].

4

Fig.1.2 Dependence of Ms temperature on Ni-content in binary NiTi alloys [4]

1.3 Effect on transformation temperatures due to ternary alloying

with NiTi

There have been numerous efforts to improve or modify the shape memory and other

properties of NiTi-based shape memory alloys through ternary alloying additions. It

is well known that the ternary alloying additions often affect the transformation

temperatures. Among all the tested elements very few have been reported to increase

the transformation temperatures i.e. Pt, Pd, Au, Zr and Hf [4-7]. All the other

elements i.e. Cr, Mn, Fe, Ag, V, Co, Al, Si etc, generally cause a decrease in

transformation temperatures [5]. Cu addition in NiTi makes the transformation

temperatures less sensitive towards the compositional changes. Addition of Cu

exceeding 5% can change the transformation route into B2 (austenite) – B19

(orthorhombic) – B19´(monoclinic) and further addition causes a slight increase in

the transformation temperatures [6].

5

1.4 High temperature shape memory alloys

The shape memory alloys have already become an important class of materials for

various biomedical applications because of their unique SME and SE. Recently, their

potential has been realized in the areas of aerospace, energy exploration and

automotive industries to be used as solid state actuators [7]. Shape memory alloys

with high transformation temperatures can simplify the design and improve the

efficiency of many mechanical components which are required to operate at

temperatures higher than 100°C in automotive, aerospace, manufacturing and energy

exploration industries. Currently, the practical usage of shape memory alloys is

limited to temperatures below 100°C [8]. This is actually the transformation

temperature limit of the two most commercially successful shape memory alloy

systems; the NiTi binary alloys and Cu based ternary alloys. The low transformation

temperatures of these systems put a limit on their usage and they cannot be used in

those applications where the minimum requirement, in terms of transformation

temperatures, exceeds 100°C. Generally, the shape memory alloys with

transformation temperatures exceeding 100°C are categorized as high temperature

shape memory alloys.

1.4.1 TiNiPd based high temperature shape memory alloys

TiNiPd based high temperature shape memory alloys have been investigated since

their discovery by Eckelmeyer in 1976 [9]. Their potential as a successful high

temperature shape memory alloy system lies in the attractive combination of

essential characteristics which they possess i.e. high transformation temperatures,

adequate workability, reasonable level of strain recovery, small hysteresis and

acceptable mechanical strength. Initially the main research focus was to improve the

transformation temperatures of TiNiPd based high temperature shape memory alloys

but in the recent years the focus has been shifted towards the improvement of their

work output, dimensional and microstructural stability and high temperature cyclic

characteristics. The transformation temperatures of this system can be altered by

replacing Ni with Pd. If the concentration of Ti is held constant at nearly 50%, the

relationship between the transformation temperatures and relative concentration of

Ni and Pd is parabolic [10], as shown in Fig. 1.3. It can be seen that replacing Ni

with Pd up to 10% causes a decrease in the transformation temperatures and then

6

they start to increase upon further replacement of Ni with Pd with an approximate

rate of 15K/at%.

Fig.1.3 Dependence of Ms temperature on addition of Pd-content by replacing Ni, in

an equi-atomic TiNiPd system [10]

The parabolic dependence of the transformation temperatures on composition

originates from the change in the structure of martensite. In Ti50Ni50-xPdx alloys the

B2 (cubic) phase transforms into B19 (orthorhombic) martensite when Pd

concentrations remain higher than 10%, and it transforms into B19ʹ (monoclinic)

martensite or R (rhombohedral) phase when Pd contents are lower than 10%.

Because of the complete miscibility of the TiNi and TiPd systems, it is possible to

develop a continuous range of transformation temperatures from room temperature to

over 500°C by adjusting the Pd contents in TiNiPd alloy. The TiNiPd alloys with Pd

contents higher than 10% are of much importance especially for high temperature

shape memory applications. In these alloys, the B2 (cubic) phase normally

transforms into B19 (orthorhombic) martensite.

7

1.5 Problem statement

For the shape memory alloys to be used as an actuator at high working temperature,

their transformation temperatures must be increased to a reasonable limit, but at the

same time it must have the property to operate with low thermal hysteresis.

Unfortunately, by increasing the transformation temperatures due to increase in Pd

content in TiNiPd, thermal hysteresis also increases due to weakening of

compatibility between the B19 martensite phase and B2 parent phase.

Similar to any other metallic material, the critical stress for slip deformation

decreases with increasing temperature in TiNiPd based high temperature shape

memory alloys. The slip deformation can easily occur simultaneously with the

reorientation of martensite variants by stressing the alloy at high temperatures. Also

the creep deformation becomes a serious problem especially when the martensitic

transformation temperature range overlaps the creep deformation temperature (~

370°C) for these alloys. Undesirable recovery and recrystallization of the

microstructure is another important concern especially in case of

thermomechanically treated TiNiPd alloys. These factors actually put a limit on the

maximum transformation temperature range up to which the transformation

temperatures of TiNiPd based alloys can be raised by changing the Pd contents. The

condition becomes further worse when the material has to work under a certain load

where transformation temperatures are increased according to the Clausius-

Clapeyron relationship.

1.6 Research aim

The aim of this research is to further increase the transformation temperatures of the

TiNiPd based high temperature shape memory alloys while maintaining the thermal

hysteresis at reasonable low level and to improve their mechanical and shape

memory properties.

8

1.7 Research objectives

A number of objectives established for this research work were:

1. To develop TiNiPdCu alloys with varying Cu percentage.

2. To characterize the microstructure, transformation temperatures, mechanical and

shape memory properties of baseline ternary TiNiPd alloy in solution treated

condition.

3. To investigate the effect of Cu addition by various composition in place of Ni in

baseline ternary TiNiPd alloys on microstructure and other properties by comparing

their results.

4. To increase the phase transformation temperatures and decrease the thermal

hysteresis of TiNiPd alloy by substitution of Cu in various composition.

5. To improve the mechanical, shape memory properties and dimensional stability of

TiNiPd alloy.

6. To study the effect of aging at various aging temperatures on microstructure and

other properties of baseline ternary TiNiPd alloy and resultant quaternary TiNiPdCu

alloys.

7. To compare the results obtained from solution treated and aged samples.

1.8 Research methodology

The methodology established for this research work can be divided into three phases

which are explained in detail below:

1.8.1 Material selection

High purity constituent elements are the basic and most essential requirement for

developing the shape memory alloys. Therefore constituent elements; Titanium,

Nickel, Palladium and Copper of highest purity (99.98 – 99.99 weight percent) were

selected for this purpose.

1.8.2 Material processing and heat treatment

Material processing includes the cleaning of constituent elements in Ultrasonic

Cleaner, precise weighing up to one-tenth of mg, melting in Vacuum Arc Melting

Furnace and casting of 20 g buttons of four alloys. Heat treatment cycles consist of

9

homogenization at 950°C for 2 hours, solution treatment at 900°C for 1 hour and

aging at different aging temperatures for 3 hours. Material processing and heat

treatment cycles are explained in detail in Chapter – 3.

1.8.3 Sample preparation and characterization

Samples were prepared by wire Electrical Discharge Machine (EDM) for various

characterizations. The samples of four alloys were characterized for their

microstructure by Optical Microscope (OP) and Scanning Electron Microscope

(SEM). Phase analysis was carried out by X-Ray Diffractometer (XRD). Differential

Scanning Calorimeter (DSC) was used to find out the phase transformation

temperatures. Mechanical properties were investigated by using the Mechanical

Testing System (MTS). Tensile Creep and Rupture Testing Machine (TCRTM) was

used to find out the various shape memory properties. Experimental setup for

complete characterization is explained in detail in Chapter – 3.

1.9 Thesis outline

The current study is focused on the improvement in transformation temperatures,

mechanical and shape memory properties of TiNiPd based high temperature shape

memory alloys through quaternary alloying addition of Cu and precipitation

hardening. The chapter wise distribution of the current study is given as follows:

Chapter – 2 gives an overview of the mechanisms and different aspects of shape

memory alloys, background and literature review.

Chapter – 3 describes materials and methods that include the material processing,

melting, heat treatment cycles, characterization techniques and experimental set up

for investigation of various properties of TiNiPdCu alloys.

Chapter – 4 presents the effect on microstructure of ternary baseline TiNiPd alloy by

addition of Cu and aging at different aging temperatures in TiNiPdCu alloys.

Chapter – 5 explains the change in phase transformation temperatures of TiNiPd

alloy due to quaternary alloying addition of Cu and aging at different aging

temperatures in TiNiPdCu alloys.

10

Chapter – 6 details the effect on mechanical properties of TiNiPd alloy due to

quaternary alloying addition of Cu and aging at different aging temperatures in

TiNiPdCu alloys.

Chapter – 7 focuses on the variation of shape memory properties of TiNiPd alloy by

addition of Cu in place of Ni.

Chapter – 8 discusses the effect of aging at different aging temperatures on shape

memory properties of TiNiPdCu alloys.

Chapter – 9 summarizes the results and discussion of all experimental work carried

out in Chapter – 4 to Chapter – 8.

Chapter – 10 enlists the major conclusions drawn from the current study and suggests

the future work.

It is expected that this research will be beneficial for the improvement of high

temperature shape memory properties of TiNiPd based alloys and will open some

new directions of research in the area of high temperature shape memory alloys.

11

Chapter – 2

Background

2.1 Introduction

Significant progress has been made both in the scientific understanding and

application of shape memory alloys (SMAs) since the discovery of these

multifunctional materials. Owing to the unique behaviors of shape memory effect

and superelasticity, SMAs have become a major materials class of choice in the

biomedical industry and are beginning to spread through other technological areas.

There is a recent revival of interest in SMAs, driven primarily by the aerospace and

automotive industries, for their potential to operate as solid state actuators.

Shape memory effect is a phenomenon whereby a deformed material could recover

its predeformed shape after being heated. When this procedure is performed against

some biasing force, the material is capable of doing work due to change in its shape.

Superelasticity is an isothermal phenomenon where the material is able to recover

high amount of strain triggered by mechanical stress. These two behaviors are the

result of reversible martensitic transformation; a diffusionless solid state phase

transformation mechanism that can be activated by temperature, stress and magnetic

field.

2.2 Discovery of shape memory alloys

The first shape memory alloy which was discovered in 1932 by a Swedish Physicist

Arne Olander was Au-Cd [11]. He observed that the plastically deformed Au-Cd

returned back to its original shape upon heating. The same shape memory effect was

then observed in 1938 by Greninger and Alden B. in a CuZn alloy [12]. In 1958,

Chang and Read demonstrated that this unique property can perform a mechanical

work [13]. They used an Au-Cd shape memory alloy to lift a weight at the Brussels

World’s Fair. The real interest in the shape memory alloys was developed in 1962

after the discovery of NiTi, exhibiting shape memory properties by William Buehler

et al. at the U.S. Naval Ordinance Research Laboratories [14]. The name for this

12

alloy was kept as ‘Nitinol’, derived from NiTi Naval Ordinance Laboratories. This

discovery was proved to be a revolution in the field of shape memory alloys due to

superior qualities as compared to previously discovered shape memory alloys. Since

the NiTi alloy has many complicated structures associated with it, the understanding

of the mechanism for SME was not possible till early 1980s. From the discovery of

SME in CuAlNi [15] alloy that made it possible to relate the SME with the

thermoelastic martensitic transformation.

2.3 Mechanism of shape memory effect and superelasticity

The mechanism of shape memory effect and superelasticity is explained by a

schematic illustration of reversible martensitic transformation, as shown in Fig. 2.1.

The parent phase is stable at high temperatures as shown at 2.1(a) as compared to the

martensite phase as shown at 2.1(b) which is a low temperature phase. When the

temperature of the parent phase is decreased then the martensite phase starts to form

at a temperature called martensite start temperature (Ms) and completely transforms

into martensite when the temperature exceeds the martensitic finish temperature

(Mf). In this case the martensite is called thermally induced martensite. To minimize

the strain energy caused by the difference between the crystal structures of parent

and martensite phases, an invariant habit plane is created between the two

transforming phases. The martensite can have different variants, each with a unique

crystallographic orientation. In Fig. 2.1 [16] only two variants, V1 and V2 are shown

for simplicity. These variants are twin related and the martensite is called twinned

martensite. The transformation strain generated during transformation from parent to

martensite is accommodated by a self-accommodation process in martensite. Self-

accommodation is realized through twinning created between the different variants

[15, 17]. The macroscopic volume in Fig. 2.1(b) remains similar as compared to that

of parent phase shown at 2.1(a) mainly due to the self-accommodation of

transformation strain in twinned martensite. When the twinned martensite is

subjected to an externally applied stress, the different variants of martensite reorient

themselves to accommodate the most favorable variant and try to transform into a

single variant which in this case is V1 as shown at 2.1(c). The martensite at this point

is called deformed or detwinned martensite. The reorientation of different variants

into a single variant actually creates the transformation strain.

13

Fig. 2.1 Mechanism of shape memory effect can be observed by following the path

a-b-c-a, while superelasticity can be realized by following the path c´-c- c´ [16]

Accordingly, shape change occurs and the volume at 2.1(c) becomes greater than that

at 2.1(a) or 2.1(b). In martensite state, the SMA is very easy to deform by exerting

force due to production of twin boundary. If the force is removed at this stage the

deformation in the martensite remains, representing the plastic deformation. Hence,

when the stress is removed, no change is occurred in the shape of deformed or

detwinned martensite because the martensite is stable at this temperature. If the

temperature of detwinned martensite is increased and it exceeds the austenite start

temperature (As) then the martensite starts the reverse transformation process and it

completes when the temperature exceeds the austenite finish temperature (Af) and

martensite completely transforms into the parent phase. As a result, the original

macroscopic shape is restored as shown at 2.1(a). This mechanism following the path

a-b-c-a is called the shape memory effect.

14

Superelasticity is another aspect of shape memory behavior. In Fig. 2.1 the

superelasticity can be realized by following the path c´-c- c´. If the temperature is

kept higher than Af, the martensite phase can be induced by the direct application of

load and is called stress induced martensite. The stress induced martensite contains

the most favorable variant as shown in 2.1(c). By applying the load the

transformation strain is created and the volume of the material is increased. The

stress induced martensite is thermodynamically unstable above Af. As a results when

the applied load is removed the materials reverts back to austenite phase and restores

the original shape. This phenomenon is called superelasticity.

2.4 Crystallography of NiTi-based shape memory alloys

Since the discovery of NiTi shape memory alloys in 1962, extensive investigations

on the important aspects of this alloy i.e. mechanism of shape memory effect,

crystallography, thermo-mechanical behavior, ternary alloying additions, fabrication

methods etc. have been carried out and a significant understanding of this system has

been achieved.

In binary NiTi based shape memory alloys the martensitic transformation occurs due

to the solid to solid phase transformation of B2-parent phase into a B19´ martensite

phase. The parent phase has a cubic-B2 ordered structure. The lattice parameter of

the parent phase is a˳ = 0.3015nm. The B19´ martensite phase on the other hand

possesses a monoclinic crystal structure, with the following lattice parameter: a =

0.2898nm, b = 0.4108 nm, c = 0.4646 nm and β = 97.78° [18, 19]. Other than a direct

transformation route from the cubic-B2 parent phase to monoclinic-B19´ martensite

phase, the parent phase can also adopt a two stage transformation path either by first

transforming into an R-phase and then into monoclinic-B19´ phase or into an

orthorhombic-B19 phase and then into monoclinic-B19´ phase as shown in Fig. 2.2

[6].

The appearance of R-phase transformation is observed mainly due to the effects of

composition (in ternary TiNiFe and TiNiAl alloys), aging (low temperature aging of

Ni rich NiTi alloys) or thermomechanical treatment of near equiatomic NiTi alloys

[20-27].

15

Fig. 2.2 Different transformation paths in NiTi based alloys [6]

R-phase possesses a trigonal crystal structure and exhibits a very narrow

transformation temperature hysteresis (about 1 – 3°C) [28]. The R-phase

transformation is martensitic and thermoelastic in nature and it gives rise to shape

memory effect and superelasticity [29, 30]. The appearance of R-phase

transformation actually suppresses the transformation to B19´ martensite and as a

result the Ms temperature decreases. By adding certain ternary alloying elements in

NiTi alloys i.e. Pd, Pt, Au and Cu, the crystal structure of the martensite phase can be

changed from monoclinic to orthorhombic. In case of orthorhombic martensite; the β

angle is reduced to 90° [6, 31, 32].

2.5 High temperature shape memory alloys

Current practical uses of SMAs are limited to temperatures below 100°C. This is the

transformation temperature limit of the two most commercially successful SMA

systems i.e. the near equiatomic NiTi binary and Cu based ternary alloys. During

thermomechanical processes required to produce stable shape memory or

superelastic behavior, the transformation temperatures are further reduced. Naturally,

such limitation obstructs the utility of SMAs in high temperature applications. On the

other hand, the unique properties of SMAs become more beneficial at high

temperatures, since it is preferable to adopt single piece adaptive and multifunctional

components over more complex multicomponent assemblies due to the higher

probability of wear or damage and the greater weight and volume required by the

latter. These issues have activated several studies on possible SMAs with

transformation temperatures above 100°C. This class of materials is simply referred

to as high temperature shape memory alloys (HTSMAs). Till now, despite intensive

16

research efforts in recent years, HTSMAs have yet to be utilized commercially in

appreciable amounts due to a number of unresolved issues.

In recent years many material scientists have paid close attention to the development

of SMAs which can operate at temperatures above 100°C because of the demand

from automotive, robotic and aerospace industries [33]. One of the simplest

examples of SME is unconstrained recovery in which the SMA wire recovers its

original shape after being deformed or twisted in any shape. The same concept has

been utilized to put forward SMA based antenna arrays. The antennae for a satellite

or a space rover can be rolled into a compact state for storage while the unit is being

transported, but when the unit is in action the antenna can be deployed by passing a

small amount of current through the SMA based antenna to prepare the antenna to

function. Most orbital satellite utilizing this application would also require a high

temperature shape memory alloy with transformation temperatures more than 100°C

to avoid unintended actuation in case of exposure to direct sunlight. Another

application area of SMAs is during constrained recovery in which the SMA

component is prevented from going back to its original shape upon heating, which

usually generates very high stresses [34]. This property can be used for fastening

multiple components, connecting pipes or tubes etc.

Due to material property degradation of currently available SMAs, they cannot be

used for above mentioned application because of their risk of failure due to material

strength [34].The high demand for HTSMAs is from the interest of utilizing them in

solid state actuators. SMA based actuators have been found to have higher energy

densities than pneumatic actuators and DC motors and at par with hydraulics while

having the advantage of weighing significantly less. Along with the advantage of

being lighter in weight, being frictionless and noiseless make SMA based solid state

actuators a formidable candidate for application in weight critical systems such as jet

turbine engines, spaceships and any other aerospace application. Hence, the

development of HTSMA will play a pivotal role in the realization of the use of SMA

in real life applications in aerospace such as clearance control in the compressor and

turbine section of jet engine, variable area and geometry inlets for subsonic jets, self-

damping components in fuel line clamps, down-well flow control valves, electrical

appliances and actuators close to engine parts in automobiles. Currently available

17

NiTi based SMAs can only operate at temperatures lower than 100°C due to their

low transformation temperatures or poor mechanical behavior at high temperatures

[34]. In the mentioned higher temperature range the martensitic transformation

occurs at a temperature range where the diffusion controlled processes like

decomposition, recrystallization, recovery, etc. could take place and might

deteriorate the shape memory behavior.

Some of the most common approaches suggested to meet the industry requirements

of HTSMAs include the following [35, 36].

Development of new metallic alloys with high transformation temperatures.

Addition of ternary or quaternary element(s).

Optimization of various thermo-mechanical treatments and process of SMAs.

To distinguish HTSMAs from SMAs, many definitions have been put forward. One

of the most widely accepted definition is given by Firstov et al. [33] where he

describes HTSMAs as alloys of which the reverse transformation temperature As

starts above 117°C in stress free conditions following any thermomechanical

treatment. Some of the most extensively researched HTSMAs are: Ni-based [37-42];

Co-based [43-46]; Zr-based [47-53]; Ti-based [54-57]; Pt-based TiPt, TiPtIr [58, 59];

Cu-based [60-64] TaRu, NbRu [65, 66] and NiTi-based TiNiPd, TiNiPt, TiNiAu,

TiNiZr, TiNiHf [67-71]. Some of the above mentioned HTSMAs are studied in detail

as under.

2.5.1 Nickel (Ni) based high temperature shape memory alloys

Ni based HTSMAs have a long history, being the first SMAs to be extensively

studied with transformation temperatures above 100°C [72, 73] . The two main

systems are NiAl and NiMn, and both have been thoroughly investigated. Relatively

low cost and a fairly thorough knowledge of related structural materials have

provided a firm basis for the study of Ni based HTSMAs.

18

2.5.1.1 Nickel Aluminum (NiAl) high temperature shape memory alloys

In 1971, Enami and Nenno first identified both shape memory and superelastic

behavior in Ni63.2Al36.8 (at%) [37]. Au and Wayman showed in 1972 that ΔT was

only 10°C and the transformation was thermoelastic. As shown in Fig. 2.3, the Ms of

NiAl increases with decreasing Al concentration by ~130°C/at%Al and reaches

100°C at Ni64Al36 [74] and 370°C in Ni66Al34 [72]. Ternary alloying in small

amounts up to 5 – 10 at% with V, Cr, Mn, Fe, Zr, Mo, Ta or W decreases

transformation temperatures moderately, alloying with Nb, Ti or Si decrease them

significantly [75] while Co, Cu or Ag additions have no effect on or slightly increase

transformation temperatures [73]. Maximum εrec in binary solution treated Ni63.1Al36.9

was reported to be 3 – 4% by Au and Wayman. When aluminum is removed to

increase transformation temperatures above room temperature, the shape memory

properties become very poor in binary NiAl alloys. No more than 0.5% εrec was

possible in solution treated Ni63.8Al36.2 prepared by powder metallurgy [40]. Even in

this small strain range, recovery rate is commonly less than 50%.

Fig. 2.3 Dependence of Ms temperature on aluminum content in NiAl alloy [73]

19

The lack of satisfactory shape memory behavior in NiAl HTSMAs has several roots;

poor tensile ductility, phase decomposition processes, and texture. While Jee et al.

[40] reported 9.5% strain to failure in solution treated Ni63.8Al36.2 in compression.

Kim et al. [72] was only able to achieve 0.5% strain to fracture in bending for

solution treated Ni63.5Al36.5, which failed by transgranular fracture in the elastic

range. Two reasons exist for this poor ductility; easy intergranular fracture due to

large grain sizes and inherently poor fracture toughness, and transgranular fracture,

due to the insufficient number of slip systems in the β´ martensite to accommodate

polycrystalline deformation at room temperature [76]. Ternary additions of Fe and

Co are highly effective in increasing ductility [77] mainly as a consequence of the

precipitation of ductile γ fcc phase. Although boron addition suppresses intergranular

failure, it does not improve ductility by itself because the alloy will fail instead by

transgranular fracture. A combination of boron and Fe or Co addition is most

effective in increasing ductility since both intergranular and transgranular failure

modes are suppressed. Grain refinement processes, such as rapid solidification, are

effective in further improving ductility in NiAl alloys containing Fe or Co.

Similar to Cu based HTSMAs, phase instabilities and decomposition cause

significant problems in NiAl SMAs. Two types of microstructural decomposition

processes are relevant. In alloys with high Al or high Fe/Co content, γ phase appears

after quenching or aging. Although the γ phase improves ductility as mentioned

previously, it does not undergo martensitic transformation and is therefore

detrimental to reversibility of the shape memory behavior [78]. The appearance of

the γ phase can be avoided by appropriate composition selection, and training can

overcome some of the deleterious effects of the γ phase in dual phase alloys.

2.5.1.2 Nickel Manganese (NiMn) high temperature shape memory alloys

The second well known NiMn alloy system is similar to the NiAl system in many

aspects. This system is based on equiatomic binary NiMn with Ms of 674°C [38].

Transformation is observed only for equiatomic NiMn and Mn rich compositions,

proceeding from B2 austenite to L10 tetragonal martensite (θ phase) [79]. Increasing

Mn content in Mn-rich compositions, decreases transformation temperatures slightly

[80]. Although the thermal hysteresis in this system is very large (75 – 150°C), the

20

martensitic transformation was observed to be thermoelastic in nature based on

microstructural studies of the resulting martensite structure. In addition, binary NiMn

alloys are extremely brittle, particularly in Mn-rich compositions such as Ni45Mn55,

where cracks were found in a specimen right after quenching.

Fig. 2.4 Dependence of martensite peak temperature, Mp on either Ni or Mn contents

in NiMnAl alloy [38]

Ternary alloying imparts profound changes to both phase transformation and

mechanical properties of binary NiMn alloys, and most are based on the NiMnAl

system, where a wide range of transformation temperatures can be obtained. It can be

noted from Fig. 2.4 that increasing the Ni content, Mp of NiMnAl alloy increases.

Although Adachi and Wayman [79] concluded that Ti, Al, Cu or B additions did not

improve ductility, results from later studies contradicted this finding. Instead, it has

been reported that elongation to failure in compression at room temperature

increased from less than 2% strain in Ni50.2Mn49.8 to above 10% for a

Ni40.3Mn50.1Ti9.6 (at%) alloy and over 15% strain in Ni39.7Mn50.3Al10 [81]. Ductility in

bending was worse than in compression in all cases, but up to 10% strain in bending

21

was still possible without cracking in Ni39.7Mn50.3Al10. The improvements due to Ti

addition were argued to be caused by a decrease in grain size, and improvements

from Al addition to be a result of the formation of the ductile γ phase. Replacement

of Ni by Al, Ti, Cu or Fe reduces transformation temperatures more than replacement

of Ni by Mn in identical amounts [82]. These ternary additions up to 10 at% also

lowered ΔT to 25°C. Addition of titanium or aluminum also led to changes in the

martensite structure. Kainuma et al. [38] indicated that in the NiMnAl system with

more than 30 at%Mn, an additional monoclinic 7R martensite can be found. Jee et al.

[40] reported that 10 at%Ti addition caused a complex 4R orthorhombic martensite

to form in the Ni–Mn–Ti system.

Shape memory properties of Ni–Mn based alloys are far superior to those of the NiAl

system. Specimens deformed in compression to 3% total strain yielded εrec of 2.8%,

and 5% εrec was obtained from 8% total applied strain in Ni45.1Mn44.9Ti10. In bending

thermal cycling experiments under constant stress of 290 MPa, 3.5% εrec (88%

recovery rate) of ~4% applied strain was also found in Ni42.5Mn50Ti7.5. Shape

memory reversibility is inferior in NiMnAl alloys compared to NiMnTi because of

non-transforming and soft γ phase precipitates. Maximum of 0.6% εrec with complete

recovery is possible in bending thermal cycling experiments of Ni39.7Mn50.3Al10, and

εrec of 1.5% is possible in the same alloy with 75% recovery rate [39]. Yang and

Mikkola [83] investigated the shape memory recovery rate of several NiMn based

ternary and quaternary alloys in the solution treated conditions under compression at

room temperature. While none of these alloys are capable of full recovery, both

NiMnTi and NiMnAl alloys showed reasonable recovery rates of around 70% up to

very high applied strain levels.

Similar to NiAl alloys, diffusion based decomposition processes also occur in NiMn

alloys when the Ni content is less than 50 at% for binary NiMn, ternary NiMnFe and

NiMnAl, and quaternary NiMnAlFe alloys [41]. Decomposition of γ phase during

aging in Ni60Mn16Al19Fe5 causes large shifts in the transformation temperatures.

Transformation temperatures in this alloy are also argued to be affected by

quenching rate, which is most likely related to the formation of γ phase during

quenching [84].

22

2.5.2 Copper (Cu) based high temperature shape memory alloys

Cu-based SMAs are the second commercialized SMAs after NiTi-based SMAs.

Although in polycrystalline form, the mechanical and shape memory properties of

Cu-based SMAs are inferior as compared to NiTi-based SMAs, however due to the

lower cost of Cu-based SMAs and ease in development process, their application is

comparable with NiTi-based SMAs in many fields. Among the two main Cu-based

binary alloy systems, (CuZn and CuAl), CuAl is considered to be more suitable

candidate for high temperature shape memory applications. As this alloy system

exhibits higher transformation temperatures and better microstructural stability as

compared to CuZn.

In the binary CuAl system, when the content of Al is reached closer to the eutectoid

concentration, the parent phase becomes unstable and decomposes to equilibrium α

and β phases. Hence for stabilization of parent phase, Ni is added. However by

ternary alloying of Ni, transformation temperature lowers by ~ 20°C/wt-% [63]. For

single crystal CuAlNi alloys, the shape memory and superelastic properties are

remarkable. In single crystal Cu81.8Al14Ni4.2 (wt-%) at deformation temperature of

205°C, perfect superelastic recovery for the applied strain of 17% is reported, due to

a distinctive stress induced martensitic transformation in two steps [60, 64]. For the

other single crystal CuAlNi alloys, complete superelastic and shape memory

recovery of about 8% is a normal behavior for different compositions as stated above

[61, 85]. However for the polycrystalline CuAlNi alloys, mechanical and shape

memory properties are significantly degraded due to transgranular fracture produced

by large grains of about 1 mm. The other problems for these alloys are the elastic

anisotropy and brittle γ phase precipitation at the grain boundaries, lowering the

ductility of material [86]. Moreover exposure at high temperature for long time

creates instability in the transformation behavior [71]. Moreover transformation is

strongly dependent on Al content as shown in Fig. 2.5. By adding Aluminum content

in place of Cu, the Ms temperature quickly drops below room temperature.

Several efforts have been made to increase the ductility by micro alloying of fourth

elements. In this regard titanium [87, 88], boron [89], vanadium [90] and zirconium

[7] have been added in small amounts, reduced the grain size significantly and

23

resultantly the improved ductility was observed. Due to reduction in diffusion of the

constituent elements, titanium is significantly effective for reducing the grain size as

compared to other elements [88]. By increasing the nickel content, precipitation of

brittle γ phase can be lowered, however significant increase can causes reduction in

ductility as well. Use of manganese in place of nickel is more effective in improving

the ductility and in lowering the brittle γ phase precipitation [91].

In solution treated condition, the shape recovery of CuAlNi alloys is poor when the

stress is applied at room temperature, however it improves by increasing the

deformation temperature [92]. Thus it was concluded that the deformation by stress

induced martensite just above Af resulted in less plastic deformation as compared to

the deformation by martensitic orientation below As [93].

Superelasticity for some CuAlNi alloys has been reported when these are deformed

above Af. It was found by Morris [91] that complete superelastic recovery can be

obtained for 2% applied strain in a solution treated Cu80.96Al12Ni4Mn3B0.04 at 150°C

and shape recovery of 4.5% can be obtained for Cu79.96Al12Ni4Mn4B0.04 at 150°C.

To improve the ductility and phase stability in CuAl binary alloys, niobium [62, 94-

96] and silver [97-99] have been added in place of nickel recently. By addition of

niobium, ductility was improved to 12.5% tensile strain at room temperature by

reducing the grain size in Cu84.44Al13Nb2.56 alloy with Ms of 305°C [95]. However, by

adding silver, the ductility was reduced to less than 1% in. Cu83.6Al10.6Ag58 alloys

[99]. Addition of either Nb or Ag reduces the decomposition process, and also

increases transformation temperatures [97, 99] Ms temperature as high as 313°C can

be reached in Cu86.23Al13.5Nb0.27 [95] , and 376°C in Cu87.5Al9.8Ag2.7 [97].

Although the increase in transformation temperatures may seem appealing at first

glance, it is important to note that the temperature for the activation of

decomposition processes is only slightly increased. For higher transformation

temperature CuAlAg alloys, the martensitic transformation vanishes after the alloys

have been heated to above their Af temperature (over 400°C), due to phase

decomposition in this temperature range [99]. Cyclic stability also suffered and as a

result the transformation temperatures shifted by 25°C after five DSC cycles in

24

CuAlNb alloys [62]. It is very difficult to suppress decomposition processes at such

high temperatures and the useful lifetime and stability of these alloys are very

limited. The shape memory behavior of these alloys are also unknown; while it was

mentioned that the Cu86.23Al13.5Nb0.27 alloy is capable of 98% recovery rate after

bending at room temperature, the applied strain level was not mentioned [95].

Fig. 2.5 Composition dependence of the martensite start temperature, Ms on

aluminum content in CuAlNi alloys [61]

Further alloying the CuAlNb system with nickel, chromium and cobalt can lower

transformation temperatures to levels where decomposition occurs more slowly, as

well as increase the intrinsic stability of the austenite and martensite [94]. While the

ductility of these quaternary alloys are worse than ternary CuAlNb alloys, tensile

elongation to failure in all compositions is still around 6–8% at room temperature

and remains superior to identically processed CuAlNi alloys [96]. Transformation

temperatures remain stable in nickel containing quaternary compositions with Af of

179°C after 30 thermal cycles, and decomposition did not affect the behavior until

after 1000 min aging at 300°C [94]. Quaternary additions have little effect on shape

25

memory behavior, with the exception of chromium, which decreases the maximum

recoverable strain. Shape memory behavior has not been studied in the CuAlAg

alloys. Other potential Cu Based SMAs include CuAlCo and CuAlZr with Mp around

250°C and thermal hysteresis of nearly 100°C and show poor cyclic and thermal

stability [7].

2.5.3 Nickel Titanium (NiTi) based high temperature shape memory alloys

NiTi alloys can be alloyed with various elements like Pd, Pt, Au, Zr, Hf, etc. to form

various ternary and quaternary alloys. The transformation temperatures (TTs) are

very strongly dependent on the composition in NiTi alloys [100]. In Ni-rich

intermetallic compounds, even a slight increase of the Ni content away from the

stoichiometry leads to a sharp decrease in the TTs while the Ti rich side of the

intermetallic compound is much less sensitive to the mentioned compositional

variation primarily due to precipitation of Ti2Ni particles which counteract the

compositional variation of the matrix. Addition of ternary element within the range

of less than 10% usually decreases the Ms or usually has very little effect.

Substitution of Fe or Co in place of Ni; or Mn, Cr, V when substituted for Ti

decreases the TTs very significantly. Addition of more than 10% Pd, Pt, Au, Hf and

Zr to NiTi increases the transformation temperature [101].

2.5.3.1 Nickel Titanium Hafnium (NiTiHf) and Nickel Titanium Zirconium (NiTiZr)

high temperature shape memory alloys

The NiTiHf and NiTiZr systems are possible alternatives to the much more costly

precious metal containing alloys like NiTiPd, NiTiPt and NiTiAu. In this case, the

ternary alloying elements (Hf and Zr) are substituted at the expense of titanium. A

prime motivation behind studying NiTi (Hf,Zr) alloys is the relatively low raw

material cost of Hf and Zr, at least compared to Pt, Pd and Au, and their greater

influence on transformation temperatures, which allows them to be used in smaller

concentrations. Since the mechanical and shape memory properties of NiTiHf and

NiTiZr alloys are very similar, they will be discussed concurrently in this section.

Additions of hafnium above 3 at% as shown in Fig. 2.6, increase the transformation

temperatures of the binary NiTi system [102]. This increase is only ~5°C/at% up to

26

5–10 at% hafnium, then elevates to over 20°C/at% thereafter. In NiTiZr alloys, any

increase in transformation temperatures occurs above 10 at% zirconium as shown in

Fig. 2.7. At this level, further zirconium addition increases Ms at a rate of ~18°C/at%

[53]. Therefore, hafnium elevates transformation temperatures more effectively than

does zirconium at identical concentrations [93]. In the solution treated state, the

martensitic peak temperature Mp is 190°C in Ni49Ti36Hf15 and 160°C for

Ni49.5Ti35.5Zr15 [103]. Transformation temperatures are not notably affected by a

change in nickel content in either system as long as the alloys were (Ti,Hf/Zr) rich,

but dropped steeply when nickel content is increased beyond the equiatomic (50 at%)

composition [104] consistent with the behavior of Ni–Ti alloys in general.

Fig. 2.6 Composition dependence of the martensite peak temperature, Mp as a

function of hafnium content in NiTiHf alloys [102]

27

Most research efforts have focused on (Ti,Hf/Zr) rich alloys with under 20 at%Hf/Zr.

In these alloys, a single stage transformation of cubic B2 parent phase to monoclinic

B19´ martensite takes place, [105] whereas in higher Hf/Zr content compositions

approximately above 20–25 at%, orthorhombic B19 martensite appears. Ti2Ni, [106]

or Ti4Ni2Ox (x˂1) type precipitates are found in (Ti,Hf) rich NiTiHf alloys depending

on interstitial oxygen level. The situation is more complicated in NiTiZr alloys. In

addition to Ti2Ni type precipitates, a phase with relatively low melting point named

λ1 appears in alloys in excess of 10 at%Zr in (Ti,Zr) rich compositions when cooled

slowly from solution [107]. NiTiZr appears to be physically less stable than NiTiHf

alloys with identical Ti and Ni content, and the volume fraction of the precipitates in

the NiTiZr alloys is generally higher. In both systems, precipitate volume fraction

increases as alloy composition moves deeper into the (Ti, Hf/Zr) rich region of the

ternary phase diagram. Both NiTiHf and NiTiZr are somewhat brittle in tension at

room temperature [108]. For example, Ni49Ti36Hf15 failed in tension at 7% strain

[109] at room temperature; fortunately, ductility is improved at higher temperatures,

and in the same alloy deformed at 260°C in full austenite, fracture did not occur until

30% tensile strain. Ductility worsens sharply with increasing Hf and Zr

concentrations [110]. Addition of 0.1 at% boron did not improve the ductility of

Ni49Ti36Hf15. Shape memory behavior of NiTiZr alloys is in general inferior to those

of NiTiHf alloys. Full recovery of up to 1.8% applied strain in bending at room

temperature was possible for Ni50Ti35Zr15, and up to 1.6% applied strain in bending

could be fully recovered in solution treated Ni50Ti30Zr20 alloys. A full recovery of

1.8% applied strain in Ni50Ti35Zr15 and an 85% recovery rate for 2.8% applied strain

in solution treated Ni49.5Ti35.5Zr15 were also observed after room temperature

deformation in compression.

Meng et al. [106] reported that aging at 700°C severely lowered the transformation

temperatures of Ni49Ti36Hf15. After 20 hours aging, Ms and As dropped by

approximately 70°C and 40°C respectively; these changes were accompanied by the

precipitation of the (Ti,Hf)2Ni phase. An increased σy and ultimate tensile strength in

martensite of the aged alloy provide further evidence of precipitate formation.

During the course of aging, these precipitates grew in size and volume fraction, and

the authors concluded that 20 h were needed for precipitates to reach the peak aged

condition and impart the best improvement in mechanical and shape memory

28

properties. Further aging coarsened the precipitates, and the strengthening effect was

lost.

Fig. 2.7 Composition dependence of the martensite peak temperature, Mp as a

function of zirconium content in NiTiZr alloys [102]

Aging in Ti rich compositions of NiTiHf alloys have shown a clear improvement in

recovery rate due to precipitation hardening when the alloy is near or at the peak

aged condition. Although Meng et al. [106] noted that shape memory recovery of

the material after aging was inferior to that found by Wang et al. [108], the initial

applied strain was larger in the former study, and improvements of recoverability of

the aged specimen over an unaged one was clearly demonstrated. More recently,

Meng et al. [36, 111] successfully achieved high transformation temperatures in Ni

rich NiTiHf SMAs through aging.

One other concern with the NiTiHf/Zr HTSMAs is that they generally have wide ΔT,

an unfavorable characteristic for actuator applications. Similar to binary NiTi alloys,

the martensite structure of the composition range of NiTiHf/Zr currently being

studied as HTSMAs is monoclinic B19´. Applying a similar strategy that proved

29

effective for the reduction of ΔT in binary NiTi [6], Meng et al. [112] alloyed NiTiHf

HTSMAs with copper and sought to change the martensite structure from B19´ to

B19, but found copper additions of 1 to 5 at% to Ni49Ti36Hf15 did not change the

martensite structure, and actually increased ΔT instead. In addition, the 5 at% copper

addition had no effect on transformation temperatures while 3% copper caused a

slight decrease. Although stress free thermal cycling in the alloy with 5 at% copper

produced double transformation peaks in DSC during heating, which was reported to

be R-phase transformation from XRD and DSC analyses, there was no obvious

improvement on the cyclic thermal stability in terms of the shift in transformation

temperatures [112].

2.5.3.2 Titanium Nickel Platinum (TiNiPt) high temperature shape memory alloys

Transformation temperatures in the TiNiPt system can be increased by replacing

nickel with platinum, but only after a threshold value of approximately 10–15% Pt is

reached [113] as shown in Fig. 2.8. At ≤10% Pt, the martensite structure is B19´

(monoclinic), and transformation temperatures are relatively insensitive to Pt content

or decreases slightly with a minimum observed at 5 – 10 at%Pt. However, at higher

levels of Pt, at least 16 at% or greater, the martensite formed is B19 (orthorhombic)

[114] and transformation temperatures increase linearly with Pt concentration until

transformation temperature near 1000°C is reached for the binary TiPt alloy. The

transformation temperatures for (Ni,Pt) rich compositions do not decline sharply with

deviations from stoichiometry and a (Ni,Pt)3Ti2 precipitate phase readily occurs in

alloys slowly cooled from elevated temperatures.

Lindquist and Wayman [115] recorded a rather small ΔT for their TiNiPt alloys,

except for the highest Pt content studied at 30%. This particular alloy exhibited ΔT

of over 80°C, which is consistent with the later data from Rios et al [113] who

reported a rather wide ΔT (40 – 80°C) for alloys containing 30% or more Pt. On the

other hand, ΔT is generally less than 20°C for alloys with 25% or less Pt, which is

desirable in any applications requiring active control of the transformation or where

cycle time and frequency is critical. Lindquist and Wayman [115] attempted to

measure the recoverable shape memory strains in TiNiPt alloys but were limited by

the extremely low room temperature ductility of their alloys. Hosoda et al. [69] have

30

also measured the room temperature tensile properties of Ti50Ni50-xPtx alloys from 0

to 50 at% Pt and found that ductility generally decreased as the amount of Pt

increased, but that σy followed a much more complicated dependence on

composition. Nickel actually appears to harden TiPt alloys considerably, while Pt

additions to NiTi soften the monoclinic martensite phase, at least initially. Compared

to the TiNiPd alloys, the stress free shape memory response of TiNiPt alloys has not

received a similar level of attention, with the exception of the attempts by Lindquist

and Wayman [115] in tension and Meisner and Sivokha [114] for a Ti50Ni34Pt16 alloy

in torsion. Instead, more effort has been focused toward the constant stress thermal

cycling behavior of TiNiPt alloys including the properties of Ti50Ni30Pt20 bar, fine

rod and wire. Alloys with transformation temperatures up to ~300°C show

reasonable work output with good reversibility for near stoichiometric compositions

and Pt contents up to 20%, even in the as processed, i.e. hot extruded, condition with

no prior training.

Fig. 2.8 Composition dependence of the transformation temperature as a function of

platinum content in TiNiPt alloys [113]

As with the other precious metal additions to NiTi, the high initial material costs will

have to be offset by the increased benefits in using HTSMAs at the system level,

31

probably limiting bulk material to high performance and demanding applications,

such as those existing in the aerospace industry. While the data in most cases are still

cursory, the TiNiPt alloys with 20% Pt or less show promise for such demanding

applications, since they have high transformation temperatures, good work output

and dimensional stability below ~350°C, low ΔT, good thermal stability and

excellent oxidation resistance below ~500°C. Certain optimization studies are

required to be performed on the TiNiPt system to further improve the properties and

consider for commercial consideration.

2.5.3.3 Titanium Nickel Paladium (TiNiPd) high temperature shape memory alloys

Interest in the TiNiPd system as potential HTSMAs was derived from three sets of

studies: the comprehensive study of phase transformations in binary B2 titanium

alloys [116], the discovery of high transformation temperatures in the TiPd binary

systems by Donkersloot [117] and the discovery of ternary alloying effects on the

transformation temperatures of binary NiTi SMAs by Eckelmeyer [9]. Based on the

results from these studies, palladium was added to the TiNi system in order to

increase transformation temperatures, and nickel to the TiPd system to improve

shape memory behavior. TiNiPd HTSMAs have received the most rigorous attention

over the years. Initial focus was centered on improving their high temperature shape

memory behavior, but more recently, the focus has shifted towards improving their

work output, as well as dimensional and microstructural stability. In this system,

transformation temperatures can be altered by replacing nickel with palladium. If the

concentration of titanium is held constant at nearly 50%, the relationship between the

transformation temperatures and relative concentration of nickel and palladium is

parabolic, as shown in Fig. 2.9. A minimum in transformation temperatures occurs at

approximately 10% Pd, although the exact composition of this minimum is still

subjected to debate. In compositions with palladium concentrations greater than the

palladium concentration at this minimum, replacing nickel with palladium increases

transformation temperatures by approximately 15°C/at%. [68, 118, 119]. On the

other hand, if the concentration of palladium is lower than the composition at the

minimum, replacing nickel with palladium actually lowers the transformation

temperatures by 4°C/at% [115]. This parabolic dependence of the transformation

temperatures on composition stems from the change in the structure of martensite.

32

On the higher palladium concentration side of the minimum, B2 austenite transforms

to B19 orthorhombic martensite, and at the lower palladium concentration side, it

transforms to B19´ monoclinic martensite or R phase. The composition at which the

transformation temperatures are at a minimum corresponds to the point of the

structure transition. Because of the complete mutual miscibility of the TiNi and TiPd

systems, the relationships between transformation temperatures and composition

hold over all ranges of palladium concentration. This enables the access to a

continuous range of transformation temperatures from room temperature to over

500°C by adjusting the amount of palladium in the alloy. Otsuka et al [120] reported

poor shape recovery of binary Ti50Pd50 HTSMA, concluding that this was primarily

caused by the low σy of both austenite and martensite. They proposed three possible

solutions to improve its shape memory behavior by strengthening the material

through:

Solid solution hardening via ternary alloying

Thermomechanical processing

Precipitation hardening or some combination of these approaches.

As the following discussion demonstrates, these solutions have met with varying

degrees of success. While the main effect of nickel addition to Ti–Pd alloys is the

reduction in transformation temperatures, it also has an indirect effect of improving

shape memory behavior by lowering the temperature range at which the alloy would

be used. At a lower temperature, σy is at a higher level. Khachin [116] observed full

recovery of 4% applied strain in high temperature torsion experiments on

Ti50Ni13Pd37. However, Lindquist and Wayman [115] studied the same alloy at room

temperature under tension, and were only able to obtain 40% recovery of 6% applied

strain. The reason for this discrepancy was not resolved, but it is likely due to the

differences in the way the materials were processed and tested, which can have a

remarkable effect on recovery rate in TiNiPd alloys. Although the σy of TiPd alloys

can be raised indirectly through nickel addition, it is still too low for perfect shape

recovery.

33

Fig. 2.9 Change in martensite start temperature; Ms with respect to palladium content

in equi-atomic TiNi50-xPdx alloys [68]

Several researchers addressed the issue of low σy in TiNiPd HTSMAs by further

solid solution strengthening. Suzuki et al. [68] and Yang et al. [121] added small

amounts of boron to Ti50Pd30Ni20 and Ti50.7Ni22.3Pd27 alloys, respectively, but neither

caused a notable reduction in εirr with boron additions of up to 0.2%. Micrometer

sized Ti2B particles were found along the grain boundaries, but they were too large

and non-uniformly distributed to possibly function as particles for precipitate

hardening. Instead, boron acted as a grain refiner by reducing the grain growth rate in

these alloys. In identically solution treated or hot rolled specimens, 0.12 at% boron

reduced the grain size from ~40 μm down to 10 μm in the Ti50.7Ni22.3Pd27 HTSMA.

Presumably for the same reason, 0.2% boron addition to Ti50Ni30Pd20 doubled the

room temperature tensile elongation to failure from 8 to 16% strain and increased the

ultimate tensile strength from 460 to 800 MPa for a sample deformed in martensite at

a temperature of 170°C. It is not clear why the grain refinement effect of boron did

not improve the shape memory behavior, since σy increases with grain size

refinement. In other studies on solid solution hardening, 5% gold or platinum

34

replacing palladium in Ti50.5Ni19.5Pd30 increased σy and mildly enhanced cyclic

stability, but had little effect on total recoverable strain. From these results, it appears

that increasing σy alone will not necessarily always result in improvements in shape

memory behavior. Another recent study similarly observed little impact on the shape

memory recovery rate after alloying Ti50.6Ni19.4Pd30 with 1 wt% cerium [122], no

reason for this observation was given.

Although boron and cerium do not appear to improve the shape memory behavior, a

recent study by Atli et al [123] showed that 0.5 at% scandium addition to

Ti50.5Ni24.5Pd25, replacing titanium, was more effective in this aspect. Although the

scandium lowered transformation temperatures by about 6 – 10°C, it was able to

reduce εirr from constant stress thermal cycling experiments under tension by half at

stress levels above 200 MPa without adversely affecting εrec. This improvement was

believed to be caused by the solution hardening effect of scandium, but the effect of

scandium on the σy of martensite and austenite was not explicitly shown.

Additionally, scandium addition also reduced ΔT and improved cyclic stability. After

10 thermal cycles at 200 MPa, the cumulative εirr was reduced by ~20% in the

scandium containing specimen.

Investigating TixPd30Ni70-x (x = 48.5 to 51.0%) alloys encompassing both sides of the

equiatomic composition, Shimizu et al. [124] realized (Fig. 2.10) that with

decreasing titanium content, transformation temperatures decrease only slightly in Ti

rich compositions, but drops off dramatically on the Ni/Pd rich side, such that Ms

declines to room temperature in a Ti48.5Pd30Ni21.5 alloy. This is similar to the

composition dependence of transformation temperatures in binary NiTi SMAs, and is

rationalized by the solubility of excess titanium or nickel atoms near the equiatomic

composition.

35

Fig. 2.10 Martensite start temperature; Ms as a function of Ti/(Ni,Pd) ratio [125]

Solubility for extra titanium atoms in near equiatomic NiTi is almost negligible, and

although the solubility for extra nickel atoms is also small, it is possible to

accommodate excess Ni concentrations under 1 at% in solution. Therefore, in

titanium rich compositions, the extra titanium atoms tend to immediately form

second phases, and do not affect the composition of the matrix and thus the

transformation temperatures of the alloy. However, a small concentration of extra

nickel atoms can dissolve in the matrix, changing its composition and transformation

temperatures. Utilizing precipitates in Ti rich compositions after proper heat

treatments, Shimizu et al. [124] were able to demonstrate that Ti50.6Pd30Ni19.4

outperformed the equiatomic Ti50Pd30Ni20 alloys with a recovery rate of 90% versus

78% respectively, on samples deformed to 6% applied strain at 200°C under tension.

The improvements were attributed to the homogeneous distribution of fine

Ti2(Ni,Pd) type precipitates formed during the annealing process.

Shirakawa et al. [126] and Nagasako et al. [127] studied the phase transformations

in the (Ni,Pd) rich TiNiPd alloys, and compensated for the decrease in

36

transformation temperature from the nickel rich matrix by increasing the palladium

content. While the two way shape memory behavior of a Ti48Pd31Ni21 alloy annealed

at 400°C was qualitatively described, no quantitative data on either one way or two

way shape memory behavior were available. Furthermore, it should also be possible

to maintain reasonably high transformation temperatures in nickel rich compositions

without increasing Pd content by applying appropriate heat treatments. The reported

precipitates are nickel rich in nickel rich compositions, and were identified to be

Ti2(Ni,Pd)3 and Ti3(Ni,Pd)4 types, so as these precipitates grow, nickel content in the

matrix should be reduced, thereby increasing transformation temperatures.

Martensite reorientation, and thus the shape memory behavior was absent in 21–29%

cold rolled Ti50Pd30Ni20 alloy during subsequent deformation at 173°C when post-

rolling annealing temperatures were below As. Moreover, shape memory behavior

deteriorated if the post-rolling annealing temperatures were very high. Golberg et al.

[128] concluded that annealing at a temperature below As does not allow the

preferentially oriented martensite inherited from the rolling process to reset to a self-

accommodated martensitic morphology. Conversely, annealing far above Af initiated

microstructural recovery, destroying the work hardening effects of cold rolling. An

ideal post-rolling annealing temperature was proposed to be above As but below the

recrystallization temperature. A follow-up study by Xu et al. [119] examined the

recovery and recrystallization processes over a full range of cold rolled and annealed

Ti50(Ni,Pd)50 alloys with palladium contents ranging from 0 to 50 at%. They

observed that recovery started at 450°C in Ti50Pd30Ni20 alloy and recrystallization

began at 550°C in Ti50Pd40Ni10 alloy. Physically, these diffusion driven mechanisms

reduce εrec by increasing grain size and reducing dislocation density from cold

working. If recrystallization takes place in martensite, reverse transformation

temperatures (As and Af) increase due to the loss of internal twins in the martensite

[129]. This occurs because during phase transformation, elastic energy is stored in

the internal twins of martensite as elastic strain. Recrystallization replaces these

martensites with the ‘new’ strain free martensite, and the stored elastic energy is lost.

Since stored energy is a driving force for the reverse transformation, its loss results in

the increase in As and Af.

In particular, recovery and recrystallization temperatures in the range of 450 – 600°C

prevent effective thermomechanical treatment of alloys with high Pd contents and

37

high transformation temperatures since recovery would occur within the operational

temperature range of the alloy. Recovery should also cause thermomechanical

training and TWSME in these HTSMAs to be almost impossible. This factor has

severely limited the development of higher temperature TiNiPd HTSMAs with Pd

contents beyond about 35 at%. Similarly, thermal cycling can also affect the stability

and properties of SMAs either by introducing a buildup of internal stresses or

through their relaxation. For example, thermal cycling of Ti50.6Pd30Ni19.4 under 200

MPa increased transformation temperatures by up to 16°C [130] and stabilized after

about 30 cycles.

In contrast, thermal cycling under no stress reduced transformation temperatures by 5

– 10°C, with alloys of higher Pd content exhibiting a greater shift in transformation

temperatures during cycling [131]. It was suggested that dislocations introduced

during cycling suppressed the martensitic transformation during stress free thermal

cycling, but the oriented internal stress fields generated from isobaric thermal cycling

assisted the external applied stress, increasing transformation temperatures [130].

TiNiPd alloys containing less than ~35% Pd have acceptable work characteristics but

tend to suffer from rather high εirr during thermal cycling under load. This

irreversibility worsens with increasing applied stress and increasing Pd concentration

at constant stress. Therefore, a significant focus of the recent research on TiNiPd

alloys has been in pursuit of mechanisms that reduce εirr. In particular, Bigelow et al

[132] have found that alloying with Pt or Au reduced εirr considerably in the absence

of thermomechanical training, but no significant reduction of εirr from these alloying

elements occurred in trained specimens. Not only are these approaches useful in

reducing εirr in TiNiPd alloys, they also have no detrimental effect on recoverable

strain and thus relative work output for these materials.

The amount of εirr for a given alloy and stress level increases as the maximum

temperature to which the sample is thermally cycled increases, with all other

considerations being constant, thus placing another criterion on the upper

temperature limit for these materials. However, certain alloying additions such as 5%

Pt increases the maximum temperature capability of TiNiPd alloys by ~30°C, thus

providing some measure of protection against overheating.

38

Fig. 2.11 (a) Specific work output for a Ti50.5Ni19.5Pd30 alloy as a function of applied

stress loaded in both tension and compression and (b) the corresponding

transformation strain versus applied stress [125]

Similar to conventional NiTi SMAs, the work output for TiNiPd alloys peaks at

some optimum stress level [125] as shown in Fig. 2.11a. This peak in work output is

a result of competing factors. As the stress level increases, more martensite variants

become favorably aligned and detwinned, forming more single variant morphology

and accommodating larger strains. At the same time, the increasing stress level

approaches σy of the alloy and causes irrecoverable deformation to occur instead of

martensitic transformation, and negatively affects the recoverability of martensite.

Consequently, these factors cause εirr to reach a maximum and then rapidly decrease

with increasing stress as shown in Fig. 2.11b for the Ti50.5Ni19.5Pd30 alloy. Therefore,

even though the applied stress continues to increase, εirr diminishes and its product

with applies stress, which is work output, reaches a maximum at a particular stress

level as well.

The maximum work output for various TiNiPd alloys as a function of the

transformation temperatures range, such that the full range over which the

transformation occurs (Mf to Af) for each alloy is clearly indicated [133], is shown in

Fig. 2.12. Large and consistent work output of between 8 and 11 J/cm3 is achieved

for alloys with transformation temperatures between 100 and 300°C. However, for

39

alloys with transformation temperatures above this range, the work output drops off

almost catastrophically so that alloys with transformation temperatures in the

neighborhood of 500°C are capable of essentially zero work output, placing another

limit on the viability of HTSMAs beyond just the need for a high transformation

temperature. This is a natural consequence of the reduction of σy at higher

temperatures and the onset of creep deformation [134]. The experiments by Bigelow

et al [135] are very useful in revealing the critical factors that would lead to the

development of superior alloys for actuator type applications. These are: a low σDT; a

high σy to any types of slip or other processes that would lead to irreversible

deformation of the martensite phase, and a high σy of austenite. Therefore, the key to

develop dimensionally stable HTSMAs with good work output is to prevent the

plastic deformation processes while not greatly affecting reorientation/detwinning

stress.

Fig. 2.12 Work output for a series of TiNiPd and TiNiPt alloys as function of the

transformation temperature range (Mf to Af) [133]

The primary methods for accomplishing this are the same as those originally

suggested by Otsuka et al. [136] for improving the stress free shape memory

40

behavior of TiPd alloys, namely: solid solution hardening, the use of

thermomechanical processing, and precipitate strengthening.

The only study on the direct effect of creep on the shape memory behavior was

carried out on Ti50Ni10Pd40 alloy with Ms of 388°C and Af of 411°C [134].

Significant amount of εirr was accumulated during tensile thermal cycles between 300

and 520°C under 200 MPa. When the heating and cooling rate was decreased from

20 to 2°C/min, εirr after one heating and cooling cycle almost doubled; an indication

of high creep activity. When the stress level was reduced to 100 MPa for heating

cooling cycles conducted within the same temperature range, εirr levels were much

lower, and did not appear to be a function of heating cooling rate. This finding was

further supported by constant temperature creep experiments under 100 and 200 MPa

at 520°C. After 2.5 hours, total creep strain of 1.2% was recorded for the specimen

under 100 MPa, while the specimen under 200 MPa showed creep strain of 12.4%.

Clearly, the combination of 200 MPa and maximum temperature of 520°C was

sufficient to activate unacceptable levels of creep activity. Although the effect of

creep under 100 MPa is comparatively much smaller, the net effect remains very

significant. Because operating temperatures in SMA components are dictated by the

transformation temperature, the Ti50Ni10Pd40 HTSMA cannot be used above 100

MPa in an actuator type application.

In summary, extensive research and development of TiNiPd alloys have produced

material with reasonable shape memory behavior, low thermal hysteresis and good

work output. On the other hand, reversibility and corresponding dimensional stability

still need to be further addressed and are the primary focus of several continuing

research efforts. However, the current system is an extremely viable alloy for

HTSMA applications in the 150 – 300°C ranges.

41

Chapter – 3

Equipment Setup and Material

Processing

3.1 Introduction

This chapter provides detailed description about the constituent elements, its

processing, equipment setup and different experimental procedures followed during

the course of the present research. A total of four alloys (SMAs) were developed by

using non-consumable electric arc furnace in the presence of argon. The prepared

alloys were then homogenized and water quenched. The homogenized button was

sliced into 0.4 mm thick strips by wire electrical discharge machine (wire-EDM).

The 0.4 mm thick strips were cold rolled by 25% and reduced their thickness to 0.3

mm. Samples for differential scanning calorimetry (DSC), X-ray diffraction (XRD),

microstructural studies, mechanical properties and shape memory measurement were

cut using wire EDM from each alloy. The detail of each process is presented in the

subsequent sections.

3.2 Development of shape memory alloys

The alloys were developed by melting of pure constituent elements. High purity

constituent elements; 99.98 wt.% Ti, 99.98 wt.% Ni, 99.99 wt.% Pd and 99.99 wt.%

Cu were used for the preparation of these alloys. The mass of constituent elements

was taken exactly as calculated from the conversion formula, i.e. atomic percent to

weight percent up to four decimal points of a gram i.e. one-tenth of mg. The

sequence of operation and material processing is shown in Fig. 3.1 and explained in

the following sections.

42

Fig. 3.1 Process flow chart of the sequence of operations for materials processing and

characterization

Mechanical Testing (Hardness + Stress-Strain Diagrams)

Thermo-mechanical Cycling Experiments (Stress of 100 – 500 MPa)

X-Ray Diffractometry (XRD)

Differential Scanning Calorimetry (DSC)

Preparation of Samples for Scanning Electron Microscopy (Polishing)

Preparation of Samples for Optical Microscopy (Polishing + Etching)

Solution Treatment of Samples at 900°C for 1 h and Water Quenched

Preparation of Samples for Different Characterization (Wire EDM)

Homogenization of Cast Buttons at 950°C for 2 h and Water Quenched

Melting of Constituent Elements in Vacuum Arc Melting Furnace

Precise Weighing of Constituent Elements (one-tenth of mg)

Cleaning of Constituent Elements by Ultrasonic Cleaner (in CH3OH)

Arrangement of Pure elements Ti, Ni, Pd, Cu (99.98 wt.% or more)

43

3.2.1 Cleaning of constituent elements

The weighed elements were cleaned by an ultrasonic cleaner in methanol (CH3OH)

for 5 minutes. After cleaning, the materials were taken to electric arc furnace for

melting.

3.2.2 Melting of constituent elements

Titanium is very reactive at high temperature and form stable chemical compounds

with oxygen or nitrogen [137, 138]. To avoid the formation of oxides or nitrides and

to get the desired composition, the melting was done by evacuating the furnace from

air (oxygen) and then partially filling it with 99.99% pure argon gas.

The alloys were prepared in the form of buttons (20 g each) in non-consumable DC

electric arc melting furnace in the presence of argon using tungsten electrode and

water cooled copper crucible. Usually the DC electric arc melting furnace is used for

melting of highly oxidizing materials like Ti. The water cooled copper crucible was

cleaned by grinding paper and acetone to remove the contamination layers from the

surface. The constituent elements were placed separately in the two hemi-spherical

cavities and pure titanium in the third cavity as getter. The function of the getter is to

absorb oxygen content and humidity from the furnace chamber before melting of the

elements. The furnace chamber was vacuumed at 1x10-2

mbar and then flushed with

99.99% pure argon gas. This process was repeated twice to remove oxygen from the

furnace chamber. For the third time the chamber was vacuumed at 1x10-4

mbar and

then small amount of argon gas was supplied to the chamber. The existence of argon

gas in the chamber is necessary to provide medium for production of electric arc

between tungsten electrode and copper crucible. Moreover it also provides shielding

medium to prevent oxidation.

Before melting the constituent elements, first the getter was preheated and melted to

absorb oxygen from the chamber. Then both the required alloys were preheated by

supplying current (60 100 A) and melted by increasing the current to 250 A, while

the voltage of 10 V was kept constant. The required alloys in form of circular buttons

with elliptical cross-section were melted six times and flipped over after each

melting to get maximum alloying homogeneity. After melting, the buttons were

44

weighed again and the mass loss was found to be less than 0.1%, therefore, the actual

composition was considered the same as before melting process i.e. nominal

composition. The same process was repeated for all four alloys. The required mass

of constituent elements for the desired four alloys is shown in Table 3.1. Hereafter

the four alloys are called 0Cu, 5Cu, 10Cu and 15Cu according to their Cu contents in

atomic percent.

Table 3.1 Chemical compositions of the four alloys given in weight percent

Alloy

Composition

(at%)

Constituent Elements (g)

Total Mass (g) Ti Ni Pd Cu

Ti50Ni25Pd25 7.3415 3.4998 8.1587 0.0000 20.0000

Ti50Ni20Pd25Cu5 7.31428 3.58646 8.12852 0.97074 20.0000

Ti50Ni15Pd25Cu10 7.2873 2.6799 8.0985 1.9343 20.0000

Ti50Ni10Pd25Cu15 7.2605 1.7800 8.0687 2.8908 20.0000

3.2.3 Homogenization

To make uniform chemical composition throughout the volume of developed alloys,

homogenization process is necessary to be carried out. Each button of the developed

alloys was kept in quartz tube, evacuated by rotary and diffusion pumps till vacuum

of 1x10-3

mbar, filled the quartz tube with pure argon gas and then sealed. Tube

furnace was heated to a temperature of 950°C and then charged the quartz tubes into

the furnace. The charging of alloy buttons at high temperature was done to avoid

precipitation process which can occur during the slow heating process. After keeping

the alloy buttons at 950°C for 2 hours, they were fast quenched in ice water to avoid

the formation of precipitates during slow cooling. To protect from oxidation, the

alloy buttons were quenched without breaking the quartz tubes. The homogenization

process is shown in Fig. 3.2.

45

3.2.4 Chemical analysis

After homogenization, the chemical composition of the four alloys was determined

by Energy Dispersive Spectroscopy (EDS). The chemical composition of impurities

(O, N, C and S) ingressed during handling and melting of constituent elements was

found by LECO Oxygen/Nitrogen Determinator and LECO Carbon/Sulphur

Determinator. The chemical compositions of all four alloys are summarized in Table

3.2.

Table 3.2 Chemical composition of homogenized alloys

Element

Composition (atomic percent)

Analysis Method 0Cu 5Cu 10Cu 15Cu

Ti 50.08 50.02 50.08 50.11

EDS Ni 25.03 19.87 15.21 9.84

Pd 23.83 25.13 23.72 23.85

Cu 0.00 3.92 9.93 15.14

O 0.022635 0.022526 0.022715 0.022526 O/N Determinator

N 0.043252 0.044535 0.043686 0.044535

C 0.005623 0.005269 0.005356 0.005269 C/S Determinator

S 0.000031 0.000027 0.000024 0.000027

3.3 Sample preparation

3.3.1 Sample dimensions

The homogenized buttons were sliced into 0.4 mm thick strips by wire Electrical

Discharge Machine (EDM). The 0.4 mm thick strips were cold rolled by 25% and

reduced their thickness to 0.3 mm. Samples for Differential Scanning Calorimetry

(DSC), X-Ray Diffractometry (XRD), microstructural studies, mechanical testing

and shape memory measurement were cut using wire EDM. Wire EDM was used for

dimensional accuracy, avoid material loss and reduce machining / cutting stresses.

Dimensions of the samples for various characterizations are shown in Table 3.3.

46

Table 3.3 Dimensions of samples for various characterizations

Characterization Length (mm) Width (mm) Thickness (mm)

DSC 2 2 0.3

XRD 10 5 0.3

Optical Microscope 10 5 0.3

SEM 10 5 0.3

Shape Memory Effect 25 3 0.3

Mechanical Properties 25 3 0.3

3.3.2 Solution treatment and aging

For solution treatment, all the prepared samples were sealed in different quartz tubes

after being evacuated and then filled with pure argon gas. Tube furnace was heated to

a temperature of 900°C and then charged with the quartz tubes. All the samples were

soaked at 900°C for 1 hour in the tube furnace and then fast quenched in cold water

without crushing the quartz tubes. The solution treatment process is shown in Fig.

3.2.

3.3.3 Aging

Tem

per

atu

re (

°C)

Time (h)

2 h at 950°C 1 h at 900°C

Ho

mogen

izti

on

So

luti

on

Tre

atm

ent

Fig. 3.2 Schematic representing homogenization and solution treatment processes

47

After solution treatment, some of the samples were aged at various aging

temperatures for investigation its properties in aged condition and comparing with

that of solution treated condition. The samples from 0Cu, 5Cu, 10Cu and 15Cu

alloys were again sealed in quartz tubes adopting the same procedure as mentioned

above. The sealed tubes were kept in the tube furnace after reaching the required

aging temperatures. All the four alloys were aged at various temperatures ranging

from 400°C to 700°C for time duration of 3 hours as shown in Fig. 3.3. After

finishing the soaking time, the samples were quenched in cold water without

breaking the quartz tubes to avoid direct contact of air and water.

Fig. 3.3 Schematic representing the aging process at various aging temperatures

3.4 Materials characterization

3.4.1 Optical microscopy

Upright Metallurgical Microscope of SINOWON manufacturer China model UMS –

300 was used for metallography. The microscope is capable to magnify the image to

1000X. The samples were first mechanically rough ground by 200, 400 grit papers

and then fine ground by 800 grit emery paper and finally fine polished by 0.5 micron

alumina slurry. After polishing, the samples were etched using reagent in the given

volume ratio 10%HF, 15%HNO3, 75%H2O to reveal the microstructure and grain

boundaries. During this process the samples were dipped in the solution for 5

Agin

g T

emp

eratu

re (

°C)

Aging Time (h)

3 h

600°C

500°C

400°C

700°C

48

minutes while stirring them regularly inside the solution. Moreover, after time

duration of each minute, the samples were washed properly by flowing water and

then dried by hot or compressed air. The samples were regularly checked in the

optical microscope to observe etching after every minute until the grain boundaries

and microstructure were become clearly visible.

3.4.2 Scanning Electron Microscopy (SEM)

For SEM, the samples were first mechanically ground using emery paper of grit

200,400 and 800. Fine polishing was carried out by using alumina slurry of 0.5

micron size. The prepared samples were then loaded in Field Emission Scanning

Electron Microscope (FE-SEM) of TESCAN made in Czech Republic, model

MIRA3 XM. It is operated at 200 V to 30 KV. It is a high resolution SEM and can

generate images of 1.2 nm while its magnification reaches to 1,000,000X at 30 KV.

Images were obtained at different magnification from various locations of the

samples. Energy Dispersive Spectroscope (EDS) attached with FE-SEM was used to

determine quantitatively the chemical composition of constituent elements in each

phase. The chemical compositions of the bulk matrix and other phases present in the

form of precipitates were analyzed by the combine study of the images of SEM and

data from EDS analysis.

3.4.3 X-Ray Diffractometry (XRD)

In the present study XRD patterns were recorded using a computer controlled X-Ray

Diffractometer model JDX-99C JEOL Japan which was operated at 40 KV and 20

mA. The radiation used was Cu Kα (λ=1.5406 Å) at room temperature with incident

angle 2Ɵ from 20° to 80°. The resulted patterns were then compared with the

standard Powder Diffraction File (PDF) of the International Centre for Diffraction

Data (ICDD) database. By comparing the d-spacing of the corresponding diffraction

angle and intensity of diffracted peak, the atomic structure of the different phases

was determined.

49

3.4.4 Differential Scanning Calorimetry (DSC)

DSC is a common technique used for determination of the thermal properties of

specimens with high degree of precision. It is a differential method which calculates

heat flow directly from the samples during heating, cooling or holding isothermally

by measuring the amount of energy required to maintain the specified temperature.

During heating or cooling, if the phase transition occurs in the sample, the heat

released or absorbed is abruptly changed and their corresponding peaks are observed.

In the present study transformation temperatures under stress free condition were

determined by using NETZCH Differential Scanning Calorimeter Germany model

214 Polyma. Using Liquid Nitrogen, it automatically controls temperature, ranging

from – 170°C to + 600°C with heating/cooling rate of 0.001 K/min to 500 K/min.

Each sample was cycled twice through full transformation range from 100°C to

250°C to ensure the reproducibility of thermoelastic transformations. It was observed

experimentally that the transformation temperatures stabilize after two complete

cycles in the solution treated samples and then remain almost constant in subsequent

cycles. The heating and cooling rate for DSC cycles was kept as 5°C per min.

Fig. 3.4 shows a typical DSC plot representing the phase transformation

temperatures. Two peaks are shown, marking the forward and reverse

transformations, where the forward transformation is exothermic and the reverse

transformation is endothermic. The temperatures at start, highest and end points of

the peak formed during forward transformation are known as martensite start,(Ms),

martensite peak(Mp) and martensite finish (Mf) temperatures respectively. Similarly

the temperatures at start, lowest and end points of peak formed during reverse

transformation are known as austenite start (As), austenite peak (Ap) and austenite

finish (Af) temperatures respectively. Phase transformation temperatures (Ms, Mf, As

and Af) of all samples are measured from the intersection of the base line and the

linear portions of exothermic or endothermic peaks as shown in Fig. 3.4. The

difference between austenite finish; Af and martensite start; Ms is called thermal

hysteresis.

50

3.5 Mechanical Testing

3.5.1 Microhardness testing

Microhardness tester of Wilson Wolpert (Model 401) was used to find out the

hardness of the samples. Pyramid indenter of 136° tip was used with 100 g load

applied for 10 seconds dwell time. The hardness values of each sample were

determined from the average of 5 measurements taken at different locations on each

sample.

3.5.2 Isothermal tensile testing

Isothermal tensile tests were performed to measure the mechanical behavior of

solution treated and aged samples taken from the four alloys. The testing was

performed on Instron Tensile Testing Machine Model 1195. In house designed and

manufactured holding grips were used to grip the samples as shown in Fig. 3.5. Load

cell of 100 KN was used and strain measurement in tension was carried out by the

movement of crosshead directly. Samples were heated by induction process and a K-

type thermocouple was attached directly to the central portion of the samples by

Fig. 3.4 Measurement scheme of transformation temperatures from DSC heating

and cooling cycles

50 100 150 200 250

He

at F

low

(w

/g)

exo

up

Temperature (°C)

Reverse Transformation

Forward Transformation

Heating

Cooling

As A

f

Mf M

s

Ap

Mp

51

wrapping the copper wire for temperature measurement. After gripping the sample

and attaching thermocouple, each sample was heated at required isothermal test

temperature while controlling load at 0 N. The samples were allowed to stay at

isothermal test temperature for 5 minutes to stabilize the thermal fluctuations. The

samples were strained at a strain rate of 1 x 10–4

mm/sec. For all alloys, only two

isothermal tensile testing were performed; one at temperature 50°C below Mf and

one at temperature 50°C above Af. All the samples were loaded continuously until

breakdown of samples, cooled down to room temperature and then unloaded.

Fig. 3.5 Special gripping arrangement for holding of 0.3 mm thick samples

Tightening Nut

Holding Jaw

Sample for mechanical and

shape memory testing Complete holding arrangement

52

3.6 Measurement of shape memory properties

3.6.1 Equipment setup

Shape memory properties were measured on Creep and Rupture Tensile Testing

System of 2330 series equipped with Applied Testing System (ATS) frame controller

and Windows Computing Creep System (WinCCS) software as shown in Figure 3.6.

Samples of 0.3 mm thickness were gripped in the sample holder as shown in Fig. 3.5.

Loading was done step by step by increasing from zero to the required maximum

load so that strain rate may be maintained as slow as possible. For tensile strain

measurement, extensometer shown in Fig. 3.6 was used. The extensometer is

consisted of dual rod-in-tube located at 180 degree apart, to which a pair of LVDT is

attached as shown in Fig. 3.6. This extensometer can give reading up to + 50% and -

20% strain in the tension and compression respectively with accuracy of thousandth

of mm. In the present experiments, the strain measurement was carried out by

loading the sample in tension.

During thermomechanical testing, the samples were heated using split tube furnace

equipped with resistive heating nichrome element which can raise the temperature of

the furnace up to 1010°C. Split tube furnace can raise the temperature of the samples

with variety of heating rate having power supply of 3300 W. In the present research,

heating rate of 10°C/min was maintained in all experimentation. The samples were

air cooled with approximate cooling rate of 102°C/min.

For temperature sensing, K-type thermocouple was used and controlled by barber

colman controller. Thermocouple was directly attached to the middle of each sample

by thin copper wire to precisely measure the sample’s temperature.

53

3.6.2 Constant stress thermal cycling tests

These tests were performed by loading the sample in tension, while keeping the

strain rate at lowest level. In these experiments the sample was loaded initially to a

stress of 50 MPa in martensite state and then heated to Af + 50°C (where Af is the

austenite finish temperature under no load). At Af + 50°C, the sample was loaded to

a stress of 100 MPa in an austenite state isothermally followed by cooling to Ms –

50°C (where Ms is the martensite start temperature under no load), while holding the

stress at constant level. The initial half heating cycle at 50 MPa was not shown in the

Fig. 3.7, because this initial behavior was not the representative of thermal cycles at

higher stresses and could not be used for evolution of shape memory characteristics.

Fig. 3.6 Internal view of lever arm creep and rupture tensile testing system

54

During cooling of sample from Af + 50°C to Ms – 50°C, initial linear response

indicate the complete austenitic phase without any deformation. However by

decreasing the temperature further, the sample was started to expand at specific

temperature and continued to remain in the expanding condition till approaching of

another specific temperature. The former and latter specific temperatures are called

martensite start and finish temperatures respectively. Between the phase

transformation temperatures, the alloy is consisted of austenite and martensite phases

and completely transformed to martensite phase at martensite finish temperature.

Upon further cooling, the strain-temperature curve was again become linear showing

no further deformation. By heating the sample from Ms – 50°C to Af + 50°C, the

initial response of the sample was found to be linear and overlap the linear portion of

cooling curve. This behavior indicates the absence of viscoplastic deformation in the

complete martensite state.

During reverse transformation process, when the temperature was increased further,

at specific temperature the sample was started to decrease in length. This behavior

was remained continuously till approaching of another specific temperature, where

the strain-temperature curve was again become linear. The temperatures at which the

shrinkage of the sample was started and then ended are called the austenite start and

finish temperatures respectively. Between the transformation temperatures, the alloy

is again consisted of martensite and austenite phases and completely transformed to

austenite phase at austenite finish temperature. When the sample was heated further,

the strain-temperature curve was again become linear. This process was repeated at

stress levels of 200, 300, 400 and 500 MPa.

Depending on the sample area, loads (in N) were calculated to produce equivalent

stress of 100, 200, 300, 400 and 500 MPa. Transformation temperatures i.e.

Martensite start (Ms), Martensite finish (Mf), Austenite start (As) and Austenite

finish (Af) temperatures under biased load were calculated from the intersection of

the linear portion of the transformation region and linear portion of martensite and

austenite region as shown in Fig. 3.7. Recovered strain (εrec) or transformation strain

was measured by subtracting the strain at the end of heating cycle from the strain at

the beginning of heating cycle. Similarly irrecovered strain (εirr) was measured by

subtracting the strain at the end of the heating cycle from the strain at the beginning

55

of the cooling cycle. The measurement scheme of transformation strains is shown in

Fig. 3.7.

The work output (J/cm3) and the recovery ratio exhibited by the alloy can be

calculated by using equations 3.1 and 3.2 respectively.

Work output = (εrec x σ) ------------- 3.1 [139]

Recovery Ratio = (εrec /( εrec + εirr)) ------------- 3.2 [140]

Where σ is the applied stress at which recoverable strain has been achieved.

Fig. 3.7 Measurement scheme of transformation temperatures, recoverable

and irrecoverable strains from typical strain-temperature curve

0 100 200 300 400

Stra

in (

%)

Temperature (°C)

Ɛirr

Ɛrec

As

Af

Mf

Ms

Heating

Cooling

56

3.7 Summary

In this chapter, material processing starting from the raw materials to the final stage

of alloys melting, heat treatment, samples preparation, characterization, mechanical

testing and constant stress thermal testing were presented. Titanium, Nickel,

Palladium and Copper of highest purity were taken and weighed precisely to get the

required alloys in accurate composition. The constituent elements were cleaned in

methanol using the ultrasonic cleaner. The cleaned constituent elements were melted

in vacuum arc melting furnace for six times to improve alloying homogeneity. Four

alloys of required composition; 0Cu, 5Cu, 10Cu and 15Cu alloys each of 20 g button

were homogenized in sealed quartz tubes separately at 950°C for 2 hours.

After homogenization, each button was sliced into 0.4 mm thick strips using wire

EDM. The same strips were then cold rolled and its thickness was reduced to 0.3

mm. The strips of each alloy were then cut into the required sample dimensions. All

the samples of four alloys were solution treated in sealed quartz tubes at 900°C for 1

hour and cold water quenched without crushing the quartz tubes. After solution

treatment, some samples of each alloy were aged at 400°C, 500°C, 600°C and 700°C

for 3 hours in sealed quarts tubes and then water quenched.

Microstructural characterization like grain size and its structure was carried out by

optical microscope in solution treated condition. By using the SEM, the size, shape

and color of second phase precipitates were found whereas chemical composition

was found by EDS in both conditions; solution treated and aged. Phase analysis of all

four alloys was carried out by XRD for both solution treated and aged conditions.

Similarly stress free transformation temperatures in both conditions of solution

treated and aged were found by DSC. Mechanical properties like microhardness and

stress strain curves were obtained for all four alloys using microhardness tester and

tensile testing machine respectively. Shape memory properties of all four alloys were

found in both conditions of solution treated and aged by using the creep and rupture

testing machine.

LVDT Pushing Supporting

57

Chapter – 4

Effect of Copper Addition and Aging

on Microstructure of TiNiPd Alloys

4.1 Introduction

This chapter presents the effect on the microstructural properties due to varying

percentage (0, 5, 10 and 15 atomic%) of copper (Cu) addition with TiNiPd alloys in

solution treated condition. It also deals with the microstructure analysis of TiNiPdCu

alloys with varying Cu percentage, aged at different aging temperatures (400°C,

500°C, 600°C and 700°C for 3 hours) which in turn changes the final microstructure

of the alloys. Microstructural analysis of these alloys was carried out by using optical

microscopy, Scanning Electron Microscopy (SEM), X-Ray Diffractometry (XRD)

and Energy Dispersive Spectroscopy (EDS).

4.2 Microstructure analysis of solution treated TiNiPdCu alloys with

varying Cu percentage

4.2.1 Second phase precipitates

Fig. 4.1 shows the SEM images of the 0Cu, 5Cu, 10Cu and 15Cu alloys in the

solution treated condition (detail of solution treatment is given in Section 3.3.2).

From Fig. 4.1, it is observed that second phase precipitates of black colors with low

density randomly distributed along the grain boundaries are present in all four alloys.

The average size of these precipitates found in the four alloys was measured to be 1.0

– 2.8 μm. These precipitates had almost the same shape which seems to be of circular

or elliptical type.

This observation shows that addition of Cu had no effect on the precipitate size.

However the density of the precipitates decreased as the Cu content in the alloys

increased. For the 0Cu alloy, the volume fraction of the precipitates was estimated to

be ~2.6%, decreased to ~1.5% for 15Cu alloy. To confirm the chemical composition

of the overall and second phase precipitates, the EDS analysis was carried out. The

58

EDS spectrums are shown in Fig. 4.2, whereas the chemical composition is shown in

Table 4.1.

Fig. 4.1 SEM images showing the second phase precipitates formed along the grain

boundaries in solution treated samples of (a) 0Cu, (b) 5Cu, (c) 10Cu and (d) 15Cu

alloys

10 μm

a

10 μm

b

10 μm

c

10 μm

d

59

The compositional data shown in Table 4.1 suggested that second phase precipitates

formed at the grain boundaries are Ti2Ni(Pd) in 0Cu alloy and Ti2Ni(Pd,Cu) in the

remaining three alloys. Such type of precipitates were commonly observed in Ti-rich

NiTi-based alloys [139]. It was believed that the Ti2Ni(Pd) second phase precipitates

were formed during solidification, because these precipitates had lower melting point

than the matrix [141]. Same type of precipitates were also reported to be formed in

ternary TiNiPd alloys having Pd contents more than 10% [139].

Table 4.1 Compositional analysis of the overall alloy and second phase precipitate in

solution treated condition of 0Cu, 5Cu, 10Cu and 15Cu alloys.

Alloy Name Analysis Region Ti (%) Ni (%) Pd (%) Cu (%)

0Cu alloy Overall alloy 49.2 23.5 27.3 0

Ti2Ni 57.5 27.4 15.1 0

5Cu alloy Overall alloy 49.4 19.1 26.8 4.7

Ti2Ni 56.8 22.3 17.3 3.6


Ti2Ni 66.8 16.4 12.2 4.6


Ti2Ni 59.3 15.4 15.2 10.1

0Cu 5Cu

Fig. 4.2 EDS spectrums shown for solution treated samples of 0Cu, 5Cu, 10Cu

and 15Cu alloys

10Cu 15Cu

60

From these observations it was concluded that by addition of Cu, the basic

microstructure of quaternary alloys did not change. The alloying element Cu

remained in the matrix of solid solution and did not generate any other second phase

precipitate, as was reported in case of Scandium (Sc) addition to TiNiPd [142, 143]

in form of Sc2O3. Similarly second phase precipitates of TiB2 is also reported by

Suzuki et al. [68] to be formed when 0.2% Boron (B) was added in Ti50Pd30Ni20 in

place of Ni.

4.2.2 Grain size

Fig. 4.3 and 4.4 show the optical micrographs of the 0Cu, 5Cu, 10Cu and 15Cu

alloys at magnification of 200X and 500X respectively in the solution treated

condition.

Fig. 4.3 Optical micrographs (at 200X) of (a) 0Cu (b) 5Cu (c) 10Cu and (d) 15Cu

alloys solution treated at 900°C for 1 hour

a b

d c

61

a b

d c

Fig. 4.4 Optical micrographs (at 500X) of (a) 0Cu (b) 5Cu (c) 10Cu and (d) 15Cu

alloys solution treated at 900°C for 1 hour

0

10

20

30

40

Gra

in S

ize

(μm

)

Cu (at%)

0 5 10 15

Fig. 4.5 Effect of increasing Cu-content on grain size

62

From these figures it can be observed that all the alloys are consisted of single phase

having typical twinned martensite structure with clearly visible grain boundaries.

Grain size was estimated by using an intercept method, described as follows. Many

straight lines all the same length were drawn through micrographs that show the

grain structure. The grains intersected by each line were counted, the line length was

then divided by an average of the number of grains intersected, taken over all the line

segments. The average grain diameter is found by dividing this result by the linear

magnification of the micrographs. The average grain size of the 0Cu, 5Cu, 10Cu and

15Cu alloys was found to be 13 μm, 16 μm, 18 μm and 20 μm respectively as shown

in Fig. 4.5. It can be concluded from the above observation that addition of Cu in

place of Ni resulted in an increase of grain size. The average grain size, 13 μm of

0Cu alloy increased to 20 μm in 15Cu alloy showing about 35% increase in grain

size.

The increase in grain size due to replacement of Ni by Cu can be attributed to the

decrease in Ni content. When the Ni content decreased in the alloy, the formation of

the second phase precipitate (Ti2Ni) also decreased. Hence, due to decrease in

formation of second phase precipitates, grain growth increased when Cu content was

increased from 0% to 15%. In case of 15Cu alloy, due to low density and

consequently less pinning effect of second phase precipitates, the grain growth was

greater as compared to grain growth in 0Cu alloy. Therefore, due to reduction in

pinning effect, the grain size of the alloys increased as the Cu content increased.

4.3 Effect of aging temperature on microstructure of TiNiPdCu

alloys with varying Cu percentage

To investigate the effect of different aging temperature on the four alloys, the

samples were analyzed by SEM with backscattered electron detector. Fig. 4.6 (a–d)

represents the back-scattered SEM images of 0Cu alloy aged under the conditions

mentioned in Section 3.3.2. These images show that only the second phase

precipitates, which were formed during solidification process, were present.

However, there was no evidence of precipitation on the grain boundaries as well as in

the grain interior due to the aging effect.

63

Fig. 4.7 (a–d) shows the back-scattered SEM images of 5Cu alloy aged under the

same conditions. It can be observed from these images that the second phase

precipitates were available along the grain boundaries. From Fig. 4.7 a, it can be seen

that due to aging at 400°C, the precipitation process was not started. Increasing the

aging temperature to 500°C, very fine precipitates of bright color were seen along the

grain boundaries. By further increasing the aging temperature to 600°C, the density

of the precipitates also increased. The size of the precipitates was further increased

when the aging temperature was increased to 700°C. These precipitates were found

along the grain boundaries as well as within the grain interior.

Fig. 4.6 Back-scattered SEM images presenting the microstructure and grain

boundaries in 0Cu alloys after aging for 3 hours at temperature of (a) 400°C, (b)

500°C, (c) 600°C and (d) 700°C

10 μm

a

10 μm

b

10 μm

c

10 μm

d

64

The backscattered SEM images shown in Fig. 4.8 (a–d) demonstrate the precipitation

behavior of 10Cu alloys aged under the same condition. Fig. 4.8a presents the

microstructure of the sample aged at 400°C. It can be seen that the precipitation

process has been started and very fine precipitates were formed on the grain

boundaries. It can also be noted that the precipitation was not observed in the grain

interior, confirming that the grain boundaries are working as nucleation sites. Fig.

4.8b shows the microstructure of the alloy aged at 500°C, indicating the population

of two types of precipitates randomly distributed along both sides of grain

boundaries. The size of these precipitates was larger as compared to the precipitates

formed at 400°C. As the aging temperature was increased to 600°C, the density of

10 μm

a

10 μm

b

10 μm

c

10 μm

d



500°C, (c) 600°C and (d) 700°C

65

the precipitates had further increased as shown in Fig. 4.8c. At 600°C, both types of

precipitates were observed to be formed heterogeneously on the grain boundaries as

well as in the grain interior. The shape of the precipitates observed in Fig. 4.8b and c

has thread-like structures with variable length.

Fig. 4.8d represents the microstructure of the alloy aged at 700°C where the

formation of low density and a relatively large size precipitate can be seen. The

precipitation behavior of 700°C-aged alloy was observed to be completely different

from the behavior observed in case of aging at 500°C and 600°C. Formation of nano-

scaled precipitates also reported by Khan et al. [140] to be formed during annealing



500°C, (c) 600°C and (d) 700°C

10 μm

a b

10 μm

d

10 μm

c

10 μm

66

at 550°C, 600°C and 650°C for 1 hour after cold deformation by 40% in the same

alloy. However the precipitates found in annealing process were of nano-scale sizes

(100 – 300nm) with different morphologies (elliptical and rod-like shapes) as

compared to precipitates formed in the present study. Moreover, the precipitation

behavior during annealing was observed to be completely different from that of

aging process. Cold deformation followed by annealing resulted in the formation of

heterogeneous nucleation sites due to deformation-induced defects [6], whereas in

case of age hardening, the high energy grain boundaries were responsible as

nucleation sites for the initiation of precipitation process.

a

10 μm 10 μm

b

c

10 μm

d

10 μm



500°C, (c) 600°C and (d) 700°C

67

In Fig. 4.9 (a–d), backscattered SEM images of aged 15Cu alloys are shown. Fig.

5.9a shows the microstructure of the sample aged at 400°C. It can be seen that very

fine precipitates were formed on both sides of grain boundaries. However the size of

the precipitates was greater as compared to 400°C-aged sample of 10Cu alloy as

shown in Fig. 4.8a. Fig. 4.9b shows the microstructure of the alloy aged at 500°C,

representing two types of precipitates, however the size of these precipitates was

larger as compared to the precipitates formed at 400°C for the same alloy. As the

aging temperature was increased to 600°C, the density of the precipitates had further

increased as shown in Fig. 4.9c. At 600°C, both types of precipitates were observed

to be formed heterogeneously on the grain boundaries as well as in the grain interior.

The shape of the precipitates observed in Fig. 4.9b and c has thread-like structures

with variable length. Fig. 4.9d represents the microstructure of the alloy aged at

700°C where the formation of low density and a relatively large size precipitates can

be seen. The precipitation behavior of 700°C-aged alloy was observed to be

completely different from the behavior observed in case of aging at 500°C and

600°C.

Table 4.2 Compositional analysis of black and white precipitates formed in 10Cu and

15Cu alloys after aging for 3 hours at various aging temperatures

It has been reported [144] that in backscattered SEM images, a precipitate with

brighter contrast represents the existence of heavier elements i.e. elements with

higher atomic number, and a precipitate with darker contrast shows the existence of

lighter elements i.e. elements with lower atomic number. It can be seen from the

periodic table that Ti has lower atomic number than Ni, Pd and Cu and thus it is

lighter as compared to other constituent elements. Therefore the presence of darker

and brighter contrast represented the existence of Ti-rich and Ti-lean precipitates,

respectively [144]. To confirm the exact composition of the two newly born

precipitates as a result of aging at various temperatures, EDS analysis was carried out

Alloy Name Analysis Region Ti (%) Ni (%) Pd (%) Cu (%)

10Cu alloy Black precipitates 60.6 9.3 23.8 6.3

White precipitates 41.1 11.4 26.7 20.8

15Cu alloy Black precipitates 57.3 10.6 24.5 7.6

White precipitates 42.5 10.7 27.3 19.5

68

and their data was summarized in Table 4.2. It can be confirmed from back-scattered

SEM images shown in Fig. 4.8, 4.9 and from EDS analysis that the darker and

brighter contrast particles are Ti2Pd (Ti-rich) and TiPdCu (Ti-lean) precipitates. It

can also be observed that in both types of precipitates, the Ni concentration is lower

than the target composition. So it is expected that by formation of these precipitates,

the Ni content was increased and Ti content was decreased in the matrix.

Comparing the precipitation behavior of 0Cu, 5Cu, 10Cu and 15Cu alloys, it can be

easily observed that stoichiometric (Ti : Ni+Pd = 50% : 50%) TiNiPd alloys were not

responding to any precipitation process within the aging temperature ranges from

400°C to 700°C. The precipitation behavior in 5Cu alloys is not much different from

the 0Cu alloy. However, very fine precipitates of brighter contrast were observed at

aging temperatures of 500°C, 600°C and 700°C as shown in Fig. 4.7 (b – d). By

adding 10% and 15% Cu in place of Ni, the precipitation behavior of stoichiometric

TiNiPd alloys remarkably changed. In both alloys, precipitation started along the

grain boundaries at lower aging temperature. As the temperature increased, the

density and size of precipitates also increased till aging temperature of 600°C. At

700°C aging temperature, the size of the precipitates further increased, however their

density decreased. It was also observed that the size of the precipitates formed in

15Cu alloys was bigger as compared to the precipitates of 10Cu alloys at

corresponding temperatures.

4.4 Phase analysis of solution treated TiNiPdCu alloys with varying

Copper percentage

Phase analysis was carried out using X-Ray Diffractometer with Cu Kα (λ = 1.5406

°A) radiation. The X-Ray Diffraction (XRD) profiles were measured in the 2Ɵ range

of 20 to 80 degree at room temperature in martensite phase. It is observed during

XRD analysis that the maximum required peaks are detected between the diffraction

angle of 30 and 60 degree, therefore XRD profiles are shown in the same range.

Fig. 4.10 shows the XRD profiles of solution treated TiNiPdCu alloys with varying

Cu percentage. All peaks of XRD profiles indicated only the B19 orthorhombic

martensite phase. From this observation, it can be confirmed that the alloys were

69

consisted of single B19 martensite phase at room temperature. Presence of second

phase particles of Ti2Ni(Pd) in 0Cu alloy and Ti2Ni(Pd, Cu) in the remaining three

alloys were not detected as these precipitates were seen in the SEM images shown in

Fig. 4.1.

The reason of unavailability of corresponding peaks of Ti2Ni(Pd) in 0Cu alloy and

Ti2Ni(Pd, Cu) in the remaining three alloys could be their low volume fraction i.e.

~2.6% to ~1.5% as discussed in Section 4.2.1. By adding Cu in TiNiPd alloys,

consistent shift of B19 martensite peaks was observed towards low angle of 2Ɵ. It

can be seen in Fig. 4.10 that the strongest B19 (111) martensite peak emerged at ~

43 degree given for 0Cu alloy, shifted towards the lower angle for 5Cu, 10Cu and

15Cu alloys. The decrease in diffraction angle by increasing the Cu addition may be

20 30 40 50 60

Inte

nsi

ty (

CP

S)

Angle (2Ɵ)

0Cu alloy

5Cu alloy

15Cu alloy

B1

9 (

11

1)

B1

9 (

00

2)

B1

9 (

02

0)

10Cu alloy

Fig. 4.10 XRD profiles at room temperature for the samples solution treated

0Cu, 5Cu, 10Cu and 15Cu alloys

70

expected by the increase in the lattice constants. The increase in lattice constants is

attributed to the addition of relatively large atomic radius of Cu (0.128 nm) as

compared to the atomic radius of Ni (0.125 nm).

4.5 Phase analysis of aged TiNiPdCu alloys with varying Copper

percentage

4.5.1 Phase analysis of aged 0Cu alloy

Fig. 4.11 shows the XRD profiles at room temperature for the samples solution

treated and aged at 400°C, 500°C, 600°C and 700°C of 0Cu alloys. It is seen that

four peaks (002), (020), (111) and (022) represent the B19 martensite phase while

one peak (011) shows the B2 austenite phase. Hence it is confirmed that at room

20 30 40 50 60

Inte

nsi

ty (

CP

S)

Angle (2Ɵ)

Solution Treated

Aged at 400°C

Aged at 500°C

Aged at 700°C

Aged at 600°C

B1

9 (

11

1)

B1

9 (

00

2)

B1

9 (

02

0)

B2

(0

11

)

B1

9 (

02

2)

Fig. 4.11 XRD profiles at room temperature for the samples solution treated and

aged at 400°C, 500°C, 600°C and 700°C of 0Cu alloys

71

temperature, the predominant phase is the B19 martensite phase where the retained

austenite B2 phase in low volume fraction is also present in the solution treated

sample. When the alloy was aged at 400°C, all the four martensite peaks are

available at the same intensities as were present in the solution treated sample.

However, the intensity of the peak corresponding to the B2 austenite phase was

increased. By further increasing the aging temperature to 500°C, 600°C and 700°C,

the XRD patterns show the same phases and no change was observed. These

observations confirm that aging at 400°C, 500°C, 600°C and 700°C for 3 hours,

stoichiometric TiNiPd alloys do not respond to any phase change.


Fig. 4.12 shows the XRD profiles at room temperature for the samples solution

treated and aged at 400°C, 500°C, 600°C and 700°C of 5Cu alloys. Here it can be

observed that all the four peaks (002), (020), (111) and (022) represent the B19

martensite phase and no phase other than martensite is detectable. Therefore, it is

confirmed that at room temperature, the material is transformed completely to the

B19 martensite phase. By aging the alloy at 400°C, all the four martensite peaks are

available at the same intensities as were present before aging. By further increasing

the aging temperature to 500°C, 600°C and 700°C, the XRD patterns show the same

phases and no change was observed. From these observations it can be confirmed

that aging at 400°C, 500°C, 600°C and 700°C for 3 hours, TiNiPdCu alloys with 5%

Cu do not respond to any phase change. It was observed from Fig. 4.7 (b – d) that by

aging the 5Cu alloys at 500°C, 600°C and 700°C, very fine precipitates are formed,

however the XRD profiles do not show the existence of those precipitates. The

reason for this observation is the low volume fraction of precipitates which cannot be

detected by XRD analysis.

72


Fig. 4.13 depicts the XRD profiles at room temperature for the samples solution

treated and aged at 400°C, 500°C, 600°C and 700°C of 10Cu alloys. The XRD

pattern for solution treated sample shows that all the five peaks (101), (002), (020),

(111) and (022) represent the B19 martensite phase. By aging the alloy at 400°C, the

XRD profile was slightly changed. At this aging temperature, the height of the

martensite peaks, except (111), was slightly decreased.

20 30 40 50 60

Inte

nsi

ty (

CP

S)

Angle (2Ɵ)

Solution Treated

Aged at 400°C

Aged at 700°C

Aged at 600°C

Aged at 500°C

B1

9 (

02

2)

B1

9 (

00

2)

B1

9 (

02

0)

B1

9 (

11

1)



73

On the other hand, emerging of new peaks, one at the left side (39.77o) and another at

right side (43.67o) of strongest B19 (111) peak were observed. Reduction in volume

fraction of martensite phase and emerging of new phase had confirmed that

precipitation has been started at aging temperature of 400°C. As the aging

temperature was increased to 500°C and 600°C, the XRD profiles were changed

significantly. At these aging temperatures, (002) and (022) reflexes of B19

martensite phase were completely disappeared due to remarkable decrease in volume

fraction of martensite phase and formation of two types of second phase precipitates

in the matrix were detected. It can be observed from the XRD profiles, that peaks

corresponding to new formed second phase precipitates are showing maximum

height at 600°C which is due to the presence of dense population of second phase

precipitates with higher volume fraction. From compositional analysis given in Table

20 30 40 50 60

Inte

nsi

ty (

CP

S)

Angle (2Ɵ)

Solution Treated

Aged at 400°C

Aged at 700°C

Aged at 600°C

Aged at 500°C B

19

(0

02

)

B1

9 (

02

2)

B1

9 (

10

1)

B1

9 (

11

1)

B1

9 (

02

0)

Ti 2

Pd

TiP

dC

u



74

4.2, XRD profile and reference [145], it was confirmed that the peaks located on the

left and right side of (111) reflex are Ti2Pd and TiPdCu, respectively. The

observations resulted from the XRD analysis regarding the precipitation behavior

was found to be completely in accordance with the microstructures given in Fig. 4.8

(a – c). When the aging temperature was further increased to 700°C, the XRD profile

was again significantly changed and the peaks corresponding to Ti2Pd and TiPdCu

were observed to be disappeared. Although it was observed from Fig. 4.8d that both

types of precipitates with low density of relatively coarser size were present in the

alloy aged at 700°C, however it could not be detected in the XRD profile, probably

due to low volume fraction. The XRD profiles for the alloys aged at 700°C were

found to be the same as was observed for solution treated sample, where all peaks

were representing the B19 orthorhombic martensite phase.


Fig. 4.14 presents the XRD profiles at room temperature for the samples solution

treated and aged at 400°C, 500°C, 600°C and 700°C of 15Cu alloys. The XRD

pattern for solution treated sample shows that all the four peaks (002), (020), (111)

and (022) represent the B19 martensite phase. These observations suggest that the

material is completely in martensite phase having the B19 orthorhombic structure at

room temperature.

By aging the alloy at 400°C, the XRD profile was significantly changed. At this

aging temperature, one martensite peak with reflex (022) was completely

disappeared. On the other hand, two new peaks, one at the left side (39.77o) and

another at right side (43.67o) of strongest B19 (111) peak were immerged. The

absence of one martensite peak and emergence of two new peaks had suggested that

volume fraction of martensite phase was decreased and precipitation was started at

aging temperature of 400°C. When the aging temperature was increased to 500°C,

remarkable change in the XRD profile was observed.

75



20 30 40 50 60

Inte

nsi

ty (

CP

S)

Angle (2Ɵ)

Solution Treated

Aged at 400°C

Aged at 700°C

Aged at 600°C

Aged at 500°C

B1

9 (

11

1)

B1

9 (

02

2)

B1

9 (

00

2)

Ti 2

Pd

TiP

dC

Uu

B1

9 (

02

0)

B1

9 (

10

1)

All the peaks corresponding to martensite phase were totally absent, however the

intensities of the new emerged peaks were observed to be increased. The XRD

pattern obtained at aging temperature of 600°C shows similar profile as was obtained

for 500°C-aged sample with slight increase in intensities. It can be observed from the

XRD profiles, that peaks corresponding to new formed second phase precipitates are

showing maximum height at 600°C which is due to the presence of dense population

of second phase precipitates with higher volume fraction. From compositional

analysis given in Table 4.2 and XRD profile, it can be confirmed that the peaks

located on the left and right side of (111) reflex are Ti2Pd and TiPdCu respectively.

The observation obtained from the XRD analysis regarding the precipitation

behavior was found to be completely in accordance with the microstructures given in

Fig. 4.9 (a – c). When the aging temperature was further increased to 700°C, the

76

XRD profile was again significantly changed. Here the four peaks corresponding to

martensite phase were again observed, however at the same time the two peaks

corresponding to Ti2Pd and TiPdCu were also present. The XRD analysis confirms

the presence of both types of precipitates as was observed from Fig. 4.9d. Here it is

also confirmed that the volume fraction of both precipitates is significantly high and

can be detected through XRD patterns.

By comparing the XRD analysis for 10Cu and 15Cu alloys, it can be concluded that

the size and density of precipitates produced in 15Cu alloys were greater as

compared to those of 10Cu alloys at the corresponding aging temperatures. This

conclusion fully supports by the backscattered SEM images shown in Fig. 4.8 (a – d)

for 10Cu alloy and Fig. 4.9 (a – d) for 15Cu alloy. Moreover the XRD profile

obtained for 700°C-aged sample of 15Cu alloy is significantly different from the

700°C-aged sample of 10Cu alloy. In 10Cu alloy, the XRD pattern shows that all the

peaks available are B19 martensite whereas the XRD pattern of obtained for 15Cu

alloy shows the martensite peaks as well as peaks for Ti2Pd and TiPdCu.

4.6 Summary

In this chapter, the effect on the microstructural properties due to varying percentage

of Cu addition with TiNiPd alloys in solution treated condition was presented.

Moreover, the effect on the microstructure due to aging at different aging

temperatures of TiNiPdCu with varying percentage of Cu was also discussed. It was

observed that the density of second phase precipitates decreased and grain size

increased when Cu was increased in TiNiPd alloys. The grain size of 15Cu alloy was

observed to be increased by 35% as compared to 0Cu alloy. By aging the 0Cu alloys,

no change in its microstructure was noticed. In 5Cu alloys, very fine precipitates

along the grain boundaries were formed at higher aging temperatures. Both 10Cu and

15Cu alloys responded significantly to precipitation process at all aging

temperatures. At low aging temperatures, precipitates of greater size were formed

along the grain boundaries in both alloys. When the aging temperature increased, the

size of the precipitates also increased. At the same time, formation of precipitates

increased not only on the grain boundaries but also spread into the grain interior.

77

However, the size of the precipitates formed in 10Cu alloys was relatively smaller

than those of 15Cu alloys.

During phase analysis it was observed that in solution treated condition, only B19

martensite phase was present in all four alloys. However the peaks representing the

martensite phase shifted consistently towards lower 2Ɵ angle by increasing the Cu in

TiNiPd alloys. By aging 0Cu and 5Cu alloys at different aging temperatures, no

phase change was observed and the phase was remained as martensite phase.

Remarkable change was noticed when 10Cu and 15Cu alloys were aged at different

aging temperatures. At low and intermediate aging temperatures, two types of second

phase precipitates (Ti2Pd, TiPdCu) were also detected, resultantly the volume of

martensite phase was reduced. However, at higher aging temperature, the formation

of those precipitates was again disappeared and only martensite phase was present.

78

Chapter – 5


on Transformation Temperatures of

TiNiPd Alloys

5.1 Introduction

This chapter presents the effect of varying percentage of copper (0Cu, 5Cu, 10Cu

and 15Cu) addition on phase transformation temperatures in TiNiPd alloys.

Moreover, it also includes the changes in transformation temperatures of TiNiPdCu

alloys occurred due to aging at different aging temperatures. These transformation

temperatures were measured by using Differential Scanning Calorimetry (DSC).

5.2 Effect of Cu addition on phase transformation temperatures

Fig. 5.1 shows the DSC heating and cooling cycles of 0Cu, 5Cu, 10Cu and 15Cu

alloys solution treated at 900°C for 1 hour. The phase transformation temperatures

measured from DSC cycles of Fig. 5.1 were re-plotted in Fig.5.2 to compare the

change in transformation temperatures for replacing of Ni with 5%, 10% and 15%

Cu with respect to the baseline 0Cu alloy. All alloys exhibited very similar behavior

with single-stage martensite transformation and demonstrated proper developed

peaks with well calculated transformation temperatures. However the transformation

heat (ΔHc) released (area under cooling curve) during forward transformation cycle

and transformation heat (ΔHh) absorbed (area under heating curve) during reverse

transformation cycle increased as the Cu contents increased in the alloy.

The martensite start temperature; Ms of 0Cu alloy increased by 12.5°C from 142.5°C

to 155°C, while the austenite finish temperature; Af increased by 11°C from 167°C to

178°C upon addition of 5% Cu in place of Ni. Moreover, it can also be noted that,

the thermal hysteresis decreased by 1.5°C from 24.5°C to 23°C. From the above

experimental results it has been observed that substitution of Ni by 5% Cu increased

the transformation temperatures significantly while thermal hysteresis decreased

slightly.

79

Similarly the transformation temperatures of 10Cu alloy further increased and it

became 27.5°C higher than the baseline 0Cu alloy. The Ms temperature of 0Cu alloy

increased by 27.5°C from 142.5°C to 170°C, while the Af temperature increased by

25°C from 167°C to 192°C. Moreover, it was also observed, that thermal hysteresis

decreased by 2.5°C from 24.5°C to 22°C. From these experimental results it was

concluded that substitution of Ni by 10% Cu, increased the transformation

temperatures significantly, while thermal hysteresis decreased slightly.

By replacing Ni with 15% Cu in 0Cu alloy, the transformation temperatures of 15Cu

alloy further increased. The Ms temperature of 0Cu alloy increased by 60.5°C from

142.5°C to 203°C, while the Af temperature increased by 52°C from 167°C to

219°C. It can also be noted that thermal hysteresis decreased by 8.5°C from 24.5°C

to 16°C. From the above experimental results it has been concluded that substitution

of Ni by 15% Cu, the transformation temperatures increased and thermal hysteresis

decreased significantly.

50 100 150 200 250

Hea

t Fl

ow

(w

/g)

exo

up

Temperature (°C)

0.5

0Cu alloy

10Cu alloy

5Cu alloy

15Cu alloy

Cooling

Heating

Fig. 5.1 DSC heating and cooling cycles of solution treated 0Cu, 5Cu, 10Cu and

15Cu alloys

80

The change in transformation temperatures and transformation heats are actually due

to the change in Ni/Pd content, because Cu content does not affect the transformation

temperatures according to Mercier et al. [146]. Suburi, T. [147] reported that

according to the TiNiTiPd pseudo-binary phase diagram, the increase in Pd-content

and decrease in Ni-content, increases the transformation temperatures. Thus the

decrease in Ni/Pd ratio (as Ni content decreased and Pd content remained constant)

resulted an increase in transformation temperatures and transformation heats

absorbed and released during forward and reverse martensitic transformation,

respectively. Moreover, according to Clausius-Clapeyron equation, the increase in

transformation temperature causes an increases in the transformation heats [3].

5.3 Effect of aging on phase transformation temperatures

To investigate the effect of aging on phase transformation temperatures, the

transformation temperatures of four alloys aged at different aging temperatures were

measured. The detail discussion is given as under.

100

135

170

205

240

Tran

sfo

rmat

ion

Tem

per

atu

re (

°C)

Cu (at%)

Ms

Mf

As

Af

0 5 10 15

Fig. 5.2 Effect of Cu addition on transformation temperatures

81

5.3.1 Effect of aging on phase transformation temperatures of 0Cu alloy

Fig. 5.3 represents the DSC heating and cooling cycles of the samples solution

treated, aged at 400°C, 500°C, 600°C and 700°C of 0Cu alloy. Fig. 5.4 shows the

variation of transformation temperatures with respect to the increasing aging

temperatures. The DSC heating and cooling cycles for the samples solution treated,

aged at 400°C, 500°C, 600°C and 700°C were showing properly developed peaks.

However the sharpness of the peaks increased as the aging temperature increased

from 400°C to 700°C. Thus the transformation heats released and absorbed during

forward and reverse transformation cycles increased as the aging temperatures

increased. The Ms of the 400°C-aged alloy increased by 8°C from 142°C to 150°C,

whereas Af increased by 9°C from 168°C to 177°C, when the aging temperature

increased from 400°C to 700°C. Moreover, it was also observed that thermal

hysteresis of the 0Cu alloy slightly increased when the aging temperatures increased.

Thermal hysteresis of the 400°C-aged alloy increased by 1°C from 26°C to 27°C,

50 100 150 200 250

Hea

t Fl

ow

(w

/g)

exo

up

Temperature (°C)

Cooling

Heating Solution Treated

Aged at 400°C

Aged at 700°C

Aged at 600°C

Aged at 500°C

0.5

Fig. 5.3 DSC heating and cooling cycles of the samples solution treated, aged at

400°C, 500°C, 600°C and 700°C of 0Cu alloy

82

when the aging temperature increased to 700°C. The increase in transformation

temperatures of the 0Cu alloy due to increase in aging temperatures can only be

attributed to the grain growth [148], as there was no precipitation detected in aging of

0Cu alloy as shown in Fig. 4.6.


Fig. 5.5 shows the DSC heating and cooling cycles of the samples solution treated,

aged at 400°C, 500°C, 600°C and 700°C of 5Cu alloy. The transformation

temperatures evaluated from DSC cycles of Fig. 5.5 were re-plotted in Fig. 5.6 which

shows the variation of transformation temperatures with respect to the increasing

aging temperatures. It can be observed from Fig. 5.6 that the transformation

temperatures increased when the aging temperatures increased from 400°C to 600°C.

However, further increase in aging temperature i.e. at 700°C, the transformation

temperatures decreased. The Ms of the 400°C-aged alloy increased by 11°C from

163°C to 174°C, when the aging temperature increased to 600°C. Similarly Af

increased by 8°C from 186°C to 194°C. At aging temperature of 700°C, the Ms

decreased by 7°C from 174°C to 167°C while the Af decreased by 2°C from 194°C

100

135

170

205

240

Tran

sfo

rmat

ion

Tem

per

atu

re (

°C)

Aging Temperature (°C)

Ms

Mf

As

Af

ST 400 500 600 700

Fig. 5.4 Effect of aging temperatures on transformation temperatures of 0Cu alloy

83

to 192°C. Moreover, it was also observed that thermal hysteresis of the 5Cu alloy

slightly increased when the aging temperatures increased. Thermal hysteresis of the

400°C-aged alloy increased by 2°C from 23°C to 25°C, when the aging temperature

increased to 700°C.

The increase in transformation temperatures of the 5Cu alloy due to increase in

aging temperatures till 600°C is the result of grain growth [148], because there was

no significant precipitation observed in aging of 5Cu alloy as shown in Fig. 4.7 (a –

c). However, by further increasing the aging temperature to 700°C, the precipitation

process started and low volume fraction of brighter precipitates along the grain

boundaries and grain interiors detected as shown in Fig. 4.7 (d). It has been discussed

in chapter 4 that the brighter precipitates were Ni-lean in nature. Due to the

formation of these precipitates, the Ni-content in the matrix increased. By increasing

the Ni-content, the transformation temperatures decreased [149].

50 100 150 200 250

Hea

t Fl

ow

(w

/g)

exo

up

Temperature (°C)

Cooling


Aged at 400°C

Aged at 700°C

Aged at 600°C

Aged at 500°C

0.5



84


Fig. 5.7 represents the DSC heating and cooling cycles of the samples solution

treated, aged at 400°C, 500°C, 600°C and 700°C of 10Cu alloy. Fig. 5.8 shows the

variation of transformation temperatures with respect to the increasing aging

temperatures. The DSC heating and cooling cycles in Fig. 5.7 for the samples

solution treated and aged at 400°C, were showing proper developed peaks with well

calculated transformation temperatures. However, the transformation peak for

solution treated sample was sharper as compared to the transformation peak of

400°C-aged sample. When the aging temperature increased to 500°C and 600°C, the

transformation peaks broadened and their height become lower and thus difficult to

calculate martensite finish temperature; Mf and austenite start temperature; As.

Therefore for comparison purpose, only Ms and Af were calculated and drawn in Fig.

5.8.

100

135

170

205

240

Tran

sfo

rmat

ion

Tem

per

atu

re (

°C)


Ms

Mf

As

Af

ST 400 500 600 700


85

By increasing the aging temperatures from 400°C to 600°C, the transformation

temperatures decreased remarkably. The Ms of the 400°C-aged sample dropped by

16°C from 170°C to 154°C when the aging temperature increased to 500°C.

Similarly the Af of the 400°C-aged sample dropped by 16°C from 192°C to 176°C

when the aging temperature increased to 500°C. By further increasing the aging

temperature to 600°C, the Ms decreased by 16°C from 154°C to 138°C and the Af

decreased by 10°C from 176°C to 166°C. At the same time thermal hysteresis

increased by 6°C from 22°C to 28°C. When the aging temperature further increased

to 700°C, again the transformation peaks of the DSC cycles resulted in well-

developed profile with proper height. Moreover, by increasing the aging temperature,

the transformation temperatures also increased remarkably as shown in Fig. 5.8. The

Ms of the 600°C-aged sample increased by 45°C from 138°C to 183°C when the

aging temperature increased to 700°C. Similarly the Af increased by 36.5°C from

166°C to 202.5°C.

50 100 150 200 250

Hea

t Fl

ow

(w

/g)

exo

up

Temperature (°C)

Cooling

Heating

Solution Treated

Aged at 400°C

Aged at 700°C

Aged at 600°C

Aged at 500°C

0.5



86

The decrease in transformation temperatures of the 10Cu alloy due to increase in

aging temperatures to 500°C and 600°C can be attributed to the formation of Ni-lean

precipitates. In Fig. 4.8, it was shown that, at aging temperatures of 500°C and

600°C, two types of precipitates (Ti2Pd and TiPdCu) were formed in the alloy. By

the compositional analysis of precipitates, it was confirmed that both types of

precipitates were Ni-lean. Due to formation of these precipitates, the Ni

concentration in the matrix increased and thus stoichiometric TiNiPdCu alloy

transformed into off- stoichiometric composition; Ti : (NiPdCu) ≠ 50 : 50. According

to Fuentes et al. [150], in the Ni-rich NiTi shape memory alloy, a change of 0.1% Ni,

varies the transformation temperature by 10°C. The remarkable increase in

transformation temperatures at aging temperature of 700°C can be attributed to the

decrease in precipitation process and grain growth [148]. Due to lower density of Ni-

lean precipitates, the Ti concentration in the matrix was again increased and resulted

in higher transformation temperatures.

100

135

170

205

240

Tran

sfo

rmat

ion

Tem

per

atu

re (

°C)


Ms

Mf

As

Af

ST 400 500 600 700


87


Fig. 5.9 shows the DSC heating and cooling cycles of the samples solution treated,

aged at 400°C, 500°C, 600°C and 700°C of 15Cu alloy. The transformation

temperatures calculated from DSC cycles of Fig. 5.9 was plotted in Fig. 5.10 which

shows the change in transformation temperatures due to change in aging

temperatures.

The DSC heating and cooling cycles in Fig. 5.9 for the samples solution treated and

aged at 400°C, were showing proper developed peaks with well calculated

transformation temperatures. However the transformation peak for solution treated

sample is sharper as compared to the transformation peak of 400°C-aged sample.

When the aging temperature increased to 500°C and 600°C, the transformation peaks

broadened and their height become lower and thus difficult to calculate Mf and As.

By increasing the aging temperatures from 400°C to 500°C and 600°C, the

transformation temperatures decreased significantly. The Ms of the 400°C-aged

sample decreased by 16°C from 184°C to 168°C when the aging temperature

increased to 500°C. Similarly the Af of the 400°C-aged sample dropped by 15°C

50 100 150 200 250

Hea

t Fl

ow

(w

/g)

exo

up

Temperature (°C)

Cooling


Aged at 400°C

Aged at 700°C

Aged at 600°C

Aged at 500°C

0.5



88

from 206°C to 191°C. By further increasing the aging temperature to 600°C, the Ms

decreased by 3°C from 168°C to 165°C and the Af decreased by 2°C from 191°C to

189°C. At the same time thermal hysteresis increased by 2°C from 22°C to 24°C.

When the aging temperature further increased to 700°C, again the transformation

peaks of the DSC cycles resulted in well-developed profile. Moreover, by increasing

the aging temperature, the transformation temperatures also increased significantly as

shown in Fig. 5.10. The Ms of the 600°C-aged sample increased by 18°C from 165°C

to 183°C when the aging temperature increased to 700°C, whereas the Af increased

by 15°C from 189°C to 204°C.

The decrease in transformation temperatures of the 15Cu alloy due to increase in

aging temperatures to 500°C and 600°C can be attributed to the formation of Ni-lean

precipitates, as discussed in section 5.3. The significant increase in transformation

temperatures at aging temperature of 700°C can be attributed to the decrease in

precipitation process and grain growth [148].

100

135

170

205

240

Tran

sfo

rmat

ion

Te

mp

erat

ure

(°C

)


Ms

Mf

As

Af

ST 400 500 600 700


89

5.4 Effect of thermal cycling on phase transformation temperatures

To investigate the effect of thermal cycling on phase transformation temperatures, all

the four; 0Cu, 5Cu, 10Cu and 15Cu alloys were thermally cycled for complete five

times. The detailed discussion is given as under.


alloy

Fig. 5.11 represents the DSC heating and cooling curves for the evolution of

transformation temperatures during five thermal cycles for solution treated sample of

0Cu alloy. The calculated transformation temperatures were plotted in Fig. 5.12 to

compare the variation in transformation temperatures due to increasing thermal

cycles. Here all the DSC curves showed sharp peaks with well estimated phase

transformation temperatures. It can be observed from Fig. 5.12 that the

transformation temperatures dropped quickly in the second cycle and then decreased

slightly in the next cycles and continue to be decreased till fifth cycle. The Ms;

142.5°C in the first cycle dropped quickly to 138.5°C in the second cycle, resulting a

net decrease of 4°C. Similarly the Af; 167°C in the first cycle dropped quickly to

159°C in the second cycle, resulting a net decrease of 8°C. Thermal hysteresis in the

second cycle decreased by 4°C from 24.5°C to 20.5°C. The decrease in

transformation temperatures and thermal hysteresis in third thermal cycle reduced as

compared to the second thermal cycle.

90

The Ms decreased by 2°C from 138.5°C to 136.5°C and Af decreased by 3°C from

159°C to 156°C in the third cycle. Similarly thermal hysteresis in the third cycle

decreased by only 1°C from 20.5°C to 19.5°C. In the fourth thermal cycle, the

decrease in transformation temperatures and thermal hysteresis further lowered as

compared to the third thermal cycle. The Ms decreased by 1°C from 136.5°C to

135.5°C and Af decreased by 2°C from 156°C to 154°C in the fourth cycle.

Similarly thermal hysteresis in the fourth cycle decreased by 1°C from 19.5°C to

18.5°C. The decrease in transformation temperatures and thermal hysteresis in fifth

thermal cycle further lowered as compared to the fourth thermal cycle. The Ms

decreased by 0.5°C from 135.5°C to 135°C and Af decreased by 1°C from 154°C to

153°C in the fifth cycle. Similarly thermal hysteresis in the fifth cycle decreased by

0.5°C from 18.5°C to 18°C. The net decrease in Ms and Af was observed to be 7.5°C

and 14°C respectively during five thermal cycles. Similarly the net decrease in

thermal hysteresis was observed to be 6.5°C.

100 150 200 250

Hea

t Fl

ow

(W

/g)

exo

up

Temperature (oC)

5th Cycle

1st Cycle

5th Cycle

Cooling

Heating

0.3

Fig. 5.11 DSC curves representing the transformation temperatures during five

thermal cycles of 0Cu alloy

91

It was noticed that, by increasing the number of thermal cycles, the transformation

temperatures of solution treated sample of 0Cu alloy decreased quickly in the second

cycle and then slow down till last thermal cycle. The faster decrease in

transformation temperatures in solution treated sample is due to the generation of

dislocations and other defects during repeated motion of the austenite-martensite

interface. This decrease in transformation temperature is unavoidable for quenched

and annealed TiNi-based alloys, however, this can be prevented by aging and

thermo-mechanical treatment [147]. It was also observed that the first heating cycle

of solution treated sample resulted in higher transformation temperatures and then

decreased faster in the second cycle. The largest drop in austenite transformation

temperatures after first heating cycle is attributed to the increase in the formation of

dislocations [151]. The dislocation densities have been reported to be increased

remarkably during the first thermal cycle and then its rate of formation decreases as

the number of cycles increases [152]. Therefore due to formation of increased

dislocation densities in the first cycle, the austenite transformation temperatures

dropped significantly in the second cycle. This effect takes place in the alloys having

low yield strength as observed in the solution treated sample. The decrease in

thermal hysteresis after the first thermal cycle is, due the stabilization effect which

decreased the As quickly and Ms slowly and resultantly their difference (thermal

hysteresis) become small.

110

140

170

200

230

Tran

sfo

rmat

ion

Tem

per

atu

res

(oC

)

Cycle Number

Ms

Mf

As

Af

1 2 3 4 5

Fig. 5.12 Effect of thermal cycles on transformation temperatures of 0Cu alloy

92


alloy




compare the change in transformation temperatures due to increasing thermal cycles.

Here all the DSC curves show sharp peaks with well estimated phase transformation

temperatures. It can be observed from Fig. 5.14 that the transformation temperatures

dropped quickly in the second cycle and then decreased slightly in the next cycles

and continue to be decreased till fifth cycle.

The Ms; 155°C in the first cycle dropped quickly to 153°C in the second cycle,

resulted a net decrease of 2°C. Similarly the Af; 178°C in the first cycle dropped

quickly to 173°C in the second cycle, resulted a net decrease of 5°C. Thermal

hysteresis in the second cycle decreased by 3°C from 23°C to 20°C. The decrease in

Af and thermal hysteresis in third thermal cycle reduced as compared to the second

thermal cycle. The Ms decreased by 2°C from 153°C to 151°C and Af decreased by

3°C from 173°C to 170°C in the third cycle. Similarly thermal hysteresis in the third

cycle decreased by only 1°C from 20°C to 19°C. In the fourth thermal cycle, the

decrease in transformation temperatures and thermal hysteresis further lowered as

compared to the third thermal cycle. The Ms decreased by 0.5°C from 151°C to

150.5°C and Af decreased by 1°C from 170°C to 169°C in the fourth cycle.

Similarly thermal hysteresis in the fourth cycle decreased by 0.5°C from 19°C to

18.5°C. The decrease in transformation temperatures and thermal hysteresis in fifth

thermal cycle further lowered as compared to the fourth thermal cycle.

93

100 150 200 250

Hea

t Fl

ow

(W

/g)

exo

up

Temperature (oC)

0.3

5th Cycle

5th Cycle

1st Cycle

Cooling

Heating



110

140

170

200

230

Tran

sfo

rmat

ion

Tem

per

atu

res

(oC

)

Cycle Number

Ms

Mf

As

Af

1 2 3 4 5

Fig. 5.14 Effect of thermal cycle on transformation temperatures of 5Cu alloy

94

The Ms decreased by 0.5°C from 150.5°C to 150°C and Af decreased by 1°C from

169°C to 168°C in the fifth cycle. Similarly thermal hysteresis in the fifth cycle

decreased by 0.5°C from 18.5°C to 18°C. During five thermal cycles, the Ms and Af

dropped by 5°C and 10°C respectively, whereas thermal hysteresis decreased by 5°C.

The reasons for decreasing the transformation temperatures and thermal hysteresis by

increasing thermal cycle in 5Cu alloy are given in section 5.4.1.


alloy

Fig. 5.15 shows the DSC heating and cooling curves for the evolution of


10Cu alloy. The calculated transformation temperatures were drawn in Fig. 5.16. It

can be observed that all the DSC curves had sharp peaks with well estimated phase

transformation temperatures. From Fig. 5.16, it can be viewed that the transformation

temperatures dropped quickly in the second cycle and then decreased slightly in the

next cycles and become almost stable in the fifth cycle.

The Ms; 170°C in the first cycle dropped quickly to 167.2°C in the second cycle,

resulted a net decrease of 2.8°C. Similarly the Af; 192.2°C in the first cycle dropped

quickly to 186.7°C in the second cycle, resulted a net decrease of 5.5°C. Thermal

hysteresis in the second cycle decreased by 2.7°C from 22.2°C to 19.5°C. The

decrease in transformation temperature and thermal hysteresis in third thermal cycle

reduced as compared to the second thermal cycle. The Ms decreased by 0.9°C from

167.2°C to 166.3°C and Af decreased by 1.5°C from 186.7°C to 185.2°C in the third

cycle. Similarly thermal hysteresis in the third cycle decreased by 0.6°C from

19.5°C to 18.9°C. In the fourth thermal cycle, the decrease in transformation

temperatures and thermal hysteresis further lowered as compared to the third thermal

cycle. The Ms remained stable at 166.3°C while the Af decreased by 0.9°C from

185.2°C to 184.3°C in the fourth cycle. Similarly thermal hysteresis in the fourth

cycle decreased by 0.9°C from 18.9°C to 18°C.

95

The decrease in Af and thermal hysteresis in fifth thermal cycle further lowered as

compared to the fourth thermal cycle. The Ms remained stable at 166.3°C while the

100 150 200 250

Hea

t Fl

ow

(W

/g)

exo

up

Temperature (oC)

0.3

5th Cycle

5th Cycle

1st Cycle

Cooling

Heating



110

140

170

200

230

Tran

sfo

rmat

ion

Tem

per

atu

res

(oC

)

Cycle Number

Ms

Mf

As

Af

1 2 3 4 5


96

Af decreased by 0.7°C from 184.3°C to 183.6°C in the fifth cycle. Similarly thermal

hysteresis in the fifth cycle decreased by 0.7°C from 18°C to 17.3°C. After

completing thermal cycles for five times, the Ms and Af lowered by 3.7°C and 8.7°C

respectively. At the same time, thermal hysteresis decreased by 4.9°C.

It was observed that the martensite start temperatures of solution treated sample of

10Cu alloy decreased quickly in the second cycle and then reduced in third cycle and

remained stable till last thermal cycle. The faster decrease in transformation

temperatures in solution treated sample was due to the generation of dislocations and

other defects during repeated motion of the austenite-martensite interface. This

decrease in transformation temperature is unavoidable for quenched and annealed

TiNi-based alloys, however, this can be prevented by aging and thermo-mechanical

treatment [147]. Rehman et al. [153] reported that by aging the same alloy at 600°C

for 3 hours, the decrease in Ms and Af temperatures is only 2°C, while thermal

hysteresis remained stable at 28°C after five thermal cycles. It was also observed that

the first heating cycle of solution treated sample resulted in higher transformation

temperatures and then decreased faster in the second cycle. The largest drop in

austenite transformation temperatures after first heating cycle was attributed to the

increase in formation of dislocations [151].


alloy




compare the variation in transformation temperatures due to increasing thermal

cycles. All the DSC curves presented sharp peaks with well estimated phase

transformation temperatures. It can be observed from Fig. 5.18 that the

transformation temperatures dropped quickly in the second cycle and then decreased

slightly in the third cycle and become stable in the fourth cycle.

97

100 150 200 250

Hea

t Fl

ow

(W

/g)

exo

up

Temperature (oC)

0.3

5th Cycle

5th Cycle

1st Cycle

Cooling

Heating



110

140

170

200

230

Tran

sfo

rmat

ion

Tem

per

atu

res

(oC

)

Cycle Number

Ms

Mf

As

Af

1 2 3 4 5


98

The Ms dropped by 2°C from 203°C to 201°C, whereas the Af dropped by 4°C from

219°C to 215°C in the second cycle. Thermal hysteresis in the second cycle

decreased by 2°C from 16°C to 14°C. However the decrease in transformation

temperature and thermal hysteresis in third thermal cycle was lower as compared to

the second thermal cycle. The Ms decreased by 1°C from 201°C to 200°C and Af

decreased by 1°C from 215°C to 214°C in third cycle. Thermal hysteresis in third

cycle remained stable at 14°C. In fourth and fifth thermal cycles, the Ms and Af

remained stable at 200°C and 214°C respectively, while thermal hysteresis also

remained unchanged at 14°C. The faster decrease in transformation temperatures and

thermal hysteresis during second thermal cycle has been explained in section 5.4.1.

During five thermal cycles, the Ms and Af decreased by 3°C and 5°C, whereas

thermal hysteresis dropped by 2°C.

5.5 Summary

In this chapter the effect of 5%, 10% and 15% Cu addition in TiNiPd and aging at

400°C, 500°C, 600°C and 700°C for 3 hours of TiNiPdCu alloys on transformation

temperatures under stress free condition were presented. Transformation

temperatures of all four alloys in both conditions; solution treated and aged were

found by differential scanning calorimetry.

Increasing the Cu concentration in TiNiPd alloys, the transformation temperatures

and transformation heats increased significantly whereas thermal hysteresis

decreased in solution treated condition. Adding the Cu content by 5 ~ 15%, the

martensite start temperatures increased by 9 ~ 42% and thermal hysteresis decreased

by 6 ~ 35%. This behavior confirmed that addition of Cu in place of Ni in TiNiPd

alloys was very beneficial to improve its transformation temperatures and decrease

its thermal hysteresis.

When 0Cu alloy was aged, the transformation temperatures slightly increased as the

aging temperature increased from 400°C to 700°C. The martensite start temperature

was increased by 1% at aging temperature of 600°C as compared to solution treated

condition. By aging the 5Cu alloy, the transformation temperatures increased slightly

when the aging temperature was increased from 400°C to 600°C and then slightly

99

decreased at aging temperature of 700°C. At aging temperature of 600°C, the

martensite start temperature was increased by 12%. For 10Cu and 15Cu alloys, the

transformation temperatures decreased significantly as the aging temperature

increased from 400°C to 600°C and then increased comparable to that of solution

treated condition when the aging temperature was increased to 700°C. The

martensite start temperatures of 10Cu and 15Cu alloys were decreased by 18% and

19% respectively at aging temperature of 600°C. From these observations it can be

concluded that aging of 0Cu and 5Cu alloys was beneficial to improve their

transformation temperatures whereas in case of 10Cu and 15Cu alloys, the

transformation temperatures decreased remarkably.

In solution treated 0Cu alloy, after completion of five thermal cycles, the Ms and Af

decreased by 7.5°C and 14°C respectively, whereas thermal hysteresis decreased by

6.5°C. In case of 5Cu alloy, the Ms, Af and thermal hysteresis dropped by 5°C,

10°C and 5°C respectively. By thermal cycling the 10Cu alloy, the Ms and Af

lowered by 3.7°C and 8.7°C respectively, whereas thermal hysteresis lowered by

4.9°C. Similarly, for 15Cu alloy the Ms, Af and thermal hysteresis decreased by 3°C,

5°C and 2°C respectively. Moreover, during thermal cycling, the transformation

temperatures of solution treated 0Cu alloys decreased by increasing the number of

thermal cycles, however the drop was faster in initial cycles and then reduced in the

final cycles. Similar behavior was also observed during thermal cycling of 5Cu,

10Cu and 15Cu alloys, however the drop in transformation temperatures was less as

compared to 0Cu alloy. This behavior confirmed that, by increasing the Cu content in

TiNiPd alloys, thermal stability improved. Hence 5Cu, 10Cu and 15Cu alloys were

more thermally stable as compared to 0Cu alloy.

100

Chapter – 6


on Mechanical Properties of TiNiPd

Alloys

6.1 Introduction

In this chapter, the mechanical properties i.e. Stress-strain relation, yield strength,

ultimate tensile strength (fracture strength) and hardness of 0Cu, 5Cu, 10Cu and

15Cu alloys for different heat treated conditions (solution treated and aged) have

been shown. It has been investigated that the mechanical properties can be improved

by applying two techniques; solid solution strengthening by 5%, 10% and 15% Cu

addition, and precipitation strengthening by aging the 0Cu, 5Cu, 10Cu and 15Cu

alloys at different temperatures ranging from 400°C to 700°C for 3 hours.

6.2 Effect of Cu addition on hardness

Fig. 6.1 represents the Vickers microhardness of 0Cu, 5Cu, 10Cu and 15Cu alloys

solution treated at 900°C for 1 hour. The hardness values of each alloy were

determined from the average of 5 measurements taken at different locations on each

sample. It was observed that the hardness increased by increasing the Cu content in

place of Ni. Hardness of 0Cu alloy; 245 Hv increased to 252 Hv when Ni was

replaced by 5% Cu. By increasing the Cu concentration from 5% to 10%, the

hardness further increased to 255 Hv. Similarly by further increasing Cu

concentration to 15%, the hardness further increased to 263 Hv. The behavior of

increasing hardness was identical to increase in transformation temperatures as

discussed in Chapter – 5. Thus it can be concluded that by increasing the substitution

of Ni by Cu, both the hardness and transformation temperatures were increased. The

increase in hardness by increasing the substitution of Ni by Cu was attributed to the

solid solution strengthening. As the atomic radius of Cu (0.128 nm) is relatively

greater than the atomic radius of Ni (0.125 nm), thus caused to increase the hardness.

101

6.3 Effect of aging on hardness

6.3.1 Effect of aging on hardness of 0Cu alloy

Fig. 6.2 represents the hardness of 0Cu alloy; solution treated and aged at given

temperatures for 3 hours. Hardness of solution treated sample; 245 Hv slightly

decreased to 236 Hv when the same alloy was aged at 400°C. The slight decrease in

hardness was attributed to the removal of thermal stresses which were produced

during quenching process due to higher temperature gradient. When the alloy was

aged at 500°C, the hardness of the alloy increased to 258 Hv, suggesting that

precipitation process has been initiated at this temperature. However due to very fine

size of precipitates, it could not be detected in backscattered SEM images given in

Fig. 4.6(a – d). By increasing the aging temperatures to 600°C, the hardness of the

alloy increased slightly to 263 Hv. This slight increase in hardness can be attributed

to the slight increase in volume fraction of the precipitates, however the size

remained the same and could not be detected. By increasing the aging temperature

further to 700°C, again the hardness of alloy reduced to 249 Hv. The decrease in

hardness showed that precipitation was not initiated at higher aging temperature.

200

220

240

260

280

300

Mic

roh

ard

nes

s (H

v)

Cu (at%)

0 5 10 15

Fig. 6.1 Microhardness of solution treated samples of 0Cu, 5Cu, 10Cu and 15Cu

alloys

102


Fig. 6.3 represents the hardness of 5Cu alloy; solution treated and aged at given


decreased to 242 Hv when the same alloy was aged at 400°C. The slight decrease in

hardness is attributed to the removal of thermal stresses. Moreover it can also be

confirmed that the formation of precipitates was not started at the said temperature

150

250

350

Har

dn

ess

(Hv)


ST 400 500 600 700

Fig. 6.2 Microhardness of 0Cu alloy aged at different temperatures

150

250

350

Har

dn

ess

(H

v)


ST 400 500 600 700


103

and thus not detected in Fig. 4.7a. When the alloy was aged at 500°C, the hardness of

the alloy increased to 280 Hv, suggesting that precipitation process has been

initiated. It can be observed from backscattered SEM images of Fig. 4.7b, that

precipitates of brighter contrast have been formed along the grain boundaries caused

to increase the hardness. By further increasing the aging temperatures to 600°C, the

hardness of the alloy decreased slightly to 276Hv. By increasing the aging

temperature further to 700°C, again the hardness of alloy reduced to 246 Hv. This

result shows that, although precipitates have been formed as shown in Fig. 4.7d,

however due to incoherency, the hardness decreased.


Fig. 6.4 shows the hardness of the samples solution treated and aged at given


increased to 260 Hv when the same alloy was aged at 400°C. It was observed

previously that, at aging temperature of 400°C, hardness of 0Cu and 5Cu alloys were

decreased due to removal of thermal stresses and absence of precipitation. However

in case of 10Cu alloy, the hardness slightly increased confirming the formation of

precipitates started at the said temperature. By increasing the aging temperatures to


150

250

350

450

Har

dn

ess

(H

v)


ST 400 500 600 700

104

500°C and 600°C, the hardness of the samples abruptly increased to 369 Hv and 408

Hv, respectively. The significant increase in the hardness values of the samples aged

at 500°C and 600°C were suggested to be the outcome of precipitation process. The

nucleation sites along the grain boundaries have been observed to be formed at

400°C as shown in Fig. 4.8a, resulted in slight increase in the hardness. By

increasing the aging temperatures to 500°C and 600°C, the sizes and densities of

both types of precipitates rapidly increased as shown in Fig. 4.8b and c , caused

further significant increase in the hardness. From these observations, it can be easily

derived that both type of precipitates formed at 400°C to 600°C were found to be

coherent.

However, by aging at 700°C, the hardness of the alloy abruptly decreased to 258 Hv

almost equal to the hardness of solution treated sample. Here the aging temperature

was higher than the recrystallization temperature of the TiNiPd alloys (~ 600°C)

[154], hence the precipitation hardening was not promoted. It can be observed from

Fig. 4.8d, that both types of precipitates were present in the alloy aged at 700°C,

however it could not cause to increase the hardness. The decrease in the hardness at

aging temperatures of 700°C can be attributed to the formation of incoherent

precipitates with comparatively low densities of relatively larger sizes.

From the above results, it can be summarized that at low aging temperatures (400°C

and below), no remarkable increase in hardness was observed. Aging at intermediate

temperatures (500°C and 600°C), increased the hardness of the alloy significantly.

However at high aging temperatures (700°C and above), the hardness of the alloy

was observed to be decreased significantly.

The hardness of the same alloy aged at 500°C for 1 hour, processed and heat treated

at the same conditions was reported by Imran et al. [140]. It was observed that the

hardness was not increased and remained at the same level as that of solution treated

sample. In this research, the hardness of the sample aged at 500°C for 3 hours

increased significantly up to 369 Hv. From these results, it can be confirmed, that for

the proper age hardening process, aging temperature as well as proper time duration

are important. From the same reference, it was also noticed that by increasing the

annealing temperature from 350°C to 400°C, the hardness of the alloy decreased

105

significantly. However, annealing at 450°C resulted in remarkable increase in the

hardness from 352 Hv to 544 Hv and then the hardness decreased continuously by

increasing the annealing temperature to 700°C. It is important to note that the

decrease in the hardness at annealing temperature of 400°C was attributed to the

recovery process, whereas formation of nano-scaled precipitates was responsible for

the remarkable increase in the hardness at 450°C. On the other hand, by increasing

the aging temperature from 400°C to 600°C, the hardness of the alloy increased

continuously from 260 to 408 Hv and then decreased in the same manner when the

aging temperature was further increased to 700°C, in the present study. By

comparing these results, it can be concluded that the age hardening behavior in the

present alloy was observed to be significantly different from that of annealing after

cold working.


Fig. 6.5 shows the hardness of the samples solution treated and aged at given


increased to 270 Hv when the same alloy was aged at 400°C. The increase in

hardness at aging temperature of 400°C indicated that precipitation process has been

started at the said temperature like 10Cu alloy. By increasing the aging temperatures

to 500°C and 600°C, the hardness of the samples abruptly increased to 442 Hv and

480 Hv respectively. The significant increase in the hardness values of the samples

aged at 500°C and 600°C were suggested to be the outcome of precipitation process.

The nucleation sites along the grain boundaries have been observed to be formed at

400°C as shown in Fig. 4.9a, resulted in slight increase in the hardness. By

increasing the aging temperatures to 500°C and 600°C, the sizes and densities of

both types of precipitates rapidly increased as shown in Fig. 4.9b and c , caused

further significant increase in the hardness. From these observations, it can be easily

derived that both type of precipitates formed at 400°C to 600°C were found to be

coherent.

106

However, by aging at 700°C, the hardness of the alloy abruptly decreased to 260 Hv

almost equal to the hardness of solution treated sample. Here the aging temperature

was higher than the recrystallization temperature of the TiNiPd alloys (~ 600°C)

[154], hence the precipitation hardening was not promoted. It can be observed from

Fig. 4.9d, that both types of precipitates were present in the alloy aged at 700°C,

however it could not increase the hardness of the alloy. The decrease in the hardness

values at aging temperatures of 700°C can be attributed to the formation of

incoherent precipitates with comparatively low densities of relatively larger sizes.

From the above results, it can be summarized that at low aging temperatures (400°C

and below), no remarkable increase in hardness was observed. Aging at intermediate

temperatures (500°C and 600°C), resulted significant increase in the hardness

whereas at high aging temperatures (700°C and above), the hardness of the alloy was

observed to be decreased significantly.

150

250

350

450

550

Har

dn

ess

(Hv)


ST 400 500 600 700

Fig. 6.5 Vickers microhardness of 15Cu alloy aged at different temperatures

107

6.4 Effect of Cu addition on mechanical strength

6.4.1 Effect of Cu addition on mechanical strength in martensite phase

It is important to investigate the mechanical properties of shape memory alloys in

both phases i.e. martensite and austenite, because the stress-strain relations in both

phases are different from each other. Other conventional structural materials like

steel, has maximum yield strength at room temperature and then decreases by

increasing the testing temperature above 100°C. Conversely, for high temperature

shape memory alloys, the yield strength at room temperature in martensite phase

(lower than Mf) must be less than the yield strength at higher temperature (greater

than Af) in austenite phase. Austenite yield stress represents the critical stress for

slip deformation while martensite yield stress represents critical stress for shear of

martensite twins. For feasible actuators, the critical stress for shear must be lower

than the critical stress for slip, so that when stress is applied, it results in shape

deformation by shear of martensite twins rather than via dislocation generation and

movement.

Tensile stress-strain curves of solution treated samples tested in the martensite phase,

50°C below Mf of 0Cu, 5Cu, 10Cu and 15Cu alloys are shown in Fig. 6.6. The yield

stress for all alloys was calculated by drawing a parallel line, 0.2% offset to the

elastic region of stress-strain curve as shown in Fig. 6.6. The yield stress (σy),

fracture stress (σf) and fracture strain (εf) calculated from these curves in the

martensitic condition for all alloys are shown in Table 6.1.

The martensite yield stress corresponds to the stress required for reorientation of

martensite twins (also called stress for detwinned martensite). From Table 6.1, it can

be observed that martensite yield stress of 5Cu alloy increased by 33 MPa from 290

to 323 MPa with respect to the baseline 0Cu alloy. Similarly, the stress at which the

fracture occurred in the alloy (σf) also increased by 60 MPa; 951.6 MPa of 0Cu alloy

increased to 1011.6 MPa for 5Cu alloy. However the strain at which the fracture

occurred in the alloy (εf) decreased by 0.75%; fracture strain of 10% for 0Cu alloy

decreased to 9.25% for 5Cu alloy.

108

In the same manner, when Ni was substituted by 10% Cu, the relevant mechanical

properties further increased. The martensite yield stress of 10Cu alloy increased by

56 MPa from 290 to 346 MPa, whereas the fracture stress increased by 88.4 MPa;

951.6 MPa of 0Cu alloy increased to 1040 MPa as compared to 0Cu alloy. However

the fracture strain decreased by 1.03%; fracture strain of 10% for 0Cu alloy

decreased to 8.97% for 10Cu alloy. The martensite yield stress of 15Cu alloy

increased by 76 MPa from 290 to 366 MPa with respect to the baseline 0Cu alloy.

Fig. 6.6 Tensile stress-strain curves of solution treated 0Cu, 5Cu, 10Cu and 15Cu alloys,

tested in martensite phase (Mf – 50°C)

0

300

600

900

1200

0 3 6 9 12

Stre

ss (

MP

a)

Strain (%)

Mf = 119°C

Testing Temperature = 69°C

0Cu alloy

0

300

600

900

1200

0 3 6 9 12Strain (%)

Mf = 133°C


5Cu alloy

0

300

600

900

1200

0 3 6 9 12

Stre

ss (

MP

a)

Strain (%)

Mf = 151°C


10Cu alloy

0

300

600

900

1200

0 3 6 9 12Strain (%)

Mf = 176°C


15Cu alloy

109

Similarly the fracture stress increased by 114.4 MPa; 951.6 MPa of 0Cu alloy

increased to 1066 MPa. However the fracture strain decreased by 1.28%; fracture

strain of 10% for 0Cu alloy decreased to 8.72%.

Table 6.1 Yield stress, fracture stress and fracture strain calculated from stress-strain

curves of Fig. 6.6 for solution treated alloys tested in martensite phase (Mf – 50°C)

Alloy Testing Temp (°C) Yield Stress, σy

(MPa)

Fracture Stress,

σf (MPa)

Fracture

Strain, εf (%)

0Cu 69 290 951.6 10

5Cu 83 323 1011.6 9.25

10Cu 101 346 1040 8.97

15Cu 126 366 1066 8.72

6.4.2 Effect of Cu addition on mechanical strength in austenite phase

Fig. 6.7 represents the tensile stress-strain curves of solution treated samples tested in

the austenite phase, 50°C above Af of 0Cu, 5Cu, 10Cu and 15Cu alloys.

Measurement of austenite yield stress was carried out by the same procedure as

discussed earlier in section 6.4.1. σy, σf and εf calculated from these curves in the

austenite phase for all alloys are shown in Table 6.2.

The austenite yield stress represents the critical stress for slip; stress required for the

deformation of material through dislocations generation and their movement. It can

be observed from Table 6.2 that the austenite yield stress of 440 MPa for 0Cu alloy

increased to 502 MPa for 5Cu alloy resulted net increase of 62 MPa. The fracture

stress also increased by 40.9 MPa; 1121.3 MPa for 0Cu alloy increased to 1162.2

MPa for 5Cu alloy. Here the fracture strain was observed to be decreased by 1.2%;

fracture strain of 8.5% for 0Cu alloy decreased to 7.3% for 5Cu alloy.

Similarly the austenite yield stress for 10Cu alloy increased by 102 MPa to 542 MPa

as compared to 440 MPa of the baseline 0Cu alloy. The fracture stress also increased

by 80.7 MPa; 1121.3 MPa for 0Cu alloy increased to 1202 MPa for 10Cu alloy. The

fracture strain was observed to be decreased by 1.54%; fracture strain of 8.5% for

110

0Cu alloy decreased to 6.96% for 10Cu alloy. The austenite yield stress for 15Cu

alloy further increased by 124 MPa from 440 MPa to 564 MPa. The fracture stress

also increased by 98.7 MPa; 1121.3 MPa for 0Cu alloy increased to 1220 MPa for

15Cu alloy. The fracture strain was observed to be decreased further by 1.74%;

fracture strain of 8.5% for 0Cu alloy decreased to 6.76% for 15Cu alloy.

0

300

600

900

1200

0 3 6 9

Stre

ss (

MP

a)

Strain (%)

Af = 167°C


0Cu alloy

0

300

600

900

1200

0 3 6 9Strain (%)

Af = 178°C


5Cu alloy

0

300

600

900

1200

0 3 6 9

Stre

ss (

MP

a)

Strain (%)

Af = 192°C


10Cu alloy

0

300

600

900

1200

0 3 6 9Strain (%)

Af = 219°C


15Cu alloy

Fig. 6.7 Tensile stress-strain curves of solution treated 0Cu, 5Cu, 10Cu and 15Cu alloys,

tested in austenite phase (Af + 50°C)

111


curves of Fig. 6.7 for solution treated alloys tested in austenite phase (Af + 50°C)


(MPa)

Fracture Stress,

σf (MPa)

Fracture

Strain, εf (%)

0Cu 217 440 1121.3 8.5

5Cu 228 502 1162.2 7.3

10Cu 242 542 1202 6.96

15Cu 269 564 1220 6.76

6.5 Comparison between the mechanical properties of martensite

and austenite phases in solution treated condition

Fig. 6.8 represents the effect of partial substitution of Ni by Cu in solution treated

0Cu alloy on martensite yield stress, tested at (Mf – 50°C) and austenite yield stress,

tested at (Af + 50°C).

150

300

450

600

Yie

ld S

tre

ss (

MP

a)

Cu (at%)

Martensite yield stress

Austenite yield stress

0 5 10 15


0Cu alloy on martensite yield stress, tested at (Mf – 50°C) and

austenite yield stress, tested at (Af + 50°C)

112

It can be observed that by substitution of Ni by 5%, 10% and 15% Cu, the yield

stress in the martensite phase increased by 33, 56 and 76 MPa respectively, whereas

the austenite yield stress increased by 62, 102 and 124 MPa respectively.


0Cu alloy on martensite fracture stress, tested at (Mf – 50°C) and austenite fracture

stress, tested at (Af + 50°C). It can be observed that by substitution of Ni by 5%, 10%

and 15% Cu, the fracture stress in the martensite phase increased by 60, 88.4 and

114.4 MPa respectively, whereas the austenite fracture stress increased by 40.9, 80.7

and 98.7 MPa respectively.


0Cu alloy on martensite fracture strain, tested at (Mf – 50°C) and austenite fracture

strain, tested at (Af + 50°C). It can be observed that by substitution of Ni by 5%, 10%

and 15% Cu, the fracture strain in the martensite phase decreased by 0.75%, 1.03%

and 1.28% respectively, whereas the austenite fracture strain decreased by 1.2%,

1.54% and 1.74% respectively.

900

1000

1100

1200

1300

Frac

ture

Str

ess

(MP

a)

Cu (at%)

Martensite fracture stress

Austenite fracture stress

0 5 10 15


0Cu alloy on martensite fracture stress, tested at (Mf – 50°C) and

austenite fracture stress, tested at (Af + 50°C)

113

The increase in both yield stresses (stress for reorientation of martensite and stress

for slip deformation) and fracture stress (maximum stress) is attributed to the solid

solution strengthening due to partial substitution of Ni by Cu. Here the atomic radius

of Cu (0.128 nm) is relatively greater than the atomic radius of Ni (0.125 nm) and

therefore responsible for solid solution strengthening. By solution strengthening

effect, the ductility of 0Cu alloy lowered and resulted in relatively low fracture strain

in both phases.

It can also be observed that the increase in stress for reorientation of martensite

variants; martensite yield stress (although it was not required for better shape

memory alloys) was observed to be 33, 56 and 76 MPa for 5Cu, 10Cu and 15Cu

alloys respectively. On the other hand, the critical stress for slip deformation;

austenite yield stress (essentially required for better shape memory alloys) was

increased by 62, 102 and 124 MPa for 5Cu, 10Cu and 15Cu alloys respectively. As

the increase in critical stress for slip was more beneficial than the increase in stress

for reorientation of martensite variants, thus it is suggested that by partial substitution

of Ni with 5%, 10% and 15% Cu, the critical stress for slip can be increased by 29,

46 and 48 MPa (difference in net increase in critical stress for slip and net increase in

the stress for reorientation of martensite variants).

5

7

9

11

Frac

ture

Str

ain

(%

)

Cu (at%)

Martensite fracture strain

Austenite fracture strain

0 5 10 15


0Cu alloy on martensite fracture strain, tested at (Mf – 50°C) and

austenite fracture strain, tested at (Af + 50°C)

114

6.6 Effect of aging on mechanical strength

6.6.1 Effect of aging on mechanical strength in martensite phase

Tensile stress-strain curves of 600°C-aged samples tested in the martensite phase,

50°C below Mf for 0Cu, 5Cu, 10Cu and 15Cu alloys are shown in Fig. 6.11. The

yield stress, fracture stress and fracture strain calculated from these curves in the

martensitic condition for all alloys are shown in Table 6.3. From Table 6.3, it can be

observed that martensite yield stress of 5Cu alloy increased by 32 MPa from 300 to

332 MPa with respect to the baseline 0Cu alloy. Similarly, the fracture stress also

increased by 65 MPa; 960 MPa of 0Cu alloy increased to 1025 MPa for 5Cu alloy.

However the fracture strain decreased by 0.6%; fracture strain of 9.5% for 0Cu alloy

decreased to 8.9% for 5Cu alloy. In the same manner, when Ni was substituted by

10% Cu, the relevant mechanical properties further increased. The martensite yield

stress of 10Cu alloy increased by 55 MPa from 300 to 355 MPa with respect to the

baseline 0Cu alloy. The fracture stress increased by 195 MPa; 960 MPa of 0Cu alloy

increased to 1155 MPa for 10Cu alloy. However the fracture strain decreased by

3.2%; fracture strain of 9.5% for 0Cu alloy decreased to 6.3% for 10Cu alloy. The

martensite yield stress of 15Cu alloy increased by 85 MPa from 300 to 385 MPa

whereas the fracture stress increased by 270 MPa; 960 MPa of 0Cu alloy increased to

1230 MPa for 15Cu alloy. However the fracture strain decreased by 3.3% from 9.5%

to 6.2% for 15Cu alloy.

115

0

300

600

900

1200

1500

0 2.5 5 7.5 10

Stre

ss (

MP

a)

Strain (%)

0Cu alloy

Mf = 121°C Testing Temperature = 71°C

0

300

600

900

1200

1500

0 2.5 5 7.5 10

Strain (%)

5Cu alloy

Mf = 144°C


0

300

600

900

1200

1500

0 2.5 5 7.5 10

Stre

ss (

MP

a)

Strain (%)

10Cu alloy


0

300

600

900

1200

1500

0 2.5 5 7.5 10

Strain (%)

15Cu alloy


Fig. 6.11 Tensile stress-strain curves of 600°C-aged 0Cu, 5Cu, 10Cu and 15Cu alloys,

tested in martensite phase (Mf – 50°C)

116


curves of Fig. 6.11 for 600°C-aged alloys tested in martensite phase (Mf – 50°C)


(MPa)

Fracture Stress,

σf (MPa)

Fracture

Strain, εf (%)

0Cu 71 300 960 9.5

5Cu 94 332 1025 8.9

10Cu 80 355 1155 6.3

15Cu 90 385 1230 6.2

6.6.2 Effect of aging on mechanical strength in austenite phase

Tensile stress-strain curves of 600°C-aged samples tested in the austenite phase,

50°C above Af for 0Cu, 5Cu, 10Cu and 15Cu alloys are shown in Fig. 6.12. The yield

stress, fracture stress and fracture strain calculated from these curves in the austenite

phase for all alloys are shown in Table 6.4. From Table 6.4, it can be observed that

austenite yield stress of 5Cu alloy increased by 95 MPa from 410 to 505 MPa

whereas the fracture stress increased by 100 MPa; 1070 MPa of 0Cu alloy increased

to 1170 MPa for 5Cu alloy. However the fracture strain decreased by 1.5% from

8.3% for 0Cu alloy to 6.8% for 5Cu alloy. In the same manner, when Ni was

substituted by 10% Cu, the relevant mechanical properties further increased. The

austenite yield stress of 10Cu alloy increased by 135 MPa from 410 to 545 MPa


to 1260 MPa for 10Cu alloy. However the fracture strain decreased by 1.8%; fracture

strain of 8.3% for 0Cu alloy decreased to 6.5% for 10Cu alloy.

117

When Ni was substituted by 15% Cu, the mechanical properties further increased.

The austenite yield stress of 15Cu alloy increased by 180 MPa from 410 to 590 MPa


to 1320 MPa for 15Cu alloy. On the hand, the fracture strain decreased by 3.05%

from 8.3% to 5.25% for 15Cu alloy.

0

300

600

900

1200

1500

0 2.5 5 7.5 10

Stre

ss (

MP

a)

Strain (%)

Af = 169°C Testing Temperature = 219°C

0Cu alloy

0

300

600

900

1200

1500

0 2.5 5 7.5 10Strain (%)


5Cu alloy

0

300

600

900

1200

1500

0 2.5 5 7.5 10

Stre

ss (

MP

a)

Strain (%)


10Cu alloy

0

300

600

900

1200

1500

0 2.5 5 7.5 10Strain (%)


15Cu alloy

Fig. 6.12 Tensile stress-strain curves of 600°C-aged 0Cu, 5Cu, 10Cu and 15Cu alloys, tested

in austenite phase (Af + 50°C)

118


curves of Fig. 6.12 for 600°C-aged alloys tested in austenite phase (Af + 50°C)


(MPa)

Fracture Stress,

σf (MPa)

Fracture

Strain, εf (%)

0Cu 219 410 1070 8.3

5Cu 244 505 1170 6.8

10Cu 216 545 1260 6.5

15Cu 239 590 1320 5.25

6.7 Comparison between the mechanical properties of martensite

and austenite phases in 600°C-aged condition

Fig. 6.13 represents the effect of partial substitution of Ni by Cu in 600°C-aged 0Cu

alloy on martensite yield stress, tested at (Mf – 50°C) and austenite yield stress,

tested at (Af + 50°C). It can be observed that by substitution of Ni by 5%, 10% and

15% Cu, the yield stress in the martensite phase increased by 32, 55 and 85 MPa

respectively, whereas the austenite yield stress increased by 95, 135 and 180 MPa

respectively.


alloy on martensite fracture stress, tested at (Mf – 50°C) and austenite fracture stress,


15% Cu, the fracture stress in the martensite phase increased by 65, 195 and 270

MPa respectively, whereas the austenite fracture stress increased by 100, 190 and

250 MPa respectively.

119

150

300

450

600

Yie

ld S

tres

s (M

Pa)

Cu (at%)

Martensite yield stress

Austenite yield stress

0 5 10 15


alloy on martensite yield stress, tested at (Mf – 50°C) and austenite

yield stress, tested at (Af + 50°C)

900

1000

1100

1200

1300

1400

Frac

ture

Str

ess

(MP

a)

Cu (at%)

Martensite fracture stress

Austenite fracture stress

0 5 10 15


alloy on martensite fracture stress, tested at (Mf – 50°C) and austenite

fracture stress, tested at (Af + 50°C)

120


alloy on martensite fracture strain, tested at (Mf – 50°C) and austenite fracture strain,


15% Cu, the fracture strain in the martensite phase decreased by 0.6%, 2.2% and

3.3% respectively, whereas the austenite fracture strain decreased by 1.5%, 1.8% and

3.05% respectively. The increase in both yield stresses of 600°C-aged alloys was

attributed to the solid solution strengthening due to partial substitution of Ni by Cu.

It can also be observed that the increase in stress for reorientation of martensite

variants was observed to be 32, 55 and 85 MPa for 5Cu, 10Cu and 15Cu alloys

respectively. On the other hand, the critical stress for slip was increased by 95, 135

and 180 MPa for 5Cu, 10Cu and 15Cu alloys respectively. As the increase in critical

stress for slip is more beneficial than the increase in stress for reorientation of

martensite variants, thus it is suggested that by partial substitution of Ni with 5%,

10% and 15% Cu, the critical stress for slip can be increased by 63, 80 and 95 MPa

(difference in net increase in critical stress for slip and net increase in the stress for

reorientation of martensite variants).

4

6

8

10

Frac

ture

Str

ain

(%

)

Cu (at%)

Martensite fracture strain

Austenite fracture strain

0 5 10 15


alloy on martensite fracture strain, tested at (Mf – 50°C) and austenite

fracture strain, tested at (Af + 50°C)

121

6.8 Summary

In this chapter the effect of 5%, 10% and 15% Cu addition and aging at 400°C,

500°C, 600°C and 700°C for 3 hours of TiNiPd alloys on mechanical properties was

presented. The hardness of all four alloys in both conditions; solution treated and

aged were found by microhardness tester whereas the yield stress, stress at fracture

and strain at fracture were found by using the tensile testing machine.

Increasing the Cu concentration in TiNiPd alloys, the hardness of the resultant alloys

increased by 3%, 4% and 7% for 5Cu, 10Cu and 15Cu alloys respectively, as

compared to 0Cu alloy, in solution treated condition. By aging the 0Cu and 5Cu

alloys at 400°C, 500°C, 600°C and 700°C for 3 hours, the hardness was slightly

increased till aging temperature of 600°C and then slightly decreased at 700°C. The

hardness of 0Cu and 5Cu alloys increased by 7% and 10% respectively when the

alloys were aged at 600°C. The hardness of 10Cu and 15Cu alloys was remarkably

increased when its aging temperature was increased from 400°C to 600°C and then

decreased remarkably almost equal to solution treated condition. It was noted that the

hardness of 10Cu and 15Cu alloys was increased by 60% and 83% respectively.

In solution treated condition, the yield stress and fracture stress were significantly

increased by increasing the Cu concentration in TiNiPd alloys in both martensite and

austenite phases, however the fracture strain was decreased. Yield stress, fracture

stress and fracture strain were found only for 600°C-aged TiNiPdCu alloys. It was

observed that the yield and fracture stresses were significantly increased by

increasing the Cu concentration in TiNiPd alloys, however fracture strain was

decreased. The yield stress of all four alloys was increased by 3 ~ 5% by aging them

at 600°C as compared to solution treated condition. From these results it can be

concluded that addition of Cu in place of Ni in TiNiPd alloys was beneficial to

improve its hardness, yield and fracture stresses.

By comparing the mechanical properties in martensite and austenite phases, it was

observed that yield and fracture stresses in austenite phase were significantly higher

than that of martensite phase. In austenite phase, the yield and fracture stresses of all

four alloys were found to be greater by 52% ~ 57% and 14% ~ 18% respectively, as

122

compared to martensite phase. However the fracture strain in austenite phase was

less as compared to martensite phase.

123

Chapter – 7

Effect of Copper Addition on Shape

Memory Properties of TiNiPd Alloys

7.1 Introduction

In this chapter, the shape memory properties; transformation temperatures under

biased load, recovered and irrecoverable strains, recovery ratio and work output of

0Cu, 5Cu, 10Cu and 15Cu alloys were determined in solution treated condition. The

detailed results and discussions are presented in the subsequent sections.

7.2 Shape memory properties of TiNiPd alloys with varying Cu

percentage

To investigate the effect of Cu addition on shape memory properties of TiNiPd high

temperature shape memory alloys, constant stress thermal cycling experiments were

conducted at various stress levels in solution treated condition of 0Cu, 5Cu, 10Cu

and 15Cu alloys.

7.2.1 Shape memory properties of 0Cu alloy

Fig. 7.1 shows the strain-temperature curves for solution treated 0Cu alloy at stress

levels of 100 – 500 MPa. Thermal cycling experiments (loading, heating and

cooling) at various stress levels were carried out by the same manner as explained in

section 3.6.2. Transformation temperatures (Ms, Mf, As, Af) at different stress levels

were measured by the tangent intersection method as shown in Fig. 3.7.

Fig. 7.2 represents the change in transformation temperatures (measured from Fig.

7.1) with respect to the applied stress levels for solution treated 0Cu alloy. It can be

observed from Fig. 7.2 that all the transformation temperatures increased as the stress

level increased from 100 MPa to 500 MPa. The Ms and Af temperatures; 150°C and

200°C at 100 MPa, increased to 163°C and 220°C respectively when the applied

stress was increased to 300 MPa. Similarly when the stress was increased to 500

MPa, the Ms and Af temperatures also increased to 177°C and 240°C respectively.

124

Moreover, a linear relationship between the transformation temperatures and applied

stress was observed for all stress levels, satisfied the Clausius-Clapeyron (Cs-Cl)

equation (dσ/dT = constant) [4].

The recovered strain [equal to the difference in strain between the Af + 50°C and

50°C of heating cycle] and irrecoverable strain [equal to the strain between heating

and cooling cycle at Af + 50°C] were calculated by the measurement scheme shown

in Fig. 4.7. The calculated transformation strains (εrec, εirr) for the solution treated

0Cu alloy were re-plotted in Fig. 7.3.

-50 50 150 250 350

Stra

in (

%)

Temperature (°C)

100 MPa

200 MPa

300 MPa

400 MPa

500 MPa

Heating

Cooling 10

%


properties of solution treated 0Cu alloy at stress levels of 100 – 500 MPa

125

It can be observed from Fig. 7.3, that by increasing the applied stress, both the

recovered and irrecoverable strains also increased. This increase in transformation

strain can be attributed to the increase in number of favored martensite variants in

direction of applied stress [155]. At 100 MPa, the recovered strain of 2.61%

increased to 4.09% when the applied stress was increased to 300 MPa. By further

50

100

150

200

250

300

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Tem

per

atu

re (

oC

)

Stress (MPa)

Ms

Mf

As

Af

Fig. 7.2 Change in transformation temperatures of solution treated

0Cu alloy at stress levels of 100 – 500 MPa

0.0

0.5

1.0

1.5

2.0

2.5

0

1

2

3

4

5

6

0 100 200 300 400 500 600

Irre

cove

rab

le S

trai

n (

%)

Re

cove

red

Str

ain

(%

)

Stress (MPa)

Ɛrec

Ɛirr


alloy at stress levels of 100 – 500 MPa

126

increasing the applied stress to 500 MPa, the recovered strain further increased to

4.8%. Similarly at 100 MPa, the irrecoverable strain of 0.82% increased to 1.51%

when the applied stress was increased to 300 MPa. By further increasing the applied

stress to 500 MPa, the irrecoverable strain further increased to 2.2%.

Fig. 7.4 represents the recovery ratio and work output for solution treated 0Cu alloy,

calculated from the transformation strains and their corresponding stress levels. It

can be noted, that the recovery ratio decreased as the applied stress was increased

from 100 MPa to 500 MPa. The recovery ratio of 76% resulted at 100 MPa

decreased to 73% at 300 MPa and then further decreased to 68% when the applied

stress was further increased to 500 MPa. As shown in Fig. 7.4, the work output

increased continuously by increasing the applied stress from 100 MPa to 500 MPa.

The work output of 2.61 J/cm3 at 100 MPa increased to 12.27 J/cm

3 when the applied

stress was increased to 300 MPa. By further increasing the applied stress to 500

MPa, the work output further increased to 24 J/cm3.

0

5

10

15

20

25

30

50

60

70

80

90

100

110

0 100 200 300 400 500 600

Wo

rk O

utp

ut

(J/c

m3)

Re

cove

ry R

atio

(%

)

Stress (MPa)

Recovery Ratio

Work Output


at stress levels of 100 – 500 MPa

127



levels of 100 – 500 MPa. The transformation temperatures at various stress levels

were calculated by the same manner as explained in section 7.2.1.

Fig. 7.6 represents the variation of transformation temperatures (measured from Fig.









equation.

-50 50 150 250 350

Stra

in (

%)

Temperature (°C)

100 MPa

200 MPa

300 MPa

400 MPa

500 MPa

Heating

Cooling 10

%



128

The recovered and irrecoverable strains were calculated from Fig. 7.5 for the solution

treated 5Cu alloy and re-plotted in Fig. 7.7. It can be observed, that by increasing the

applied stress, both the recovered and irrecoverable strains also increased. At 100

MPa, the recovered strain of 2.56% increased to 4.15% when the applied stress was

increased to 300 MPa. By further increasing the applied stress to 500 MPa, the

recovered strain further increased to 4.9%. Similarly at 100 MPa, the irrecoverable

strain of 0.6% increased to 1.2% when the applied stress was increased to 300 MPa.

By further increasing the applied stress to 500 MPa, the irrecoverable strain further

increased to 1.7%.

50

100

150

200

250

300

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Tem

per

atu

re (

oC

)

Stress (MPa)

Ms

Mf

As

Af



129

Fig. 7.8 represents the recovery ratio and work output for solution treated 5Cu alloy,


can be noted, that the recovery ratio decreased as the applied stress was increased



0.0

0.5

1.0

1.5

2.0

2.5

0

1

2

3

4

5

6

0 100 200 300 400 500 600

Irre

cove

rab

le S

trai

n (

%)

Re

cove

red

Str

ain

(%

)

Stress (MPa)

Ɛrec

Ɛirr



0

5

10

15

20

25

30

50

60

70

80

90

100

110

0 100 200 300 400 500 600

Wo

rk O

utp

ut

(J/c

m3)

Re

cove

ry R

atio

(%

)

Stress (MPa)

Recovery Ratio

Work Output



130




3 when the applied


MPa, the work output further increased to 24.5 J/cm3.



levels of 100 – 500 MPa.



observed from Fig. 7.10 that all the transformation temperatures increased as the

stress level increased from 100 MPa to 500 MPa. The Ms and Af temperatures;

198°C and 245°C at 100 MPa, increased to 207°C and 263°C respectively when the

applied stress was increased to 300 MPa. Similarly when the stress was increased to

500 MPa, the Ms and Af temperatures also increased to 234°C and 283°C

respectively. Moreover, a linear relationship between the transformation

temperatures and applied stress was observed for all stress levels, satisfied the

Clausius-Clapeyron (Cs-Cl) equation.

-50 50 150 250 350

Stra

in (

%)

Temperature (°C)

100 MPa

200 MPa

300 MPa

400 MPa

500 MPa

Heating

Cooling 10

%



131

The recovered and irrecoverable strains were calculated from Fig. 7.9 for the solution

treated 10Cu alloy and re-plotted in Fig. 7.11. It can be observed, that by increasing

the applied stress, both the recovered and irrecoverable strains also increased. At 100

MPa, the recovered strain of 3.95% increased to 5.2% when the applied stress was

50

100

150

200

250

300

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Tem

per

atu

re (

oC

)

Stress (MPa)

Ms

Mf

As

Af



0.0

0.5

1.0

1.5

2.0

2.5

0

1

2

3

4

5

6

0 100 200 300 400 500 600

Irre

cove

rab

le S

trai

n (

%)

Re

cove

red

Str

ain

(%

)

Stress (MPa)

Ɛrec

Ɛirr



132


recovered strain remained stable at 5.2%.

Similarly at 100 MPa, the irrecoverable strain of 0.1% increased to 0.6% when the

applied stress was increased to 300 MPa. By further increasing the applied stress to

500 MPa, the irrecoverable strain further increased to 1.6%.

Fig. 7.12 represents the recovery ratio and work output for solution treated 10Cu

alloy, calculated from the transformation strains and their corresponding stress

levels. It can be noted that the recovery ratio decreased as the applied stress was

increased from 100 MPa to 500 MPa. The recovery ratio of 97% resulted at 100 MPa





3 when the applied


MPa, the work output further increased to 26 J/cm3.

0

5

10

15

20

25

30

50

60

70

80

90

100

110

0 100 200 300 400 500 600W

ork

Ou

tpu

t (J

/cm

3)

Re

cove

ry R

atio

(%

)

Stress (MPa)

Recovery Ratio

Work Output

Fig. 7.12 Recovery ratio and work output of solution treated 10Cu


133



levels of 100 – 500 MPa. Fig. 7.14 represents the variation of transformation

temperatures (measured from Fig. 7.13) with respect to the applied stress levels for

solution treated 15Cu alloy. It can be observed from Fig. 7.14 that all the

transformation temperatures increased as the stress level increased from 100 MPa to

500 MPa. The Ms and Af temperatures; 214°C and 261°C at 100 MPa, increased to

226°C and 277°C respectively when the applied stress was increased to 300 MPa.

Similarly when the stress was increased to 500 MPa, the Ms and Af temperatures also

increased to 244°C and 295°C respectively. Moreover, a linear relationship between

the transformation temperatures and applied stress was observed for all stress levels,

satisfied the Clausius-Clapeyron (Cs-Cl) equation.

-50 50 150 250 350

Stra

in (

%)

Temperature (°C)

100 MPa

200 MPa

300 MPa

400 MPa

500 MPa

Heating

Cooling 10

%



134

The recovered and irrecoverable strains were calculated from Fig. 7.13 for the

solution treated 15Cu alloy and re-plotted in Fig. 7.15. It can be observed, that by

increasing the applied stress, both the recovered and irrecoverable strains also

increased. At 100 MPa, the recovered strain of 3.65% increased to 4.8% when the

applied stress was increased to 300 MPa. By further increasing the applied stress to

500 MPa, the recovered strain further increased to 5.1%. Similarly at 100 MPa, the

irrecoverable strain of 0.2% increased to 0.62% when the applied stress was


irrecoverable strain further increased to 1.7%.

Fig. 7.16 represents the recovery ratio and work output for solution treated 15Cu

alloy, calculated from the transformation strains and their corresponding stress

levels. It can be noted, that the recovery ratio decreased as the applied stress was

increased from 100 MPa to 500 MPa. The recovery ratio of 95% resulted at 100 MPa





3 when the applied



50

100

150

200

250

300

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Tem

per

atu

re (

oC

)

Stress (MPa)

Ms

Mf

As

Af



135

0.0

0.5

1.0

1.5

2.0

2.5

0

1

2

3

4

5

6

0 100 200 300 400 500 600

Irre

cove

rab

le S

trai

n (

%)

Re

cove

red

Str

ain

(%

)

Stress (MPa)

Ɛrec

Ɛirr



0

5

10

15

20

25

30

50

60

70

80

90

100

110

0 100 200 300 400 500 600

Wo

rk O

utp

ut

(J/c

m3)

Re

cove

ry R

atio

(%

)

Stress (MPa)

Recovery Ratio

Work Output

Fig. 7.16 Recovery ratio and work output of solution treated 15Cu


136

7.3 Effect of Cu addition on transformation temperatures

Fig. 7.17 and 7.18 represent the martensite start and austenite finish temperatures of

solution treated 0Cu, 5Cu, 10Cu and 15Cu alloys under stress levels of 100 – 500

MPa, respectively. The Ms and Af temperatures of 0Cu alloy increased each by 25°C

upon addition of 5% Cu in place of Ni at stress level of 100 MPa. The Ms and Af

temperatures of 0Cu alloy further increased by 48°C and 45°C respectively when Ni

was replaced by 10% Cu. When Ni was replaced by 15% Cu, the Ms and Af

temperatures further increased by 64°C and 61°C respectively as compared to

baseline 0Cu alloy.

Similarly in the same manner, it can be observed from Fig. 7.17 that the martensite

start temperatures at different stress levels increased by increasing the Cu-content in

place of Ni. The increase in transformation temperatures can be attributed to the

decrease in Ni-content and Ni/Pd ratio, because Cu content does not affect the

transformation temperatures [146]. It has been reported that according to the

TiNiTiPd pseudo-binary phase diagram, the increase in Pd-content and decrease in

Ni-content, increases the transformation temperatures [147]. Thus the decrease in

Ni/Pd ratio (as Ni content decreased and Pd content remained constant) resulted in

increase in transformation temperatures.

150

200

250

300

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Te

mp

erat

ure

(°C

)

Stress (MPa)

Ms

0Cu

5Cu

10Cu

15Cu

Fig. 7.17 Comparison of martensite start temperatures of solution treated

0Cu, 5Cu, 10Cu and 15Cu alloys under stress level of 100 – 500 MPa

137

7.4 Effect of Cu addition on transformation strains

Fig. 7.19 and 7.20 show the transformation strain (εrec) and irrecoverable strain (εirr)

respectively of solution treated 0Cu, 5Cu, 10Cu and 15Cu alloys under stress levels

of 100 – 500 MPa. It can be observed from Fig. 7.19 that the εrec of 0Cu and 5Cu

alloys were remained almost at the same level whereas it was increased by 1.34%

when Ni was replaced by 10% Cu at stress level of 100 MPa. However, the εrec of

15Cu alloy was slightly decreased as compared to 10Cu alloy. Similarly in the same

manner, the εrec of 0Cu and 5Cu alloys were remained at the same level and then it

increased for 10Cu alloy and then decreased for 15Cu alloy at all stress levels.

150

200

250

300

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Tem

per

atu

re (

°C)

Stress (MPa)

Af

0Cu

5Cu

10Cu

15Cu

Fig. 7.18 Comparison of austenite finish temperatures of solution treated

0Cu, 5Cu, 10Cu and 15Cu alloys under stress level of 100 – 500 MPa

138

It can be observed from Fig. 7.20 that εirr of 5Cu and 10Cu alloys was decreased by

0.22% and 0.72% respectively as compared to 0Cu alloy at stress level of 100 MPa.

However, εirr of 15Cu was remained almost at the same level as of 10Cu alloy.

Similarly in the same manner, the εirr of 5Cu and 10Cu alloys was decreased as

compared to 0Cu alloy at all stress levels. Moreover it can also be observed from Fig.

2

3

4

5

6

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Str

ain

(%

)

Stress (MPa)

εrec

0Cu

5Cu

10Cu

15Cu

Fig. 7.19 Comparison of recovered strain of solution treated 0Cu, 5Cu,


0

1

2

3

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Str

ain

(%

)

Stress (MPa)

εirr

0Cu

5Cu

10Cu

15Cu

Fig. 7.20 Comparison of irrecoverable strain of solution treated 0Cu,

5Cu, 10Cu and 15Cu alloys under stress level of 100 – 500 MPa

139

7.19 and 7.20 that the transformation strains increased by increasing the stress level

from 100 MPa to 500 MPa.

7.5 Effect of Cu addition on recovery ratio and work output

Fig. 7.21 and 7.22 show the recovery ratio and work output of solution treated 0Cu,

5Cu, 10Cu and 15Cu alloys under stress levels of 100 – 500 MPa. It can be observed

from Fig. 7.21 that the recovery ratio of 5Cu and 10Cu alloys was increased by 5%

and 21% as compared to 0Cu alloy at stress level of 100 MPa. However, the recovery

ratio of 15Cu was slightly decreased by 2% as compared to 10Cu alloy and the same

trend remained till higher stress levels. Moreover it can also be observed that for

each alloy the recovery ratio was decreased as the applied stress level was increased

from 100 MPa to 500 MPa.

Work output of 5Cu alloy was observed to be remained almost the same as of 0Cu

alloy, however it was increased by 1.34 J/cm3

for 10Cu alloy as compared to 0Cu

alloy. Again for 15Cu alloy the work output was remained almost the same as of

10Cu alloy. Similarly in the same manner, the work output of 0Cu and 5Cu alloys

was observed to be remained at the same level at higher stress levels, whereas it was

increased for 10Cu alloy. For 15Cu alloy the work output was again slightly

decreased as compared to 10Cu alloy. Moreover it can also be observed that for each

alloy the work output was increased as the applied stress level was increased from

100 MPa to 500 MPa.

140

0

10

20

30

0 100 200 300 400 500 600 700

Wo

rk O

utp

ut

(J/c

m3)

Stress (MPa)

0Cu

5Cu

10Cu

15Cu

Fig. 7.22 Comparison of work output of solution treated 0Cu, 5Cu,


60

70

80

90

100

0 100 200 300 400 500 600 700

Re

cove

ry R

atio

(%

)

Stress (MPa)

0Cu

5Cu

10Cu

15Cu

Fig. 7.21 Comparison of recovery ratio of solution treated 0Cu, 5Cu,


141

7.6 Summary

To study the effect of Cu addition on shape memory properties of TiNiPd high

temperature shape memory alloys, constant stress thermomechanical cycling

experiments in solution treated condition were conducted under various stress levels

of 100 – 500 MPa. By substitution of Cu in place of Ni, the shape memory

properties; transformation temperatures, recovered and irrecoverable strains,

recovery ratio and work output of the alloy were changed. The transformation

temperatures of baseline 0Cu alloy increased significantly when Ni was replaced by

5% Cu. Increasing the Cu content to 10% and 15%, the transformation temperatures

of the alloy were further increased and got maximum values at 15% Cu. The

martensite start temperatures of 5Cu, 10Cu and 15Cu alloys increased by 16%, 32%

and 38% respectively as compared to 0Cu alloy.

The recovered and irrecoverable strains of alloys were increased up to addition of

10% Cu and then slightly decreased when 15% Cu was added. The strain recovery of

5Cu, 10Cu and 15Cu alloys were observed to be increased by 2%, 8% and 6%

respectively as compared to 0Cu alloy. In the same manner, the irrecoverable strain

was decreased up to addition of 10% Cu and then slightly increased by addition of

15% Cu. It was observed that the irrecoverable strain was decreased by 23%, 27%

and 23% for 5Cu, 10Cu and 15Cu alloys respectively. Like recovered strain, the

recovery ratio was also increased by increasing the Cu content up to 10% and then

slightly decreased by addition of 15% Cu. Recovery ratio of 5Cu, 10Cu and 15Cu

alloys was increased by 6%, 27% and 25% respectively as compared to 0Cu alloy.

Similarly the work output of the alloy was increased up to addition of 10% Cu and

then slightly decreased when 15% Cu was added. The work output of 5Cu, 10Cu and

15Cu alloys was increased by 2%, 8% and 6% respectively as compared to the

baseline 0Cu alloy. From these observations it can be firmly concluded that addition

of 10% Cu in place of Ni in TiNiPd alloy resulted an improved shape memory

properties in terms of recovered and irrecoverable strains, recovery ratio and work

output.

142

Chapter – 8

Effect of Aging on Shape Memory

Properties of TiNiPd Alloys

8.1 Introduction

In this chapter, the shape memory properties; transformation temperatures under

biased load, recovered and irrecoverable strains, recovery ratio and work output were

investigated for of 0Cu, 5Cu, 10Cu and 15Cu alloys, aged at aging temperatures of

600°C for 3 hours. The detailed results and discussion are given in the following

sections.

8.2 Shape memory properties of 600°C-aged TiNiPd alloys with

varying Cu percentage

To investigate the effect of aging on shape memory properties, constant stress

thermal cycling experiments were performed. Later on the transformation

temperatures, recovered and irrecoverable strains, recovery ratio and work output of

600°C-aged 0Cu, 5Cu, 10Cu and 15Cu alloys were compared to that of solution

treated samples presented in Chapter – 7.

8.2.1 Shape memory properties of 600°C-aged 0Cu alloy

Fig. 8.1 shows the strain-temperature curves of 600°C-aged 0Cu alloy at stress levels

of 100 – 500 MPa. Thermal cycling experiments (loading, heating and cooling) at

various stress levels were carried out by the same manner as explained in section

3.6.2. Transformation temperatures (Ms, Mf, As, Af) at different stress levels were

measured by the tangent intersection method as shown in Fig. 3.7.


8.1) with respect to the applied stress levels for 600°C-aged 0Cu alloy. It can be




143





equation [4].

The recovered and irrecoverable strains were calculated by the measurement scheme

shown in Fig. 3.7. The calculated transformation strains (εrec, εirr) for the 600°C-aged

0Cu alloy were re-plotted in Fig. 8.3.

It can be observed, that by increasing the applied stress, both the recovered and

irrecoverable strains also increased. This increase in transformation strains can be

attributed to the increase in number of favored martensite variants in direction of

applied stress [155].

-50 50 150 250 350

Stra

in (

%)

Temperature (°C)

100 MPa

200 MPa

300 MPa

400 MPa

500 MPa

Heating

Cooling 10

%


properties of 600°C-aged 0Cu alloy at stress levels of 100 – 500 MPa

144

At 100 MPa, the recovered strain of 2.59% increased to 4.00% when the applied


MPa, the recovered strain further increased to 4.7%. Similarly at 100 MPa, the

irrecoverable strain of 0.8% increased to 1.4% when the applied stress was increased

to 300 MPa. By further increasing the applied stress to 500 MPa, the irrecoverable

strain further increased to 2.05%.

50

100

150

200

250

300

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Tem

per

atu

re (

oC

)

Stress (MPa)

Ms

Mf

As

Af

Fig. 8.2 Change in transformation temperatures of 600°C-aged


0.0

0.5

1.0

1.5

2.0

2.5

0

1

2

3

4

5

6

0 100 200 300 400 500 600

Irre

cove

rab

le S

trai

n (

%)

Re

cove

red

Str

ain

(%

)

Stress (MPa)

Ɛrec

Ɛirr



145

Fig. 8.4 represents the recovery ratio and work output of 600°C-aged 0Cu alloy,


can be noted that the recovery ratio decreased as the applied stress was increased






3 when the applied




The strain-temperature curves of 600°C-aged 5Cu alloy at stress levels of 100 – 500

MPa is shown in Fig. 8.5.

0

5

10

15

20

25

30

50

60

70

80

90

100

110

0 100 200 300 400 500 600

Wo

rk O

utp

ut

(J/c

m3)

Re

cove

ry R

atio

(%

)

Stress (MPa)

Recovery Ratio

Work Output


stress levels of 100 – 500 MPa

146

The variation of transformation temperatures (measured from Fig. 8.5) with respect

to the applied stress levels for 600°C-aged 5Cu alloy is shown in Fig. 8.6. It can be




-50 50 150 250 350

Stra

in (

%)

Temperature (°C)

100 MPa

200 MPa

300 MPa

400 MPa

500 MPa

Heating

Cooling 10

%



50

100

150

200

250

300

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Te

mp

erat

ure

(oC

)

Stress (MPa)

Ms

Mf

As

Af



147





equation.

The recovered and irrecoverable strains were calculated from Fig. 8.5 for the 600°C-

aged 5Cu alloy and then it were re-plotted in Fig. 8.7.


irrecoverable strains also increased. At 100 MPa, the recovered strain of 2.5%



4.8%. Similarly at 100 MPa, the irrecoverable strain of 0.5% increased to 1.0% when

the applied stress was increased to 300 MPa. By further increasing the applied stress

to 500 MPa, the irrecoverable strain further increased to 1.5%.

0.0

0.5

1.0

1.5

2.0

2.5

0

1

2

3

4

5

6

0 100 200 300 400 500 600

Irre

cove

rab

le S

trai

n (

%)

Re

cove

red

Str

ain

(%

)

Stress (MPa)

Ɛrec

Ɛirr



148

Fig. 8.8 represents the recovery ratio and work output of 600°C-aged 5Cu alloy,

calculated from the transformation strains at their corresponding stress levels. It can

be noted, that the recovery ratio decreased as the applied stress was increased from

100 MPa to 500 MPa. The recovery ratio of 83% resulted at 100 MPa decreased to

80% at 300 MPa and then further decreased to 76% when the applied stress was

further increased to 500 MPa. As shown in Fig. 8.8, the work output increased

continuously by increasing the applied stress from 100 MPa to 500 MPa. The work

output of 2.5 J/cm3 at 100 MPa increased to 12.3 J/cm

3 when the applied stress was

increased to 300 MPa. By further increasing the applied stress to 500 MPa, the work

output further increased to 24.0 J/cm3.


Fig. 8.9 shows the strain-temperature curves of 600°C-aged 10Cu alloy at stress

levels of 100 – 500 MPa.

0

5

10

15

20

25

30

50

60

70

80

90

100

110

0 100 200 300 400 500 600

Wo

rk O

utp

ut

(J/c

m3)

Re

cove

ry R

atio

(%

)

Stress (MPa)

Recovery Ratio

Work Output



149


8.9) with respect to the applied stress levels for 600°C-aged 10Cu alloy. It can be

observed from Fig. 8.10 that all the transformation temperatures increased as the

stress level increased from 100 MPa to 500 MPa. The Ms and Af temperatures;

162°C and 225°C at 100 MPa, increased to 175°C and 230°C respectively when the

applied stress was increased to 300 MPa. Similarly when the stress was increased to

500 MPa, the Ms and Af temperatures also increased to 205°C and 233°C

respectively. Moreover, a linear relationship between the transformation

temperatures and applied stress was observed for all stress levels, satisfied the

Clausius-Clapeyron (Cs-Cl) equation.

The recovered and irrecoverable strains were calculated from Fig. 8.9 were re-plotted

in Fig. 8.11. It can be observed, that by increasing the applied stress, both the

recovered and irrecoverable strains also increased. At 100 MPa, the recovered strain

of 0.45% increased to 0.96% when the applied stress was increased to 300 MPa. By

further increasing the applied stress to 500 MPa, the recovered strain further

increased to 1.4%. Similarly at 100 MPa, the irrecoverable strain of 0% increased to

0.03% when the applied stress was increased to 300 MPa. By further increasing the

applied stress to 500 MPa, the irrecoverable strain further increased to 0.09%.

-50 50 150 250 350

Stra

in (

%)

Temperature (°C)

2.5

%

300 MPa

100 MPa

500 MPa

400 MPa

200 MPa

Heating

Cooling



150

50

100

150

200

250

300

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Tem

per

atu

re (

oC

)

Stress (MPa)

Ms

Mf

As

Af



0.0

0.5

1.0

1.5

2.0

2.5

0

1

2

3

4

5

6

0 100 200 300 400 500 600

Irre

cove

rab

le S

trai

n (

%)

Re

cove

red

Str

ain

(%

)

Stress (MPa)

Ɛrec

Ɛirr

Fig. 8.11 Recovered and irrecoverable strains of 600°C-aged 10Cu


151

Fig. 8.12 represents the recovery ratio and work output of 600°C-aged 10Cu alloy. It

can be noted that the recovery ratio decreased as the applied stress was increased






3 when the applied




Fig. 8.13 shows the strain-temperature curves of 600°C-aged 15Cu alloy at stress

levels of 100 – 500 MPa. Fig. 8.14 represents the variation of transformation

temperatures (measured from Fig. 8.13) with respect to the applied stress levels for

600°C-aged 15Cu alloy. It can be observed from Fig. 8.14 that all the transformation

temperatures increased as the stress level increased from 100 MPa to 500 MPa. The

Ms and Af temperatures; 178°C and 241°C at 100 MPa, increased to 187°C and

251°C respectively when the applied stress was increased to 300 MPa.

0

5

10

15

20

25

30

50

60

70

80

90

100

110

0 100 200 300 400 500 600

Wo

rk O

utp

ut

(J/c

m3)

Re

cove

ry R

atio

(%

)

Stress (MPa)

Recovery Ratio

Work Output



152

-50 50 150 250 350

Stra

in (

%)

Temperature (°C)

2.5

%

300 MPa

100 MPa

500 MPa

400 MPa

200 MPa

Heating

Cooling



50

100

150

200

250

300

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Te

mp

erat

ure

(oC

)

Stress (MPa)

Ms

Mf

As

Af



153

Similarly when the stress was increased to 500 MPa, the Ms and Af temperatures also

increased to 211°C and 258°C respectively. Moreover, a linear relationship between

the transformation temperatures and applied stress was observed for all stress levels,

satisfied the Clausius-Clapeyron (Cs-Cl) equation.

0.0

0.5

1.0

1.5

2.0

2.5

0

1

2

3

4

5

6

0 100 200 300 400 500 600

Irre

cove

rab

le S

trai

n (

%)

Re

cove

red

Str

ain

(%

)

Stress (MPa)

Ɛrec

Ɛirr

Fig. 8.15 Recovered and irrecoverable strains of 600°C-aged 15Cu


0

5

10

15

20

25

30

50

60

70

80

90

100

110

0 100 200 300 400 500 600

Wo

rk O

utp

ut

(J/c

m3)

Re

cove

ry R

atio

(%

)

Stress (MPa)

Recovery Ratio

Work Output



154

The transformation strains were calculated from Fig. 8.13 for the 600°C-aged 15Cu

alloy and their values were re-plotted in Fig. 8.15.


irrecoverable strains also increased. At 100 MPa, the recovered strain of 0.43%



1.34%. Similarly at 100 MPa, the irrecoverable strain of 0% remained at 0% when

the applied stress was increased to 300 MPa. By further increasing the applied stress

to 500 MPa, the irrecoverable strain increased to only 0.07%.

Fig. 8.16 represents the recovery ratio and work output for 600°C-aged 15Cu alloy,

calculated from the transformation strains at their corresponding stress levels. It can

be noted that the recovery ratio decreased as the applied stress was increased from

100 MPa to 500 MPa. However the recovery ratio remained stable till the stress level

of 300 MPa. The recovery ratio of 100% resulted at 100 MPa remained at the same

level at 300 MPa and then decreased to 94% when the applied stress was increased to

500 MPa. As shown in Fig. 8.16, the work output increased continuously by

increasing the applied stress from 100 MPa to 500 MPa. The work output of 0.43

J/cm3 at 100 MPa increased to 2.73 J/cm

3 when the applied stress was increased to

300 MPa. By further increasing the applied stress to 500 MPa, the work output

further increased to 6.7 J/cm3.

8.3 Comparison of shape memory properties between solution

treated and 600°C-aged 0Cu alloys

8.3.1 Comparison of transformation temperatures

Fig. 8.17 shows the comparison of transformation temperatures of solution treated

and 600°C-aged 0Cu alloys under stress levels of 100 – 500 MPa. It can be observed

that the Ms and Af temperatures of solution treated and 600°C-aged samples

remained unchanged at lower stress levels. However at higher stress levels the Ms

and Af temperatures of 600°C-aged sample were slightly increased as compared to

solution treated condition. At 500 MPa, the Ms and Af temperatures of 600°C-aged

sample were increased by 3°C and 5°C respectively. This result showed that in case

155

of 0Cu alloy, there was no effect on transformation temperatures when the alloy aged

at 600°C for 3 hours. The slight increase in transformation temperatures was due to

the grain growth as discussed in section 5.3.1.

8.3.2 Comparison of transformation strains

The comparison of transformation strains between solution treated and 600°C-aged

0Cu alloys under stress levels of 100 – 500 MPa is shown in Fig. 8.18. At lower

stress levels, the transformation strains of solution treated and 600°C-aged samples

remained same whereas at higher stress levels the transformation strains of 600°C-

aged sample were slightly decreased as compared to solution treated condition. At

500 MPa, the εrec and εirr of 600°C-aged sample were decreased by 0.1% and 0.15%

respectively. This result indicated that there was no effect on transformation strains

by aging the 0Cu alloy at 600°C for 3 hours.

100

150

200

250

300

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Te

mp

erat

ure

(°C

)

Stress (MPa)

Soln Treated

Ms

Af

600 °C-aged Ms

Af

Fig. 8.17 Comparison of transformation temperatures of solution

treated and 600°C-aged 0Cu alloy under stress level of 100 – 500

MPa

156

8.3.3 Comparison of recovery ratio and work output

Fig. 8.19 represents the comparison of recovery ratio and work output between

solution treated and 600°C-aged 0Cu alloys under stress levels of 100 – 500 MPa.

Here again it can be observed that at lower stress levels, the recovery ratio and work

output of solution treated and 600°C-aged samples remained same whereas at higher

stress levels the work output of 600°C-aged sample was slightly decreased as

compared to solution treated condition. However the recovery ratio of 600°C-aged

sample was slightly improved. At 500 MPa, the work output was decreased by 0.02

J/cm3 and recovery ratio was increased by 2% when the alloy was aged at 600°C for

3 hours. This result suggested that there was no significant effect on recovery ratio

and work output by aging the 0Cu alloy at 600°C for 3 hours.

0

2

4

6

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Str

ain

(%

)

Stress (MPa)

Soln Treated εrec

εirr

600 °C-aged εrec

εirr



157




The comparison of transformation temperatures of solution treated and 600°C-aged

samples under stress levels of 100 – 500 MPa for 5Cu alloys is shown in Fig. 8.20.

By comparing their transformation temperatures, it can be observed that the Ms and

Af temperatures of 600°C-aged sample were slightly increased than that of solution

treated sample at lower stress levels. Upon further increasing the stress level to 500

MPa, the Ms and Af temperatures of 600°C-aged sample were further increased as

compared to solution treated condition. At 100 MPa, the Ms and Af temperatures of

600°C-aged sample were increased by 5°C each whereas at 500 MPa, the Ms and Af

temperatures of 600°C-aged sample were further increased by 13°C and 12°C

respectively. This result showed that by aging the 5Cu alloy at 600°C for 3 hours,

slight increase occurred in transformation temperatures at lower as well as at higher

stress levels. The increase in transformation temperatures of 600°C-aged 5Cu alloy

can be attributed to the grain growth as discussed in section 5.3.2.

0

10

20

30

60

70

80

90

100

110

0 100 200 300 400 500 600 700

Wo

rk O

utp

ut

(J/c

m3)

Re

cove

ry R

atio

(%

)

Stress (MPa)

Soln Treated

600 °C-aged

Rec Ratio

Work Output

Rec Ratio

Work Output


treated and 600°C-aged 0Cu alloy under stress level of 100 – 500 MPa

158


Fig. 8.21 shows the comparison of transformation strains between solution treated

and 600°C-aged 5Cu alloys under stress levels of 100 – 500 MPa. It can be observed

that at lower stress levels, the recovered and irrecoverable strains of solution treated

and 600°C-aged samples were remained same whereas at higher stress levels the

recovered strain of 600°C-aged sample were slightly decreased as compared to

solution treated condition. At 500 MPa, the εrec and εirr of 600°C-aged sample were

decreased by 0.1% and 0.2% respectively. This result indicated that there was no

significant effect on transformation strains when the 5Cu alloy was aged at 600°C for

3 hours.

100

150

200

250

300

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Te

mp

erat

ure

(°C

)

Stress (MPa)

Soln Treated

Ms

Af

600 °C-aged Ms

Af



159



solution treated and 600°C-aged 5Cu alloys under stress levels of 100 – 500 MPa.

Work output of solution treated and 600°C-aged samples were remained same at

lower stress levels whereas at higher stress levels the work output of 600°C-aged

sample was slightly decreased as compared to solution treated condition. However

the recovery ratio of 600°C-aged sample was slightly improved at all stress levels. At

500 MPa, the work output of 600°C-aged sample was decreased by 0.06 J/cm3. This

result suggests that there was no significant effect on recovery ratio and work output

by aging the 5Cu alloy at 600°C for 3 hours.

0

2

4

6

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Str

ain

(%

)

Stress (MPa)

Soln Treated εrec

εirr

600 °C-aged εrec

εirr



160




Fig. 8.23 shows the comparison of transformation temperatures of solution treated

and 600°C-aged samples under stress levels of 100 – 500 MPa for 10Cu alloys. It can

be noted that the transformation temperatures of 600°C-aged sample were observed

to be significantly lower than that of solution treated sample at each stress levels. At

stress level of 100 MPa, the Ms and Af temperatures of 600°C-aged sample decreased

by 36°C and 20°C respectively as compared to solution treated condition and the

same trend remained till higher stress levels. This result suggested that due to aging

at 600°C for 3 hours, precipitates were formed, caused to decrease their

transformation temperatures. The same result can be supported from back-scattered

SEM images shown in Fig. 4.8, where both types of precipitates can be observed on

the grain boundaries as well as in the grain interiors.

0

10

20

30

60

70

80

90

100

110

0 100 200 300 400 500 600 700

Wo

rk O

utp

ut

(J/c

m3)

Re

cove

ry R

atio

(%

)

Stress (MPa)

Soln Treated

600 °C-aged

Rec Ratio

Work Output

Rec Ratio

Work Output



161


Fig. 8.24 shows the comparison of transformation strains between solution treated

and 600°C-aged 10Cu alloys under stress levels of 100 – 500 MPa. From Fig. 8.24 it

is evident that at all stress levels, the recoverable strain of 600°C-aged sample

remarkably decreased as compared to solution treated one. At 100 MPa, the εrec of

600°C-aged sample was decreased by 3.5% whereas at 500 MPa, the same value

further decreased to 3.8%. Similarly the irrecoverable strain of 600°C-aged sample

was decreased slightly at lower stress level whereas at higher stress levels the same

value further decreased. At 100 MPa, the εirr of 600°C-aged sample was decreased

by 0.1% whereas at 500 MPa, the same value was further decreased by 1.51% as

compared to that of solution treated sample. The remarkable change in

transformation strains confirms the formation of both types of precipitates when the

10Cu alloy was aged at 600°C for 3 hours.

100

150

200

250

300

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Tem

per

atu

re (

°C)

Stress (MPa)

Soln Treated

Ms

Af

600 °C-aged Ms

Af



162



solution treated and 600°C-aged 10Cu alloys under stress levels of 100 – 500 MPa. It

can be observed that recovery ratio of 600°C-aged sample was found to be greater at

all stress levels. However at lower stress levels the difference in their recovery ratio

was smaller and then become greater when the applied stress level was increased to

500 MPa. At 100 MPa the recovery ratio of 600°C-aged sample was increased by 3%

whereas at 500 MPa the same value further increased by 18% as compared to

solution treated sample. Conversely the work output of 600°C-aged sample was

significantly decreased at all stress levels as compared to solution treated condition.

At 100 MPa, the work output of 600°C-aged sample was significantly decreased by

3.5 J/cm3 and then further decreased by 19 J/cm

3 at 500 MPa as compared to that of

solution treated sample.

0

2

4

6

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Str

ain

(%

)

Stress (MPa)

Soln Treated εrec

εirr

600 °C-aged εrec

εirr



163




The comparison of transformation temperatures of solution treated and 600°C-aged

samples under stress levels of 100 – 500 MPa for 15Cu alloys is shown in Fig. 8.26.

By comparing their transformation temperatures, it can be observed that the Ms and

Af temperatures were significantly decreased when the alloy was aged at 600°C for 3

hours as compared to solution treated sample at each stress levels. At stress level of

100 MPa, the Ms and Af temperatures of 600°C-aged sample decreased by 36°C and

20°C respectively as compared to solution treated condition. Similarly the decrease

in transformation temperatures of 15Cu alloy due to aging at 600°C for hours was

remained till higher stress levels. This result indicated that by aging the 15Cu alloy at

600°C for 3 hours, significant decrease in transformation temperatures was observed

at each stress levels. This result confirmed that due to aging at 600°C for 3 hours,

precipitates were formed, caused to lower their transformation temperatures. The

same result can be supported from back-scattered SEM images shown in Fig. 4.9,

where both types of precipitates can be observed on the grain boundaries as well as

in the grain interiors.

0

10

20

30

60

70

80

90

100

110

0 100 200 300 400 500 600 700

Wo

rk O

utp

ut

(J/c

m3)

Re

cove

ry R

atio

(%

)

Stress (MPa)

Soln Treated

600 °C-aged

Rec Ratio

Work Output

Rec Ratio

Work Output



164


The comparison of transformation strains between solution treated and 600°C-aged

15Cu alloys under stress levels of 100 – 500 MPa is shown in Fig. 8.27. It can be

observed that at all stress levels, the recoverable strain of 600°C-aged sample

remarkably decreased as compared to solution treated one. At 100 MPa, the εrec of

600°C-aged sample was decreased by 3.21% whereas at 500 MPa, the same value

further decreased to 3.76%. Similarly the irrecoverable strain of 600°C-aged sample

was decreased slightly at lower stress level whereas at higher stress levels the same

value further decreased. At 100 MPa, the εirr of 600°C-aged sample was decreased

by 0.2% whereas at 500 MPa, the same value was further decreased by 1.63% as

compared to that of solution treated sample.

100

150

200

250

300

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Tem

per

atu

re (

°C)

Stress (MPa)

Soln Treated

Ms

Af

600 °C-aged Ms

Af



165



solution treated and 600°C-aged 0Cu alloys under stress levels of 100 – 500 MPa. It

can be observed that recovery ratio of 600°C-aged sample was increased at all stress

levels. However at lower stress levels the difference in their recovery ratio was

smaller and then become greater when the applied stress level was increased to 500

MPa. At 100 MPa the recovery ratio of 600°C-aged sample was increased by 5%

whereas at 500 MPa the same value further increased by 20% as compared to

solution treated sample. Conversely the work output of 600°C-aged sample was

significantly decreased at all stress levels as compared to solution treated condition.

At 100 MPa, the work output of 600°C-aged sample was significantly decreased by

3.22 J/cm3 and then further decreased by 18.8 J/cm

3 at 500 MPa as compared to that

of solution treated sample.

0

2

4

6

0 100 200 300 400 500 600 700

Tran

sfo

rmat

ion

Str

ain

(%

)

Stress (MPa)

Soln Treated εrec

εirr

600 °C-aged εrec

εirr



166

0

10

20

30

60

70

80

90

100

110

0 100 200 300 400 500 600 700

Wo

rk O

utp

ut

(J/c

m3)

Re

cove

ry R

atio

(%

)

Stress (MPa)

Soln Treated

600 °C-aged

Rec Ratio

Work Output

Rec Ratio

Work Output

Fig. 8.28 Comparison of recovery ratio and work output of solution treated


167

8.7 Summary

To investigate the effect of aging at 600°C for 3 hours on transformation

temperatures, transformation strains, recovery ratio and work output of TiNiPdCu

alloys, thermal cycling experiments were performed. The shape memory properties

of 600°C-aged alloys were then compared with properties of solution treated alloys

given in Chapter – 7.

By aging 0Cu alloy at 600°C for 3 hours, no change was observed on shape memory

properties as compared to solution treated samples. In case of 5Cu alloy, the

transformation strains and work output were remained unchanged. However

transformation temperatures and recovery ratio of 600°C-aged 5Cu alloy were

slightly increased as compared to that of solution treated one. The martensite start

temperature and recovery ratio increased by 12% and 6% respectively. In view of

these results it can be concluded that aging of 5Cu alloy was beneficial to enhance its

shape memory properties.

Comparing the shape memory properties of 10Cu alloy, it was noted that by aging

the alloy at 600°C for 3 hours, their properties were significantly changed.

Transformation temperatures, transformation strains and work output of 600°C-aged

samples were significantly decreased whereas recovery ratio was slightly increased

as compared to solution treated samples. The recovered strain and work output of the

alloy decreased each by 73% whereas recovery ratio increased by 23%. For 15Cu

alloy the same trend was observed as found in case of 10Cu alloy. Hence for 15Cu

alloy the recovered strain and work output of the alloy decreased each by 74%

whereas recovery ratio increased by 27%. These observations confirmed that aging

of 10Cu and 15Cu alloys was not beneficial in terms of transformation temperatures,

recovered strains and work output, however it was beneficial in terms of

irrecoverable strains and recovery ratio.

168

Chapter – 9

Summary of Results and Discussion

9.1 Introduction

In this chapter, the results and discussion already presented in Chapters 4, 5, 6, 7 and

8 are summarized. In Chapter – 4, the effect of Cu addition and aging on

microstructure has been discussed. The effect of Cu addition and aging on stress free

transformation temperatures has been presented in Chapter – 5. Dependence of

mechanical properties on Cu addition and aging has been discussed in Chapter – 6.

The effect of Cu addition on shape memory properties in solution treated condition

has been shown in Chapter – 7. Lastly, in Chapter – 8, the effect of aging at 600°C

temperature on shape memory properties has been presented.

9.2 Effect of Cu addition on microstructure

To investigate the effect of 5%, 10% and 15% Cu addition on microstructure of

TiNiPd alloys, optical microscopy, scanning electron microscopy, X-ray

diffractometry and energy dispersive spectroscopy were carried out. It was observed

that all the four alloys; 0Cu, 5Cu, 10Cu and 15Cu were consisted of single phase

having typical twinned martensite structure, B19 (orthorhombic) with clearly visible

grain boundaries. However by increasing the Cu concentration in TiNiPd alloys, the

grain size of the resultant TiNiPdCu alloys increased up to 35%. Moreover,

consistent shift of B19 phase was observed towards lower 2Ɵ angle in XRD patterns.

Second phase precipitates were formed in all four alloys at its grain boundaries with

same size having circular or elliptical shapes. This observation proved that addition

of Cu had no effect on the precipitate size and shape, however its density decreased

as the Cu concentration increased in TiNiPd alloys. The chemical composition of

second phase precipitates clearly indicated that the precipitates formed at the grain

boundaries were Ti2Ni which were formed during solidification.

169

9.3 Effect of Cu addition on transformation temperatures

It was observed that by increasing the Cu concentration in TiNiPd alloys, the

transformation temperatures and transformation heats were increased significantly

whereas thermal hysteresis was decreased. By adding 5 ~ 15% Cu in TiNiPd alloys,

the martensite start temperature increased by 9 ~ 42% and thermal hysteresis

decreased by 6 ~ 35%. From this result it can be confirmed that by addition of Cu in

place of Ni in TiNiPd alloys, the transformation temperatures could be significantly

increased whereas its thermal hysteresis could be decreased.

During thermal cycling, the transformation temperatures of solution treated 0Cu and

5Cu alloys decreased by increasing the number of thermal cycles, however the drop

was faster in initial cycles and then reduced in the final cycles. In 0Cu and 5Cu

alloys the drop in martensite start temperature after 5 thermal cycles was 7.5°C and

5°C respectively. Similar behavior was also observed during thermal cycling of

solution treated 10Cu and 15Cu alloys; however the drop in transformation

temperatures was less as compared to 0Cu and 5Cu alloys. In 10Cu and 15Cu alloys

the drop in martensite start temperature was observed to be 3.7°C and 3°C

respectively. This behavior showed that 10Cu and 15Cu alloys were more

dimensionally stable as compared to 0Cu and 5Cu alloys.

9.4 Effect of Cu addition on mechanical and shape memory

properties

During microhardness testing, it was observed that the hardness increased by 7%

when Cu concentration was increased up to 15% in TiNiPd alloys. Similarly, the

yield stress and fracture stress were increased by 26% and 12% respectively, in

martensite phase whereas the same values increased by 28% and 8% in austenite

phase, however the fracture strain decreased. From these results it can be concluded

that addition of Cu in place of Ni in TiNiPd alloys was beneficial to improve its

hardness, yield and fracture stress.

By comparing the mechanical properties in martensite and austenite phases, it was

observed that yield and fracture stresses in austenite phase were significantly higher

than that of martensite phase by 52% ~ 57% and 14% ~ 18% respectively, in all four

170

alloys. By substitution of Cu in place of Ni, the shape memory properties; recovered

and irrecoverable strains, recovery ratio and work output of the alloy were

significantly changed. It was observed that the recovered strain of alloy was

increased up to addition of 10% Cu and then slightly decreased when 15% Cu was

added. The strain recovery of 5Cu, 10Cu and 15Cu alloys was increased by 2%, 8%

and 6% respectively, with respect to the baseline 0Cu alloy. Similarly the

irrecoverable strain was decreased up to addition of 10% Cu and then slightly

increased when 15% Cu was added. The irrecoverable strain of 5Cu, 10Cu and 15Cu

alloys was decreased by 23%, 27% and 23% respectively. The recovery ratio was

observed to be increased by increasing the Cu content up to 10% and then slightly

decreased by addition of 15% Cu. Recovery ratio of 5Cu, 10Cu and 15Cu alloys was

improved by 6%, 27% and 25% respectively as compared to 0Cu alloy. Similarly the

work output of the alloy was increased up to addition of 10% Cu and then slightly

decreased when 15% Cu was added. The work output of 5Cu, 10Cu and 15Cu alloys

was increased by 2%, 8% and 6% respectively. From these observations it can be

firmly concluded that addition of 10% Cu in place of Ni in TiNiPd alloy resulted an

improved shape memory properties in terms of recovered and irrecoverable strains,

recovery ratio and work output.

9.5 Effect of aging on microstructure and transformation

temperatures

Aging at 400°C, 500°C, 600°C and 700°C for 3 hours had no significant effect on

microstructure of 0Cu and 5Cu alloys. However when 10Cu and 15Cu alloys were

aged at above mentioned temperatures, remarkable change in microstructure was

observed. By aging 10Cu and 15Cu alloys, two types of (darker and brighter

contrast) precipitates Ti2Pd and TiPdCu were formed at the grain boundaries and

grain interiors. The density of the precipitates were remained the same, however, the

size in 15Cu alloy was bigger as compared to that of 10Cu alloy.

When 0Cu alloy was aged, the transformation temperatures slightly increased as the

aging temperature increased from 400°C to 700°C. The martensite start temperature

was increased by 1% at aging temperature of 600°C as compared to solution treated

condition. By aging the 5Cu alloy, the transformation temperatures were slightly

171

increased when the aging temperature was increased from 400°C to 600°C and then

slightly decreased at aging temperature of 700°C. The martensite start temperature

increased by 12% when it was aged at 600°C. Conversely, for 10Cu and 15Cu alloys,

the transformation temperatures were significantly decreased as the aging

temperature was increased from 400°C to 600°C and then increased comparable to

that of solution treated condition when the aging temperature was increased to

700°C. The martensite start temperatures of 10Cu and 15Cu alloys were decreased

by 18% and 19% respectively at aging temperature of 600°C. These observations

indicated that aging of 0Cu and 5Cu alloys at 600°C was beneficial to improve its

transformation temperatures whereas aging of 10Cu and 15Cu alloys remarkably

decreased their transformation temperatures.

9.6 Effect of aging on mechanical and shape memory properties

By aging the 0Cu and 5Cu alloys at 400°C, 500°C, 600°C and 700°C for 3 hours, the

hardness was slightly increased till aging temperature of 600°C and then slightly

decreased at 700°C aging temperature. The hardness of 0Cu and 5Cu alloys

increased by 7% and 10% respectively when the alloys were aged at 600°C as

compared to solution treated condition. The hardness of 10Cu and 15Cu alloys was

remarkably increased when its aging temperature was increased from 400°C to

600°C and then decreased remarkably almost equal to solution treated condition. The

hardness of 10Cu and 15Cu alloys increased by 60% and 83% respectively when the

alloys were aged at 600°C. This behavior showed that by aging of TiNiPdCu alloys

having Cu content up to 5%, their hardness increased slightly whereas alloys having

higher Cu content, their hardness increased remarkably.

Yield stress, fracture stress and fracture strain were investigated only for 600°C-aged

TiNiPdCu alloys. It was observed that yield stress was slightly increased by aging

0Cu, 5Cu, 10Cu and 15Cu alloys at 600°C, however fracture strain was decreased as

compared to solution treated condition. The yield stress of all four alloys was

increased by 3 ~ 5% by aging them at 600°C as compared to solution condition.

When 0Cu alloy was aged at 600°C for 3 hours, no change in shape memory

properties was observed between solution treated and 600°C-aged samples. Similarly

172

for 5Cu alloy, the shape memory properties like transformation strains and work

output were remained unchanged. However transformation temperatures and

recovery ratio of 600°C-aged 5Cu alloy were slightly increased as compared to that

of solution treated one. The martensite start temperature and recovery ratio increased

by 12% and 6% respectively. From this discussion it can be concluded that aging of

5Cu alloy was beneficial to enhance its shape memory properties. Comparing the

shape memory properties of 10Cu alloy, it was noted that by aging the alloy at 600°C

for 3 hours, their properties were significantly changed. Transformation

temperatures, transformation strains and work output of 600°C-aged samples were

significantly decreased whereas recovery ratio was slightly increased as compared to

solution treated samples. The recovered strain and work output of the alloy decreased

by 73% whereas recovery ratio increased by 23%. For 15Cu alloy the same trend was

observed as found in case of 10Cu alloy. Hence for 15Cu alloy the recovered strain

and work output of the alloy decreased by 74% whereas recovery ratio increased by

27%. These observations confirmed that aging of 10Cu and 15Cu alloys was not

beneficial in terms of transformation temperatures, recovered strains and work

output, however it was beneficial in terms of irrecoverable strains and recovery ratio.

173

Chapter – 10

Conclusions and Recommendations for

Future Work

10.1 Summary of experimentation

The aim of this research was to improve the transformation temperatures, mechanical

and shape memory properties of TiNiPd high temperature shape memory alloys by

varying percentage of Cu and Ni in the alloy. The effect of aging (at different

temperatures of 400°C, 500°C, 600°C and 700°C for 3 hours) on the microstructure,

mechanical and shape memory properties of the alloys were also investigated. To

investigate the effect of Cu addition and aging on transformation temperatures,

mechanical and shape memory properties, a comprehensive characterization and

thermomechanical cycling experiments were conducted on solution treated and aged

alloys. Experimentation is summarized as below:

Four types of NiTi-based high temperature shape memory alloys i.e.

Ti50Ni25Pd25, Ti50Ni20Pd25Cu5, Ti50Ni15Pd25Cu10 and Ti50Ni10Pd25Cu15 (all in

atomic percent) were melted in vacuum arc melting furnace in desired

composition.

The required heat treatment i.e. homogenization at 950°C for 2 hours was

carried out in sealed quartz tubes.

Button of each alloy (20 g) was sliced in 0.4 mm thick strips using wire

EDM.

Cold rolling of 0.4 mm thick strips was done by 25% and its thickness was

reduced to 0.3 mm.

Solution treatment at 900°C for 1 hour of all samples and aging at 400°C,

500°C, 600°C and 700°C for 3 hours of selected samples were done in sealed

quartz tubes and then water quenched without breaking the quartz tubes.

Size, structure of grain and second phase precipitates was studied using

optical microscope and SEM. EDS was used for compositional analysis.

XRD analysis was carried out to investigate the presence of different phases.

174

Microhardness Tester and Mechanical Testing Machine were used to

investigate the hardness and yield strength.

Tensile Rupture and Creep Testing Machine was used to evaluate the shape

memory properties.

10.2 Conclusions

Four NiTi-based; Ti50Ni25-xPd25Cux (x=0, 5, 10 and 15) high temperature shape

memory alloys were developed. From the microstructural, mechanical

characterization and thermomechanical cycling experiments, the following

conclusions were drawn:

10.2.1 Effect of Cu addition

In solution treated condition, all four alloys consisted of single phase having

typical martensite B19 (orthorhombic) structure with clearly visible grain

boundaries. However by increasing the Cu content till 15%, the grain size

also increased up to 35%. Second phase precipitates of the same size formed

in all alloys, however the density of precipitates decreased with increasing

percentage of Cu. Ti2Ni type precipitates were observed to be formed along

the grain boundaries. Comparing the XRD patterns of all four alloys, it was

observed that the martensite phases consistently shifted towards lower 2Ɵ

angle with increasing content of Cu.

By increasing the Cu content up to 15%, the transformation temperatures

increased by 42%, whereas thermal hysteresis decreased by 35%. During

thermal cycling, the drop in transformation temperatures of 10Cu and 15Cu

alloys were less as compared to 0Cu and 5Cu alloys, hence increasing the Cu

content, the dimensional stability increased.

It was observed that addition of 15% Cu in the alloy, the hardness and yield

strength increased by 7% and 26% respectively in martensite phase, whereas

in austenite phase the yield strength increased by 28%. By comparing, it was

observed that the yield strength in austenite phase of all alloys was higher

than that of martensite phase by 52% ~ 57%.

The recovered strain, recovery ratio and work output increased by 2%, 6%

and 2% respectively when 5% Cu was added in TiNiPd alloy. For 10Cu alloy

175

the same values increased by 8%, 27% and 8% and similarly for 15Cu alloy

increased by 6%, 25% and 6% as compared to 0Cu alloy.

10.2.2 Effect of aging

By aging 0Cu and 5Cu alloys at 400°C, 500°C, 600°C and 700°C for 3 hours,

there was no significant effect on their microstructures in terms of martensite

phase, second phase precipitates and grain size. However, two types of

precipitates; Ti2Pd and TiPdCu (darker and brighter contrast respectively)

produced after aging of 10Cu and 15Cu alloys at mentioned aging

temperatures.

For 0Cu and 5Cu alloys, the martensite start temperature increased by 1% and

12% respectively, however in 10Cu and 15Cu alloys, the same temperature

decreased by 18% and 19% respectively by aging at 600°C. These results

indicated that aging of 0Cu and 5Cu alloys was beneficial to increase their

transformation temperatures, however the same temperatures were

significantly reduced in case of 10Cu and 15Cu alloys.

Hardness of the four alloys increased with increasing Cu content; however,

the increase was more pronounced in case of 10Cu and 15Cu alloys. For

10Cu and 15Cu alloys the hardness increased by 60% and 82% respectively

at aging temperature of 600°C. Similarly the yield strength of all alloys was

also increased by 3 ~ 5% by aging them at 600°C aging temperature.

The shape memory properties remained unchanged after aging 0Cu alloy at

600°C aging temperature. The recovered strain and work output remained

stable whereas the martensite start temperature and recovery ratio increased

by 12% and 6% respectively after aging the 5Cu alloy. By aging the 10Cu

alloy at 600°C, recovered strain and work output decreased by 73% whereas

recovery ratio increased by 23%. Similarly for 15Cu alloy, the same values

decreased by 74% and recovery ratio increased by 27% when it was aged at

600°C.

At the end, it can be concluded that addition of 5%, 10% and 15% Cu in place of Ni

in TiNiPd high temperature shape memory alloys was very useful to improve their

transformation temperatures, dimensional stability, mechanical and shape memory

176

properties. Moreover, it can also be concluded that aging of 0Cu and 5Cu alloy was

beneficial to improve their transformation temperatures and shape memory

properties. However, aging of 10Cu and 15Cu alloys produced an adverse effect on

their properties in terms of shape memory and transformation temperatures.

10.3 Recommendations for future work

Based upon these results, certain recommendations for future endeavors are detailed

below:

This research can be extended by varying the composition of Ni, Pd and Cu

in TiNiPdCu high temperature shape memory alloys.

In the recent research, mechanical properties like stress-strain curves have

been obtained till fracture of samples. If the samples are strained by pre-

defined limit before fracture and thermomechanical cycling (loading and

unloading) are performed for many cycles, it will be helpful to determine the

stress hysteresis and dimensional stability under biased load.

In the present research work, aging at different aging temperatures for

constant time duration has been studied. However it will be of significant

importance if aging at constant aging temperature for variable time duration

is studied.

To investigating the stresses produced due to addition of Cu in TiNiPd alloys

Application of TiNiPdCu alloys in actuators

177

References:

[1] K. Yamauchi, I.O., K. Tsuchiya and S. Miyazaki, Shape memory and

superelastic alloys: Technologies and Application2011: Woodhead

Publishing Limited,. pp. 4-7

[2] Van Humbeeck, J., Advanced engineering materials, 2001. 3(11), p. 837-850.

[3] Otsuka K, W.C.M., Shape Memory Materials Cambridge University Press,

Cambridge, 1998, p. 3

[4] Otsuka, K. and X. Ren, Progress in materials science, 2005. 50(5), p. 511-

678.

[5] Huma, T., : Funakubo H, Shape Memory Materials, Gordon and Breach

Scientific Publisher, New York; 1987. p.83.

[6] Nam, T.H., T. Saburi, and K. Shimizu, Mater Trans JIM, 1990. 31(11), p.

959-967.

[7] Ma, J., I. Karaman, and R.D. Noebe, International Materials Reviews, 2010.

55(5), p. 257-315.

[8] K.N. Melton; in Engineering Aspects of Shape Memory Alloys (Eds: T. W.

Duerig, K.N.M., D. Stockel, C. M. Waymann), 1990: Butterworth –

Heinemann, London pp. 21-34.

[9] Eckelmeyer, K., Scripta metallurgica, 1976. 10(8), p. 667-672.

[10] Lindquist, P. and C. Wayman, Mats, 1989. 9, p. 123-128.

[11] Olander, A., Zeitschrift Fur Kristallographie, 1932. 83(1/2), p. 145-148.

[12] Greninger, A.B., AIME TRANS, 1938. 128, p. 369-377.

[13] Chang, L. and T. Read, Transactions of the American Institute of Mining and

Metallurgical Engineers, 1951. 191(1), p. 47-52.

[14] Buehler, W. and R. Wiley, Transactions of the American Society of Metals,

1962. 55, p. 269-276.

[15] Miyazaki, S. and K. Otsuka, ISIJ International, 1989. 29(5), p. 353-377.

[16] Khan, A., Characterization and Application of Shape Memory Alloy Wires

for Micro and Meso Positioning Systems: D.Sc. dissertation, The Geoarge

Washington University, Washington DC, USA, 2008.

[17] Miyazaki, S., K. Otsuka, and C. Wayman, Acta Metallurgica, 1989. 37(7), p.

1873-1884.

[18] Kudoh, Y., M. Tokonami, S. Miyazaki, and K. Otsuka, Acta Metallurgica,

1985. 33(11), p. 2049-2056.

178

[19] Hara, T., T. Ohba, and K. Otsuka, Journal de physique. IV, 1995. 5(8), p. C8.

641-C8. 645.

[20] Saburi, T., T. Tatsumi, and S. Nenno, Le Journal de Physique Colloques,

1982. 43(C4), p. C4-261-C4-266.

[21] Miyazaki, S., Y. Ohmi, K. Otsuka, and Y. Suzuki, Suppl, 1982. 12, p. 255-

260.

[22] Ling, H.C. and R. Kaplow, Metallurgical and Materials Transactions A, 1980.

11(1), p. 77-83.

[23] Miyazaki, S. and K. Otsuka, Metallurgical Transactions A, 1986. 17(1), p.

53-63.

[24] Salamon, M., M. Meichle, and C. Wayman, Physical Review B, 1985.

31(11), p. 7306.

[25] Shapiro, S., Y. Noda, Y. Fujii, and Y. Yamada, Physical Review B, 1984.

30(8), p. 4314.

[26] Hwang, C. and C. Wayman, Scripta metallurgica, 1983. 17(3), p. 381-384.

[27] Miyazaki, S. and K. Otsuka, Philosophical Magazine A, 1985. 50(3), p. 393-

408.

[28] Goo, E. and R. Sinclair, Acta Metallurgica, 1985. 33, p. 1717-1723.

[29] Todoroki, T. and H. Tamura, Japan Institute of Metals, Journal, 1986. 50, p.

538-545.

[30] Stachowiak, G.B. and P.G. McCormick, Acta Metallurgica, 1988. 36(2), p.

291-297.

[31] Shugo, Y., F. Hasegawa, and T. Honma, Bull. Res. Inst. Min. Dressing

Metall. Tohoku Univ., 1981. 37(1), p. 79-88.

[32] Lindquist, P. and C. Wayman. Transformation Behavior of Ti sub 50--(Ni

sub(50--X) Pd sub X) and Ti sub 50--(Ni sub(50--X) Pt sub X) Alloys. in

Proceedings of the MRS International Meeting on Advanced Materials. 1988.

[33] Firstov, G., J. Van Humbeeck, and Y.N. Koval, Journal of Intelligent

Material Systems and Structures, 2006. 17(12), p. 1041-1047.

[34] Ronald, N., B. Tiffany, and A.P. Santo, NiTi-Based High-Temperature

Shape-Memory Alloys, in Advanced Structural Materials2006, CRC Press. p.

145-186.

[35] Otsuka, K. and X. Ren, Intermetallics, 1999. 7(5), p. 511-528.

[36] Meng, X., W. Cai, Y. Zheng, and L. Zhao, Materials Science and

Engineering: A, 2006. 438, p. 666-670.

179

[37] Enami, K. and S. Nenno, Metallurgical and Materials Transactions B, 1971.

2(5), p. 1487-1490.

[38] Kainuma, R., K. Ishida, and H. Nakano, Metallurgical and Materials

Transactions A, 1996. 27(12), p. 4153-4162.

[39] Potapov, P. and V. Udovenko, Journal de physique. IV, 1995. 5(8), p. C8.

1059-C8. 1063.

[40] Jee, K., P. Potapov, S. Song, and M. Shin, Scripta Materialia, 1997. 36(2), p.

207-212.

[41] He, Z.-R. and J.-E. Zhou, Materials Science and Engineering: A, 2003.

360(1), p. 183-190.

[42] Karaca, H., I. Karaman, B. Basaran, D. Lagoudas, Y. Chumlyakov, and H.

Maier, Acta materialia, 2007. 55(13), p. 4253-4269.

[43] Omori, T., W. Ito, K. Ando, K. Oikawa, R. Kainuma, and K. Ishida,

Materials transactions, 2006. 47(9), p. 2377.

[44] Omori, T., Y. Sutou, K. Oikawa, R. Kainuma, and K. Ishida, Materials

Science and Engineering: A, 2006. 438, p. 1045-1049.

[45] Omori, T., Y. Sutou, K. Oikawa, R. Kainuma, and K. Ishida, Materials

transactions, 2003. 44(12), p. 2732-2735.

[46] Omori, T., Y. Sutou, K. Oikawa, R. Kainuma, and K. Ishida, Scripta

Materialia, 2005. 52(7), p. 565-569.

[47] Semenova, E.L. and Y.V. Kudryavtsev, Journal of Alloys and Compounds,

1994. 203, p. 165-168.

[48] Koval, Y.N., G. Firstov, and A. Kotko, Scripta metallurgica et materialia,

1992. 27(11), p. 1611-1616.

[49] Firstov, G., Y.N. Koval, and J. Van Humbeeck, Le Journal de Physique IV,

1997. 7(C5), p. C5-549-C5-554.

[50] Koval, Y.N., G. Firstov, J.V. Humbeeck, L. Delaey, and W. Jang, Journal de

physique. IV, 1995. 5(8), p. C8. 1103-C8. 1108.

[51] Waterstrat, R., J. Stalick, X. Meng-Burany, A. Curzon, and M. Estermann,

Scripta metallurgica et materialia, 1995. 33(5), p. 695-703.

[52] Firstov, G., J. Van Humbeeck, and Y.N. Koval, Materials Science and

Engineering: A, 2004. 378(1), p. 2-10.

[53] Firstov, G., J. Van Humbeeck, and Y.N. Koval, Scripta Materialia, 2004.

50(2), p. 243-248.

180

[54] Maeshima, T., S. Ushimaru, K. Yamauchi, and M. Nishida, Materials Science

and Engineering: A, 2006. 438, p. 844-847.

[55] Kim, H.Y., H. Satoru, J.I. Kim, H. Hosoda, and S. Miyazaki, Materials

transactions, 2004. 45(7), p. 2443-2448.

[56] Kim, H., Y. Ikehara, J. Kim, H. Hosoda, and S. Miyazaki, Acta materialia,

2006. 54(9), p. 2419-2429.

[57] Ikeda, M., S.-Y. Komatsu, I. Sowa, and M. Niinomi, Metallurgical and

Materials Transactions A, 2002. 33(3), p. 487-493.

[58] Biggs, T., M. Cortie, M. Witcomb, and L. Cornish, Platinum Metals Review,

2003. 47(4), p. 142-156.

[59] Yamabe-Mitarai, Y., T. Hara, S. Miura, and H. Hosoda, Materials

transactions, 2006. 47(3), p. 650.

[60] Kato, H., R. Stalmans, and J. Van Humbeeck, Materials Transactions,

JIM(Japan), 1998. 39(3), p. 378-386.

[61] Kannarpady, G.K. and A. Bhattacharyya, Materials Science and Engineering:

A, 2006. 438, p. 719-722.

[62] Lelatko, J., N. Morawiec, Y.N. Koval, and V. Kolomyttsev, Metal science

and heat treatment, 1999. 41(8), p. 351-353.

[63] Recarte, V., R. Pérez-Sáez, J. San Juan, E. Bocanegra, and M. Nó,

Metallurgical and Materials Transactions A, 2002. 33(8), p. 2581-2591.

[64] Otsuka, K., H. Sakamoto, and K. Shimizu, Acta Metallurgica, 1979. 27(4), p.

585-601.

[65] Fonda, R., H. Jones, and R. Vandermeer, Scripta Materialia, 1998. 39(8), p.

1031-1037.

[66] He, Z.-r., J.-e. Zhou, and Y. Furuya, Materials Science and Engineering: A,

2003. 348(1), p. 36-40.

[67] Lindquist, P. and C. Wayman, Butterworth-Heinemann, Engineering Aspects

of Shape Memory Alloys(UK), 1990, 1990, p. 58-68.

[68] Suzuki, Y., Y. Xu, S. Morito, K. Otsuka, and K. Mitose, Materials Letters,

1998. 36(1), p. 85-94.

[69] Hosoda, H., M. Tsuji, Y. Takahashi, T. Inamura, K. Wakashima, Y. Yamabe-

Mitarai, S. Miyazaki, and K. Inoue. Phase Stability and Mechanical

Properties of Ti-Ni Shape Memory Alloys Containing Platinum Group

Metals. in Materials Science Forum. 2003. Trans Tech Publ.

[70] Kockar, B., I. Karaman, J. Kim, and Y. Chumlyakov, Scripta Materialia,

2006. 54(12), p. 2203-2208.

181

[71] Dagdelen, F., T. Gokhan, A. Aydogdu, Y. Aydogdu, and O. Adigüzel,

Materials Letters, 2003. 57(5), p. 1079-1085.

[72] Kim, Y. and C. Wayman, Scripta metallurgica et materialia, 1990. 24(2), p.

245-250.

[73] Kainuma, R., H. Ohtani, and K. Ishida, Metallurgical and Materials

Transactions A, 1996. 27(9), p. 2445-2453.

[74] Au, Y. and C. Wayman, Scripta metallurgica, 1972. 6(12), p. 1209-1214.

[75] Kim, S., D. Wee, and M. Oh, Metallurgical and Materials Transactions A,

2003. 34(10), p. 2089-2095.

[76] George, E., C. Liu, J. Horton, C. Sparks, M. Kao, H. Kunsmann, and T. King,

Materials characterization, 1997. 39(2), p. 665-686.

[77] Ishida, K., R. Kainuma, N. Ueno, and T. Nishizawa, Metallurgical

Transactions A, 1991. 22(2), p. 441-446.

[78] Ono, N., A. Tsukahara, R. Kainuma, and K. Ishida, Materials Science and

Engineering: A, 1999. 273, p. 420-424.

[79] Adachi, K. and C. Wayman, Metallurgical Transactions A, 1985. 16(9), p.

1567-1579.

[80] Kainuma, R., M. Ise, K. Ishikawa, I. Ohnuma, and K. Ishida, Journal of

Alloys and Compounds, 1998. 269(1), p. 173-180.

[81] Potapov, P., N. Polyakova, V. Udovenko, and E. Svistunova, Zeitschrift fur

metallkunde, 1996. 87(1), p. 33-39.

[82] Potapov, P., Metal science and heat treatment, 1993. 35(9), p. 520-525.

[83] Yang, W.S., Mikhola, D.E. in Proceedings of the MRS International Meeting

on Advanced Materials. 1992.

[84] He, Z.-r. and J.-e. Zhou, Journal of Alloys and Compounds, 2003. 358(1–2),

p. 245-251.

[85] Sade, M., C. Damiani, R. Gastien, F. Lovey, J. Malarría, and A. Yawny,

Smart Materials and Structures, 2007. 16(1), p. S126.

[86] Tadaki, T., ‘Shape Memory Materials’ (Eds: K. Otsuka, C. M. Wayman),

Cambridge University Press, Cambridge, 1998.

[87] Van Humbeeck, J., M. Chandrasekaran, and L. Delaey, ISIJ International,

1989. 29(5), p. 388-394.

[88] Adachi, K., K. Shoji, and Y. Hamada, ISIJ International, 1989. 29(5), p. 378-

387.

182

[89] Font, J., J. Muntasell, J. Pons, and E. Cesari, Journal of materials research,

1997. 12(09), p. 2288-2297.

[90] Gao, Y., M. Zhu, and J. Lai, Journal of materials science, 1998. 33(14), p.

3579-3584.

[91] Morris, M., Acta Metallurgica et Materialia, 1992. 40(7), p. 1573-1586.

[92] Eucken, S., E. Kobus, and E. Hornbogen, 1991.

[93] Olier, P., J. Brachet, J. Bechade, C. Foucher, and G. Guenin, Journal de

physique. IV, 1995. 5(8), p. C8. 741-C8. 746.

[94] Lelatko, J. and H. Morawiec, Le Journal de Physique IV, 2001. 11(PR8), p.

Pr8-487-Pr8-492.

[95] Lelatko, J., H. Morawiec, Y. Koval, and V. Kolomytsev, Materialoved.,

2000. 6, p. 23-25.

[96] Lelatko, J. and H. Morawiec, Materials chemistry and physics, 2003. 81(2), p.

472-475.

[97] Guilemany, J., J. Fernandez, R. Franch, A. Benedetti, and A. Adorno, Le

Journal de Physique IV, 1995. 5(C2), p. C2-361-C2-365.

[98] Guilemany, J., J. Fernández, R. Franch, A. Benedetti, and A. Adorno, Le

Journal de Physique IV, 1995. 5(C2), p. C2-361-C2-365.

[99] Ma, Y.-q., C.-b. Jiang, L.-F. Deng, and H.-b. Xu, Cailiao Kexue Yu

Jishu(Journal of Materials Science & Technology)(China), 2003. 19, p. 431-

434.

[100] Wasilewski, R., S. Butler, J. Hanlon, and D. Worden, Metallurgical

Transactions, 1971. 2(1), p. 229-238.

[101] Ronald, N., B. Tiffany, and A.P. Santo, NiTi-Based High-Temperature

Shape-Memory Alloys, in Advanced Structural Materials, S.S.T. WO, Editor

2007, Taylor & Francis Group: NW.

[102] Angst, D., P. Thoma, and M. Kao, Journal de physique. IV, 1995. 5(8), p. C8.

747-C8. 752.

[103] Hsieh, S. and S. Wu, Materials characterization, 1998. 41(4), p. 151-162.

[104] Thoma, P.E. and J.J. Boehm, Materials Science and Engineering: A, 1999.

273, p. 385-389.

[105] Han, S., S. Jin, W. Zou, Z. Zhang, and D. Yang, 1995.

[106] Meng, X., Y. Zheng, Z. Wang, and L. Zhao, Scripta Materialia, 2000. 42(4),

p. 341-348.

183

[107] Hsieh, S. and S. Wu, Journal of Alloys and Compounds, 1998. 266(1), p.

276-282.

[108] Wang, Y., Y. Zheng, W. Cai, and L. Zhao, Scripta Materialia, 1999. 40(12),

p. 1327-1331.

[109] Meng, X.-L., Y.-F. Zheng, Z. Wang, and L. Zhao, Materials Letters, 2000.

45(2), p. 128-132.

[110] Sivokha, V. and L. Meisner, Physica B Condensed Matter, 2001. 296, p. 329-

333.

[111] Meng, X., W. Cai, F. Chen, and L. Zhao, Scripta Materialia, 2006. 54(9), p.

1599-1604.

[112] Meng, X., W. Cai, K. Lau, L. Zhao, and L. Zhou, Intermetallics, 2005. 13(2),

p. 197-201.

[113] O. Rios, R. D. Noebe, T. Biles, A. Garg, A. Palczer, D. Scheiman, H. J.

Seifert, and M. Kaufman. Proc. SPIE, 2005, 5761, 376 - 387.

[114] Meisner, L.L. and V.P. Sivokha, Physica B: Condensed Matter, 2004. 344(1–

4), p. 93-98.

[115] Lindquist, P. and C. Wayman, Butterworth-Heinemann, Engineering Aspects

of Shape Memory Alloys(UK), 1990, p. 58-68.

[116] Khachin, V., Revue de Physique Appliquée, 1989. 24(7), p. 733-739.

[117] Donkersloot, H.C. and J.H.N. Van Vucht, Journal of the Less Common

Metals, 1970. 20(2), p. 83-91.

[118] Golberg, D., Y. Xu, Y. Murakami, K. Otsuka, T. Ueki, and H. Horikawa,


[119] Xu, Y., S. Shimizu, Y. Suzuki, K. Otsuka, T. Ueki, and K. Mitose, Acta

materialia, 1997. 45(4), p. 1503-1511.

[120] Otsuka, K., K. Oda, Y. Ueno, M. Piao, T. Ueki, and H. Horikawa, Scripta

Metallurgica et Materialia; Journal Volume: 29; Journal Issue: 10; Other

Information: PBD: 15 Nov 1993, 1993, p. Medium: X; Size: pp. 1355-1358.

[121] Yang, W.S. and D. Mikkola, Scripta metallurgica et materialia, 1993. 28(2),

p. 161-165.

[122] Tian, Q., J. Wu, and Y. Cheng, Material Science and Technology, 2003. 19,

p. 179.

[123] Atli, K., I. Karaman, R. Noebe, A. Garg, Y. Chumlyakov, and I. Kireeva,


184

[124] Shimizu, S., Y. Xu, E. Okunishi, S. Tanaka, K. Otsuka, and K. Mitose,


[125] Noebe, R.D., S.A. Padula, G.S. Bigelow, O. Rios, A. Garg, and B. Lerch.

Proc. SPIE, 2006, 6170, 617010.

[126] Shirakawa, Y., Y. Morizono, and M. Nishida. New precipitate phase in Pd

and Ni rich Ti-Pd-Ni shape memory alloys. in Materials science forum. 2000.

Trans Tech Publ.

[127] Nagasako, M., M. Nishida, Y. Murakami, and D. Shindo, Materials Science

and Engineering: A, 2006. 438, p. 848-851.

[128] Golberg, D., Y. Xu, Y. Murakami, S. Morito, K. Otsuka, T. Ueki, and H.

Horikawa, Intermetallics, 1995. 3(1), p. 35-46.

[129] Cai, W., K. Otsuka, and M. Asai, Materials transactions-JIM, 1999. 40(9), p.

895-898.

[130] Cai, W., S. Tanaka, and K. Otsuka. Thermal cyclic characteristics under load

in a Ti50. 6Pd30Ni19. 4 alloy. in Materials science forum. 2000. Trans Tech

Publ.

[131] Cai, W. and K. Otsuka, Journal of Materials Science and Technology, 2001.

17, p. 359.

[132] Bigelow, G.S., S.A. Padula II, A. Garg, D. Gaydosh, and R.D. Noebe,


[133] S. A. Padula, G.S. Bigelow, R.D. Noebe, D. Gaydosh, and A. Garg.

'Challenges and progress in the develoment of high-temperature shape

memory alloys based on NiTiX compositions for high-force actuator

applications' in Proc. Int. conf. on 'shape memory and superelastic

technologies', 787, 2008. Materials Park, OH, ASM International.

[134] Kumar, P.K. and D.C. Lagoudas, Acta materialia, 2010. 58(5), p. 1618-1628.

[135] G. S. Bigelow, R.D. Noebe, S.A. Padula, and A. Garg. 'Develoment and

characterization of improved high-temperature shape memory alloys by solid

solution strengthening and thermomechanical processing of NiTiPd alloys' in

Proc. Int. conf. on 'shape memory and superelastic technologies', 113, 2008.

Materials Park, OH, ASM International.

[136] Otsuka, K., K. Oda, Y. Ueno, M. Piao, T. Ueki, and H. Horikawa, Scripta

metallurgica et materialia, 1993. 29(10), p. 1355-1358.

[137] Suzuki, Y., ‘Shape Memory Materials’ (Eds: K. Otsuka, C. M. Wayman),

Cambridge University Press, Cambridge 1998, pp. 133 – 148 1998.

[138] K. Yamauchi, I.O., K. Tsuchiya and S. Miyazaki, Shape memory and

superelastic alloys: Technologies and Application2011: Woodhead

Publishing Limited,. pp 53 - 62.

185

[139] Kumar, P.K., U. Desai, J.A. Monroe, D.C. Lagoudas, I. Karaman, G.

Bigelow, and R.D. Noebe, Materials Science and Engineering: A, 2011. 530,

p. 117-127.

[140] Khan, M.I., H.Y. Kim, T.-h. Nam, and S. Miyazaki, Acta materialia, 2012.

60(16), p. 5900-5913.

[141] Hiroaki Okamoto, P.R.S., Linda Kacprzak, Binary alloy phase diagrams,, ed.

T B Massalski; H Okamoto1990: Materials Park, OH ASM International,. pp.

2874–2876.

[142] Atli, K., I. Karaman, R. Noebe, A. Garg, Y. Chumlyakov, and I. Kireeva,

Acta materialia, 2011. 59(12), p. 4747.

[143] Ramaiah, K., C. Saikrishna, J. Bhagyaraj, and S. Bhaumik, Intermetallics,

2013. 40, p. 10-18.

[144] J. I. Goldstein, D.E.N., P. Echlin, D. C. Joy, C. Fiori, E. Lifshin, Scanning

electron microscopy and X-ray microanalysis, 1981: Plenum Press, New

York.

[145] Rehman, S.u., M. Khan, A. Nusair Khan, L. Ali, S. Zaman, M. Waseem, and

S.H.I. Jaffery, Materials Science and Engineering: A, 2014. 619(0), p. 171-

179.

[146] Mercier, O. and K.N. Melton, Metallurgical and Materials Transactions A,

1979. 10(3), p. 387-389.

[147] Suburi, T., ‘Shape Memory Materials’ (Eds: K. Otsuka, C. M. Wayman),

Cambridge University Press, Cambridge, 1998.

[148] Lin, K.-N., S.-K. Wu, and L.-M. Wu, Materials transactions, 2009. 50(10), p.

2384-2390.

[149] Fuentes, J.G., P. Gumpel, and J. Strittmatter, Advanced engineering

materials, 2002. 4(7), p. 437-452.

[150] Fuentes, J.M.G., P. Gumpel, and J. Strittmatter, Advanced engineering

materials, 2002. 4(7), p. 437-452.

[151] Pelton, A., Journal of materials engineering and performance, 2011. 20(4-5),

p. 613-617.

[152] Pelton, A., G. Huang, P. Moine, and R. Sinclair, Materials Science and

Engineering: A, 2012. 532, p. 130-138.

[153] Rehman, S.u., M. Khan, A. Nusair Khan, M. Imran Khan, L. Ali, and S.H.

Imran Jaffery, Journal of Alloys and Compounds, 2014. 616(0), p. 275-283.

[154] Xu, Y., S. Shimizu, Y. Suzuki, K. Otsuka, T. Ueki, and K. Mitose, Acta

materialia, 1997. 45(4), p. 1503-1511.

186

[155] Karaca, H., E. Acar, B. Basaran, R. Noebe, G. Bigelow, A. Garg, F. Yang, M.

Mills, and Y. Chumlyakov, Scripta Materialia, 2012. 67(7), p. 728-731.

Development, Characterization and Testing of Nickel ...

Documents

Transcript of Development, Characterization and Testing of Nickel ...