Material characterization of next generation shape memory alloys (Cu-Al-Mn, Ni-Ti-Co and Fe-Mn-Si) for use in bridges in seismic regions

PDF

Download

Open document

Flip pages

Summarize

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Material Characterization of Next Generation Shape Memory Alloys
(Cu-Al-Mn, Ni-Ti-Co and Fe-Mn-Si) for Use in
Bridges in Seismic Regions
by
Huanpeng Hong
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(CIVIL ENGINEERING)
May 2024
Copyright 2024 Huanpeng Hong

ii
Acknowledgements
First of all, I would like to express my sincere gratitude to my advisor Dr. Bora Gencturk for
his meticulous and unreserved guidance, patience, and help during my Ph.D. studies. The guidance
and insights from Dr. Gencturk during the course of my studies were invaluable and will be a
priceless treasure throughout my life. I am also sincerely grateful to Dr. M. Saiid Saiidi for his
guidance, support, and encouragement throughout the research projects. Dr. Saiidi’s instructions
in my studies will benefit me throughout my future career.
Deep thanks are given to Dr. Yoshikazu Araki from Kyoto University, Japan, and Dr. Sumio
Kise from Furukawa Techno Material Co. Ltd. I would like to sincerely thank them for their
valuable time, help and advice during my research process.
I also would like to thank my defense committee members: Dr. Iraj Ershaghi, Dr. Qiming
Wang, and Dr. Chukwuebuka Nweke for their precious time and valuable insights on my research.
I am very grateful to all the members of the Structures and Materials Research Laboratory
(SMRL): Dr. Botong Zheng, Dr. Xiaoying Pan, Dr. Hadi Aryan, Dr. Zhidong Zhang, Mr. Fidel
Hurtado, and Mr. Juan Tuchan, as well as the staff of the Viterbi/Dornsife Machine Shop: Dr. Seth
Wieman, Mr. Alex Flores, and Mr. Scott Collom. Without their professional help, I would not be
able to complete my experimental work smoothly. Some of the testing activities presented in this
dissertation were performed by other researchers including Dr. Hadi Aryan, Dr. Amit Jain, Dr.
Farshid Hosseini, and Ms. Susan Alexis Brown. Their contributions are greatly appreciated. I

iii
would also like to thank my colleagues: Bozhou Zhuang, Anna Arcaro, and Yang Chu for their
help in my work.
The author thanks Dr. Fumiyoshi Yamashita from Furukawa Techno Material, Co., Ltd. for
the help with the design of the electrochemical experiments. The author acknowledges Dr. Daichi
Minami of Kanagawa Institute of Industrial Science and Technology (KISTEC) for the help in the
Electron Back-Scattered Diffraction (EBSD) analysis.
The author would like to thank Furukawa Techno Material Co., LTD. for providing the CuAl-Mn and Ni-Ti materials, SAES Smart Materials for providing the Ni-Ti-Co and Ni-Ti materials,
and Re-fer AG. for providing the Fe-Mn-Si material. The author acknowledges the Headed
Reinforcement Corp. (HRC) for performing the bar heading and providing necessary fixtures for
testing. The stainless and epoxy coated steels were donated by the Contractors Material Co. in
Cincinnatio, OH, and the high chromium steel was donated by the MMFX Steel Corp. of America
in Irvine, CA.
The research presented in this dissertation was in part funded by the National Cooperative
Highway Research Program (NCHRP) under Project NCHRP IDEA 210, in part by the United
States National Science Foundation (NSF) under the award No. 1642488, and in part by the
Washington State Department of Transportation (WSDOT) under the Project TPF-5(491) and the
Federal Highway Administration. Any opinions, findings, conclusions, or recommendations are
those of the author and do not necessarily reflect the views of the funding agencies.

iv
Table of Contents
Acknowledgements.................................................................................................................... ii
List of Tables..............................................................................................................................x
List of Figures........................................................................................................................... xi
Abstract ........................................................................................................................... xxiii
Chapter 1 - Objectives and scope.................................................................................................1
1.1 Research objectives.............................................................................................1
1.2 Research scope....................................................................................................4
1.3 Organization of the dissertation...........................................................................9
Chapter 2 - Background information .........................................................................................12
2.1 Overview of shape memory alloy (SMA) ..........................................................12
2.2 Ni-Ti-based SMA..............................................................................................14
2.3 Cu-based SMA..................................................................................................16
2.4 Fe-based SMA ..................................................................................................20
2.5 Summary...........................................................................................................24
Chapter 3 - Corrosion behavior of Cu-Al-Mn SMA...................................................................26
3.1 Research motivation..........................................................................................26
3.2 Experimental program.......................................................................................29
3.2.1 Long-term corrosion tests ..........................................................................29
3.2.2 Electrochemical corrosion tests..................................................................37
3.3 Results and discussion of long-term corrosion tests...........................................39
3.3.1 Corrosion surface condition .......................................................................39
3.3.2 Mass loss analyses.....................................................................................45
3.3.3 Mechanical test results of SCB ..................................................................47

v
3.3.4 Mechanical test results of PCP and PCP-C.................................................52
3.3.5 Mechanical test results of steel rebar..........................................................59
3.4 Results and discussion of electrochemical corrosion tests..................................70
3.4.1 Tafel curves...............................................................................................70
3.4.2 Corrosion rate............................................................................................73
3.5 Summary of findings.........................................................................................75
Chapter 4 - Low-cycle fatigue behavior of Cu-Al-Mn SMA at different temperatures ...............77
4.1 Research motivation..........................................................................................77
4.2 Materials and methods.......................................................................................79
4.2.1 Materials....................................................................................................79
4.2.2 Methods.....................................................................................................82
4.3 Results of SCB..................................................................................................87
4.3.1 SCB at room temperature, 25 ℃................................................................87
4.3.2 SCB at -40 ℃............................................................................................91
4.3.3 SCB at 50 ℃ .............................................................................................94
4.4 Results of PCP ..................................................................................................97
4.5 Results of NTB ...............................................................................................100
4.5.1 NTB at room temperature, 25 °C .............................................................100
4.5.2 NTB at -40 ℃ and -10 ℃........................................................................103
4.5.3 NTB at 50 ℃...........................................................................................106
4.6 Summary of findings.......................................................................................109
Chapter 5 - Machinability characteristics of Cu-Al-Mn SMA..................................................110
5.1 Research motivation........................................................................................110
5.2 Experimental program.....................................................................................112
5.2.1 Vickers hardness......................................................................................113
5.2.2 Machinability tests...................................................................................114
5.3 Results and discussion of single machining tests .............................................117
5.3.1 Chip formation.........................................................................................117

vi
5.3.2 Tool wear ................................................................................................123
5.3.3 Surface roughness....................................................................................130
5.4 Results and discussion of continuous machining tests......................................132
5.4.1 Tool wear ................................................................................................133
5.4.2 Surface roughness and dimensional variation...........................................134
5.5 Summary of findings.......................................................................................136
Chapter 6 - Headed coupling behavior of Cu-Al-Mn SMA......................................................138
6.1 Research motivation........................................................................................138
6.2 Experimental program.....................................................................................142
6.2.1 Specimen preparation ..............................................................................142
6.2.2 Test methods............................................................................................144
6.3 Results of mechanical tests..............................................................................147
6.4 Results of microstructural analyses..................................................................151
6.4.1 Electron backscatter diffraction (EBSD) ..................................................151
6.4.2 Metallographic analysis along bar length .................................................153
6.4.3 Fractography............................................................................................157
6.4.4 Metallographic analysis on headed ends...................................................160
6.4.5 Comparison with previous studies............................................................164
6.5 Summary of findings.......................................................................................166
Chapter 7 - Cost estimation of bridge columns reinforced with Cu-Al-Mn SMA .....................168
7.1 Research motivation........................................................................................168
7.2 Design of bridge columns................................................................................169
7.2.1 Material and geometric properties............................................................169
7.2.2 Capacity check.........................................................................................171
7.3 Analysis of bridge columns.............................................................................172
7.3.1 Moment-curvature analysis......................................................................172
7.3.2 Drift capacity...........................................................................................175
7.4 Cost estimation of typical columns..................................................................177

vii
7.5 Summary of findings.......................................................................................184
Chapter 8 - Effect of temperature on superelasticity and ductility of Ni-Ti-Co SMA ...............186
8.1 Research motivation........................................................................................186
8.2 Material and methods......................................................................................187
8.2.1 Materials..................................................................................................187
8.2.2 Methods...................................................................................................189
8.3 Results and discussion.....................................................................................191
8.3.1 Ni-Ti-Co SMA.........................................................................................191
8.3.2 Ni-Ti SMA ..............................................................................................194
8.3.3 Cu-Al-Mn SMA.......................................................................................196
8.3.4 Comparison of Ni-Ti-Co, Ni-Ti and Cu-Al-Mn SMA ..............................199
8.4 Summary of findings.......................................................................................201
Chapter 9 - Low-cycle fatigue behavior of Ni-Ti-Co SMA at different temperatures...............202
9.1 Research motivation........................................................................................202
9.2 Materials and methods.....................................................................................203
9.2.1 Materials..................................................................................................203
9.2.2 Methods...................................................................................................203
9.3 Results and discussion.....................................................................................205
9.3.1 Ni-Ti-Co and Ni-Ti at room temperature, 23 ℃.......................................205
9.3.2 Ni-Ti-Co at -40 ℃ and Ni-Ti at -0 ℃ ......................................................210
9.3.3 Ni-Ti-Co and Ni-Ti at 50 ℃ ....................................................................213
9.4 Summary of findings.......................................................................................216
Chapter 10 - Moment-curvature analysis of bridge columns reinforced with Ni-Ti-Co SMA...218
10.1 Research motivation......................................................................................218
10.2 Modeling methods.........................................................................................219
10.2.1 RC columns...........................................................................................219
10.2.2 SMA columns........................................................................................221
10.3 Methodology.................................................................................................224

viii
10.4 Results and discussion...................................................................................228
10.4.1 Sections with different diameters...........................................................228
10.4.2 Sections with different reinforcement ratios...........................................231
10.4.3 Sections with different axial force ratios ................................................232
10.5 Summary of findings.....................................................................................234
Chapter 11 - Effect of temperature on ductility and recovery strain of Fe-Mn-Si SMA............236
11.2 Research motivation......................................................................................236
11.3 Materials and methods...................................................................................237
11.3.1 Materials................................................................................................237
11.3.2 Methods.................................................................................................238
11.4 Results ..........................................................................................................239
11.5 Discussion.....................................................................................................242
11.6 Summary of findings.....................................................................................244
Chapter 12 - Cyclic actuation behavior of Fe-Mn-Si SMA for use in self-centering columns...246
12.1 Research motivation......................................................................................246
12.2 Experimental program...................................................................................247
12.2.1 Test setup ..............................................................................................247
12.2.2 Test methods..........................................................................................248
12.3 Results and discussion...................................................................................252
12.3.1 Effect of post-actuation temperature.......................................................252
12.3.2 Effect of prestrain level..........................................................................253
12.3.3 Low-cycle fatigue resistance ..................................................................256
12.4 Summary of findings.....................................................................................261
Chapter 13 - Summary and conclusions...................................................................................263
13.1 Summary of completed research....................................................................263
13.1.1 Research on Cu-Al-Mn SMA.................................................................263
13.1.2 Research on Ni-Ti-Co SMA...................................................................266
13.1.3 Research on Fe-Mn-Si SMA ..................................................................267

ix
13.2 Conclusions...................................................................................................268
13.2.1 Conclusions on Cu-Al-Mn SMA............................................................268
13.2.2 Conclusions on Ni-Ti-Co SMA..............................................................272
13.2.3 Conclusions on Fe-Mn-Si SMA .............................................................274
13.3 Recommendations for future research............................................................276
References .............................................................................................................................278
Appendix A - Long-term corrosion test results........................................................................295
A.1 Single crystal Cu-Al-Mn SMA bar (SCB) ......................................................296
A.2 Polycrystal Cu-Al-Mn SMA plate (PCP)........................................................297
A.3 Mild steel (MS)..............................................................................................298
A.4 High chromium steel (XS)..............................................................................301
A.5 Epoxy coated steel (ES)..................................................................................304
A.6 Stainless steel (SS) .........................................................................................307
Appendix B - Low-cycle fatigue test results of Cu-Al-Mn and Ni-Ti SMAs............................310
B.1 SCB at room temperature, 25 ℃.....................................................................311
B.2 SCB at -40 ℃.................................................................................................315
B.3 SCB at 50 ℃..................................................................................................316
B.4 PCP at room temperature, 25 ℃ .....................................................................317
B.5 NTB at room temperature, 25 ℃ ....................................................................320
B.6 NTB at -10 ℃ ................................................................................................321
B.7 NTB at 50 ℃..................................................................................................322

x
List of Tables
Table 3-1 Chemical composition of steel rebar..........................................................................32
Table 3-2 Summary of materials and dimensions of samples used in long-term corrosion
tests . ........................................................................................................................33
Table 3-3 Summary of materials and dimensions of samples used in electrochemical
corrosion tests...........................................................................................................38
Table 4-1 Summary of low-cycle fatigue tests on CAM and NiTi SMAs...................................86
Table 6-1 Properties extracted from mechanical tests. .............................................................151
Table 7-1 Summary of the cost estimation results....................................................................182
Table 9-1 Summary of low-cycle fatigue tests on Ni-Ti-Co SMA and Ni-Ti SMAs.................205
Table 10-1 Details of properties used to model SMA bars. ......................................................223
Table 12-1 Test matrix of cyclic actuation tests on Fe-Mn-Si SMA.........................................251
Table A-1 Mass loss and mechanical behavior degradation of single crystal Cu-Al-Mn
SMA bar (SCB) during long-term corrosion tests....................................................296
Table A-2 Mass loss and mechanical behavior degradation of polycrystal Cu-Al-Mn SMA
plate (PCP) during long-term corrosion tests...........................................................297
Table A-3 Mass loss of mild steel (MS) during long-term corrosion tests. ...............................298
Table A-4 Mechanical property degradation of mild steel (MS) during long-term corrosion
tests. .......................................................................................................................299
Table A-5 Mass loss of high chromium steel (XS) during long-term corrosion tests. ...............301
Table A-6 Mechanical property degradation of high chromium steel (XS) during long-term
corrosion tests.........................................................................................................302
Table A-7 Mass loss of epoxy coated steel (ES) during long-term corrosion tests....................304
Table A-8 Mechanical property degradation of epoxy coated steel (ES) during long-term
corrosion tests.........................................................................................................305
Table A-9 Mass loss of stainless steel (SS) during long-term corrosion tests. ..........................307
Table A-10 Mechanical property degradation of stainless steel (SS) during long-term
corrosion tests.........................................................................................................308

xi
List of Figures
Fig. 1-1 Typical plastic hinge damage of conventional reinforced concrete (RC) bridge
columns [4]..................................................................................................................2
Fig. 1-2 (a) Schematic response of a bridge column with and without SMA bars, and (b)
arrangement of SMAs in a bridge column [10]. Note: SEA is referring to SMA
with superelastic effect at room temperature. ...............................................................3
Fig. 1-3 A schematic diagram of research activities conducted in this dissertation.......................6
Fig. 2-1 Schematic diagram of (a) superelastic effect, and (b) shape memory effect in
SMAs. .......................................................................................................................13
Fig. 2-2 Martensitic and austenitic phase transformation in SMAs [42].. ...................................14
Fig. 2-3 Effect of Al content on cold workability, tensile elongation and shape recovery of
Cu-Al-Mn alloys [39]. ...............................................................................................17
Fig. 2-4 (a) AGG process, (b) optical micrograph of a Cu-Al-Mn alloy sheet before and
after AGG, and (c) crystal orientation of a Cu-Al-Mn sheet after AGG [59]...............18
Fig. 2-5 (a) Effect of relative grain size on maximum superelastic strain of Cu-Al-Mn SMA
[58], and (b) effect of crystal orientation on stress-strain curves of Cu-Al-Mn
SMA [61]. .................................................................................................................19
Fig. 2-6 Phase transformation in (a) Ni-Ti-based (or Cu-based) SMA, and (b) Fe-based
SMA..........................................................................................................................21
Fig. 2-7 Schematic diagram of the actuation process of Fe-Mn-Si SMA. ...................................22
Fig. 2-8 Images of some typical applications using Fe-Mn-Si SMA to strengthen or repair
existing concrete structures: (a) actively confining of a concrete column [65], (b)
shear strengthening of a concrete beam [66], and (c) flexural strengthening of
concrete slabs [45].....................................................................................................22
Fig. 2-9 Schematic diagrams of Fe-Mn-Si SMA application in: (a) flexural strengthening
beams, (b) shear strengthening beams, (c) active confining columns, and (d) selfcentering columns......................................................................................................23

xii
Fig. 3-1 Preparation of test specimens for long-term corrosion tests: (a) dimensions of SCB
and PCP, (b) machined SCB, (c) surface grain boundaries of as received
polycrystal CAM SMA, and (d) preparation of PCP-i and PCP-iC. ............................30
Fig. 3-2 Surface grain distribution of as received polycrystal CAM SMA rods: (a) PC-1
rod, (b) PC-2 rod, (c) PC-3 rod, (d) PC-4 rod, (e) PC-5 rod........................................31
Fig. 3-3 Arrangement of specimens in corrosion chamber for long-term salt spray
corrosion....................................................................................................................34
Fig. 3-4 Timeline of long-term corrosion and mechanical tests..................................................35
Fig. 3-5 Setup for electrochemical corrosion tests......................................................................39
Fig. 3-6 Surface conditions of CAM SMA specimens prior to cleaning: (a) SCB after 20
days of exposure, (b) PCP after 20 days of exposure, (c) SCB after 75 days of
exposure, and (d) PCP after 75 days of exposure........................................................40
Fig. 3-7 Corrosion surface of two CAM SMA specimens at different ages after cleaning:
(a) SCB-2, and (b) PCP-2. Number inside the parenthesis indicates the
corresponding mass loss percentage...........................................................................41
Fig. 3-8 Two CAM specimens that broke naturally at 1051 days. ..............................................43
Fig. 3-9 Surface conditions of four types of steel rebar after 296 days of corrosion after
cleaning: (a) MS, (b) XS, (c) ES, and (d) SS. .............................................................44
Fig. 3-10 Mass loss of six materials during long-term corrosion tests: (a) MS, (b) SS, (c)
ES, (d) SS, (e) PCP, and (f) SCB. ..............................................................................46
Fig. 3-11 Ranked normalized mass loss of PCP, SCB, MS, XS, ES, and SS after 296 days
of exposure................................................................................................................47
Fig. 3-12 Mechanical test results of SCB at different levels of corrosion: (a) SCB-1, (b)
SCB-2, and (c) SCB-3. ..............................................................................................48
Fig. 3-13 Definition of critical mechanical properties of superelastic CAM SMA......................49
Fig. 3-14 Variation of mechanical properties of SCB at different levels of corrosion: (a)
Ms, norm, (b) Af, norm, (c) hy, norm and (d) re, norm. ........................................................50

xiii
Fig. 3-15 Full range of stress-strain curves after 1051 days of exposure: (a) SCB-2, and (b)
SCB-3........................................................................................................................51
Fig. 3-16 Failure tests of SCB after 1051 days of exposure: (a) SCB-2, and (b) SCB-3..............51
Fig. 3-17 Mechanical test results of PCP at different levels of corrosion: (a) PCP-1, (b)
PCP-2, (c) PCP-3, (d) PCP-4, and (e) PCP-5..............................................................53
Fig. 3-18 Mechanical properties of PCP at different levels of corrosion: (a) Ms, norm, (b) Af,
norm, (c) hy, norm, and (d) re, norm..................................................................................54
Fig. 3-19 Full range stress-strain curves after 1051 days of exposure: (a) PCP-1, (b) PCP-2,
(c) PCP-4 and (d) PCP-5............................................................................................55
Fig. 3-20 Failure of PCP and PCP-C after 1051 corrosion exposure: (a) Exposed PCP-4,
(b) Unexposed PCP-4C, (c) Exposed PCP-5, and (d) Unexposed PCP-5C..................56
Fig. 3-21 Full range of stress-strain curves of unexposed samples: (a) PCP-1C, (b) PCP2C, (c) PCP-3C, (d) PCP-4C, and (e) PCP-5C............................................................58
Fig. 3-22 Compassion of rupture strain and plastic stress between PCP and PCP-C: (a)
Rupture strain, and (b) Ultimate stress. ......................................................................59
Fig. 3-23 Stress-strain curves of MS and XS at predetermined levels of corrosion: (a)
#3MS, (b) #5MS, (c) #10MS, (d) #3XS, (e) #5XS, and (f) #10XS. Note: numbers
inside parenthesis indicate the mass loss percentage...................................................60
Fig. 3-24 Stress-strain curves of ES and SS at predetermined levels of corrosion: (a) #3ES,
(b) #5ES, (c) #10ES, (d) #3SS, (e) #5SS, and (f) #10SS. Note: numbers inside
parenthesis indicate the mass loss percentage.............................................................62
Fig. 3-25 Mechanical properties of MS at different levels of corrosion: (a) fy, norm, (b) fm,
norm, (c) E, norm, and (d) r, norm......................................................................................65
Fig. 3-26 Mechanical properties of XS at different levels of corrosion: (a) fy, norm, (b) fm,
norm, (c) E, norm, and (d) r, norm. .....................................................................................67
Fig. 3-27 Mechanical properties of ES at different levels of corrosion: (a) fy, norm, (b) fm, norm,
(c) E, norm, and (d) r, norm.............................................................................................68

xiv
Fig. 3-28 Mechanical properties of SS at different levels of corrosion: (a) fy, norm, (b) fm, norm,
(c) E, norm, and (d) r, norm. ............................................................................................70
Fig. 3-29 Potentiodynamic polarization curves of MS, XS, ES, SS, CAM and NiTi...................71
Fig. 3-30 Corrosion potential of MS, XS, ES, SS, CAM and NiTi. ............................................72
Fig. 3-31 Corrosion current density of MS, XS, ES, SS, CAM and NiTi: (a) corrosion
current density, icorr, (b) corrosion current density, icorr within the range from 0-
0.5 A/cm2................................................................................................................73
Fig. 3-32 Corrosion rate of MS, XS, ES, SS, CAM and NiTi.....................................................74
Fig. 4-1 Transformation temperatures of SMA used in this study: (a) CAM, and (b) NiTi
SMA..........................................................................................................................80
Fig. 4-2 Dimensions of: (a) SCB, NTB, and (b) PCP specimens. ...............................................81
Fig. 4-3 Optical microscope image of polycrystal NiTi SMA. ...................................................81
Fig. 4-4 Setup for low-cycle fatigue test: (a) setup for SCB and NTB at room temperature,
(b) zoom-in view of customized grip fixture for SCB and NTB, (c) zoom-in view
of grip fixture for PCP, (d) inside view of environmental chamber and setup for
PCP, and (d) outside view of environmental chamber connected with liquid
nitrogen tank..............................................................................................................82
Fig. 4-5. Schematic diagram of grip fixture used for SCB, NTB: (a) picture of setup, (b)
front view, (c) section A-A, and (d) section B-B. .......................................................83
Fig. 4-6. Schematic diagram of grip fixture used for PCP: (a) picture of setup, (b) front
view, and (c) section A-A. .........................................................................................84
Fig. 4-7 Definition of mechanical properties considered in low-cycle fatigue test......................87
Fig. 4-8 Stress-strain curves of SCB at room temperature, 25 °C: (a) SCB-25C-1, and (b)
SCB-25C-2................................................................................................................88
Fig. 4-9 Variation in mechanical properties of SCB at room temperature, 25 °C: (a) SCB25C-1, and (b) SCB-25C-2. .......................................................................................90
Fig. 4-10 Stress-strain curves of SCB at -40 °C: (a) SCB-m40C-1, and (b) SCB-m40C-2..........92

xv
Fig. 4-11 Variation in mechanical properties of SCB at -40 °C: (a) SCB-m40C-1, and (b)
SCB-m40C-2.............................................................................................................93
Fig. 4-12 Stress-strain curves of SCB at 50 °C: (a) SCB-50C-1, and (b) SCB-50C-2. ................95
Fig. 4-13 Variation in mechanical properties of SCB at 50 °C: (a) SCB-50C-1, and (b)
SCB-50C-2................................................................................................................96
Fig. 4-14 Stress-strain curves of PCP at room temperature, 25 °C: (a) PCP-25C-1, and (b)
PCP-25C-2. ...............................................................................................................98
Fig. 4-15 Variation in mechanical properties of PCP at room temperature, 25 °C: (a) PCP25C-1, and (b) PCP-25C-2.........................................................................................99
Fig. 4-16 Grain distribution and fracture locations of PCP: (a) PCP-25C-1, (b) PCP-25C-3,
(c) PCP-25C-4, and (d) PCP-25C-5. The blue bracket indicates the gauge length
of extensometer. The red arrow indicates the fracture location. ................................100
Fig. 4-17 Stress-strain curves of NTB at room temperature, 25 °C: (a) NTB-25C-1, and (b)
NTB-25C-2. ............................................................................................................101
Fig. 4-18 Stress-strain curves comparison of NTB-25C-1 and results reported by Kang et
al. [113]...................................................................................................................102
Fig. 4-19 Variation in mechanical properties of NTB at room temperature, 25 °C: (a) NTB25C-1, and (b) NTB-25C-2. .....................................................................................103
Fig. 4-20 Stress-strain curves at typical cycles of NTB-m40C. ................................................104
Fig. 4-21 Stress-strain curves at typical cycles of NTB at -10 °C: (a) NTB-m10C-1, and (b)
NTB-m10C-2. .........................................................................................................105
Fig. 4-22 Variation in mechanical properties of NTB at -10 °C: (a) NTB-m10C-1, and (b)
NTB-m10C-2. .........................................................................................................106
Fig. 4-23. Stress-strain curves of NTB at 50 °C: (a) NTB-50C-1, and (b) NTB-50C-2.............107
Fig. 4-24 Variation in mechanical properties of NTB at 50 °C: (a) NTB-50C-1, and (b)
NTB-50C-2. ............................................................................................................108

xvi
Fig. 5-1 Vickers hardness of Cu-Al-Mn SMA (CAM) in comparison with other commonly
used materials: Ni-Ti SMA (NiTi), 304 stainless steel (SS), mild steel (MS),
brass, aluminum, and copper....................................................................................113
Fig. 5-2 (a) Machinability test setup, and (b) definition of key parameters in single point
turning tests. ............................................................................................................115
Fig. 5-3 Test matrix of machinability tests: (a) single, and (b) continuous machining tests.......117
Fig. 5-4 Images and classification of chips generated when machining CAM, MS and SS at
different Vc, fr, and dc...............................................................................................119
Fig. 5-5 Images and classification of chips generated when machining NiTi at different Vc,
fr, and dc. .................................................................................................................121
Fig. 5-6 Comparison of cutting temperature when machining CAM and NiTi at Vc = 60
m/min......................................................................................................................122
Fig. 5-7 Typical tool wear patterns in single point turning: (a) Typical wear patterns on an
insert, and (b) Flank wear characteristics according to ISO-3685 [141]....................124
Fig. 5-8 Nose and average flank wear after machining CAM, MS, SS and NiTi at varying
Vc and dc: (a) Nose wear after machining CAM, MS and SS, (b) Nose wear after
machining NiTi, (c) Average flank wear after machining CAM, MS and SS, and
(d) Average flank wear after machining NiTi...........................................................126
Fig. 5-9 Typical images of tool wear after machining: (a) CAM, (b) MS, (c) SS, and (d)
NiTi at dc =1.5 mm, Vc =60 m/min, and fr =0.1 mm/rev. ..........................................127
Fig. 5-10 Nose and average flank wear after machining CAM, MS, SS and NiTi at varying
Vc and fr: (a) Nose wear after machining CAM, MS and SS, (b) Nose wear after
machining NiTi, (c) Average flank wear after machining CAM, MS and SS, and
(d) Average flank wear after machining NiTi...........................................................129
Fig. 5-11 Typical images of tool wear after machining: (a) CAM, (b) MS, (c) SS, and (d)
NiTi at fr =0.2 mm/rev, Vc =60 m/min and dc =0.5 mm. ...........................................130
Fig. 5-12 Surface roughness (arithmetical average roughness Ra and maximum peak to
valley height Rt) of machined CAM, MS and SS at varying Vc, dc and fr:: (a) Ra at

xvii
varying Vc and dc, (b) Rt at varying Vc and dc, (c) Ra at varying Vc and fr, and (d)
Rt at varying Vc and fr...............................................................................................132
Fig. 5-13 Images of tool wear after continuous machining tests on: (a) MS and (b) CAM........134
Fig. 5-14 Surface roughness (maximum peak to valley height Rt) of CAM and MS in
continuous machining tests. .....................................................................................135
Fig. 5-15 Relative diameter difference of CAM and MS in continuous machining tests. ..........136
Fig. 6-1 Illustration of headed mechanical couplers. ................................................................140
Fig. 6-2 Dimensions of headed specimens: (a) 30 mm diameter, and (b) 20 mm diameter
CAM SMAs. ...........................................................................................................142
Fig. 6-3 Heading process: (a) heating the specimen with a blowtorch, and (b) heading the
specimen using a mechanical extruder. ....................................................................143
Fig. 6-4 Surface temperature of headed end of five CAM SMAs after removing from
extruder. RT: room temperature...............................................................................144
Fig. 6-5 Test setup for mechanical property characterization: (a) LP-1 and LP-2 and (b)
LP-3. .......................................................................................................................145
Fig. 6-6 Mechanical test results: (a) S30-1 during LP-1, (b) S20-1 during LP-1, (c) S20-2
during LP-1, (d) S20-3 during LP-1 and LP-2, (e) S20-4 during LP 1 and LP 2,
(f) S20-3 and S20-4 connected in series during LP-3, and (g) schematic diagram
of three loading protocols (LP). ...............................................................................148
Fig. 6-7 Photographs of fractured samples after mechanical tests: (a) S30-1, (b) S20-1, (c)
S20-2, and (d) S20-3................................................................................................149
Fig. 6-8 Schematical diagram of an ideal stress-strain curve of single crystal CAM SMAs.
Stage I: elastic deformation of Austenitic phase (A), Stage II: transformation
from A to Martensitic phase (M), Stage III: elastic deformation of M, Stage IV:
plastic deformation. .................................................................................................150
Fig. 6-9 Transformation strains of S30-1, S20-1, S20-2 and S30-3 measured from EBSD. ......152
Fig. 6-10 Microstructural images at different distances along the length of S30-1 from the
fracture. ...................................................................................................................155

xviii
Fig. 6-11 Microstructural images at different distances along the length of S20-3 from the
fracture. ...................................................................................................................156
Fig. 6-12 Fracture surface of S30-1. ........................................................................................158
Fig. 6-13 Fracture surface of S20-1. ........................................................................................159
Fig. 6-14 Fracture surface of S20-2. ........................................................................................159
Fig. 6-15 Fracture surface of S20-3. ........................................................................................160
Fig. 6-16 Microstructures of S20-2 (with brittle fracture) in the headed end. ...........................161
Fig. 6-17 Microstructures of S20-1 (with brittle fracture) in the headed end. ...........................163
Fig. 6-18 Microstructures of S20-3 (with ductile fracture) in the headed end. ..........................164
Fig. 7-1 Sectional configuration of designed columns: (a) CAM RC, (b) NiTi RC, and (c)
steel RC...................................................................................................................171
Fig. 7-2 Axial force-bending moment interaction diagram of designed columns: (a) CAM
RC, (b) NiTi RC, and (c) steel RC. ..........................................................................172
Fig. 7-3 Constitutive models used in sectional analysis: (a) CAM SMA, (b) NiTi SMA, (c)
confined concrete, (d) unconfined concrete, and (e) steel rebar. ...............................173
Fig. 7-4 Moment-curvature diagrams of three types of columns: (a) CAM RC, (b) NiTi RC,
and (c) steel RC. ......................................................................................................174
Fig. 7-5 Arrangement of longitudinal reinforcement for three types of columns: (a) CAM
RC, (b) NiTi RC, and (c) steel RC. ..........................................................................177
Fig. 7-6 Cost comparison of steel RC, CAM RC, and NiTi RC................................................184
Fig. 8-1 Dimensions of (a) Ni-Ti-Co and Ni-Ti, and (b) Cu-Al-Mn SMA specimens...............189
Fig. 8-2 Test setup for mechanical tests on Ni-Ti-Co, Ni-Ti and Cu-Al-Mn SMA samples. .....190
Fig. 8-3 Results of 1% strain incremental cyclic loading until failure on Ni-Ti-Co SMA at
23 ℃: (a) stress-strain curve, and (b) residual strain versus maximum applied
strain. ......................................................................................................................192

xix
Fig. 8-4 Stress-strain curves of Ni-Ti-Co SMA at different temperatures: (a) 50 ℃, (b) 40
℃, (c) 23 ℃, (d) 0 ℃, (e) -20 ℃, (f) -40 ℃, and (g) -60 ℃. ...................................194
Fig. 8-5 Results of 1% strain incremental cyclic loading until failure on Ni-Ti SMA at 23
℃: (a) stress-strain curve, (b) residual strain versus maximum applied strain...........195
Fig. 8-6 Stress-strain curves of Ni-Ti SMA at different temperatures: (a) 50 ℃, (b) 23 ℃,
(c) 0 ℃, and (d) -20 ℃. ...........................................................................................196
Fig. 8-7 Result of 1% strain incremental cyclic loading until failure on Cu-Al-Mn SMA at
25 ℃: (a) stress-strain curve, and (b) residual strain versus applied strain. ...............197
Fig. 8-8 Stress-strain curves of Cu-Al-Mn SMA at different temperatures...............................199
Fig. 8-9 Relationship between yield stress and temperature of Ni-Ti-Co, Ni-Ti, and Cu-AlMn SMA..................................................................................................................200
Fig. 9-1 Dimensions of Ni-Ti-Co and Ni-Ti SMA samples used in low-cycle fatigue tests. .....203
Fig. 9-2 Stress-strain curves of NiTiCo SMA at room temperature, 23 °C: (a) NiTiCo-1,
and (b) NiTiCo-2. ....................................................................................................206
Fig. 9-3 Stress-strain curves of Ni-Ti SMA at room temperature, 23°C: (a) NiTi-1, and (b)
NiTi-2......................................................................................................................207
Fig. 9-4 Variation in mechanical properties of Ni-Ti-Co SMA at room temperature, 23 °C:
(a) NiTiCo-1, and (b) NiTiCo-2. ..............................................................................209
Fig. 9-5 Variation in mechanical properties of Ni-Ti SMA at room temperature, 23 °C: (a)
NiTi-1, and (b) NiTi-2. ............................................................................................210
Fig. 9-6 Stress-strain curves of Ni-Ti-Co SMA at -40 °C.........................................................211
Fig. 9-7 Stress-strain curves of Ni-Ti SMA at 0 °C..................................................................211
Fig. 9-8 Variation in mechanical properties of Ni-Ti-Co SMA at -40 °C. ................................212
Fig. 9-9 Variation in mechanical properties of Ni-Ti SMA at 0 °C. .........................................213
Fig. 9-10 Stress-strain curves of Ni-Ti-Co SMA at 50 °C. .......................................................214
Fig. 9-11 Stress-strain curves of Ni-Ti SMA at 50 °C..............................................................214

xx
Fig. 9-12 Variation in mechanical properties of Ni-Ti-Co SMA at 50 °C.................................215
Fig. 9-13 Variation in mechanical properties of Ni-Ti SMA at 50 °C.......................................216
Fig. 10-1 Constitutive models used in RC columns: (a) confined concrete, (b) unconfined
concrete, and (c) steel rebar. ....................................................................................220
Fig. 10-2 (a) Section details of reference column, and (b) Validation of established
OpenSees model. .....................................................................................................221
Fig. 10-3 Definition of key parameter used to model SMA bars. .............................................222
Fig. 10-4 Stress-strain curves used to model SMA columns: (a) CAM SMA, (b) NiTiCo
SMA, (c) NiTi SMA, (d) confined ECC, and (e) unconfined ECC. ..........................223
Fig. 10-5 Typical moment-curvature curves of: (a) RC column, and (b) SMA column.............225
Fig. 10-6 Test matrix of moment-curvature analyses. ..............................................................226
Fig. 10-7 M-Phi results of RC and SMA columns with different section diameter, D: (a)
RC-1 with D = 4 ft, (c) to (d) NiTi, CAM and NiTiCo sections iterating to match
RC-1; (e) RC-2 with D = 5 ft, (f) to (h) NiTi, CAM and NiTiCo sections iterating
to match RC-2; (i) RC-3 with D = 6 ft, (j) to (l) NiTi, CAM and NiTiCo sections
iterating to match RC-3............................................................................................230
Fig. 10-8 M-Phi results of RC and SMA columns with different reinforcement ratio, r: (a)
RC-4 with r = 1%, (c) to (d) NiTi, CAM and NiTiCo sections iterating to match
RC-4; (e) RC-5 with r = 2%, (f) to (h) NiTi, CAM and NiTiCo sections iterating
to match RC-5; (i) RC-6 with r = 3%, (j) to (l) NiTi, CAM and NiTiCo sections
iterating to match RC-6............................................................................................232
Fig. 10-9 M-Phi results of RC and SMA columns with different axial force ratio, a: (a) RC7 with a = 5%, (c) to (d) NiTi, CAM and NiTiCo sections iterating to match RC7; (e) RC-8 with a = 10%, (f) to (h) NiTi, CAM and NiTiCo sections iterating to
match RC-8; (i) RC-9 with a = 15%, (j) to (l) NiTi, CAM and NiTiCo sections
iterating to match RC-8............................................................................................234
Fig. 11-1 Dimensions of Fe-Mn-Si SMA samples. All dimensions are in mm. ........................238

xxi
Fig. 11-2 Schematic diagram of the loading protocols and definition of key parameters: (a)
monotonic loading, (b) 1% strain incremental cyclic loading. ..................................239
Fig. 11-3 Monotonic loading stress-strain curves of Fe-Mn-Si SMA at different temperatures:
(a) full view, and (b) zoomed-in view.........................................................................240
Fig. 11-4 Incremental cyclic tests on Fe-Mn-Si SMA at different temperatures: (a) 23 ℃,
(b) -40 ℃, and (c) 50 ℃..........................................................................................242
Fig. 11-5 Mechanical properties extracted from the incremental cyclic loading tests: (a)
Eload vs. maximum applied strain, (b) r vs. maximum applied stress, and (c) r
vs. maximum applied strain. ....................................................................................243
Fig. 12-1 Schematic diagrams of: (a) actuation process of Fe-Mn-Si SMA, (b) incremental
cyclic loading after actuation, and (c) low-cycle fatigue loading after actuation. ......249
Fig. 12-2 Labeling rules of the Fe-Mn-Si SMA specimens. .....................................................251
Fig. 12-3 Results of incremental cyclic loading tests on Fe-Mn-Si SMA under different
post-actuation temperatures: (a) 23 ℃, P(4%)-C(0.1%)-23C, (b) -40 ℃, P(4%)-
C(0.1%)-m40C, and (c) 50 ℃, P(4%)-C(0.1%)-50C................................................253
Fig. 12-4 Results of incremental cyclic loading on Fe-Mn-Si SMA at different prestrain
levels: (a) 15%, P(15%)-C(0.1%)-23C, (b) 20%, P(20%)-C(0.1%)-23C, (c) 25%,
P(25%)-C(0.1%)-23C, and (d) 30%, P(30%)-C(0.1%)-23C. ....................................255
Fig. 12-5 Results of monotonic loading tests on Fe-Mn-Si SMA after incremental cyclic
loading: (a) full view, and (b) zoomed-in view. ..........................................................256
Fig. 12-6 Results of low-cycle fatigue tests on Fe-Mn-Si SMA: (a) prestrain 15% & fatigue
0.5%, P(15%)-F(0.5%)-23C, (b) prestrain 15% & fatigue 1.0%, P(15%)-
F(1.0%)-23C, and (c) prestrain 20% & fatigue 1.0%, P(20%)-F(1.0%)-23C.
Note: ‘C1’ means cycle No.1. ..................................................................................258
Fig. 12-7 Variation in mechanical properties of Fe-Mn-Si SMA during low-cycle fatigue
loading: (a) P(15%)-F(0.5%)-23C, (b) P(15%)-F(1.0%)-23C, and (c) P(20%)-
F(1.0%)-23C............................................................................................................260
Fig. 12-8 Results of monotonic loading tests on Fe-Mn-Si SMA after low-cycle fatigue
loading: (a) full view, and (b) zoomed-in view.........................................................261

xxii
Fig. B-1 Stress-strain curves of single crystal Cu-Al-Mn SMA bar (SCB) at room
temperature, 25 °C: (a) SCB-25C-3, (b) SCB-25C-4, (c) SCB-25C-5, and (d)
SCB-25C-6..............................................................................................................312
Fig. B-2 Variation in mechanical properties of SCB at room temperature, 25 °C: (a) SCB25C-3, (b) SCB-25C-4, (c) SCB-25C-5, and (d) SCB-25C-6....................................314
Fig. B-3 Stress-strain curves of SCB-m10-3 at -40 °C. ............................................................315
Fig. B-4 Variation in mechanical properties of SCB-m10C-3 at -40 °C. ..................................315
Fig. B-5 Stress-strain curves of SCB-50C-3 at 50 °C...............................................................316
Fig. B-6 Variation in mechanical properties of SCB-50C-3 at 50 °C........................................316
Fig. B-7 Stress-strain curves of polycrystal Cu-Al-Mn plate (PCP) at room temperature, 25
°C: (a) PCP-25C-3, (b) PCP-25C-4, and (c) PCP-25C-5. .........................................317
Fig. B-8 Variation in mechanical properties of PCP at room temperature, 25 °C: (a) PCP25C-3, (b) PCP-25C-4, and (c) PCP-25C-5..............................................................319
Fig. B-9 Stress-strain curves of NTB-25C-3 at room temperature, 25 °C. ................................320
Fig. B-10 Variation in mechanical properties of NTB-25C-3 at room temperature, 25 °C........320
Fig. B-11 Stress-strain curves of NTB-m10C-3 at -10 °C. .......................................................321
Fig. B-12 Variation in mechanical properties of NTB-m10C-3 at -10 °C.................................321
Fig. B-13 Stress-strain curves of NTB-50C-3 at 50 °C. ...........................................................322
Fig. B-14 Variation in mechanical properties of NTB-50C-3 at 50 °C. ....................................322

xxiii
Abstract
Shape memory alloys (SMAs) are considered for use in bridges due to their unique
superelastic and shape memory effects. The first successful implementation of SMAs in the SR99
Alaskan Viaduct Bridge in Seattle was completed in 2017. Existing research on the application of
SMAs in bridges mainly focused on binary Ni-Ti compositions due to their earlier discovery. NiTi SMAs have the necessary characteristics to be used as plastic hinge reinforcement in bridge
columns, such as strength, ductility, superelasticity, corrosion resistance, and energy dissipation
capacity. However, certain properties of the Ni-Ti SMAs, such as the difficulty in machining,
potential loss of superelasticity at low temperatures, and the high cost, still drive the search for
alternate materials.
The next generation SMAs considered for use in bridge applications include: Cu-Al-Mn SMA,
Ni-Ti-Co SMA, and Fe-Mn-Si SMA. This dissertation performed comprehensive material
characterizations on these SMAs for use in bridges under extreme environments, such as corrosion,
varying ambient temperature, and earthquake loading. To benchmark each SMA composition,
conventional Ni-Ti SMA and commonly used reinforcing steels, such as mild steel, stainless steel,
high chromium steel, and epoxy coated steel, were tested under the same conditions.
The main research activities conducted on Cu-Al-Mn SMA, Ni-Ti-Co SMA, and Fe-Mn-Si
SMA are described as follows in three parts. For each target material, research activities selected
for investigation were determined based on the needs in real bridge applications and the knowledge

xxiv
gaps in literature. Frist, the following material characteristics of Cu-Al-Mn SMA were studied: (i)
corrosion behavior, (ii) low-cycle fatigue behavior at different temperatures, (iii) machinability
characteristics, (iv) headed coupling behavior, and (v) cost effectiveness when applied in typical
bridge columns. Comparisons with conventional Ni-Ti SMA and commonly used reinforcing
steels including mild steel, stainless steel, high chromium steel, and epoxy coated steel were made
to benchmark the behavior of Cu-Al-Mn SMA. Second, the following material characteristics of
Ni-Ti-Co SMA were studied: (i) effect of temperature on superelasticity, strength and ductility, (ii)
low-cycle fatigue behavior at different temperatures, and (iii) moment-curvature response when
applied in typical bridge columns. Comparisons with Ni-Ti SMA and Cu-Al-Mn SMA were made
to benchmark the behavior of Ni-Ti-Co SMA. Third, the following material characteristics of FeMn-Si SMA were studied: (i) effect of temperature on strength, ductility and recovery strain, (ii)
cyclic actuation behavior, and (iii) low-cycle fatigue behavior.
The main findings on Cu-Al-Mn SMA included the following five aspects. First, Cu-Al-Mn
SMA had higher corrosion resistance than mild steel but lower than conventional Ni-Ti SMA. The
superelasticity, particularly the strain recovery capacity, of Cu-Al-Mn SMA showed almost no
degradation after long-term corrosion. Second, when subjected 5% strain low-cycle fatigue loading,
the fatigue life of Cu-Al-Mn SMA (over 50,000 cycles) could be 500 times higher than that of NiTi SMA (around 100 cycles). However, it is noted that the yield strength and energy dissipation
of Cu-Al-Mn SMA was lower than that of Ni-Ti SMA. Third, the machinability of Cu-Al-Mn
SMA was over 20 times better than Ni-Ti SMA. The difficulty of machining Cu-Al-Mn SMA was

xxv
overall close to that of stainless steel but higher than that of mild steel. Fourth, using headed
coupling method to connect large diameter Cu-Al-Mn SMA with conventional steel was feasible.
The key to ensure the superelasticity and ductility of headed Cu-Al-Mn SMA was to ensure a
consistent cooling rate in the central and peripheral regions of the headed end after the heading
process. Fifth, compared to the cost of conventional RC columns, the additional cost of Cu-Al-Mn
SMA reinforced columns was only about 1/4 of the cost on Ni-Ti SMA reinforced columns.
The main findings on Ni-Ti-Co SMA included the following three aspects. First, at room
temperature 23 ℃, the yield strength of Ni-Ti-Co SMA was around 2.6 times that of Ni-Ti SMA
and 3.3 times that of Cu-Al-Mn SMA. The superelastic temperature range of Ni-Ti-Co SMA (-
40 ℃ to 40 ℃) was wider than that of Ni-Ti SMA (0 ℃ to 50 ℃) and close to that of Cu-Al-Mn
SMA (-40 ℃ to 50 ℃). It is noted that when temperature increased to 50 ℃, the stress-strain
curves of Ni-Ti-Co SMA was no longer flag shaped and had slight residual strain accumulation
(0.5% after unloading from 4%). Second, Ni-Ti-Co SMA has great potential for use in lowtemperature seismic applications. At low temperature, the energy dissipation and strain recovery
of Ni-Ti-Co SMA at -40 ℃ was even better than Ni-Ti SMA at 0 ℃. At room temperature 23 ℃
and high temperature 50 ℃, the Ni-Ti-Co SMA showed close low-cycle fatigue resistance (in
terms of superelasticity degradation and fatigue life) to Ni-Ti SMA. Third, due to the high strength
and large size availability of Ni-Ti-Co SMA, columns reinforced with it could have comparable
flexural capacity with conventional steel reinforced concrete columns without increasing the
section diameter or longitudinal reinforcement ratio.

xxvi
The main findings on Fe-Mn-Si SMA included the following three aspects. First, nonactivated Fe-Mn-Si SMA showed excellent deformability under a wide range of temperatures from
-40 ℃ to 50 ℃. The maximum fracture strain of Fe-Mn-Si SMA at -40 ℃ reached 58%. In addition,
the recovery strain of non-activated Fe-Mn-Si SMA increased as the increasing of maximum
applied strain. At 23 ℃, the recovery strain of Fe-Mn-Si SMA at 5% strain was 0.83%, while at
25% strain, the recovery strain increased to 1.64%, almost doubled. Second, increasing the
prestrain level could efficiently increase the post-actuation strain amplitude before the actuation
stress reduces to zero. Specifically, increasing the prestrain level from 4% (commonly used in past
research) to 25%, the strain amplitude when actuation stress decreased to zero showed a 110%
increase. Third, increasing the prestrain level did not sacrifice the post-actuation low-cycle fatigue
resistance and deformability of Fe-Mn-Si SMA. After 500 cycles of 0.5% strain fatigue loading,
Fe-Mn-Si SMA with a prestrain of 15% still exhibited a fracture strain over 19% in the subsequent
monotonic loading.

1
Chapter 1 - Objectives and scope
1.1 Research objectives
Traditional designs of reinforced concrete (RC) bridge columns rely on the formation of
plastic hinges at predetermined locations to absorb the seismic energy and achieve a ductile
seismic response. Although this design philosophy can effectively prevent collapse, the plastic
hinge damage after strong earthquakes causes many problems. Some typical plastic hinge damage
observations from conventional RC bridge columns are shown in Fig. 1-1. The yielding of
reinforcing bars accompanied by spalling and crushing of concrete could lead to severe damage
and large residual drifts, which could seriously inhibit the traffic flow and cause indirect economic
losses [1,2]. Furthermore, repairing the plastic hinge damage or replacing the whole bridge could
exacerbate the situation and cause greater financial losses. As reported by Kawashima et al. [3],
due to the difficulty of repairing the plastic hinge damage, more than one hundred RC columns
with residual inclinations greater than one degree were demolished after the 1995 Kobe earthquake.
To alleviate the plastic hinge damage and residual drift, possible strategies include using
smart materials, post-tensioning elements, damage-resistant cementitious materials, and base
isolation among others. Shape memory alloys (SMAs) are one of the smart materials considered
for use in bridges due to their two special material properties, namely the superelastic and shape
memory effects. The superelastic effect allows SMAs to recover inelastic strain and dissipate
energy upon unloading, and the shape memory effect allows SMAs to return to the original shape

2
upon thermal stimulation, also referred to as actuation. More details on the mechanism of the
superelastic and shape memory effects of SMA will be introduced in Chapter 2.
(a) (b) (c)
Fig. 1-1 Typical plastic hinge damage of conventional reinforced concrete (RC) bridge
columns [4].
For bridge applications, the superelastic and shape memory effects of SMA can be
respectively used to provide strain recovery and post-tensioning performance in bridge columns,
thereby, improving the seismic resilience of bridge systems [5–14]. The feasibility of applying
SMAs in bridge columns has been proven by extensive experimental studies [6,7,9,10,15–27].
According to Hosseini et al. [10], the strain recovery capacity of SMA bars are shown to reduce
the permanent deformation of columns up to 91% compared with the conventional RC columns
after being subjected to a peak drift of 7%. Schematic responses of a bridge column with and
without SMA bars and the arrangement of SMA bars in a bridge column reported by Hosseini et
al. [10] are shown in Fig. 1-2.

3
Past research on SMAs mainly focused on the binary Ni-Ti alloy compositions, which show stable
superelasticity and corrosion resistance. Nevertheless, for large-scale applications, Ni-Ti SMAs still
have the following limitations. First, the Ni-Ti SMAs are very expensive compared to steel and
they are difficult to process [28]. Second, the martensitic transformation start temperature of NiTi SMAs is normally higher than -25 ℃ [29], which means they will lose superelasticity at low
temperatures. More commonly, Ni-Ti SMAs lose superelasticity at around 0 ℃. Third, the Ni-Ti
SMAs are available commonly in thin wire forms. Manufacturing of Ni-Ti SMAs in large diameter
with stable thermo-mechanical behavior is challenging and adds to the cost. The high cost,
difficulty in machining, and limited operating temperature restrict the wide application of Ni-Ti
(a) (b)
Fig. 1-2 (a) Schematic response of a bridge column with and without SMA bars, and (b)
arrangement of SMAs in a bridge column [10]. Note: SEA is referring to SMA with
superelastic effect at room temperature.

4
SMAs in civil engineering. To overcome these limitations of conventional Ni-Ti SMAs, alternate
materials are being studied.
Next generation SMAs that have attracted attention from the bridge engineering community
include Cu-Al-Mn, Ni-Ti-Co, and Fe-Mn-Si SMAs. The advantages of Cu-Al-Mn SMA mainly
include excellent low-cycle fatigue stability, superelasticity under a wide temperature range, ease of
machinability, and relatively low cost [30]. The advantages of Ni-Ti-Co SMA mainly include high
strength, superelasticity under a wide temperature range, and availability in large size. Cu-Al-Mn and
Ni-Ti-Co SMAs are promising materials for bridge applications due to their superior superelasticity.
The Fe-Mn-Si SMA has received increasing research attention in recent years due to their relatively
low cost, shape memory effect, and potential for use in post-tensioning of concrete members [31].
Due to the short history, studies on Cu-Al-Mn, Ni-Ti-Co, and Fe-Mn-Si SMAs are still limited.
Most of the key mechanical behaviors of these SMA compositions related to bridge applications have
never been studied. In this context, the main objectives of this dissertation are to investigate the
mechanical behaviors of the next generation SMAs: Cu-Al-Mn, Ni-Ti-Co, and Fe-Mn-Si SMAs,
and to determine the feasibility of using them in bridges in seismic regions.
1.2 Research scope
The extreme conditions that may threaten the safety and functionality of bridges mainly include:
long-term exposure to corrosive conditions, ambient climatic temperature variations, and earthquake
loads. The corrosion of reinforcement could lead to degradation of flexural capacity and concrete

5
spalling, both of which compromise the serviceability and strength of bridges [32–35]. Variation in
ambient temperature (if combined with seismic loads) could cause partial or complete loss of
superelastic or shape memory effect in certain composition of SMA materials. In severe cases, both of
these two properties may disappear, threatening the seismic resistance of bridges. With regards to
earthquake loads, a typical earthquake event may involve hundreds of cyclic loadings [36]. To
ensure the bridges reinforced with SMA bars remain operational after an earthquake, it is important
for the SMA bars to maintain stable low-cycle fatigue resistance without fracture or losing the
superelastic or shape memory effect.
In addition to the challenges from the abovementioned extreme conditions, since SMA bars
are typically applied only in the plastic hinge regions to take full advantage of their strain recovery
and energy dissipation capacity while maintaining cost efficiency [37,38], machining and coupling
SMA bars with conventional steel bars become necessary. Therefore, the feasibility and cost
effectiveness of machining and coupling SMA bars with conventional steel bars also need to be
evaluated.
In this dissertation, to characterize the behavior of next generation Cu-Al-Mn, Ni-Ti-Co and
Fe-Mn-Si SMAs for use in bridges in seismic regions, the following investigations were conducted,
namely long-term corrosion behavior, electrochemical corrosion behavior, low-cycle fatigue behavior,
temperature dependence of the strength and ductility, machinability characteristics, headed coupling
behavior, cost effectiveness, and cyclic actuation behavior. Based on the target SMA, the research
activities conducted in this dissertation can be divided into three parts, as shown in Fig. 1-3.

6
The research activities conducted in this dissertation are briefly described below.
Part I: Cu-Al-Mn SMA
➢ Long-term salt spray corrosion and electrochemical corrosion tests were conducted. The
following materials were tested and compared: single crystal Cu-Al-Mn SMA bars, polycrystal
Cu-Al-Mn SMA plates, polycrystal Ni-Ti SMA bars, as well as four types of commonly used
steel reinforcing bars, namely mild steel, high chromium steel, epoxy coated steel, and stainless
steel. The corrosion surface, mass loss, degradation in mechanical properties, corrosion
potential, and corrosion rate were investigated.
➢ Low-cycle fatigue tests at room temperature, 25 ℃, low temperatures, -40 ℃ and -10 ℃, and
high temperature, 50 ℃ were conducted. Three types of materials were tested and compared,
Fig. 1-3 A schematic diagram of research activities conducted in this dissertation.
Cu-Al-Mn
SMA
Long-term corrosion
behavior
Electrochemical
corrosion behavior
Low-cycle fatigue
behavior
Temperature
dependance
Machinability
characteristics
Headed coupling
behavior
Cost estimation
Ni-Ti SMA
Mild steel
High chromium steel
Epoxy coated steel
Stainless steel
Ni-Ti SMA
Ni-Ti SMA
Mild steel
Stainless steel
Ni-Ti SMA
Mild steel
Ni-Ti-Co
SMA
Ni-Ti SMA
Low-cycle fatigue
behavior
Temperature
dependance
Fe-Mn-Si
SMA
Low-cycle fatigue
behavior
Temperature
dependance
Strength and
ductility
Cyclic actuation
behavior
Strength and
ductility
Cu-Al-Mn
SMA
Note: is the target SMA is the behaviors characterized is the materials used to make comparisons

7
namely single crystal Cu-Al-Mn SMA bars, polycrystal Cu-Al-Mn SMA plates, and
polycrystal Ni-Ti bars. Degradation in key superelastic properties of these materials under
different temperatures with respect to low-cycle fatigue loading was investigated.
➢ Single point turning machinability tests, including both one-workpiece single machining and
thirty workpieces continuous machining, were conducted. The materials tested were Cu-AlMn SMA, Ni-Ti SMA, mild steel, and stainless steel. Effects of a wide range of cutting
parameters, such as cutting speed ranging from 15 to 120 m/min, feed rate ranging from 0.1 to
0.2 mm/rev, and depth of cut ranging from 0.5 to 1.5 mm, were investigated. The chip
formation, tool wear, surface roughness, and work piece diameter deviation were analyzed.
➢ Mechanical tests and microstructural analyses on headed Cu-Al-Mn SMA coupled with
conventional steel rebar were conducted. Mechanical tests including monotonic and cyclic
loadings were applied to simulate earthquake loading. Microstructural analyses including
electron backscatter diffraction (EBSD), metallographic imaging, Vickers hardness testing,
and fractographic evaluation were performed. The effect of heading on the stress induced
martensitic transformation (SIMT), phase composition, and failure modes were investigated.
➢ Cost estimation analyses on representative bridge columns reinforced with Cu-Al-Mn SMA
were performed. Comparisons with conventional RC columns and columns reinforced with
Ni-Ti SMA were made. The cost of producing, processing, and coupling SMA bars were

8
considered, and the cost effectiveness of using Cu-Al-Mn SMA in bridge columns was
evaluated.
Part II: Ni-Ti-Co SMA
➢ Cyclic loading tests on Ni-Ti-Co SMA at temperatures ranging from -60 ℃ to 50 ℃ were
conducted. Comparisons with Ni-Ti and Cu-Al-Mn SMAs were made. The effect of
temperature on the superelasticity and ductility of Ni-Ti-Co, Ni-Ti and Cu-Al-Mn SMAs was
evaluated. Key mechanical properties such as yield strength, ductility, and maximum recovery
strain of Ni-Ti-Co SMA were extracted and compared with Ni-Ti and Cu-Al-Mn SMAs.
➢ Low-cycle fatigue tests on Ni-Ti-Co SMA atroom temperature 23 ℃, low temperature -40 ℃,
and high temperature 50 ℃ were conducted. Comparisons with Ni-Ti SMA were made. The
effect of low-cycle fatigue loading on superelastic properties of Ni-Ti-Co and Ni-Ti SMAs
was analyzed and compared. The stress-strain curves, Young’s modulus, yield strength,
damping ratio, and recovery strain were extracted and discussed.
➢ Moment-curvature analyses on representative bridge columns reinforced with Ni-Ti-Co SMA
bars were performed. Comparisons with bridge columns reinforced with Ni-Ti and Cu-Al-Mn
SMAs, as well as conventional RC columns were made. The influence of column diameter,
longitudinal reinforcement ratio, and axial force ratio on the flexural capacity of bridge
columns reinforced with Ni-Ti-Co SMA bars was analyzed, and compared with those
reinforced with Ni-Ti and Cu-Al-Mn SMAs.

9
Part III: Fe-Mn-Si SMA
➢ Monotonic and incremental cyclic loading tests on non-activated Fe-Mn-Si SMA at room
temperature 23 ℃, low temperature -40 ℃, and high temperature 50 ℃ were conducted. The
influence of temperature on the key mechanical properties of Fe-Mn-Si SMA, such as
Young’s modulus, yield strength, ductility, and recovery strain was analyzed.
➢ Cyclic, low-cycle fatigue, and monotonic loading tests on activated Fe-Mn-Si SMA were
conducted. Post-actuation temperatures ranging from -40 ℃ to 50 ℃, prestrain levels ranging
from 4% to 30%, and low-cycle fatigue loading amplitudes ranging from 0.5% to 1.0% were
tested to study the cyclic actuation behavior of Fe-Mn-Si SMA. Influence of cyclic and lowcycle fatigue loading on the deformability and actuation stress degradation of Fe-Mn-Si SMA
was investigated.
1.3 Organization of the dissertation
This dissertation contains thirteen chapters. The organization of the dissertation and the
content of each chapter are described as follows.
Chapter 1 outlines the research objectives, scope and organization of this dissertation.
Chapter 2 presents the background information on SMAs studied in this dissertation, and
literature review on the development history and characteristics of each SMA type.
Chapter 3 ~ 7 presents research on Cu-Al-Mn SMA.

10
Chapter 3 presents the long-term corrosion and electrochemical corrosion behavior of Cu-AlMn SMA, and comparisons with Ni-Ti SMA, mild steel, high chromium steel, epoxy
coated steel, and stainless steel.
Chapter 4 presents the low-cycle fatigue behavior of Cu-Al-Mn SMA under different
temperatures, and comparisons with Ni-Ti SMA.
Chapter 5 presentsthe machinability characteristics of Cu-Al-Mn SMA, and comparisons with
Ni-Ti SMA, mild steel, and stainless steel.
Chapter 6 presents the headed coupling behavior of Cu-Al-Mn SMA with steel rebar.
Chapter 7 presents the cost estimation of typical bridge columns reinforced with Cu-Al-Mn
SMA, and comparisons with conventional RC column and bridge columns
reinforced with Ni-Ti SMA.
Chapter 8 ~ 10 presented research on Ni-Ti-Co SMA.
Chapter 8 presents the temperature dependence of strength and superelasticity of Ni-Ti-Co
SMA, and comparisons with Ni-Ti and Cu-Al-Mn SMAs.
Chapter 9 presents the low-cycle fatigue behavior of Ni-Ti-Co SMA under different
temperatures, and comparisons with Ni-Ti SMA.

11
Chapter 10 presents the moment-curvature response of typical bridge columns reinforced with
Ni-Ti-Co SMA, and comparisons with columns reinforced with Ni-Ti and Cu-AlMn SMAs.
Chapter 11 ~ 12 presents research on Fe-Mn-Si SMA.
Chapter 11 presents the temperature dependence of strength, ductility and recovery strain of
non-activated Fe-Mn-Si SMA.
Chapter 12 presents the effect of prestrain level, post-actuation temperature, cyclic and lowcycle fatigue loading on the actuation behavior of Fe-Mn-Si SMA.
Chapter 13 presents the summary of completed research and main findings of this dissertation, and
recommendations for future research.

12
Chapter 2 - Background information
The background information on shape memory alloys (SMAs) studied in this dissertation is
presented. The basic mechanical properties of SMA, i.e., the superelastic and shape memory
effects, as well as the advantages of their application in bridges, are introduced. The development
history and characteristics of Ni-Ti-based, Cu-based and Fe-based SMAs are reviewed.
2.1 Overview of shape memory alloy (SMA)
Shape memory alloys (SMAs) are unique materials that can recover inelastic strains upon
stress removal, known as superelastic effect or superelasticity, see Fig. 2-1 (a), or return to their
original shape upon thermal stimulation, known as shape memory effect, see Fig. 2-1 (b). SMAs
that exhibit superelasticity at room temperature are also known as superelastic alloys (SEAs).
Typically, SEAs are used as flexural reinforcement in concrete bridge columns because the
superelastic effect is particularly advantageous in dissipating seismic energy and reducing
permanent drifts of bridge columns [11,39–41]. The shape memory effect is most commonly used
to provide post-tensioning in concrete structures. In contrast to traditional technologies using high
strength steel tendons post-tensioned by hydraulic devices, the post-tensioning of SMAs is not
performed through application of an external load, but through their internal martensitic phase

13
transformation via thermal stimulation, also known as “actuation”. Through actuation, SMAs can
generate post-tensioning forces with no friction losses. Furthermore, since no heavy hydraulic
devices are required, manpower and construction space can be saved. All these features are
advantages for bridge applications.
The superelastic and shape memory effect of SMAs originate from the internal reversible
transformation between the austenitic and martensitic phases, also known as martensitic
transformation, as shown in Fig. 2-2. The austenitic phase has a highly symmetric crystal structure
with stability at low stress or high temperature, whereas the martensitic phase has a low-symmetry
crystal structure with stability at high stress or low temperature. When the ambient temperature
decreases or the stress increases, the austenitic phase transforms into the martensitic phase and
vice versa. The martensitic phase transformation does not involve any compositional change, and
it can be triggered repeatedly with proper thermal-mechanical stimulations. Temperature is one of
the most important factors affecting the martensitic phase transformation in SMAs.
(a) (b)
Fig. 2-1 Schematic diagram of (a) superelastic effect, and (b) shape memory effect in SMAs.
Strain
Stress Stress
Strain
Unloading Heating (Actuation)

14
Fig. 2-2 Martensitic and austenitic phase transformation in SMAs [42].
Materials with shape memory effect were first documented in Au-Cd alloys in 1932 [43].
Since then, shape memory effect has been discovered in various types of alloys. To date, SMAs
have been widely used in various industries such as aerospace, automotive, biomedical, and civil
engineering [42,44–46]. This dissertation mainly focused on the application of SMAs in bridge
engineering, among which the most commonly used or received the most research attention are
Ni-Ti-based, Cu-based, and Fe-based SMAs.
2.2 Ni-Ti-based SMA
Ni-Ti-based SMAs are the most widely investigated shape memory alloys due to their earlier
discovery. Binary Ni-Ti SMAs, which are composed of near equiatomic Nickel (Ni) ad Titanium
(Ti), i.e., 50at % Ni and 50at% Ti, where “at %” represents the atomic percent, have been proven
to exhibit stable superelasticity, shape memory effect, corrosion resistance and biocompatibility.
Superelastic

15
In the early stages of discovery, Ni-Ti SMAs were widely used as medical devices such as
orthodontic archwires, dental arenas, and neurovascular stents due to their biocompatibility [42].
Later, Ni-Ti SMAs gained more attention in the aerospace and civil engineering applications. The
high strength, ductility and energy dissipation capacity of Ni-Ti SMAs make them suitable for use
as plastic hinge reinforcement in bridge columns. The first successful implementation of Ni-Ti
SMAs in the SR99 Alaskan Viaduct Bridge in Seattle was completed in 2017 [47].
Nevertheless, for large-scale applications, Ni-Ti SMAs still have the following limitations.
First, the Ni-Ti SMAs are prohibitively expensive and difficult to process [28,48]. Moreover, the
martensitic transformation start temperature of Ni-Ti SMAs is normally higher than -25 ℃ [29],
which means they will lose superelasticity at low temperatures. More commonly, Ni-Ti SMAs lose
superelasticity at around 0 ℃. Third, the Ni-Ti SMAs are commonly available in thin wire forms.
The manufacturing of large size Ni-Ti SMAs with stable thermal-mechanical behavior remains
challenging. The high cost, difficulty in machining, and narrow operating temperature restrict the
wide application of Ni-Ti SMAs in civil engineering.
The limitations of Ni-Ti SMAs have prompted the search for alternate materials. To produce
Ni-Ti-based SMAs with lower martensitic transformation start temperature, ternary elements such
as Hafnium (Hf), Iron (Fe) or Niobium (Nb) were added to Ni-Ti SMAs [49,50]. However, the
ternary Ni-Ti-Hf and Ni-Ti-Fe SMAs are normally only available in small sizes and the decrease
in their martensitic transformation start temperature is still limited (normally cannot be lower than

16
-20 ℃). The balance between cost and thermal-mechanical of Ni-Ti-based SMAs remains
challenging.
When applying SMAs as flexural reinforcement in bridges, the Young's modulus, yield
strength, and ductility are critical to resist seismic loadings. The yield strength and ductility of
SMAs are in general comparable to conventional steels, however, the Young's modulus of most
SMAs is only around 1/7 to 1/3 of that of conventional steels [19]. In this case, Ni-Ti-Co SMA,
which has higher strength and Yong’s modulus than binary Ni-Ti SMA, is developed and
considered for bridge applications. Kishi et al. [51] found that the addition of Cobalt (Co) increases
the yield strength and decreases the martensitic transformation start temperature. Compared with
conventional binary Ni-Ti SMA, the yield strength of Ni-Ti-Co SMA can be more than 50% higher
[52]. Another advantage of Ni-Ti-Co SMA over the conventional Ni-Ti-based SMA is its
availability in large sizes. At the time of writing this dissertation, large Ni-Ti-Co SMA bars with
diameter over 32 mm have been developed. The availability of large bars indicates the possibility
of using Ni-Ti-Co SMA in bridge columns. However, the basic mechanical properties of large size
Ni-Ti-Co SMA have never been reported in literature.
2.3 Cu-based SMA
Compared with Ni-Ti-based SMAs, the Cu-based SMAs are commercially attractive because
of their low cost and availability in large sizes. Traditional Cu-based SMAs such as Cu-Al-Ni and
Cu-Zn-Al SMAs are brittle and show poor fatigue resistance, which have been restricting their

17
widespread applications [53,54]. Research has been conducted to improve the ductility and fatigue
stability of Cu-based SMAs through texture control, grain refinement, and heat treatment [55–57].
Kainuma et al. [39] found that the performance of Cu-Al-Mn alloys can be improved by controlling
the Aluminum (Al) content. As shown in Fig. 2-3, the cold-workability and elongation of Cu-AlMn alloys increase with decreasing the Al content. Within the composition range < 17 at.-% Al,
the Cu-Al-Mn alloy with cold workability > 60% and tensile elongation > 10% can be obtained
because of a low degree of order in the parent phase [39]. Based on this discovery, research on
developing reliable Cu-Al-Mn alloys has accelerated. Today, it is well agreed that Cu-Al-Mn
alloys with an Al content between 16 and 18 at.-% show a good balance between workability and
shape memory properties [58].
Fig. 2-3 Effect of Al content on cold workability, tensile elongation and shape recovery of Cu-AlMn alloys [39].

18
Another important discovery that made the high-performance Cu-Al-Mn alloys possible is
the abnormal grain growth (AGG) though cyclic heat treatment [59]. A typical AGG process is
shown in Fig. 2-4 (a), where a Cu-Al-Mn alloy sheet was annealed at 750 ~ 900 °C for 10 min,
followed by cooling to 500 ~ 600 °C and subsequent heating to 750 ~ 900 °C with a duration of
60 min and a final water quenching. By doing so repeatedly, the single grain size can grow from
the micrometer-level to millimeter-level and even to centimeter-level, as shown in Fig. 2-4 (b).
Using such techniques, the largest grain size can now be over 700 mm in length [60]. Fig. 2-4 (c)
shows the crystallographic orientation of a Cu-Al-Mn sheet after AGG. Consistent grain
orientations was obtained using such a heat treatment method.
Fig. 2-4 (a) AGG process, (b) optical micrograph of a Cu-Al-Mn alloy sheet before and after
AGG, and (c) crystal orientation of a Cu-Al-Mn sheet after AGG [59].
The mechanical properties of Cu-Al-Mn SMAs have a strong dependence on the grain size
and crystal orientation [56]. Generally speaking, Cu-Al-Mn SMAs with larger grain size and more
Grain size
before AGG
Grain size
after AGG
(a) (c)
(b)

19
consistent intergranular orientations show higher superelastic strain and ductility [58,59,61]. The
effect of relative grain size on the maximum superelastic strain and the effect of crystal orientation
on the stress-strain curves of Cu-Al-Mn SMA are shown in Fig. 2-5. It is seen that Cu-Al-Mn SMA
with a relative grain size, d/t, over 6 and <001> oriented texture shows a maximum superelastic
strain over 6%. In Fig. 2-5 (a), d denotes the grain size, t denotes the thickness of SMA sheet, and
PEmax denotes the maximum superelastic strain.
After two decades of research and development, the existing methods can produce Cu-Al-Mn
SMAs with a large grain size, high ductility, and consistent crystal orientation, all of which
Fig. 2-5 (a) Effect of relative grain size on maximum superelastic strain of Cu-Al-Mn SMA
[58], and (b) effect of crystal orientation on stress-strain curves of Cu-Al-Mn SMA [61].

20
guarantee the superelasticity and cold workability of Cu-Al-Mn SMAs and are advantages of using
them in civil engineering. Cu-Al-Mn SMAs are currently available in single crystal and polycrystal
forms [36,41]. The manufacturing process of single crystal Cu-Al-Mn SMA is more complex and
therefore the cost is higher. Polycrystal Cu-Al-Mn SMA is more cost efficient because of not
requiring extensive heat treatment. Both single crystal and polycrystal Cu-Al-Mn SMAs were
investigated in this dissertation.
2.4 Fe-based SMA
Apart from the abovementioned Ni-Ti-based and Cu-based SMAs, Fe-based SMAs are also
gaining increasing attention in recent years due to their shape recovery capacity and relatively low
cost [62]. Applications of Ni-Ti-based and Cu-based SMAs in civil engineering mainly focused
on using their superelastic effects, i.e., strain recovery and energy dissipation capacity. While for
Fe-based SMAs, existing applications mainly focused on using their shape memory effect to
provide post-tensioning in concrete structures.
The martensitic transformation mechanism of Fe-based SMAs is fundamentally different
from that of Ni-Ti-based and Cu-based SMAs. A comparison of the phase transformation between
Ni-Ti-based (or Cu-based) and Fe-based SMA is shown in Fig. 2-6. The full austenite to martensite
transformation in Fe-based SMAs cannot be realized upon stress removal, which means the Febased SMAs are inferior to Ni-Ti-based or Cu-based SMAs in terms of superelasticity. However,
because Fe-based SMAs have large thermal hysteresis and are prone to forming plastic

21
deformation, their shape memory effect can be easily triggered and maintained at room
temperature. The shape memory effect of Fe-based SMAs enables them to generate posttensioning stress via thermal stimulation, i.e., actuation. There are several different types of FeSMAs available so far, such as Fe-Pt, Fe-Pd, Fe-Ni-Co, and Fe-Mn-Si SMAs. Among them, FeMn-Si SMA has received more research attention in civil engineering applications [62].
A schematic diagram of the actuation process of Fe-Mn-Si SMA is shown in Fig. 2-7. First,
stretch the as-manufactured Fe-Mn-Si SMA to a certain strain level and then unload to zero force.
This process, also known as prestraining, could be performed prior to mounting Fe-Mn-Si SMA
on the target structure. Second, install the Fe-Mn-Si SMA on the target structure and restrain its
two ends to keep the strain constant. Third, apply an external heat stimulation through a blowtorch
or electric resistance until the designed heating temperature is achieved. After that, cool down the
Fe-Mn-Si SMA naturally, and a recovery stress is developed in the material and applied to the
target structure.
(a) (b)
Fig. 2-6 Phase transformation in (a) Ni-Ti-based (or Cu-based) SMA, and (b) Fe-based SMA.
Temperature
Stress
Austenite
Detwinned
martensite
Twinned
martensite
Mf Ms As Af Temperature
Stress
Mf Ms As Af
Austenite
Plasticity with irreversible slip
Austenite
Austenite +
stress-induced
martensite f
s

22
Fig. 2-7 Schematic diagram of the actuation process of Fe-Mn-Si SMA.
In such an actuation process, post-tensioning forces can be generated with no friction losses,
furthermore, since no heavy hydraulic devices are required, manpower and construction space can
be saved. In addition to these advantages, Fe-Mn-Si SMA also has better weldability and lower
cost compared with other SMA compositions for use as post-tensioning elements, such as Ni-TiNb alloys [63,64], all of which indicate the potential of using Fe-Mn-Si SMA in bridges [62].
Images of some typical applications using Fe-Mn-Si SMA to strengthen or repair existing concrete
structures are shown in Fig. 2-8.
Fig. 2-8 Images of some typical applications using Fe-Mn-Si SMA to strengthen or repair
existing concrete structures: (a) flexural strengthening [65], (b) shear strengthening [66].
As received Fe-Mn-Si SMA
Prestrain, then unload to zero force
Restrain both ends
Apply external heat stimulation
Cooling down to room temperature
Recovery stress
(a) (b)

23
Existing research on the application of Fe-Mn-Si SMA mainly focused on structural
rehabilitation, such as flexural, shear strengthening of beams, and active confining of columns, as
shown in Fig. 2-9 (a) to (c). In flexural strengthening, Fe-Mn-Si SMA is used to generate a posttensioning force along the longitudinal direction of the beam, which improves the stiffness and
flexural capacity, see Fig. 2-9 (a). Some examples of this application are presented in Schranz et
al. [65] and Vůjtěch et al.[67]. In shear strengthening, Fe-Mn-Si SMA is used to generate posttensioning force along the transverse direction of the beam. Transverse Fe-Mn-Si SMA hoops or
strips are installed and actuated to improve the shear strength of beams, see Fig. 2-9 (b). Some
examples of this application are presented in Czaderski et al. [66] and Cladera et al.[68]. In active
confining, actuated FeSMA is wrapped around the column to generate active lateral confining
stress and improve the axial capacity, see Fig. 2-9 (c). Some examples of this application are
presented in Zerbe et al. [69] and Vieira et al. [70].
Fig. 2-9 Schematic diagrams of Fe-Mn-Si SMA application in: (a) flexural strengthening
beams, (b) shear strengthening beams, (c) active confining columns, and (d) self-centering
columns.
(a)
(b)
(c)
Fe-Mn-Si SMA
(d)

24
In addition to strengthening or repairing existing structures, Fe-Mn-Si SMA also has potential
for use in self-centering bridge columns, see Fig. 2-9 (d). However, due to the short research
history, studies on using Fe-Mn-Si SMA in self-centering columns are still in the conceptual stage.
Only two pilot studies have been reported on the application of Fe-Mn-Si SMA in self-centering
bridge columns [15,71]. Research conducted in this dissertation mainly focused on using Fe-MnSi SMA in self-centering bridge columns.
2.5 Summary
SMAs are promising materials considered for use in bridges to improve the seismic
resistance. The superelastic effect of SMAs can be used to dissipate seismic energy and reduce
permanent drifts of bridge columns. The shape memory effect of SMAs can be used to provide
post-tensioning performance in bridge columns. Existing studies on SMAs for bridge applications
mainly focused on binary Ni-Ti ones due to their early discovery. However, the high cost, difficult
machinability, and narrow application temperature range of binary Ni-Ti SMAs restrict their
widespread application. To overcome the disadvantages and limitations of conventional binary NiTi SMAs, alternate materials are being studied.
Next generation SMAs considered for use in bridges include Ni-Ti-Co, Cu-Al-Mn, and FeMn-Si SMAs. Each SMA type has its own advantage. The Ni-Ti-Co SMA is gaining attention due
to its high strength and availability in large size. The Cu-Al-Mn SMA is promising for bridge
applications due to its low cost and availability in large size. The Fe-Mn-Si SMA is becoming

25
increasingly popular due to its capacity in generating post-tensioning forces and low cost. Limited
research on Cu-Al-Mn, Ni-Ti-Co and Fe-Mn-Si SMAs has been reported. The feasibility of
applying these SMAs in bridges subjected to harsh environments remains to be investigated.
To ensure safe and economical applications of Cu-Al-Mn, Ni-Ti-Co and Fe-Mn-Si SMAs in
bridges subjected to extreme environments, the corrosion resistance, low-cycle fatigue stability,
machinability, feasibility of coupling with conventional rebar, cost effectiveness, and temperature
dependence of these SMAs need to be characterized. Research in the following chapters addresses
these knowledge gaps.

26
Chapter 3 - Corrosion behavior of Cu-Al-Mn SMA
Long-term salt spray and electrochemical corrosion behaviors of Cu-Al-Mn SMA were
investigated. Comparisons with Ni-Ti SMA and four types of commonly used steel rebar were also
made. First, long-term accelerated corrosion testing on Cu-Al-Mn (CAM) SMA, mild steel (MS),
high chromium steel (XS), epoxy coated steel (ES), and stainless steel (SS), was conducted up to
1,051 days. For each steel, three different diameters: U.S. #3,  = 9.53 mm, U.S. #5,  = 15.88
mm, and U.S. #10,  = 32.26 mm, were tested to determine the effect of bar size on the corrosion
behavior. Mechanical tests were conducted after specimens reached predetermined corrosion
levels. Second, electrochemical corrosion tests were performed on Cu-Al-Mn SMA, the
abovementioned four types of steel, as well as Ni-Ti SMA. Tafel curves were employed to determine
the corrosion rates of Cu-Al-Mn SMA, four types of steel, and Ni-Ti SMA.
3.1 Research motivation
Corrosion is well-known as the leading problems for civil infrastructure around the world, it
is estimated to be responsible for approximately 3% of the world’s annual gross domestic product
(GDP) [72]. When bridges are subjected to chemicals such as deicing salts or airborne chlorides,
the corrosion of the steel rebar leads to mechanical degradation, and concrete spalling that could

27
compromise the serviceability and strength of the structure [32–35]. To delay the initiation of
corrosion and slow down the corrosion progression, certain elements, such as chromium (Cr),
nickel (Ni) and molybdenum (Mo), are added to low-carbon steel. With these elements, a compact
passive film is formed on the rebar surface thereby improving the corrosion resistance of steel
[73]. Generally, higher Cr content yields better corrosion resistance due to the formation of an
amorphous FeCO3 layer [73], meanwhile increasing the cost. A minimum addition of 3% Cr in
low carbon steel forms weathering steel, which forms a protective rust cover under operating
conditions and reduces the further rusting of the substrate [74,75]. The high chromium steel (XS)
is obtained when the Cr content is increased up to 9~11% [76]. The special microstructure of XS
formed during the production process avoids the carbide formation and improves the corrosion
resistance and strength. Increasing the Cr content to approximately 10 ~ 20% yields stainless steels
(SS). It is known that the stainless steel can reach a corrosion resistance of up to 200 ~ 900 times
higher than that of common carbon steel [77,78]. Epoxy coated steels (ES) achieve corrosion
resistance using a different mechanism. The epoxy cover acts as a physical barrier between steel
and the external agents and protect the substrate from corrosion reactions. It is known that the
epoxy coating can remain intact and protect the rebar for up to 20 years of service [79,80].
Extensive research has been conducted on the corrosion performance of Ni-Ti SMAs [81–
85]. It has been proven that the Ni-Ti SMAs show excellent corrosion resistance in aggressive
environments. The primary reason for that is the formation of a compact and stable oxide film on
the surface [86,87]. Due to the presence of certain alloying metals in SMAs such as nickel (Ni)

28
and niobium (Nb), a compact passive film is formed on the surface, thereby improving the
corrosion resistance of Ni-Ti SMAs.
Despite a well understanding of the corrosion behavior of Ni-Ti SMA, there is very limited
research on the behavior of Cu-Al-Mn SMA under harsh environmental conditions that may be
encountered by civil engineering structures. Saud et al. [88] conducted potentialdynamic
polarization tests on Cu-Al-Ni alloys with different amounts of manganese (Mn) addition. It was
found that the Mn content shifts the alloy’s corrosion potential to nobler direction because the Mn
is easier to oxidize and the formation of corrosion products on the electrode surface prevents
further corrosion propagation. Pareek et al. [89] investigated the chemical resistance and yield load
of Cu-Al-Mn SMAs in acidic and alkaline environments. It was found that the Cu-Al-Mn SMAs
have a higher corrosion resistance and a lower reduction in the yield stress than those of mild steel
in most acidic environment, such as H2SO4, HCl, and HNO3. Nevertheless, except for these limited
studies, the degradation in the superelastic properties of Cu-Al-Mn SMA subjected to long-term
corrosion is still unknown. To address this research gap, this study investigated the effect of longterm corrosion on the superelasticity of Cu-Al-Mn SMAs and compared it with Ni-Ti SMA and
four types of commonly used steel rebar.

29
3.2 Experimental program
3.2.1 Long-term corrosion tests
The composition of Cu-Al-Mn SMA used in this study was Cu-8.38Al-11.32Mn (wt. %),
obtained from Furukawa Techno Material Co., Ltd. For brevity, the Cu-Al-Mn SMA is referred to
as CAM SMA hereafter in this chapter. The as received CAM SMAs (both single crystal and
polycrystal) were in the form of 20 mm diameter smooth rods. Two types of CAM SMA specimens
were investigated in the long-term corrosion tests: single crystal CAM SMA bars (SCB) with a
gage length diameter of 12.7 mm, and polycrystal CAM SMA plates (PCP) with a constant
thickness of 6 mm. The SCB was directly machined from as received single crystal CAM SMA
rods to 140 mm length and 16 mm diameter, then the middle section was reduced to 12.7 mm
diameter, as shown in Fig. 3-1 (a) and (b). Three SCB specimens were tested, labeled as SCB-i (i
ranges from 1 to 3).
To obtain PCP, the surface grain boundaries of the as received polycrystal CAM rods was
firstly marked based on a visual inspection, as shown in Fig. 3-1 (c). After that, the polycrystal
CAM rods were sliced along the length into four parts as shown in Fig. 3-1 (d), the middle two
larger parts, denoted as PCP-i and PCP-iC (i ranges from 1 to 5), were used for long-term corrosion
testing. The PCP-i was mechanically tested after designed corrosion levels and its counterpart
PCP-iC was mechanically tested without corrosion. Since PCP-i and PCP-iC were sliced from the
same as received rod and had almost the same grain distribution, their behavior was compared to

30
determine the effect of corrosion on fracture behavior of polycrystal CAM SMA. The grain
boundaries of the as received polycrystal CAM SMA rods are shown in Fig. 3-2.
Fig. 3-1 Preparation of test specimens for long-term corrosion tests: (a) dimensions of SCB
and PCP, (b) machined SCB, (c) surface grain boundaries of as received polycrystal CAM
SMA, and (d) preparation of PCP-i and PCP-iC.
In order to benchmark the behavior of CAM SMAs with commonly used steel rebar in
construction, four types of steel rebar, namely mild steel (MS), high chromium steel (XS), epoxy
coated steel (ES), and stainless steel (SS), were also studied in the long-term corrosion tests, as
shown in Table 3-1. For brevity, the above listed four types of steel rebar are referred to as MS,
XS, ES, and SS respectively hereafter. MS, XS, ES and SS used this study respectively conformed
30
12
30
12
140
R45
R45
56
~20
~17
~6
250
16
65
60
65
60
Grip length
Grip length
20
As received polycrystal CAM rod
Slicing the as received polycrystal CAM rod to obtain PCP-i and PCP-iC
PCP-i
PCP-iC
Note:
All dimensions
are in mm
T
(a) (b) Machined SCB
(c)
(d)
~12.7
~12.7
Surface grain
boudaries
Surface grain boudaries

31
to ASTM A516 Grade 60 [90], ASTM A1035 CS Grade 100 [76], ASTM A615 Grade 60 [91],
and ASTM A955 S32304 [92]. The chemical compositions of these steel rebar are presented in
Table 3-1. For each type of steel rebar, three diameters were tested, namely U.S. #3 ( = 9.53
mm), U.S. #5 ( = 15.88 mm), and U.S. #10 ( = 32.26 mm). These diameters represent the most
commonly used sizes and provide an opportunity to determine the rate of corrosion as a function
of rebar diameter. Three samples of each size and material were tested to ensure the repeatability
of the results. As mentioned in the next subsection, the samples were tested at four different levels
of corrosion. Therefore, in total, 4 (steel types) × 3 (sizes) ×4 (corrosion degrees) ×3 (repetitions)
= 144 steel rebar specimens were tested in this study.
(a) (b) (c) (d) (e)
Fig. 3-2 Surface grain distribution of as received polycrystal CAM SMA rods: (a) PC-1 rod,
(b) PC-2 rod, (c) PC-3 rod, (d) PC-4 rod, (e) PC-5 rod.

32
Table 3-1 Chemical composition of steel rebar.
Steel type Diameter Chemical composition, %
MS
C Mn P S Si
#3 0.21 0.77 0.03 0.03 0.29
#5 0.23 1.05 0.03 0.03 0.29
#10 0.23 1.05 0.03 0.03 0.29
ES
C Mn Ni Si Cr
#3 0.44 0.33 0.15 0.28 0.23
#5 0.43 0.81 0.23 0.23 0.11
#10 0.35 0.79 0.21 0.23 0.11
SS
Cr Ni Cu Mn Mo
#3 22.77 3.69 0.3 1.68 0.18
#5 22.56 4.18 0.35 1.77 0.19
#10 22.65 4.02 0.31 1.65 0.2
XS
Cr Mn Si Cu Ni
#3 9.54 0.68 0.46 0.17 0.09
#5 9.86 0.57 0.43 0.19 0.1
#10 9.51 0.53 0.44 0.14 0.08
To sum up, the materials considered in the long-term corrosion tests included two types of
CAM SMA, namely single crystal CAM bars (SCB), polycrystal CAM plates (PCP); and four
types of steel rebar, namely mild steel (MS), epoxy coated steel (ES), and stainless steel (SS), high
chromium steel (XS). A summary of the materials and dimensions of samples used in the longterm corrosion tests is shown in Table 3-2.

33
Table 3-2 Summary of materials and dimensions of samples used in long-term corrosion
tests .
Material Acronym Dimension of samples
Single crystal CAM SMA bar SCB
Dog-bone bars with a diameter of 12.7
mm in gauge length
Polycrystal CAM SMA plate
exposed to corrosion
PCP
Plates with a thickness of 6 mm
Polycrystal CAM SMA plate not
exposed to corrosion
PCP-C
Mild steel MS For each material, the following three
diameters were considered: U.S. #3 ( =
9.53 mm), U.S. #5 ( = 15.88 mm) and
U.S. #10 ( = 32.26 mm)
High chromium steel XS
Epoxy coated steel ES
Stainless steel SS
The arrangement of test specimens in the corrosion chamber for long-term salt spray
corrosion tests is shown in Fig. 3-3. Using ASTM B117 [93] as a reference to determine the details
of the exposure conditions, a fine mist of 5% by weight NaCl solution was sprayed constantly at
40 °C in a WEICE WTC/A160 cyclic corrosion chamber. The average rate of salt spray was 1.34
mL/hour. As shown in Fig. 3-3, a funnel was put on top of a measuring cylinder, and the solution
collected from the spray within a unit time duration was used to determine this salt spray rate. A
red corrosion resistant coating was used to prevent corrosion of the ends of the specimens that
were later gripped for mechanical testing.

34
Fig. 3-3 Arrangement of specimens in corrosion chamber for long-term salt spray corrosion.
The timeline of the long-term corrosion tests and mechanical tests is shown in Fig. 3-4. Longterm corrosion of CAM SMAs was performed in two phases: first, the specimens were subjected
to a continuous salt spray in a corrosion chamber for 296 days; next, the salt spray was stopped,
and the specimens were removed (without cleaning) to normal laboratory environment until 1051
days. The corrosion of steel rebar (MS, XS, ES and SS) was performed in a single phase: they
were subjected to salt spray in a corrosion chamber for 296 days. It is noted that the PCP and SCB
specimens were taken out of the corrosion chamber at predetermined days, cleaned, mechanically
tested without failure, and then put back into the chamber for subsequent corrosion and mechanical
testing. On the other hand, the steel rebar specimens that had reached the predetermined days were
taken out, cleaned, and monotonically tested until failure. Mass loss and stress-strain curves were
collected from these specimens before disposal.
ES SS XS
mm
mm
#5 steel (d=15.88mm) #3 steel (d=9.53mm)
MS ES SS XS
#10 steel (d=32.26mm)
140mm
PCP
SCB
CAM SEA
mm
MS
ES
SS
XS
250mm
MS
Corrosion
chamber
Salt spray nozzle
Spray rate measuring cylinder

35
Fig. 3-4 Timeline of long-term corrosion and mechanical tests.
As shown in Fig. 3-4, when different materials reached predetermined corrosion days,
gravimetric measurements were taken to calculate their mass loss percent. The mass loss percent
was obtained as the ratio of mass loss at a given age of exposure and the original mass at day zero.
At each of the predetermined corrosion days, specimens were cleaned as per ASTM G1 [94] using
the following steps: (1) rinse in deionized water; (2) scrub with plastic bristle brush to loosen the
corrosion products; (3) clean in an ultrasonic bath for 15 minutes; (4) place in a 500 mL solution
of 37% by weight HCl for 5 min; (5) remove HCl by rinsing with deionized water and then acetone;
and (6) air-dry prior to the mass loss measurements. Steps 1 to 3 were repeated until the mass loss
value became constant. A Branson CPX8800H ultrasonic bath was used to clean SCB, PCP, #3
and #5 steel rebar. For #10 steel rebar, no ultrasonic bath was used in the third step of cleaning
because the specimens were too large to fit into the bath. After thorough cleaning, gravimetric
measurements were conducted to determine the mass loss and mass loss percentage. For SCB,
PCP, and #3 steel rebar, a Mettler Toledo ME303E scale with 320 g capacity and 1 mg sensitivity
Day 0 Day 9 Day 20 Day 30 Day 75 Day 296 Day 1051
Corrosion chamber exposure
Mechanical tests on all corroded steel
specimens were completed at Day 296
Ambient air exposure
PCP
PCP-C
SCB
MS
XS
E S
SS
PCP
---
---
---
---
---
---
PCP
---
SCB
MS
XS
E S
SS
PCP
---
SCB
---
---
---
---
PCP
---
SCB
MS
XS
E S
SS
---
---
---
MS
XS
E S
SS
PCP
PCP-C
SCB
All corroded CAM specimens were
removed from the chamber and kept in
ambient lab conditions after Day 296
Note: Corrosion time not drawn to scale Gravimetric measurements Mechanical tests

36
was used. For #5 and #10 steel rebar, a Cole-Parmer SK-10000-53 scale with 15 kg capacity and
0.5 g sensitivity was used.
An MTS 370.5 dynamic servo-hydraulic frame was used to test all the CAM SMA specimens
and four types of steel rebar except for #10 rebar. The #10 steel rebar was tested by a static servohydraulic SATEC model 600HVL universal testing machine. The tensile force was directly applied
to all specimens and an extensometer was used to measure the deformation on the samples. Due
to the different specimen lengths, three types of extensometers were used in the long-term
corrosion tests: (a) Instron Model No. 2630-115 with 50.8 mm gauge length; (b) Epsilon Model
No. 3542-0200-050-ST with 50.8 mm gauge length; and (c) Epsilon Model No. 3543-0400-400TST with 203.2 mm gauge length. For all SCB and PCP with exposure duration less than 75 days,
extensometer (a) was used; for all SCB and PCP with exposure duration of 1051 days,
extensometer (b) was used. Regarding steel rebar, for all #3 and #5 steel rebar with exposure
duration less than 75 days, the extensometer (a) was used; for all #3 and #5 steel rebar with
exposure duration of 296 days, extensometer (b) was used; for all #10 steel rebar, extensometer
(c) was used.
For PCP and SCB, cyclic tension tests were performed. Before the corrosion test, all CAM
specimens were trained at room temperature to stabilize the martensitic transformation. The
training consisted of five tensile cycles with increasing strain amplitudes from 1% to 5% at 1%
strain increment. The loading rate was 0.4 mm/sec. The cyclic loading tests on PCP and SCB were
conducted in 1% strain increments up to a maximum strain amplitude, and at the end of each cycle,

37
specimens were unloaded to a near-zero force. The maximum strain amplitude and the control
method of each test on CAM SMA varied. All tests up to 75 days of exposure were conducted in
displacement control with a maximum strain amplitude of 4%. All the tests at 1051 days were
conducted in strain control with maximum strain amplitude of 5%; after 5% strain cycle was
completed, the specimens were monotonically stretched to failure. For the four types of steel rebar,
monotonic loading tests was applied to the bars directly until rupture occurred. The loading rate
for testing #3, #5 and #10 bars was 0.01 mm/sec, 0.0043 mm/sec, and 0.0043 mm/sec, respectively,
according to ASTM A370 [95]. In this study, all stress values were computed based on the nominal
cross-sectional area of the specimens in uncorroded conditions.
3.2.2 Electrochemical corrosion tests
In the electrochemical corrosion tests, commonly used binary Ni-Ti SMA was considered to
make comparison with CAM SMA and four types of steel rebar. The composition of NiTi SMAs
is 55.93Ni-Ti (wt. %), obtained from the same manufacturer of CAM SMA: Furukawa Techno
Material Co., Ltd. As shown in Table 3-3, six materials were considered in the electrochemical
corrosion tests: single crystal CAM SMA, polycrystal NiTi SMA, MS, ES, SS, and XS. All six
materials were firstly machined into 1 cm3
cubes and then embedded into an epoxy resin mount to
leave one side of the cube exposed with a surface area of 1 cm2
area. The exposed surface was
polished before submerging into the solution. Three specimens were prepared for each material to
ensure the repeatability of the results.

38
Table 3-3 Summary of materials and dimensions of samples used in
electrochemical corrosion tests.
Material Acronym Dimension of samples
Single crystal CAM SMA CAM
For each material, a 1 cm3
cube
sample was used
NiTi SMA NiTi
Mild steel MS
High chromium steel XS
Epoxy coated steel ES
Stainless steel SS
The electrochemical corrosion test setup designed according to ASTM G5 [96] is shown in
Fig. 3-5. The electrochemical measurements were conducted at 30 ℃ inside a MultiPort Gamry
Corrosion Cell containing 800 ml of 3.5% by weight NaCl solution. A WBE10 PolyScience water
bath was used to maintain a constant temperature of the NaCl solution. An INTERFACE1000E
Gamry Potentiostat was used to perform the tests and collect data. A classic three-electrode cell
was used for electrochemical corrosion tests where platinum was used as the counter electrode,
standard Ag/AgCl electrode was used as the reference electrode and the specimen as the working
electrode. Prior to the electrochemical tests, the specimens were submerged in 1100 ml of 1M
NaCl solution inside a 30 ℃ water bath for 1 hour and the open circuit corrosion potential (OCP)
was determined. Then a polarization from -200 mV to +200 mV with respect to OCP was applied
to the specimens with a constant scan rate of 0.33 mV/s.

39
Fig. 3-5 Setup for electrochemical corrosion tests.
3.3 Results and discussion of long-term corrosion tests
3.3.1 Corrosion surface condition
The surface conditions of corroded SCB and PCP before cleaning is shown in Fig. 3-6. The
corrosion of both SCB and PCP started with speckles. Some visible dull red speckles were
observed after nine days of exposure when the mass loss was approximately 0.15%. Then, the
speckled corrosion areas gradually merged into some bigger spots. As shown in Fig. 3-6 (a) and
(b), after 20 days of exposure at a mass loss of approximately 0.2%, some obvious local corrosion
spots were observed, but the spots on PCP were smaller and more uniform than those on SCB. As
corrosion progressed, the local spots turned into some deep-black solid and crumbly corrosion
products, which easily fell off with a brush. After 75 days of exposure at a mass loss of
Counter electrode
Reference electrode
Working electrode
Corrosion cell Water bath
Potentiostat

40
approximately 1%, the surface of both PCP and SCB were covered with dark corrosion products,
as shown in Fig. 3-6 (c) and (d).
(a) (b)
(c) (d)
Fig. 3-6 Surface conditions of CAM SMA specimens prior to cleaning: (a) SCB after 20 days
of exposure, (b) PCP after 20 days of exposure, (c) SCB after 75 days of exposure, and (d)
PCP after 75 days of exposure.
10 mm 10 mm
10 mm
10 mm

41
(a)
(b)
Fig. 3-7 Corrosion surface of two CAM SMA specimens at different ages after cleaning: (a)
SCB-2, and (b) PCP-2. Number inside the parenthesis indicates the corresponding mass loss
percentage.
Day 9
(0.16%)
Day 30
(0.31%)
Day 75
(0.93%)
Day 296
(4.24%)
Day 1051
(5.80%)
Day 0
(0%)
10 mm
Day9
(0.16%)
Day 20
(0.16%)
Day 30
(0.16%)
Day 75
(1.01%)
Day 296
(6.82%)
Day 1051
(12.10%)
Day0
(0%)
10 mm

42
The surface conditions of SCB and PCP after cleaning are shown in Fig. 3-7. Two typical
specimens, SCB-2 and PCP-2 were selected for this presentation while the rest of the CAM
specimens showed similar surface patterns. Number in the parenthesis indicates the mass loss
percentage after certain exposure durations. From the clean surfaces shown in Fig. 3-7, it is
confirmed that the corrosion on both PCP and SCB did not occur uniformly. The corrosion initiated
in some local pits and grew bigger and deeper, which is consistent with Fig. 3-6. Some large
corrosion spots were observed after 296 days of exposure when the mass loss of SCB-2 and PCP2 reached 4.24% and 6.82%, respectively. At final corrosion stage after 1051 days at a mass loss
of 5.80% and 12.10%, respectively, for SCB-2 and PCP-2, both SCB and PCP corroded seriously,
the local corrosion pits merged along the length, and the cross-sectional area decreased
accordingly.
It is worth noting that two of the specimens broke without an application of external force
before (PCP-3) or during (SCB-1) the cleaning at 1051 days, as shown in Fig. 3-8. These two
specimens were different in that they broke without any external force after corrosion while the
rest of the specimens were able to withstand the subsequent cyclic mechanical loading. This
problem is attributed to the localized corrosion concentrated on the surface defects of the
specimens, such as grain boundaries and shear slip band induced by cyclic loading. Since all CAM
SMA specimens were placed back into the corrosion chamber after cyclic loading at predetermined
days, the progressive deepening of the localized corrosion can lead to propagation of microcracks
that have occurred during previous tests and penetration along the intrusion of slip bands on SCB-

43
1 or the intergranular regions on PCP-3. This interpretation is consistent with the 57° fracture
surface angle observed for SCB-1 with respect to the loading direction as shown in Fig. 3-8 (a).
According to Kato et al. [97], the cyclic loading of CAM SMAs results in a 57° angle of residual
martensite habit plane trace. However, the frequency of occurrence of this phenomenon is unclear,
and the dependence of such a localized corrosion crack growth on the crystal orientation of CAM
SMAs remains unknown. Further investigation on this matter is required.
Fig. 3-8 Two CAM specimens that broke naturally at 1051 days.
The typical corrosion surface of four types of steel rebar after 296 days of exposure and
cleaning is shown in Fig. 3-9. It is seen in Fig. 3-9 that after 296 days of exposure, MS and XS
corroded more severely than ES and SS. Furthermore, the rebar with a smaller diameter showed a
higher corrosion than rebar with larger diameter. The corrosion on MS was found to be relatively
uniform; however, the corrosion on XS specimens was highly inhomogeneous. Apart from the
significant cross-sectional area reduction along the whole specimen, some very deep pits were
observed. The interconnected deep pits led to a porous corrosion surface on the XS specimens.
The similar porous surface was also observed for XS by Nachiappan et al. [98]. The ES and SS
Broken during cleaning after 1051 days
(b) PCP-3
(a) SCB-1
Broken before cleaning after 1051 days
20 mm
57°

44
specimens were both in good surface condition after corrosion, almost no damage could be
observed on the surface after cleaning.
(a) MS (b) XS
(c) ES (d) SS
Fig. 3-9 Surface conditions of four types of steel rebar after 296 days of corrosion after
cleaning: (a) MS, (b) XS, (c) ES, and (d) SS.
#3 #5 #10
(21.91%) (10.91%) (6.44%)
10 mm
#3 #5 #10
(14.88%) (9.26%) (2.04%)
10 mm
#3 #5 #10
(0.25%) (0.17%) (0.09%)
10 mm
#3 #5 #10
(0.23%) (0.17%) (0.06%)
10 mm

45
3.3.2 Mass loss analyses
The mass loss of CAM SMA and steel rebar over the testing duration is shown in Fig. 3-10.
Original data of the mass loss for different materials can be found in Appendix A. In Fig. 3-10, it
is seen that during the salt spray exposure up to 296 days, the MS shows the highest mass loss, but
the mass loss values show large dependence on the rebar diameter, the mass loss of #3 MS is
around three times that of #10 MS. The mass loss of XS is smaller than that of MS but it shows
similar diameter dependency. The mass loss of #3 XS is around 60% of #3 MS. The dependence
of corrosion on rebar diameter is consistent with previous research [33,99–102]. When the steel
rebar is exposed to corrosion, the depth of corrosion attack remains constant, therefore, the smaller
diameter rebar shows a higher percentage of volume loss and thus a higher percentage of mass loss
[101].
The ES and SS showed negligible mass loss (around 1/100 of MS and XS) throughout the
testing program, which agrees with the visual observations in Fig. 3-6. The reason why ES and SS
specimens show a large scatter in their mass loss is that their mass loss is very small compared to
their initial weight. The mass loss of PCP is close to the #3 XS and it is around twice that of the
SCB. During corrosion in ambient laboratory conditions from 296 days to 1051 days, the mass
loss of PCP is almost doubled (increased by 82%) while that of SCB increased by 40%.

46
The average mass loss of six types of materials at 296 days of exposure is normalized by the
mean value of #3 MS and compared in Fig. 3-11. It is seen that the mass loss of PCP is about the
(a) (b)
(c) (d)
(e) (f)
Fig. 3-10 Mass loss of six materials during long-term corrosion tests: (a) MS, (b) SS, (c) ES,
(d) SS, (e) PCP, and (f) SCB.
0.0
4.0
8.0
12.0
16.0
20.0
24.0
0 20 75 296
Mass loss (%)
Exposure duration (days)
MS#3 MS#5 MS#10
Corrosion chamber
exposure
0.0
4.0
8.0
12.0
16.0
20.0
24.0
0 20 75 296
Mass loss (%)
Exposure duration (days)
XS#3 XS#5 XS#10
Corrosion chamber
exposure
0.00
0.08
0.16
0.24
0.32
0.40
0 20 75 296
Mass loss (%)
Exposure duration (days)
ES#3
ES#5
ES#10
Corrosion chamber
exposure
0.00
0.08
0.16
0.24
0.32
0.40
0 20 75 296
Mass loss (%)
Exposure duration (days)
SS#3
SS#5
SS#10
Corrosion chamber
exposure
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
0 9 20 30 75 296 1051
Mass loss (%)
Exposure duration (days)
PCP
Corrosion chamber
exposure
Ambient air
exposure
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
0 9 20 30 75 296 1051
Mass loss (%)
Exposure duration (days)
SCB
Corrosion chamber
exposure
Ambient air
exposure

47
same as that of #10 MS and it is around 1/3 of that of the #3 MS. The mass loss percent of SCB
has been observed to be around 1/5 of #3 MS.
Fig. 3-11 Ranked normalized mass loss of PCP, SCB, MS, XS, ES, and SS after 296 days
of exposure.
3.3.3 Mechanical test results of SCB
The stress-strain curves of SCB at different levels of corrosion are shown in Fig. 3-12. The
results shown in Fig. 3-12 for 1051 days of exposure exclude the final failure loading so that the
results can be compared clearly. The full stress-strain curves are shown in Fig. 3-15. It is observed
from Fig. 3-12 that, in early stages up to 75 days of corrosion, there is no obvious deterioration of
superelasticity of SCB. The hysteretic loops remain flag-shaped, and the mechanical behavior is
almost unchanged. At 1051 days, the hysteresis loops are still flag-shaped but become slightly
narrower.

48
(a)
(b)
(c)
Fig. 3-12 Mechanical test results of SCB at different levels of corrosion: (a) SCB-1, (b) SCB2, and (c) SCB-3.
Four critical mechanical properties that represent the superelasticity of SCB were extracted
from the stress strain curves, namely martensitic transformation start stress (also referred to as
yield stress), Ms, austenitic transformation finish stress, Af, stress hysteresis, hy, and residual
strain, re. The definition of these parameters is shown in Fig. 3-13. To present a more intuitive
comparison of the effect of long-term corrosion, these parameters were normalized by the test
results obtained before subjecting the specimens to corrosion exposure. The original data of mass
loss and mechanical properties of SCB before normalization can be found in Appendix A.1.
0
100
200
300
Stress (MPa)
0 5 (0) 5 (0) 5 (0)
5% Strain
Day 0
(0%)
Day 20
(0.25%)
Day 30
(0.30%)
Day 75
(0.87%)
Strain (%)
5 (0) 5
Day 1051
(Broken)*
0
100
200
300
Stress (MPa)
0 5 (0) 5 (0) 5 (0)
5% Strain
Day 0
(0%)
Day 20
(0.28%)
Day 30
(0.31%)
Day 75
(0.93%)
Strain (%)
5 (0) 5
Day 1051
(5.80%)
0
100
200
300
Stress (MPa)
0 5 (0) 5 (0) 5 (0)
5% Strain
Day 0
(0%)
Day 20
(0.29%)
Day 30
(0.34%)
Day 75
(1.03%)
Strain (%)
5 (0) 5
Day 1051
(5.50%)

49
Fig. 3-13 Definition of critical mechanical properties of superelastic CAM SMA.
Fig. 3-14 shows the normalized mechanical properties of SCB with increasing mass loss.
From Fig. 3-14 (a) to (d), it is seen that with 6% mass loss, the Ms, norm decreases around 10%
while Af, norm increases slightly. The decrease of Ms, norm and increase of Af, norm led to a 40%
reduction of the hy, norm of SCB. The re, norm shows large scatter but it is noted that the nonnormalized residual strains for the three specimens are all less than 0.1%, as shown in Fig. 3-12.
(a) (b)
Ms
re
Strain
Stress
T>Af
0
Af
hy= Ms- Af
Ms Martensitic transformation start stress
Af Austenitic transformation finish stress
hy Stress hysteresis
re Residual strain
0.0
0.5
1.0
1.5
0.00 2.00 4.00 6.00 8.00
Ms, norm
Mass loss (%)
SCB-1
SCB-2
SCB-3
0.0
0.5
1.0
1.5
0.00 2.00 4.00 6.00 8.00
Af, norm
Mass loss (%)
SCB-1
SCB-2
SCB-3

50
(c) (d)
Fig. 3-14 Variation of mechanical properties of SCB at different levels of corrosion: (a) Ms,
norm, (b) Af, norm, (c) hy, norm and (d) re, norm.
The full-range of the stress-strain curve of corroded SCB after 1051 days of exposure are
shown in Fig. 3-15. An ‘x’ mark indicates the fracture of specimens at the end of the loading.
Because SCB-1 was broken during cleaning at 1051 days of exposure (see Fig. 3-8), only two SCB
were tested: SCB-2 and SCB-3. It is seen in Fig. 3-15 that these two SCB show still excellent
superelasticity and ductility after long-term corrosion. The rupture strain of SCB-2 and SCB-3
were obtained as 25% and 36% after 1051 days of exposure. The pictures of SCB-2 and SCB-3
after fracture are shown in Fig. 3-16. Ductile necking is observed in the rupture zone.
0.0
0.5
1.0
1.5
0.00 2.00 4.00 6.00 8.00
hy, norm
Mass loss (%)
SCB-1
SCB-2
SCB-3
0.0
0.5
1.0
1.5
0.00 2.00 4.00 6.00 8.00
re, norm
Mass loss (%)
SCB-1
SCB-2
SCB-3

51
(a) (b)
Fig. 3-15 Full range of stress-strain curves after 1051 days of exposure: (a) SCB-2, and (b)
SCB-3.
(a) (b)
Fig. 3-16 Failure tests of SCB after 1051 days of exposure: (a) SCB-2, and (b) SCB-3.
0 8 16 24 32 40
Strain (%)
0
100
200
300
400
500
Stress (MPa)
Day 1051
(5.80%)
Strain (%)
0
100
200
300
0 1 2 3 4 5 6
Stress (MPa)
0 8 16 24 32 40
Strain (%)
0
100
200
300
400
500
Stress (MPa)
Day 1051
(5.50%)
Strain (%)
0 1 2 3 4 5 6
0
100
200
300
Stress (MPa)

52
3.3.4 Mechanical test results of PCP and PCP-C
The stress-strain curves of exposed PCP are shown in Fig. 3-17. The results for 1051 days
shown in Fig. 3-17 are partial responses from 0 to 5% strain while the complete stress-strain curves
are shown later. It is seen from Fig. 3-17 that up to 75 days of exposure, the maximum strain,
transformation stress, and stress hysteresis of PCP vary from specimen to specimen. The main
reason is that each PCP has a unique grain size distribution and orientation. The grain distribution
within the extensometer gauge length was different for each specimen, thereby, leading to differing
transformation stress, modulus, and superelastic strain among different specimens [41,56].
Additionally, because the elastic modulus of CAM SMA depends on the grain orientation [11,103],
the displacement controlled loading protocol resulted in different strain levels on the specimens.
Regarding the effect of corrosion, the five exposed PCP showed similar results to those of the
SCB: in early corrosion up to 75 days, the superelastic behavior of PCP remained unchanged,
except some minor deviations in the residual strain; at 1051 days, a remarkable decrease in yield
stress (i.e., the martensitic transformation start stress) was observed due to the reduction of the
effective cross-sectional area.

53
(a)
(b)
(c)
(d)
(e)
Fig. 3-17 Mechanical test results of PCP at different levels of corrosion: (a) PCP-1, (b) PCP-2,
(c) PCP-3, (d) PCP-4, and (e) PCP-5.
0
100
200
300 Stress (MPa)
0 5 (0) 5 (0) 5 (0) 5 (0)
5% Strain
Strain (%)
5 (0) 5
Day 0
(0%)
Day 9
(0.21%)
Day 20
(0.21%)
Day 30
(0.21%)
Day 75
(1.11%)
Day 1051
(12.17%)
0
100
200
300 Stress (MPa)
0 5 (0) 5 (0) 5 (0) 5 (0)
5% Strain
Strain (%)
5 (0) 5
Day 0
(0%)
Day 9
(0.16%)
Day 20
(0.16%)
Day 30
(0.16%)
Day 75
(1.01%)
Day 1051
(12.10%)
0
100
200
300 Stress (MPa)
0 5 (0) 5 (0) 5 (0) 5 (0)
5% Strain
Strain (%)
Day 0
(0%)
Day 9
(0.21%)
Day 20
(0.21%)
Day 30
(0.21%)
Day 75
(1.01%)
5 (0) 5
Day 1051
(Broken)*
0
100
200
300 Stress (MPa)
0 5 (0) 5 (0) 5 (0) 5 (0)
5% Strain
Strain (%)
5 (0) 5
Day 0
(0%)
Day 9
(0.23%)
Day 20
(0.23%)
Day 30
(0.23%)
Day 75
(1.11%)
Day 1051
(15.56%)
0
100
200
300 Stress (MPa)
0 5 (0) 5 (0) 5 (0) 5 (0)
Day 0
(0%)
Day 9
(0.19%)
Day 20
(0.19%)
Day 30
(0.19%)
Day 75
(0.97%)
Strain (%)
5 (0) 5
Day 1051
(10.39%)
5% Strain

54
(a) (b)
(c) (d)
Fig. 3-18 Mechanical properties of PCP at different levels of corrosion: (a) Ms, norm, (b) Af,
norm, (c) hy, norm, and (d) re, norm.
The normalized mechanical properties of PCP with increasing mass loss are shown in Fig.
3-18. The original data of PCP before normalization can be found in Appendix A.2. It is seen in
Fig. 3-18 (a) that the corrosion leads to an obvious decrease in Ms, norm of PCP, but the level of
decreases varies from specimen to specimen. For PCP-4, the mass loss was around 10% after 296
days of exposure, and the Ms, norm decreased by 50%; for PCP-1 and PCP-3, the mass loss was
around 12% after 296 days of exposure, but the Ms, norm decreased by 20%. The effect of corrosion
on Af, norm is similar to Ms, norm, as shown in Fig. 3-18 (b). More than 50% of decrease in Af, norm
was seen after 296 days of exposure but large scatter in data exists. The hy, norm of PCP shown in
Fig. 3-18 (c) did not show an obvious decrease, the hy, norm of PCP-1, PCP-2 and PCP-4 even
0.0
0.5
1.0
1.5
2.0
0.00 5.00 10.00 15.00 20.00
Ms, norm
Mass loss (%)
PCP-1
PCP-2
PCP-3
PCP-4
PCP-5
0.0
0.5
1.0
1.5
2.0
0.00 5.00 10.00 15.00 20.00
Af, norm
Mass loss (%)
PCP-1
PCP-2
PCP-3
PCP-4
PCP-5
0.0
1.0
2.0
3.0
4.0
0.00 5.00 10.00 15.00 20.00
hy, norm
Mass loss (%)
PCP-1
PCP-2
PCP-3
PCP-4
PCP-5
0.0
2.0
4.0
6.0
8.0
10.0
0.00 5.00 10.00 15.00 20.00
re, norm
Mass loss (%)
PCP-1
PCP-2
PCP-3
PCP-4
PCP-5

55
showed an increasing trend. The effect of corrosion on re, norm shown in Fig. 3-18 (d) was not
apparent. But from the non-normalized data shown in Fig. 3-17, the residual strain of all PCP
specimens was less than 0.65%.
(a) (b)
(c) (d)
Fig. 3-19 Full range stress-strain curves after 1051 days of exposure: (a) PCP-1, (b) PCP-2,
(c) PCP-4 and (d) PCP-5.
The full-range of the stress-strain curves of the samples at 1051 days of exposure are shown
in Fig. 3-19. Because PCP-3 was broken before cleaning at 1051 days, only four exposed PCP
were tested. As seen in Fig. 3-19, all four PCP showed superelastic behavior after more than 10%
mass loss. The hysteresis loops up to 5% strain are flag-shaped, and all four exposed PCP ruptured
0 4 8 12 16
Strain (%)
0
100
200
300
400
500
Stress (MPa)
Day 1051
(12.17%)
0 1 2 3 4 5 6
Strain (%)
0
100
200
300
Stress (MPa)
0 4 8 12 16
Strain (%)
0
100
200
300
400
500
Stress (MPa)
Day 1051
(12.10%)
0 1 2 3 4 5 6
Strain (%)
0
100
200
300
Stress (MPa)
0 4 8 12 16
Strain (%)
0
100
200
300
400
500
Stress (MPa)
Day 1051
(15.56%)
0 1 2 3 4 5 6
Strain (%)
0
100
200
300
Stress (MPa)
0 4 8 12 16
Strain (%)
0
100
200
300
400
500
Stress (MPa)
Day 1051
(10.39%)
0 1 2 3 4 5 6
Strain (%)
0
100
200
300
Stress (MPa)

56
at over 8% strain. Typical pictures of the two corroded PCP (PCP-4 and PCP-5) after fracture are
shown in Fig. 3-20. No visible necking was observed around the rupture zone.
(a) (b)
(c) (d)
Fig. 3-20 Failure of PCP and PCP-C after 1051 corrosion exposure: (a) Exposed PCP-4, (b)
Unexposed PCP-4C, (c) Exposed PCP-5, and (d) Unexposed PCP-5C.

57
The unexposed PCP-C and exposed PCP are compared to investigate the effect of corrosion
on the fracture behavior of polycrystal CAM SMA. The test results of unexposed PCP
counterparts: PCP-C are shown in Fig. 3-21. A comparison of Fig. 3-19 and Fig. 3-21 indicates
that PCP and PCP-C showed comparable superelasticity and deformability. The long-term
corrosion mainly affects the martensitic transformation stress and plastic deformation stress of
CAM SMAs, which is caused by the loss of effective cross-sectional area. Typical pictures of the
two unexposed PCP-C (PCP-4C, PCP-5C) after fracture are shown in Fig. 3-20 (b) and (d). The
PCP-4C shown in Fig. 3-20 (b) failed due to stress concentrations in one of the grips. The fracture
of PCP-5C shown in Fig. 3-20 (d) is ductile with obvious necking in the rupture zone.
(a) (b)
(c) (d)
0 4 8 12 16
Strain (%)
0
100
200
300
400
500
Stress (MPa)
0 1 2 3 4 5 6
Strain (%)
0
100
200
300
Stress (MPa)
0 4 8 12 16
Strain (%)
0
100
200
300
400
500
Stress (MPa)
0 1 2 3 4 5 6
Strain (%)
0
100
200
300
Stress (MPa)
0 4 8 12 16
Strain (%)
0
100
200
300
400
500
Stress (MPa)
0 1 2 3 4 5 6
Strain (%)
0
100
200
300
Stress (MPa)
0 4 8 12 16
Strain (%)
0
100
200
300
400
500
Stress (MPa)
0 1 2 3 4 5 6
Strain (%)
0
100
200
300
400
Stress (MPa)

58
(e)
Fig. 3-21 Full range of stress-strain curves of unexposed samples: (a) PCP-1C, (b) PCP-2C,
(c) PCP-3C, (d) PCP-4C, and (e) PCP-5C.
The rupture strain and ultimate stress of PCP and PCP-C are extracted and plotted in Fig.
3-22. From Fig. 3-22 (a), it is seen that the deformability of polycrystal CAM SMA did not
necessarily decrease after long-term corrosion. PCP-1, PCP-4 and PCP-5 showed almost the same
or even higher rupture strains than their unexposed counterparts. From Fig. 3-22 (b), it is seen that
the ultimate stress deteriorated after corrosion, which coincides with the drop in yield stress
resulting from the reduction of the cross-sectional area. It should be noted that the PCP-4 showed
an unexpected increase than PCP-4C, because PCP-4 fractured before entering the plastic range
due to stress concentration in the grips as shown in Fig. 3-20 (b).
0 4 8 12 16
Strain (%)
0
100
200
300
400
500
Stress (MPa)
0 1 2 3 4 5 6
Strain (%)
0
100
200
300
Stress (MPa)

59
3.3.5 Mechanical test results of steel rebar
Fig. 3-23 shows the stress-strain curves of MS and XS at predetermined dates of corrosion.
The numbers inside parenthesis in legend indicate the mass loss. Because only two decimal places
were kept, the mass loss of specimen less than 0.01% were expressed as 0.00% in the legend. From
Fig. 3-23 (a) to (c), it is seen that the stress strain curves of #3 MS show more significant
degradation than #5 MS and #10 MS. In terms of ultimate stress, the #3 MS decreases around 40%
after 296 days of exposure, while the decrease of #5 MS and # MS is around 20% and 10%,
respectively. The stress-strain curves of XS show similar trend to MS: smaller diameter behave
more degradation. The ultimate stress of #3 XS degrades around 45% after 296 days of exposure,
while the degradation of ultimate stress of #5 XS is 20%, as shown in Fig. 3-23 (d) and (e). The
stress-strain curves of #10 XS show some fluctuation, see Fig. 3-23 (f), but the overall degradation
is smaller than #5 XS and #3 XS.
(a) (b)
Fig. 3-22 Compassion of rupture strain and plastic stress between PCP and PCP-C: (a) Rupture
strain, and (b) Ultimate stress.
0.0
4.0
8.0
12.0
16.0
1 2 3 4 5
Rupture strain (%)
Specimen No.
Rupture strain of PCP-i and PCP-iC
PCP
PCP-C
0
100
200
300
400
500
1 2 3 4 5
Ultimate stress (MPa)
Specimen No.
Plastic stress of PCP-i and PCP-iC
PCP
PCP-C

60
(a) (d)
(b) (e)
(c) (f)
Fig. 3-23 Stress-strain curves of MS and XS at predetermined levels of corrosion: (a) #3MS,
(b) #5MS, (c) #10MS, (d) #3XS, (e) #5XS, and (f) #10XS. Note: numbers inside parenthesis
indicate the mass loss percentage.

61
Fig. 3-24 shows the stress-strain curves of ES and SS at predetermined dates of corrosion.
From ES shown in Fig. 3-24 (a) to (c) and SS shown in Fig. 3-24 (d) to (f), it is seen that the stress
strain curves of all ES and SS specimens show little degradation after corrosion, because their
mass loss is very small (less than 0.5%). Besides, there is no obvious difference of mass loss or
stress-strain degradation between #3 diameter bars and #10 diameter for ES and SS, the influence
of rebar diameter on stress-strain curves of ES and SS is trivial.
In a word, from Fig. 3-23 and Fig. 3-24, it can be concluded that the overall degradation trend
of the stress-strain curves of four types of steel rebar is consistent with the magnitude of their mass
loss. The bigger the mass loss, the more their yield stress and ultimate stress degrades. The reason
is that the more the mass loss, the more cross-section area loss will be. This study uses engineering
stress based on nominal cross section area, thereby leading to smaller calculation of yield stress
and ultimate stress.
Based on the above mechanical tests, four key properties are extracted, namely elastic
modulus, yield stress, maximum stress, and rupture strain. A reduction in the elastic modulus, yield
stress and maximum stress could affect the stiffness and strength while the rupture strain is an
important indicator of ductility. These mechanical properties serve as important benchmarks for
evaluating CAM SMA. The determination of yield stress follows the standard recommendations
in ASTM E8 [104] where a 0.2% offset method is used for steels that do not have an obvious yield
plateau.

62
(a) (d)
(b) (e)
(c) (f)
Fig. 3-24 Stress-strain curves of ES and SS at predetermined levels of corrosion: (a) #3ES,
(b) #5ES, (c) #10ES, (d) #3SS, (e) #5SS, and (f) #10SS. Note: numbers inside parenthesis
indicate the mass loss percentage.

63
To present a more intuitive comparison with CAM SEAs, the average of the three test results
obtained before subjecting the specimens to corrosion exposure was used to normalize the
mechanical properties. The normalized mechanical properties are: elastic modulus, E, norm, yield
stress, fy, norm, maximum stress, fm, norm, and rupture strain, r, norm. The original data of mechanical
properties of MS, XS, ES and SS specimens before normalization are compiled in Appendix A.3
to A.6, which form a detailed experimental database for further research.
The mechanical properties of MS with respect to mass loss are shown in Fig. 3-25. A linear
curve fit to the data points are also shown along with the coefficient of determination, R
2
, values.
It is seen from Fig. 3-25 (a) and (b) that the fy, norm and fm, norm of MS almost show a perfectly linear
decreasing trend with increasing mass loss, the R
2
of these three parameters are all over 0.88 for
three diameter specimens. Besides, the slope of decreasing of #3MS is generally larger than the #5
and #10MS. Take fm, norm for instance, with 1% of mass loss, the fm, norm of #3MS, #5MS and #10MS
decreases by 1.71%, 1.32% and 1.41%, respectively. From Fig. 3-25 (c) and (d), it is seen that the
E, norm and r, norm of MS also show a decreasing trend with respect to mass loss, but the relationship
is not well captured by a linear fitting. Except for the E, norm for #3 MS and r, norm for #10 MS with
R
2
larger than 0.8, the R
2
of the rest of linear fitting is all less than 0.5.

64
(a)
(b)
(c)
y = -0.0182x + 0.9866
R² = 0.9774
y = -0.011x + 0.9941
R² = 0.9209
y = -0.0139x + 0.9921
R² = 0.8846
0.00
0.25
0.50
0.75
1.00
1.25
1.50
0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0
fy, norm
Mass loss (%)
(b) MS-fy
#3 MS #5 MS #10 MS
y = -0.0171x + 0.9904
R² = 0.9774
y = -0.0132x + 1.0016
R² = 0.9792
y = -0.0141x + 0.9979
R² = 0.9608
0.00
0.25
0.50
0.75
1.00
1.25
1.50
0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0
fm, norm
Mass loss (%)
(c) MS-fm
#3 MS #5 MS #10 MS
y = -0.0138x + 1.0035
R² = 0.8126
y = -0.0246x + 0.9783
R² = 0.1605
y = -0.0383x + 1.0548
R² = 0.3594
0.00
0.25
0.50
0.75
1.00
1.25
1.50
0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0
E, norm
Mass loss (%)
(a) MS-E
#3 MS #5 MS #10 MS

65
(d)
Fig. 3-25 Mechanical properties of MS at different levels of corrosion: (a) fy, norm, (b) fm, norm,
(c) E, norm, and (d) r, norm.
The mechanical properties of XS with respect to mass loss are shown in Fig. 3-26. The
decreasing trend of fy, norm and fm, norm of XS is similar to that of MS: an almost perfectly linear
relationship exists, but the scatter in the results of XS is larger than MS and it increase for #5 and
#10 XS, as shown in Fig. 3-26 (a) and (b). Take fy, norm for instance, the R
2
of #3MS, #5MS, and
#10MS is 0.98, 0.92, and 0.88, respectively, while the R
2
of #3XS, #5XS, and #10XS is 0.98, 0.79,
0.35, respectively. Besides, the decreasing trend of XS is overall more significant than MS. Take
fm, norm for instance, with 1% of mass loss, the fm, norm of #3XS, #5XS, and #10XS decreases by
3.1%, 2.0% and 2.1%, respectively. From Fig. 3-26 (c) and (d), it is seen that the E, norm and r, norm
of XS show an overall decreasing trend with respect to mass loss, but the linear relationship is
more obvious for #3XS. The R
2 of E, norm and r, norm for #3XS reaches 0.91 and 0.81, respectively.
With 1% of mass loss, the E, norm and r, norm of #3XS decreases by 2.4% and 4.2%, respectively.
y = -0.0216x + 1.1886
R² = 0.4853
y = -0.007x + 1.0275
R² = 0.0149
y = -0.0369x + 0.9892
R² = 0.8077
0.00
0.25
0.50
0.75
1.00
1.25
1.50
0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0
r, norm
Mass loss (%)
(d) MS-e_rup
#3 MS #5 MS #10 MS

66
(a)
(b)
(c)
y = -0.0341x + 1.0035
R² = 0.9809
y = -0.0191x + 0.9901
R² = 0.7856
y = -0.0471x + 0.944
R² = 0.3501
0.00
0.25
0.50
0.75
1.00
1.25
1.50
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00
fy, norm
Mass loss (%)
(b) XS-fy
#3 XS #5 XS #10 XS
y = -0.0306x + 1.0044
R² = 0.9841
y = -0.0203x + 1.0161
R² = 0.9649
y = -0.0207x + 1.0037
R² = 0.7073
0.00
0.25
0.50
0.75
1.00
1.25
1.50
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00
fm, norm
Mass loss (%)
(c) XS-fm
#3 XS #5 XS #10 XS
y = -0.024x + 0.9691
R² = 0.9051
y = -0.0267x + 0.8874
R² = 0.3606
y = -0.0227x + 0.9635
R² = 0.1095
0.00
0.25
0.50
0.75
1.00
1.25
1.50
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00
E, norm
Mass loss (%)
(a) XS-E
#3 XS #5 XS #10 XS

67
(d)
Fig. 3-26 Mechanical properties of XS at different levels of corrosion: (a) fy, norm, (b) fm, norm,
(c) E, norm, and (d) r, norm.
The mechanical properties of ES and SS with respect to mass loss are shown in Fig. 3-27 and
Fig. 3-28, respectively. It is seen that the fy, norm and fm, norm of ES and SS show almost no visible
degradation, which is consistent with the stress-strain curves shown in Fig. 3-24. Besides, the E,
norm and r, norm of ES and SS exhibit large scatter in the data and no obvious degradation trend can
be observed. Therefore, no linear fitting was conducted on the mechanical properties of ES and
SS.
(a)
y = -0.0417x + 0.9098
R² = 0.8052
y = -0.1782x + 1.0482
R² = 0.8928
0.00
0.25
0.50
0.75
1.00
1.25
1.50
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00
r, norm
Mass loss (%)
(d) XS-e_rup
#3 XS #5 XS #10 XS
0.00
0.25
0.50
0.75
1.00
1.25
1.50
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
fy, norm
Mass loss (%)
(b) ES-fy
#3 ES #5 ES #10 ES

68
(b)
(c)
(d)
Fig. 3-27 Mechanical properties of ES at different levels of corrosion: (a) fy, norm, (b) fm,
norm, (c) E, norm, and (d) r, norm.
0.00
0.25
0.50
0.75
1.00
1.25
1.50
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
fm, norm
Mass loss (%)
(c) ES-fm
#3 ES #5 ES #10 ES
0.00
0.25
0.50
0.75
1.00
1.25
1.50
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
E, norm
Mass loss (%)
(a) ES-E
#3 ES #5 ES #10 ES
0.00
0.25
0.50
0.75
1.00
1.25
1.50
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
r, norm
Mass loss (%)
(d) ES-e_rup
#3 ES #5 ES #10 ES

69
(a)
(b)
(c)
0.50
0.75
1.00
1.25
1.50
0.00 0.05 0.10 0.15 0.20 0.25 0.30
fy, norm
Mass loss (%)
(b) SS-fy
#3 SS #5 SS #10 SS
0.50
0.75
1.00
1.25
1.50
0.00 0.05 0.10 0.15 0.20 0.25 0.30
fm, norm
Mass loss (%)
(c) SS-fm
#3 SS #5 SS #10 SS
0.50
0.75
1.00
1.25
1.50
0.00 0.05 0.10 0.15 0.20 0.25 0.30
E, norm
Mass loss (%)
(a) SS-E
#3 SS #5 SS #10 SS

70
(d)
Fig. 3-28 Mechanical properties of SS at different levels of corrosion: (a) fy, norm, (b) fm, norm, (c)
E, norm, and (d) r, norm.
3.4 Results and discussion of electrochemical corrosion tests
3.4.1 Tafel curves
The potentialdynamic polarization curves (also known as Tafel curves) of six materials: MS,
XS. ES, SS, CAM SMA, and NiTi SMA are shown in Fig. 3-29, where Ecorr denotes corrosion
potential, and icorr denotes corrosion current density. From these curves, it is seen that: in terms of
Ecorr, six materials rank from largest to smallest as: MS, ES, NiTi, CAM, XS and SS; while in
terms of icorr, they rank from largest to smallest as: NiTi, SS, XS, CAM, MS, and ES. The CAM
showed a close Ecorr to that of NiTi but its icorr was higher than that of NiTi.
0.50
0.75
1.00
1.25
1.50
0.00 0.05 0.10 0.15 0.20 0.25 0.30
r, norm
Mass loss (%)
(d) SS-e_rup
#3 SS #5 SS #10 SS

71
Fig. 3-29 Potentiodynamic polarization curves of MS, XS, ES, SS, CAM and NiTi.
Based on the potentialdynamic polarization curves, the corrosion potential, Ecorr, and
corrosion current density, icorr, of six materials were extracted, as shown in Fig. 3-30 and Fig. 3-31,
respectively. An average of the three samples was used. The average corrosion potential of MS,
XS. ES, SS, CAM and NiTi were obtained as -453 mV, -113 mV, -446 mV, 8 mV, -160 mV, and

72
-196 mV, respectively. The average corrosion current density of MS, XS. ES, SS, CAM and NiTi
were obtained as 16.26 A/cm2
, 0.31 A/cm2
, 14.37 A/cm2
, 0.08 A/cm2
, 4.75 A/cm2
, and 0.07
A/cm2
, respectively.
Fig. 3-30 Corrosion potential of MS, XS, ES, SS, CAM and NiTi.
(a)
-500
-400
-300
-200
-100
0
100
Ecorr (mV)
MS XS ES SS CAM NiTi
E_corr
Specimen 1
Specimen 2
Specimen 3
0
5
10
15
20
25
30
icorr (uA/cm2
)
MS XS ES SS CAM NiTi
i_corr
Specimen 1
Specimen 2
Specimen 3

73
(b)
Fig. 3-31 Corrosion current density of MS, XS, ES, SS, CAM and NiTi: (a) corrosion
current density, icorr, (b) corrosion current density, icorr within the range from 0-0.5 A/cm2.
3.4.2 Corrosion rate
Based on Faraday’s law and the average corrosion current density, the corrosion rate of six
materials was calculated and ranked in the order from the largest to the smallest as shown in Fig.
3-32. The corrosion rate of CAM was 0.0565 mm/year close to 0.032 mm/year reported in [89].
From Fig. 3-32, it is seen that the corrosion rate of CAM is higher than those of SS and NiTi but
it is only 1/3 of that of MS. This trend is consistent with the mass loss measured from the longterm corrosion tests, as shown in Fig. 3-10 and Fig. 3-11.
0
0.1
0.2
0.3
0.4
0.5
icorr (uA/cm2
)
MS XS ES SS CAM NiTi
i_corr
Specimen 1
Specimen 2
Specimen 3

74
Fig. 3-32 Corrosion rate of MS, XS, ES, SS, CAM and NiTi.
It is worth noting that in the long-term corrosion exposure, the ES showed a comparable mass
loss to SS (see Fig. 3-10). However, in the electrochemical test results, the corrosion rate of ES
(0.1687 mm/year) was close to that of MS (0.1910 mm/year). The reason is that ES relies on the
exterior epoxy coating for corrosion resistance [105]. While in the electrochemical tests, the epoxy
coating was removed, and the base metal was exposed to the solution. Without the protective
exterior physical barrier, ES showed a similar corrosion resistance to that of MS.
The XS in the long-term tests showed surface corrosion (deep corrosion pits and a porous
surface condition as shown in Fig. 3-9) and a high mass loss (see Fig. 3-10), its mechanical
properties also degraded. However, in the electrochemical tests, the corrosion rate of XS (0.0035
mm/year) was close to SS (0.0008 mm/year) and the corrosion rate of both materials was much
lower than that of MS (0.1910 mm/year). This is attributed to the sensitivity of XS to sodium
chloride concentration. According to Nachiappan et al. [98], at 0.1% NaCl concentration, the
corrosion rate of XS is 1/12 of that of MS, but when the concentration of NaCl is increased to 3%,
0.1910
0.1687
0.0565
0.0035 0.0008
0.0006
0.000
0.040
0.080
0.120
0.160
0.200
0.240
Corrosion rate (mm/year)
MS ES CAM XS SS NiTi

75
the corrosion rate of XS increases to 1/2 of that of MS. In this study, the NaCl concentration in the
electrochemical tests is 3.5%, and lower than the 5% used in the salt spray tests. Moreover, the
long-term spray of the fine mist of salt solutions creates a larger contact between chloride ions and
the porous surface of XS, thereby leading to more corrosion.
3.5 Summary of findings
It was found that, during long-term salt spray tests, corrosion on both polycrystal Cu-Al-Mn
SMA plate (PCP) and single crystal Cu-Al-Mn SMA bar (SCB) were local, initiating with local
speckles and then growing to some deep pits. The local corrosion pits merged together along the
length and the cross-sectional area decreased accordingly.
In general, the mass loss of PCP was higher than that of SCB, but both of them were lower
than that of mild steel (MS) and high chromium steel (XS), and higher than that of epoxy coated
steel (ES) and stainless steel (SS). Specifically, after three years of salt spray and ambient air
exposure, the XS was around 60% of MS; the PCP and SCB was around 1/3 of MS; the ES and
SS was only around 1/100 of MS. The mass loss of MS and XS showed large dependency on its
diameter: the corrosion degree increased in order from #3 diameter to #5 diameter and #10
diameter. The ES and SS showed negligible mass loss and their mass loss dependency on rebar
diameter was trivial.
The mechanical properties degradation of Cu-Al-Mn SMA varied from specimen to
specimen. After three years of corrosion, the yield stress of SCB decreased about 20% at mass loss

76
of around 6%; the yield stress of PCP decreased about 40% at mass loss of around 15%. Besides,
because of the different grain distributions, the mechanical property of PCP showed large scatter
among specimens. But it should be noted that after three years of corrosion, both PCP and SCB
still showed excellent superelasticity. Specifically, the superelastic strain recovery showed almost
no degradation for all tested specimens. The fracture strain of one corroded SCB was over 32%.
From the electrochemical corrosion tests, it was seen that the corrosion potential of Cu-AlMn SMA was comparable to that of Ni-Ti SMA and less than that of SS and XS, but its corrosion
current density was close to that of MS and ES. Based on Faraday’s law, the corrosion rate of CuAl-Mn SMA was obtained to be only 1/3 of that of the MS. The corrosion rate of MS, ES, Cu-AlMn SMA, XS, SS and Ni-Ti SMA were respectively obtained as 0.1910 mm/year, 0.1687
mm/year, 0.0565 mm/year, 0.0035 mm/year, 0.0008 mm/year, and 0.0006 mm/year.

77
Chapter 4 - Low-cycle fatigue behavior of Cu-Al-Mn SMA at
different temperatures
The low-cycle fatigue performance of Cu-Al-Mn SMA was studied at room temperature, 25 ℃,
low temperature, -40 ℃, and high temperature, 50 ℃. The temperature range took into account
the annual ambient temperature variation in most parts of the world. Both single crystal and
polycrystal Cu-Al-Mn SMAs were investigated. Strain cycles up to 50,000 was applied at a tensile
strain amplitude of 5%. Comparisons with conventional Ni-Ti SMA under the same low-cycle
fatigue loading conditions were also made. Variations in the superelastic properties were observed
and analyzed, including the stress-strain curves, Young’s modulus, yield stress, damping ratio,
and recovery strain.
4.1 Research motivation
Due to their excellent strain recovery and energy dissipating capacity, SMAs have been widely
studied as column reinforcement, bracing systems, and energy dissipating devices [6,15–17,22]. Such
loading conditions require the SMAs to have stable mechanical behavior under fatigue loading.
According to Eggeler et al. [106], fatigue on SMAs can be categorized into structural fatigue and
functional fatigue. Structural fatigue, also referred to as high-cycle fatigue, typically involves low

78
strain/stress levels and a large number of loading cycles (over 100,000). Research interests in structural
fatigue include crack propagation, energy release rate, and fatigue life of materials. Functional fatigue,
also referred to as low-cycle fatigue, on the other hand, typically focuses on the functional property’s
degradation under high strain/stress levels and a small number of loading cycles (less than 10,000).
Research interests in functional fatigue include the degradation of superelastic and shape memory
effects of SMAs under cyclic loadings. In this dissertation, only low-cycle fatigue was investigated,
because when applying SMAs in bridges, a typical earthquake event may involve hundreds of cyclic
loading with tens of high strain cycles [36]. The degradation of superelastic or shape memory effects
of SMA under cyclic loadings is the main objective of this dissertation, as they are the key to the selfcentering capacity of bridges.
Existing research on the low-cycle fatigue behavior of SMAs mainly focused on binary Ni-Ti
SMAs [107–112]. The research on low-cycle fatigue performance of Cu-Al-Mn SMA is very limited.
Only two studies have been reported so far. Kato et al. [97] investigated the fatigue behavior of single
crystal Cu-Al-Mn SMA through cyclic tension, compression, and tension-compression tests. Small
specimens with 1.8 mm width and 5 mm gauge length were tested. It was found that the stress-strain
curves of single crystal Cu-Al-Mn SMA show almost no degradation in the first 100 cycles under a
10% strain amplitude. Shrestha et al. [36]studied the fatigue behavior of Cu-Al-Mn SMA bars through
cyclic tensile tests. Cu-Al-Mn specimens with 11 mm diameter and 100 mm length were tested. It was
seen that the residual strain starts to accumulate in the range from 200 to 400 cycles and no fracture
was observed by the end of 1,000 loading cycles for all tested Cu-Al-Mn SMA specimens.

79
These two studies confirmed the fatigue resistance of Cu-Al-Mn SMAs. However, in these
studies, only small samples were tested up to a limited number of 1,000 loading cycles at room
temperature. The fatigue performance of larger diameter Cu-Al-Mn SMAs under low and high
temperatures that are applicable to civil engineering have never been reported. To fill this knowledge
gap, this study investigated the low-cycle fatigue behavior of Cu-Al-Mn SMAs at different
temperatures including room temperature, 25 ℃, low temperature, -40 ℃, and high temperature, 50
℃. Conventional Ni-Ti SMA was tested under the same conditions for comparison purposes.
4.2 Materials and methods
4.2.1 Materials
The materials tested in this study were single crystal and polycrystal Cu-Al-Mn SMA and
polycrystal Ni-Ti SMA. For brevity, Cu-Al-Mn and Ni-Ti SMA is respectively referred to as CAM
and NiTi SMA hereafter. The composition and supplier of CAM and NiTi SMA used in this study
were the same as previously mentioned in Chapter 3. The transformation temperatures of CAM
and NiTi SMA measured by differential scanning calorimetry (DSC) are shown in Fig. 4-1. From
Fig. 4-1, it is seen that the austenitic finish temperature, Af, of CAM and NiTi SMA is -67.3 °C
and -15.5 °C, respectively. Both materials are in austenitic phase at room temperature.

80
Two forms of CAM specimens and one form of NiTi specimen were studied, namely single
crystal CAM SMA bars (SCB), polycrystal CAM SMA plates (PCP), and polycrystal NiTi bars
(NTB). The machining process for SCB and PCP were the same as previously mentioned in
Chapter 3. The NTB was machined from as received rods to 140 mm long and 16 mm diameter
dog bone samples with a diameter of 12.7 mm in the gauge length. The dimensions of SCB, NTB
(a)
(b)
Fig. 4-1 Transformation temperatures of SMA used in this study: (a) CAM, and (b) NiTi
SMA.

81
and PCP used in this study are shown in Fig. 4-2. The optical microscope image of polycrystal
NiTi SMA used in this study is shown in Fig. 4-3. It is seen that the average grain size of NiTi
SMA is around 100 m. The grain distribution of PCP will be discussed in detail in Section 4.4.
Fig. 4-2 Dimensions of: (a) SCB, NTB, and (b) PCP specimens.
Fig. 4-3 Optical microscope image of polycrystal NiTi SMA.
30 12 12 30
140
R45 R45
56
17 20
6
250
16
65 60 60 65
Toothed
zone
Toothed
zone
Note:
All dimensions
are in mm
12.7
12.7
(a)
(b)
100 μm
88 μm
76 μm
114 μm

82
4.2.2 Methods
The setup of low-cycle fatigue tests is shown in Fig. 4-4. An MTS 370.5 dynamic servohydraulic frame was used to apply the cyclic loading. For non-room temperature tests, an MTS
651.06E-04 environmental chamber was equipped to the MTS frame. An Epsilon Model No. 3542-
0200-050-ST with 50.8 mm gauge length extensometer was used to measure the strain in the
specimens. A BMS16HR-53 Data Acquisition System (DAQ) was used to collect the data.
Fig. 4-4 Setup for low-cycle fatigue test: (a) setup for SCB and NTB at room temperature, (b)
zoom-in view of customized grip fixture for SCB and NTB, (c) zoom-in view of grip fixture
for PCP, (d) inside view of environmental chamber and setup for PCP, and (d) outside view of
environmental chamber connected with liquid nitrogen tank.
To reduce the stress concentrations at the grip, customized fixtures were designed and
fabricated. For SCB and NTB, the round bars were mounted in a conical grip with a conical hole,
which had a smooth contact with the specimens, as shown in Fig. 4-4 (a) and (b). The details of
(b) (d)
(c)
(a)
PCP
Extensometer
(e)
Environmental
chamber
MTS load
frame
Liquid
nitrogen
tank
Environmental
chamber
Extensometer
Customized grip fixture
for SCB / NTB
SCB/NTB

83
the grip fixture for SCB and NTB are shown in Fig. 4-5. This mechanical contact-based grip
effectively reduced the stress concentrations and eliminated premature failures. The PCP was
gripped by two steel plates at the tooth ends, where two steel plates were tightened by four high
strength steel bolts, as shown in Fig. 4-4 (c) and (d). The details of the grip fixture for PCP are
shown in Fig. 4-6. Although not shown in Fig. 4-6 for clarity, heavy duty c-clamps were also used
to increase the clamping force applied on the specimens. There was no slip observed during the
tests.
Fig. 4-5. Schematic diagram of grip fixture used for SCB, NTB: (a) picture of setup, (b) front
view, (c) section A-A, and (d) section B-B.
B-B
(c) Section A-A
(d) Section B-B
Clamp
Part-1
Clamp
Part-2
Bolt
Threaded fixture
Part-1
Conical
surface
Clamp
Part-1
Bolt
Threaded fixture
Part-2
Threaded fixture
Part-1
A-A
(b) Front view
Threaded
fixture
Clamp
Hydraulic
wedge grip
Specimen hole
Extension
rod
(a) Picture of setup
SCB/NTB

84
Fig. 4-6. Schematic diagram of grip fixture used for PCP: (a) picture of setup, (b) front view,
and (c) section A-A.
Prior to low-cycle fatigue tests, all specimens were trained at room temperature to stabilize
the martensitic transformation. The training consisted of five tensile cycles with increasing
amplitude of 1% to 5% at 1% increments. During the low-cycle fatigue tests, specimens were
stretched to 5% target strain at a speed of 0.4% strain/sec, then unloaded to near-zero force within
around 10 seconds. The target force in unloading was set close to zero but not exactly zero to avoid
any slack or compressive loading in the specimens. This process was cyclically repeated until
fracture occurred. The loading was strain controlled while the unloading was under force
controlled. For low temperature tests, the test condition was produced using a liquid nitrogen tank,
as shown in Fig. 4-4 (e). With the existing setup, it was not feasible to replace the liquid nitrogen
A-A
(c) Section A-A
Clamp
Part-1
Clamp
Part-2
PCP
Steel
plates
(b) Front view
(a) Picture of setup
Steel
plates
PCP
Bolt

85
tank without disrupting the test environment. Therefore, the cold temperature tests had to be
completed before the tank ran out (within approximately 24 hours).
The main purpose of this study is to investigate the low-cycle fatigue behavior of CAM SMA
compared with NiTi SMA, focusing on the degradation of their superelastic properties with respect
to increasing loading cycles. Therefore, the tests were not designed to load all specimens until
failure. It was found that after a certain numbers of cyclic loading (around 2,000 cycles for CAM
and 1,000 cycles for NiTi), the stress-strain curves of SMA bars degraded into a line (which means
that the energy dissipation disappeared completely), and the residual strain reached a plateau and
remained unchanged with the increasing loading cycles (this is referred to as stabilized condition
hereafter). Therefore, fatigue tests on some samples were terminated after a stabilized condition
had been observed unless the specimen ruptured prior to that.
The fatigue loading was terminated when either of the following conditions was achieved:
(1) the specimen fractures, (2) the residual strain becomes stable, or (3) the liquid nitrogen tank
runs out. A summary of the final test condition of each specimen and the test matrix of low-cycle
fatigue tests is provided in Table 4-1. The specimen labels indicate ‘the material type’ – ‘the test
temperature (‘m’ denotes minus)’ – ‘specimen number’. For example, ‘SCB-25C-1’ indicates the
first specimen of SCB tested at 25 °C and ‘NTB-m10C-2’ indicates second specimen of NTB
tested at -10 °C.

86
Table 4-1 Summary of low-cycle fatigue tests on CAM and NiTi SMAs.
Material Temperature Sample label
Total loading
cycles
Reason for test
termination
SCB
Room temperature
(25 ℃)
SCB-25C-1 50,000 Condition 2
SCB-25C-2 10,000 Condition 2
SCB-25C-3 4,285 Condition 2
SCB-25C-4 5,326 Condition 2
SCB-25C-5 7,000 Condition 2
SCB-25C-6 10,000 Condition 2
Low temperature
(-40 ℃)
SCB-m40C-1 5,500 Condition 2
SCB-m40C-2 2,600 Condition 2
SCB-m40C-3 7,200 Condition 3
High temperature
(50 ℃)
SCB-50C-2 9,100 Condition 2
SCB-50C-3 2,100 Condition 2
SCB-50C-1 2,100 Condition 2
PCP
Room temperature
(25 ℃)
PCP-25C-1 15,001 Condition 1
PCP-25C-2 1,281 Condition 1
PCP-25C-3 358 Condition 1
PCP-25C-4 111 Condition 1
PCP-25C-5 105 Condition 1
NTB
Room temperature
(25 ℃)
NTB-25C-1 90 Condition 1
NTB-25C-2 31 Condition 1
NTB-25C-3 54 Condition 1
Low temperature
(-40 ℃)
NTB-m40C 7 Condition 1
Low temperature
(-10 ℃)
NTB-m10C-1 1,982 Condition 3
NTB-m10C-2 451 Condition 1
NTB-m10C-3 1,133 Condition 1
High temperature
(50 ℃)
NTB-50C-1 88 Condition 1
NTB-50C-2 16 Condition 1
NTB-50C-3 29 Condition 1
After obtaining the stress-strain curves from low-cycle fatigue tests, eight mechanical
properties were extracted and analyzed to quantify the effect of fatigue loading and temperature
variation on the superelasticity of CAM and NiTi SMAs. They are: elastic modulus, Eload, elastic
modulus after yielding, Eload2, maximum stress, max, yield stress, y, damping ratio, R, maximum

87
strain, max, residual strain, resi, and recovery strain, reco. The definition of these parameters is
shown in Fig. 4-7.
The results of low-cycle fatigue tests on CAM and NiTi SMA are presented in the following
subsections. Since a large number of specimens were tested, for the sake of brevity and ease of
reading, only two typical specimens for each test scenario and material were selected and discussed
in this chapter. The remaining test results are presented in the Appendix B.
4.3 Results of SCB
4.3.1 SCB at room temperature, 25 ℃
Fig. 4-8 shows the stress-strain curves of SCB-25C-1 and SCB-25C-2 at room temperature,
25 °C. It is seen that both samples showed excellent superelasticity until about 200 cycles. In the
first 100 cycles, the stress strain curves remained ideal flag-shaped and almost no residual strain
accumulation was observed. From 100 to 1,000 cycles, the stress-strain curves narrowed down,
Fig. 4-7 Definition of mechanical properties considered in low-cycle fatigue test.

88
and the residual strain increased. Moreover, the damping ratio showed an obvious increase from
100 to 1,000 cycles. The reason for this increase is the reduction of the transformation stress and
increase in the area of the stress-strain curves.
(a)
(b)
Fig. 4-8 Stress-strain curves of SCB at room temperature, 25 °C: (a) SCB-25C-1, and (b)
SCB-25C-2.

89
Fig. 4-9 shows the variation of extracted parameters with increasing loading cycles. In Fig.
4-8, it is seen that the yield stress first decreased and then almost disappeared at around 1,000
cycles. However, because the yield stress was obtained as the intersection point of Eload and Eload2,
this disappearance did not show in the extracted results in Fig. 4-9. Furthermore, it was found that
when the ratio of Eload2 to Eload was over 0.85, the Eload2 and y were largely influenced from
numerical error; therefore, the results for these parameters were truncated when the ratio of Eload2
to Eload reached 0.85.
From Fig. 4-9, it is seen that the residual strain and damping ratio of SCB started to statute at
around 1,000 cycles, and after that, they remained stable until 50,000 cycles. It is an important
finding that even after 50,000 cycles of 5% strain loading, the CAM SMA bar can still sustain
loading with no visible crack or fracture. This demonstrates excellent low-cycle fatigue resistance
of CAM SMA.

90
(a)
(b)
Fig. 4-9 Variation in mechanical properties of SCB at room temperature, 25 °C: (a) SCB-25C1, and (b) SCB-25C-2.

91
4.3.2 SCB at -40 ℃
Fig. 4-10 shows the stress-strain curves of SCB-m40C-1 and SCB-m40C-2 at -40 ℃. It is
seen that at -40 ℃, the SCB showed an ideal flag-shaped stress strain curve and excellent damping
capacity until 200 cycles. Furthermore, the specimen SCB-m40C-1 showed no rupture after 5,500
cycles of fatigue loading. The test was terminated because the liquid nitrogen tank ran out and the
environmental chamber cannot keep sustaining stable low temperature condition. The envelope
curve from training at room temperature was presented with a grey dashed line at Cycle 1. As seen
in Fig. 4-10, the transformation stress at -40 ℃ was lower than that at room temperature. However,
area inside the stress strain cycles at -40 ℃ was larger than that at room temperature (refer to Fig.
4-8), which indicates higher energy dissipation capacity of CAM SMA at low temperature.
(a)

92
(b)
Fig. 4-10 Stress-strain curves of SCB at -40 °C: (a) SCB-m40C-1, and (b) SCB-m40C-2.
Fig. 4-11 shows the variation of extracted parameters with respect to loading cycles at -40
℃. The variation in the mechanical properties at -40 ℃ is similar to that at room temperature.
What is worth noting is that, from the damping ratio and residual strain shown in Fig. 4-9 and Fig.
4-11, the energy dissipation and strain recovery capacity at -40 ℃ were superior to those at room
temperature, demonstrating the promise of applying CAM SMA in low temperatures.

93
(a)
(b)
Fig. 4-11 Variation in mechanical properties of SCB at -40 °C: (a) SCB-m40C-1, and (b)
SCB-m40C-2.

94
4.3.3 SCB at 50 ℃
The stress-strain curves of SCB-50C-1 and SCB-50C-2 at 50 ℃ are shown in Fig. 4-12. The
envelope curve from training at room temperature was presented with a grey dashed line. SCB50C-1 showed no fracture after 9,100 cycles of loading. Compared with the stress-strain curves at
room temperature, 25 ℃ (shown in Fig. 4-8), and -40 ℃ (shown in Fig. 4-10), the yield stress of
SCB at 50 ℃ was higher and the hysteresis loop become narrower.
The variation in mechanical properties of SCB at 50 ℃ is shown in Fig. 4-13. The general
variation in the mechanical properties at 50 ℃ was similar to those at room temperature, 25 ℃
and -40 ℃. An important observation is that the damping ratio at 50 ℃ showed an increase after
100 cycles. The reason is that in the first several cycles, the stress-strain curves were very narrow.
While as the residual strain increases and the yield stress decreased, the hysteresis loop became
wider and led to an increase in the damping ratio. After around 1,000 cycles, the response of SCB
became almost linear.

95
(a)
(b)
Fig. 4-12 Stress-strain curves of SCB at 50 °C: (a) SCB-50C-1, and (b) SCB-50C-2.

96
(a)
(b)
Fig. 4-13 Variation in mechanical properties of SCB at 50 °C: (a) SCB-50C-1, and (b) SCB50C-2.

97
4.4 Results of PCP
Fig. 4-14 shows the stress-strain curves of PCP-25C-1 and PCP-25C-2 at 25 ℃. The first
specimen ruptured after 15,000 cycles of fatigue loading. Fig. 4-15 shows the extracted mechanical
properties of two PCP with respect to loading cycles. From Fig. 4-14 and Fig. 4-15, it is seen that
the fatigue resistance of PCP-25C-1 was similar to that of SCB. In the initial 100 cycles, the
superelasticity showed almost no degradation, flag-shaped stress-strain curves and energy
dissipation were observed. From 100 to 1,000 cycles, the stress-strain curves became narrower,
and the residual strain accumulated rapidly. After 1,000 cycles, the response became almost linear
until fracture.
(a)

98
(b)
Fig. 4-14 Stress-strain curves of PCP at room temperature, 25 °C: (a) PCP-25C-1, and (b)
PCP-25C-2.
(a)

99
(b)
Fig. 4-15 Variation in mechanical properties of PCP at room temperature, 25 °C: (a) PCP25C-1, and (b) PCP-25C-2.
Even though a long fatigue life (i.e., 15,000 cycles) exhibited by PCP-25C-1, the fatigue life
of the rest of PCP showed a large scatter. The rest of four PCP, respectively, fractured at 1,280,
357, 110 and 104 cycles (see Appendix B). Fig. 4-16 shows the grain distribution and fracture
location of PCP specimens. The black solid lines indicate the grain boundaries on the specimen’s
surface, obtained by visual observation. The blue dash line and bracket indicate the gauge length
where the extensometer was placed during the test. Compared with SCB, the reason why the PCP
showed a large scatter in fatigue life was that each polycrystal plate was composed of several small
and randomly distributed grains with different sizes, as shown in Fig. 4-16. In fatigue loading, the
boundaries of neighboring grains, particularly when the neighboring grains have different

100
orientations, may cause stress concentrations and result in crack formation and propagation
[13,41,113]. The fracture location of each PCP is marked with a red arrow in Fig. 4-16, which
supports this explanation. Because of the large dependence of fatigue behavior on the grain size,
distribution and orientation, no further tests on PCP were conducted at low or high temperatures.
(a) (b) (c) (d)
Fig. 4-16 Grain distribution and fracture locations of PCP: (a) PCP-25C-1, (b) PCP-25C-3,
(c) PCP-25C-4, and (d) PCP-25C-5. The blue bracket indicates the gauge length of
extensometer. The red arrow indicates the fracture location.
4.5 Results of NTB
4.5.1 NTB at room temperature, 25 °C
Fig. 4-17 shows the stress-strain curves of NTB-25C-1 and NTB-25C-2 at 25 ℃. The three
tested NTB at 25 ℃ all fractured rapidly with fatigue lives of 89, 31 and 54 cycles, respectively.

101
It is seen that the stress-strain curve of NTB at 1
st cycle was not ideal flag-shaped: the residual
strain of NTB at the first cycle reached 1%. The reason is that for SMA, the reversible stress
induced martensitic transformation is completed at the maximum recovery strain, i.e., superelastic
limit, after that, further loading triggers plastic deformation and increased residual strain [58,114].
The superelastic limit of NTB used in this study was only around 3%. The 5% cyclic loading
exceeded this superelastic limit and led to plastic deformation and remarkable residual strain.
Similar stress-strain curve for NiTi were reported by Kang et al. [115]. As shown in Fig. 4-18, the
initial stress-strain curve of NTB-25C-1 is compared with the stress-strain curves of NiTi SMA
tested by Kang et al. [115] at peak stress peak of 450, 500, 550 and 600 MPa, respectively.
(a)
(b)
Fig. 4-17 Stress-strain curves of NTB at room temperature, 25 °C: (a) NTB-25C-1, and (b)
NTB-25C-2.

102
Fig. 4-18 Stress-strain curves comparison of NTB-25C-1 and results reported by Kang et al.
[115]
Fig. 4-19 shows the extracted mechanical properties of NTB with respect to loading cycles.
As the fatigue loading cycle increased, the stress-strain curve of NTB narrowed down rapidly,
leading to a sharp decrease in energy dissipation. The residual strain also increased in the first few
cycles and then stabilized until fracture.
(a)

103
(b)
Fig. 4-19 Variation in mechanical properties of NTB at room temperature, 25 °C: (a) NTB25C-1, and (b) NTB-25C-2.
4.5.2 NTB at -40 ℃ and -10 ℃
The low temperature fatigue tests on NTB were conducted at -40 ℃ and -10 ℃. The reason
was that no superelasticity was observed at -40 ℃, therefore, -10 ℃ was selected to investigate
the fatigue resistance of NTB. Stress-strain curves of NTB-m40C at -40 ℃ are shown in Fig. 4-20,
where the envelope curve of training at room temperature, 25 ℃ was shown in grey. From Fig.
4-20, it is seen that the NTB was in pure martensitic phase at -40 ℃. No superelasticity was seen
because no austenite-martensite phase transformation was triggered. The specimen NTB-m40C
fractured after 6 cycles of fatigue loading.

104
Fig. 4-20 Stress-strain curves at typical cycles of NTB-m40C.
Fig. 4-21 shows the stress-strain curves of NTB-m10C-1 and NTB-m10C-2 at -10 ℃. Fig.
4-22 shows the extracted mechanical properties with respect to loading cycles. The energy
dissipation and strain recovery capacity of NTB at -10 ℃ started to degrade rapidly at around 10
cycles and stabilized at around 100 cycles. Compared with the results at 25 ℃, NTB specimens at
-10 ℃ showed higher fatigue life. This trend is consistent with the results reported by Iasnii et al.
[116] and Mahtabi et al. [117].
(a)

105
(b)
Fig. 4-21 Stress-strain curves at typical cycles of NTB at -10 °C: (a) NTB-m10C-1, and
(b) NTB-m10C-2.
(a)

106
(b)
Fig. 4-22 Variation in mechanical properties of NTB at -10 °C: (a) NTB-m10C-1, and (b)
NTB-m10C-2.
4.5.3 NTB at 50 ℃
The stress-strain curves of NTB-50C-1 and NTB-50C-2 at 50 ℃ are shown in Fig. 4-23. The
yield stress of NTB at 50 ℃ reached 500 MPa, which was higher than training results. All NTB at
50 ℃ fractured only after a few cycles of loading. The first one ruptured after 87 cycles of fatigue
loading while the second and the third ones fractured, only after 15 and 29 cycles, respectively.
Furthermore, the stress-strain curves narrowed down very rapidly and then remained unchanged
until fracture.

107
(a)
(b)
Fig. 4-23. Stress-strain curves of NTB at 50 °C: (a) NTB-50C-1, and (b) NTB-50C-2.
Fig. 4-24 shows the extracted mechanical properties of two selected NTB with respect to
loading cycles. It is observed that the Young’s modulus and yield stress of NTB at 50 ℃ showed
negligible degradation. The damping ratio and residual strain degraded rapidly and then stabilized.
The NTB-50C-2 stabilized in the second cycle. Even in the first cycle, significant residual strain
was observed, indicating the poor superelasticity of NTB at high temperature, 50 ℃.

108
(a)
(b)
Fig. 4-24 Variation in mechanical properties of NTB at 50 °C: (a) NTB-50C-1, and (b) NTB50C-2.

109
4.6 Summary of findings
It was found that at 25 ℃ and 50 ℃, the stress-strain curves of Ni-Ti SMA narrowed down
rapidly in the first few cycles, leading to a significant decrease in its energy dissipation capacity.
Additionally, at -40 ℃, the Ni-Ti SMA lost superelasticity completely. The single crystal Cu-AlMn SMA showed excellent superelasticity and fatigue resistance at all tested temperatures: -40 ℃,
25 ℃, and 50 ℃. Specifically, in the first 100 cycles, almost no degradation was observed in the
superelastic properties. From 100 to 1,000 cycles, the yield stress started to decrease, the residual
strain started to accumulate, and the damping ratio started to reduce. At -40 ℃, 25 ℃, and 50 ℃,
the hysteresis loop of single crystal Cu-Al-Mn SMA became linear after around 8,000, 2,000, and
2,000 cycles, respectively. The superelasticity and fatigue resistance of polycrystal Cu-Al-Mn
SMA were found to largely depend on the grain distributions. The five polycrystal Cu-Al-Mn
SMA specimens all showed comparable superelasticity to single crystal Cu-Al-Mn SMA at the
early stage of fatigue loading (i.e., 1 to 100 cycles), but the fatigue life of polycrystal Cu-Al-Mn
SMA showed a large scatter (difference of more than two orders of magnitude).
In summary, compared with Ni-Ti SMA, Cu-Al-Mn SMA had a longer fatigue life and slower
superelastic properties degradation with increasing loading cycles at a wide range of temperature
from -40 ℃ to 50 ℃. However, the yield stress and energy dissipation of Cu-Al-Mn SMA were
lower than that of Ni-Ti SMA. When applying Cu-Al-Mn SMA in bridge columns, attention
should be paid to its low yield stress and low energy dissipation.

110
Chapter 5 - Machinability characteristics of Cu-Al-Mn SMA
The machinability characteristics of Cu-Al-Mn SMA, such as chip formation, cutting
temperature, tool wear, workpiece surface roughness and diameter deviation, were studied and
compared with conventional Ni-Ti SMA, as well as commonly used steels: mild steel (MS), and
304 stainless steel (SS). Both single machining tests on one workpiece and continuous machining
tests on multiple workpieces were conducted. A wide range of cutting parameters were
investigated including the cutting speed ranging from 15 to 120 m/min, feed rate ranging from 0.1
to 0.2 mm/rev, and depth of cut ranging from 0.5 to 1.5 mm.
5.1 Research motivation
Cu-Al-Mn SMA is cost effective, has excellent low-cycle fatigue resistance, and a wide
temperature application range compared to other types of SMAs. These advantages of Cu-Al-Mn
SMAs have resulted in an increased research interest in their use in civil engineering applications,
particularly as reinforcement in concrete structures. Such applications could require machining of
the Cu-Al-Mn SMA bars for connecting with other structural elements. Therefore, it is important
to understand their machinability using conventional metal processing methods.

111
However, there is no comprehensive study on the machineability of Cu-Al-Mn SMAs in
comparison to other metal alloys so far. Due to the special thermal-mechanical behavior induced
by the martensitic phase transformation, the machinability of SMAs is different from that of steels
or other traditional difficult-to-machine materials such as stainless steel [118], Inconel 718 [119],
and Ti-6Al-4V [120]. SMAs normally exhibit a flag-shaped stress-strain curve until the martensitic
transformation is complete, after which they enter into the plastic range and exhibit high strain
hardening and ductility. The special stress-strain behavior of SMAs could cause a series of
problems in machining such as excessive formation of burrs, poor chip breakage, rapid tool wear,
and poor product surface [121–124]. The rapid tool wear in turn requires frequent tool change,
resulting in increased production time and cost. Furthermore, both the yield stress and elastic
modulus of SMAs increase with increasing ambient temperature according to the Clausius–
Clapeyron relationship [41]. This means that the material becomes stronger as the cutting
temperature increases [125], which can lead to increased cutting force and stability issues, and
make the processing of SMAs even more challenging [126].
Existing research on the machineability of SMAs mainly focused on Ni-Ti-based alloys
because of their longer history [126–129]. Ni-Ti-based SMAs are considered a difficult-tomachine material because of several reasons such as high strain hardening and ductility, varying
elastic modulus, low thermal conductivity, and high specific heat [130]. According to Weinert et
al. [127], machining Ni-Ti SMAs usually produces only one continuous chip during the whole
process, which occurs even when the cutting length exceeds 100 m. The poor chip breakage and

112
strain recovery property causes massive burr formation, rapid tool wear and poor workpiece
quality. Guo et al. [128] studied the dynamic mechanical property and machinability of Ni-Ti
SMAs. It was found that the very high dynamic strength of Ni-Ti SMAs (up to 3 GPa) applies a
large mechanical loading on the tool-workpiece interface, which leads to high cutting temperature,
strong adhesion wear and tool flaking. The high cutting temperature is detrimental because it in turn
affects the functional behavior of SMAs [131,132]. To reduce the cutting temperature, special
strategies such as liquid nitrogen cooling [121] and flood cooling [129] are required when machining
Ni-Ti-based SMAs.
In contrast to the extensive studies on the machinability of Ni-Ti-based SMAs, the
machineability of Cu-Al-Mn SMAs has never been systematically studied. To fill this knowledge
gap, the machinability characteristics of Cu-Al-Mn SMA was investigated in this study and
compared with Ni-Ti SMA and conventional steels including mild steel and 304 stainless steel.
5.2 Experimental program
The materials tested in this study were polycrystal Cu-Al-Mn SMA, polycrystal Ni-Ti SMA,
mild steel, and 304 stainless steel (SS). For brevity, Cu-Al-Mn SMA, Ni-Ti SMA, mild steel, and
304 stainless steel was respectively referred to as CAM, NiTi, MS and SS hereafter in this chapter.
The composition and supplier of CAM and NiTi were the same as previously mentioned in Chapter
3. Both MS and SS were received as 20 mm diameter round bars, from McMaster-Carr Supply
Company.

113
5.2.1 Vickers hardness
Hardness is usually used as a preliminary indicator of the material machinability [130]. The
Vickers hardness of CAM was measured and compared with NiTi, ASTM A276 [133]
multipurpose 304 SS, ASTM A108 [134] MS, ASTM B16 [135] ultra-machinable 360 brass,
ASTM B209 [136] multipurpose 6061 aluminum, and ASTM B152 [137] multipurpose 110
copper, as shown in Fig. 5-1. A LECO Model No. LM-100 microindentation hardness tester was
used, with a load of 1 kg and dwell time of 10 s. Each material was tested five times to obtain the
average hardness. From Fig. 5-1, it is seen that the hardness of CAM is lower than that of NiTi but
close to that of 304 SS and MS. Therefore, to present a more intuitive comparison, based on the
Vickers hardness values shown in Fig. 5-1, NiTi, MS and 304 SS were chosen to benchmark the
machinability of CAM.
Fig. 5-1 Vickers hardness of Cu-Al-Mn SMA (CAM) in comparison with other commonly
used materials: Ni-Ti SMA (NiTi), 304 stainless steel (SS), mild steel (MS), brass,
aluminum, and copper.
438
321
280
237
166
137 122
0
100
200
300
400
500
NiTi 304 SS CAM MS Brass Aluminum Copper
Vickers hardness

114
5.2.2 Machinability tests
Standard single point turning experiments were performed to evaluate machinability.
Specifically, the chip formation, cutting temperature, tool wear, and workpiece surface roughness
were analyzed and compared under different cutting parameters, including eight cutting speeds Vc
= 15, 30, 45, 60, 75, 90, 105 and 120 m/min, three feed rates fr = 0.1, 0.15, 0.2 mm/rev, and three
depths of cut dc = 0.5, 1.0, 1.5 mm. The experimental setup and definition of cutting parameters
Vc, fr, and dc are shown in Fig. 5-2. A TRAK 1630SX lathe CNC machine with 2.24 kW power
and a spindle speed range of 150-2500 rpm was used. The KOOL MIST FORMULA #77 mist
spray coolant was applied during the machining at a spray rate of 480 ml/h. A Hertel
CCMT32.50.5BH HT115CR uncoated cermet turning insert was used as the cutting tool. The
cermet tools are commonly used for steel machining because of their high transverse rupture
strength and toughness, high-temperature resistance, and finish quality [138,139]. The dimensions
of the inserts were: nose radius r= 0.2 mm, tip angle = 80°, relief angle = 7°, and rake angle
=5°. During machining, a REED R2020 infrared thermometer with a temperature range from -50
℃ to 3992 ℃ was used to monitor the cutting temperature. After each test, a HAYEAR 48MP 2K
optical microscope with a maximum magnification of 150x was used to examine the tool wear and
chip morphology. The surface roughness of workpiece was measured according to ISO 4287 [140]
using a Mitutoyo 178-561-02A Surftest SJ-210 Surface Roughness Tester (with an accuracy of 1
m).

115
The machinability test matrix is shown in Fig. 5-3. Two types of machining tests were
conducted. First, single machining tests were performed on only one workpiece at a time to
characterize the effect of cutting parameters, Vc, fr, and dc, on CAM, NiTi, MS and SS. At the end
of each single machining test, the insert and workpiece were removed. The next test started with a
fresh sample and a new insert. The chip formation, tool wear and product surface roughness were
examined and analyzed. The combination of parameters tested in the single machining tests is
shown in Fig. 5-3 (a). A constant cutting volume of 7,068 mm3 was used to control all the single
machining tests, and only one parameter was changed at a time. The constant cutting volume
control method is also commonly used by others [127,141]. Eight Vc: 15, 30, 45, 60, 75, 90, 105,
120 m/min were considered. For CAM and NiTi, the fr values of 0.10, 0.15 to 0.20 mm/rev, and
dc values of 0.5, 1.0 to 1.5 mm were used. The values of Vc and dc were chosen according to
previous studies on NiTi [121,127,129] and the values of fr were chosen according to the geometry
Fig. 5-2 (a) Machinability test setup, and (b) definition of key parameters in single point
turning tests.
Workpiece
Machined
surface
Direction of
feed motion
fr
Tool holder
Chip
Work
surface
Direction of
cutting speed
Vc
Major cutting
edge
Minor cutting
edge
Depth of cut
dc
Insert
Workpiece
Spindle
Chip
Tool holder
Coolant supply
(a) (b)

116
of the insert (the optimum feed rate fr is usually set to 1/2 of the nose radius rc [142]). The
maximum Vc for NiTi was limited to 60 m/min because when Vc reached 60 m/min, the tool wear
was already far beyond the tool life criterion limit recommended by ISO-3685 [143]. For MS and
SS, only the most aggressive fr =0.2 mm/rev and dc =1.5 mm were used.
Next, to simulate the real material processing which involves large stock removal, continuous
machining tests on thirty workpieces (each workpiece with a cutting volume of 7,068 mm3
) were
conducted on CAM and MS. The so called ‘continuously’ means that thirty specimens were
machined one after another without changing the insert or the cutting parameters. In an actual
machining project, in order to ensure efficiency and product quality, a roughing operation is
usually performed first to establish the rough shape with a coarse finish, subsequently, a finishing
operation with a smaller dc is used to improve the finish and obtain the final dimensions [144]. The
following parameters that are commonly used to machine MS were adopted in the continuous
machining tests: Vc =60 m/min, fr =0.1 mm/rev, and rough path dc =1.0 and finish path dc =0.5
mm, as shown in Fig. 5-3 (b). The total cutting length and cutting volume was 4.5 m and 212,040
mm3
, respectively. Since only one insert was used in one test, as the tip of the insert wore, the
diameter of the machined sample increased. At the end of the continuous machining tests, the insert
was removed and measured under the optical microscope; and the diameter of each machined
sample was measured by a Mitutoyo IP65 micrometer (with an accuracy of 1 m).

117
5.3 Results and discussion of single machining tests
5.3.1 Chip formation
When machining metals, shearing occurs in the contact region between the chip and the insert,
creating stress concentrations and an increase in the temperature. The high temperature, pressure
Fig. 5-3 Test matrix of machinability tests: (a) single, and (b) continuous machining tests.
15
30
45
60
75
0.10
0.15
0.20
0.5
1.0
1.5 90
105
120
CAM
(40 conditions)
15
30
45
60
0.10
0.15
0.20
0.5
1.0
1.5
NiTi
(20 conditions)
15
30
45
60
75
0.10
0.20
0.5
1.5 90
105
120
MS or SS
(16 conditions for
each material)
60 CAM
(30 workpieces)
60
MS
(30 workpieces)
0.10
0.10
Rough path: 1.0
Finish path: 0.5
Rough path: 1.0
Finish path: 0.5
Single
machining
tests
Continuous
machining
tests
Depth of cut
dc
(mm)
Feed rate
fr
(mm/rev)
Cutting speed
Vc
(m/min)
(a)
(b)

118
and the inhomogeneous plastic deformation result in severe work hardening of the chip, which in
turn wears the insert and affects the product surface integrity. Depending on the workpiece material
and cutting conditions, the chip will show distinct shapes and colors. In general, the chip formation
is mainly affected by the workpiece material, insert geometry, cutting temperature, and cutting
parameters (Vc, fr, and dc) [121,142].
To characterize the machinability of CAM, the chip formation under different Vc, fr, and dc
were investigated in terms of chip shape, length, and color, and the results were compared with
MS, SS and NiTi. The chips generated from CAM machining in comparison with those from MS
and SS are shown in Fig. 5-4, where the description on top of each picture is the chip classification
based on ISO-3685 [143]. For example, short chips generated when machining MS at Vc= 15
m/min, fr = 0.2 mm/rev, and dc= 0.5 mm are classified as “6.2 loose arc” as per the ISO-3685
[143]. Details of the chip characteristics can be found in ISO-3685 [143].

119
Fig. 5-4 Images and classification of chips generated when machining CAM, MS and SS at different Vc, fr, and dc.

120
From Fig. 5-4, it is seen that although the hardness of CAM is close to those of MS and SS
(see Fig. 5-1), the chip formation is very different. Machining MS and SS produced short and loose
arc-shaped chips (except for some cases when Vc < 60 m/min and dc = 1.5 mm) while machining
CAM produced long and continuous chips in snarled tubular form, which is believed to occur due
to the superelastic behavior and high ductility of the material. As reported in [121,127], for NiTi
superelastic alloys, the long and continuous chips may lead to a series of problems such as severe
chip flow damage on the major cutting edge, long chips wrapping around the workpiece and
interfering with coolant delivery, and damage to the product surface [121,127]. Furthermore, the
snarled chips rotating with the workpiece can be very dangerous for manual operation. Therefore,
special attention needs to be paid to improve the chip breakage when machining CAM. The effect
of cutting parameters (Vc, fr, and dc) on the shape and length of chips when machining CAM can
be summarized as follows. A change in Vc has almost no influence on the shape and length of
chips. An increase in dc decreases the chip breakage because higher dc generates a larger contact
area at the major cutting edge, resulting in wider snarled-tubular-formed chips that are stronger
and intertwined tighter. Increasing fr can remarkably improve the chip breakage. At aggressive fr=
0.2 mm/rev, discontinuous chips were created in short conical form with an average length of 20
mm. Therefore, using an aggressive fr can be considered as an effective way to improve the chip
breakage when machining CAM.
The chips collected from machining NiTi are shown in Fig. 5-5. It is seen that machining
NiTi generally created snarled ribbon chips in blueviolet color, and notably as the Vc increased,

121
the color became more purple. While for CAM, shown in Fig. 5-4, an increase of Vc up to 120
m/min had almost no influence on the color of the chips. The color of the chips mainly reflects the
cutting temperature during machining [145]. The ribbon shape and violet color indicate that the
chips were melted by the extremely high cutting temperature during machining NiTi. As shown in
Fig. 5-5, when machining NiTi at Vc= 60 m/min, fr = 0.1 mm/rev and dc= 1.5 mm, the chips were
welded together and formed a nest-shaped lump wrapped around the workpiece.
Fig. 5-5 Images and classification of chips generated when machining NiTi at different Vc, fr,
and dc.

122
A comparison of the cutting temperatures when machining CAM and NiTi at Vc = 60m/min
is shown in Fig. 5-6. It is seen that at dc= 1.5 mm and fr = 0.1 mm/rev, the cutting temperature
when machining NiTi reached 630 ℃, while the same for CAM was only 200 ℃. The cutting
temperature of CAM was much lower than that of NiTi, which may be due to the high thermal
conductivity of copper. Previous research confirmed that the high cutting temperature is a major
concern when machining NiTi [131]. High cutting temperature during machining can lead to
excessive tool wear and adhesion as well as altered functionality of the workpiece [130]. Shape
memory alloy materials are temperature sensitive and high temperatures may trigger the phase
transformation and alter the thermomechanical behavior of the material. The high thermal
conductivity and low cutting temperature when machining CAM make it easier to process
compared to NiTi.
Fig. 5-6 Comparison of cutting temperature when machining CAM and NiTi at Vc = 60 m/min.
0
100
200
300
400
500
600
700
0.5 1.0 1.5
Temperature (℃)
dc
(mm)
Vc=60m/min
fr=0.1mm/rev
CAM
NiTi
0
100
200
300
400
500
600
700
0.10 0.15 0.20
Temperature (℃)
f
r (mm/rev)
Vc=60m/min
dc=0.5mm
CAM
NiTi

123
5.3.2 Tool wear
During machining, the gradual friction between the tool-workpiece and tool-chip interfaces
leads to a progressive wear of the insert. Tool wear is mainly affected by the following factors:
tool geometry and material, coolant supply, workpiece material, cutting parameters, and chip
formation [142]. An excessive tool wear may lead to a series of problems such as frequent tool
change (which increases the machining time and cost), increased cutting force and energy
consumption, and inaccurate product dimensions. The typical tool wear patterns and characteristics
of flank wear as per ISO-3685 [143] are shown in Fig. 5-7. An average flank wear of 300 m is
specified as the tool life limit in ISO-3685 [143]. When using single point turning to process CAM,
the nose wear of the insert will influence the product diameter and surface quality, and the flank
wear will influence the chip formation, tool life and burs at the edge of the product. Therefore,
understanding the patterns and degree of nose wear and average flank wear under different cutting
parameters is essential to optimize the productivity and ensure a desired product quality when
machining CAM.

124
The nose and average flank wear at various Vc and dc values after machining CAM, MS, SS
and NiTi are shown in Fig. 5-8, where a conservative fr = 0.1 mm/rev is considered for all cases.
It is seen in Fig. 5-8 (a) and (b) that increasing Vc led to a uniform increase in the nose wear for all
four materials, and a higher dc produced a higher nose wear. Notably, the variation of nose wear
from machining CAM was overall slightly higher than that from machining SS and MS but much
smaller than that from machining NiTi. For example, at Vc = 60 m/min and dc = 1.5 mm, the nose
wear from machining CAM, SS, MS and NiTi was 102, 88, 83, and 1965 m, respectively. From
Fig. 5-8 (a) and (b), it is seen that the increase in the average flank wear from machining CAM
with increasing values of Vc and dc was less obvious when Vc reached 60 m/min. The average flank
wear variation from machining CAM was close to that of SS, higher than that of MS, but still much
smaller than that of NiTi. At Vc = 60 m/min, dc = 1.5 mm, the average flank wear from machining
CAM, SS, MS and NiTi was 106, 83, 52 and 734 m, respectively. When Vc varied from 15 to
120 m/min and dc varied from 0.5 to 1.5 mm, the nose wear from machining CAM was around 0.7
(a) (b)
Fig. 5-7 Typical tool wear patterns in single point turning: (a) Typical wear patterns on an
insert, and (b) Flank wear characteristics according to ISO-3685 [143].
Minor
cutting edge
Major
cutting edge
Tool holder
Insert
Nose
wear
Average flank
wear
C B N
Notch wear
Major cutting
edge
Zone

125
to 1.5 times that from machining SS, 0.8 to 1.7 times of that from machining MS, and 1/7 to 1/21
of that from machining NiTi; the average flank wear from machining CAM was around 1.2 to 1.7
times of that from machining SS, 1.4 to 2.4 times of that from machining MS, and 1/7 to 1/12 of
that from machining NiTi.
Typical images of tool wear after machining four materials at an aggressive dc = 1.5 mm are
shown in Fig. 5-9 for Vc = 60 m/min and fr = 0.1 mm/rev. An aggressive dc = 1.5 mm meant a
larger contact area at the major cutting edge [144] and a wider snarled-tubular-form chip formation
(see Fig. 5-4). It is seen from Fig. 5-9 (a) that CAM showed a typical abrasion wear mechanism in
the nose region, which was similar to MS and SS shown in Fig. 5-9 (b) and (c). However, slight
chipping caused by the continuous chip flow could be observed on the major cutting edge from
machining CAM, which was not seen in the cases of MS and SS. SS is known as a hard-to-machine
material due to its low thermal conductivity (resulting from the relatively high Ni content) and
high strain hardening and ductility [118]. Previous research has shown that machining SS usually
results in excessive adhesion of cutting debris, which may in turn cause built-up edge (BUE)
formation [146,147]. This was also observed in this study, as shown in Fig. 5-9 (c). Although CAM
has a similar hardness to SS, the BUE observed at the major cutting edge when machining SS did
not appear when machining CAM. The reason may be that the long, ductile, and snarled tubular
chips generated when machining CAM rubbed against the major cutting edge, thereby avoiding
the BUE accumulation. From Fig. 5-9 (d), it is seen that machining NiTi showed, respectively,
adhesion and chipping wear in the nose and flank region of the insert. The high cutting temperature

126
resulting from the low thermal conductivity and high specific heat of NiTi melted the chips and
debris and welded them on the insert, thereby leading to adhesion [129]. The chipping at the major
cutting edge was caused by the high contact stress at the insert-workpiece interface [128].
(a) (b)
(c) (d)
Fig. 5-8 Nose and average flank wear after machining CAM, MS, SS and NiTi at varying Vc and
dc: (a) Nose wear after machining CAM, MS and SS, (b) Nose wear after machining NiTi, (c)
Average flank wear after machining CAM, MS and SS, and (d) Average flank wear after
machining NiTi.
0
40
80
120
160
200
15 30 45 60 75 90 105 120
Nose wear (m)
Vc
(m/min)
fr=0.1mm/rev CAM dc=0.5mm
CAM dc=1.0mm
CAM dc=1.5mm
SS dc=1.5mm
MS dc=1.5mm
0
300
600
900
1200
1500
1800
2100
15 30 45 60
Nose wear (m)
Vc
(m/min)
NiTi dc=0.5mm fr=0.1mm/rev
NiTi dc=1.0mm
NiTi dc=1.5mm
0
40
80
120
160
200
15 30 45 60 75 90 105 120
Average flank wear (m)
Vc
(m/min)
fr=0.1mm/rev CAM dc=0.5mm
CAM dc=1.0mm
CAM dc=1.5mm
SS dc=1.5mm
MS dc=1.5mm
0
300
600
900
1200
1500
1800
2100
15 30 45 60
Average flank wear (m)
Vc
(m/min)
fr=0.1mm/rev NiTi dc=0.5mm
NiTi dc=1.0mm
NiTi dc=1.5mm

127
Fig. 5-9 Typical images of tool wear after machining: (a) CAM, (b) MS, (c) SS, and (d) NiTi
at dc =1.5 mm, Vc =60 m/min, and fr =0.1 mm/rev.
The nose and average flank wear at various Vc and fr values after machining CAM, MS, SS
and NiTi are shown in Fig. 5-10, where a conservative dc =0.5 mm is considered for all cases. It is
seen in Fig. 5-10 (a) and (b) that an increase of Vc and fr led to an increase of the nose wear when
machining CAM. At Vc < 60 m/min, the nose wear from machining CAM was close to that from
machining MS but smaller than that from machining SS. At Vc > 60 m/min, the nose wear from
machining CAM was larger than that from machining SS and MS, but much smaller than that from
machining NiTi. For example, at Vc =60 m/min and fr =0.2 mm/rev, the nose wear from machining

128
CAM, SS, MS and NiTi was 109, 85, 79 and 1502 m, respectively. From Fig. 5-10 (c), it is seen
that at a conservative fr =0.1 mm/rev, an increase of Vc led to a uniform increase in the average
flank wear for CAM, while at higher fr, the influence of Vc was less remarkable. The influence of
fr on the average flank wear from machining CAM became less obvious when Vc reached 60
m/min. At Vc =60 m/min and fr =0.2 mm/rev, the average flank wear from machining CAM, SS,
MS and NiTi was 114, 78, 65 and 1439 m, respectively. When Vc varied from 15 to 120 m/min
and fr varied from 0.1 to 0.2 mm/rev, the nose wear from machining CAM was around 0.6 to 1.8
times of that from machining SS, 0.9 to 1.8 times of that from machining MS, and 1/7 to 1/15 of
that from machining NiTi; the average flank wear from machining CAM was around 0.8 to 1.7
times of that from machining SS, 1.1 to 2.1 times of that from machining MS, and 1/8 to 1/13 of
that from machining NiTi.
Typical images of tool wear after machining four materials at an aggressive fr =0.2 mm/rev
are shown in Fig. 5-11 for Vc =60 m/min and dc =0.5 mm. From Fig. 5-11 (a) and (b), it is seen
that machining CAM at an aggressive fr =0.2 mm/rev showed abrasion wear of the nose and flank
regions, which was similar to MS, but more serious notch wear and chipping were observed for
CAM. The reason is that at aggressive fr, a larger contact pressure is generated at the toolworkpiece interface on the major cutting edge. Machining SS showed less BUE at aggressive fr,
as shown in Fig. 5-11 (c), because the high-speed feed motion impacted the accumulation of BUE.
From Fig. 5-11 (d), it is seen that at aggressive fr, machining NiTi showed adhered debris and
welded chips in the nose region, which was caused by the high cutting temperature. Grooves and

129
chipping caused by the high contact pressure could also be observed. In the flank region, severe
abrasion, chipping and notches could be observed after machining NiTi at aggressive fr.
(a) (b)
(c) (d)
Fig. 5-10 Nose and average flank wear after machining CAM, MS, SS and NiTi at varying
Vc and fr: (a) Nose wear after machining CAM, MS and SS, (b) Nose wear after machining
NiTi, (c) Average flank wear after machining CAM, MS and SS, and (d) Average flank
wear after machining NiTi.
0
40
80
120
160
200
15 30 45 60 75 90 105 120
Nose wear (m)
Vc
(m/min)
dc=0.5mm CAM fr=0.10mm/rev
CAM fr=0.15mm/rev
CAM fr=0.20mm/rev
SS fr=0.20mm/rev
MS fr=0.20mm/rev
0
300
600
900
1200
1500
1800
2100
15 30 45 60
Nose wear (m)
Vc
(m/min)
NiTi fr=0.10mm/rev dc=0.5mm
NiTi fr=0.15mm/rev
NiTi fr=0.20mm/rev
0
40
80
120
160
200
15 30 45 60 75 90 105 120
Average flank wear (m)
Vc
(m/min)
dc=0.5mm CAM fr=0.10mm/rev
CAM fr=0.15mm/rev
CAM fr=0.20mm/rev
SS fr=0.20mm/rev
MS fr=0.20mm/rev
0
300
600
900
1200
1500
1800
2100
15 30 45 60
Average flank wear (m)
Vc
(m/min)
dc=0.5mm NiTi fr=0.10mm/rev
NiTi fr=0.15mm/rev
NiTi fr=0.20mm/rev

130
Fig. 5-11 Typical images of tool wear after machining: (a) CAM, (b) MS, (c) SS, and
(d) NiTi at fr =0.2 mm/rev, Vc =60 m/min and dc =0.5 mm.
5.3.3 Surface roughness
The progressive tool wear and continuous chip flow during machining can lead to a
deteriorated product surface integrity. To characterize the surface state of CAM machined under
different cutting parameters, the surface roughness of each machined sample was measured along
the direction of feed motion as per ISO 4287-1997 [140] and compared with those measured on
MS and SS. NiTi was not considered for these measurements because, as presented in the previous
section, the tool wear when machining NiTi has exceeded the tool life criterion limit by more than

131
three times. Under such large tool wears, a meaningful comparison of surface roughness cannot
be made. Two commonly used roughness parameters were considered, namely arithmetical
average roughness Ra, and maximum peak to valley height Rt. Ra is the most popular parameter
used for product quality assessment and is adopted by many standards due to its ease of
measurement [140]. But since Ra has no physical meaning and gives no information on the
difference between peaks and valleys, and in-length features, it is commonly used along with Rt,
which indicates the deviation from peaks to valleys and is sensitive to scratches [148].
Surface roughness Ra and Rt of machined CAM samples under different Vc, dc and fr are
compared with MS and SS in Fig. 5-12. It is seen that the Ra and Rt showed an overall consistent
trend of variation. Both Ra and Rt of CAM were close to those of SS but less than those of MS.
The surface roughness of MS showed a rapid reduction when Vc reached 60 m/min, the same
observation was made by Ozcatalbas et al. [149]. Increasing Vc had little effect on the surface
roughness of CAM regardless of the change of dc and fr. Comparing Fig. 5-12 (a) and (c), or (b)
and (d), it is seen that that the surface roughness of CAM was mainly affected by fr, while an
increase of dc had no apparent influence on the surface roughness. The reason is that the ideal
surface roughness is a step-like trajectory of the forward movement of the insert; therefore, fr has
a higher influence on the surface roughness. From Fig. 5-4, it is seen that using an aggressive fr
may be an effective way to improve the chip breakage when machining CAM. However, an
aggressive fr will lead to a higher surface roughness. At Vc = 60 m/min and dc= 0.5 mm, when the

132
fr was increased from 0.1 mm/rev to 0.2 mm/rev, the Ra and Rt respectively increased by 2.2 and
1.7 times.
(a) (b)
(c) (d)
Fig. 5-12 Surface roughness (arithmetical average roughness Ra and maximum peak to valley
height Rt) of machined CAM, MS and SS at varying Vc, dc and fr:: (a) Ra at varying Vc and dc, (b)
Rt at varying Vc and dc, (c) Ra at varying Vc and fr, and (d) Rt at varying Vc and fr.
5.4 Results and discussion of continuous machining tests
In above discussion, it is found that the effects of cutting parameters Vc, dc and fr on chip
formation, tool wear and surface roughness when machining CAM were investigated and
compared with MS, SS and NiTi. However, the cutting volume of one workpiece considered in
single machining tests is relatively small and cannot well represent the continuous operations in
0.0
1.0
2.0
3.0
4.0
5.0
15 30 45 60 75 90 105 120
Ra (um)
Vc
(m/min)
fr=0.1mm/rev CAM dc=0.5mm
CAM dc=1.0mm
CAM dc=1.5mm
MS dc=1.5mm
SS dc=1.5mm
0.0
5.0
10.0
15.0
20.0
25.0
30.0
15 30 45 60 75 90 105 120
Rt (um)
Vc
(m/min)
fr=0.1mm/rev CAM dc=0.5mm
CAM dc=1.0mm
CAM dc=1.5mm
MS dc=1.5mm
SS dc=1.5mm
0.0
2.0
4.0
6.0
8.0
15 30 45 60 75 90 105 120
Ra (um)
Vc
(m/min)
dc=0.5mm CAM fr=0.10mm/rev
CAM fr=0.15mm/rev
CAM fr=0.20mm/rev
MS fr=0.20mm/rev
SS fr=0.20mm/rev
0.0
10.0
20.0
30.0
40.0
50.0
15 30 45 60 75 90 105 120
Rt (um)
Vc
(m/min)
dc=0.5mm CAM fr=0.10mm/rev
CAM fr=0.15mm/rev
CAM fr=0.20mm/rev
MS fr=0.2mm/rev
SS fr=0.2mm/rev

133
actual machining. As seen in Fig. 5-9, machining CAM showed abrasion in the nose region of the
insert, which is similar what was observed when machining MS. But on the major cutting edge, in
contrast to CAM, no chipping or chip flow induced notch wear was observed for MS. When a
larger volume of material removal is involved, the long-and-continuous snarled tubular chips
generated when machining CAM may lead to a more severe flank wear than machining MS. To
better understand performance under longer machining conditions, continuous tests on thirty
workpieces were performed on CAM and MS.
5.4.1 Tool wear
The images of nose and flank wear after continuous machining tests on CAM and MS are
shown in Fig. 5-13. As seen in Fig. 5-13 (a), uniformly distributed abrasion wear were observed
in the nose and flank region after continuous machining of MS. The nose and average flank wear
were 156 m and 116 m, respectively, for MS. The nose wear was more pronounced than the
flank wear because mainly loose arc chip were formed during machining MS. As seen in Fig. 5-13
(b), adhered debris and grooves were observed in the nose region after continuous machining
CAM, and significant chip flow damage was observed on the major cutting edge. The apparent
abrasion induced grooves and adhered debris on the nose and significant chip flow induced notches
on the flank when machining CAM indicate that the insert was dulling during the machining and
this may affect the surface quality and final geometry of the workpieces. The nose and average
flank wear after continuous machining CAM were 251 m and 349 m, respectively, which
exceeded the 300 m tool life limit. Overall, after continuous machining with a total cutting length

134
of 4.5 m, the nose wear of machining CAM was 1.6 times that of machining MS, and the average
flank wear of machining CAM was 3 times that of machining MS.
Fig. 5-13 Images of tool wear after continuous machining tests on: (a) MS and (b) CAM.
5.4.2 Surface roughness and dimensional variation
The surface roughness Rt variation of CAM and MS with respect to cutting length/workpiece
number in continuous machining tests is shown in Fig. 5-14. With an increase in the cutting length,
the Rt of MS showed negligible variation. The Rt of CAM was smaller than that of MS at the
beginning and started to increase when the cutting length reached 2.7 m. When cutting length
reached 3.45 m, the Rt from machining CAM was comparable to that from MS. The relative
diameter difference of CAM and MS with respect to cutting length/workpiece number is shown in
Fig. 5-15. The relative diameter difference was obtained as the difference in the final product

135
diameter of each workpiece after machining and the final product diameter of the first workpiece
after machining. Three measurements were taken on each workpiece to obtain the average
diameter before calculating the relative difference. A higher relative diameter difference indicates
a larger deviation in the workpiece geometry due to tool wear as machining continued. As shown
in Fig. 5-15, the relative diameter difference showed a consistent trend with the surface roughness
Rt shown in Fig. 5-14. The relative diameter difference growth of MS was relatively stable. At a
cutting length of 4.5 m, the relative diameter difference of MS was 42 m. Machining CAM
showed a larger scatter in workpiece diameter than MS. A sudden increase of relative diameter
difference was observed when the cutting length reached 3.3 m, the same sudden increase was also
observed on the surface roughness Rt, which may be caused by a sudden chipping of the insert. At
a cutting length of 4.5 m, the relative diameter difference of CAM was 52 m, 10 m larger than
that of MS.
Fig. 5-14 Surface roughness (maximum peak to valley height Rt) of CAM and MS in continuous
machining tests.
0.0
4.0
8.0
12.0
16.0
20.0
0.00 0.45 0.90 1.35 1.80 2.25 2.70 3.15 3.60 4.05 4.50
Rt (m)
Cutting length (m)
MS
CAM
0 3 6 9 12 15 18 21 24 27 30
Workpiece No.

136
Fig. 5-15 Relative diameter difference of CAM and MS in continuous machining tests.
5.5 Summary of findings
It was found that machining Cu-Al-Mn SMA (CAM) mainly showed an abrasion wear
mechanism in the nose region, and chipping and notch wear caused by continuous chip flow was
observed in the flank region. The chip flow damage is considered to be a major challenge when
machining CAM, particularly when performing continuous machining for large stock removal.
Compared with Ni-Ti SMA (NiTi), CAM has a higher machinability. The cutting temperature
during machining CAM was around 1/3 of that when machining NiTi. Chip melting and welding
that occurred when machining NiTi did not occur when machining CAM. The tool wear from
machining CAM was only around 1/7 to 1/21 of that from machining NiTi.
Compared with mild steel (MS) and 304 stainless steel (SS), machining CAM showed
different chip formation, tool wear and surface roughness patterns due to superelasticity and high
-10.0
0.0
10.0
20.0
30.0
40.0
50.0
60.0
0.00 0.45 0.90 1.35 1.80 2.25 2.70 3.15 3.60 4.05 4.50
Relative diameter difference (m)
Cutting length (m)
MS
CAM
0 3 6 9 12 15 18 21 24 27 30
Workpiece No.

137
ductility. Specifically, machining CAM produced long and continuous chips in snarled tubular
form but machining MS and SS produced short and loose arc-shaped chips. The tool wear from
machining CAM was overall close to (0.6 to 1.8 times) that from SS and higher than (0.8 to 2.4
times) that from MS; furthermore, unlike MS and SS, the long and continuous chips generated
when machining CAM led to higher notch wear and chipping in the flank region. The surface
roughness of machined CAM was overall close to that of SS but smaller than that of MS.

138
Chapter 6 - Headed coupling behavior of Cu-Al-Mn SMA
The feasibility of using headed coupling to connect large diameter Cu-Al-Mn SMA with steel
rebar was investigated. Five large diameter Cu-Al-Mn SMA samples with headed ends were
prepared, including one with 30 mm diameter and four with 20 mm diameter. First, mechanical
tests were performed on the five headed samples where each sample was coupled with one steel
rebar at each end. Monotonic, incremental and constant strain cyclic loading was applied to
simulate earthquake loading. The key mechanical properties were extracted and discussed. Second,
microstructural analyses including electron backscatter diffraction (EBSD), metallographic
imaging, Vickers hardness testing, and fractographic evaluation were performed. The crystal
orientation, phase composition, and fracture surfaces were investigated to understand the effect
of the heading process on the stress induced martensitic transformation (SIMT), phase
composition, and failure of the headed Cu-Al-Mn SMA.
6.1 Research motivation
In concrete structures, SMAs are often applied as reinforcement in localized regions to
achieve the desired structural performance while minimizing the use of expensive materials. Most
commonly, SMAs are only applied as flexural reinforcement in the plastic hinge regions where

139
deformation demands are concentrated to take full advantage of their strain recovery and energy
dissipation characteristics. This partial use of the SMAs requires an effective connection with
conventional steel bars [37]. One of the conventionally used methods to connect two rebar is
through mechanical coupling, such as shear screw coupling, threaded coupling or headed coupling,
and adhesive coupling, such as grouted sleeve coupling [150,151]. The shear screw coupling is a
convenient approach; however, it has been reported to suffer from lower ductility and fatigue
resistance because of the premature rupture resulting from the high stress concentrations on and
around the screws [37]. The threaded coupling provides a more reliable connection. Multiple
studies on Cu-Al-Mn SMA adopted this form of connection [10,41,48]. However, the cross-section
needs to be reduced in the middle of the bars to avoid stress concentration at the threaded ends and
premature failure in the threads. For SMAs, this reduction of the cross-section in the middle of the
bars causes significant loss of expensive material and additional machining costs. The grouted
sleeve coupling is commonly used in non-seismic applications. However, this connection form
does not work well for SMAs for the following reasons: quick degradation of the grouted bond
under reserved cyclic loading particularly due to the smooth surface of the SMAs. The smooth
surface cannot provide sufficient friction and engage the confinement, which prevents the use of
grouted sleeve couplers for SMA bars. Therefore, alternative methods for coupling are being
studied.
Headed coupling (also referred to as upset headed coupling) is a cost-effective method widely
used in connecting conventional steel rebar. During the heading process, two prefabricated male

140
threaded steel collars are first inserted on the bars, then each end of the bar (one at a time) is heated,
and the shape of the bar ends is deformed by mechanical extrusion. The deformed headed ends are
connected to each other through a female-to-female threaded steel collar, as shown in Fig. 6-1.
With headed coupling, both tensile and compressive forces can be transferred through the
connection [23,27]. Furthermore, the cross-sectional area of the bars is not reduced, which avoids
material loss and allows the full capacity development of the bars. As such, this coupling method
has great potential in coupling SMAs with each other as well as with conventional steel rebar.
Fig. 6-1 Illustration of headed mechanical couplers.
Due to the short history and limited research on coupling of Cu-Al-Mn SMA, data on headed
coupling behavior of Cu-Al-Mn SMA is very limited. There are only two studies in open literature
on headed coupling of Cu-Al-Mn SMA. Kise et al. [7] investigated the mechanical behavior of
three headed Cu-Al-Mn SMA with a diameter of 13 mm. The small diameter specimens exhibited
Male threaded
steel collar
Female threaded
steel collar
Headed end

141
excellent superelasticity and ductility. One of the specimens had a maximum fracture strain of
more than 35%. In bridges, however, most rebar are over 20 mm and use of 30 mm diameter rebar
is common. Although headed Cu-Al-Mn SMA showed no reduction in superelasticity and ductility
for small diameter bars, the reliability of headed coupling for large diameter Cu-Al-Mn SMA is
still not well understood. Mohebbi et al. [18] performed mechanical tests on two 30 mm diameter
Cu-Al-Mn SMA bars under monotonic and cyclic tensile loading. Premature fracture was observed
on both headed samples. However, since no microstructural analysis was performed, the reasons
for such premature failures on the headed Cu-Al-Mn SMA bar are still unknown.
Previous studies found that the mechanical properties of Cu-Al-Mn SMA exhibit a strong
dependence on their crystal structure and phase composition [28,60]. It is expected that the crystal
structure and phase composition of large diameter Cu-Al-Mn SMA that have undergone high
temperature heading process will be altered. However, conclusive evidence on this matter has been
lacking. To fully investigate this issue, further experiments, especially microstructural analyses,
are needed. To fill the above knowledge gaps, in this study, the headed coupling behavior of large
diameter Cu-Al-Mn SMA was investigated by both mechanical testing and microstructural
analyses.

142
6.2 Experimental program
6.2.1 Specimen preparation
The material tested in this study was single crystal Cu-Al-Mn (CAM) SMA. The composition
and supplier of CAM SMA were the same as previously mentioned in Chapter 3. Five headed
CAM SMAs were prepared: one with 30 mm (labeled as S30-1) and four with 20 mm diameter
(labeled as S20-1, S20-2, S20-3 and S20-4, respectively). The dimensions of the headed CAM
SMAs are shown in Fig. 6-2 (a) and (b). Both 30 mm diameter and 20 mm diameter CAM SMAs
were received in round bars with a length of 30 cm. To simulate the real coupling in bridge
columns, headed Grade 60 mild steel rebar with 32.3 mm diameter (U.S. #10) and 19 mm diameter
(U.S. #6) were also prepared and respectively connected to headed CAM SMAs with diameters of
30 mm and 20 mm. The headed end dimensions of the steel rebar were the same as those of the
headed CAM SMAs shown in Fig. 6-2.
Fig. 6-2 Dimensions of headed specimens: (a) 30 mm diameter, and (b) 20 mm diameter CAM
SMAs.
20
226
30
15
(b)
(a) 42
29
15 6 184 6
255
15 4.5 216 4.5 15
Note: All dimensions are in mm

143
To better simulate the real-life conditions, the same industrial heading process for steel rebar
was followed in this study. The heading process is illustrated in Fig. 6-3. The samples were first
heated by a blowtorch and then headed using a mechanical extruder. After that, the headed
specimens were air cooled naturally to the ambient temperature. After heading, a REED R2020
infrared thermometer was used to monitor the surface temperature of the headed region of each
specimen during cooling. The surface temperatures of the headed end of five CAM specimens
after removing from the extruder with respect to time are shown in Fig. 6-4. It is seen that all
headed specimens cooled down to room temperature in about 30 min and the maximum
temperature during the heading process has exceeded 600 °C.
Fig. 6-3 Heading process: (a) heating the specimen with a blowtorch, and (b) heading the
specimen using a mechanical extruder.
(a) (b)
Mechanical extrusion Blowtorch heating

144
Fig. 6-4 Surface temperature of headed end of five CAM SMAs after removing from extruder.
RT: room temperature.
6.2.2 Test methods
The test setup used for characterizing the mechanical properties of the CAM SMAs is shown
in Fig. 6-5. An MTS 370.5 dynamic servo-hydraulic frame was used to apply the load and two
Epsilon extensometers (with model numbers 3542-0200-100-ST and 3542-0200-050-ST,
respectively) were used to measure the strain. A BMS16HR-53 Mars Labs data acquisition (DAQ)
system was used to record the data. Two different loading configurations were considered in the
mechanical tests. In the first configuration, two different loading protocols were investigated:
tensile strain cycles with 1% increments up to a maximum of 5% strain (LP-1), and five constant
tensile strain cycles with an amplitude of 5% (LP-2). The incremental and constant strain cyclic
loading were selected to simulate earthquake loading [25] and the 5% amplitude was chosen to
cover most of the seismic loading conditions [152]. The test setup for the first two loading
conditions is shown in Fig. 6-5 (a), both of which were performed under extensometer strain
control. An extensometer was used to monitor the strain of the headed CAM SMAs, while two
0
200
400
600
800
1000
0 15 30 45 60
Temperature ( C)
Time (min)
S30-1
S20-1
S20-2
S20-3
S20-4
RT

145
Omega KFH-6-350-C1-11L3M3R strain gauges were applied on the top and bottom steel rebar.
The monitoring of the strain in the steel rebar was performed to ensure that no inelasticity occurs
in the steel rebar and all the inelastic deformation is concentrated in the CAM SMAs as desired in
a real application.
The second loading condition was monotonic tensile loading until fracture as shown in Fig.
6-5 (b). An extensometer was installed in the middle of each of the two headed CAM SMAs.
Additionally, a Celesco CLWG-225-MC4 linear potentiometer was installed between the two ends
of the grip fixture to measure the overall deformation. The loading rate for all the three protocols
Fig. 6-5 Test setup for mechanical property characterization: (a) LP-1 and LP-2, and (b) LP-3.
Bottom headed
steel bar
Top headed
steel bar
Extensometer
Extensometer
LVDT
Headed CAM
SEA sample
(a) (b)
Strain gauge
Strain gauge
Extensometer

146
was 0.4 mm/sec. Samples S30-1, S20-1 and S20-2 fractured during the first loading (LP-1); while
samples S20-3 and S20-4 were still intact after LP-1 (tensile strain cycles with 1% increments up
to a maximum of 5% strain) and subsequent loading LP-2 (five constant tensile strain cycles with
an amplitude of 5%). In the third test, LP-3, S20-3 and S20-4 were connected in series and
monotonically loaded until fracture. The S20-3 fractured during LP-3, and S20-4 remained intact.
After mechanical testing, microstructural analyses were performed to understand the
underlying reasons causing failure. First, the crystal orientations of S30-1, S20-1, S20-2 and S20-
3 were determined by electron backscatter diffraction (EBSD) at the midpoint of each specimen.
Second, metallographic analysis and Vickers hardness tests along the length of S30-1 and S20-3
were performed to determine the effect of heading on the stress induced martensitic transformation
(SIMT). The samples were sliced longitudinally into two halves, and metallographic and Vickers
hardness tests were performed at increasing distances from the fracture location. Third, a fracture
analysis was performed on the broken surfaces of S30-1, S20-1, S20-2 and S20-3. Finally, a
metallographic analysis and Vickers hardness tests were performed on the headed end portion of
S20-1, S20-2 and S20-3. An Mitutoyo HM 200B micro-Vickers hardness machine was used for
the Vickers hardness tests. A JEOL JSM-IT300LV scanning electron microscope (SEM) was used
for the fractographic and metallographic analyses and an Oxford instrument Nordlys Nano EBSD
was used for the crystal orientation measurements.

147
6.3 Results of mechanical tests
The stress-strain curves for all samples are shown in Fig. 6-6 (a) through (f) and a schematic
of the loading protocols, LP-1, LP-2, and LP-3 is shown in Fig. 6-6 (g). S30-1, S20-1 and S20-2
fractured during LP-1 with fracture strains of 4.5%, 4.2% and 4.6%, respectively. S20-3 and S20-
4 remained intact after LP-1 and subsequent LP-2, then in LP-3, they were connected together and
the series system failed when S20-3 broke. The fracture strain of the series system and S20-3 was
11.8% and 10.5%, respectively. All five headed samples exhibited superelastic behavior until
failure. The residual strains of all headed samples after unloading were less than 0.3%, indicating
their strain recovery capacity. The yield strains of top and bottom steel rebar coupled with CAM
SMAs are indicated with vertical pink dashed lines in each subfigure in Fig. 6-6. Except for S20-
1, the top and bottom steel in all the samples did not reach yielding, which indicates that the
deformation is concentrated in the CAM SMAs as expected.
The photographs of fractured samples after mechanical tests are provided in Fig. 6-7. S30-1,
S20-1 and S20-2 fractured in the headed end and S20-3 fractured in the middle portion with
obvious necking. Since S20-4 remained intact after the series test with S20-3, see Fig. 6-6 (f), it is
concluded that S20-4 has equivalent or better deformability than S20-3.

148
Fig. 6-6 Mechanical test results: (a) S30-1 during LP-1, (b) S20-1 during LP-1, (c) S20-2
during LP-1, (d) S20-3 during LP-1 and LP-2, (e) S20-4 during LP 1 and LP 2, (f) S20-3 and
S20-4 connected in series during LP-3, and (g) schematic diagram of three loading protocols
(LP).
 = 0.2%
LP 1
Fracture
 = 0.2%
LP 1
 = 0.2%
LP 1
(a) (b) (c)
(d)
(e)
(f)
(g)
0
5
0
12
Pseudo Time Strain (%)
LP 1
LP 2
LP 3
0
5
1
2
3
4
 = 0.2%
LP 2
 = 0.2%
LP 1
LP 3
S20-4
S20-3
 = 0.2%
LP 2
 = 0.2%
LP 1
Pseudo Time
Pseudo Time

149
It is known that the superelasticity and ductility of CAM SMAs largely depend on the crystal
characteristics such as grain orientation and grain boundaries [11]. The crystal characteristics are
mainly affected by the heat treatment (such as annealing and aging) and forming (such as cold
rolling and extrusion) during alloy manufacturing [59]. Since during the heading process,
blowtorch heating and mechanical extrusion are involved, such operations may have changed the
crystal characteristics of CAM SMAs. To better understand the effect of heading on the mechanical
behavior of CAM SMAs, a schematic diagram of the stress-strain curves of a virgin (without any
disturbance such as heading) single crystal CAM SMA is shown in Fig. 6-8. An ideal stress-strain
curve of CAM SMA can be categorized into four stages. Stage I is the elastic deformation of the
austenitic phase (A). In stage II, the SIMT is triggered, and the A starts to transform into a
martensitic phase (M). In stage III, the SIMT is completed and the material is in full M. In stage
IV, plastic deformation with irreversible slips is developed until fracture. Prior to stage IV, the
deformation can be recovered after unloading as a result of the reversible SIMT.
Fig. 6-7 Photographs of fractured samples after mechanical tests: (a) S30-1, (b) S20-1, (c) S20-
2, and (d) S20-3.
(a)
S30-1
(b)
S20-1
(c)
S20-2
(d)
S20-3
20mm

150
According to the definitions shown in Fig. 6-8, eight key properties were extracted from the
stress-strain curves shown in Fig. 6-6 and listed in Table 6-1. These are: Young’s modulus, Eload,
Young’s modulus after yielding, Eload2, martensitic start stress (also referred to as yield stress),
Ms, austenitic finish stress, Af, stress hysteresis, hy, residual strain r, transformation strain ,
and fracture strain f. Comparing the measured stress-strain curves of headed CAM SMAs (Fig.
6-6) with the ideal one (Fig. 6-8), it is seen that amongst the five samples, the stress-strain curves
of S20-3 and S20-4 are very close to the ideal condition, that is, the SIMT is completed and the
plastic deformation has developed. However, S30-1, S20-1 and S20-2 fractured before entering
stage IV. Therefore, whether the SIMT is completed or not cannot be concluded from the stressstrain curves. Besides, in Fig. 6-6 (b), S20-1 exhibits higher Ms and Eload (which led to the yielding
of the steel rebar), but whether this is due grain orientation or caused by the heading process
Fig. 6-8 Schematical diagram of an ideal stress-strain curve of single crystal CAM SMAs.
Stage I: elastic deformation of Austenitic phase (A), Stage II: transformation from A to
Martensitic phase (M), Stage III: elastic deformation of M, Stage IV: plastic deformation.
Strain
Stress
0 r
Ms
Af
Eload
Eload2
hy=Ms-Af
f
 T>Af
Stage I Stage II Stage III Stage IV

151
remains to be confirmed. Microstructural analyses were performed as presented in the next section
to answer these questions.
Table 6-1 Properties extracted from mechanical tests.
Label
Loading
protocol
Eload Eload2 Ms Af hy r f 
(GPa) (GPa) (MPa) (MPa) (MPa) (%) (%) (%)
S30-1 LP1 35 1.8 215 169 46 0.14 4.5 4.0
S20-1 LP1 99 2.0 446 358 88 0.25 4.2 3.9
S20-2 LP1 36 0.7 279 225 54 0.27 4.6 4.1
S20-3
LP1 37 0.6 256 209 47 0.16 NA NA
LP2 34 0.9 238 205 34 0.03 NA NA
LP3 46 0.7 256 NA NA NA 10.9 8.3
S20-4
LP1 49 0.6 271 237 34 0.23 NA NA
LP2 51 0.5 266 233 33 0.04 NA NA
LP3 40 0.5 272 NA NA NA NA 7.5
Series LP3 28 1.1 248 NA NA NA 11.8 7.2
Note: ‘NA’ means the value is not applicable.
6.4 Results of microstructural analyses
6.4.1 Electron backscatter diffraction (EBSD)
The grain orientation along the loading direction, and the transformation strain measured by
EBSD, ,,EBSD, of S30-1, S20-1, S20-2 and S20-3 are shown in Fig. 6-9. Existing studies have
found that the  of CAM SMAs can be calculated by phenomelogical theory of martensitic
formations [58,114,153], and the obtained isotransformation strains can be shown on an inverse
pole figure, as depicted in Fig. 6-9. After determining , the Ms and Eload of CAM SMAs can be

152
obtained according to the Clausius-Clapeyron relationship [154]. In general, CAM SMAs with
near <0 0 1> orientation have higher , but lower Ms and Eload; and CAM SMAs with near <1 1
1> orientation have higher Ms and Eload, but lower . More details can be found in Kise et al.
[11].
From Fig. 6-9, the ,EBSD of S30-1, S20-1, S20-2 and S20-3 were determined as 8.3%, 4.2%,
6.7% and 8.4%, respectively. From Fig. 6-6 and Table 6-1, the transformation strains extracted
from mechanical tests ,,mechanical of these samples were 4.0%, 3.9%, 4.1%, 8.3%, respectively.
The difference between ,,mechanical and ,,EBSD can be used to evaluate if the SIMT was completed
in the mechanical tests. A large difference means that the headed sample failed before the
completion of SIMT and vice versa. The ,,mechanical of S20-1 and S20-3 was close to ,,EBSD,
indicating that SIMT was completed in these samples prior to fracture. Besides, it is noted that
S20-1 had a near <1 1 1> orientation. According to Kise et al. [11], CAM SMAs with near <1 1
1> orientation exhibit lower  and higher Ms and Eload. Therefore, it is concluded that the high
Fig. 6-9 Transformation strains of S30-1, S20-1, S20-2 and S30-3 measured from EBSD.
10.3
10
9
8
7
6
5
4
3
111
001 101
S30-1
S20-1
S20-2
S20-3

153
Ms and Eload shown in Fig. 6-6 (b) is attributed to the original grain orientation, which eliminates
the effect of the heading process. While for S30-1 and S20-2 the ,,mechanical is substantially
different than ,,EBSD. This difference indicates that the samples fractured before the completion
of SIMT. The premature failure of S30-1 and S20-2 may be due to the heading process on the
crystal structure or phase composition of the samples; however, further metallographic analysis is
needed to confirm this assumption.
6.4.2 Metallographic analysis along bar length
To investigate the effect of heading process on the phase composition and SIMT of CAM
SMAs, metallographic analysis was conducted along the length of two representative samples:
S30-1 and S20-3. For S30-1, four locations (L1~4) at distances of 0 mm, 30 mm, 50 mm and 80
mm (close to the midpoint) from the fracture location were investigated. For S20-3, four locations
(L1~4) at distances of 0 mm, 35 mm, 55 mm and 75 mm (close to the midpoint of the sample)
from the fracture location were investigated. The microstructural images along the length of S30-
1 and S20-3 are shown in Fig. 6-10 and Fig. 6-11, respectively.
From Fig. 6-10 (a) to (c) and Fig. 6-11 (a), it is seen that at L1, L2 and L3 of S30-1 and at L1
of S20-2, high density of bainite phases precipitated, and the density of the bainite phase was seen
to decrease with an increasing distance from the headed end. It is known that the volume fraction
of the bainite phase has a linear relationship with the Vickers hardness, Hv [155]. Therefore, the
Vickers hardness can be used as an indicator of the density of the bainite phase. The

154
interconnection of the bainite phase limits the development of stress induced martensitic plates. In
addition, from Fig. 6-11 (a), cracks between the randomly distributed bainite and martensitic
phases are also observed. The reason for these cracks is that the bainite phase has a higher hardness
and strength than the matrix β phase [58,114]. The random distribution of the bainite phase forms
a near-frame structure, which increases the friction at phase boundary, thereby hindering the
formation and development of stress induced martensite plates. Cracks are generated when the
deformation exceeds the deformation limits. At L4 of both S30-1 and S20-3, the material was
composed of pure β phase (i.e., the phase of the as-received CAM SMAs prior to heading). No
bainite phase precipitation or stress induced martensitic plates were observed. The β phase is the
source of the superelasticity in CAM SMAs, which has a body-centered cubic (BCC) structure
[156]. Therefore, it is concluded that the heading process did not influence the midpoint (along the
length) of both 30 mm diameter and 20 mm diameter specimens.

155
Fig. 6-10 Microstructural images at different distances along the length of S30-1 from the
fracture.
Hv 254
Hv 255
Hv 241
Hv 237
(a)
(b)
(c)
(d)
L2: 30 mm from the fracture
L3: 50 mm from the fracture
L4: 80 mm from the fracture
L1: 0 mm from the fracture
Loading direction
Stress induced martensite plate
Bainite phase
Bainite phase
Stress induced martensite plate
Bainite phase
Stress induced martensite plate
β phase
β phase
β phase

156
Fig. 6-11 Microstructural images at different distances along the length of S20-3 from the
fracture.
ベイナイト相無以降観察出来
ず
(a)
(b)
(c)
(d)
L2: 35 mm from the fracture
L4: 75 mm from the fracture
Hv 247
L1: 0 mm from the fracture
Hv 246
L3: 55 mm from the fracture
Hv 241
Hv 237
Loading direction
Stress induced martensite plate
Stress induced martensite plate
Stress induced martensite plate
Bainite phase
Crack
Plastic deformation
β phase

157
6.4.3 Fractography
In the previous section, it was shown that the heading process precipitated high density bainite
phase near the headed end, and the prevalence of bainite phase decreased with increasing distance
from the headed end. The middle portion of the headed sample was still in β phase. Since the
bainite phase has a higher strength and hardness than the β phase, the headed ends of the specimens
are expected to be stronger, and the failure should occur in away from the end (i.e., the middle
portion). However, as shown in Fig. 6-7, only S20-3 and S20-4 behaved as expected while S30-1,
S20-1 and S20-2 failed at the headed end. To analyze the cause of this unexpected behavior of
S30-1, S20-1 and S20-2, and the underlying mechanisms, in this section and the following section,
fractographic analyses were performed on the headed end.
The fracture surfaces of S30-1, S20-1, S20-2 and S20-3 are respectively shown in Fig. 6-12
to Fig. 6-15. For S30-1 shown in Fig. 6-12, obvious internal defects generated by the heading
process are observed in the middle part of the fracture surface. The long cracks ran almost the full
diameter of the specimen and extend to the surface. In addition, the surface oxidation during the
high temperature mechanical extrusion is also be observed. Along the crack direction, the section
was roughly divided into two parts and four representative locations were selected. Scanning
electron microscopy images at these locations are shown in Fig. 6-12 (a)~(d). The entire fracture
surface of S30-1 exhibited dimple patterns, which are typical features of a ductile fracture [157–
159]. However, the oxidized surface and the cracks reduced the effective cross-sectional area,
thereby intensified the stress concentration at the neck of the headed end. Therefore, the defects

158
generated by the heading process were found to have caused of the premature failure of S30-1 at
the headed end.
For S20-1 and S20-2 respectively shown in Fig. 6-13 and Fig. 6-14, a large river patterned
area and a small, elongated dimple patterned area were observed on the fracture surface. The river
patterns are caused by the interconnection of cleavage planes as the crack moves across the grains,
which is a typical characteristic mechanism for a brittle fracture [157–159]. Whereas for S20-3
Fig. 6-12 Fracture surface of S30-1.
5 mm 5 mm
Cracks
Cracks
a
b
c
d
50 µm
(a) (b)
(c) (d)
50 µm
50 µm 50 µm
Shallow equiaxed dimple
Elongated dimple Deep equiaxed dimple
Deep equiaxed dimple
5 mm
Oxidized surface

159
shown in Fig. 6-15, a pronounced necking occurred, and the circular cross section was deformed
into an ellipsoidal shape. The central portion had equiaxed dimples and the outer portion had
elongated dimples. These are ductile fracture characteristics [159], indicating that the fracture of
S20-3 started from the central portion under tensile stress and then propagated to the outer portion
due to shear stress.
Fig. 6-13 Fracture surface of S20-1.
Fig. 6-14 Fracture surface of S20-2.
1 mm
a
b
c
(a)
10 µm
50 µm
(c)
10 µm
(b)
River pattern
River pattern
Elongated dimple
pattern
1 mm
a
b
c
(a)
10 µm
(b)
(c)
10 µm
10 µm
River pattern
River pattern
Elongated dimple
pattern
Shear lip

160
6.4.4 Metallographic analysis on headed ends
In order to determine the reason why S20-1 and S20-2 showed brittle fractures and S20-3
showed ductile fracture, a metallographic analysis was performed on the headed ends of these
samples. The headed end of each sample was cut into two parts along the longitudinal direction,
then the microstructures were investigated. The results of metallographic analysis on the headed
ends of S20-2, S20-1 and S20-3 are shown in Fig. 6-16 to Fig. 6-18. The results are discussed
below starting with S20-2.
Fig. 6-15 Fracture surface of S20-3.
Elongated dimple
pattern
1 mm
b
c
d
a
(a) (b)
(c) (d)
50 µm 50 µm
50 µm 50 µm
Equiaxed dimple
pattern
Elongated dimple
pattern
Shear lip

161
Fig. 6-16 Microstructures of S20-2 (with brittle fracture) in the headed end.
From the microstructures of S20-2 shown in Fig. 6-16, precipitation of a high density bainite
phase was observed, which is consistent with the observations Fig. 6-10 and Fig. 6-11. However,
it is worth noting that the density of the bainite phase was apparently higher in the end region Fig.
Bainite phase Bainite phase
Bainite phase
Bainite phase
α phase
α phase
Bainite phase
Bainite phase Bainite phase
(a) Hv 292 (b) Hv 269 (c) Hv 262
(d) Hv 277 (e) Hv 259 (f) Hv 248
(g) Hv 280 (h) Hv 283 (i) Hv 264
α phase
b phase
b phase
b phase
b phase
a
b
c
d
e
f
g h i
10 µm 10 µm 10 µm
10 µm 10 µm 10 µm
10 µm 10 µm 10 µm

162
6-16 than in locations away from the end (Fig. 6-10 and Fig. 6-11). In addition, an  phase
precipitation was also observed in the central region, as shown in Fig. 6-16 (c), (e), (f). It is known
that the  phase with a face centered cubic structure precipitates at around 700 ℃, which does not
show superelasticity but has high ductility [156]. It is also noted based on Fig. 6-16 that the density
of the bainite phase and the Vickers hardness were higher in the outer regions, as shown in Fig.
6-16 (a), (b), (d), (g), (h), (i), while the same was lower in the central regions, as shown in Fig.
6-16 (c), (e), (f).
Previous studies found that the bainite phase precipitates from 200 °C to 300 ℃, and the
increase of bainite phase density increases the strength while reducing the ductility [40]. Based on
the microstructures shown in Fig. 6-16, the brittle failure of S20-2 was caused by the
inhomogeneous distribution of the bainite and  phases. The outer portion had a higher density of
bainite phase, which has a high strength but low ductility while the central region had a higher
density of  phase, which has a lower strength but higher ductility. The incompatibility between
the outer bainite phase and the central  phase led to the brittle failure of S20-2 at the headed end
as shown in Fig. 6-14. The microstructures of the headed end of S20-1 and S20-3 shown in Fig.
6-17 and Fig. 6-18, respectively, confirmed this explanation.

163
Fig. 6-17 Microstructures of S20-1 (with brittle fracture) in the headed end.
From Fig. 6-17, it is seen that the bainite phase precipitated in the outer region while the 
phase precipitated in the central region. The incompatibility between the outer bainite phase and
the central  phase led to the brittle fracture of S20-1 at the headed end. For S20-3, it is seen in
Fig. 6-18 that its headed region is uniformly precipitated with high density bainite phase. Since
the bainite phase distribution was homogeneous in the central and outer regions, a uniform
strengthening effect was introduced in the headed ends of the specimen, thus, enabling the sample
to fail with a ductile fracture (see Fig. 6-15).
Bainite phase
(a)
Bainite phase
α phase
α phase
Bainite phase
(b) (c) Hv 283 Hv 239 Hv 241
a b c
b phase
b phase

164
Fig. 6-18 Microstructures of S20-3 (with ductile fracture) in the headed end.
6.4.5 Comparison with previous studies
In Kise et al. [7], three 13 mm diameter CAM SMAs with headed ends were tested and no
premature failure was observed. Whereas in this study, among the five prepared headed CAM
SMAs, S20-1, S20-3 and S20-4 fractured after the completion of the SIMT (the fracture strains of
S20-3 and S20-4 was over 10.9% indicating their high deformation capacity); and S30-1 and S20-
2 fractured prematurely at the headed ends before the completion of SIMT (the fracture strains of
S30-1 and S20-2 was 4.5% and 4.6%, respectively). From the microstructural analyses, it was
Bainite phase
Bainite phase
Bainite phase Bainite phase
(a) Hv 292 (b) Hv 280
a
c
b
d
(c) Hv 308 (d) Hv 278
10 µm 10 µm
10 µm 10 µm

165
found that the premature failure of S30-1 is due to internal defects generated during the heading
process; and that of S20-2 is due to the inhomogeneous distribution of the bainite phase in the
central and peripheral regions of the specimen. Compared with the results reported by Kise et al.
[7], the premature failure observed in this study is attributed to the slower cooling rate of large
diameter bars after heading.
For small diameter headed CAM SMAs investigated in Kise et al. [7], the temperature
variation during cooling in the central and peripheral regions of the sample was uniform. The
uniformly precipitated bainite phase strengthened the headed end of the samples such that the
fracture occurred in the middle portion of the specimens. Whereas for the large diameter CAM
SMAs investigated in this paper, due to the larger cross section, a difference in the cooling rate
between the central and peripheral regions occurred as confirmed with microstructural analysis.
Specifically, during cooling, when the temperature reached 200 °C to 300 ℃, the high density
bainite phase precipitated in the outer regions of the headed ends. At the same time, the temperature
of the central region remained over 500 ℃, leading to  phase precipitation. The  phase is more
ductile than the bainite phase; therefore, the inhomogeneous distribution of bainite phase led to
the strength incompatibility of the cross section in the headed ends and rendered it prone to a
premature failure. Therefore, it is concluded that key to effective headed coupling of large diameter
CAM SMAs is to ensure a consistent cooling rate in the central and peripheral regions of the
headed region.

166
6.5 Summary of findings
It was found that, for both 30 mm diameter and 20 mm diameter headed Cu-Al-Mn SMAs,
the heading process only affected the near end portions. Specifically, after heading, high density
bainite phase precipitated near the headed end regions, and the density of bainite phase decreased
with an increasing distance from the headed ends. The middle portion of the sample was not
affected, and it was still in the original b phase. The interconnection of the high density bainite
phase limited the motion and development of martensitic plates. The randomly distributed bainite
phase formed a near-frame structure, which increased the friction of the phase boundary
movement, thereby hindering the formation and development of stress induced martensite plates.
However, since the bainite phase has a higher strength and hardness than the matrix, its highdensity precipitation played an overall strengthening role in the headed end portion of Cu-Al-Mn
SMA.
In the mechanical tests with two ends coupled to steel rebar, all five headed samples coupled
with steel rebar exhibited superelasticity with no degradation until failure. The residual strains of
all the headed samples after unloading were less than 0.3%, indicating their strain recovery
capacity. The 30 mm sample, S30-1, fractured at a strain of 4.5% before the completion of the
stress induced martensitic transformation (SIMT). The premature failure of S30-1 was due to the
internal defects generated during heading. Among the four 20 mm diameter samples, S20-1, S20-
3 and S20-4 fractured after the completion of SIMT, indicating that the effect of the heading
process on their superelasticity is negligible. The fracture strains of S20-3 and S20-4 were over

167
10.9%, indicating their high deformation capacity. S20-2 fractured before the completion of SIMT
with a strain of 4.6%. The premature failure of S20-2 was attributed to the inhomogeneous
distribution of the bainite phase in the central and peripheral regions of the headed portions of the
sample.
The key to effective headed coupling of large diameter Cu-Al-Mn SMA bars was found to be
a consistent cooling rate in the central and peripheral regions of the headed end during cooling.
The reason is that for large diameter Cu-Al-Mn SMA bars after heading, due to the larger crosssectional area, a differential cooling rate occurs between the central and peripheral regions after
heading. The peripheral region cools faster, thus precipitating a high density of bainite phase, while
the central region cools more slowly, resulting in the precipitation of the  phase. The  phase has
a higher ductility than the bainite phase. Therefore, the inhomogeneous distribution of the bainite
phase leads to the strength incompatibility within the cross section of the headed ends, which is
prone to a premature failure.

168
Chapter 7 - Cost estimation of bridge columns reinforced with
Cu-Al-Mn SMA
The cost effectiveness of applying Cu-Al-Mn SMA in typical bridge columns was investigated.
Comparisons with columns reinforced with Ni-Ti SMA and conventional steels were made. The
cost of producing, processing, and coupling SMA bars were considered. Three typical bridge
columns were designed, namely a conventional reinforced concrete (RC) column, a concrete
column reinforced with Cu-Al-Mn SMA, and a concrete column reinforced with Ni-Ti SMA. The
three typical bridge columns were designed to have the same flexural capacity, and the cost
effectiveness of each type of column was evaluated.
7.1 Research motivation
Existing research on the application of SMA in bridges mainly focuses on the binary Ni-Ti
alloy compositions due to their earlier discovery. However, it is known that the Ni-Ti SMAs are
expensive when used in large quantities. In addition, they are difficult to process, and the effective
connection or coupling of Ni-Ti-based SMAs with conventional steel rebar remains challenging.
The high material cost, poor machinability, and difficulty in connecting with steel rebar, make the
application of Ni-Ti SMAs more expensive.

169
Cu-Al-Mn SMA, which shows excellent low-cycle fatigue stability and superelasticity in a
wide range of temperature from -40 ℃ to 50 ℃, has potential for use in bridges. Since the
manufacturing of Cu-Al-Mn SMA does not involve any noble metal elements, its material cost is
lower than conventional Ni-Ti-based SMAs. Furthermore, the better machinability of Cu-Al-Mn
SMA further reduces the cost when applied in bridges compared with Ni-Ti-based SMAs.
However, as mentioned in Chapter 4, the yield strength of Cu-Al-Mn SMA is substantially lower
than that of Ni-Ti SMA, that means the required amount of material to satisfy the same flexural
capacity will be higher when using Cu-Al-Mn and Ni-Ti SMAs in bridges. However, due to the
short application history, no studies have been reported on this topic. Research is needed to
determine the cost effectiveness of Cu-Al-Mn SMA over conventional Ni-Ti SMA when applied
in bridge columns. To fill this knowledge gap, this study presented the cost estimation analysis of
typical bridge columns reinforced with Cu-Al-Mn SMA. Comparisons with columns reinforced
with Ni-Ti SMA and conventional reinforced concrete (RC) columns were also made.
7.2 Design of bridge columns
7.2.1 Material and geometric properties
Three typical bridge columns were designed, namely a conventional reinforcement concrete
(RC) column, a concrete column reinforced with Cu-Al-Mn SMA, and a concrete column
reinforced with Ni-Ti SMA. For brevity, Cu-Al-Mn SMA and Ni-Ti SMA are respectively referred
to as CAM SMA and NiTi SMA hereafter. Columns reinforced with steel rebar, CAM SMA and
NiTi SMA are respectively referred to as steel RC, CAM RC and NiTi RC hereafter.

170
The loads, geometry and material properties of the typical bridge column were selected based
on the NCHRP Research Report 864 [160]. The factored loads applied on the column were 2577
kN-m bending moment and 5542 kN axial load. The column diameter was D = 1219 mm (4 ft);
the column height was L = 5791 mm and the concrete cover was c = 51 mm. The transverse
reinforcement was U.S. #7, dtr = 22.2 mm, spaced at 89 mm, with a transverse reinforcement ratio
of 1.3%. Regarding the material properties, the compressive strength of concrete was fcc = 27.6
MPa, and the yield strength of longitudinal and transverse reinforcement was fyl = fytr = 414 MPa.
Both CAM and NiTi SMA were 30 mm diameter round bars. The yield strengths of CAM and
NiTi SMA were fyCAM = 280 MPa and fyNiTi = 380 MPa, respectively. The Young’s modulus of
CAM and NiTi SMA were ECAM = 51.2 GPa and ENiTi = 37.9 GPa, respectively. The parameters
of CAM were obtained from the sample shown in Fig. 8-8 (a), which will be discussed later. The
parameters of NiTi were determined based on NCHRP Research Report 864 [160]. Because the yield
strength of steel rebar is higher than those of CAM and NiTi SMA, columns reinforced with SMAs
have a lower flexural strength than those reinforced with steel. Therefore, to make a fair
comparison of bridge columns with different reinforcements, all columns were designed for the
same flexural capacity. This resulted in steel RC, NiTi RC and CAM RC columns to have,
respectively, 12, 33 and 42 longitudinal bars and 1.04%, 1.49% and 1.90% longitudinal
reinforcement ratios, as shown in Fig. 7-1. The clear spacing between the longitudinal bars for
steel RC, CAM RC and NiTi RC was 280.1 mm, 83.0 mm, and 59.1 mm, respectively.

171
(a) (b) (c)
Fig. 7-1 Sectional configuration of columns: (a) CAM RC, (b) NiTi RC, and (c) steel RC.
7.2.2 Capacity check
According to the NCHRP Research Report 864 [160], the axial capacity of the columns was
calculated by
= 0.75(1( − ) + ) Eq. 7-1
where fcc was the nominal compressive strength of concrete, z1 was the upper limit strength
modifier, Ag was the gross area of the cross-section, Al was the area of the longitudinal
reinforcement, and fyl was the nominal austenitic yield strength of SMA bars.
The axial demand to axial capacity ratio, D/R, for the three columns were, respectively,
D/RCAM = 0.196, D/RNiTi = 0.188, and D/RSteel = 0.204. Fig. 7-2 shows the axial force-bending
moment interaction curves of the three designed columns, which were generated using SAP2000
[161]. The constitutive model for different materials was the same as those used in the moment-

172
curvature analysis, which will be presented in the following section. As seen from the interaction
diagrams, all three designed columns satisfy the requirements.
(a) (b)
(c)
Fig. 7-2 Axial force-bending moment interaction diagram of designed columns: (a)
CAM RC, (b) NiTi RC, and (c) steel RC.
7.3 Analysis of bridge columns
7.3.1 Moment-curvature analysis
A sectional analysis using SAP2000 [161] was conducted to obtain the moment-curvature
response of the designed columns. Constitutive models for different material were adopted
following the NCHRP Research Report 864 [160]. In total, three different materials were
-10000
0
10000
20000
30000
40000
0 2000 4000 6000 8000
Aixal force (kN)
Moment (kN-m)
CAM
Demand
-10000
0
10000
20000
30000
40000
0 2000 4000 6000 8000
Aixal force (kN)
Moment (kN-m)
NiTi
Demand
-10000
0
10000
20000
30000
40000
0 2000 4000 6000 8000
Aixal force (kN)
Moment (kN-m)
Steel
Demand

173
considered, namely, unconfined concrete cover, confined concrete core, and the longitudinal
reinforcement (including CAM, NiTi SMA bars, and steel rebar). The stress-strain curves of
different materials used in the sectional analysis are shown in Fig. 7-3.
(a) (b)
(c) (d)
(e)
Fig. 7-3 Constitutive models used in sectional analysis: (a) CAM SMA, (b) NiTi SMA, (c)
confined concrete, (d) unconfined concrete, and (e) steel rebar.
0
200
400
600
800
0 0.02 0.04 0.06 0.08 0.1
Stress (MPa)
Strain
CAM
0
250
500
750
1000
0 0.02 0.04 0.06 0.08 0.1
Stress (MPa)
Strain
NiTi
-50
-40
-30
-20
-10
0
-0.012 -0.009 -0.006 -0.003 0 Stress (MPa)
Strain
Confined Conc
-40
-30
-20
-10
0
-0.005 -0.004 -0.003 -0.002 -0.001 0 Stress (MPa)
Strain
Unconfined Conc
-750
-500
-250
0
250
500
750
-0.06 -0.04 -0.02 0 0.02 0.04 0.06
Stress (MPa)
Strain
Steel

174
Using the above material properties and the applied axial load, the moment-curvature
diagrams of three columns were obtained as shown in Fig. 7-4, where the idealized bilinear curves
were shown in dashed lines. From the idealized bilinear moment-curvature diagrams, it is seen that
with the selected cross-sectional configurations, the three types of columns had almost identical
plastic moment capacity.
(a) (b)
(c)
Fig. 7-4 Moment-curvature diagrams of three types of columns: (a) CAM RC, (b) NiTi
RC, and (c) steel RC.
0
1500
3000
4500
6000
0.0E+00 1.0E-02 2.0E-02 3.0E-02 4.0E-02 5.0E-02
Moment (kN-m)
Curvature
Idealized
CAM
0
1500
3000
4500
6000
0.0E+00 1.0E-02 2.0E-02 3.0E-02 4.0E-02 5.0E-02
Moment (kN-m)
Curvature
Idealized
NiTi
0
1500
3000
4500
6000
0.0E+00 1.0E-02 2.0E-02 3.0E-02 4.0E-02 5.0E-02 6.0E-02
Moment (kN-m)
Curvature
Idealized
Steel

175
7.3.2 Drift capacity
Available formulations of plastic hinge theory were used to roughly estimate the drift ratio of
the columns based on the idealized bilinear moment-curvature diagrams. The deformation was
assumed to be localized in the plastic hinge region. This simplification works well for conventional
RC columns for two reasons. First, deformed steel rebar shows strain compatibility with the
concrete. Second, the localized deformation assumption well approximates the actual deformation
along the length of the column. On the other hand, SMA bars have a smooth surface, and they are
usually intentionally deboned from the concrete to further localize cracks [10]. The development
of more accurate formulations to predict the SMA reinforced column response is outside the scope
of this study. The objective here is to perform a cost comparison of the three types of RC columns;
therefore, only a rough estimation of the column deformations is required.
According to NCHRP Research Report 864 [160], the plastic hinge length Lp of steel RC
column was obtained as:
= max{ 0.3, 0.08 + 0.15 } Eq. 7-2
where fyl and dyl were, respectively, the yield strength and the diameter of the steel rebar; and L
was the column height. Similarly, the displacement, , at the top of the column was obtained as
 =

2
3
+ ( − )( −

2
) Eq. 7-3
where y was the idealized yield curvature and u was the idealized ultimate curvature when
concrete reaches the ultimate strain 0.01.

176
Based on the above calculation, for steel RC, the plastic hinge length, Lp, steel, the displacement
at the top end of the column, steel, and the drift ratio, steel, were, respectively, obtained as 786
mm, 166 mm and 2.863%. For SMA reinforced columns, targeting the same displacement at the
top end of the column as in the steel RC column, the length of the CAM SMA and NiTi SMA bars
were obtained as 674 mm and 514 mm, respectively.
Fig. 7-5 shows the arrangement of longitudinal reinforcement for the three types of bridge
columns. For SMA reinforced columns, the CAM SMA and NiTi SMA were only applied in the
plastic hinge region while the remaining length was steel rebar. Because the most commonly
available CAM SMA and NiTi SMA are 30 cm long, the CAM SMA and NiTi SMA have to be
spliced to reach the desired length. Hypothetically, mechanical couplers were used to splice the
CAM SMA and NiTi SMA, as shown in Fig. 7-5. For CAM RC, three pieces of SMA bars and
four mechanical couplers were needed. For NiTi RC, two pieces of SMA bars and three mechanical
couplers were needed.

177
Fig. 7-5 Arrangement of longitudinal reinforcement for three types of columns: (a)
CAM RC, (b) NiTi RC, and (c) steel RC.
After obtaining the length of the SMA bars, the price of each column can be evaluated based
on how much materials and coupling methods are used.
7.4 Cost estimation of typical columns
It was assumed here that the difference in the cost of the three columns was only due to the
longitudinal reinforcement. This assumption is approximate because the use of SMA
reinforcement may alter other aspects of construction. The cost of SMA reinforcement includes:
Steel rebar
CAM
SMA bar
NiTi
SMA bar
Lcolumn
(a) CAM RC (b) NiTi RC (c) Steel RC
LNiTi
LCAM
Note: This figure is not drawn to scale

178
SMA bars, machining or heading, and mechanical couplers. In the following sections, the cost of
three types of reinforcement was estimated.
The cost of steel RC was obtained as follows. Price of each steel reinforcement Esteel was
computed by
= × Eq. 7-4
= × × (

2
)
2 ×
1
1000
× ×
1
1000
× Eq. 7-5
where Wsteel was the weight of each steel reinforcement, Nsteel was the number of steel rebar (i.e.,
12), dsteel was the diameter of steel rebar (i.e., 36 mm), steel was the density of steel rebar (i.e., 7.8
g/cm3
), and Psteel was the price of steel rebar $2.2 /kg. The total cost for all steel rebar Tsteel was
computed by
For CAM RC and NiTi RC, two types of coupling methods were considered, namely headed
end coupling (referred to as heading method hereafter) and dog-bone machined coupling (referred
to machining method hereafter). The heading method involves firstly heading the two ends of
SMA bars and steel rebar, then using mechanical couplers to connect them. The machining method
involves firstly machining the SMA bars into dog-bone shape, then threading the two ends, and
finally using mechanical couplers to connect SMA bars together. Because the machining method
results in a smaller cross-sectional area, the cost on SMA material and coupling are different. The
= × Eq. 7-6

179
main differences between heading and machining methods include: the effective cross-sectional
area of the SMA bars, the type of couplers, and the labor involved in the coupling.
When using machining method, the 30 mm diameter as received CAM SMA bars were
reduced into an effective diameter of 26 mm. The steel in the remaining length was U.S. #9 rebar
with a diameter of 28.7 mm. The cost of CAM RC columns was calculated as shown below. The
cost of each reinforcement bar ECAM-machine constituted four parts: the cost of CAM SMA material
Ematerial, the cost of machining Emachining, the cost of mechanical coupler Ecoupler, and the cost of the
remaining steel rebar Eremainsteel. The cost of CAM SMA material Ematerial was obtained as
= × × (
−ℎ
2
)
2 ×
1
1000
× ×
1
1000
× Eq. 7-7
where NCAM was the number of CAM SMA bars (i.e., 42), LCAM was the length of CAM SMA in
the plastic hinge region (i.e., 674 mm), nCAM was the number of CAM SMA bars in each
longitudinal rebar (i.e., 3), dsteel was the diameter of steel rebar (i.e., 28.7 mm (U.S. #9)), CAM was
the density of CAM SMA (i.e., 7.1 g/cm3
), and PCAM was the price of CAM SMA (i.e., $66.14
/kg). Note that the diameter of CAM SMA before machining was 30 mm, after machining it was
26 mm.
The cost of machining CAM SMA Emachining was obtained as
ℎ = ℎ− × Eq. 7-8

180
where Pmachine-CAM was the price of machining each piece of CAM SMA. The cost of machining
usually includes four parts; namely, cost of setup, Csetup, cost of machining, Cturning (turning is
considered in this study), cost of inserts, Cinsert, and cost of cutting the machined specimens into
designed length, Ccutting.
When machining CAM SMA, the four parts of machining cost were respectively taken as
Csetup = $120 /h, Cturning = $120 /h, Cinsert = $10 /each, Ccutting = $0 (cutting CAM SMA is as easy
as cutting steel, therefore, no extra cost was added). For each piece of CAM SMA machining, only
one setup work and one piece of insert was needed. The time required for each piece of CAM
SMA machining was taken as 0.25 h. 100 specimens were considered to compute the average price
of machining CAM SMA. Therefore, the price of machining each piece of CAM SMA was
obtained as
ℎ− =
+++
100
Eq. 7-9
The cost of machining CAM SMA Emachining was obtained as
ℎ = 46 × Eq. 7-10
The cost of mechanical coupler Ecoupler was obtained as
= × ( + 1) Eq. 7-11
where Pcoupler was the cost the cost of couplers (both for machining and heading), estimated as
$10/bar, including two couplers per bar, one for each end.

181
The cost of remaining steel rebar Eremainsteel was obtained as
= ( − ) × × (

2
)
2 ×
1
1000
× ×
1
1000
× Eq. 7-12
Therefore, the cost of each CAM reinforcement ECAM-machine was obtained as
−ℎ = + ℎ + + Eq. 7-13
Total cost of all reinforcement when using machining TCAM-machine was computed by
−ℎ = −ℎ × Eq. 7-14
When using heading, a smaller number of SMA bars is required, because the heading method
does not reduce the effective cross-sectional area. The number of CAM SMA bars NCAM was 32.
The diameter of CAM SMA before and after heading was all 30 mm. The diameter of remaining
steel rebar dsteel was 32.3 mm (U.S. #10). All the other variables were the same as those above for
the machining method. The cost of heading each bar, Pheading-CAM and Pheading-NiTi was estimated as
$10/bar, which included both ends of the bar.
Using the same approach, the cost of NiTi RC columns was obtained for machining and
heading methods. When using machining method, the number of NiTi SMA bars NNiTi was 33. The
length of NiTi SMA in plastic region LNiTi was 514 mm, which required nNiTi = 2 as received NiTi
SMA bars (each as received NiTi SMA bar was 300 mm in length). The diameter of NiTi SMA
before machining was 30 mm, after machining it was 26 mm. The diameter of remaining steel
rebar dsteel was 28.7 mm (U.S. #9) for machining and it was 32.3 mm (U.S. #10) for heading. The

182
density of NiTi SMA NiTi was 6.5 g/cm3
. The price of NiTi SMA PNiTi was $154.3 /kg. When
machining NiTi SMA, the four parts of machining cost were respectively taken as Csetup = $120
/h, Cturning = $120 /h, Cinsert = $100 /each, Ccutting = $300. For each piece of NiTi SMA machining,
one setup work and two pieces of inserts were needed. The time required for each piece of NiTi
SMA machining was taken as 3.25 h. 100 specimens were considered to compute the average price
of machining NiTi SMA. When using heading method, the number of NiTi SMA bars NNiTi was
25. The same calculation process was adopted, and for simplicity, it is not repeated here.
To sum up, a summary of the cost estimation results is provided in Table 7-1.
Table 7-1 Summary of the cost estimation results.
N steel 12 N CAM 42 N CAM 32 N NiTi 33 N NiTi 25
d steel (mm) 36 L CAM (mm) 674 L CAM (mm) 674 L NiTi (mm) 514 L NiTi (mm) 514
r hosteel (g/cm3) 7.8 n CAM 3 n CAM 3 n NiTi 2 n NiTi 2
P steel ($/kg) 2.2 d steel (mm) 28.7 d steel (mm) 32.3 d steel (mm) 28.7 d steel (mm) 32.3
E steel ($) 101 ρ CAM (g/cm3) 7.1 ρ CAM (g/cm3) 7.1 ρ NiTi (g/cm3) 6.5 ρ NiTi (g/cm3) 6.5
T steel ($) 1212 P CAM ($/kg) 66.14 P CAM ($/kg) 66.14 P NiTi ($/kg) 154.3 P NiTi ($/kg) 154.3
E material ($) 224 E material ($) 224 E material ($) 364 E material ($) 364
P machine-CAM ($/each) 46 P heading-CAM ($/each) 10 P machine-NiTi ($/each) 46 P heading-NiTi ($/each) 10
E machining ($) 138 E heading ($) 30 E machining ($) 896 E heading ($) 20
P coupler ($/each) 10 P coupler ($/each) 10 P coupler ($/each) 10 P coupler ($/each) 10
E coupler ($) 40 E coupler ($) 40 E coupler ($) 30 E coupler ($) 30
E remainsteel ($) 57 E remainsteel ($) 72 E remainsteel ($) 59 E remainsteel ($) 74
E CAM-machine ($) 459 E CAM-heading ($) 336 E NiTi-machine ($) 2245 E NiTi-heading ($) 493
T CAM-machine ($) 19278 T CAM-heading ($) 11712 T NiTi-machine ($) 74085 T NiTi-heading ($) 12325
Conventional RC CAM with Machining CAM with Heading NiTi with Machining NiTi with Heading

183
Based on the above calculations, the cost comparison of steel RC, CAM RC, and NiTi RC is
shown in Fig. 7-6. The total cost of longitudinal reinforcement of a steel RC was $1212. While for
CAM RC, the total cost was $19,278 when using machining and $11,712 when using heading. The
cost increased by $18,066 (when using machining) and $10,500 (when using heading) in
comparison with the conventional steel RC column. For each NiTi RC column, the total cost was
$74,085 when using machining and $12,325 when using heading. The cost increased by $72,873
(when using machining) and $11,113 (when using heading) in comparison with conventional steel
RC column.
N steel 12 N CAM 42 N CAM 32 N NiTi 33 N NiTi 25
d steel
(mm) 36 L CAM (mm) 674 L CAM (mm) 674 L NiTi (mm) 514 L NiTi (mm) 514
r hosteel
(g/cm3) 7.8 n CAM 3 n CAM 3 n NiTi 2 n NiTi 2
P steel
($/kg) 2.2 d steel
(mm) 28.7 d steel
(mm) 32.3 d steel (mm) 28.7 d steel (mm) 32.3
E steel
($) 101 ρ CAM (g/cm3) 7.1 ρ CAM (g/cm3) 7.1 ρ NiTi (g/cm3) 6.5 ρ NiTi (g/cm3) 6.5
T steel
($) 1212 P CAM ($/kg) 66.14 P CAM ($/kg) 66.14 P NiTi ($/kg) 154.3 P NiTi ($/kg) 154.3
E material
($) 224 E material
($) 224 E material ($) 364 E material ($) 364
P machine-CAM ($/each) 46 P heading-CAM ($/each) 10 P machine-NiTi ($/each) 46 P heading-NiTi ($/each) 10
E machining
($) 138 E heading
($) 30 E machining ($) 896 E heading ($) 20
P coupler
($/each) 10 P coupler
($/each) 10 P coupler ($/each) 10 P coupler ($/each) 10
E coupler
($) 40 E coupler
($) 40 E coupler ($) 30 E coupler ($) 30
E remainsteel
($) 57 E remainsteel
($) 72 E remainsteel ($) 59 E remainsteel ($) 74
E CAM-machine
($) 459 E CAM-heading
($) 336 E NiTi-machine ($) 2245 E NiTi-heading ($) 493
T CAM-machine
($) 19278 T CAM-heading
($) 11712 T NiTi-machine ($) 74085 T NiTi-heading ($) 12325
Conventional RC CAM with Machining CAM with Heading NiTi with Machining NiTi with Heading

184
Fig. 7-6 Cost comparison of steel RC, CAM RC, and NiTi RC.
7.5 Summary of findings
It was found that the Cu-Al-Mn SMA reinforced column showed economic advantage over
the Ni-Ti SMA reinforced column particularly if the machining method was used to connect SMA
bars with the steel rebar. Compared to the cost of conventional RC columns, the additional cost of
Cu-Al-Mn SMA reinforced column was only about 1/4 of the cost on Ni-Ti SMA reinforced
column, indicating the cost effectiveness of Cu-Al-Mn SMA.
Even though the cost of Cu-Al-Mn SMA reinforced column and Ni-Ti SMA reinforced
columns was very close when using heading method, it should be noted that the reliable coupling
between large diameter Ni-Ti SMA bars with steel rebar to resist seismic loading remains
challenging and requires more studies. Therefore, considering the most commonly used coupling
method at present, i.e., threaded coupling by machining, it is concluded that when using Cu-AlMn SMA in bridges columns, the total extra cost reduces by up to 75% than using Ni-Ti SMA.
1,212
19,278
11,712
74,085
12,325
0
20,000
40,000
60,000
80,000
100,000
Steel RC CAM RC using
machining
CAM RC using
heading
NiTi RC using
machining
NiTi RC using
heading
Cost ($)
Column type

185
It should be noted that, in this study, it was assumed that the difference in the cost of the three
columns was only due to the longitudinal reinforcement. This assumption is approximate because
the use of SMA reinforcement may alter other aspects of construction. The added cost to the total
cost of the column with Cu-Al-Mn SMA using heading is $11,712, which is expected to be a trivial
increase in the overall cost of the column and the bridge. The reason is that for conventional RC
column, its permanent deformation after earthquake is still a problem. The cost used to repair or
rehabilitate the column after an earthquake should also be considered. The strain recovery and
energy dissipation capacity of SMA bars are shown to reduce the permanent deformation of SMA
reinforced columns up to 91% compared with the conventional RC column after being subjected
to a peak drift of 7%, as reported by Hosseini et al. [10]. Therefore, in comparison with
conventional steel RC, columns reinforced with superelastic SMA bars still save the total cost in
the long term.

186
Chapter 8 - Effect of temperature on superelasticity and ductility
of Ni-Ti-Co SMA
The effect of temperature on superelasticity and ductility of Ni-Ti-Co SMA was investigated.
Cyclic loadings at different temperatures ranging from -60 ℃ to 50 ℃ were performed. To
benchmark the behavior of Ni-Ti-Co SMA, Ni-Ti and Cu-Al-Mn SMAs were also tested. Key
mechanical properties such as yield strength, ductility, maximum recovery strain, and temperature
dependence of the yield strength of Ni-Ti-Co, Ni-Ti and Cu-Al-Mn SMAs were extracted and
discussed.
8.1 Research motivation
As mentioned in the previous chapters, existing research on Ni-Ti-based SMA mainly focused
on binary Ni-Ti compositions due to their early discovery. However, since the manufacturing of
large size Ni-Ti SMA remains challenging, most of existing research on Ni-Ti SMA was limited
to either thin-strip or wire-formed samples. In addition to the size limitation, existing Ni-Ti SMAs
normally have a martensitic transformation start temperature higher than -25 ℃ [29], which means
they will lose superelasticity at low temperatures. Due to these limitation of Ni-Ti SMAs, alternate
materials are being developed.

187
Ni-Ti-Co SMA is potentially attractive for application in bridge columns due to its high
strength and availability in large diameters. As reported by Kishi et al. [51], the addition of Cobalt
(Co) could increase the yield strength and decrease the martensitic transformation start temperature.
Compared with conventional binary Ni-Ti SMA, the yield strength of Ni-Ti-Co SMA could be
more than 50% higher [52]. At the time of writing this dissertation, Ni-Ti-Co SMA bars with a
diameter over 32 mm has been developed. However, due to the short research history, only very
limited publications on thin wire-shaped Ni-Ti-Co SMA are available, and the basic mechanical
properties of large size Ni-Ti-Co SMA have not been reported. To fill this knowledge gap, the
mechanical behavior of large size Ni-Ti-Co SMA was characterized in this chapter. Comparisons
with other existing SMAs including Ni-Ti and Cu-Al-Mn SMAs were also made.
8.2 Material and methods
8.2.1 Materials
Three types of materials were included in this study, namely Ni-Ti-Co SMA, Ni-Ti SMA and
single crystal Cu-Al-Mn SMA. Both Ni-Ti-Co SMA and Ni-Ti SMA were obtained from SAES
Smart Materials. The manufacturer of single crystal Cu-Al-Mn SMA used in this study was the
same as that mentioned in Chapter 3. The as received dimensions of three SMAs were as follows.
The Ni-Ti-Co SMA was received as round bars with 32 mm diameter and approximately 1.8 m
length. The Ni-Ti SMA was received as round bars with 27.2 mm diameter and approximately 1.9
m length. The Cu-Al-Mn SMA was received as round bars with 20 mm diameter and 0.3 m length.

188
The dimensions of Ni-Ti-Co, Ni-Ti and Cu-Al-Mn SMA samples used in this study are shown
in Fig. 8-1. To compare Ni-Ti-Co, Ni-Ti and Cu-Al-Mn SMAs under the same conditions, all three
materials were machined into cylindrical dog-bone samples with the same effective diameter of
12.7 mm. The thread length of Ni-Ti-Co and Ni-Ti SMA samples was 5 mm longer than that of
Cu-Al-Mn SMA samples considering their higher strength. The following machining methods
were adopted. For Cu-Al-Mn SMA, only traditional computer numerical control (CNC) machining
was used for the sample preparation. Whereas for Ni-Ti-Co and Ni-Ti SMA, due to their difficult
machinability, both traditional CNC machining and non-traditional machining techniques were
used. First, they were cut into short pieces with a water jet; then, electro discharge machining
(EDM) was used to reduce them into round bars with a diameter of 18 mm; finally, traditional
CNC machining was used to form the dog bone shape and add threads on both ends. After
machining all Ni-Ti-Co and Ni-Ti SMA samples into designed shape, they were heat treated to
stabilize their martensitic transformation. The heat treatment temperature was 375 ℃ with a
duration of 40 mins.

189
Fig. 8-1 Dimensions of (a) Ni-Ti-Co and Ni-Ti, and (b) Cu-Al-Mn SMA specimens.
8.2.2 Methods
The test setup used in this study is shown in Fig. 8-2. An MTS 370.5 dynamic servo-hydraulic
frame was used to apply the load. An MTS 651.06E−04 environmental chamber was used along
with the MTS load frame to house the specimens during testing. A liquid nitrogen tank was used
for low temperature tests, as shown in Fig. 8-2 (a). Two extension rods were used to connect the
sample to hydraulic griping systems. A 50.8 mm gauge length Epsilon extensometer (model
number 3542-0200-100-LHT) was used to measure the strain. A BMS16HR-53 Mars Labs data
acquisition (DAQ) system was used to record the data.
1 8 12.7
35 1 2 5 6
R=45
150
12 3 5
M18-1.5 thread 12.7 M18-1.5 thread
30 1 2 5 6
R=45
140
12 30
12.7
Note: All dimensions are in mm
M18-1.5 thread M18-1.5 thread
(a)
(b)
1 8 12.7

190
Fig. 8-2 Test setup for mechanical tests on Ni-Ti-Co, Ni-Ti and Cu-Al-Mn SMA samples.
Four different loading protocols (LPs) were performed in this study. The LP-1 was 1% strain
incremental cyclic loading until failure at room temperature, 23 ℃. The LP-2 was 1% strain
incremental cyclic loading until 5% (without breaking the sample) under different temperatures.
The temperatures for LP-2 were different for different materials, which will be discussed in detail
in the next section. Both LP-1 and LP-2 were performed under extensometer control. The LP-3
was 1% strain incremental cyclic loading until failure at room temperature, 23 ℃, using crosshead
displacement control. For incremental cyclic loading in LP-1, LP-2 and LP-3, after reaching the
target strain at each cycle, the sample was unloaded to near-zero force under force control, then
the next increment was initiated. The LP-4 was loading the sample to 7% strain then unloading to
near-zero force under different temperatures. The loading rate was 0.04 mm/min for all four LPs.
For LP-2 and LP-4 performed at different temperatures, the following procedures were adopted.
First, install and place the sample in the chamber; then, adjust the chamber temperature to the
Environmental
chamber
MTS load
frame
Liquid
nitrogen
tank
Extension rod
Extensometer
Sample
Extension rod
(a) (b)

191
target value, wait 40 minutes until the specimen temperature reaches equilibrium; and finally,
intimate the loading.
8.3 Results and discussion
8.3.1 Ni-Ti-Co SMA
The results of 1% strain incremental cyclic loading until failure on Ni-Ti-Co SMA at 23 ℃ is
shown in Fig. 8-3 (a). Ideal flag shape stress-strain curves were observed up to 7% strain. Strain
hardening (after the martensitic transformation finish point) was observed at around 5%. When the
applied strain reached 8%, the residual strain was only 0.7%, indicating the excellent strain recovery
of Ni-Ti-Co SMA. The sample fractured during the 8th cycle when the applied strain was 8% (as
marked by the red cross). According to the parameter definitions shown in Chapter 4 (Fig. 4-7), the
Young’s modulus and yield strength of Ni-Ti-Co SMA at 23 ℃ were 40 GPa and 631 MPa,
respectively. The variation of residual strain with respect to applied strain is shown in Fig. 8-3 (b). It
is seen that when the applied strain was smaller than 7%, the Ni-Ti-Co sample exhibited negative
residual strain. The negative residual strain indicates the sample shortened during unloading, which
may be caused by the release of residual stress generated when manufacturing or heat treating the alloy.
More research is needed to further understand this phenomenon.

192
(a) (b)
Fig. 8-3 Results of 1% strain incremental cyclic loading until failure on Ni-Ti-Co SMA at
23 ℃: (a) stress-strain curve, and (b) residual strain versus maximum applied strain.
The stress-strain curves of Ni-Ti-Co SMA under different temperatures are shown in Fig. 8-4.
From high to low, the temperatures tested were 50 ℃, 40 ℃, room temperature 23 ℃, 0 ℃, -20 ℃, -
40 ℃ and -60 ℃. It is noted that results in Fig. 8-4 were obtained using the same Ni-Ti-Co sample.
That is, after testing at one target temperature, the same sample was tested again at the other
temperatures. During the tests at 50 ℃ and -60 ℃, as shown in Fig. 8-4 (a) and (g), the incremental
cyclic loading was manually terminated when a large residual strain was observed after unloading. The
circles shown in Fig. 8-4 (a) and (g) indicate the manual termination.
From Fig. 8-4, it is seen that Ni-Ti-Co SMA showed ideal flag-shape stress-strain curves at 23 ℃,
0 ℃, -20 ℃ and -40 ℃. There was almost no residual strain up to a cyclic loading amplitude of 5%,
indicating the excellent strain recovery of Ni-Ti-Co SMA. At -60 ℃, the strain recovery of Ni-Ti-Co
SMA disappeared almost completely. Regarding high temperatures, at 40 ℃, the stress-strain curves
were no longer flag-shaped: the austenitic transformation finish point disappeared and the shape of

193
unloading curve became nearly linear. However, it is noted that the residual strain at 40 ℃ was still
small: after unloading from 5%, the residual strain was less than 0.2%. When the temperature increased
to 50 ℃, the residual strain started accumulating when unloading from 3%; after unloading from 4%,
the residual strain was around 0.5%. Overall, the yield strength, i.e., martensitic transformation start
stress, of Ni-Ti-Co SMA decreased as the temperature decreased. The relationship between yield
strength and temperature of Ni-Ti-Co SMA will be discussed in detail and compared with other SMAs
in Section 8.3.4.
(a) (b)
(c) (d)

194
(e) (f)
(g)
Fig. 8-4 Stress-strain curves of Ni-Ti-Co SMA at different temperatures: (a) 50 ℃, (b) 40 ℃,
(c) 23 ℃, (d) 0 ℃, (e) -20 ℃, (f) -40 ℃, and (g) -60 ℃.
8.3.2 Ni-Ti SMA
The results of 1% strain incremental cyclic loading until failure on Ni-Ti SMA at 23 ℃ are shown
in Fig. 8-5 (a). Similar to the results of Ni-Ti-Co SMA shown in Fig. 8-3 (a), Ni-Ti SMA showed ideal
flag-shaped stress-strain curves up to 7% strain. The Young’s modulus and yield strength of Ni-Ti
SMA at 23 ℃ were lower than that of Ni-Ti-Co SMA, which were 35 GPa and 396 MPa, respectively.
The residual strain versus maximum applied strain of Ni-Ti SMA is shown in Fig. 8-5 (b). It is seen
that before the strain reached 7%, the Ni-Ti SMA also showed negative residual strain, which was

195
similar to that of Ni-Ti-Co SMA. When unloading from 8%, the residual strain of Ni-Ti SMA was
around 0.4%. The Ni-Ti SMA fractured when the strain reached 8.5%.
(a) (b)
Fig. 8-5 Results of 1% strain incremental cyclic loading until failure on Ni-Ti SMA at 23
℃: (a) stress-strain curve, (b) residual strain versus maximum applied strain.
The stress-strain curves of Ni-Ti SMA under different temperatures are shown in Fig. 8-6. From
high to low, the temperatures tested were 50 ℃, 23 ℃, 0 ℃, and -20 ℃. Same as the Ni-Ti-Co SMA
results shown in Fig. 8-4, the results of Ni-Ti SMA shown in Fig. 8-6 were also obtained using the
same sample. At -20 ℃, the test was terminated manually when the strain reached 2%, as marked by
the circle. From Fig. 8-6 (a) to (c), it is seen that Ni-Ti SMA showed flag-shaped hysteresis loopsfrom
50 ℃ to 0 ℃. It is noted that, at 50 ℃ and 23 ℃, the sample showed remarkable shortening after
unloading: the residual strain after unloading from 5% reached -0.8%. When temperature decreased to
0 ℃, residual strain accumulation was observed but the negative residual strain disappeared. At -20 ℃,
the Ni-Ti SMA exhibited no flag-shaped hysteresis loop and the strain recovery disappeared almost
completely.

196
(a) (b)
(c) (d)
Fig. 8-6 Stress-strain curves of Ni-Ti SMA at different temperatures: (a) 50 ℃, (b) 23 ℃, (c) 0
℃, and (d) -20 ℃.
8.3.3 Cu-Al-Mn SMA
The results of 1% strain incremental cyclic loading until failure on Cu-Al-Mn SMA at 25 ℃ are
shown in Fig. 8-7. It is noted that this test was performed under crosshead displacement control,
which is different from the test methods used on Ni-Ti-Co (Fig. 8-3) and Ni-Ti (Fig. 8-5) SMAs.
In Fig. 8-7 (a), the stress-strain curve at 5% strain was shown in thick red line. The residual strain
versus maximum applied strain is shown in Fig. 8-7 (b). The Cu-Al-Mn SMA showed residual strain

197
accumulation since around 7.1% strain and fractured at 11.7% strain. Assuming a residual strain of
0.25% as the threshold, the maximum recovery strain of Cu-Al-Mn SMA can be taken as 7.6%. The
stress-strain curve at the maximum recovery strain is shown in thick black line in Fig. 8-7 (a).
(a) (b)
Fig. 8-7 Result of 1% strain incremental cyclic loading until failure on Cu-Al-Mn SMA at
25 ℃: (a) stress-strain curve, and (b) residual strain versus applied strain.
The stress-strain curves of Cu-Al-Mn SMA under different temperatures are shown in Fig. 8-8.
It is known that the superelastic behavior of Cu-Al-Mn SMA is sensitive to its crystalline orientations
[11,154]. To eliminate the influence of crystalline orientation and obtain a better understanding on the
temperature effect, results of each subfigure in Fig. 8-8 were obtained from Cu-Al-Mn samples with
the same crystalline orientation.
From Fig. 8-8 (a) to (c), it is seen that Cu-Al-Mn SMA showed ideal flag-shaped hysteresis loops
with almost no residual strain at 50 ℃, 23 ℃, and -40 ℃, indicating its superelasticity under a wide
range of temperatures. In Fig. 8-8 (d), the Cu-Al-Mn SMA lost strain recovery significantly at -40 ℃
and a large residual strain accumulation was observed. This might be caused by the unique crystal

198
orientation of this sample, because none of the other samples lost strain recovery at -40 ℃. However,
the influence of crystal orientation on the temperature dependence of superelasticity of Cu-Al-Mn
SMA has never been reported in the literature. More detailed research is needed to further understand
this issue. In Fig. 8-8 (e), the Cu-Al-Mn SMA was loaded to 7% strain and then unloaded (i.e., LP-4)
at 50 ℃, 23 ℃, and -40 ℃, respectively. It is seen that Cu-Al-Mn SMA showed 100% strain recovery
up to 7% at all three temperatures.
(a) (b)
(c) (d)

199
(e)
Fig. 8-8 Stress-strain curves of Cu-Al-Mn SMA at different temperatures.
8.3.4 Comparison of Ni-Ti-Co, Ni-Ti and Cu-Al-Mn SMA
Based on the above test results, the superelastic limit, i.e., maximum recovery strain, and
temperature dependence of Ni-Ti-Co, Ni-Ti and Cu-Al-Mn SMAs were extracted and compared.
Regarding the superelastic limit, it is seen from Fig. 8-3 and Fig. 8-5 that, for Ni-Ti-Co and NiTi SMAs, there was no residual strain accumulation when the applied strain was up to 7% (both of
them fractured during the 8th cycle when strain was about to reach or slightly exceeded 8%). Therefore,
7% strain can be taken as the superelastic limit of Ni-Ti-Co and Ni-Ti SMA. For Cu-Al-Mn SMA
shown in Fig. 8-7, its superelastic limit was taken as 7.6% (when the residual strain reached 0.25%).
Therefore, it can be concluded that the superelastic limit of Ni-Ti-Co SMA is close to that of Ni-Ti
SMA, and slightly smaller than that of Cu-Al-Mn SMA.
Regarding the temperature dependence, the relationship between temperature and yield
strength of Ni-Ti-Co, Ni-Ti, and Cu-Al-Mn SMAs is compared in Fig. 8-9. The data points in Fig.

200
8-9 were extracted from the abovementioned stress-strain curves: Ni-Ti-Co from Fig. 8-4, Ni-Ti
from Fig. 8-6, and Cu-Al-Mn from Fig. 8-8 (b). A linear fitting was performed in Fig. 8-9 to better
illustrate the trend. According to Clausius–Clapeyron relationship [41], the martensitic start
transformation stress (i.e., the yield stress) of SMA is proportional to the ambient temperature.
Assuming the material properties (such as entropy difference and molar volume) are constant, the
lower the temperature, the lower the yield stress will be. From Fig. 8-9, it is seen that the yield
strength of Ni-Ti-Co SMA showed larger scatter but it overall followed a linear relationship with
the temperature. The slope of Ni-Ti-Co SMA was close to that of Ni-Ti SMA. The yield strength
of Cu-Al-Mn SMA was perfectly linear to temperature and it had the smallest slope. The slope of
the linear fitting for Ni-Ti-Co, Ni-Ti and Cu-Al-Mn SMAs was 6.5 MPa/℃, 6.8 MPa/℃, and 2.0
MPa/℃, respectively. This indicates that the temperature dependence of Ni-Ti-Co SMA was close
to that of Ni-Ti SMA but around 3.3 times that of Cu-Al-Mn SMA.
Fig. 8-9 Relationship between yield stress and temperature of Ni-Ti-Co, Ni-Ti, and Cu-Al-Mn
SMAs.
0
200
400
600
800
-60 -40 -20 0 20 40 60
Yield stress (MPa)
Temperature (℃)
Ni-Ti-Co
Cu-Al-Mn
Ni-Ti

201
8.4 Summary of findings
It was found that, at room temperature 23 ℃, Ni-Ti-Co SMA exhibited excellent
superelasticity in terms of flag-shaped stress-strain curves and strain recovery capacity. Compared
with Ni-Ti and Cu-Al-Mn SMAs, the maximum recovery strain of Ni-Ti-Co SMA (around 7%)
was close to that of Ni-Ti SMA (around 7%) and smaller than that of Cu-Al-Mn SMA (around
7.6%). The yield stress of Ni-Ti-Co SMA was around 2.6 times that of Ni-Ti SMA and 3.3 times
that of Cu-Al-Mn SMA. The fracture strain of Ni-Ti-Co SMA (8%) was lower than that of Ni-Ti
SMA (8.5%) and Cu-Al-Mn SMA (12.7%).
Under different ambient temperatures, Ni-Ti-Co SMA exhibited ideal flag-shaped stress-strain
curves at 40 ℃, 23 ℃, 0 ℃, -20 ℃ and -40 ℃. There was almost no residual strain when the
incremental cyclic loading reached 5%, indicating its excellent strain recovery capacity and wide
application temperatures. Ni-Ti SMA lost superelasticity when the temperature dropped to 0 ℃,
and Cu-Al-Mn SMA showed stable superelasticity in the range of -40 ℃ to 50 ℃. Therefore, it can
be concluded that the temperature range of Ni-Ti-Co SMA is close to that of Cu-Al-Mn SMA and
wider than that of Ni-Ti SMA. In the yield stress versus temperature diagram, the slope of the linear
fitting for Ni-Ti-Co, Ni-Ti and Cu-Al-Mn SMAs was 6.5 MPa/℃, 6.8 MPa/℃, and 2.0 MPa/℃,
respectively. The dependence of yield stress on temperature of Ni-Ti-Co SMA was close to that of
Ni-Ti SMA but was 3.3 times that of Cu-Al-Mn SMA.

202
Chapter 9 - Low-cycle fatigue behavior of Ni-Ti-Co SMA at
different temperatures
The low-cycle fatigue behavior of Ni-Ti-Co SMA was investigated and compared with Ni-Ti
SMA at room temperature 23 ℃, low temperature -40 ℃, 0 ℃, and high temperature 50 ℃. Lowcycle fatigue loading with a constant strain amplitude of 5% was performed and the effect of
fatigue loading on the key superelastic properties was evaluated. The stress-strain curves, Young’s
modulus, yield stress, damping ratio, and recovery strain of Ni-Ti-Co SMA were extracted and
compared with Ni-Ti SMA.
9.1 Research motivation
From Chapter 8, it is seen that Ni-Ti-Co SMA has high strength and excellent superelasticity
in a wide temperature range, both of which are advantageous for bridge applications. As mentioned
previously in Chapter 1 and Chapter 4, bridges reinforced with SMA bars are intended to remain
operational after an earthquake, and it is important that the SMA bars have low-cycle fatigue
stability. Therefore, it is necessary to understand how the mechanical properties of Ni-Ti-Co SMA
degrade with increasing number of cyclic loadings before adopting it in bridges. However, due to
the short history, the low-cycle fatigue behavior of Ni-Ti-Co SMA has never been reported. In addition,

203
the temperature dependence of its low-cycle fatigue resistance is still unknown. To fill theses
knowledge gaps, this study investigated the low-cycle fatigue behavior of Ni-Ti-Co SMA under
different temperatures. Comparisons with Ni-Ti SMA was also made.
9.2 Materials and methods
9.2.1 Materials
Materials tested in this study were Ni-Ti-Co and Ni-Ti SMAs. The supplier and sample
machining method of these two materials used in this study were the same as previously provided
in Chapter 8. Cylinder dog-bone samples with the same dimensions were machined for both NiTi-Co and Ni-Ti SMA. The dimensions of the specimens are shown in Fig. 9-1.
Fig. 9-1 Dimensions of Ni-Ti-Co and Ni-Ti SMA samples used in low-cycle fatigue tests.
9.2.2 Methods
The set up used in this study was the same as that used in Chapter 8 (see Fig. 8-2). An MTS
370.5 dynamic servo-hydraulic frame was used to apply the load. An MTS 651.06E−04
environmental chamber was used along with the MTS load frame to house the specimens during
testing. Two extension rods were used to connect the sample to hydraulic griping systems. A 50.8
1 8 12.7
35 1 2 5 6
R=45
150
12 3 5
M18-1.5 thread 12.7 M18-1.5 thread
30 1 2 5 6
R=45
140
12 30
12.7
Note: All dimensions are in mm
M18-1.5 thread M18-1.5 thread
(a)
(b)
1 8 12.7

204
mm gauge length Epsilon extensometer (model number 3542-0200-100-LHT) was used to
measure the strain. A BMS16HR-53 Mars Labs data acquisition (DAQ) system was used to record
the data. A liquid nitrogen tank was used for low temperature tests.
A summary of low-cycle fatigue tests on Ni-Ti-Co and Ni-Ti SMA is provided in Table 9-1.
Three temperatures were considered for Ni-Ti-Co SMA, namely room temperature 23 ℃, high
temperature 50 ℃, and low temperature -40 ℃. Whereas for Ni-Ti SMA, it is seen in Chapter 8
that the Ni-Ti SMA shows flag-shape hysteresis loop from 50 ℃ to 0 ℃, and when the temperature
drops down to -20 ℃, it loses superelasticity completely. Therefore, the low temperature
considered for Ni-Ti SMA was 0 ℃.
The following loading protocols were used in the low-cycle fatigue tests. Cyclic loading with
a constant strain amplitude of 5% was performed on the sample until failure. For each cycle, the
loading was performed under extensometer control until 5% strain; then the sample was unloaded
to near-zero force under force control. The loading rate was 0.1 mm/min. For NiTiCo-1, NiTi-1
and NiTi-3, 1% strain incremental cyclic loading up to 5%, i.e., so called training, was performed
before low-cycle fatigue loading to investigate the effect of training. Training is a commonly used
way to stabilize the martensitic transformation of SMA [41,162]. For NiTi-4 at 0 ℃, the test was
manually terminated at 800 cycles when a stabilized stress-strain response was observed.

205
After obtaining the stress-strain curves, eight mechanical properties were extracted and
analyzed to quantify the effect of low-cycle fatigue and temperature on the superelasticity of NiTi-Co and Ni-Ti SMAs. The mechanical properties considered in this study were consistent with
those considered in Chapter 4. They are: elastic modulus, Eload, elastic modulus after yielding,
Eload2, maximum stress, max, yielding stress, y, damping ratio, R, maximum strain, max, residual
strain, resi, and recovery strain, reco. The definition of these parameters can be found in Fig. 4-7.
9.3 Results and discussion
9.3.1 Ni-Ti-Co and Ni-Ti at room temperature, 23 ℃
The stress-strain curves of Ni-Ti-Co SMA samples at room temperature, 23 ℃ are shown in Fig.
9-2. The material exhibited a flag-shape stress-strain curve during training or first fatigue cycle,
however, the width of hysteresis loops and the yield strength degraded rapidly within five cycles.
Comparing the training result with cycle-1 result in Fig. 9-2 (a), it is seen that after training, the yield
Table 9-1 Summary of low-cycle fatigue tests on Ni-Ti-Co SMA and Ni-Ti SMAs.
Material Temperature
Sample
label
Loading protocol
Fatigue cycle
numbers
Ni-Ti-Co
SMA
Room temperature
(23 ℃)
NiTiCo-1 Training + Fatigue 44
NiTiCo-2 Fatigue 91
High temperature (50 ℃) NiTiCo-3 Fatigue 48
Low temperature (-40 ℃) NiTiCo-4 Fatigue 471
Ni-Ti
SMA
Room temperature
(23 ℃)
NiTi-1 Training + Fatigue 44
NiTi-2 Fatigue 92
High temperature (50 ℃) NiTi-3 Training + Fatigue 48
Low temperature (0 ℃) NiTi-4 Fatigue 800

206
strength and hysteresis loop width of Ni-Ti-Co SMA decreased significantly. Therefore, training is not
recommended for Ni-Ti-Co SMA when applied in bridges. From fifth to 10th cycles, the yield strength
showed a continuous decrease and the width of hysteresis loop narrowed gradually. After 10th cycles,
the stress-strain curves of Ni-Ti-Co SMA stabilized until fracture. The NiTiCo-1 and NiTiCo-2
fractured at 44 and 92 cycles, respectively. It is noted that, even though the yield strength and width of
hysteresis loop of Ni-Ti-Co SMA showed a rapid reduction and fractured within 100 cycles, its strain
recovery capacity was maintained up to failure. Almost no residual strain accumulation was observed
during the low-cycle fatigue loading of Ni-Ti-Co SMA at 23 ℃.
(a)
(b)
Fig. 9-2 Stress-strain curves of NiTiCo SMA at room temperature, 23 °C: (a) NiTiCo-1, and
(b) NiTiCo-2.
The stress-strain curves of Ni-Ti SMA samples at room temperature, 23 ℃ are shown in Fig. 9-3.
Compared with the results of Ni-Ti-Co SMA shown in Fig. 9-2, Ni-Ti SMA showed a similar

207
superelasticity degradation trend during fatigue loading. The difference is that the yield strength and
hysteresis loop width of Ni-Ti SMA decreased slower than that of Ni-Ti-Co SMA; furthermore, the
yield strength of Ni-Ti SMA did not show much reduction after training. From first to 10th cycles, the
stress-strain curves of Ni-Ti SMA showed slightly degradation and the stabilization was reached at
20th cycles. The fatigue life of Ni-Ti SMA was slightly higher than that of Ni-Ti-Co SMA. The two
tested samples NiTi-1 and NiTi-2 fractured at 152 and 104 cycles, respectively. It is worth noting that
as the fatigue cycle number increases, the maximum stress of Ni-Ti SMA at 5% showed a remarkable
increase. This phenomenon may be caused by the strain hardening effect of Ni-Ti SMA and did not
occur when testing Ni-Ti-Co SMA.
(a)
(b)
Fig. 9-3 Stress-strain curves of Ni-Ti SMA at room temperature, 23 °C: (a) NiTi-1, and (b)
NiTi-2.
The extracted parameters of Ni-Ti-Co and Ni-Ti SMA with respect to loading cycles at 23 °C are
shown in Fig. 9-4 and Fig. 9-5, respectively. Comparing the extracted parameters of Ni-Ti-Co with

208
Ni-Ti, it is seen that Ni-Ti-Co SMA showed an overall higher degradation rate than Ni-Ti SMA
(average of the two tested samples of each material was used to make comparisons). Take cycle 10
as an example, Eload, y, and R of Ni-Ti-Co SMA were respectively 75%, 69% and 62% of those
during the first cycle; while for Ni-Ti SMA, its Eload showed almost no decrease, and its y and R
were respectively 80% and 89% of those during the first cycle. At the last cycle, the y and R of
Ni-Ti-Co SMA were respectively 53% and 52% of those during the initial cycle. Whereas for NiTi SMA, its y and R were respectively 36% and 64% of those during the first cycle. The resi of
both Ni-Ti-Co and Ni-Ti SMA remained near-zero throughout the low-cycle fatigue tests. It is
noted that Ni-Ti SMA exhibited a higher negative strain during fatigue loading, which led to a
slight increase of reco. In the last cycle, resi and reco of Ni-Ti-Co were -0.24% and 5.2%,
respectively; while the resi and reco of Ni-Ti were -0.56% and 5.5%, respectively.
(a)

209
(b)
Fig. 9-4 Variation in mechanical properties of Ni-Ti-Co SMA at room temperature, 23 °C:
(a) NiTiCo-1, and (b) NiTiCo-2.
(a)

210
(b)
Fig. 9-5 Variation in mechanical properties of Ni-Ti SMA at room temperature, 23 °C: (a)
NiTi-1, and (b) NiTi-2.
9.3.2 Ni-Ti-Co at -40 ℃ and Ni-Ti at -0 ℃
The stress-strain curves of Ni-Ti-Co SMA at low temperature -40 °C are shown in Fig. 9-6.
Compared with the results at room temperature, it is seen that Ni-Ti-Co SMA showed a much wider
hysteresis loop width at -40 °C. Furthermore, it is worth noting that the stress-strain curves exhibit
minor degradation up to 100 cycles (although the yield stress decreased slightly, the narrowing of the
hysteresis loop was negligible). From 150 to 471 cycles, at which fracture occurred, the hysteresis loop
width narrowed down slightly and the residual strain accumulated gradually. Overall, compared with
room temperature behavior, Ni-Ti-Co SMA showed improved low-cycle fatigue resistance in terms
of both fatigue life and stress-strain hysteresis curves at -40 °C.

211
Fig. 9-6 Stress-strain curves of Ni-Ti-Co SMA at -40 °C.
The stress-strain curves of Ni-Ti SMA at low temperature 0 °C are shown in Fig. 9-7. Different
from the Ni-Ti-Co SMA shown in Fig. 9-6, the hysteresis loop of Ni-Ti SMA showed significant
narrowing and the yield stress of Ni-Ti SMA decreased more significantly in the first 100 cycles. From
100 to 800 cycles, at which the test was terminated, the Ni-Ti SMA response was stabilized and the
shape of hysteresis loop exhibited little degradation.
Fig. 9-7 Stress-strain curves of Ni-Ti SMA at 0 °C.

212
The extracted parameters of Ni-Ti-Co and Ni-Ti SMA with respect to loading cycles at -40
°C and 0 °C are shown in Fig. 9-7 and Fig. 9-8, respectively. It is seen that the superelasticity
degradation of Ni-Ti-Co SMA at 40 °C was slower than that of Ni-Ti SMA at 0 °C. For example,
at 100 cycles, Eload, y, and R of Ni-Ti-Co SMA were respectively 94%, 83% and 90% of those
during the first cycle; reco of Ni-Ti-Co SMA showed almost no decrease. While for Ni-Ti SMA,
Eload, y, and R at 100 cycles were respectively 77%, 51% and 36% of those during the first cycle;
resi up to 2.3% was observed and the reco of Ni-Ti SMA at 100 cycles was 53% of that in the first
cycle. In the last cycle, Eload, y, and R of Ni-Ti-Co SMA were respectively 78%, 70% and 77%
of those during the first cycle; while for Ni-Ti SMA, Eload, y, and R were respectively 75%, 39%
and 23% of those in the first cycle. The reco of Ni-Ti-Co SMA decreased by 12% when it fractured,
while the same value of Ni-Ti SMA was 45%.
Fig. 9-8 Variation in mechanical properties of Ni-Ti-Co SMA at -40 °C.

213
Fig. 9-9 Variation in mechanical properties of Ni-Ti SMA at 0 °C.
9.3.3 Ni-Ti-Co and Ni-Ti at 50 ℃
The stress-strain curves of Ni-Ti-Co SMA at 50 °C are shown in Fig. 9-10. The stress-strain curve
of Ni-Ti-Co SMA is not ideally flag-shaped even in the first cycle because the austenitic transformation
finish point disappeared during unloading and a residual strain of approximately 0.8% was observed.
Up to fifth cycle, large reduction in the area of hysteresis loops was observed, and the martensitic
transformation start point also disappeared. From fifth to 48th, at which the bar fractured, the hysteresis
loop showed almost no change, indicating the Ni-Ti-Co SMA wasstabilized condition after five cycles
of fatigue loading. Compared with the results at 23 °C and -40 °C, Ni-Ti-Co SMA showed faster
stabilization, more significant hysteresis narrowing, and a larger residual strain accumulation at 50
°C.

214
Fig. 9-10 Stress-strain curves of Ni-Ti-Co SMA at 50 °C.
The stress-strain curves of Ni-Ti SMA at 50 °C are shown in Fig. 9-11. Similar to Ni-Ti-Co
SMA, the Ni-Ti SMA exhibited rapid stabilization and reduction in the area of the hysteresis curves
within the first five cycles. From fifth to 71st cycles, at which fracture occurred, the hysteresis loop
width narrowed gradually. The difference from Ni-Ti-Co SMA was that Ni-Ti SMA had less
residual strain accumulation, and the maximum stress when loading to 5% showed a more
significant increase with respect to the first cycle.
Fig. 9-11 Stress-strain curves of Ni-Ti SMA at 50 °C.
The extracted parameters of Ni-Ti-Co and Ni-Ti SMA with respect to loading cycles at 50 °C
are shown in Fig. 9-12 and Fig. 9-13, respectively. The Ni-Ti-Co SMA overall had more significant
superelasticity degradation in terms of energy dissipation and strain recovery than Ni-Ti SMA.
Take the 10th for example, Eload, y, R and reco of Ni-Ti-Co SMA were respectively 54%, 78%,
31% and 87% of those during the first cycle; while Eload, y, R and reco of Ni-Ti-Co SMA were
respectively 53%, 72%, 67% and 97% of those during the first cycle. In the last cycle, Eload, y,

215
and R of Ni-Ti-Co SMA were respectively 47%, 69% and 29% of those during the first cycle,
whereas those values for Ni-Ti SMA were 41%, 62% and 53%, respectively. The reco of Ni-Ti
SMA decreased by 8% when it fractured, while the value for Ni-Ti SMA was less than 1%.
Fig. 9-12 Variation in mechanical properties of Ni-Ti-Co SMA at 50 °C.

216
Fig. 9-13 Variation in mechanical properties of Ni-Ti SMA at 50 °C.
9.4 Summary of findings
It was found that at room temperature, 23 ℃, Ni-Ti-Co SMA exhibited similar low-cycle fatigue
resistance to Ni-Ti SMA in terms of fatigue life and degradation of superelasticity. Both materials
fractured after about 100 cycles of 5% constant strain fatigue loading, and neither materials had
appreciable residual strain accumulation before fracture, meaning that the strain recovery capacity
throughout the fatigue loading was stable. It is noted that the yield strength and energy dissipation
capacity of Ni-Ti-Co SMA decreased faster than those of Ni-Ti SMA. Due to the rapid degradation of
yield strength during the first few fatigue loading cycles, the training at room temperature, which is
commonly used to stabilize the martensitic transformation of SMA, is not recommended for Ni-Ti-Co
SMA.

217
At low temperature, Ni-Ti-Co SMA showed excellent fatigue resistance when temperature was
reduced to -40 ℃. Compared to Ni-Ti SMA at 0 ℃, the yield strength, energy dissipation and strain
recovery of Ni-Ti-Co SMA showed much slower degradation. There was no loss of yield strength,
energy dissipation, and strain recovery of Ni-Ti-Co SMA during the first 100 cycles of fatigue loading.
From 100 to 471 cycles, at which fracture occurred, the strain recovery of Ni-Ti-Co SMA showed
almost no degradation.
At 50 ℃, the low-cycle fatigue resistance of Ni-Ti-Co SMA was lower than that at 23 ℃ and
lower than that of Ni-Ti SMA at 50 ℃. The hysteresis loops of Ni-Ti-Co at 50 ℃ narrowed quickly
(within five cycles), leading to a disappearance of the martensitic transformation start and austenitic
transformation finish points, as well as a significant decrease in the energy dissipation capacity. It is
noted that the strain recovery of Ni-Ti-Co at 50 ℃ had no degradation during the fatigue loading.
In summary, Ni-Ti-Co SMA has comparable low-cycle fatigue resistance to Ni-Ti SMA at room
temperature 23 ℃ and high temperature 50 ℃. While at low temperature, Ni-Ti-Co SMA showed
much better low-cycle fatigue resistance than Ni-Ti SMA. These characteristics indicate the feasibility
of applying Ni-Ti-Co SMA in bridges subjected to extreme temperatures, particularly low
temperatures.

218
Chapter 10 - Moment-curvature analysis of bridge columns
reinforced with Ni-Ti-Co SMA
Moment-curvature analyses using Open System for Earthquake Engineering Simulation
(OpenSees) were performed to investigate the flexural behavior of Ni-Ti-Co SMA reinforced
sections for possible implementation in typical bridge columns. For comparison purposes,
columns reinforced with Ni-Ti SMA and Cu-Al-Mn SMA bars were also investigated. Conventional
reinforced concrete (RC) bridge columns were analyzed as benchmarks to determine the reference
plastic moments. The influence of key parameters: column section diameter, longitudinal
reinforcement ratio, and axial force ratio was investigated.
10.1 Research motivation
From Chapter 8, it is seen that Ni-Ti-Co SMA shows much higher yield strength than that of
Ni-Ti SMA (1.6 times) and Cu-Al-Mn SMA (2.3 times). Furthermore, the availability in large
sizes of Ni-Ti-Co SMA (over 32 mm diameter) is superior to Ni-Ti or Cu-Al-Mn SMAs. Both the
high strength and availability in large diameter of Ni-Ti-Co SMA make it potentially suitable for
bridge applications. However, no research has been performed on structural properties of bridge
columns reinforced with Ni-Ti-Co SMA.

219
To fill this knowledge gap and investigate the feasibility of applying Ni-Ti-Co SMA in bridge
columns, in this chapter, the moment-curvature response of typical bridge columns reinforced NiTi-Co SMA was studied. For comparison purposes, conventional reinforced concrete columns,
and columns reinforced with Ni-Ti SMA and Cu-Al-Mn SMA were also investigated when a same
flexural capacity was reached.
10.2 Modeling methods
10.2.1 RC columns
OpenSees [163] was used to perform the moment curvature analyses in this study. The
analysis of RC column sections includes three components, namely unconfined concrete cover,
longitudinal bars, and confined concrete core. The effect of the transverse reinforcement was
accounted for in the confined concrete properties. Concrete01 material was used to model the
unconfined concrete cover and Concrete04 material was used to model the confined concrete core.
Mander’s model [164] was used to determine the properties of both unconfined concrete cover and
confined concrete core. The tensile strength of concrete was ignored in all analyses.
ReinforcingSteel material was used to model the longitudinal steel bars. Constitutive models used
in RC columns are shown in Fig. 10-1.

220
(a) (b) (c)
Fig. 10-1 Constitutive models used in RC columns: (a) confined concrete, (b) unconfined
concrete, and (c) steel rebar.
A full-scale RC bridge column [165] with a section diameter D = 1.22 m (4 ft) was used as
the reference column to validate the OpenSees model. This column model represents the typical
single-column bridge bents commonly used in California and is designed according to Caltrans
Bridge Design Specifications and Seismic Design Criteria. The cross section of the reference
column is shown in Fig. 10-2 (a). Grade 60 steel was used with a longitudinal reinforcement ratio
of 1.55% and transverse reinforcement ratio of 0.95%. The axial load was 2530 kN, with an axial
force ratio of a = 5.3%. Normal weight concrete with a compressive strength of 40.3 MPa (5.8 ksi)
was used in the entire column. Additional information can be found in Schoettler et al. [165]. The
simulated result is compared with the idealized test result in Fig. 10-2 (b). It is seen that the result
from the OpenSees model matches well with the idealized test result.
(a) (b) (c)

221
(a) (b)
Fig. 10-2 (a) Section details of reference column, and (b) Validation of established OpenSees
model.
10.2.2 SMA columns
Engineered cementitious composite (ECC) is a fiber reinforced fine mortar often used in the
plastic hinge regionsto reduce and delay plastic hinge damage such as that due to concrete spalling.
Compared with conventional concrete, ECC has a high tensile strength and ductility. Multiple
cracks with a width of less than 100 m form under tensile loading giving the ECC ability to
deform significantly before failure. The advantages of ECC complement the strain recovery of
SMA and the combination of the two could help keep bridge columns in service even after strong
earthquakes. The moment curvature analyses conducted in this study used this combination when
the longitudinal steel reinforcement was replaced with SMA. The transverse reinforcement was
assumed to be mild steel in the SMA columns.
The SMA column section includes three components, namely unconfined ECC cover,
longitudinal SMA bars, and confined ECC core. Similar to the RC columns, the effect of transverse
D = 1.22 m (4 ft)
51 mm (2 in)
clear cover
18 pcs U.S. #11
( = 35.8 mm) rebar
Double U.S. #5
( = 15.9 mm) hoops
(a) (b)

222
reinforcement is implicit in the confined ECC properties. Both the unconfined ECC cover and
confined ECC core were modeled by Concrete02 material, using constitutive models developed
by Motaref et al. [166]. To be consistent with the modeling methods used in the RC column, the
tensile strength of ECC was also ignored. SelfCentering material was used to model the SMA bars,
i.e., Ni-Ti-Co, Ni-Ti and Cu-Al-Mn SMA. For brevity, they are referred to as NiTiCo, NiTi and
CAM SMA, respectively, hereafter.
Schematic diagram of the key parameters in SelfCentering material is shown in Fig. 10-3,
blue solid line. SMA bars were connected to the steel rebar that were embedded in the footing and
cap beam. The bond slip of the steel rebar was accounted for using the method developed by Tazarv
et al. [167]. The modified stress-strain curves of SMA are shown in Fig. 10-3, green dashed line.
The parameters used to model NiTiCo, NiTi and CAM SMA bars are listed in Table 10-1. The
stress-strain curves used to model SMA columns are shown in Fig. 10-4.
Fig. 10-3 Definition of key parameter used to model SMA bars.
Strain
Stress
0
fy
k1
k2
k3
L
Modified
Original

223
Fig. 10-4 Stress-strain curves used to model SMA columns: (a) CAM SMA, (b) NiTiCo
SMA, (c) NiTi SMA, (d) confined ECC, and (e) unconfined ECC.
(a) (b) (c)
(d) (e)
Table 10-1 Details of properties used to model SMA bars.
Properties NiTiCo NiTi CAM
Yield stress fy (MPa) - 631 396 273
Martensitic finish strain L (%) - 4.9 4.5 6.2
Austenite modulus k1 (GPa)
Original 40.0 35.0 65.0
Modified 24.2 22.2 46.2
Post-yield stiffness k2 (GPa)
Original 1.0 2.3 1.4
Modified 2.0 2.2 1.4
Post-hardening stiffness k3 (GPa)
Original 9.3 9.1 34.1
Modified 8.1 7.9 30.1

224
10.3 Methodology
The criterion to design SMA reinforced column sections was to match the plastic moment
with their corresponding reference RC column sections. It is realized that matching the plastic
moment ignores the fact that the relatively low SMA reinforced column stiffness could affect the
plastic moment demand in the bridge columns. Typical moment-curvature (M-Phi) diagrams of
RC and SMA columns are shown in Fig. 10-5. For the RC column shown in Fig. 10-5 (a), the
actual M-Phi curve (O-A-B) can be idealized with an elasto-plastic response (O-C-E) according to
AASHTO LRFD Bridge Design Specifications [168]. Point A in Fig. 10-5 (a) is at the first
longitudinal reinforcing bar yielding, and the idealized plastic moment at point C, Mp, is obtained
by equating the area between the actual and the idealized response beyond point A, i.e., the blue
shaded region (ACD) and the red shaded region (BDE). The ultimate curvature, u, is determined
when the compressive strain of concrete reaches 0.018, which is 1.5 times of the value in Mander’s
model [164], because as reported by Motaref et al. [166], the ultimate strain of concrete measured
from the column tests is significantly higher than the value computed by the Mander’s model [164].

225
(a) (b)
Fig. 10-5 Typical moment-curvature curves of: (a) RC column, and (b) SMA column.
The shape of the M-Phi response for SMA reinforced column sections is generally very
different than that of RC section as seen in Fig. 10-5. Therefore, the idealization method in LRFD
Bridge Design Specifications [168] was modified as follows. First, the ultimate curvature, u, is
reached when the compressive strain at the edge of the ECC fiber reaches 0.03, which is 1.5 times
of the value computed using the method in Motaref et al. [166]. Second, the elastic portion of the
idealized curve passes through the 0.5Mu, i.e., point F’ in Fig. 10-5 (b). The same approach was
also adopted by Pulido et al. [169] for concrete sections with properties significantly different than
that of RC. Third, to obtain the idealized plastic moment at point C’, Mp, the red shaded region in
O’F’ plus B’D’E’ was added up to equate the blue shaded region C’D’F’.
Using the above methods, the idealized plastic moment of RC column and SMA column was
obtained and compared for a range of parameters. For RC columns, nine conditions were
considered, including three section diameters (D = 4, 5 and 6 ft), three reinforcement ratios (r =
1%, 2% and 3%) and three axial force ratios (a = 5%, 10% and 15%). The control RC column had
Curvature
Moment
y yi
MP
My
u
O
A
C
D E
B
Curvature
y yi
MP
0.5Mu
u
Moment
O’
F’
C’
E’
B’
D’
A’
Mu

226
D = 5 ft, r = 2%, a = 10%. The steel rebar was 35.8 mm diameter (U.S. #11), with a yield strength
of 461MPs (67 ksi). Conventional concrete was used with a compressive strength of 40.3 MPa
(5.8 ksi). The clear cover was 51 mm (2 in). The test matrix of moment-curvature analyses is
shown in Fig. 10-6. One parameter was changed at a time in the moment-curvature analysis. It is
noted that only circular column section was considered in this study.
Fig. 10-6 Test matrix of moment-curvature analyses.
Three types of SMA reinforced columns were analyzed, namely NiTiCo, NiTi and CAM
SMA. The largest diameters of SMA bars used in past studies were assumed: 32 mm NiTiCo, 27
mm NiTi, and 30 mm CAM SMA bars. Concrete cover of all SMA reinforced columns was the
same as that of RC columns, but the number of longitudinal bars or the section diameter was
adjusted to match Mp of the corresponding RC sections.
Section
diameter
D (ft)
Reinforcem
ent ratio
r (%)
Axial froce
ratio
a (%) NiTiCo NiTi CAM
RC
RC-1 4
2 10
NiTiCo-1.x NiTi-1.x CAM-1.x
RC-2 5 NiTiCo-2.x NiTi-2.x CAM-2.x
RC-3 6 NiTiCo-3.x NiTi-3.x CAM-3.x
RC-4
5
1
10
NiTiCo-4.x NiTi-4.x CAM-4.x
RC-5 2 NiTiCo-5.x NiTi-5.x CAM-5.x
RC-6 3 NiTiCo-6.x NiTi-6.x CAM-6.x
RC-7
5 2
5 NiTiCo-7.x NiTi-7.x CAM-7.x
RC-8 10 NiTiCo-8.x NiTi-8.x CAM-8.x
RC-9 15 NiTiCo-9.x NiTi-9.x CAM-9.x
Note: The section diameter or number of longitudinal bars of NiTiCo, NiTi and CuAlMn SMA section
is iterated until its plastic moment matches the corresponding RC column. ‘-.x’ means no. x of iteration.

227
Matching of Mp required consideration of the difference between the SMA and RC sections.
From Fig. 10-5, it is seen that the SMA column has a different M-Phi response than the RC column.
Because the yield strength and Young’s modulus of SMA bars are different than those of steel bars,
the number of SMA bars was adjusted so that the SMA and RC section Mp values are the same.
However, there is an upper limit on the number of longitudinal bars due to the required minimum
spacing [168,170].
An iterative process was used to design the RC-equivalent SMA reinforced sections. The
number of SMA bars was increased first, when the clear spacing between the longitudinal bars
reached the upper limit and the moment capacity is still not sufficient, the section diameter of SMA
columns was increased with an increment of 152.4 mm (0.5 ft). The increased column diameter
led to a reduction of the axial force ratio a because the gravity load from the superstructures
remains the same. The increase in the column weight that slightly increases the load on column
section at the base was neglected. The iterations were repeated until the target Mp was reached.
It is worth noting that increasing the diameter will change the stiffness and weight of the
column, which is undesirable for practical applications because it will alter the seismic demand of
the bridge system. The scope of this research was to understand the M-Phi response of NiTiCo,
NiTi and CAM SMA reinforced bridge column sections, and compare their flexural behavior with
that of conventional RC columns. Further research is needed to address the dynamic response of
the SMA columns and the effect of their lower Young’s modulus and possible larger diameter.

228
10.4 Results and discussion
10.4.1 Sections with different diameters
M-Phi results of RC and SMA columns with different diameters, D, are shown in Fig. 10-7.
For ease of comparison, the plots in Fig. 10-7 were placed vertically. The diameter of RC columns,
varying from D = 1219 mm (4 ft), D = 1524 mm (5 ft) and D = 1829 mm (6 ft), were used as the
control variable, and the r= 2%, and a= 10%. The r and a of the SMA columns were iterated using
the abovementioned principles until the idealized plastic moment, Mp, is equivalent to the
corresponding RC columns.
The meaning of the annotations in each subfigure of Fig. 10-7 is explained as follows. Take
Fig. 10-7 (l) for instance: ‘NiTiCo-3.1’ means the first iteration of NiTiCo column corresponds to
RC-3; ‘#68, max’ means it has 68 longitudinal NiTiCo reinforcement bars, which is the maximum
number of bars that can be placed in the section considering the clearing space limit; ‘F’ means
axial force. Because the Mp of NiTiCo-3.1 is 27582 kN-m, less than the target ‘RC-3’ (Mp = 27516
kN-m), a second iteration (labeled as ‘NiTiCo-3.2’) was performed.
From Fig. 10-7 (a) to (c), it is seen that to match the Mp with RC-1 (D = 4 ft), the NiTi and
CAM columns with a diameter of 4 ft were not strong enough when the maximum number of
longitudinal reinforcements was used. The smaller bar diameter and lower yield strength of NiTi
and CAM SMA than steel rebar led to the lower Mp of NiTi and CAM SMA column. Therefore, a
second iteration with a larger section diameter was needed. When the section diameter was

229
increased to 4.5 ft, the Mp of NiTi and CAM columns was equivalent to RC column and the r was
1.8% and 2.3%, respectively for NiTi and CAM columns.
For the NiTiCo column shown in Fig. 10-7 (d), to match Mp with the RC column, the r of
NiTiCo columns was 2.2%, slightly higher than the RC (2%). The yield strength of NiTiCo SMA
is higher than steel rebar, but it is noted that the diameter of NiTiCo SMA bars is 3.8 mm smaller
than that of the steel rebar, the smaller diameter of SMA bars led to a larger number of bars required
for the same plastic moment, thus resulting in a higher r of NiTiCo columns. However, unlike NiTi
and CAM columns, the section diameter of NiTiCo column did not need to increase, indicating
the advantage of NiTiCo SMA over the other SMA bars.
(a) (e) (i)
(b) (f) (j)
RC-1
Idealized
RC-1: Mp=7986 kN-m
D=4ft, r=2% (#23), a=10% (F=4705kN)
RC-2: Mp=15809 kN-m
D=5ft, r=2% (#36), a=10% (F=7351kN)
RC-2
Idealized
RC-3: Mp=27516 kN-m
D=6ft, r=2% (#52), a=10% (F=10586kN)
RC-3
Idealized
NiTi-1.1
NiTi-1.2
NiTi-1.1:Mp=6986kN-m
D=4ft, r=2.6% (#52, max), a=10% (F=4705kN)
NiTi-1.2:Mp=8020kN-m
D=4.5ft, r=1.8% (#46), a=7.9% (F=4705kN)
NiTi-2.1:Mp=12196kN-m
D=5ft, r=2.1% (#66, max), a=10% (F=7351kN)
NiTi-2.2:Mp=15336kN-m
D=5.5ft, r=1.9% (#73, max), a=8.3% (F=7351kN)
NiTi-2.3:Mp=15864kN-m
D=6ft, r=1.3% (#59), a=6.9% (F=7351kN)
NiTi-2.1
NiTi-2.3 NiTi-2.2
NiTi-3.1:Mp=19275kN-m
D=6ft, r=1.7% (#80, max), a=10% (F=10586kN)
NiTi-3.2:Mp=23350kN-m
D=6.5ft, r=1.6% (#87, max), a=8.5% (F=10586kN)
NiTi-3.3:Mp=27608kN-m
D=7ft, r=1.5% (#93), a=7.3% (F=10586kN)
NiTi-3.1
NiTi-3.2
NiTi-3.3

230
(c) (g) (k)
(d) (h) (l)
Fig. 10-7 M-Phi results of RC and SMA columns with different section diameter, D: (a)
RC-1 with D = 4 ft, (c) to (d) NiTi, CAM and NiTiCo sections iterating to match RC-1; (e)
RC-2 with D = 5 ft, (f) to (h) NiTi, CAM and NiTiCo sections iterating to match RC-2; (i)
RC-3 with D = 6 ft, (j) to (l) NiTi, CAM and NiTiCo sections iterating to match RC-3.
When increasing the diameter of RC to 5 ft and 6 ft, as shown in Fig. 10-7 (e) to (l), the
overall trend in the SMA columns was similar to that shown in Fig. 10-7 (a) to (d): when NiTi and
CAM columns have the same D as RC, their Mp was smaller than RC columns when the maximum
number of longitudinal reinforcements was reached. The diameter in NiTi and CAM sections need
to increase by 1 ft or 1.5 ft to match the RC columns Mp. The r of NiTiCo column at D= 5 ft was
2.2%, which was slightly higher than RC-2 with the same Mp. The NiTiCo-3.2, with D= 6.5 ft and
r=1.5%, matched the Mp of RC-3, with D= 5 ft and r=2%.
CAM-1.1:Mp=6314 kN-m
D=4ft, r=2.8% (#47, max), a=10% (F=4705kN)
CAM-1.2:Mp=8064 kN-m
D=4.5ft, r=2.3% (#49), a=7.9% (F=4705kN)
CAM-1.1
CAM-1.2
CAM-2.1:Mp=11139 kN-m
D=5ft, r=2.3% (#60, max), a=10% (F=7351kN)
CAM-2.2:Mp=13789 kN-m
D=5.5ft, r=2.1% (#66, max), a=8.3% (F=7351kN)
CAM-2.3:Mp=15784 kN-m
D=6ft, r=1.7% (#64), a=6.9% (F=7351kN)
CAM-2.3
CAM-2.2
CAM-2.1 CAM-3.1
CAM-3.2
CAM-3.3
CAM-3.4
CAM-3.1:Mp=17628 kN-m
D=6ft, r=1.9% (#72, max), a=10% (F=10586kN)
CAM-3.2:Mp=25053 kN-m
D=7ft, r=1.7% (#85, max), a=7.3% (F=10586kN)
CAM-3.3:Mp=27498 kN-m
D=7.5ft, r=1.4% (#80), a=6.4% (F=10586kN)
CAM-3.1
CAM-3.2
CAM-3.3
NiTiCo-1.1:Mp=7958 kN-m
D=4ft, r=2.2% (#32), a=10% (F=4705kN)
NiTiCo-1.1
NiTiCo-2.1: Mp=15857 kN-m
D=5ft, r=2.2% (#51), a=10% (F=7351kN)
NiTiCo-2.1
NiTiCo-3.1: Mp=26221 kN-m
D=6ft, r=2.1% (#68, max), a=10% (F=10586kN)
NiTiCo-3.2
NiTiCo-3.1
NiTiCo-3.2: Mp=27582kN-m
D=6.5ft, r=1.5% (#59), a=8.5% (F=10586kN)

231
10.4.2 Sections with different reinforcement ratios
M-Phi results of RC and SMA columns with different reinforcement ratios, r, are shown in
Fig. 10-8. The r of RC columns was the control variable, varying from 1% to 2% and 3%, and the
D= 5 ft and a= 10%. In Fig. 10-8 (a) to (d), RC-4 has an r value of 1%. To match Mp of RC-4, r
had to be increased to 1.6%, 2.1% and 1.1% in NiTi, CAM and NiTiCo columns, respectively.
There was no need to increase the SMA column diameters.
In Fig. 10-8 (i) to (l), when the r in RC-4 was increased to 3%, D in NiTi, CAM and NiTiCo
columns needed to be increased to match the Mp. Specifically, D in NiTi, CAM and NiTiCo
columns needed to be 6.5 ft, 6.5 ft and 5.5 ft; the r was reduced to 1.4%, 1.9% and 2.2% in NiTi6.3, CAM-6.3 and NiTiCo-6.2, respectively. When the RC columns had a high r of 3%, SMA
columns needed to have a larger diameter to match the flexural capacity, even for NiTiCo which
had a higher yield strength than the steel bars. A similar trend was observed when the longitudinal
steel ratio was 2% in the RC section, although the required increases in the SMA section diameter
was less than of those matching the plastic moment in the RC column with 3% reinforcement ratio.
(a) (e) (i)
RC-4: Mp=10518 kN-m
D=5ft, r=1% (#18), a=10% (F=7351kN)
RC-4
Idealized
RC-5: Mp=15809 kN-m
D=5ft, r=2% (#36), a=10% (F=7351kN)
RC-5
Idealized
RC-6: Mp=20678 kN-m
D=5ft, r=3% (#54), a=10% (F=7351kN)
RC-6
Idealized

232
(b) (f) (j)
(c) (g) (k)
(d) (h) (l)
Fig. 10-8 M-Phi results of RC and SMA columns with different reinforcement ratio, r: (a)
RC-4 with r = 1%, (c) to (d) NiTi, CAM and NiTiCo sections iterating to match RC-4; (e)
RC-5 with r = 2%, (f) to (h) NiTi, CAM and NiTiCo sections iterating to match RC-5; (i)
RC-6 with r = 3%, (j) to (l) NiTi, CAM and NiTiCo sections iterating to match RC-6.
10.4.3 Sections with different axial force ratios
M-Phi results of RC and SMA columns with different axial force ratio, a, are shown in Fig.
10-9. The a for RC columns was the control variable, varying from 5% to 10% and 15%. The RC
section diameter was kept at 5 ft and the longitudinal reinforcement ratio was kept at 2%. Overall,
NiTi-4.1:Mp=10556 kN-m
D=5ft, r=1.6% (#51), a=10% (F=7351kN)
NiTi-4.1
NiTi-5.1:Mp=12196kN-m
D=5ft, r=2.1% (#66, max), a=10% (F=7351kN)
NiTi-5.2:Mp=15336kN-m
D=5.5ft, r=1.9% (#73, max), a=8.3% (F=7351kN)
NiTi-5.3:Mp=15864kN-m
D=6ft, r=1.3% (#59), a=6.9% (F=7351kN)
NiTi-5.1
NiTi-5.3 NiTi-5.2
NiTi-6.1:Mp=12196 kN-m
D=5ft, r=2.1% (#66, max), a=10% (F=7351kN)
NiTi-6.2:Mp=18856 kN-m
D=6ft, r=1.7% (#80, max), a=6.9% (F=7351kN)
NiTi-6.3:Mp=20722 kN-m
D=6.5ft, r=1.4% (#74), a=5.9% (F=7351kN)
NiTi-6.3 NiTi-6.2
NiTi-6.1
CAM-4.1:Mp=10532 kN-m
D=5ft, r=2.1% (#54), a=10% (F=7351kN)
CAM-4.1
CAM-5.1:Mp=11139 kN-m
D=5ft, r=2.3% (#60, max), a=10% (F=7351kN)
CAM-5.2:Mp=13789 kN-m
D=5.5ft, r=2.1% (#66, max), a=8.3% (F=7351kN)
CAM-5.3:Mp=15784 kN-m
D=6ft, r=1.7% (#64), a=6.9% (F=7351kN)
CAM-5.3
CAM-5.2
CAM-5.1
CAM-6.1:Mp=11134 kN-m
D=5ft, r=2.3% (#60, max), a=10% (F=7351kN)
CAM-6.2:Mp=16781 kN-m
D=6ft, r=1.9% (#72, max), a=6.9% (F=7351kN)
CAM-6.3:Mp=20636 kN-m
D=6.5ft, r=1.9% (#81), a=5.9% (F=7351kN)
CAM-6.1
CAM-6.2
CAM-6.3
NiTiCo-4.1:Mp=10562kN-m
D=5ft, r=1.1% (#26), a=10% (F=7351kN)
NiTiCo-4.1
NiTiCo-5.1: Mp=15857 kN-m
D=5ft, r=2.2% (#51), a=10% (F=7351kN)
NiTiCo-5.1
NiTiCo-6.1:Mp=17430kN-m
D=5ft, r=2.6% (#60,max), a=10% (F=7351kN)
NiTiCo-6.2:Mp=20704kN-m
D=5.5ft, r=2.2% (#60), a=8.3% (F=7351kN)
NiTiCo-6.2
NiTiCo-6.1

233
it is seen that as a increases, the Mp of RC columns increased and the ultimate curvature reduced,
a similar trend was also observed in NiTi, CAM and NiTiCo columns. This trend is well known.
From Fig. 10-9 (a) to (d), it is seen that to match the Mp of RC-7 with a= 5%, the D for NiTi
and CAM columns needed to be 0.5 ft and 1 ft larger, respectively, reducing r in NiTi-7.2 and
CAM-7.3 columns to 1.9%, and 1.7%, respectively. There was no need to increase D for NiTiCo7.1 to match the Mp of RC-7, but the reinforcement ratio was increased slightly to 2.1%. For RC9 with a= 15%, the D of NiTi and CAM columns needed to be 1 ft larger; the r of NiTi-9.3 and
CAM-7.3 columns was 1.3%, and 1.7%, respectively. Similar to NiTiCo-7.1, D in NiTiCo-8.1,
and NiTiCo-9.1 remained at 5 ft, but the reinforcement ratio changed to 2.2% and 2.5%,
respectively. The necessary diameter increases in NiTi and CAM sections under 10% axial force
ratio were the same as those under 15% ratio.
(a) (e) (i)
(b) (f) (j)
RC-7: Mp=14755 kN-m
D=5ft, r=2% (#36), a=5% (F=3676kN)
RC-7
Idealized
RC-8: Mp=15809 kN-m
D=5ft, r=2% (#36), a=10% (F=7351kN)
RC-8
Idealized
RC-9: Mp=16742 kN-m
D=5ft, r=2% (#36), a=15% (F=11027kN)
RC-9
Idealized
NiTi-7.1:Mp=11804 kN-m
D=5ft, r=2.1% (#66, max), a=5% (F=3676kN)
NiTi-7.2:Mp=14773 kN-m
D=5.5ft, r=1.9% (#73), a=4.1% (F=3676kN)
NiTi-7.1
NiTi-7.2
NiTi-8.1:Mp=12196kN-m
D=5ft, r=2.1% (#66, max), a=10% (F=7351kN)
NiTi-8.2:Mp=15336kN-m
D=5.5ft, r=1.9% (#73, max), a=8.3% (F=7351kN)
NiTi-8.3:Mp=15864kN-m
D=6ft, r=1.3% (#59), a=6.9% (F=7351kN)
NiTi-8.1
NiTi-8.3 NiTi-8.2
NiTi-9.1:Mp=12459 kN-m
D=5ft, r=2.1% (#66, max), a=15% (F=11027kN)
NiTi-9.2:Mp=15686 kN-m
D=5.5ft, r=1.9% (#73, max), a=12.4% (F=11027kN)
NiTi-9.3:Mp=16770 kN-m
D=6ft, r=1.3% (#61), a=10.4% (F=11027kN)
NiTi-9.1
NiTi-9.2
NiTi-9.3

234
(c) (g) (k)
(d) (h) (l)
Fig. 10-9 M-Phi results of RC and SMA columns with different axial force ratio, a: (a) RC-7
with a = 5%, (c) to (d) NiTi, CAM and NiTiCo sections iterating to match RC-7; (e) RC-8
with a = 10%, (f) to (h) NiTi, CAM and NiTiCo sections iterating to match RC-8; (i) RC-9
with a = 15%, (j) to (l) NiTi, CAM and NiTiCo sections iterating to match RC-8.
10.5 Summary of findings
The criterion to design SMA reinforced column sections that are equivalent to their
counterpart RC sections in this study was to match the plastic moment capacity. Since the yield
strength and the Young’s modulus of SMA bars are different compared to those of steel
reinforcement, in general the stiffness and plastic moment in SMA reinforced sections were lower
than those of reinforced concrete (RC) sections. To match the idealized plastic moment of SMA
columns with the RC columns, more SMA bars are needed that correspond to a higher longitudinal
reinforcement ratio. Depending on the available SMA bar sizes, the section diameter may have to
CAM-7.1:Mp=10350 kN-m
D=5ft, r=2.3% (#60, max), a=5% (F=3676kN)
CAM-7.2:Mp=12904 kN-m
D=5.5ft, r=2.1% (#66, max), a=4.1% (F=3676kN)
CAM-7.3:Mp=14794 kN-m
D=6ft, r=1.7% (#62), a=3.5% (F=3676kN)
CAM-7.1
CAM-7.2
CAM-7.3
CAM-8.1:Mp=11139 kN-m
D=5ft, r=2.3% (#60, max), a=10% (F=7351kN)
CAM-8.2:Mp=13789 kN-m
D=5.5ft, r=2.1% (#66, max), a=8.3% (F=7351kN)
CAM-8.3:Mp=15784 kN-m
D=6ft, r=1.7% (#64), a=6.9% (F=7351kN)
CAM-8.3
CAM-8.1
CAM-8.2
CAM-9.1:Mp=11908 kN-m
D=5ft, r=2.3% (#60, max), a=15% (F=11027kN)
CAM-9.2:Mp=14640 kN-m
D=5.5ft, r=2.1% (#66, max), a=12.4% (F=11027kN)
CAM-9.3:Mp=16750 kN-m
D=6ft, r=1.7% (#64), a=10.4% (F=11027kN)
CAM-9.3
CAM-9.1
CAM-9.2
NiTiCo-7.1:Mp=14794kN-m
D=5ft, r=2.1% (#47), a=5% (F=3676kN)
NiTiCo-7.1
NiTiCo-8.1: Mp=15857 kN-m
D=5ft, r=2.2% (#51), a=10% (F=7351kN)
NiTiCo-8.1
NiTiCo-9.1:Mp=16749kN-m
D=5ft, r=2.5% (#56), a=15% (F=11027kN)
NiTiCo-9.1

235
be increased to maintain sufficient bar spacing. These trends are more likely in Ni-Ti and Cu-AlMn columns.
The diameter of Ni-Ti-Co reinforced sections in most cases remained the same as the RC
section diameter due to the higher yield strength of Ni-Ti-Co SMA bars. For example, for the
control RC column with a section diameter of 5 ft, a reinforcement ratio of 2%, and an axial force
ratio of 10%, to match the plastic moment of Ni-Ti-Co column with the RC column, the
reinforcement ratio of Ni-Ti-Co column was 2.2%, slightly higher than that of the RC column
(2%). Only when the corresponding RC column had a very large diameter or a very high
reinforcement ratio, the Ni-Ti-Co column diameter had to be larger than the corresponding RC
column. Specifically, when the corresponding RC column had a large diameter of 6 ft or a high
reinforcement ratio of 3%, the Ni-Ti-Co column diameter had to be respectively 0.5 ft and 1 ft
larger than the corresponding RC column. In such cases, the reinforcement ratio in Ni-Ti-Co
column was 1.5% and 2.2%, respectively. Compared with Ni-Ti and Cu-Al-Mn columns, the
plastic moment of Ni-Ti-Co columns were much higher for the same column diameter and
longitudinal reinforcement ratio, leading to a much lower reinforcement ratio and smaller column
diameter. All these indicate the advantages of Ni-Ti-Co SMA and the potential of using it in real
bridge applications.

236
Chapter 11 - Effect of temperature on ductility and recovery strain
of Fe-Mn-Si SMA
The effect of temperature on the strength, ductility and recovery strain of Fe-Mn-Si SMA
before thermal stimulation, i.e., actuation, was investigated. Monotonic and incremental cyclic
loading tests were respectively performed at room temperature 23 ℃, -40 ℃, and 50 ℃. The
influence of temperature on the key mechanical properties of Fe-Mn-Si SMA, such as Young’s
modulus, yield strength, ductility, and recovery strain were extracted and analyzed.
11.2 Research motivation
As mentioned in Chapter 2, Fe-Mn-Si SMA has received increasing research attention for
post-tensioning concrete structures. In contrast to traditional approach using high strength steel
tendons, the post-tensioning with Fe-Mn-Si SMA is not performed through application of an
external load, but through their internal martensitic phase transformation via thermal stimulation,
also known as actuation. The advantage of using Fe-Mn-Si SMA is that the post-tensioning forces
can be generated with no friction losses. Furthermore, since no heavy hydraulic devices are
required, manpower and construction space can be saved. In addition to these advantages, Fe-MnSi SMA also has good machinability, weldability [62] and low cost compared with other

237
conventionally used post-tensioning SMA compositions such as Ni-Ti-Nb alloy [63,64], all of
which guarantee the wide application of Fe-Mn-Si SMA in bridge applications.
When applying Fe-Mn-Si SMA as post-tensioning elements in bridges, the material has to be
prestrained to a certain strain level prior to thermal stimulation. The prestrain level and variation
of ambient temperature could potentially affect the strength, ductility and prestressing behavior of
Fe-Mn-Si SMA after activation. However, the effect of temperature on these mechanical properties
has never been investigated. To fill this knowledge gap, in this chapter, monotonic and incremental
cyclic loading on Fe-Mn-Si SMA under different temperatures was studied. The strength, ductility,
and strain recovery variation of Fe-Mn-Si SMA at room temperature 23 ℃, -40 ℃ and 50 ℃ were
extracted and analyzed.
11.3 Materials and methods
11.3.1 Materials
Material tested in this chapter was Fe-Mn-Si SMA. The composition was Fe–17Mn–5Si–
10Cr–4Ni–1(V,C) (mass%), obtained from Re-fer AG, Switzerland. The as received material was
18 mm diameter round bars. Cylindrical dog bone specimens with threads on both ends were
prepared by machining. Metric threads with a major diameter of 18 mm and a pitch of 1.5 mm
were machined on both ends of the sample; and the middle portion was reduced to a diameter of
12.7 mm, with a gauge length of 56 mm. The geometry of the specimens is shown in Fig. 11-1

238
Fig. 11-1 Dimensions of Fe-Mn-Si SMA samples. All dimensions are in mm.
11.3.2 Methods
The test setup used in this study was the same as that mentioned in Chapter 8, see Fig. 8-2.
An MTS 370.5 dynamic servo-hydraulic frame was used to apply the load. An MTS 651.06E−04
environmental chamber was used along with the MTS load frame to house the specimens during
testing. Two extension rods were used to connect the sample to hydraulic griping systems. A 50.8
mm gauge length Epsilon extensometer (model number 3542-0200-100-LHT) was used to
measure the strain. A BMS16HR-53 Mars Labs data acquisition (DAQ) system was used to record
the data.
Two loading protocols were adopted, namely monotonic loading and 1% strain incremental
cyclic loading, see Fig. 11-2. In the monotonic loading shown in Fig. 11-2 (a), the sample was
uniaxially stretched until fracture. In the incremental cyclic loading, a 1% tensile strain incremental
cyclic loading was applied until fracture. In each cycle, the sample was unloaded to near-zero
force. Extensometer controlled loading with a rate of 0.015 mm/s was adopted. Three temperatures
were investigated, namely 23 ℃, -40 ℃, and 50 ℃. When testing at different temperatures, the
following procedures were adopted. First, install and place the sample in the chamber; then, adjust
1 8 12.7
35 1 2 5 6
R=45
150
12 3 5
M18-1.5 thread 12.7 M18-1.5 thread
30 1 2 5 6
R=45
140
12 30
12.7
Note: All dimensions are in mm
M18-1.5 thread M18-1.5 thread
(a)
(b)
1 8 12.7

239
the chamber temperature to the target value, wait 40 minutes until the specimen temperature
reaches equilibrium with the targe temperature; finally, start the loading protocol.
After the tests at different temperatures, three key mechanical properties were extracted and
analyzed, namely Young’s modulus Eload, yield strength extracted by the 0.2% offset method fy,
and recovery strain r. The definition of these parameters are shown in Fig. 11-2.
11.4 Results
The monotonic loading stress-strain curves of Fe-Mn-Si SMA under different temperatures
are shown in Fig. 11-3. The yield strength and ultimate strength of Fe-Mn-Si SMA increases as the
temperature decreases, which is consistent with conventional reinforcing steels [171–173]. The
fracture strain of Fe-Mn-Si SMA at -40 ℃, 23 ℃ and 50 ℃ is 58%, 48% and 54%, respectively,
indicating its excellent deformability under a wide range of temperatures. From the zoomed-in
view shown in Fig. 11-3 (b), the yield strength extracted by the 0.2% offset method [104] at -40 ℃,
(a) (b)
Fig. 11-2 Schematic diagram of the loading protocols and definition of key parameters: (a)
monotonic loading, (b) 1% strain incremental cyclic loading.
Eload
0.2 %
fy
f Strain
Stress
Eload
Strain
Stress
r

240
20 ℃ and 50 ℃ is 532 MPa, 496 MPa and 472 MPa, respectively. The Young's modulus of FeMn-Si SMA is 168 GPa, which is insensitive to temperature variations.
(a) (b)
Fig. 11-3 Monotonic loading stress-strain curves of Fe-Mn-Si SMA at different temperatures: (a)
full view, and (b) zoomed-in view.
The cyclic loading stress-strain curves of Fe-Mn-Si SMA at different temperatures are shown in
Fig. 11-4. Under cyclic loading, the fracture strain of Fe-Mn-Si SMA at 23 ℃, -40 ℃ and 50 ℃
reached 27%, 45% and 40%, respectively. Such big fracture strains under incremental cyclic loading
indicates the excellent deformability of Fe-Mn-Si SMA. Similar to the monotonic loading results
shown in Fig. 11-3, Fe-Mn-Si SMA exhibited lower fracture strain at 23 ℃. The reason for this
phenomenon may be that the Fe-Mn-Si SMA exhibits lower deformability at room temperature.
More detailed research is needed to confirm this issue.
In most of existing studies using Fe-Mn-Si SMA to repair or strength concrete structures, the
material is only prestrained to 4% before activation [62,65]. From the cyclic loading results shown in
Fig. 11-4, it is seen that after 4% strain, the Fe-Mn-Si SMA still have a large deformability reservation.

241
For example, at 23 ℃ shown in Fig. 11-4 (a), if only being prestrained to 4% strain, the sample can
still withstand more than 20% of subsequent cyclic loading strain (i.e., the deformability
reservation). The deformability reservation at -40 ℃ and 50 ℃ is over 35% strain, see Fig. 11-4
(b) and (c). Therefore, it is concluded that a larger prestrain level can be adopted to take full use of
the deformability of Fe-Mn-Si SMA.
(a)
(b)

242
(c)
Fig. 11-4 Incremental cyclic tests on Fe-Mn-Si SMA at different temperatures: (a) 23
℃, (b) -40 ℃, and (c) 50 ℃
11.5 Discussion
Based on the definition of key parameters shown in Fig. 11-2 (a) and (b), the Young’s modulus
Eload and strain recovery r of Fe-Mn-Si SMA were extracted, as shown in Fig. 11-5. From Fig.
11-5 (a), it is seen that the Eload of Fe-Mn-Si SMA decreased gradually with the increase of applied
strain, and the rate of decreasing was slower at -40 ℃. At 23 ℃ and 50 ℃, the decreasing of Eload
exhibited a close decreasing pattern and rate. Specifically, in the first cycle, Eload = 165 GPa, when
the applied strain increased to 20%, Eload at -40 ℃, 23 ℃ and 50 ℃ decreased to 125 GPa, 92 GPa
and 90 GPa, respectively.
The r of Fe-Mn-Si SMA is an important indicator of its post-actuation deformability.
Specifically, a larger r indicates the Fe-Mn-Si SMA after actuation can withstand a larger cyclic
loading amplitude without losing the actuation stress. From Fig. 11-5 (b), it is seen that at all three
temperatures, the r showed an almost linear relationship with the maximum applied stress and a
higher the r was observed at a higher temperature.

243
(a)
(b)
(c)
Fig. 11-5 Mechanical properties extracted from the incremental cyclic loading tests: (a)
Eload vs. maximum applied strain, (b) r vs. maximum applied stress, and (c) r vs.
maximum applied strain.
0
50
100
150
200
0 5 10 15 20 25 30 35 40 45 50
Eload (GPa)
Maximum applied strain (%)
-40℃
23℃
50℃
0.00
0.50
1.00
1.50
2.00
400 600 800 1000 1200
r (%)
Maximum applied stress (MPa)
50℃
23℃
-40℃
0.00
0.50
1.00
1.50
2.00
0 5 10 15 20 25 30 35 40 45 50
r (%)
Maximum applied strain (%)
50℃
23℃
-40℃

244
The relationship between r and maximum applied strain is shown in Fig. 11-5 (c). It is seen
that at all three temperatures, the r increases as the maximum applied strain increases. The trends
at 23 ℃ and 50 ℃ are almost the same. The same is true at -40 ℃ up to 20% strain, but the increase
is smaller afterwards. For example, r at 5% maximum applied strain at -40 ℃, 23 ℃ and 50 ℃ is
0.86%, 0.83% and 0.92%, respectively. When the maximum applied strain was 25%, the r at -
40 ℃, 23 ℃ and 50 ℃ was 1.48%, 1.64% and 1.66%, respectively.
11.6 Summary of findings
It was found that Fe-Mn-Si SMA before actuation exhibited excellent deformability under a
wide range of temperatures from -40 ℃ to 50 ℃. Under both monotonic and incremental cyclic
loadings, the fracture strains of Fe-Mn-Si SMA at -40 ℃, 23 ℃ and 50 ℃ were all over 25%. The
maximum fracture strain of Fe-Mn-Si SMA at -40 ℃ reached 58%.
Under monotonic loading, the yield strength and ultimate strength of Fe-Mn-Si SMA
increased as the temperature decreased, which is consistent with traditional reinforcing steel. The
Young’s modulus of Fe-Mn-Si SMA under monotonic loading was insensitive to the ambient
temperature variations. In incremental cyclic loading, the Young’s modulus of Fe-Mn-Si SMA
decreased gradually with increasing strain, and at -40 ℃, the rate of this decrease was the smallest.
The recovery strain of Fe-Mn-Si SMA increased significantly with the increase of maximum
applied strain or maximum applied stress at -40 ℃, 23 ℃ and 50 ℃. Since the recovery strain of
Fe-Mn-Si SMA is an important indicator of its post-actuation deformability (i.e., a larger recovery

245
strain indicates the Fe-Mn-Si SMA after actuation can maintain its prestress force under larger
maximum strain amplitude during earthquakes), the increase of recovery strain along with the
maximum applied strain means that larger prestrain should be applied to maintain the recentering
capacity of columns reinforced with Fe-Mn-Si SMA.

246
Chapter 12 - Cyclic actuation behavior of Fe-Mn-Si SMA for use
in self-centering columns
The actuation behavior of Fe-Mn-Si SMA was investigated through application of cyclic, lowcycle fatigue, and monotonic loading tests. Different post-actuation temperatures (23 ℃, -40 ℃,
and 50 ℃), prestrain levels (4%, 15%, 20%, 25%, and 30%), and low-cycle fatigue loading
amplitudes (0.5% and 1%) were tested to determine the feasibility of using Fe-Mn-Si SMA in selfcentering bridge columns. The influence of cyclic loading and low-cycle fatigue loading on the
deformability and actuation stress degradation of Fe-Mn-Si SMA were analyzed.
12.1 Research motivation
Fe-Mn-Si SMA is a promising material for self-centering bridge column applications. In
Chapter 11, it was found Fe-Mn-Si SMA exhibits excellent ductility and temperature stability.
When apply Fe-Mn-Si SMA in self-centering bridge columns subjected to earthquake loadings,
the self-recentering of columns relies on the actuation stress of Fe-Mn-Si SMA. To ensure the
column reinforced with Fe-Mn-Si SMA remains a self-centering capacity under moderate or strong
earthquakes, Fe-Mn-Si SMA bars need to withstand large strain before the actuation stress reduces
to zero and without fracture. In addition, the ambient temperature variations may affect the

247
actuation stress and post-actuation deformability of Fe-Mn-Si SMA, and thus threaten the safety
of the bridges. Therefore, it is important to understand the cyclic actuation stability of Fe-Mn-Si
SMA under seismic loadings and variant ambient temperatures before adopting it in real bridges.
However, the basic mechanical properties of Fe-Mn-Si SMA regarding these aspects have not been
investigated. To fill this knowledge gap, the cyclic actuation behavior of Fe-Mn-Si SMA under
different temperatures, prestrain levels, and low-cycle fatigue loading amplitudes were
investigated in this study.
12.2 Experimental program
Material tested in this study was Fe-Mn-Si SMA. The material composition, supplier and
sample dimensions used in this study were the same as previously provided in Chapter 11.
12.2.1 Test setup
The test setup used in this study was also the same as previously mentioned in Chapter 11.
An MTS 370.5 dynamic servo-hydraulic frame was used to apply the load. An MTS 651.06E−04
environmental chamber was used along with the MTS load frame to house the specimens during
testing. Two extension rods were used to connect the sample to hydraulic griping systems due to
the size limitation of the environmental chamber. A 50.8 mm gauge length Epsilon extensometer
(model number 3542-0200-100-LHT) was used to measure the strain. A BMS16HR-53 Mars Labs
data acquisition (DAQ) system was used to record the data.

248
12.2.2 Test methods
A schematic diagram of the Fe-Mn-Si SMA actuation process is shown in Fig. 12-1 (a). First,
at room temperature (23 ℃), the sample was stretched from point O to a certain strain level (point
A) under extensometer control and then unloaded to near-zero force (point B) under force control.
This process is so called ‘prestraining’. For brevity, the strain level at point A is referred to as
prestrain level hereafter. Second, the sample was stretched to 200 MPa (point C). This was done
to avoid buckling when heating the sample under a constant strain, which causes thermal expansion
and development of a compressive force. Third, keep the strain of the sample constant, raise the
temperature in the chamber from 23 ℃ (point C) to 200 ℃ (point D), maintain this temperature
for 15 mins, then cool down to a certain temperature (point E). The end of cooling temperature at
point E is referred to as post-actuation temperature hereafter. The rate of both heating and cooling
during actuation (point C to D to E) was 3 ℃/min.
(a)
Strain
Stress
Temperature
Stress
Cooling
Heating
T2 T0 T1
O Prestrain level
A
B
C
D
E
C
D
E

249
(b) (c)
Fig. 12-1 Schematic diagrams of: (a) actuation process of Fe-Mn-Si SMA, (b) incremental
cyclic loading after actuation, and (c) low-cycle fatigue loading after actuation.
After the above actuation process, three different types of cyclic loading were applied
separately on the activated Fe-Mn-Si SMA samples, namely incremental cyclic loading, low-cycle
fatigue loading, and monotonic loading. In the incremental cyclic loading shown in Fig. 12-1 (b),
0.1% strain incremental cyclic loading was applied on the activated sample (from point E). In each
cycle, the material was loaded and unloaded with respect to the end of actuation state, i.e., point E
in Fig. 12-1 (b). As the cyclic strain amplitudes increased incrementally, the actuation stress
decreased gradually upon unloading. The incremental cyclic loading was stopped when the
actuation stress decreased to zero, as shown in Fig. 12-1 (b). In the low-cycle fatigue loading
shown in Fig. 12-1 (c), constant strain cyclic loading was applied for 500 cycles. In each cycle,
the material was loaded to a constant strain amplitude with respect to the end of actuation state,
i.e., point E in Fig. 12-1 (c), then unloaded by the same constant strain amplitude. After incremental
cyclic loading or low-cycle fatigue loading, the sample was monotonically stretched to failure.
Strain
Stress
B
C
D
E
O
A
0.1%
F
B
C
D
E
Strain
Stress
O
A
0.5% or 1.0%
F

250
Extensometer controlled method with a rate of 0.015 mm/s was used to for the loading and
unloading process during the incremental cyclic loading, low-cycle fatigue loading, and monotonic
loading.
The test matrix of cyclic actuation tests on Fe-Mn-Si SMA are shown in Table 12-1. The
sample labeling rule is shown in Fig. 12-2. The label of each sample consists of three parts: (1) the
prestrain level before actuation; (2) the loading protocol applied after actuation; (3) the postactuation temperature, i.e., the ambient temperature used for post-actuation tests. For example,
P(4%)-C(0.1%)-23C means: the sample was prestrained to 4% strain before actuation, and after
actuation, a 0.1% strain incremental cyclic loading was applied at 23 ℃.

251
Table 12-1 Test matrix of cyclic actuation tests on Fe-Mn-Si SMA.
No. Loading
type
Prestrain
value
Cyclic / Fatigue
loading strain
amplitude
Post-actuation
temperature
Sample
1
Incremental
cyclic
loading
4% 0.1% 23 ℃ P(4%)-C(0.1%)-23C
2 4% 0.1% 50 ℃ P(4%)-C(0.1%)-50C
3 4% 0.1% -40 ℃ P(4%)-C(0.1%)-m40C
4 15% 0.1% 23 ℃ P(15%)-C(0.1%)-23C
5 20% 0.1% 23 ℃ P(20%)-C(0.1%)-23C
6 25% 0.1% 23 ℃ P(25%)-C(0.1%)-23C
7 30% 0.1% 23 ℃ P(30%)-C(0.1%)-23C
8 Low-cycle
fatigue
loading
15% 0.5% 23 ℃ P(15%)-F(0.5%)-23C
9 15% 1.0% 23 ℃ P(15%)-F(1.0%)-23C
10 20% 1.5% 23 ℃ P(20%)-F(1.0%)-23C
Fig. 12-2 Labeling rules of the Fe-Mn-Si SMA specimens.
P (4%) – C (0.1%) – 23C
P (x%): Prestrain
level of x% before
actuation.
C (x%): Cyclic loading with
x% strain increment until
actuation stress disappear
F (x%): Fatigue loading with
x% constant strain amplitude
for 500 cycles
– –
Post-actuation
temperature
23C: 23 ℃
50C: 50 ℃
m40C: -40 ℃

252
12.3 Results and discussion
12.3.1 Effect of post-actuation temperature
The results of incremental cyclic loading on Fe-Mn-Si SMA at different post-actuation
temperatures are shown in Fig. 12-3. The prestrain level of all these three samples was 4%. From
Fig. 12-3, it is seen that the initial actuation stress (i.e., the stress at point E shown in Fig. 12-1 (b)),
decreased as the post-actuation temperature increased. Specifically, when the post-actuation
temperatures were-40 ℃, 23 ℃ and 50 ℃, the actuation stresses were 521 MPa, 402 MPa, and
188 MPa, respectively.
The reason why Fe-Mn-Si SMA exhibited a higher actuation stress at a lower post-actuation
temperature is that the actuation stress is mainly developed during the cooling process, as shown
in Fig. 12-1 (a), point E to D. During cooling, both phase transformation and thermal cooling
shorten the sample; thereby, generating the actuation stress. A low post-actuation temperature
means the cooling ends at a lower temperature, therefore, the phase transformation together with
the thermal shrinkage can be fully developed, leading to a higher actuation stress. Regarding the
actuation stress degradation with respect to incremental cyclic loadings, it is seen from Fig. 12-3
that, at 23 ℃, -40 ℃ and 50 ℃, the actuation stress decreased to zero when the cyclic loading
amplitude reached 0.82%, 0.80% and 0.73%, respectively. In Chapter 11, it is found that when the
maximum applied strain was 4%, the recovery strain at 23 ℃, -40 ℃ and 50 ℃ was 0.84%, 0.74%,
and 0.76%, respectively. The strain amplitudes when actuation stress disappear were consistent
with the recovery strain obtained in Chapter 11, see Fig. 11-5 (c).

253
(a) (b)
(c)
Fig. 12-3 Results of incremental cyclic loading tests on Fe-Mn-Si SMA under different postactuation temperatures: (a) 23 ℃, P(4%)-C(0.1%)-23C, (b) -40 ℃, P(4%)-C(0.1%)-m40C,
and (c) 50 ℃, P(4%)-C(0.1%)-50C.
12.3.2 Effect of prestrain level
In most of the past studies and applications, Fe-Mn-Si SMA is prestrained to only 4% prior
to actuation. This prestrain level is sufficient when Fe-Mn-Si SMA is used to strengthen exiting
beams in flexure or shear. In the research reported by Lee et al. [174] and Shahverdi et al. [62],
the actuation stress of Fe-Mn-Si SMA shows an apparent increase when the prestrain level
increases up to 2%. A prestrain level exceeding 4% has negligible influence on the actuation stress

254
magnitude. The magnitude of actuation stress is the main target when strengthening or repairing
existing structures. However, for self-centering bridge column applications, the prestressing force
in the Fe-Mn-Si SMA bars decreases as the Fe-Mn-Si SMA strain increases under seismic loads.
To maintain a reasonable level of recentering force in the column, it is necessary to increase the
prestrain level beyond 4%.
In Chapter 11, it is seen that the recovery strain of Fe-Mn-Si SMA increases as the maximum
applied strain increases. Specifically, at room temperature 23 ℃, when the maximum applied
strain is tripled from 5% to 15%, the recovery strain changes from 0.92% to 1.36%, increased by
48%. Therefore, a prestrain level of 4% does not take full advantage of the strain recovery capacity
of Fe-Mn-Si SMA. A larger prestrain level needs to be explored to fully utilize the strain recovery
capacity and ensure recentering of the bridge columns. To determine the effect of prestrain level
on strain recovery of Fe-Mn-Si SMA, prestrain levels of 15%, 20%, 25% and 30% were studied.
The results of incremental cyclic loading on Fe-Mn-Si SMA at different prestrain levels are
shown in Fig. 12-4. The post-actuation temperature of these samples were all 23 ℃. From Fig. 12-4
(a) to (c), it is seen that, at prestrain levels of 15%, 20% and 25%, the actuation stress decreased
to zero when the cyclic loading amplitude reached 1.3%, 1.5% and 1.7%, respectively. Compared
with the results at 4% prestrain level shown in Fig. 12-3 (a), it is seen that raising the prestrain level
effectively increased the post-actuation strain amplitude at which the actuation stress (that provides
recentering forces to the column) reduced to zero.

255
This is an important finding for self-centering bridge column applications, because this means
that, by increasing the prestrain level of Fe-Mn-Si SMA, the columns can withstand greater seismic
deformations without losing the recentering force. However, the prestrain level cannot be too high
because it could fracture the bar under seismic loads. As shown in Fig. 12-4 (d), when the prestrain
level was 30%, although the actuation stress did not reduce to zero, the sample fractured at 1.2%
strain (total strain experienced by the sample was 31.2%) during the incremental cyclic loading.
(a) (b)
(c) (d)
Fig. 12-4 Results of incremental cyclic loading on Fe-Mn-Si SMA at different prestrain
levels: (a) 15%, P(15%)-C(0.1%)-23C, (b) 20%, P(20%)-C(0.1%)-23C, (c) 25%, P(25%)-
C(0.1%)-23C, and (d) 30%, P(30%)-C(0.1%)-23C.

256
This means the 30% prestrain level is too high and significantly reduces the post-actuation
deformability of Fe-Mn-Si SMA.
The monotonic tests results of sample with 15%, 20% and 25% prestrain are shown in Fig.
12-5. At prestrain levels of 15%, 20% and 25%, the fracture strain was 25.8%, 16.1%, and 15.4%,
respectively. From the monotonic test results, it is confirmed that when increasing the prestrain
level to 25%, the Fe-Mn-Si SMA still has a post-actuation deformability reservation of more than
15%, which is on the same order of magnitude of a pristine deformed steel reinforcing bar.
However, in order to avoid the possible fracture of Fe-Mn-Si SMA bars, a prestrain level of 15%
or 20% is recommended for the practical applications.
12.3.3 Low-cycle fatigue resistance
Based on the incremental cyclic and subsequent monotonic loading test results, prestrain
levels of 15% and 20% were used to conduct low-cycle fatigue tests on Fe-Mn-Si SMA. To
(a) (b)
Fig. 12-5 Results of monotonic loading tests on Fe-Mn-Si SMA after incremental cyclic
loading: (a) full view, and (b) zoomed-in view.

257
simulate the seismic loading condition when applying Fe-Mn-Si SMA in self-centering bridge
columns, two strain amplitudes: 0.5% and 1.0% were applied. According to Motaref et al. [166],
when the drift ratio of a self-centering precast segmental column reaches 10%, the strain of posttensioning tendons is around 0.5%. Raza et al. [71] performed quasi-static cyclic loading on RC
columns strengthened with Fe-Mn-Si SMA bars and found that when the drift ratio of the column
reached 5%, the maximum strain of Fe-Mn-Si SMA was around 0.8%. Therefore, it is assumed
that the low-cycle fatigue strain amplitudes of 0.5% and 1.0% cover most of the earthquake
excitations when applying Fe-Mn-Si SMA in self-centering bridge columns.
The low-cycle fatigue test results of Fe-Mn-Si SMA are shown in Fig. 12-6. All specimens
were tested at room temperature. The specimens in Fig. 12-6 (a) and (b) were prestrained to 15%
and the one in Fig. 12-6 (c) was prestrained to 20%. Overall, it is found that Fe-Mn-Si SMA
exhibited excellent low-cycle fatigue resistance in terms of actuation stress and energy dissipation
stability up to 500 cycles. Specifically, from Fig. 12-6 (a) and (b), it is seen that under a prestrain
level of 15%, when the fatigue amplitude was 0.5%, the hysteresis loop of Fe-Mn-Si SMA showed
almost no degradation after 500 cycles: the stress-strain curves at cycle five, 100 and 500 were
almost the same. When the fatigue amplitude was 1.0%, the degradation of the hysteresis loop was
still negligible; after unloading, the residual actuation stress was around 1/2 of that under 0.5%
fatigue; besides, the hysteresis loop narrowed slightly under fatigue amplitude. From Fig. 12-6 (c),
it is seen that when the prestrain level increased to 20%, the hysteresis loop were still stable with

258
no degradation after 500 cycles. Comparing with Fig. 12-6 (b), it is noted that increasing the
prestrain level to 20% led to slightly narrower hysteresis loops.
(a) (b)
(c)
Fig. 12-6 Results of low-cycle fatigue tests on Fe-Mn-Si SMA: (a) prestrain 15% & fatigue 0.5%,
P(15%)-F(0.5%)-23C, (b) prestrain 15% & fatigue 1.0%, P(15%)-F(1.0%)-23C, and (c)
prestrain 20% & fatigue 1.0%, P(20%)-F(1.0%)-23C. Note: ‘C1’ means cycle No.1.
The variation of maximum and minimum actuation stress max and min, and damping ratio
R of Fe-Mn-Si SMA during fatigue loading were extracted and shown in Fig. 12-7. In bridge
applications, max and min guarantee the self-centering capacity and R guarantees the energy
dissipation capacity. From Fig. 12-7 (a), it is seen that the min and min of sample P(15%)-F(0.5%)-

259
23C showed almost no degradation up to 500 cycles; the R degraded slightly within 20 cycles and
then remained almost unchanged. At 500 cycles, R of sample P(15%)-F(0.5%)-23C degraded by
24% compared with the first cycle. For sample P(15%)-F(1.0%)-23C shown in Fig. 12-7 (b), after
500 cycles, the max, min and R decreased by 4%, 38%, and 25%, respectively. The min and R
of sample P(20%)-F(1.0%)-23C after 500 cycles decreased by 8% and 24%, respectively; and the
max showed a slightly increase of 3%, as shown in Fig. 12-7 (c).
(a)

260
(b)
(c)
Fig. 12-7 Variation in mechanical properties of Fe-Mn-Si SMA during low-cycle fatigue
loading: (a) P(15%)-F(0.5%)-23C, (b) P(15%)-F(1.0%)-23C, and (c) P(20%)-F(1.0%)-23C.
After low-cycle fatigue loading, monotonic tensile loading was performed on Fe-Mn-Si SMA
until failure. The results of monotonic loading test are shown in Fig. 12-8. The fracture strain of
sample P(15%)-F(0.5%)-23C, P(15%)-F(1.0%)-23C, and P(20%)-F(1.0%)-23C was 19.8%,
13.7%, and 9.3%, respectively. From Fig. 12-8 (b), it is seen that after 500 cycles of low-cycle

261
fatigue loading, the sample P(20%)-F(1.0%)-23C with 20% prestrain level showed higher strength
and Young’s modulus during the subsequent monotonic loading. It is worth noting that even at a
high prestrain level of 20%, after 500 cycles of 1.0% strain low-cycle fatigue loading, the sample
still showed excellent deformability with a fracture strain exceeding 9%.
(a) (b)
Fig. 12-8 Results of monotonic loading tests on Fe-Mn-Si SMA after low-cycle fatigue
loading: (a) full view, and (b) zoomed-in view.
12.4 Summary of findings
It was found that Fe-Mn-Si SMA exhibits excellent post-actuation deformability and lowcycle fatigue resistance, both of which are advantageous for self-centering bridge columns
subjected to strong earthquakes. Existing studies mostly adopted a 4% prestrain level when
applying Fe-Mn-Si SMA to strengthen concrete beams under gravity loads. However, it was found
that a small prestrain level of 4% did not fully utilize the strain recovery capacity of Fe-Mn-Si
SMA. Increasing the prestrain level could efficiently increase the post-actuation strain amplitude
before the actuation stress reduces to zero. Specifically, at the prestrain levels of 15%, 20% and

262
25%, the actuation stress decreased to zero when the cyclic loading amplitude reached 1.3%, 1.5%
and 1.7%, respectively. A 60% to 110% increase compared to the bars prestraining to 4% strain
(commonly used in past research). This means that, by increasing the prestrain level of Fe-Mn-Si
SMA, the columns could withstand greater seismic deformations without completely losing the
post-tensioning force, which is important for self-centering bridge column applications. However,
in order to avoid the possible fracture of Fe-Mn-Si SMA bars, a prestrain level of 15% or 20% is
recommended for the practical applications.
The post-actuation temperature was found to have a significant effect on the initial actuation
stress of Fe-Mn-Si SMA. Specifically, the initial actuation stress decreased as the post-actuation
temperature increased. When the post-post actuation temperatures were -40 ℃, 23 ℃ and 50 ℃,
the actuation stresses of Fe-Mn-Si SMA with a prestrain of 4% were 521 MPa, 402 MPa and 188
MPa, respectively.
When subjected to low-cycle fatigue loading, it was found that Fe-Mn-Si SMA exhibited
excellent fatigue resistance in terms of actuation stress and energy dissipation stability up to 500
cycles. Specifically, at a fatigue loading amplitude of 1.0% and a prestrain level of 20%, the
hysteresis loop of Fe-Mn-Si SMA remained almost unchanged up to 500 cycles. The minimum
actuation stress and damping ratio respectively decreased by 8% and 24% after 500 cycles; and
the maximum actuation stress showed a slight increase of 3%. After 500 cycles of 0.5% strain
fatigue loading, Fe-Mn-Si SMA with a prestrain of 15% exhibited a fracture strain over 19% in
the subsequent monotonic loading.

263
Chapter 13 - Summary and conclusions
Shape memory alloys (SMAs) are unique materials that could recover inelastic strains upon
stress removal, referred to as superelastic effect or superelasticity, or generate actuation stress if
restrained upon thermal stimulation, referred to as shape memory effect. The superelastic and
shape memory effects of SMA could be respectively used to provide strain recovery and posttensioning in bridge columns, thereby, eliminating permanent drifts and improving seismic
resistance.
Cu-Al-Mn SMA, Ni-Ti-Co SMA, and Fe-Mn-Si SMA are considered as the next generation
SMAs for use in bridges. In this dissertation, the material properties of these SMAs related to
bridge applications were characterized. Conventional Ni-Ti SMA and commonly used reinforcing
steels were also tested to benchmark the behavior of Cu-Al-Mn SMA, Ni-Ti-Co SMA, and FeMn-Si SMA.
13.1 Summary of completed research
13.1.1 Research on Cu-Al-Mn SMA
The following material properties of Cu-Al-Mn SMA were investigated: (i) corrosion
behavior, (ii) low-cycle fatigue behavior at different temperatures, (iii) machinability
characteristics, (iv) headed coupling behavior, and (v) cost effectiveness when applied in typical

264
bridge columns. Summaries of the completed research on these five aspects are presented as
follows.
First, the corrosion behavior of Cu-Al-Mn SMA was investigated by both long-term salt spray
corrosion and electrochemical corrosion tests. In the long-term corrosion tests, the effect of
corrosion on the mass loss and mechanical properties degradation of single crystal Cu-Al-Mn SMA
bar (SCB) and polycrystal Cu-Al-Mn SMA plate (PCP) samples was investigated, and compared
with four types of commonly used reinforcing steels: mild steel (MS), high chromium steel (XS),
epoxy coated steel (ES), and stainless steel (SS). For each steel, three different diameters (U.S. #3,
#5, and #10) were tested. In the electrochemical corrosion tests, potentialdynamic polarization
curves were employed to determine the corrosion rate of the abovementioned Cu-Al-Mn SMA,
four types of steels, and conventional Ni-Ti SMA.
Second, the low-cycle fatigue behavior of Cu-Al-Mn SMA was investigated at different
temperatures ranging from -40 ℃ to 50 ℃. The temperature range took into accounts the annual
ambient temperature variation in most parts of the world. Low-cycle fatigue loading up to 50,000
cycles was applied at a strain amplitude of 5%. The effect of low-cycle fatigue loading and
temperature variation on the superelastic properties of Cu-Al-Mn SMA was investigated. Both
single crystal Cu-Al-Mn SMA bar (SCB) and polycrystal Cu-Al-Mn SMA plate (PCP) samples
were tested, and compared with conventional Ni-Ti SMA.

265
Third, the machinability characteristics of Cu-Al-Mn SMA were investigated by single point
turning tests, and compared with conventional Ni-Ti SMA, mild steel (MS), and stainless steel
(SS). The fundamental machining characteristics of these materials such as chip formation, cutting
temperature, tool wear, workpiece surface roughness and diameter deviation were studied and
compared. Both single machining tests on one workpiece and continuous machining tests on
multiple workpieces were conducted. A wide range of cutting parameters were investigated,
including the cutting speed ranging from 15 to 120 m/min, feed rate ranging from 0.1 to 0.2
mm/rev, and depth of cut ranging from 0.5 to 1.5 mm.
Fourth, the headed coupling behavior of large diameter Cu-Al-Mn SMA was investigated by
both mechanical testing and microstructural analysis. In the mechanical testing, monotonic,
incremental, and constant strain cyclic loadings were applied, and the key mechanical properties
were extracted and discussed. In the microstructural analysis, electron backscatter diffraction
(EBSD), metallographic imaging, Vickers hardness testing, and fractographic analysis were
performed to investigate the crystal orientation, phase composition, and fracture surfaces of the
Cu-Al-Mn SMA after heading.
Fifth, the cost effectiveness of applying Cu-Al-Mn SMA in typical bridge columns was
investigated. Comparisons with columns reinforced with Ni-Ti SMA and conventional steels were
made. The cost of producing, processing, and coupling SMA bars were considered.

266
13.1.2 Research on Ni-Ti-Co SMA
The following material properties of Ni-Ti-Co SMA were investigated: (i) effect of
temperature on superelasticity, strength and ductility, (ii) low-cycle fatigue behavior at different
temperatures, and (iii) moment-curvature analysis when applied in typical bridge columns.
Summaries of the completed research on these three aspects are presented as follows.
First, cyclic tensile loading tests at temperatures ranging from -60 ℃ to 50 ℃ were performed
to investigate the effect of temperature on the superelasticity and ductility of Ni-Ti-Co SMA. NiTi SMA and Cu-Al-Mn SMA were also tested to benchmark the behavior of Ni-Ti-Co SMA.
Second, low-cycle fatigue behavior of Ni-Ti-Co SMA was investigated and compared with
Ni-Ti SMA at room temperature 23 ℃, low temperature -40 ℃, 0 ℃, and high temperature 50 ℃.
The effect of low-cycle fatigue loading and temperature variation on the superelastic properties of
Ni-Ti-Co SMA was analyzed and compared with Ni-Ti SMA.
Third, moment-curvature analyses were performed to investigate the flexural behavior of NiTi-Co SMA reinforced sections for possible implementation in typical bridge columns. For
comparison purposes, columns reinforced with Ni-Ti SMA and Cu-Al-Mn SMA bars were also
investigated. Conventional reinforced concrete (RC) bridge columns were analyzed as benchmarks
to determine the reference plastic moments. The influence of key parameters: column section
diameter, longitudinal reinforcement ratio, and axial force ratio was investigated.

267
13.1.3 Research on Fe-Mn-Si SMA
The following material properties of Fe-Mn-Si SMA were investigated: (i) effect of
temperature on ductility and recovery strain, (ii) cyclic actuation behavior, and (iii) low-cycle
fatigue behavior. Summaries of the completed research on these three aspects are presented as
follows.
First, the effect of temperature on the strength, ductility, and recovery strain of Fe-Mn-Si
SMA before thermal stimulation, i.e., actuation, was investigated. Monotonic and incremental
cyclic loading tests were respectively performed at 23 ℃, -40 ℃, and 50 ℃. The effect of
temperature on the key mechanical properties of Fe-Mn-Si SMA, such as Young’s modulus, yield
strength, ductility, and recovery strain was analyzed.
Second, strain-controlled incremental cyclic tensile loading tests were performed to
investigate the actuation stress degradation of Fe-Mn-Si SMA. The effect of different postactuation temperatures (23 ℃, -40 ℃, and 50 ℃) and prestrain levels (4%, 15%, 20%, 25%, and
30%) on the actuation stress degradation of Fe-Mn-Si SMA was investigated.
Third, low-cycle fatigue behavior of Fe-Mn-Si SMA after actuation was investigated by
strain-controlled cyclic tensile loadings up to 500 cycles. Effect of different prestrain levels (15%
and 20%) and different low-cycle fatigue loading amplitudes (0.5% and 1.0%) on the actuation
stress degradation and post-actuation deformability of Fe-Mn-Si SMA was analyzed.

268
13.2 Conclusions
13.2.1 Conclusions on Cu-Al-Mn SMA
From the long-term corrosion tests, it was found that corrosion on Cu-Al-Mn SMA was
localized, initiating with some local speckles and then growing to some deep pits. The local
corrosion pits merged together along the length, and the cross-sectional area decreased
accordingly. The mass loss resulting from the surface corrosion of Cu-Al-Mn SMA was overall
only around 1/3 of the MS. The mass loss and mechanical properties degradation of MS and XS
showed a dependence on the bar diameters: smaller diameter bars had higher mass loss and
mechanical properties degradation. Unlike the MS and XS, the mass loss and mechanical
properties degradation of Cu-Al-Mn SMA showed a dependence on the grain characteristics.
Specifically, the mass loss and mechanical properties degradation of PCP showed larger scatter
than SCB due to the variation in grain size, boundaries and orientations. The superelasticity,
particularly the recovery strain, of Cu-Al-Mn SMA had almost no degradation after long-term
corrosion. The strength degradation of Cu-Al-Mn SMA was less than that of MS and XS, but
higher than that of ES and SS.
From the electrochemical corrosion tests, it was found that the corrosion potential of Cu-AlMn SMA was comparable to that of Ni-Ti SMA and less than that of SS and XS; and the corrosion
current density of Cu-Al-Mn SMA was close to that of MS and ES. The corrosion rate of Cu-AlMn SMA obtained from Faraday’s law was only 1/3 that of the MS. From high to low, the
corrosion rate of MS, ES, Cu-Al-Mn SMA, XS, SS and Ni-Ti SMA were respectively obtained as

269
0.1910 mm/year, 0.1687 mm/year, 0.0565 mm/year, 0.0035 mm/year, 0.0008 mm/year, and
0.0006 mm/year.
From the low-cycle fatigue tests at different temperatures, it was found that Cu-Al-Mn SMA
showed flag shaped stress-strain curves with 100% strain recovery at low temperature -40 ℃, room
temperature 23 ℃, and high temperature 50 ℃. Conventional Ni-Ti SMA lost superelasticity
completely at -40 ℃. The superelastic temperature range of Cu-Al-Mn SMA was wider than that
of Ni-Ti SMA, but it is noted that the yield stress and energy dissipation of Cu-Al-Mn SMA were
lower than those of Ni-Ti SMA. Under low-cycle fatigue loading, the yield stress, damping ratio,
and recovery strain of Ni-Ti SMA degraded and stabilized within the first 100 loading cycles in
the temperature range from -10 ℃ to 50 ℃. Single crystal Cu-Al-Mn SMA showed slower
superelastic properties degradation and longer fatigue life than Ni-Ti SMA. Specifically, no
degradation in the energy dissipation or strain recovery capacity of SCB was seen within the first
100 loading cycles. Fifty thousand, 7,200, and 9,100 loading cycles were reached respectively at
25 ℃, -40 ℃ and 50 ℃ with no failure of the SCB specimens. The damping ratio of SCB at -40 ℃
and 50 ℃ showed no degradation until 1,000 loading cycles, indicating its potential for use in a
wide range of climate conditions. The superelastic properties degradation of PCP was close to that
of SCB within the first100 cycles, but the PCP showed larger scatters in the low-cycle fatigue life
due to the variation in grain structures.
From the machinability tests, it was found that Cu-Al-Mn SMA had much higher
machinability than Ni-Ti SMA. The cutting temperature during machining Cu-Al-Mn SMA was

270
around 1/3 of that when machining Ni-Ti SMA. Chip melting and welding that occurred when
machining Ni-Ti SMA did not occur when machining Cu-Al-Mn SMA. The tool wear from
machining Cu-Al-Mn SMA was only around 1/7 to 1/21 of that from machining Ni-Ti SMA.
Compared with MS and SS, the tool wear from machining Cu-Al-Mn SMA was overall close to
(0.6 to 1.8 times) that from SS and higher than (0.8 to 2.4 times) that from MS. Due to the
superelasticity and high ductility, machining Cu-Al-Mn SMA mainly showed an abrasion wear
mechanism in the nose region, and chipping and notch wear caused by continuous chip flow was
observed in the flank region, which was different from that of machining SS or MS. The chip flow
damage was considered to be a major challenge when machining Cu-Al-Mn SMA, particularly
when performing continuous machining for large stock removal.
From the headed coupling tests, it was found that using headed coupling method to connect
large diameter Cu-Al-Mn SMA bars with conventional steel rebar was feasible. The key was found
to be a consistent cooling rate in the central and peripheral regions of the headed end during
cooling. The reason was that, for large diameter Cu-Al-Mn SMA bars after heading, due to the
larger cross-sectional area, a differential cooling rate occurred between the central and peripheral
regions after heading. The peripheral region cooled faster, thus precipitating a high density of
bainite phase, while the central region cooled more slowly, resulting in the precipitation of the 
phase. The  phase had a higher ductility than the bainite phase. Therefore, the inhomogeneous
distribution of the bainite and  phase led to the strength incompatibility within the cross section
of the headed ends, which was prone to a premature failure.

271
From the cost effectiveness analysis, it was found that the Cu-Al-Mn SMA reinforced
columns showed economic advantage over the Ni-Ti SMA reinforced columns, particularly if a
machining method was used to connect SMA bars with the steel rebar. Compared with the cost of
conventional RC columns, the additional cost of Cu-Al-Mn SMA reinforced column was only
about 1/4 of the cost on Ni-Ti SMA reinforced column, indicating the cost effectiveness of
applying Cu-Al-Mn SMA in bridge columns.
In summary, it is concluded that Cu-Al-Mn SMA is a promising material for use in bridges.
The superelasticity, particularly the strain recovery capacity, of Cu-Al-Mn SMA had almost no
degradation after long-term corrosion, which indicates the feasibility of applying Cu-Al-Mn SMA
in bridges subjected to corrosive environments. Under low-cycle fatigue loading, flag shaped
stress-strain curves were observed for Cu-Al-Mn SMA from -40 ℃ to 50 ℃, and the
superelasticity showed almost no degradation within 100 cycles. This indicates the feasibility of
applying Cu-Al-Mn SMA in bridges subjected to earthquakes and extreme temperature variations.
In addition, the machinability and cost effectiveness Cu-Al-Mn SMA were higher than those of
conventional Ni-Ti SMA. However, it is noted that Cu-Al-Mn SMA has relatively low yield stress
and damping ratio. Additionally, the mechanical behavior of Cu-Al-Mn SMA has a large
dependence on the grain characteristics such as grain size, boundary, and orientations. These
limitations of Cu-Al-Mn SMA should be noted when applying it in bridges.

272
13.2.2 Conclusions on Ni-Ti-Co SMA
From the cyclic loading tests at different temperatures, it was found that Ni-Ti-Co SMA
showed flag shaped stress-strain curves from -40 ℃ to 40 ℃. The superelastic temperature range
of Ni-Ti-Co SMA (-40 ℃ to 40 ℃) was close to that of Cu-Al-Mn SMA (-40 ℃ to 50 ℃) and
wider than that of Ni-Ti SMA (0 ℃ to 50 ℃). It is noted that at high temperature 50 ℃, the stressstrain curves of Ni-Ti-Co SMA was no longer flag shaped and slight residual strain accumulation
(0.5% after unloading from 4%) was observed. At room temperature 23 ℃, the maximum recovery
strain of Ni-Ti-Co SMA (around 7%) was close to that of Ni-Ti SMA (around 7%) and smaller
than that of Cu-Al-Mn SMA (around 7.6%). The fracture strain of Ni-Ti-Co SMA (8%) was lower
than that of Ni-Ti SMA (8.5%) and Cu-Al-Mn SMA (12.7%). The dependence of yield stress on
temperature of Ni-Ti-Co SMA (6.5 MPa/ ℃) was close to that of Ni-Ti SMA (6.8 MPa/ ℃) but
was 3.3 times that of Cu-Al-Mn SMA (2.0 MPa/ ℃). The yield strength of Ni-Ti-Co SMA was
2.6 times that of Ni-Ti SMA and 3.3 times that of Cu-Al-Mn SMA. The wide superelastic
temperature range and high strength of Ni-Ti-Co SMA are advantageous for application in bridges.
From the 5% constant strain low-cycle fatigue tests at different temperatures, it was found
that at room temperature 23 ℃ and high temperature 50 ℃, Ni-Ti-Co SMA showed similar lowcycle fatigue resistance to Ni-Ti SMA in terms of superelasticity degradation and fatigue life.
While at low temperature, the strain recovery and energy dissipation of Ni-Ti-Co SMA showed
better low-cycle fatigue resistance than Ni-Ti SMA. Ni-Ti SMA lost superelasticity at -40 ℃.
Compared to Ni-Ti SMA at 0 ℃, the yield strength, energy dissipation, and strain recovery of Ni-

273
Ti-Co SMA at -40 ℃ showed much slower degradation. At -40 ℃, there was no loss of yield
strength, energy dissipation, and strain recovery of Ni-Ti-Co SMA during the first 100 cycles of
fatigue loading. This indicates that Ni-Ti-Co SMA has greater potential for use in low temperature
seismic applications.
From the moment-curvature analysis on typical bridge columns, it was found that, due to the
yield strength and Young’s modulus of Ni-Ti and Cu-Al-Mn SMA bars were smaller than those
of steel reinforcement, in general the stiffness and plastic moment in Ni-Ti and Cu-Al-Mn SMA
reinforced sections were lower than those of RC sections. To match the idealized plastic moment
of SMA columns with the RC columns, overall, more Ni-Ti and Cu-Al-Mn SMA bars were needed
that correspond to a higher longitudinal reinforcement ratio. While for Ni-Ti-Co SMA reinforced
sections, due to the higher yield strength and larger size availability of Ni-Ti-Co SMA bars, the
diameter of Ni-Ti-Co reinforced sections in most cases remained the same as that of the RC
sections, and the longitudinal reinforcement ratio of Ni-Ti-Co reinforced sections could remain the
same as that of the RC sections, which indicates the advantages of Ni-Ti-Co SMA and the potential
of using it in real bridges.
In summary, it is concluded that Ni-Ti-Co SMA is a promising material for use in bridges.
The Ni-Ti-Co SMA had a wide superelastic temperature range and low-cycle fatigue resistance
from -40 ℃ to 50 ℃, indicating the feasibility of applying it in bridges subjected to extreme
environments. In particular, the energy dissipation and strain recovery of Ni-Ti-Co SMA had
excellent low-cycle fatigue resistance at -40 ℃, which was much better than those of Ni-Ti SMA

274
at 0 ℃, indicating the greater potential of applying Ni-Ti-Co SMA in low temperature applications.
In addition, the high strength and large size availability of Ni-Ti-Co SMA enable columns
reinforced with it to have comparable flexural capacity with conventional RC columns without
increasing the longitudinal reinforcement ratio. This is better than Ni-Ti SMA and Cu-Al-Mn SMA,
which is advantageous for bridge applications.
13.2.3 Conclusions on Fe-Mn-Si SMA
From the monotonic and cyclic tests at different temperatures, it was found that Fe-Mn-Si
SMA before actuation exhibited excellent ductility under a wide range of temperatures from -40 ℃
to 50 ℃. Under both monotonic and incremental cyclic loadings, the fracture strains of Fe-Mn-Si
SMA at -40 ℃, 23 ℃ and 50 ℃ were all over 25%. The maximum fracture strain of Fe-Mn-Si
SMA at -40 ℃ reached 58%. The recovery strain of Fe-Mn-Si SMA showed a significant increase
with the increasing of applied strain at -40 ℃, 23 ℃ and 50 ℃. The increase of recovery strain
along with the maximum applied strain meant that a higher prestrain level could be applied to
improve the post-actuation deformation capacity of Fe-Mn-Si SMA.
From the incremental cyclic tests on activated Fe-Mn-Si SMA, it was found that the postactuation temperature had a significant effect on the initial actuation stress of Fe-Mn-Si SMA.
Specifically, the initial actuation stress decreased as the post-actuation temperature increased. In
addition, increasing the prestrain level could efficiently increase the post-actuation strain
amplitude before the actuation stress reduced to zero. Specifically, at the prestrain levels of 15%,

275
20% and 25%, the actuation stress decreased to zero when the cyclic loading amplitude reached
1.3%, 1.5% and 1.7%, respectively. A 60% to 110% increase compared to the bars prestrained to
4% strain (commonly used in past research). This means that, by increasing the prestrain level of
Fe-Mn-Si SMA, the columns could withstand greater seismic deformations without losing the
post-tensioning force, which is important for self-centering bridge column applications. However,
in order to avoid the possible fracture of Fe-Mn-Si SMA bars, a prestrain level of 15% or 20% is
recommended for the practical applications.
From the low-cycle fatigue tests, it was found that Fe-Mn-Si SMA with a prestrain level up
to 20% still exhibited excellent low-cycle fatigue resistance in terms of actuation stress and energy
dissipation stability. Specifically, at a prestrain level of 20% and a fatigue loading amplitude of
1.0%, the hysteresis loop of Fe-Mn-Si SMA remained almost unchanged up to 500 cycles. In
addition, increasing the prestrain levels to 15% or 20% did not sacrifice the post-actuation
deformability of Fe-Mn-Si SMA. For example, after 500 cycles of 0.5% strain fatigue loading, FeMn-Si SMA with a prestrain of 15% still exhibited a fracture strain over 19% in the subsequent
monotonic loading. This indicates that using a prestrain level up to 15% or 20% is feasible for real
bridge applications.
In summary, it is concluded that Fe-Mn-Si SMA is a feasible material for use in self-centering
bridge columns. Increasing the prestrain level was proved to be an effective way to improve the
post-actuation strain amplitude before the actuation stress reduced to zero. Furthermore, due to the
high deformability of Fe-Mn-Si SMA, increasing the prestrain level did not sacrifice the post-

276
actuation deformability. A prestrain level of 15% or 20% is recommended for the practical
applications of Fe-Mn-Si SMA in self-centering bridge columns.
13.3 Recommendations for future research
Based on the research findings from this dissertation, the recommendations for future research
are described as follows:
➢ Feasible headed coupling and heat treating methods for connecting large diameter Ni-TiCo SMA with steel rebar should be investigated. Microstructural analyses should be
performed to determine the effect of heading on the superelasticity of Ni-Ti-Co SMA.
➢ Low-cycle fatigue tests on large diameter Ni-Ti-Co SMA with headed coupling at
different temperatures should be performed to further facilitate the application of Ni-TiCo SMA in bridge applications.
➢ Effect of grain characteristics, such as grain boundary and orientation, on the temperature
dependence of Cu-Al-Mn SAM should be investigated.
➢ Microstructural analyses of the effect of grain characteristics (e.g., grain boundary and
orientation) on the corrosion behavior of Cu-Al-Mn SAM should be investigated.
➢ Full-scale experimental studies of bridge columns reinforced with Cu-Al-Mn, Ni-Ti-Co
and Fe-Mn-Si SMAs should be conducted to confirm the structural-level performance of
bridge columns using these SMAs.

277
➢ Experimental studies of using Cu-Al-Mn, Ni-Ti-Co and Fe-Mn-Si SMAs together with
other high performance materials, such as ultra-high-performance-concrete (UHPC) or
Engineered Cementitious Composites (ECC), should be conducted to further explore the
possibility of using these next generation SMAs to improve the seismic resistance of
bridges.
➢ Dynamic analysis on bridge columns reinforced with Cu-Al-Mn, Ni-Ti-Co and Fe-Mn-Si
SMAs should be performed to further evaluate the influence of SMA bars on the dynamic
response of bridges. Specifically, the effect of the following characteristics of SMA on
the dynamic response of bridges should be investigated: effect of strain rate dependence,
effect of low stiffness, effect of varying strength at different temperatures.
➢ More detailed constitutive models for superelastic Cu-Al-Mn and Ni-Ti-Co SMAs
considering the low-cycle fatigue degradation should be developed. The effect of
superelasticity degradation of Cu-Al-Mn and Ni-Ti-Co SMA on the seismic performance
of bridge columns should be investigated.
➢ Life-cycle assessment of bridges reinforced with Cu-Al-Mn, Ni-Ti-Co and Fe-Mn-Si
SMAs should be performed to evaluate the long-term benefits of using the next generation
SMA bars in bridge columns.

278
References
[1] O. Youssf, M.A. ElGawady, J.E. Mills, Experimental Investigation of Crumb Rubber
Concrete Columns under Seismic Loading, Structures. 3 (2015) 13–27.
https://doi.org/10.1016/j.istruc.2015.02.005.
[2] G. Ma, H. Li, Experimental Study of the Seismic Behavior of Predamaged ReinforcedConcrete Columns Retrofitted with Basalt Fiber–Reinforced Polymer, J. Compos. Constr.
19 (2015). https://doi.org/10.1061/(asce)cc.1943-5614.0000572.
[3] K. Kawashima, G.A. MacRae, J. Hoshikuma, K. Nagaya, Residual Displacement Response
Spectrum, J. Struct. Eng. 124 (1998) 523–530. https://doi.org/10.1061/(asce)0733-
9445(1998)124:5(523).
[4] W.F. Chen, L. Duan, Bridge engineering handbook: Fundamentals, Boca Rat. CRC Press.
(2014). https://doi.org/10.1201/b15616.
[5] Y. Araki, T. Endo, T. Omori, Y. Sutou, Y. Koetaka, R. Kainuma, K. Ishida, Potential of
superelastic Cu-Al-Mn alloy bars for seismic applications, Earthq. Eng. Struct. Dyn. 40
(2011) 107–115. https://doi.org/10.1002/eqe.1029.
[6] S. Varela, M.S. Saiidi, Dynamic performance of novel bridge columns with superelastic
CuAlMn shape memory alloy and ECC, Int. J. Bridg. Eng. 2 (2014) 29–58.
[7] S. Kise, A. Mohebbi, M.S. Saiidi, T. Omori, R. Kainuma, K.C. Shrestha, Y. Araki,
Mechanical splicing of superelastic Cu-Al-Mn alloy bars with headed ends, Smart Mater.
Struct. 27 (2018) 65–25. https://doi.org/10.1088/1361-665X/aabf0d.
[8] H. Hong, B. Gencturk, S.A. Brown, F. Hosseini, A. Jain, H. Aryan, S. Saiidi, Y. Araki, S.
Kise, Long-term corrosion resistance of Cu-Al-Mn superelastic alloys and steel rebar for
use in bridges, Constr. Build. Mater. 350 (2022) 128795.
https://doi.org/10.1016/j.conbuildmat.2022.128795.
[9] S. Varela, M. Saiidi, Resilient deconstructible columns for accelerated bridge construction
in seismically active areas, J. Intell. Mater. Syst. Struct. 28 (2017) 1751–1774.
https://doi.org/10.1177/1045389X16679285.
[10] F. Hosseini, B. Gencturk, S. Lahpour, D.I. Gil, An experimental investigation of innovative
bridge columns with engineered cementitious composites and Cu-Al-Mn super-elastic
alloys, Smart Mater. Struct. 24 (2015) 85029–16. https://doi.org/10.1088/0964-
1726/24/8/085029.

279
[11] S. Kise, Y. Araki, T. Omori, R. Kainuma, Orientation Dependence of Plasticity and Fracture
in Single-Crystal Superelastic Cu-Al-Mn SMA Bars, J. Mater. Civ. Eng. 33 (2021)
04021027. https://doi.org/10.1061/(asce)mt.1943-5533.0003568.
[12] S. Kise, N. Kataoka, R. Takamatsu, M. Nishida, T. Omori, R. Kainuma, Y. Araki,
Feasibility of Roll-Threading Superelastic Cu-Al-Mn SMA Rods, J. Mater. Civ. Eng. 33
(2021) 04021254. https://doi.org/10.1061/(asce)mt.1943-5533.0003874.
[13] Y. Araki, K.C. Shrestha, N. Maekawa, Y. Koetaka, T. Omori, R. Kainuma, Shaking table
tests of steel frame with superelastic Cu-Al-Mn SMA tension braces, Earthq. Eng. Struct.
Dyn. 45 (2016) 297–314. https://doi.org/10.1002/eqe.2659.
[14] K.C. Shrestha, Y. Araki, T. Nagae, T. Omori, Y. Sutou, R. Kainuma, K. Ishida,
Applicability of Cu-Al-Mn shape memory alloy bars to retrofitting of historical masonry
constructions, Earthq. Struct. 2 (2011) 233–256. https://doi.org/10.12989/eas.2011.2.3.233.
[15] M. Vahedi, S. Zolfagharysaravi, H. Ebrahimian, M. Saiid Saiidi, Experimental-analytical
investigation of accelerated bridge construction concrete columns with self-centering FeSMA bars subjected to near-fault ground motions, Eng. Struct. 299 (2024) 117127.
https://doi.org/https://doi.org/10.1016/j.engstruct.2023.117127.
[16] M. Tazarv, M. Saiidi, Low-Damage Precast Columns for Accelerated Bridge Construction
in High Seismic Zones, J. Bridg. Eng. 21 (2015) 4015056.
https://doi.org/10.1061/(ASCE)BE.1943-5592.0000806.
[17] C. Noguez, M.S. Saiidi, Shake-Table Studies of a Four-Span Bridge Model with Advanced
Materials, J. Struct. Eng. 138 (2012) 183–192. https://doi.org/10.1061/(ASCE)ST.1943-
541X.0000457.
[18] A. Mohebbi, E.J. Jordan, And, M.S. Saiidi, Exploratory experimental studies of spliced
CAM Shape memory alloy bars for seismic application, Tech. Rep. CCEER 18-04. (2018).
https://doi.org/hdl.handle.net/11714/7328.
[19] M. Tazarv, M. Saiid Saiidi, Reinforcing NiTi Superelastic SMA for Concrete Structures, J.
Struct. Eng. 141 (2015) 04014197. https://doi.org/10.1061/(asce)st.1943-541x.0001176.
[20] K.C. Shrestha, Y. Araki, T. Nagae, Y. Koetaka, Y. Suzuki, T. Omori, Y. Sutou, R. Kainuma,
K. Ishida, Feasibility of Cu-Al-Mn superelastic alloy bars as reinforcement elements in
concrete beams, Smart Mater. Struct. 22 (2013) 025025. https://doi.org/10.1088/0964-
1726/22/2/025025.

280
[21] M.S. Saiidi, M. O’Brien, S.Z. Mahmoud, Cyclic response of concrete bridge columns using
superelastic nitinol and bendable concrete, ACI Struct. J. 106 (2009) 69–77.
https://doi.org/10.14359/56285.
[22] W.S. Chang, Y. Araki, Use of shape-memory alloys in construction: A critical review, Proc.
Inst. Civ. Eng. Civ. Eng. 169 (2016) 87–95. https://doi.org/10.1680/jcien.15.00010.
[23] M. Tazarv, M. Saiid Saiidi, Low-Damage Precast Columns for Accelerated Bridge
Construction in High Seismic Zones, J. Bridg. Eng. 21 (2016) 04015056.
https://doi.org/10.1061/(asce)be.1943-5592.0000806.
[24] M.S. Saiidi, H. Wang, Exploratory study of seismic response of concrete columns with
shape memory alloys reinforcement, ACI Struct. J. 103 (2006) 436–443.
https://doi.org/10.14359/15322.
[25] S. Varela, M.S. Saiidi, A bridge column with superelastic NiTi SMA and replaceable rubber
hinge for earthquake damage mitigation, Smart Mater. Struct. 25 (2016) 5012.
https://doi.org/10.1088/0964-1726/25/7/075012.
[26] H.P. Hong., B. Gencturk., H. Aryan., A. Jain., Y. Araki., M.S. Saiidi., S. Kise, Low-cycle
fatigue behavior of Cu-Al-Mn superelastic alloys at different temperatures, Smart Mater.
Struct. 31 (2022) 115022. https://doi.org/10.1088/1361-665X/ac97d3.
[27] Z.B. Haber, M.S. Saiidi, D.H. Sanders, Seismic performance of precast columns with
mechanically spliced column-footing connections, ACI Struct. J. 111 (2014) 639–650.
https://doi.org/10.14359/51686624.
[28] K. Otsuka, X. Ren, Physical metallurgy of Ti-Ni-based shape memory alloys, Prog. Mater.
Sci. 50 (2005) 511–678. https://doi.org/10.1016/j.pmatsci.2004.10.001.
[29] Y. Zhang, J.A. Camilleri, S. Zhu, Mechanical properties of superelastic Cu-Al-Be wires at
cold temperatures for the seismic protection of bridges, Smart Mater. Struct. 17 (2008)
025008. https://doi.org/10.1088/0964-1726/17/2/025008.
[30] B. Gencturk, M.S. Saiidi, Material Characteristics of Cu-Based Superelastic Alloys for
Applications in Bridge Columns to Improve Seismic Performance, NCHRP-IDEA-Project210. Progr. Proj. Final Rep. (2022).
[31] A. Cladera, B. Weber, C. Leinenbach, C. Czaderski, M. Shahverdi, M. Motavalli, Ironbased shape memory alloys for civil engineering structures: An overview, Constr. Build.
Mater. 63 (2014) 281–293. https://doi.org/10.1016/j.conbuildmat.2014.04.032.

281
[32] W. Zhu, R. François, C.S. Poon, J.G. Dai, Influences of corrosion degree and corrosion
morphology on the ductility of steel reinforcement, Constr. Build. Mater. 148 (2017) 297–
306. https://doi.org/10.1016/j.conbuildmat.2017.05.079.
[33] E. Moreno, A. Cobo, G. Palomo, M.N. González, Mathematical models to predict the
mechanical behavior of reinforcements depending on their degree of corrosion and the
diameter of the rebars, Constr. Build. Mater. 61 (2014) 156–163.
https://doi.org/10.1016/j.conbuildmat.2014.03.003.
[34] L. Xiao, J. Peng, J. Zhang, Y. Ma, C.S. Cai, Comparative assessment of mechanical
properties of HPS between electrochemical corrosion and spray corrosion, Constr. Build.
Mater. 237 (2020) 117735. https://doi.org/10.1016/j.conbuildmat.2019.117735.
[35] I. Fernandez, J.M. Bairán, A.R. Marí, Corrosion effects on the mechanical properties of
reinforcing steel bars. Fatigue and σ-ε behavior, Constr. Build. Mater. 101 (2015) 772–783.
https://doi.org/10.1016/j.conbuildmat.2015.10.139.
[36] K.C. Shrestha, Y. Araki, T. Kusama, T. Omori, R. Kainuma, Functional Fatigue of
Polycrystalline Cu-Al-Mn Superelastic Alloy Bars under Cyclic Tension, J. Mater. Civ.
Eng. 28 (2016) 4015194. https://doi.org/10.1061/(asce)mt.1943-5533.0001417.
[37] M.S. Alam, M.A. Youssef, M.L. Nehdi, Exploratory investigation on mechanical anchors
for connecting SMA bars to steel or FRP bars, Mater. Struct. Constr. 43 (2010) 91–107.
https://doi.org/10.1617/s11527-010-9601-0.
[38] D. Hillis, M.S. Saiidi, Design, Construction, and Nonlinear Dynamic Analysis of Three
Bridge Bents Used in a Bridge System Test, CCEER-09-03. Proj. Rep. (2009).
[39] Y. Sutou, T. Omori, R. Kainuma, K. Ishida, Ductile Cu-Al-Mn based shape memory alloys:
General properties and applications, Mater. Sci. Technol. 24 (2008) 896–901.
https://doi.org/10.1179/174328408X302567.
[40] Y. Sutou, N. Koeda, T. Omori, R. Kainuma, K. Ishida, Effects of aging on stress-induced
martensitic transformation in ductile Cu-Al-Mn-based shape memory alloys, Acta Mater.
57 (2009) 5759–5770. https://doi.org/10.1016/j.actamat.2009.08.011.
[41] B. Gencturk, Y. Araki, T. Kusama, T. Omori, R. Kainuma, F. Medina, Loading rate and
temperature dependency of superelastic Cu-Al-Mn alloys, Constr. Build. Mater. 53 (2014)
555–560. https://doi.org/10.1016/j.conbuildmat.2013.12.002.
[42] J. Mohd Jani, M. Leary, A. Subic, M.A. Gibson, A review of shape memory alloy research,
applications and opportunities, Mater. Des. 56 (2014) 1078–1113.
https://doi.org/10.1016/j.matdes.2013.11.084.

282
[43] A. Ölander, An electrochemical investigation of solid cadmium-gold alloys, J. Am. Chem.
Soc. 54 (1932) 3819–3833. https://doi.org/10.1021/ja01349a004.
[44] L. Sun, W.M. Huang, Z. Ding, Y. Zhao, C.C. Wang, H. Purnawali, C. Tang, Stimulusresponsive shape memory materials: A review, Mater. Des. 33 (2012) 577–640.
https://doi.org/10.1016/j.matdes.2011.04.065.
[45] S. Raza, B. Shafei, M. Saiid Saiidi, M. Motavalli, M. Shahverdi, Shape memory alloy
reinforcement for strengthening and self-centering of concrete structures—State of the art,
Constr. Build. Mater. 324 (2022) 126628.
https://doi.org/10.1016/j.conbuildmat.2022.126628.
[46] A.H.M. Muntasir Billah, J. Rahman, Q. Zhang, Shape memory alloys (SMAs) for resilient
bridges: A state-of-the-art review, Structures. 37 (2022) 514–527.
https://doi.org/10.1016/j.istruc.2022.01.034.
[47] J. Ge, M.S. Saiidi, S. Varela, Computational studies on the seismic response of the State
Route 99 bridge in Seattle with SMA/ECC plastic hinges, Front. Struct. Civ. Eng. 13 (2019)
149–164. https://doi.org/10.1007/s11709-018-0482-6.
[48] S. Pareek, Y. Suzuki, Y. Araki, M.A. Youssef, M. Meshaly, Plastic hinge relocation in
reinforced concrete beams using Cu-Al-Mn SMA bars, Eng. Struct. 175 (2018) 765–775.
https://doi.org/10.1016/j.engstruct.2018.08.072.
[49] W. Trehern, H. Ozcan, B. Franco, N. Hite, N. Malone, B. Loveall, T.D. Morrison, O.
Benafan, I. Karaman, Exploring thermomechanical functionality of CuAlMn as an extreme
low temperature shape memory alloy, Mater. Lett. 308 (2022) 131246.
https://doi.org/10.1016/j.matlet.2021.131246.
[50] J. Frenzel, E.P. George, A. Dlouhy, C. Somsen, M.F.X. Wagner, G. Eggeler, Influence of
Ni on martensitic phase transformations in NiTi shape memory alloys, Acta Mater. 58
(2010) 3444–3458. https://doi.org/10.1016/j.actamat.2010.02.019.
[51] Y. Kishi, Z. Yajima, K. Shimizu, Relation between tensile deformation behavior and
microstructure in a Ti-Ni-Co shape memory alloy, Mater. Trans. 43 (2002) 834–839.
https://doi.org/10.2320/matertrans.43.834.
[52] R.M. Manjeri, D. Norwich, F. Sczerzenie, X. Huang, M. Long, M. Ehrlinspiel, A Study of
Thermo-mechanically Processed High Stiffness NiTiCo Shape Memory Alloy, J. Mater.
Eng. Perform. 25 (2016) 894–900. https://doi.org/10.1007/s11665-016-1932-8.

283
[53] T. Omori, N. Koeda, Y. Sutou, R. Kainuma, K. Ishida, Superplasticity of Cu-Al-Mn-Ni
shape memory alloy, Mater. Trans. 48 (2007) 2914–2918.
https://doi.org/10.2320/matertrans.D-MRA2007879.
[54] M. Miki, Y. Ogino, Effects of additional elements on the superplasticity of a Cu-14Al-3Ni
shape memory alloy, Mater. Trans. JIM. 30 (1989) 999–1008.
https://doi.org/10.2320/matertrans1989.30.999.
[55] S. Karthick, S. Shalini, S.S. Mani Prabu, K. Suhel, A. Vandan, C. Puneet, S. Manoj Kumar,
R. Venkatesh, I.A. Palani, Influence of quaternary alloying addition on transformation
temperatures and shape memory properties of Cu–Al–Mn shape memory alloy coated
optical fiber, Measurement. 153 (2020) 107379.
https://doi.org/10.1016/j.measurement.2019.107379.
[56] J.L. Liu, H.Y. Huang, J.X. Xie, The roles of grain orientation and grain boundary
characteristics in the enhanced superelasticity of Cu71.8Al17.8Mn10.4 shape memory
alloys, Mater. Des. 64 (2014) 427–433. https://doi.org/10.1016/j.matdes.2014.07.070.
[57] J. li Liu, Z.H. Chen, H. you Huang, J. xin Xie, Microstructure and superelasticity control
by rolling and heat treatment in columnar-grained Cu-Al-Mn shape memory alloy, Mater.
Sci. Eng. A. 696 (2017) 315–322. https://doi.org/10.1016/j.msea.2017.04.085.
[58] Y. Sutou, T. Omori, R. Kainuma, K. Ishida, Grain size dependence of pseudoelasticity in
polycrystalline Cu-Al-Mn-based shape memory sheets, Acta Mater. 61 (2013) 3842–3850.
https://doi.org/10.1016/j.actamat.2013.03.022.
[59] T. Omori, T. Kusama, S. Kawata, I. Ohnuma, Y. Sutou, Y. Araki, K. Ishida, R. Kainuma,
Abnormal grain growth induced by cyclic heat treatment, Science (80-. ). 341 (2013) 1500–
1502. https://doi.org/10.1126/science.1238017.
[60] T. Kusama, T. Omori, T. Saito, S. Kise, T. Tanaka, Y. Araki, R. Kainuma, Ultra-large single
crystals by abnormal grain growth, Nat. Commun. 8 (2017) 354–9.
https://doi.org/10.1038/s41467-017-00383-0.
[61] Y. Sutou, T. Omori, J.J. Wang, R. Kainuma, K. Ishida, Characteristics of Cu-Al-Mn-based
shape memory alloys and their applications, Mater. Sci. Eng. A. 378 (2004) 278–282.
https://doi.org/10.1016/j.msea.2003.12.048.
[62] M. Shahverdi, J. Michels, C. Czaderski, M. Motavalli, Iron-based shape memory alloy strips
for strengthening RC members: Material behavior and characterization, Constr. Build.
Mater. 173 (2018) 586–599. https://doi.org/10.1016/j.conbuildmat.2018.04.057.

284
[63] M. Shin, B. Andrawes, Experimental investigation of actively confined concrete using
shape memory alloys, Eng. Struct. 32 (2010) 656–664.
https://doi.org/10.1016/j.engstruct.2009.11.012.
[64] H. Zhao, B. Andrawes, Innovative prestressing technique using curved shape memory alloy
reinforcement, Constr. Build. Mater. 238 (2020) 117687.
https://doi.org/10.1016/j.conbuildmat.2019.117687.
[65] B. Schranz, J. Michels, C. Czaderski, M. Motavalli, T. Vogel, M. Shahverdi, Strengthening
and prestressing of bridge decks with ribbed iron-based shape memory alloy bars, Eng.
Struct. 241 (2021) 112467. https://doi.org/10.1016/j.engstruct.2021.112467.
[66] C. Czaderski, M. Shahverdi, J. Michels, Iron based shape memory alloys as shear
reinforcement for bridge girders, Constr. Build. Mater. 274 (2021) 121793.
https://doi.org/10.1016/j.conbuildmat.2020.121793.
[67] J. Vůjtěch, P. Ryjáček, J. Campos Matos, E. Ghafoori, Iron-Based shape memory alloy for
strengthening of 113-Year bridge, Eng. Struct. 248 (2021) 113231.
https://doi.org/10.1016/j.engstruct.2021.113231.
[68] A. Cladera, L.A. Montoya-Coronado, J.G. Ruiz-Pinilla, C. Ribas, Shear strengthening of
slender reinforced concrete T-shaped beams using iron-based shape memory alloy strips,
Eng. Struct. 221 (2020) 111018. https://doi.org/10.1016/j.engstruct.2020.111018.
[69] L. Zerbe, D. Vieira, A. Belarbi, A. Senouci, Uniaxial compressive behavior of circular
concrete columns actively confined with Fe-SMA strips, Eng. Struct. 255 (2022) 113878.
https://doi.org/10.1016/j.engstruct.2022.113878.
[70] D. Vieira, L. Zerbe, A. Belarbi, Numerical modeling of iron-based SMA confined concrete
columns under axial compressive loading, Eng. Struct. 275 (2023).
https://doi.org/10.1016/j.engstruct.2022.115185.
[71] S. Raza, R. Widmann, J. Michels, M. Saiid Saiidi, M. Motavalli, M. Shahverdi, Selfcentering technique for existing concrete bridge columns using prestressed iron-based shape
memory alloy reinforcement, Eng. Struct. 294 (2023) 116799.
https://doi.org/10.1016/j.engstruct.2023.116799.
[72] J. Bhandari, F. Khan, R. Abbassi, V. Garaniya, R. Ojeda, Modelling of pitting corrosion in
marine and offshore steel structures - A technical review, J. Loss Prev. Process Ind. (2015).
https://doi.org/10.1016/j.jlp.2015.06.008.

285
[73] K. Asami, M. Kikuchi, In-depth distribution of rusts on a plain carbon steel and weathering
steels exposed to coastal-industrial atmosphere for 17 years, Corros. Sci. (2003).
https://doi.org/10.1016/S0010-938X(03)00070-2.
[74] X.Y. Guo, J.F. Kang, J.S. Zhu, M.H. Duan, Corrosion Behavior and Mechanical Property
Degradation of Weathering Steel in Marine Atmosphere, J. Mater. Civ. Eng. (2019).
https://doi.org/10.1061/(ASCE)MT.1943-5533.0002814.
[75] A.K. Singh, N. Rani, S. Kundu, Optimized Cr concentration in reinforcement steel for high
corrosion resistance in simulated pore solution by electrochemical and Cyclic corrosion
tests, ISIJ Int. (2016). https://doi.org/10.2355/isijinternational.ISIJINT-2016-088.
[76] ASTM A1035/A1035M, Standard Specification for Deformed and Plain, Low-carbon,
Chromium, Steel Bars for Concrete Reinforcement, West Conshohocken, PA, USA. (2015).
https://doi.org/10.1520/A1035.
[77] X. Wu, D. Ye, H. Li, H. Yu, Corrosion characteristics of S23043 duplex stainless steel bars,
Constr. Build. Mater. (2018). https://doi.org/10.1016/j.conbuildmat.2018.05.008.
[78] M. Corradi, A. Di Schino, A. Borri, R. Rufini, A review of the use of stainless steel for
masonry repair and reinforcement, Constr. Build. Mater. (2018).
https://doi.org/10.1016/j.conbuildmat.2018.06.034.
[79] J.L. Smith, Y.P. Virmani, Performance of epoxy-coated rebars in bridge decks, Public
Roads. 60 (1996).
[80] B. Goffin, N. Banthia, N. Yonemitsu, Use of infrared thermal imaging to detect corrosion
of epoxy coated and uncoated rebar in concrete, Constr. Build. Mater. (2020).
https://doi.org/10.1016/j.conbuildmat.2020.120162.
[81] K. Zhang, H. Zhang, P. Liu, C. Zhang, W. Li, X. Chen, F. Ma, Electrophoretic deposition
of graphene oxide on NiTi alloy for corrosion prevention, Vacuum. 161 (2019) 276–282.
https://doi.org/10.1016/j.vacuum.2018.12.032.
[82] H. Wang, J. Jürgensen, P. Decker, Z. Hu, K. Yan, E.L. Gurevich, A. Ostendorf, Corrosion
behavior of NiTi alloy subjected to femtosecond laser shock peening without protective
coating in air environment, Appl. Surf. Sci. 501 (2020) 144338.
https://doi.org/10.1016/j.apsusc.2019.144338.
[83] Y. Say, B. Aksakal, Enhanced corrosion properties of biological NiTi alloy by
hydroxyapatite and bioglass based biocomposite coatings, J. Mater. Res. Technol. 9 (2020)
1742–1749. https://doi.org/10.1016/j.jmrt.2019.12.005.

286
[84] E.H. Habar, F. Tatengkeng, The difference of corrosion resistance between NiTi archwires
and NiTi with additional cooper archwires in artificial saliva, J. Dentomaxillofacial Sci. 5
(2020) 120–123. https://doi.org/10.15562/jdmfs.v5i2.1089.
[85] S. Tamilselvi, V. Raman, N. Rajendran, Evaluation of corrosion behavior of surface
modified Ti-6Al-4V ELI alloy in hanks solution, J. Appl. Electrochem. 40 (2010) 285–293.
https://doi.org/10.1007/s10800-009-9972-5.
[86] H. Zhao, B. Andrawes, Mechanical properties of NiTiNb shape memory alloy subjected to
a harsh corrosive environment, J. Mater. Civ. Eng. 29 (2017) 04016232.
https://doi.org/10.1061/(ASCE)MT.1943-5533.0001725.
[87] S.N. Saud, E. Hamzah, T. Abubakar, H.R. Bakhsheshi-Rad, Correlation of microstructural
and corrosion characteristics of quaternary shape memory alloys Cu-Al-Ni-X (X=Mn or
Ti), Trans. Nonferrous Met. Soc. China. 25 (2015) 1158–1170.
https://doi.org/10.1016/S1003-6326(15)63711-6.
[88] S.N. Saud, E. Hamzah, T. Abubakar, H.R. Bakhsheshi-Rad, M. Zamri, M. Tanemura,
Effects of Mn additions on the structure, mechanical properties, and corrosion behavior of
Cu-Al-Ni shape memory alloys, J. Mater. Eng. Perform. 23 (2014) 3620–3629.
https://doi.org/10.1007/s11665-014-1134-1.
[89] S. Pareek, S. Kise, F. Yamashita, B. Gencturk, F. Hosseini, S. Alexis Brown, Y. Araki,
Chemical Resistance of Cu-Al-Mn Superelastic Alloy Bars in Acidic and Alkaline
Environments, J. Mater. Civ. Eng. 33 (2021) 04020394.
https://doi.org/10.1061/(asce)mt.1943-5533.0003478.
[90] ASTM A516/A516M, Standard Specification for Pressure Vessel Plates, Carbon Steel, for
Moderate- and Lower-Temperature Service, West Conshohocken, PA, USA. 10 (2011).
https://doi.org/10.1520/A0516_A0516M-10R15.
[91] ASTM A615/A615M, Standard Specification for Deformed and Plain Carbon-Steel Bars
for Concrete Reinforcement, West Conshohocken, PA, USA. (2022).
https://doi.org/10.1520/A0615_A0615M-22.
[92] ASTM A955/A955M, A955/A955M Standard Specification for Deformed and Plain
Stainless Steel Bars for Concrete Reinforcement, West Conshohocken, PA, USA. (2020).
https://doi.org/10.1520/A0955_A0955M-20C.
[93] ASTM B117-19, Standard Practice for Operating Salt Spray (FOG) Apparatus, West
Conshohocken, PA, USA. (2019). https://doi.org/10.1520/B0117-19.

287
[94] ASTM G1-03, Standard Practice for Preparing, Cleaning, and Evaluating Corrosion Test
Specimens., West Conshohocken, PA, USA. (2017). https://doi.org/10.1520/G0001-
03R17E01.
[95] ASTM A370-23, Standard Test Methods and Definitions for Mechanical Testing of Steel
Products, West Conshohocken, PA, USA. (2019). https://doi.org/10.1520/A0370-23.
[96] ASTM G5-14, Standard Reference Test Method for Making Potentiodynamic Anodic
Polarization Measurements, West Conshohocken, PA, USA. (2014).
https://doi.org/10.1520/G0005-14R21.
[97] H. Kato, T. Ozu, S. Hashimoto, S. Miura, Cyclic stress-strain response of superelastic CuAl-Mn alloy single crystals, Mater. Sci. Eng. A. 264 (1999) 245–253.
https://doi.org/10.1016/s0921-5093(98)01095-8.
[98] V. Nachiappan, E.H. Cho, Corrosion of High Chromium and Conventional Steels
Embedded in Concrete, J. Perform. Constr. Facil. 19 (2005) 56–61.
https://doi.org/10.1061/(asce)0887-3828(2005)19:1(56).
[99] S. Imperatore, Z. Rinaldi, C. Drago, Degradation relationships for the mechanical properties
of corroded steel rebars, Constr. Build. Mater. 148 (2017) 219–230.
https://doi.org/10.1016/j.conbuildmat.2017.04.209.
[100] C.A. Apostolopoulos, The Influence of Corrosion and Cross-Section Diameter on the
Mechanical Properties of B500c Steel, J. Mater. Eng. Perform. 18 (2009) 190–195.
https://doi.org/10.1007/s11665-008-9281-x.
[101] C.A. Apostolopoulos, The Influence of Corrosion and Cross-Section Diameter on the
Mechanical Properties of B500c Steel, J. Mater. Eng. Perform. 18 (2009) 190–195.
https://doi.org/10.1007/s11665-008-9281-x.
[102] A.A. Almusallam, Effect of degree of corrosion on the properties of reinforcing steel bars,
Constr. Build. Mater. 15 (2001) 361–368. https://doi.org/10.1016/S0950-0618(01)00009-5.
[103] S. Xu, R. Tsukuda, M. Zhao, X. Xu, T. Omori, K. Yoshimi, R. Kainuma, Negative Poisson’s
ratio in Heusler-type Cu–Al–Mn shape memory alloy, Scr. Mater. 177 (2020) 74–78.
https://doi.org/10.1016/j.scriptamat.2019.10.015.
[104] ASTM E8/E8M, Standard Test Methods for Tension Testing of Metallic Materials, West
Conshohocken, PA, USA. (2013). https://doi.org/10.1520/E0008_E0008M-13A.

288
[105] M.G. Sohail, R. Kahraman, N.A. Alnuaimi, B. Gencturk, W. Alnahhal, M. Dawood, A.
Belarbi, Electrochemical behavior of mild and corrosion resistant concrete reinforcing
steels, Constr. Build. Mater. 232 (2020) 117205.
https://doi.org/10.1016/j.conbuildmat.2019.117205.
[106] G. Eggeler, E. Hornbogen, A. Yawny, A. Heckmann, M. Wagner, Structural and functional
fatigue of NiTi shape memory alloys, Mater. Sci. Eng. A. 378 (2004).
https://doi.org/10.1016/j.msea.2003.10.327.
[107] Y. Zhang, Y. You, Z. Moumni, G. Anlas, J. Zhu, W. Zhang, Experimental and theoretical
investigation of the frequency effect on low cycle fatigue of shape memory alloys, Int. J.
Plast. 90 (2017) 1–30. https://doi.org/10.1016/j.ijplas.2016.11.012.
[108] C. Maletta, E. Sgambitterra, F. Furgiuele, R. Casati, A. Tuissi, Fatigue properties of a
pseudoelastic NiTi alloy: Strain ratcheting and hysteresis under cyclic tensile loading, Int.
J. Fatigue. 66 (2014) 78–85. https://doi.org/10.1016/j.ijfatigue.2014.03.011.
[109] S. Miyazaki, K. Mizukoshi, T. Ueki, T. Sakuma, Y. Liu, Fatigue life of Ti-50 at.% Ni and
Ti-40Ni-10Cu (at.%) shape memory alloy wires, Mater. Sci. Eng. A. 273 (1999) 658–663.
https://doi.org/10.1016/s0921-5093(99)00344-5.
[110] C. Maletta, E. Sgambitterra, F. Furgiuele, R. Casati, A. Tuissi, Fatigue of pseudoelastic NiTi
within the stress-induced transformation regime: A modified Coffin-Manson approach,
Smart Mater. Struct. 21 (2012) 112001. https://doi.org/10.1088/0964-1726/21/11/112001.
[111] L.C. Brinson, I. Schmidt, R. Lammering, Stress-induced transformation behavior of a
polycrystalline NiTi shape memory alloy: Micro and macromechanical investigations via
in situ optical microscopy, J. Mech. Phys. Solids. 52 (2004) 1549–1571.
https://doi.org/10.1016/j.jmps.2004.01.001.
[112] G. Kang, D. Song, Review on structural fatigue of NiTi shape memory alloys: Pure
mechanical and thermo-mechanical ones, Theor. Appl. Mech. Lett. 5 (2015) 245–254.
https://doi.org/10.1016/j.taml.2015.11.004.
[113] Y. Araki, K. Kimura, T. Asai, T. Masui, T. Omori, R. Kainuma, Integrated mechanical and
material design of quasi-zero-stiffness vibration isolator with superelastic Cu-Al-Mn shape
memory alloy bars, J. Sound Vib. 358 (2015) 74–83.
https://doi.org/10.1016/j.jsv.2015.08.018.
[114] Y. Sutou, T. Omori, K. Yamauchi, N. Ono, R. Kainuma, K. Ishida, Effect of grain size and
texture on pseudoelasticity in Cu–Al–Mn-based shape memory wire, Acta Mater. 53 (2005)
4121–4133. https://doi.org/10.1016/j.actamat.2005.05.013.

289
[115] G. Kang, Q. Kan, L. Qian, Y. Liu, Ratchetting deformation of super-elastic and shapememory NiTi alloys, Mech. Mater. 41 (2009) 139–153.
https://doi.org/10.1016/j.mechmat.2008.09.001.
[116] V. Iasnii, P. Yasniy, D. Baran, A. Rudawska, The effect of temperature on low-cycle fatigue
of shape memory alloy, Frat. Ed Integrita Strutt. 13 (2019) 310–318.
https://doi.org/10.3221/IGF-ESIS.50.26.
[117] M.J. Mahtabi, N. Shamsaei, M.R. Mitchell, Fatigue of Nitinol: The state-of-the-art and
ongoing challenges, J. Mech. Behav. Biomed. Mater. 50 (2015) 228–254.
https://doi.org/10.1016/j.jmbbm.2015.06.010.
[118] U.S. Patel, S.K. Rawal, A.F.M. Arif, S.C. Veldhuis, Influence of secondary carbides on
microstructure, wear mechanism, and tool performance for different cermet grades during
high-speed dry finish turning of AISI 304 stainless steel, Wear. 452–453 (2020) 203285.
https://doi.org/10.1016/j.wear.2020.203285.
[119] F. Pusavec, H. Hamdi, J. Kopac, I.S. Jawahir, Surface integrity in cryogenic machining of
nickel based alloy - Inconel 718, J. Mater. Process. Technol. 211 (2011) 773–783.
https://doi.org/10.1016/j.jmatprotec.2010.12.013.
[120] S. Chaabani, P.J. Arrazola, Y. Ayed, A. Madariaga, A. Tidu, G. Germain, Comparison
between cryogenic coolants effect on tool wear and surface integrity in finishing turning of
Inconel 718, J. Mater. Process. Technol. 285 (2020) 116780.
https://doi.org/10.1016/j.jmatprotec.2020.116780.
[121] Y. Kaynak, H.E. Karaca, R.D. Noebe, I.S. Jawahir, Tool-wear analysis in cryogenic
machining of NiTi shape memory alloys: A comparison of tool-wear performance with dry
and MQL machining, Wear. 306 (2013) 51–63. https://doi.org/10.1016/j.wear.2013.05.011.
[122] H. Huang, H.Y. Zheng, Y. Liu, Experimental investigations of the machinability of
Ni50.6Ti49.4 alloy, Smart Mater. Struct. 14 (2005) S297–S301.
https://doi.org/10.1088/0964-1726/14/5/019.
[123] G.J. Wang, Z.Q. Liu, J.T. Niu, W.M. Huang, Q.Z. Xu, Work hardening influencing on shape
memory effect of NiTi alloy by varying milling speeds, Smart Mater. Struct. 28 (2019)
105034. https://doi.org/10.1088/1361-665X/ab3de2.
[124] G.J. Wang, Z.Q. Liu, X. Ai, W.M. Huang, J.T. Niu, Effect of cutting parameters on strain
hardening of nickel-titanium shape memory alloy, Smart Mater. Struct. 27 (2018) 75027.
https://doi.org/10.1088/1361-665X/aac43d.

290
[125] O. Benafan, R.D. Noebe, S.A. Padula, D.J. Gaydosh, B.A. Lerch, A. Garg, G.S. Bigelow,
K. An, R. Vaidyanathan, Temperature-dependent behavior of a polycrystalline NiTi shape
memory alloy around the transformation regime, Scr. Mater. 68 (2013) 571–574.
https://doi.org/10.1016/j.scriptamat.2012.11.042.
[126] H.C. Lin, K.M. Lin, Y.C. Chen, Study on the machining characteristics of TiNi shape
memory alloys, J. Mater. Process. Technol. 105 (2000) 327–332.
https://doi.org/10.1016/S0924-0136(00)00656-7.
[127] K. Weinert, V. Petzoldt, Machining of NiTi based shape memory alloys, Mater. Sci. Eng.
A. 378 (2004) 180–184. https://doi.org/10.1016/j.msea.2003.10.344.
[128] Y. Guo, A. Klink, C. Fu, J. Snyder, Machinability and surface integrity of Nitinol shape
memory alloy, CIRP Ann. - Manuf. Technol. 62 (2013) 83–86.
https://doi.org/10.1016/j.cirp.2013.03.004.
[129] K.E. Kirmacioglu, Y. Kaynak, O. Benafan, Machinability of Ni-rich NiTiHf high
temperature shape memory alloy, Smart Mater. Struct. 28 (2019) 55008.
https://doi.org/10.1088/1361-665X/ab02a2.
[130] E. Kaya, İ. Kaya, A review on machining of NiTi shape memory alloys: the process and
post process perspective, Int. J. Adv. Manuf. Technol. 100 (2019) 2045–2087.
https://doi.org/10.1007/s00170-018-2818-8.
[131] Y. Kaynak, H.E. Karaca, R.D. Noebe, I.S. Jawahir, The Effect of Active Phase of the Work
Material on Machining Performance of a NiTi Shape Memory Alloy, Metall. Mater. Trans.
A Phys. Metall. Mater. Sci. 46 (2015) 2625–2636. https://doi.org/10.1007/s11661-015-
2828-1.
[132] Y. Kaynak, H.E. Karaca, I.S. Jawahir, Cutting Speed Dependent Microstructure and
Transformation Behavior of NiTi Alloy in Dry and Cryogenic Machining, J. Mater. Eng.
Perform. 24 (2015) 452–460. https://doi.org/10.1007/s11665-014-1247-6.
[133] ASTM A276, Standard Specification for Stainless Steel Bars and Shapes, West
Conshohocken, PA, USA. (2015). https://doi.org/10.1520/A0276-13A.
[134] ASTM A108, Standard Specification for Steel Bar, Carbon and Alloy, Cold-Finished, West
Conshohocken, PA, USA. (2018). https://doi.org/10.1520/A0108-18.
[135] ASTM B16/B16M, Standard Specification for Free-Cutting Brass Rod, Bar and Shapes for
Use in Screw Machines, West Conshohocken, PA, USA. (2015).
https://doi.org/10.1520/B0016_B0016M-19.

291
[136] ASTM B209/B209M, Standard Specification for Aluminum and Aluminum-Alloy Sheet
and Plate, West Conshohocken, PA, USA. (2022).
https://doi.org/10.1520/B0209_B0209M-21A.
[137] ASTM B152/B152M, Standard Specification for Copper Sheet, Strip, Plate, and Rolled Bar,
West Conshohocken, PA, USA. (2019). https://doi.org/10.1520/B0152_B0152M-19.
[138] U. Patel, S. Rawal, B. Bose, A.F.M. Arif, S. Veldhuis, Performance evaluations of Ti-based
PVD coatings deposited on cermet tools for high-speed dry finish turning of AISI 304
stainless steel, Wear. 492 (2022) 204214. https://doi.org/10.1016/j.wear.2021.204214.
[139] Y. Peng, H. Miao, Z. Peng, Development of TiCN-based cermets: Mechanical properties
and wear mechanism, Int. J. Refract. Met. Hard Mater. 39 (2013) 78–89.
https://doi.org/10.1016/j.ijrmhm.2012.07.001.
[140] ISO 21920-2:2021, Geometrical product specifications (GPS) -Surface texture: Profile
method: Terms, definitions and surface texture parameters, Int. Stand. ISO. (2007).
[141] L. Osmond, D. Curtis, T. Slatter, Chip formation and wear mechanisms of SiAlON and
whisker-reinforced ceramics when turning Inconel 718, Wear. 486–487 (2021) 204128.
https://doi.org/10.1016/j.wear.2021.204128.
[142] G.T. Smith, Cutting Tool Technology: Industrial Handbook, Springer London. (2008).
https://doi.org/10.1007/978-1-84800-205-0.
[143] ISO 3685:1993, Tool life testing with single-point turning tools, Int. Stand. ISO. (1993).
[144] J.P. Davim, Modern Machining Technology: A Practical Guide, Woodhead Publ. (2011).
https://doi.org/10.1533/9780857094940.1.
[145] I.S. Jawahir, R. Ghosh, X.D. Fang, P.X. Li, An investigation of the effects of chip flow on
tool-wear in machining with complex grooved tools, Wear. 184 (1995) 145–154.
https://doi.org/10.1016/0043-1648(94)06572-1.
[146] M. Kaladhar, K.V. Subbaiah, C.S. Rao, Determination of Optimum Process Parameters
during turning of AISI 304 Austenitic Stainless Steels using Taguchi method and ANOVA,
Int. J. Lean Think. 3 (2012) 1.
[147] I. Korkut, M. Kasap, I. Ciftci, U. Seker, Determination of optimum cutting parameters
during machining of AISI 304 austenitic stainless steel, Mater. Des. 25 (2004) 303–305.
https://doi.org/10.1016/j.matdes.2003.10.011.

292
[148] V.P. Astakhov, Surface Integrity - Definition and Importance in Functional Performance,
Surf. Integr. Mach. Springer, London. (2009). https://doi.org/10.1007/978-1-84882-874-
2_1.
[149] Y. Ozcatalbas, F. Ercan, The effects of heat treatment on the machinability of mild steels,
J. Mater. Process. Technol. 136 (2003) 227–238. https://doi.org/10.1016/S0924-
0136(03)00273-5.
[150] M. Tazarv, M.S. Saiidi, Seismic design of bridge columns incorporating mechanical bar
splices in plastic hinge regions, Eng. Struct. 124 (2016) 507–520.
https://doi.org/10.1016/j.engstruct.2016.06.041.
[151] P.K. Dahal, M. Tazarv, Mechanical bar splices for incorporation in plastic hinge regions of
RC members, Constr. Build. Mater. 258 (2020) 120308.
https://doi.org/10.1016/j.conbuildmat.2020.120308.
[152] H. Hong, B. Gencturk, H. Aryan, A. Jain, Y. Araki, M.S. Saiidi, S. Kise, Low-cycle fatigue
behavior of Cu–Al–Mn superelastic alloys at different temperatures, Smart Mater. Struct.
31 (2022) 115022. https://doi.org/10.1088/1361-665X/ac97d3.
[153] Y. Sutou, T. Omori, R. Kainuma, K. Ishida, N. Ono, Enhancement of superelasticity in CuAl-Mn-Ni shape-memory alloys by texture control, Metall. Mater. Trans. A Phys. Metall.
Mater. Sci. 33 (2002) 2817–2824. https://doi.org/10.1007/s11661-002-0267-2.
[154] T. Omori, S. Kawata, R. Kainuma, Orientation dependence of superelasticity and stress
hysteresis in Cu-Al-Mn alloy, Mater. Trans. 61 (2020) 55–60.
https://doi.org/10.2320/matertrans.MT-MJ2019008.
[155] Y. Sutou, N. Koeda, T. Omori, R. Kainuma, K. Ishida, Effects of ageing on bainitic and
thermally induced martensitic transformations in ductile Cu-Al-Mn-based shape memory
alloys, Acta Mater. 57 (2009) 5748–5758. https://doi.org/10.1016/j.actamat.2009.08.003.
[156] R. Kainuma, N. Satoh, X.J. Liu, I. Ohnuma, K. Ishida, Phase equilibria and Heusler phase
stability in the Cu-rich portion of the Cu-Al-Mn system, J. Alloys Compd. 266 (1998) 191–
200. https://doi.org/10.1016/S0925-8388(97)00425-8.
[157] J.L. González-Velázquez, Fractography and failure analysis, Springer Int. Publ. AG. (2018).
https://doi.org/10.1007/978-3-319-76651-5.
[158] M.E. Stevenson, P.D. Umberger, S.F. Uchneat, Fracture Appearance and Mechanisms of
Deformation and Fracture, ASM Int. (2021).
https://doi.org/10.31399/asm.hb.v11.a0006774.

293
[159] W.T. Becker, S. Lampman, Fracture Appearance and Mechanisms of Deformation and
Fracture, ASM Int. (2018). https://doi.org/10.31399/asm.hb.v11.a0003537.
[160] M.S. Saiidi, M. Tazarv, S. Varela, S. Bennion, M.L. Marsh, I. Ghorbani, T.P. Murphy,
Seismic Evaluation of Bridge Columns with Energy Dissipating Mechanisms, Volume 1:
Research Overview and Volume 2: Guidelines . National Cooperative Highway Research
Program (NCHRP) Research Report 864. (2018).
[161] CSI SAP2000, Integrated software for structural analysis and design, Comput. Struct. Inc.,
Berkeley, CA, USA. (2009).
[162] D.C. Lagoudas, Shape Memory Alloys: modeling and engineering applications, SpringerVerlag US. (2008). https://doi.org/10.1007/978-0-387-47685-8.
[163] PEER, Open System for Earthquake Engineering Simulation (OpenSees), Univ. California,
Berkeley, CA. (2000).
[164] J.B. Mander, M.J.N. Priestley, R. Park, Theoretical Stress‐Strain Model for Confined
Concrete, J. Struct. Eng. 114 (1988). https://doi.org/10.1061/(ASCE)0733-
9445(1988)114:8(1804).
[165] D.C. Schoettler, Restrepo, Guerrini, A Full-Scale, Single-Column Bridge Bent Tested by
Shake-Table Excitation, PEER Rep. No. 2015/02. (2015).
[166] S. Motaref, Seismic Response of Precast Bridge Columns with Energy Dissipating Joints,
Univ. Nevada, Reno. (2011).
[167] M. Tazarv, Next Generation of Bridge Columns for Accelerated Bridge Construction in
High Seismic Zones, Univ. Nevada, Reno. (2014).
[168] AASHTO, Guide Specifications for LRFD Seismic Bridge Design, Am. Assoc. State
Highw. Transp. Off. (2011).
[169] C. Pulido, M. Saiid Saiidi, D. Sanders, A. Itani, S. El-Azazy, Seismic performance of twocolumn bents - Part I: Retrofit with carbon fiber-reinforced polymer fabrics, ACI Struct. J.
101 (2004) 558–568. https://doi.org/10.14359/13342.
[170] ACI 318-14, Building code requirements for structural concrete and commentary, Am.
Concr. Inst. (2014).

294
[171] J.B. Yan, J. Cao, P. Xie, J. Xie, Concrete-filled circular stainless-steel tubular columns
subjected to low-temperature axial compression, Structures. 55 (2023).
https://doi.org/10.1016/j.istruc.2023.06.054.
[172] L.A. Montejo, J.E. Sloan, M.J. Kowalsky, T. Hassan, Cyclic Response of Reinforced
Concrete Members at Low Temperatures, J. Cold Reg. Eng. 22 (2008).
https://doi.org/10.1061/(asce)0887-381x(2008)22:3(79).
[173] L. Dahmani, A. Khenane, S. Kaci, Behavior of the reinforced concrete at cryogenic
temperatures, Cryogenics (Guildf). 47 (2007) 517–525.
https://doi.org/10.1016/j.cryogenics.2007.07.001.
[174] W.J. Lee, B. Weber, G. Feltrin, C. Czaderski, M. Motavalli, C. Leinenbach, Stress recovery
behaviour of an Fe–Mn–Si–Cr–Ni–VC shape memory alloy used for prestressing, Smart
Mater. Struct. 22 (2013) 125037. https://doi.org/10.1088/0964-1726/22/12/125037.

295
Appendix A - Long-term corrosion test results
This appendix presents the supplementary results of long-term corrosion tests that are not
shown in Chapter 3. Appendix A.1 and A.2 respectively provide the original data of mass loss and
mechanical behavior degradation of single crystal Cu-Al-Mn SMA bar (SCB), and polycrystal CuAl-Mn SMA plate (PCP) during long-term corrosion tests. Appendix A.3 to A.6 respectively
provide the original data of mass loss and mechanical behavior degradation of mild steel (MS),
high chromium steel (XS), epoxy coated steel (ES), and stainless steel (SS) during long-term
corrosion tests.

296
A.1 Single crystal Cu-Al-Mn SMA bar (SCB)
Table A-1 Mass loss and mechanical behavior degradation of single crystal Cu-Al-Mn SMA bar (SCB)
during long-term corrosion tests.
Specimen
label
Corrosion
time
(Day)
Mass
loss
(g)
Mass loss
percent
(%)
Ms Af resi hy
Ori
(MPa)
Norm
Ori
(MPa)
Norm
Ori
(%)
Norm
Ori
(MPa)
Norm
SCB-1
0 170.8 0.00 248.8 1.00 188.2 1.00 0.03 1.00 60.60 1.00
20 170.4 0.25 247.1 0.99 195.8 1.04 0.01 0.27 51.30 0.85
30 170.3 0.30 248.7 1.00 194.9 1.04 0.02 0.64 53.80 0.89
75 169.3 0.87 249.1 1.00 200.0 1.06 0.02 0.60 49.10 0.81
1051 NA NA NA NA NA NA NA NA NA NA
SCB-2
0 171.3 0.00 210.4 1.00 136.1 1.00 0.08 1.00 74.30 1.00
20 170.9 0.28 207.2 0.98 139.2 1.02 0.07 0.08 68.00 0.92
30 170.8 0.31 207.4 0.99 143.5 1.05 0.06 0.75 63.90 0.86
75 169.7 0.93 209.6 1.00 146.0 1.07 0.03 0.34 63.60 0.86
1051 161.4 5.80 185.5 0.88 143.9 1.06 0.03 0.35 41.60 0.56
SCB-3
0 170.8 0.00 193.8 1.00 126.2 1.00 0.06 1.00 67.60 1.00
20 170.3 0.29 191.6 0.99 130.3 1.03 0.06 0.97 61.30 0.91
30 170.2 0.34 192.4 0.99 134.3 1.06 0.03 0.47 58.10 0.86
75 169.0 1.03 193.3 1.00 137.6 1.09 0.04 0.62 55.70 0.82
1051 161.4 5.50 166.3 0.86 126.5 1.00 0.05 0.94 39.80 0.59
Note: ‘NA’ means the data is not available; ‘Ori’ means the original data; ‘Norm’ means the normalized data.

297
A.2 Polycrystal Cu-Al-Mn SMA plate (PCP)
Table A-2 Mass loss and mechanical behavior degradation of polycrystal Cu-Al-Mn SMA plate (PCP)
during long-term corrosion tests.
Specimen
label
Corrosion
time
(Day)
Mass
loss
(g)
Mass loss
percent
(%)
Ms Af resi hy
Ori
(MPa)
Norm
Ori
(MPa)
Norm
Ori
(%)
Norm
Ori
(MPa)
Norm
PCP-1
0 202.3 0.00 225.7 1.00 217.9 1.00 0.08 1.00 7.8 1.00
9 201.8 0.21 210.0 0.93 197.1 0.90 0.06 0.70 12.9 1.65
20 201.8 0.21 242.0 1.07 220.0 1.01 0.28 3.48 22.0 2.82
30 201.8 0.21 243.0 1.08 221.0 1.01 0.28 3.56 22.0 2.82
75 200.0 1.11 217.1 0.96 193.5 0.89 0.12 1.53 23.6 3.03
1051 177.6 12.17 185.2 0.82 160.5 0.74 0.02 0.30 24.7 3.17
PCP-2
0 199.5 0.00 206.6 1.00 183.5 1.00 0.05 1.00 23.1 1.00
9 199.2 0.16 196.6 0.95 177.1 0.97 0.65 13.52 19.5 0.84
20 199.2 0.16 200.2 0.97 177.5 0.97 0.15 3.19 22.7 0.98
30 199.2 0.16 201.2 0.97 177.0 0.96 0.07 1.50 24.2 1.05
75 197.4 1.01 199.0 0.96 172.4 0.94 0.05 1.04 26.6 1.15
1051 175.3 12.10 165.7 0.80 134.7 0.73 0.24 5.06 31.0 1.34
PCP-3
0 198.6 0.00 194.7 1.00 170.2 1.00 0.04 1.00 24.5 1.00
9 198.2 0.21 188.3 0.97 169.1 0.99 0.16 3.66 19.2 0.78
20 198.3 0.15 190.3 0.98 171.3 1.01 0.20 4.52 19.0 0.78
30 198.3 0.14 189.6 0.97 167.1 0.98 0.05 1.05 22.5 0.92
75 196.6 1.01 178.0 0.91 156.7 0.92 0.08 1.75 21.3 0.87
1051 NA NA NA NA NA NA NA NA NA NA
PCP-4
0 184.6 0.00 251.0 1.00 233.9 1.00 0.17 1.00 17.1 1.00
9 184.2 0.23 238.4 0.95 219.9 0.94 0.04 0.26 18.5 1.08
20 184.2 0.23 259.1 1.03 240.9 1.03 0.06 0.36 18.2 1.06
30 184.2 0.23 248.6 0.99 229.8 0.98 0.09 0.53 18.8 1.10
75 182.6 1.11 234.7 0.94 214.8 0.92 0.05 0.29 19.9 1.16
1051 155.9 15.56 150.0 0.60 129.3 0.55 0.11 0.67 20.7 1.21
PCP-5
0 200.9 0.00 127.7 1.00 103.4 1.00 0.03 1.00 24.3 1.00
9 200.5 0.19 121.8 0.95 93.0 0.90 0.08 2.53 28.8 1.18
20 200.5 0.19 117.7 0.92 94.8 0.92 0.22 7.27 22.9 0.94
30 200.5 0.19 114.0 0.89 87.4 0.85 0.05 1.63 26.6 1.09
75 198.9 0.97 104.7 0.82 76.2 0.74 0.25 8.30 28.5 1.17
1051 180.0 10.39 61.3 0.48 41.7 0.40 0.02 0.57 19.6 0.81
Note: ‘NA’ means the data is not available; ‘Ori’ means the original data; ‘Norm’ means the normalized data.

298
A.3 Mild steel (MS)
Table A-3 Mass loss of mild steel (MS) during long-term corrosion tests.
Specimen label
Diameter
Corrosion
days
Initial
length
Initial
mass
Mass
after
corrosion
Mass loss
Mass loss
percent
(mm) (Day) (mm) (g) (g) (g) (%)
MS#3D0-1
9.53
(US#3)
0
364 \ \ \ 0.00
MS#3D0-2 369 \ \ \ 0.00
MS#3D0-3 357 \ \ \ 0.00
MS#3D20-1
20
359 190.60 189.60 1.10 0.56
MS#3D20-2 356 188.70 187.60 1.10 0.59
MS#3D20-3 360 190.20 189.00 1.30 0.66
MS#3D75-1
75
348 188.60 179.20 9.30 4.96
MS#3D75-2 350 189.00 180.20 8.80 4.65
MS#3D75-3 351 189.70 181.10 8.60 4.53
MS#3D296-1
296
371 189.70 148.10 41.60 21.91
MS#3D296-2 351 189.60 146.70 42.90 22.63
MS#3D296-3 371 190.10 145.60 44.50 23.42
MS#5D0-1
15.88
(US#5)
0
433 \ \ \ 0.00
MS#5D0-2 434 \ \ \ 0.00
MS#5D0-3 440 \ \ \ 0.00
MS#5D20-1
20
428 577.50 572.50 5.00 0.87
MS#5D20-2 429 579.00 574.00 5.00 0.86
MS#5D20-3 427 567.00 562.50 4.50 0.79
MS#5D75-1
75
422 573.50 558.50 15.00 2.62
MS#5D75-2 418 574.50 558.00 16.50 2.87
MS#5D75-3 419 578.00 562.50 15.50 2.68
MS#5D296-1
296
381 572.50 510.00 62.50 10.92
MS#5D296-2 381 578.50 511.50 67.00 11.58
MS#5D296-3 383 577.00 514.50 62.50 10.83
MS#10D0-1
32.26
(US#10)
0
662 \ \ \ 0.00
MS#10D0-2 661 \ \ \ 0.00
MS#10D0-3 663 \ \ \ 0.00
MS#10D20-1
20
665 4047.50 4032.00 15.50 0.38
MS#10D20-2 663 4041.00 4024.00 17.00 0.42
MS#10D20-3 666 4054.00 4035.00 19.00 0.47

299
MS#10D75-1
32.26
(US#10)
75
662 4014.50 3920.00 94.50 2.35
MS#10D75-2 661 4043.00 3951.50 91.50 2.26
MS#10D75-3 658 4040.50 3947.00 93.50 2.31
MS#10D296-1
296
663 4040.00 3780.00 260.00 6.44
MS#10D296-2 665 4063.50 3777.00 286.50 7.05
MS#10D296-3 660 4097.00 3787.00 310.00 7.57
Note: ‘\’ means that entry is not applicable
Table A-4 Mechanical property degradation of mild steel (MS) during long-term corrosion tests.
Specimen label
E E, norm fy fy, norm fm fm, norm rup rup, norm
Ori (GPa) Norm
Ori
(MPa)
Norm
Ori
(MPa)
Norm Ori (%) Norm
MS#3D0-1 203 1.01 528 1.00 678 1.01 12.20 1.07
MS#3D0-2 194 0.97 518 0.99 672 1.00 14.60 1.28
MS#3D0-3 204 1.02 531 1.01 663 0.99 7.50 0.65
MS#3D20-1 184 0.92 513 0.98 659 0.98 13.30 1.16
MS#3D20-2 217 1.08 513 0.97 657 0.98 14.20 1.24
MS#3D20-3 202 1.01 513 0.98 654 0.98 16.20 1.42
MS#3D75-1 174 0.87 473 0.90 608 0.91 15.30 1.34
MS#3D75-2 202 1.01 468 0.89 606 0.90 15.00 1.31
MS#3D75-3 186 0.93 457 0.87 597 0.89 11.10 0.97
MS#3D296-1 117 0.58 286 0.54 383 0.57 7.80 0.68
MS#3D296-2 149 0.74 297 0.56 396 0.59 7.40 0.65
MS#3D296-3 148 0.74 326 0.62 436 0.65 7.70 0.68
MS#5D0-1 244 1.21 441 1.01 628 1.01 15.50 0.97
MS#5D0-2 225 1.12 443 1.01 626 1.01 15.30 0.96
MS#5D0-3 134 0.67 427 0.98 613 0.99 17.00 1.07
MS#5D20-1 195 0.97 433 0.99 619 1.00 16.20 1.01
MS#5D20-2 223 1.11 440 1.01 622 1.00 18.00 1.13
MS#5D20-3 201 1.00 426 0.97 609 0.98 16.60 1.04
MS#5D75-1 120 0.60 417 0.95 600 0.96 14.70 0.92
MS#5D75-2 172 0.86 412 0.94 600 0.96 15.50 0.97
MS#5D75-3 189 0.94 416 0.95 605 0.97 17.40 1.09
MS#5D296-1 156 0.78 388 0.89 541 0.87 NA NA
MS#5D296-2 NA NA 377 0.86 524 0.84 NA NA
MS#5D296-3 NA NA 382 0.87 532 0.85 NA NA

300
MS#10D0-1 NA NA NA NA NA NA NA NA
MS#10D0-2 219 1.10 460 1.01 658 1.00 14.00 1.00
MS#10D0-3 179 0.90 446 0.99 654 1.00 13.90 1.00
MS#10D20-1 242 1.22 450 0.99 652 0.99 13.70 0.98
MS#10D20-2 186 0.93 445 0.98 652 0.99 13.50 0.97
MS#10D20-3 244 1.23 447 0.99 648 0.99 14.00 1.01
MS#10D75-1 154 0.77 435 0.96 636 0.97 13.00 0.93
MS#10D75-2 175 0.88 425 0.94 632 0.96 12.90 0.93
MS#10D75-3 209 1.05 434 0.96 630 0.96 10.80 0.78
MS#10D296-1 164 0.82 404 0.89 586 0.89 10.60 0.76
MS#10D296-2 164 0.82 413 0.91 599 0.91 10.40 0.75
MS#10D296-3 NA NA NA NA NA NA NA NA
Note: 'Ori' means the original data; 'Norm' means the normalized data; ‘NA’ means the data is not available.

301
A.4 High chromium steel (XS)
Table A-5 Mass loss of high chromium steel (XS) during long-term corrosion tests.
Specimen label
Diameter
Corrosion
days
Initial
length
Initial
mass
Mass
after
corrosion
Mass loss
Mass loss
percent
(mm) (Day) (mm) (g) (g) (g) (%)
XS#3D0-1
9.53
(US#3)
0
362 \ \ \ 0.00
XS#3D0-2 364 \ \ \ 0.00
XS#3D0-3 357 \ \ \ 0.00
XS#3D20-1
20
354 189.10 188.20 0.90 0.47
XS#3D20-2 347 189.70 188.90 0.80 0.42
XS#3D20-3 346 182.50 181.80 0.70 0.41
XS#3D75-1
75
349 190.80 186.70 4.10 2.14
XS#3D75-2 348 189.70 185.40 4.30 2.25
XS#3D75-3 341 185.50 182.00 3.50 1.88
XS#3D296-1
296
464 187.70 159.70 27.90 14.88
XS#3D296-2 361 185.80 162.10 23.70 12.74
XS#3D296-3 413 187.40 157.80 29.60 15.80
XS#5D0-1
15.88
(US#5)
0
413 \ \ \ 0.00
XS#3D0-2 409 \ \ \ 0.00
XS#5D0-3 407 \ \ \ 0.00
XS#5D20-1
20
423 587.00 584.50 2.50 0.43
XS#5D20-2 422 584.00 581.00 3.00 0.51
XS#5D20-3 403 593.00 589.50 3.50 0.59
XS#5D75-1
75
396 590.50 581.50 9.00 1.52
XS#5D75-2 411 603.50 595.00 8.50 1.41
XS#5D75-3 402 586.00 577.00 9.00 1.54
XS#3D296-1
296
387 594.00 539.00 55.00 9.26
XS#5D296-2 389 592.00 541.00 51.00 8.61
XS#5D296-3 383 587.00 533.50 53.50 9.11
XS#10D0-1
32.26
(US#10)
0
647 \ \ \ 0.00
XS#10D0-2 639 \ \ \ 0.00
XS#10D0-3 667 \ \ \ 0.00
XS#10D20-1
20
660 4052.00 4046.50 5.50 0.14
XS#10D20-2 663 4080.00 4075.00 5.00 0.12
XS#10D20-3 663 4069.50 4063.50 6.00 0.15

302
XS#10D75-1
32.26
(US#10)
75
\ 3920.00 3897.00 23.00 0.59
XS#10D75-2 \ 3971.50 3947.00 24.50 0.62
XS#10D75-3 \ 4079.50 4055.50 24.00 0.59
XS#10D296-1
296
664 4086.00 4002.50 83.50 2.04
XS#10D296-2 646 3987.00 3866.50 120.50 3.02
XS#10D296-3 660 4046.00 3927.00 119.00 2.94
Note: ‘\’ means that entry is not applicable
Table A-6 Mechanical property degradation of high chromium steel (XS) during long-term corrosion tests.
Specimen label
E E, norm fy fy, norm fm fm, norm rup rup, norm
Ori (GPa) Norm
Ori
(MPa)
Norm
Ori
(MPa)
Norm Ori (%) Norm
XS#3D0-1 229 1.03 971 1.00 1223 0.99 7.20 1.02
XS#3D0-2 228 1.02 984 1.01 1253 1.01 7.30 1.03
XS#3D0-3 211 0.95 968 0.99 1229 1.00 6.70 0.94
XS#3D20-1 213 0.96 972 1.00 1231 1.00 7.50 1.06
XS#3D20-2 213 0.96 1007 1.03 1255 1.02 6.20 0.88
XS#3D20-3 215 0.97 952 0.98 1195 0.97 6.50 0.92
XS#3D75-1 194 0.87 892 0.92 1156 0.94 5.00 0.71
XS#3D75-2 194 0.87 897 0.92 1177 0.95 3.60 0.51
XS#3D75-3 201 0.90 908 0.93 1175 0.95 5.20 0.74
XS#3D296-1 117 0.53 435 0.45 679 0.55 2.10 0.30
XS#3D296-2 150 0.67 534 0.55 693 0.56 2.40 0.34
XS#3D296-3 151 0.68 516 0.53 694 0.56 2.40 0.34
XS#5D0-1 189 0.77 873 0.98 1084 1.00 5.10 1.05
XS#3D0-2 245 1.00 854 0.96 1064 0.98 6.00 1.24
XS#5D0-3 300 1.23 935 1.05 1106 1.02 3.40 0.71
XS#5D20-1 201 0.82 897 1.01 1109 1.02 6.20 1.27
XS#5D20-2 231 0.94 898 1.01 1108 1.02 5.80 1.20
XS#5D20-3 181 0.74 834 0.94 1097 1.01 6.00 1.24
XS#5D75-1 168 0.69 805 0.91 1067 0.98 5.30 1.09
XS#5D75-2 204 0.83 888 1.00 1087 1.00 5.30 1.09
XS#5D75-3 182 0.75 810 0.91 1069 0.99 6.10 1.26
XS#3D296-1 161 0.66 733 0.83 900 0.83 NA NA
XS#5D296-2 166 0.68 750 0.85 915 0.84 NA NA
XS#5D296-3 161 0.66 705 0.79 892 0.82 NA NA

303
XS#10D0-1 177 0.90 1045 1.07 1068 1.00 NA 0.98
XS#10D0-2 218 1.10 1021 1.04 1074 1.00 NA 0.99
XS#10D0-3 199 1.01 877 0.89 1067 1.00 NA 1.03
XS#10D20-1 188 0.95 876 0.89 1073 1.00 NA 1.08
XS#10D20-2 199 1.01 855 0.87 1071 1.00 NA 1.05
XS#10D20-3 184 0.93 864 0.88 1073 1.00 NA 1.11
XS#10D75-1 180 0.91 967 0.99 1057 0.99 NA 0.89
XS#10D75-2 195 0.98 891 0.91 1057 0.99 NA 1.02
XS#10D75-3 170 0.86 816 0.83 1057 0.99 NA 0.95
XS#10D296-1 164 0.83 868 0.88 1076 1.01 NA 0.55
XS#10D296-2 199 1.00 846 0.86 989 0.92 NA 0.51
XS#10D296-3 172 0.87 716 0.73 996 0.93 NA 0.60
Note: 'Ori' means the original measured data; 'Norm' means the normalized data; ‘NA’ means the data is not available.

304
A.5 Epoxy coated steel (ES)
Table A-7 Mass loss of epoxy coated steel (ES) during long-term corrosion tests.
Specimen
label
Diameter
Corrosion
days
Initial
length
Initial
mass
Mass
after
corrosion
Mass loss
Mass loss
percent
(mm) (Day) (mm) (g) (g) (g) (%)
ES#3D0-1
9.53
(US#3)
0
360 \ \ \ 0.00
ES#3D0-2 359 \ \ \ 0.00
ES#3D0-3 359 \ \ \ 0.00
ES#3D20-1
20
359 185.40 185.40 0.00 0.00
ES#3D20-2 356 183.00 182.90 0.10 0.06
ES#3D20-3 365 184.00 184.00 0.00 0.00
ES#3D75-1
75
360 183.40 183.30 0.10 0.03
ES#3D75-2 360 184.40 184.40 0.00 0.00
ES#3D75-3 356 184.50 184.40 0.10 0.04
ES#3D296-1
296
330 184.00 183.50 0.50 0.25
ES#3D296-2 341 183.90 183.20 0.70 0.36
ES#3D296-3 351 184.00 183.70 0.30 0.19
ES#5D0-1
15.88
(US#5)
0
422 \ \ \ 0.00
ES#3D0-2 422 \ \ \ 0.00
ES#5D0-3 426 \ \ \ 0.00
ES#5D20-1
20
424 572.00 571.50 0.50 0.09
ES#5D20-2 425 571.50 571.50 0.00 0.00
ES#5D20-3 421 574.50 574.50 0.00 0.00
ES#5D75-1
75
419 572.00 571.50 0.50 0.09
ES#5D75-2 420 572.50 572.50 0.00 0.00
ES#5D75-3 424 571.50 571.50 0.00 0.00
ES#3D296-1
296
383 573.00 572.00 1.00 0.17
ES#5D296-2 383 571.00 570.00 1.00 0.18
ES#5D296-3 383 573.00 572.00 1.00 0.17
ES#10D0-1
32.26
(US#10)
0
659 \ \ \ 0.00
ES#10D0-2 660 \ \ \ 0.00
ES#10D0-3 659 \ \ \ 0.00
ES#10D20-1
20
659 4052.50 4051.00 1.50 0.04
ES#10D20-2 667 4036.00 4035.00 1.00 0.02
ES#10D20-3 657 4032.50 4030.00 2.50 0.06

305
ES#10D75-1
32.26
(US#10)
75
659 4043.00 4041.00 2.00 0.05
ES#10D75-2 657 4070.50 4068.50 2.00 0.05
ES#10D75-3 660 4100.00 4098.50 1.50 0.04
ES#10D296-1
296
660 4065.00 4061.50 3.50 0.09
ES#10D296-2 657 4057.00 4051.50 5.50 0.14
ES#10D296-3 656 4052.00 4049.00 3.00 0.07
Note: ‘\’ means that entry is not applicable
Table A-8 Mechanical property degradation of epoxy coated steel (ES) during long-term corrosion tests.
Specimen
label
E E, norm fy fy, norm fm fm, norm rup rup, norm
Ori
(GPa)
Norm
Ori
(MPa)
Norm
Ori
(MPa)
Norm Ori (%) Norm
ES#3D0-1 204 1.04 483 1.00 746 1.00 12.50 0.94
ES#3D0-2 196 1.00 486 1.00 751 1.00 13.80 1.04
ES#3D0-3 188 0.96 486 1.00 749 1.00 13.50 1.02
ES#3D20-1 205 1.05 485 1.00 752 1.00 12.80 0.96
ES#3D20-2 197 1.01 483 1.00 745 1.00 13.10 0.98
ES#3D20-3 210 1.07 491 1.01 748 1.00 11.50 0.87
ES#3D75-1 218 1.11 480 0.99 745 1.00 8.90 0.67
ES#3D75-2 206 1.05 485 1.00 756 1.01 13.30 1.00
ES#3D75-3 202 1.03 487 1.00 757 1.01 11.90 0.90
ES#3D296-1 202 1.03 493 1.02 755 1.01 10.30 0.78
ES#3D296-2 201 1.02 483 1.00 745 1.00 11.60 0.87
ES#3D296-3 205 1.05 489 1.01 757 1.01 11.10 0.84
ES#5D0-1 180 0.88 439 0.98 697 0.99 15.40 1.08
ES#3D0-2 192 0.94 466 1.04 726 1.03 13.00 0.91
ES#5D0-3 238 1.17 439 0.98 697 0.99 14.40 1.01
ES#5D20-1 176 0.87 464 1.03 722 1.02 12.40 0.87
ES#5D20-2 166 0.82 438 0.98 694 0.98 15.00 1.05
ES#5D20-3 180 0.89 467 1.04 726 1.03 14.90 1.04
ES#5D75-1 278 1.37 442 0.99 699 0.99 13.40 0.94
ES#5D75-2 216 1.06 442 0.99 699 0.99 18.00 1.26
ES#5D75-3 193 0.95 460 1.03 721 1.02 13.00 0.91
ES#3D296-1 195 0.96 471 1.05 729 1.03 NA NA
ES#5D296-2 180 0.89 470 1.05 731 1.03 NA NA
ES#5D296-3 190 0.93 466 1.04 727 1.03 NA NA

306
ES#10D0-1 247 1.10 469 1.00 664 1.00 11.10 1.07
ES#10D0-2 244 1.09 479 1.02 674 1.01 9.30 0.90
ES#10D0-3 180 0.80 464 0.98 654 0.98 10.70 1.03
ES#10D20-1 253 1.13 472 1.00 662 1.00 10.30 1.00
ES#10D20-2 151 0.68 471 1.00 662 1.00 10.30 1.00
ES#10D20-3 170 0.76 467 0.99 658 0.99 9.80 0.94
ES#10D75-1 213 0.95 460 0.98 652 0.98 9.30 0.90
ES#10D75-2 173 0.77 485 1.03 646 0.97 5.40 0.52
ES#10D75-3 172 0.77 484 1.03 645 0.97 5.50 0.54
ES#10D296-1 156 0.70 461 0.98 660 0.99 11.30 1.10
ES#10D296-2 174 0.78 484 1.03 665 1.00 10.00 0.96
ES#10D296-3 252 1.13 462 0.98 658 0.99 9.40 0.90
Note: 'Ori' means the original measured data; 'Norm' means the normalized data; ‘NA’ means the data is not available.

307
A.6 Stainless steel (SS)
Table A-9 Mass loss of stainless steel (SS) during long-term corrosion tests.
Specimen label
Diameter
Corrosion
days
Initial
length
Initial
mass
Mass
after
corrosion
Mass loss
Mass loss
percent
(mm) (Day) (mm) (g) (g) (g) (%)
SS#3D0-1
9.53
(US#3)
0
389 \ \ \ 0.00
SS#3D0-2 395 \ \ \ 0.00
SS#3D0-3 389 \ \ \ 0.00
SS#3D20-1
20
394 186.20 186.20 0.00 0.00
SS#3D20-2 394 185.90 185.90 0.00 0.00
SS#3D20-3 397 184.00 184.00 0.00 0.01
SS#3D75-1
75
389 182.30 182.30 0.00 0.02
SS#3D75-2 390 184.60 184.60 0.00 0.01
SS#3D75-3 384 182.80 182.80 0.00 0.02
SS#3D296-1
296
433 186.60 186.20 0.40 0.23
SS#3D296-2 413 185.20 184.80 0.50 0.26
SS#3D296-3 371 183.00 182.70 0.40 0.20
SS#5D0-1
15.88
(US#5)
0
457 \ \ \ 0.00
SS#3D0-2 436 \ \ \ 0.00
SS#5D0-3 445 \ \ \ 0.00
SS#5D20-1
20
442 590.50 590.50 0.00 0.00
SS#5D20-2 448 575.00 575.00 0.00 0.00
SS#5D20-3 460 574.50 574.50 0.00 0.00
SS#5D75-1
75
456 600.00 599.00 1.00 0.17
SS#5D75-2 436 581.50 581.00 0.50 0.09
SS#5D75-3 448 593.00 592.00 1.00 0.17
SS#3D296-1
296
398 604.00 604.00 0.00 0.00
SS#5D296-2 381 571.50 570.50 1.00 0.17
SS#5D296-3 383 574.00 573.50 0.50 0.09
SS#10D0-1
32.26
(US#10)
0
651 \ \ \ 0.00
SS#10D0-2 649 \ \ \ 0.00
SS#10D0-3 653 \ \ \ 0.00
SS#10D20-1
20
655 3994.50 3993.00 1.50 0.04
SS#10D20-2 648 3975.00 3973.50 1.50 0.04
SS#10D20-3 646 3960.00 3959.00 1.00 0.03

308
SS#10D75-1
32.26
(US#10)
75
647 3972.50 3971.00 1.50 0.04
SS#10D75-2 646 3978.00 3976.50 1.50 0.04
SS#10D75-3 649 3974.00 3972.50 1.50 0.04
SS#10D296-1
296
648 3970.00 3967.50 2.50 0.06
SS#10D296-2 648 3954.00 3951.50 2.50 0.06
SS#10D296-3 646 3967.50 3964.50 3.00 0.08
Note: ‘\’ means that entry is not applicable
Table A-10 Mechanical property degradation of stainless steel (SS) during long-term corrosion tests.
Specimen label
E E, norm fy fy, norm fm fm, norm rup rup, norm
Ori
(GPa)
Norm
Ori
(MPa)
Norm
Ori
(MPa)
Norm Ori (%) Norm
SS#3D0-1 145 0.99 513 0.98 767 0.99 29.40 1.10
SS#3D0-2 147 1.00 534 1.02 778 1.00 26.00 0.97
SS#3D0-3 149 1.01 522 1.00 782 1.01 24.80 0.93
SS#3D20-1 139 0.94 503 0.96 779 1.01 27.90 1.05
SS#3D20-2 147 1.00 514 0.98 770 0.99 26.40 0.99
SS#3D20-3 145 0.98 517 0.99 783 1.01 28.60 1.07
SS#3D75-1 144 0.98 516 0.99 787 1.01 25.00 0.94
SS#3D75-2 138 0.94 510 0.98 785 1.01 22.20 0.83
SS#3D75-3 149 1.01 532 1.02 785 1.01 24.20 0.90
SS#3D296-1 163 1.11 530 1.01 774 1.00 27.00 1.01
SS#3D296-2 163 1.11 545 1.04 771 0.99 25.30 0.95
SS#3D296-3 164 1.12 532 1.02 768 0.99 24.50 0.92
SS#5D0-1 134 0.96 542 1.01 811 1.00 20.50 1.06
SS#3D0-2 141 1.00 549 1.02 812 1.00 18.40 0.95
SS#5D0-3 147 1.04 516 0.96 806 1.00 19.30 1.00
SS#5D20-1 172 1.22 578 1.08 795 0.98 17.60 0.91
SS#5D20-2 162 1.15 599 1.12 799 0.99 18.30 0.94
SS#5D20-3 139 0.99 526 0.98 795 0.98 18.20 0.94
SS#5D75-1 130 0.92 514 0.96 809 1.00 18.70 0.97
SS#5D75-2 136 0.97 522 0.97 816 1.01 18.20 0.94
SS#5D75-3 130 0.93 513 0.96 801 0.99 19.80 1.02
SS#3D296-1 141 1.00 649 1.21 821 1.01 NA NA
SS#5D296-2 139 0.99 642 1.20 807 1.00 NA NA
SS#5D296-3 132 0.94 663 1.24 805 0.99 NA NA

309
SS#10D0-1 146 0.98 588 1.00 734 0.99 16.50 0.98
SS#10D0-2 144 0.97 604 1.02 759 1.02 16.50 0.98
SS#10D0-3 157 1.05 579 0.98 735 0.99 17.50 1.04
SS#10D20-1 369 1.13 562 0.95 736 0.99 16.80 1.00
SS#10D20-2 137 0.92 593 1.01 735 0.99 18.70 1.11
SS#10D20-3 139 0.93 597 1.01 734 0.99 16.80 1.00
SS#10D75-1 143 0.96 625 1.06 760 1.02 14.20 0.85
SS#10D75-2 NA NA NA NA NA NA NA NA
SS#10D75-3 165 1.11 591 1.00 735 0.99 17.30 1.03
SS#10D296-1 154 1.03 598 1.01 738 0.99 16.40 0.98
SS#10D296-2 159 1.07 594 1.01 735 0.99 14.30 0.85
SS#10D296-3 140 0.94 625 1.06 760 1.02 13.50 0.80
Note: 'Ori' means the original measured data; 'Norm' means the normalized data; ‘NA’ means the data is not available.

310
Appendix B - Low-cycle fatigue test results of Cu-Al-Mn and
Ni-Ti SMAs
This appendix presents the supplementary results of low-cycle fatigue tests that are not shown
in Chapter 4. Appendix B.1 to B.3 respectively provide the stress-strain curves and variation in
mechanical properties of single crystal Cu-Al-Mn SMA bar (SCB) at room temperature, 25 ℃,
low temperature, -40 ℃, and high temperature, 50 ℃. Appendix B.4 provides the stress-strain
curves and variation in mechanical properties of polycrystal Cu-Al-Mn SMA plate (PCP) at room
temperature, 25 ℃. Appendix B.5 to B.7 respectively provide the stress-strain curves and variation
in mechanical properties of Ni-Ti bar (NTB) at room temperature, 25 ℃, low temperature, -10 ℃,
and high temperature, 50 ℃.

311
B.1 SCB at room temperature, 25 ℃
(a)
(b)

312
(c)
(d)
Fig. B-1 Stress-strain curves of single crystal Cu-Al-Mn SMA bar (SCB) at room temperature,
25 °C: (a) SCB-25C-3, (b) SCB-25C-4, (c) SCB-25C-5, and (d) SCB-25C-6.

313
(a)
(b)

314
(c)
(d)
Fig. B-2 Variation in mechanical properties of SCB at room temperature, 25 °C: (a) SCB25C-3, (b) SCB-25C-4, (c) SCB-25C-5, and (d) SCB-25C-6.

315
B.2 SCB at -40 ℃
Fig. B-4 Variation in mechanical properties of SCB-m10C-3 at -40 °C.
Fig. B-3 Stress-strain curves of SCB-m10-3 at -40 °C.

316
B.3 SCB at 50 ℃
Fig. B-5 Stress-strain curves of SCB-50C-3 at 50 °C.
Fig. B-6 Variation in mechanical properties of SCB-50C-3 at 50 °C.

317
B.4 PCP at room temperature, 25 ℃
(a)
(b)
(c)
Fig. B-7 Stress-strain curves of polycrystal Cu-Al-Mn plate (PCP) at room temperature, 25
°C: (a) PCP-25C-3, (b) PCP-25C-4, and (c) PCP-25C-5.

318
(a)
(b)

319
(c)
Fig. B-8 Variation in mechanical properties of PCP at room temperature, 25 °C: (a) PCP-25C3, (b) PCP-25C-4, and (c) PCP-25C-5.

320
B.5 NTB at room temperature, 25 ℃
Fig. B-9 Stress-strain curves of NTB-25C-3 at room temperature, 25 °C.
Fig. B-10 Variation in mechanical properties of NTB-25C-3 at room temperature, 25 °C.

321
B.6 NTB at -10 ℃
Fig. B-11 Stress-strain curves of NTB-m10C-3 at -10 °C.
Fig. B-12 Variation in mechanical properties of NTB-m10C-3 at -10 °C.

322
B.7 NTB at 50 ℃
Fig. B-14 Variation in mechanical properties of NTB-50C-3 at 50 °C.
Fig. B-13 Stress-strain curves of NTB-50C-3 at 50 °C.

Abstract (if available)

Abstract Shape memory alloys (SMAs) are considered for use in bridges due to their unique superelastic and shape memory effects. The first successful implementation of SMAs in the SR99 Alaskan Viaduct Bridge in Seattle was completed in 2017. Existing research on the application of SMAs in bridges mainly focused on binary Ni-Ti compositions due to their earlier discovery. Ni-Ti SMAs have the necessary characteristics to be used as plastic hinge reinforcement in bridge columns, such as strength, ductility, superelasticity, corrosion resistance, and energy dissipation capacity. However, certain properties of the Ni-Ti SMAs, such as the difficulty in machining, potential loss of superelasticity at low temperatures, and the high cost, still drive the search for alternate materials.

The next generation SMAs considered for use in bridge applications include: Cu-Al-Mn SMA, Ni-Ti-Co SMA, and Fe-Mn-Si SMA. This dissertation performed comprehensive material characterizations on these SMAs for use in bridges under extreme environments, such as corrosion, varying ambient temperature, and earthquake loading. To benchmark each SMA composition, conventional Ni-Ti SMA and commonly used reinforcing steels, such as mild steel, stainless steel, high chromium steel, and epoxy coated steel, were tested under the same conditions.

The main research activities conducted on Cu-Al-Mn SMA, Ni-Ti-Co SMA, and Fe-Mn-Si SMA are described as follows in three parts. For each target material, research activities selected for investigation were determined based on the needs in real bridge applications and the knowledge gaps in literature. Frist, the following material characteristics of Cu-Al-Mn SMA were studied: (i) corrosion behavior, (ii) low-cycle fatigue behavior at different temperatures, (iii) machinability characteristics, (iv) headed coupling behavior, and (v) cost effectiveness when applied in typical bridge columns. Comparisons with conventional Ni-Ti SMA and commonly used reinforcing steels including mild steel, stainless steel, high chromium steel, and epoxy coated steel were made to benchmark the behavior of Cu-Al-Mn SMA. Second, the following material characteristics of Ni-Ti-Co SMA were studied: (i) effect of temperature on superelasticity, strength and ductility, (ii) low-cycle fatigue behavior at different temperatures, and (iii) moment-curvature response when applied in typical bridge columns. Comparisons with Ni-Ti SMA and Cu-Al-Mn SMA were made to benchmark the behavior of Ni-Ti-Co SMA. Third, the following material characteristics of Fe-Mn-Si SMA were studied: (i) effect of temperature on strength, ductility and recovery strain, (ii) cyclic actuation behavior, and (iii) low-cycle fatigue behavior.

The main findings on Cu-Al-Mn SMA included the following five aspects. First, Cu-Al-Mn SMA had higher corrosion resistance than mild steel but lower than conventional Ni-Ti SMA. The superelasticity, particularly the strain recovery capacity, of Cu-Al-Mn SMA showed almost no degradation after long-term corrosion. Second, when subjected 5% strain low-cycle fatigue loading, the fatigue life of Cu-Al-Mn SMA (over 50,000 cycles) could be 500 times higher than that of Ni-Ti SMA (around 100 cycles). However, it is noted that the yield strength and energy dissipation of Cu-Al-Mn SMA was lower than that of Ni-Ti SMA. Third, the machinability of Cu-Al-Mn SMA was over 20 times better than Ni-Ti SMA. The difficulty of machining Cu-Al-Mn SMA was overall close to that of stainless steel but higher than that of mild steel. Fourth, using headed coupling method to connect large diameter Cu-Al-Mn SMA with conventional steel was feasible. The key to ensure the superelasticity and ductility of headed Cu-Al-Mn SMA was to ensure a consistent cooling rate in the central and peripheral regions of the headed end after the heading process. Fifth, compared to the cost of conventional RC columns, the additional cost of Cu-Al-Mn SMA reinforced columns was only about 1/4 of the cost on Ni-Ti SMA reinforced columns.

The main findings on Ni-Ti-Co SMA included the following three aspects. First, at room temperature 23 ℃, the yield strength of Ni-Ti-Co SMA was around 2.6 times that of Ni-Ti SMA and 3.3 times that of Cu-Al-Mn SMA. The superelastic temperature range of Ni-Ti-Co SMA (-40 ℃ to 40 ℃) was wider than that of Ni-Ti SMA (0 ℃ to 50 ℃) and close to that of Cu-Al-Mn SMA (-40 ℃ to 50 ℃). It is noted that when temperature increased to 50 ℃, the stress-strain curves of Ni-Ti-Co SMA was no longer flag shaped and had slight residual strain accumulation (0.5% after unloading from 4%). Second, Ni-Ti-Co SMA has great potential for use in low-temperature seismic applications. At low temperature, the energy dissipation and strain recovery of Ni-Ti-Co SMA at -40 ℃ was even better than Ni-Ti SMA at 0 ℃. At room temperature 23 ℃ and high temperature 50 ℃, the Ni-Ti-Co SMA showed close low-cycle fatigue resistance (in terms of superelasticity degradation and fatigue life) to Ni-Ti SMA. Third, due to the high strength and large size availability of Ni-Ti-Co SMA, columns reinforced with it could have comparable flexural capacity with conventional steel reinforced concrete columns without increasing the section diameter or longitudinal reinforcement ratio.

The main findings on Fe-Mn-Si SMA included the following three aspects. First, non-activated Fe-Mn-Si SMA showed excellent deformability under a wide range of temperatures from -40 ℃ to 50 ℃. The maximum fracture strain of Fe-Mn-Si SMA at -40 ℃ reached 58%. In addition, the recovery strain of non-activated Fe-Mn-Si SMA increased as the increasing of maximum applied strain. At 23 ℃, the recovery strain of Fe-Mn-Si SMA at 5% strain was 0.83%, while at 25% strain, the recovery strain increased to 1.64%, almost doubled. Second, increasing the prestrain level could efficiently increase the post-actuation strain amplitude before the actuation stress reduces to zero. Specifically, increasing the prestrain level from 4% (commonly used in past research) to 25%, the strain amplitude when actuation stress decreased to zero showed a 110% increase. Third, increasing the prestrain level did not sacrifice the post-actuation low-cycle fatigue resistance and deformability of Fe-Mn-Si SMA. After 500 cycles of 0.5% strain fatigue loading, Fe-Mn-Si SMA with a prestrain of 15% still exhibited a fracture strain over 19% in the subsequent monotonic loading.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Next generation seismic resistant structural elements using high-performance materials

Asset Metadata

Creator Hong, Huanpeng (author)

Core Title Material characterization of next generation shape memory alloys (Cu-Al-Mn, Ni-Ti-Co and Fe-Mn-Si) for use in bridges in seismic regions

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Civil Engineering

Degree Conferral Date 2024-05

Publication Date 01/24/2024

Defense Date 01/08/2024

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

Tag actuation,Bridge,corrosion,cost estimation,fatigue,machinability,mechanical coupling,OAI-PMH Harvest,seismic,shape memory alloy,superelasticity,temperature dependence

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Gencturk, Bora (committee chair), Ershaghi, Iraj (committee member), Nweke, Chukwuebuka (committee member), Saiidi, Saiid (committee member), Wang, Qiming (committee member)

Creator Email 892505622@qq.com,huanpeng@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113810136

Unique identifier UC113810136

Identifier etd-HongHuanpe-12624.pdf (filename)

Legacy Identifier etd-HongHuanpe-12624

Document Type Dissertation

Format theses (aat)

Rights Hong, Huanpeng

Internet Media Type application/pdf

Type texts

Source 20240125-usctheses-batch-1122 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

Repository Name University of Southern California Digital Library

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

Repository Email cisadmin@lib.usc.edu