AbimbolaOladokun PhDThesis

MECHANISM OF FRETTING CORROSION AT THE MODULAR
TAPER INTERFACE OF HIP PROSTHESIS
By
Abimbola Oluwawemimo Oladokun
Submitted in accordance with the requirements for the degree of

Doctor of philosophy
The University of Leeds

School of Mechanical Engineering
Leeds, UK
September, 2017
i
The candidate confirms that the work submitted is his own, except where work
which has formed part of jointly authored publications has been included. The
contribution of the candidate and other co-authors in published work from this
thesis has been clearly indicated. The candidate confirms that appropriate
credit has been given within the thesis where reference has been made to the
work of others.
In the papers contributing to this thesis, the candidate (first author) carried out
all the experiments, analysis and preparation of the manuscripts. All other
authors contributed by proof reading and providing insight for the discussions.
This copy has been supplied on the understanding that it is copyright material
and that no quotation from this thesis may be published without proper
acknowledgement.
© 2017 The University of Leeds and Abimbola Oluwawemimo Oladokun
ii
Papers and Conference proceedings contributing to this thesis
 A. Oladokun, M. Pettersson, M. Bryant, H. Engqvist, C. Persson, R.

Hall & A. Neville (2015) Fretting of CoCrMo and Ti6Al4V alloys in
modular prostheses, Tribology - Materials, Surfaces & Interfaces, 9:4,
165-173, DOI: 10.1179/1751584X15Y.0000000014
 Oladokun, A., M. Pettersson, M. Bryant, R. Hall, and A. Neville. "SUB-

SURFACE INVESTIGATION OF FRETTED CO28CR6MO AND
TI6AL4V ALLOYS." Bone Joint J 98, no. SUPP 9 (2016): 99-99.
 Oladokun, A., M. Bryant, R. Hall, and A. Neville. "The effect of cyclic

load on the evolution of fretting current at the interface of Metal-on-
Metal and Ceramic-on-Metal taper junction of hip prostheses." Bone
Joint J 99, no. SUPP 5 (2017): 68-68.
iii
Acknowledgements
Firstly, I thank Christ Jesus for walking with me throughout this journey. I have
experienced grace, strength, specific guidance and tenacity. And therefore, I
am increasingly realising that with God, nothing shall be impossible!
I would like to appreciate my core supervisors: Prof. Anne Neville and Dr. Mike
Bryant for their support towards my development in this journey. Special
thanks to Anne for granting me the opportunity to do this PhD and to Mike for
always being available to talk about oxides and amorphous transformations.
Thanks to Prof. Richard Hall for making my experience even richer by
supporting my attendance to ISTA 2015 and 2016 conference – they served
as defining moments in my PhD journey, rekindling my passion for the
research.
Special thanks to Dr. Todd Stewart who supervised my bachelors’ research

project on Spider silk; his encouraging words inspired me to consider a PhD
without a Masters’ degree. A big ‘thank you’ to Dr. Saurabh Lal and Dr.
Nilanjan for inspiring me with their academic excellence and words of advice.
I am truly grateful to the iFS technicians that supported my research work:

Ron, Mick, Andrew, Paul and Jordan. Special appreciation to Tony Weise and
Adrian Eagles who were great support of my work from my undergraduate
days until now. I am also grateful to Mike Ward and Zabeada of LEMAS;
Maria, Ligia, Bianca, Corentin and Ogbemi for their support of this research.
I would like to appreciate all my office mates in Room 332a and Cemetery
lodge as well as the entire PhD Eindhoven football team (I claim credit for the
name). I have had 4 wonderful years spent with Andrew Barber, Joseph
Norton, William Stokes, Benjamin Pickles and Joshua Owen.
Special thanks to my Pastors: Paul and Natasha Okhuoya who supported me

throughout the 7 years I’ve been in Leeds. I am sincerely grateful to the entire
Cloverleaf family for being my family in Leeds. I would specifically like to
iv
appreciate John Ojukwu my dearest brother who also contributed to my desire
to doing a PhD.
I am so grateful to my parents who sacrificed their jobs and settlement in

Nigeria so that I and all my siblings can have a better education. My grand-
parents, aunties, uncles and in-laws, I love you all so much. Thank you for the
endless prayers; none of them fell to the ground.
And to my wife, Fiona Oladokun, who married me during this PhD journey and
supported both of us for over 1 year on her meagre salary when I had no
income; I am beyond grateful for your faith, strength and support. By the way,
these dots tends to infinity……………………………… → ∞
And to all others whose names I have not mentioned. I am sincerely grateful.
v
Abstract
Modularity of total hip arthroplasty (THA) has been linked to various forms of
adverse local tissue reaction (ALTR). ALTR is often a result of metallic
particles and ions released from corroding implant materials to the peri-
prosthetic tissues and blood stream of the human body. More so, it is a
consequence of several fretting and crevice-induced corrosion mechanisms
occurring simultaneously. Fretting corrosion and fatigue damage of the
modular taper is initiated and sustained by micromotions at the taper interface
through the multi-directional loads applied onto the prosthesis during daily
living activities.
Subsequent to the failure of the implant, analysis of retrieved explants

generally offer information regarding the mode of wear and to an extent, the
types of corrosion damage. However, retrieval studies are limited in that, they
do not provide a holistic insight into the in-vivo degradation mechanisms which
ultimately led to the early failure of the implant. On the other hand, controlled
in-vitro studies proves useful for replicating the evolution of wear and
corrosion in-situ. In addition, the role of multiple individual factors which
contribute to fretting corrosion can be elucidated through in-vitro methods.
In this study, metal – metal and ceramic – metal fretting interfaces were
investigated. Advance microscopy and spectroscopy techniques were
employed in the characterisation of passive films, corrosion products and
metallurgical transformations of CoCrMo and Ti6Al4V alloys. Both mechanical
and electrochemical data consisting of interfacial energy, open circuit potential
(OCP) and fretting corrosion currents were measured using an in-situ
tribocorrosion cell. Other surface analytical techniques were used to quantify
wear and obtain surface topography.
The results showed that CoCrMo and Ti6Al4V display independent

characteristic behaviours when in a metal – metal or ceramic – metal fretting
contact. For example, fretting contacts involving Ti6Al4V experienced higher
contact compliance than those of CoCrMo. The higher interfacial compliance
vi
thus led to a significant proportion of wear being redistributed at the contacts
involving Ti6Al4V alloys. It was subsequently observed that the redistributed
wear at the interface leads to a mixed fretting regime whereby, the contact
appear to be ‘cold-welded’ during a partial-slip regime and subsequently
transitions to a gross slip fretting regime when the ‘welded’ interfacial material
fractures. The phenomena was evident in both metal – metal and ceramic –
metal interfaces involving Ti6Al4V.
The subsurface transformation in CoCrMo alloy when subjected to fretting was

observed to be typically strain-induced twinning and loss of nano-crystalline
region. On the other hand, Ti6Al4V alloy was observed to be strain-induced
recrystallization, mechanical mixing, crack initiation and propagation.
It was also observed that the chemical composition of fretting corrosion

products at the metal – metal interfaces were dependent on the contact
condition and the specific area within the contact that the wear product is
located. For example, metal–oxides and chlorides were prominent within
creviced regions whilst precipitation of metal-phosphates and metal–oxides
were present in well aerated regions. Evidence of Cr6+ and pitting corrosion
products were also identified at the most severe creviced environment of the
CoCrMo – Ti6Al4V fretting contact.
This research also showed that the evolution of fretting corrosion current
(specifically in self-mated CoCrMo contact) can be linked to the surface history
of wear and corrosion. Furthermore, studies conducted on realistic taper
components revealed that the use of ceramic bearings in order to eliminate
one of two conductive components only resulted in the reduction of static
corrosion currents but not fretting corrosion currents. Rather, it was deduced
that interference fit at the ceramic – metal interface (relative to metal – metal)
increases the surface area that is susceptible to passive oxide abrasion under
fretting conditions. Therefore, the use of a ceramic bearing did not reduce
fretting current at the modular taper interface.
vii
viii
Table of Contents
Acknowledgements.................................................................................... iv
Abstract ....................................................................................................... vi
List of Figures ........................................................................................ xviii
List of Tables ........................................................................................ xxxiii
Nomenclature ....................................................................................... xxxiv
Chapter 1 Introduction ................................................................................ 1
1.1 Research objectives ...................................................................... 2
1.2 Thesis outline ................................................................................ 3
Chapter 2 Background to biomaterials, corrosion and tribology ........... 5
2.1 Introduction ................................................................................... 5
2.2 Metallic biomaterials ...................................................................... 5
2.2.1 Cobalt based Alloys.............................................................. 5
2.2.2 Titanium and Titanium based alloys ..................................... 7
2.2.3 Stainless steel ...................................................................... 9
2.3 Ceramic biomaterials .................................................................. 10
2.3.1 Alumina (Al2O3) and alumina composite ............................ 11
2.3.2 Silicon Nitride (Si3N4) ........................................................ 13
2.4 Corrosion..................................................................................... 14
2.4.1Electrochemistry of corrosion .............................................. 14
2.4.2 Types of corrosion effective at the modular junction of

THA ..................................................................................... 16
2.4.3Passivity of biomaterials ...................................................... 20
2.4.4 Faraday’s law ..................................................................... 21
2.5 Tribology ..................................................................................... 22
2.5.1 Wear................................................................................... 22
ix
2.5.2 Adhesive wear .................................................................... 22
2.5.3 Abrasive wear..................................................................... 23
2.5.4 Fatigue Wear ...................................................................... 24
2.5.5 Corrosive wear ................................................................... 25
2.6 Fretting ........................................................................................ 26
2.6.1 Fretting regimes ................................................................. 27
2.6.2 Fretting corrosion, wear and fatigue ................................... 31
Chapter 3 Literature review ...................................................................... 33
3.1 Total hip arthroplasty ................................................................... 33
3.1.1 Rationale for modularity in THA.......................................... 34
3.1.2 The Morse taper (head-neck taper) .................................... 36
3.1.3 The neck–stem taper .......................................................... 37
3.1.4 Assembly of modular tapers ............................................... 37
3.2 Challenges with modularity ......................................................... 38
3.2.1 Fretting and corrosion scores in retrieval studies ............... 39
3.2.2 Review of retrieval studies – head-neck interface .............. 40
3.2.3 Review of retrieval studies – the neck-stem interface ........ 43
3.2.4 Corrosion products and metal ion releases ........................ 46
3.2.5 Metallurgical transformations in explants ........................... 49
3.3 In-vitro studies of fretting corrosion ............................................. 50
3.3.1 Common in-vitro methods for assessing fretting

corrosion in THA ................................................................. 50
3.3.2 Review of tribometer studies .............................................. 52
3.3.3 Review of in-vitro studies using realistic taper

components......................................................................... 55
3.4 Effect of proteins on corrosion of biomaterials ............................ 58
3.5 Tribocorrosion ............................................................................. 59
x
3.5.1 Tribocorrosion in fretting..................................................... 62
3.5.2 Mechanically assisted crevice corrosion ............................ 63
3.5.3 Repassivation characteristics of CoCrMo and Ti6Al4V ...... 64
3.6 Summary ..................................................................................... 65
3.6.1 Gaps in literature ................................................................ 65
Chapter 4 Experimental methodology and surface analytical

techniques ......................................................................................... 67
4.1 Introduction ................................................................................. 67
4.2 Experimental materials ................................................................ 67
4.2.1 Metallic and ceramic sample preparation ........................... 67
4.2.2 Test environments and lubricants ...................................... 68
4.3 Electrochemical technique .......................................................... 68
4.3.1 Open circuit potential measurement ................................... 69
4.3.2 Potentio-static polarisation ................................................. 69
4.4 Experimental set-up .................................................................... 71
4.4.1 Fretting tribometer with in-situ electrochemistry ................. 71
4.4.2 Tribometer fretting corrosion studies (test protocol) ........... 73
4.4.3 Fatigue loading set – up with in-situ electrochemistry ........ 78
4.5 Surface, sub-surface and chemical analysis ............................... 83
4.5.1 Optical light microscopy ..................................................... 83
4.5.2 White light interferometry ................................................... 83
4.5.3 Scanning electron microscopy (SEM) ................................ 84
4.5.4 Focused Ion Beam (FIB) .................................................... 85
4.5.5 Transmission Electron Microscope (TEM) .......................... 87
4.5.6 X-Ray Powder Diffraction (XRD) Analysis .......................... 88
4.5.7 Energy Dispersed X-Ray Spectroscopy (EDS) .................. 89
4.5.8 X-Ray Photoelectron Spectroscopy (XPS) ......................... 89

xi
4.5.9 Nano-indentation hardness ................................................ 91
4.5.10 Coordinate Measurement Machine (CMM)....................... 91
Chapter 5 Metallurgy and surface chemistry of CoCrMo and Ti6Al4V

............................................................................................................ 92
5.1 Introduction ................................................................................. 92
5.2 Crystalline structure of CoCrMo and Ti6Al4V .............................. 93
5.3 Subsurface microscopy and spectroscopy .................................. 94
5.3.1 TEM and SAED of polished CoCrMo and Ti6Al4V ............. 94
5.3.2 TEM – EDS of polished CoCrMo and Ti6Al4V ................... 97
5.4 Surface chemistry...................................................................... 100
5.4.1 XPS assessment of CoCrMo passive film ........................ 100
5.4.2 XPS assessment of Ti6Al4V passive film ......................... 103
5.5 Nano–indentation hardness of polished surface ....................... 105
5.6 Discussion and summary .......................................................... 106
Chapter 6 Fretting wear mechanism of metal – metal material

combinations ................................................................................... 109
6.1 Introduction ............................................................................... 109
6.2 Tribological assessment of metal – metal fretting contacts ....... 111
6.2.1 Fretting loop analysis ....................................................... 112
6.3 Open circuit potential measurement .......................................... 115
6.4 Wear volume ............................................................................. 118
6.5 Surface Analysis........................................................................ 120
6.5.1 3D and 2D surface profilometry of fretted contacts .......... 120
6.5.2 Back scatter SEM of CoCrMo and Ti6Al4V surface ......... 125
6.6 Spectroscopy of CoCrMo and Ti6Al4V surface from the metal

– metal interface ........................................................................ 127
xii
6.6.1 EDS analysis of CoCrMo and Ti6Al4V surfaces (metal –
metal) ................................................................................ 128
6.6.2 XPS analysis of CoCrMo and Ti6Al4V surfaces (metal –

metal) ................................................................................ 131
6.7 Characterisation of metallurgical transformations (metal –

metal) ........................................................................................ 138
6.7.1 TEM of Fretted CoCrMo and Ti6Al4V alloy ...................... 138
6.7.2 TEM-EDS and SAED analysis ......................................... 143
6.8 Nano-indentation ....................................................................... 157
Chapter 7 Fretting wear mechanism of ceramic – metal material

combinations ................................................................................... 165
7.1 Introduction ............................................................................... 165
7.2 Tribological assessment of Ceramic – Metal and Ceramic –

Ceramic Fretting Contacts......................................................... 167
7.3 Open circuit potential measurement .......................................... 173
7.4 Wear volume ............................................................................. 177
7.5 Surface analysis ........................................................................ 179
7.5.1 3D and 2D surface profilometry of fretted wear surface ... 179
7.5.2 BS-SEM analysis ............................................................. 186
7.6 Spectroscopy of CoCrMo and Ti6Al4V surface from the

ceramic – metal interface .......................................................... 188
7.6.1 EDS analysis of CoCrMo and Ti6Al4V surfaces (ceramic

– metal) ............................................................................. 188
7.6.2 XPS analysis of CoCrMo and Ti6Al4V surface (ceramic

– metal) ............................................................................. 190
xiii
7.7 Characterisation of metallurgical transformations (ceramic –
metal) ........................................................................................ 196
7.8 Nano-indentation ....................................................................... 204
7.9 Fretting of Biolox Ceramic on CoCrMo and Ti6Al4V alloy ......... 205
7.9.1 Fretting wear mechanism of Biolox – CoCrMo ................. 206
7.9.2 Fretting wear mechanism of Biolox – Ti6Al4V (mixed

regime) .............................................................................. 210
7.9.3 Fretting wear mechanism of Biolox – Ti6Al4V (gross slip

regime) .............................................................................. 214
Chapter 8 Quantification of fretting corrosion current using self-

mated CoCrMo material combination ............................................ 224
8.1 Introduction ............................................................................... 224
8.2 Anodic current in static condition ............................................... 226
8.3 Method 1 ................................................................................... 227
8.3.2 Electrochemical results .................................................... 229
8.3.3 Surface analysis ............................................................... 230
8.3.4 Tribocorrosion mass loss.................................................. 231
8.4 Method 2 ................................................................................... 234
8.4.4 Tribocorrosion mass loss.................................................. 238
8.5 Method 3 ................................................................................... 239
xiv
8.5.4 Tribocorrosion mass loss ................................................. 242
Chapter 9 Assessment of fretting current from the head-neck metal-

metal and ceramic-metal taper interface ....................................... 248
9.1 Introduction ............................................................................... 248
9.2 Electrochemical results ............................................................. 253
9.2.1 Open Circuit Potential ...................................................... 253
9.2.2 Static Anodic Current Transient ....................................... 255
9.2.3 Frequency variation .......................................................... 256
9.2.4 Mean load variation .......................................................... 258
9.2.5 Cyclic load variation ......................................................... 260
9.2.6 Systemic checks .............................................................. 261
9.3 Mechanical results .................................................................... 263
9.4 Surface Analysis ....................................................................... 266
9.4.1 CoCrMo – CoCrMo........................................................... 266
9.4.2 CoCrMo – Ti6Al4V ........................................................... 267
9.4.3 Biolox – CoCrMo .............................................................. 268
9.4.4 Biolox – Ti6Al4V ............................................................... 270
9.5 The case of misalignment ......................................................... 271
Chapter 10 Overall discussion ............................................................... 276
10.1 Introduction ............................................................................... 276
10.2 Linking fretting loops with the fretting interface ......................... 276
10.2.1 Fretting loop deviations at the self-mated CoCrMo

interface ............................................................................ 278
10.2.2 Fretting loop deviations at ceramic – metal interface ..... 280
xv
10.2.3 Fretting loop deviations at CoCrMo – Ti6Al4V interface
.......................................................................................... 282
10.2.4 Summary ........................................................................ 284
10.3 Aspects of fretting and corrosion mechanisms .......................... 284
10.3.1 Mechanism of cold-welding and pseudo-amorphous

structures at CoCrMo – Ti6Al4V interface ......................... 284
10.3.2 Mechanism of stiction at the Si3N4 – Ti6Al4V interface .. 290
10.4 Characteristics of biomedical materials at the modular taper

junction of hip implants .............................................................. 293
10.4.1 CoCrMo at the modular taper interface .......................... 293
10.4.2 Ti6Al4V at the modular taper interface ........................... 295
10.4.3 Si3N4 at the modular taper interface .............................. 295
10.4.4 Biolox at the modular taper interface .............................. 297
10.5 Relevance to clinical application ............................................... 298
Chapter 11 Conclusions and Future work............................................. 301
11.1 Conclusions ............................................................................... 301
11.1.1 Material characterisation, metal – metal and ceramic –

metal systems (Chapters 5 – 7) ........................................ 301
11.1.2 Quantification of fretting corrosion current on self-

mated CoCrMo (Chapter 8) ............................................... 303
11.1.3 Component testing of metal – metal and ceramic –

metal tapers (Chapter 9) ................................................... 304
11.2 Future work ............................................................................... 305
xvi
References ............................................................................................... 306
Appendix A– Composition of FBS and PBS ......................................... 334
Appendix B– Determination of contact compliance............................. 336
Appendix C – Interactions between energy, wear and oxidation ........ 340
Appendix D – List of oral presentations................................................ 341
xvii
List of Figures
Figure 2-1 – Generations of alumina ceramic [13]....................................... 12
Figure 2-2 Mechanism of improved toughness in composite alumina

ceramic [15]......................................................................................... 12
Figure 2-3 – Electrical interactions at the metallic surface a) Schematic

diagram of electric double layer [23] b) Anodic and Cathodic sites
during corrosion [21]............................................................................ 16
Figure 2-4 – Intergranular attack on CoCrMo alloy [27] ............................... 17
Figure 2-5 – Pitting corrosion attack on CoCrMo alloy [27]. ........................ 18
Figure 2-6 – A simplified 3 stage depiction of crevice corrosion: a)

conducting electrolyte in contact with crevice. b) cathodic and
anodic sites are formed, c) corrosion products create a severer
crevice [21]. ......................................................................................... 19
Figure 2-7 – Theoretical plot of a potentio-dynamic scan in the anodic

direction for a passive alloy [32] .......................................................... 21
Figure 2-8 – Adhesive wear mechanism [35] .............................................. 23
Figure 2-9 – Abrasive wear mechanism [35] ............................................... 24
Figure 2-10 – Fatigue wear mechanism [35] ............................................... 24
Figure 2-11 – Corrosive wear mechanism [35] ............................................ 25
Figure 2-12 – Difference between a fretting contact (left) and a

reciprocating sliding contact (right). Adapted from [39] ....................... 27
Figure 2-13 – Fretting hysteresis loops from a plot of tangential force and
fretting displacement amplitude [41], beneath the fretting regimes
are the characteristic surface damage profiles [38]. ............................ 28
Figure 2-14 – Fretting regimes: a) characterisation of the fretting contact

of a partial-slip regime in terms of tangential force (Ft ) and traction
force (qx) [42]....................................................................................... 29
Figure 2-15 – Schematic representation of Mindlin’s hypothesis................. 30
xviii
Figure 2-16 – Fretting damage modes: a) wear and fatigue modes as a
function of displacement, b) wear and fatigue modes as a function
of load [38]. ......................................................................................... 32
Figure 3-1 – Percentage of bearing combinations a) cemented b)

uncemented [1]. .................................................................................. 34
Figure 3-2 – Evolution of modularity in THA ................................................ 35
Figure 3-3 – Increasing bearing sizes from left to right and decreasing
taper sizes in the same direction [57]. ................................................. 37
Figure 3-4 – Fixation of both head–neck and neck–stem [62] ..................... 38
Figure 3-5 – Fretting corrosion at the head-neck interface: a) evidence

of Ti-Cr-Mo interfacial layer [27] b) evidence of corrosion products
on the trunnion surface [67]. ............................................................... 41
Figure 3-6 – Fracture of the Ti6Al4V trunnion resulting from corrosion

[74] ...................................................................................................... 42
Figure 3-7 – Wear pattern commonly observed at the head-neck taper

interface: Assymetric and Axisymmetric [75]. ...................................... 43
Figure 3-8 – Neck components of bi-modular tapers a) distribution of

stresses across the neck component at the head-neck and neck-
stem interfaces [77] b) Corrosion at the neck-stem interface of a
circular neck design [79]. .................................................................... 44
Figure 3-9 – Fracture of the neck component at the neck-stem interface

[81] ...................................................................................................... 46
Figure 3-10 – Solid corrosion products in retrieval. The region ‘A’ is

typically mixed metal-oxides and chlorides while the region ‘B’ is
typically metal-phosphates [87]. .......................................................... 47
Figure 3-11 – Subsurface transformation in the partial-slip region of a)

CoCrMo and b) TMZF [98] .................................................................. 50
Figure 3-12 – Fretting tribometer instrumented with in-situ

electrochemistry for the assessment of fretting corrosion [99] ............ 51
xix
Figure 3-13 – Fatigue assembly of the Morse taper interface
instrumented with in-situ electrochemistry [101]. ................................. 52
Figure 3-14 – Considerations of a tribocorrosion systems [137].................. 60
Figure 3-15 – A model of tribocorrosion system in a sliding contact [138]

............................................................................................................ 61
Figure 3-16 – Stack and Chi’s criteria of degradation mechanism in a

tribocorrosion system [136]. ................................................................ 62
Figure 4-1 – Schematic representation of a tribocorrosion current – time

plot. ..................................................................................................... 70
Figure 4-2 – Schematic diagram of the fretting tribometer .......................... 71
Figure 4-3 – Image of the fretting corrosion tribometer set-up .................... 72
Figure 4-4 – Schematic diagram of fretting corrosion protocol under OCP

conditions ............................................................................................ 75
Figure 4-5 – Schematic diagram of potentio-static fretting corrosion

protocol ............................................................................................... 77
Figure 4-6 – Taper assembly: a) Dimension of spigot and surface profile

of the grooved taper, b) assembly and pull-off method, c) an
example of the taper separation force. ................................................ 79
Figure 4-7 – 3D representation of the fatigue set-up for assessing fretting

corrosion of realistic taper geometry. .................................................. 80
Figure 4-8 – Protocol for fatigue loading tests on realistic taper

components......................................................................................... 82
Figure 4-9 – Schematic for determining wear volume ................................. 84
Figure 4-10 – Elastic contact stress pattern for a ball-on- flat contact [42]
............................................................................................................ 86
Figure 4-11 – FIB SEM micrographs of TEM sample preparation stages

(a) Pt deposited on area of interest (b) removal of material from
either side of Pt deposit (c) removal of lamella from bulk material
(d) attachment of lamella to TEM post (e) final thinning of sample
to electron transparency ...................................................................... 87
xx
Figure 4-12 – Schematic diagram of atoms in planes separated by a
space ‘d’ as incident X-ray of wavelength λ diffracts at angle ‘θ’. ........ 88
Figure 5-1 – Outline of analyses performed in chapter 5............................. 92
Figure 5-2 – Crystalline structure of polished CoCrMo alloy ....................... 93
Figure 5-3 Crystalline structure of polished Ti6Al4V alloy ........................... 94
Figure 5-4 – TEM and SAED of polished CoCrMo: a) Bright-field

micrograph. b) Dark-field micrograph c) SAED of nano-crystalline
region ‘A’ d) SAED of single crystal ‘B’ ................................................ 96
Figure 5-5 – TEM and SAED of Polished Ti6Al4V: a) Bright-field

region ‘A’ d) SAED of single crystal ‘B’ ................................................ 97
Figure 5-6 – TEM–EDS map of polished CoCrMo showing Co, Cr, Mo,
O. ........................................................................................................ 98
Figure 5-7 – TEM–EDS map of polished Ti6Al4V showing Ti, Al, V, O. ...... 98
Figure 5-8 – XPS Depth profile analysis of CoCrMo passive film .............. 101
Figure 5-9 – Proportion of species in CoCrMo protective layer: a) O

species, b) Cr species, c) Mo species ............................................... 102
Figure 5-10 – XPS Depth profile analysis of Ti6Al4V passive film ............ 103
Figure 5-11 – Proportion of species in Ti6Al4V protective layer: a) O

species, b) Ti species, c) Al species, d) V species ............................ 104
Figure 5-12 – Nano-indentation hardness of CoCrMo and Ti6Al4V a) P

vs. h plot for CoCrMo b) P vs. h plot for Ti6Al4V c) calculated
hardness of CoCrMo and Ti6Al4V..................................................... 106
Figure 5-13 – Schematic diagram of a three-layer surface structure of

CoCrMo and Ti6Al4V. Surface contaminants are not included in the
three layers. ...................................................................................... 108
Figure 6-1 – Outline of the fretting experiments and analyses performed

in chapter 6. C and T represents CoCrMo and Ti6Al4V respectively.
The letters that appear in red represent the samples analysed. ....... 110
xxi
Figure 6-2 – Slip ratio of CoCrMo – CoCrMo and CoCrMo – Ti6Al4V
under OCP conditions from ±10 to ±150 µm displacement................ 111
Figure 6-3 – Fretting of CoCrMo – CoCrMo: a) Fretting loop b) Energy

dissipated per unit cycle. ................................................................... 113
Figure 6-4 – Fretting of CoCrMo – Ti6Al4V: a) Fretting loop b) Energy

dissipated per unit cycle. ................................................................... 114
Figure 6-5 – Open circuit potential of CoCrMo – CoCrMo: a) OCP plot

for ±10 µm, ±25 µm, ±50 µm and ±150 µm. b) Max. cathodic shift
of all displacements. .......................................................................... 116
Figure 6-6 – Open circuit potential of CoCrMo – Ti6Al4V: a) OCP plot for
±10 µm, ±25 µm, ±50 µm and ±150 µm. b) Max. cathodic shift of all
displacements. .................................................................................. 118
Figure 6-7 – Sum of ball and flat wear volume and redistributed volume
for CoCrMo – CoCrMo fretting contact. ............................................. 119
for CoCrMo – Ti6Al4V fretting contact. .............................................. 120
Figure 6-9 – Surface profilometry of CoCrMo ball and CoCrMo flat for all
four displacements under OCP conditions: a) 3D surface profile, b)
2D cross-section of the vertical axis of the wear surface................... 122
Figure 6-10 – Surface profilometry of CoCrMo ball and Ti6Al4V flat for
all four displacements under OCP conditions: a) 3D surface profile,
b) 2D cross-section of the vertical axis of the wear surface .............. 124
Figure 6-11 – Back scatter-SEM of CoCrMo flat from the CoCrMo –

CoCrMo interface at ±50 µm under OCP conditions. ........................ 126
Figure 6-12 – Back scatter-SEM of Ti6Al4V flat from the CoCrMo –

Ti6Al4V interface at ±50 µm under OCP conditions. ......................... 127
Figure 6-13 – SEM-EDS of CoCrMo flat at ±50 µm from CoCrMo –

CoCrMo interface under OCP conditions. The inset table outlines
the composition of a specific third-body region. ................................ 129
xxii
Figure 6-14 – SEM-EDS of Ti6Al4V flat at ±50 µm from CoCrMo –
Ti6Al4V interface under OCP conditions. .......................................... 130
Figure 6-15 – XPS analysis of wear and corrosion product from the wear
surface of CoCrMo – CoCrMo combination at ±50 µm under OCP
conditions. ......................................................................................... 132
surface of CoCrMo – Ti6Al4V combination at ±50 µm. Creviced
region – Region ‘C’............................................................................ 135
surface of CoCrMo – Ti6Al4V combination at ±50 µm under OCP
conditions. Region of mechanically induced crack – region ‘D’. ........ 137
Figure 6-18 – TEM bright and dark field micrographs of CoCrMo flat from
CoCrMo – CoCrMo interface at ±10 µm (a & b), ±25 µm (c & d),
±50 µm (e & f), ±150 µm (g & h). ....................................................... 141
Figure 6-19 – TEM bright and dark field micrographs of Ti6Al4V flat
component from CoCrMo – Ti6Al4V interface at ±10 µm (a & b),
±25 µm (c & d), ±50 µm (e & f), ±150 µm (g & h). ............................. 143
Figure 6-20 – TEM and SAED of CoCrMo – CoCrMo at ±150 µm: a)

Bright field TEM micrograph of CoCrMo subsurface b) Close-up
image of tribochemical products on surface c) SAED of structure 1
d) SAED of structure 2 e) SAED of CoCrMo bulk............................ 145
Figure 6-21 – TEM-EDS map of CoCrMo – CoCrMo flat surface at ±150

µm ..................................................................................................... 146
Figure 6-22 – TEM and EDS analysis of the transferred third-body at the
interface of CoCrMo – CoCrMo, ±150 µm. ........................................ 148
Figure 6-23 – TEM and SAED of mechanically mixed layers on Ti6Al4V

alloy: a) TEM micrograph of Ti6Al4V subsurface b) high-resolution
micrographs of the subsurface structures c) SAED of the
mechanically mixed layer d) SAED of mechanically mixed Ti6Al4V
structures e) SEAD of Ti6Al4V nano-crystalline structure. ................ 150
xxiii
Figure 6-24 – TEM-EDS map of the mechanically mixed layer and
transformed Ti6Al4V alloy from CoCrMo – Ti6Al4V interface............ 151
Figure 6-25 – TEM and SAED of Ti6Al4V surface (mechanically cracked

region): a) TEM micrograph of Ti6Al4V subsurface b) high
resolution image of the region in red dotted square c) SAED of the
nano-crystalline mesh d) SAED of a single large particle. ................. 153
Figure 6-26 – TEM-EDS map of the Ti6Al4V flat at the mechanically

induced crack region of the CoCrMo – Ti6Al4V interface. ................. 154
Figure 6-27 – TEM and SAED of CoCrMo ball surface: a) Dark field TEM
micrograph of CoCrMo subsurface b) Bright field image c) SAED of
the transferred material d) SAED of near-surface CoCrMo bulk. ...... 156
Figure 6-28 – TEM-EDS analysis of the CoCrMo ball surface from the
CoCrMo – Ti6Al4V interface.............................................................. 156
Figure 6-29 – Nano-indentation of layer formed at the CoCrMo – Ti6Al4V

interface: a) image of the indented cross-section b) P vs. h graph
of indentations c) average hardness of the layer. .............................. 157
Figure 6-30 – Comparing Max. cathodic shift of the two metal – metal
material combinations. ...................................................................... 159
Figure 6-31 – Schematic diagram of fretting corrosion products and SEM

micrographs of CoCrMo and Ti6Al4V from the CoCrMo – Ti6Al4V
interface at ±50 µm displacement...................................................... 164

in chapter 7 for Si3N4 ceramic. C and T represents CoCrMo and
Ti6Al4V respectively. The letters that appear in red represent the
samples analysed. The green ticks represent the analysis
performed on self-mated Si3N4. ......................................................... 166

in chapter 7 for Biolox ceramic. ......................................................... 167
Figure 7-3 – Slip ratio of Si3N4 – CoCrMo, Si3N4 – Ti6Al4V and Si3N4 –
Si3N4 under OCP conditions for ±10, ±25 ±50 µm displacements. ... 168
xxiv
Figure 7-4 – Fretting contact analysis for Si3N4 – CoCrMo: a) Fretting
loop b) Energy dissipated per cycle. ................................................. 169
Figure 7-5 – Fretting contact analysis for Si3N4 – Ti6Al4V: a) Fretting

loop b) Energy dissipated per cycle. ................................................. 171
Figure 7-6 – Fretting contact analysis for Si3N4 – Si3N4: a) Fretting loop
b) Energy dissipated per cycle. ......................................................... 173
Figure 7-7 – Open circuit potential of Si3N4 – CoCrMo: a) OCP plot for
±10 μm, ±25 μm, ±50 μm. b) Max. cathodic shift for all
displacements. .................................................................................. 175
Figure 7-8 – Open circuit potential of Si3N4 – Ti6Al4V: a) OCP plot for
±10 μm, ±25 μm, ±50 μm. b) Max. cathodic shift for all
displacements. .................................................................................. 176
Figure 7-9 – Wear and redistributed volume from the Si3N4 – CoCrMo
fretting contact. A summation of both ball and flat components. ....... 177
Figure 7-10 – Wear and redistributed volume from the Si3N4 – Ti6Al4V
Figure 7-11 – Wear and redistributed volume from the Si3N4 – Si3N4
Figure 7-12 – Surface profilometry of Si3N4 ball and CoCrMo flat

surfaces: a) 3D surface profile of all displacements, b) 2D cross-
section of the vertical axis of the wear surface. ................................. 181
Figure 7-13 – Surface profilometry of Si3N4 ball and Ti6Al4V flat surfaces:
a) 3D surface profile of all displacements, b) 2D cross-section of
the vertical axis of the wear surface. ................................................. 183
Figure 7-14 – Surface profilometry of Si3N4 ball and Si3N4 flat surfaces:
a) 3D surface profile of all displacements, b) 2D cross-section of
the vertical axis of the wear surface. ................................................. 185
Figure 7-15 – Back scatter-SEM micrograph of CoCrMo flat from the

fretting against Si3N4 ball at ±50 μm. ................................................. 186
xxv
Figure 7-16 – Back scatter-SEM micrograph of Ti6Al4V flat fretted
against Si3N4 ball at ±50 μm. ............................................................. 187
Figure 7-17 – SEM-EDS of CoCrMo flat fretted against Si3N4 ball at ±50
μm. .................................................................................................... 188
Figure 7-18 – SEM-EDS of Ti6Al4V flat fretted against Si3N4 ball at ±50
μm. .................................................................................................... 189
surface of Si3N4 – CoCrMo combination at ±50 μm. Si&O-rich third-
body product on CoCrMo. ................................................................. 191
surface of Si3N4 – Ti6Al4V combination at ±50 μm. Mechanically
mixed layer on Ti6Al4V – Region ‘C’. ................................................ 193
surface of Si3N4 – Ti6Al4V combination at ±50 μm. Ti6Al4V surface
which was once covered by the mechanically mixed layer – Region
‘D’ ...................................................................................................... 195
Figure 7-22 – TEM and SAED of CoCrMo from the Si3N4 – CoCrMo
combination: a) TEM micrograph of CoCrMo subsurface b) high
resolution image of the uppermost subsurface c) SAED of Si and
O-rich material d) SAED of CoCrMo single-crystal. ........................... 197
Figure 7-23 – TEM-EDS map of the CoCrMo subsurface from the Si3N4
– CoCrMo material combination. Above the white-line are
deposited Pt and empty space. ......................................................... 198
Figure 7-24 – TEM and SAED of Ti6Al4V from the Si3N4 – Ti6Al4V
combination (region of mechanically mixed layer): a) TEM
micrograph of the Ti6Al4V subsurface b) high resolution image of
the mechanically mixed layer c) high resolution image of the
transformed amorphous Ti6Al4V structure d) SAED of the
mechanically mixed layer e) SAED of the suspending large particle
f) SAED of fractured region. .............................................................. 200
xxvi
Figure 7-25 – TEM-EDS map of the Ti6Al4V subsurface from the Si3N4
– Ti6Al4V material combination – region of mechanically mixed
layer. ................................................................................................. 201
combination (region absent of mechanically mixed layer): a) TEM
micrograph of CoCrMo subsurface b) high resolution image of the
nano-crystalline region c) SAED of the uppermost region d) SAED
of the totally transformed region e) SAED of nano-crystalline area. .. 203
Figure 7-27 – TEM-EDS map of the Ti6Al4V subsurface from the Si3N4
– Ti6Al4V material combination – region absent of mechanically
mixed layer. ....................................................................................... 204
Figure 7-28 – Nano-indentation of layer formed at the CoCrMo – Ti6Al4V

interface: a) P vs. h graph of indentations b) average hardness of
the layer. ........................................................................................... 205
Figure 7-29 – Fretting of Biolox – CoCrMo material combination at ±50

µm, Pmax = 1 GPa: a) 2D fretting loop b) Energy dissipated per
cycle c) OCP measurement. ............................................................. 207
Figure 7-30 – 3D and 2D surface profilometry of Biolox – CoCrMo at ±50

µm displacement. .............................................................................. 208
Figure 7-31 – SEM-EDS map of Biolox – CoCrMo flat surface at ±50 µm

a) BS-SEM micrograph b) SEM – EDS ............................................. 209
Figure 7-32 – Fretting of Biolox – Ti6Al4V material combination at ±50

µm (Pmax = 1 GPa): a) 2D fretting loop b) Energy dissipated per
cycle c) OCP measurement. ............................................................. 211
Figure 7-33 – 3D and 2D surface profilometry of Biolox – Ti6Al4V at ±50

µm displacement (Pmax = 1 GPa). ................................................... 212
Figure 7-34 – SEM-EDS map of Biolox – Ti6Al4V flat surface at ±50 µm

(1 GPa) a) BS-SEM micrograph b) SEM – EDS................................ 214

µm (0.77 GPa): a) 2D fretting loop b) Energy dissipated per cycle
c) OCP measurement. ...................................................................... 215
xxvii
µm displacement (Pmax = 0.77 GPa). .............................................. 216

(0.77 GPa) a) BS-SEM micrograph b) SEM – EDS ........................... 218
Figure 7-38 – Comparing Max. cathodic shift in ceramic – metal

combinations. .................................................................................... 223
Figure 8-1 – Outline of the fretting experiments performed in chapter 8 ... 225
Figure 8-2 – Slip ratio of CoCrMo – CoCrMo under potentio-static

conditions for ±10 – ±150 µm in ‘Methods 1 – 3’. .............................. 226
Figure 8-3 – Anodic current transient of loaded CoCrMo – CoCrMo

contact subjected to 0V vs. Ag/AgCl of potentio-static polarisation
for 4000 seconds in static conditions. ............................................... 227
Figure 8-4 – Method 1: fretting loop analysis of CoCrMo – CoCrMo at

±10 µm, ±25 µm, ±50 µm, ±150 µm displacement in potentio-static
conditions. a) Fretting loop b) Energy dissipated per cycle. .............. 228
Figure 8-5 – Method 1: fretting current from CoCrMo – CoCrMo contact

at ±10 µm, ±25 µm±, 50 µm, ±150 µm displacement, a) Fretting
current transient b) Cumulative charge ............................................. 230
Figure 8-6 – Method 1: 3D surface profilometry of CoCrMo ball and flat

fretting interface assessed under potentio-static conditions. ............. 231
Figure 8-7 – Method 1: the sum of tribocorrosion mass loss from ball and
flat components; a) Wear volume (blue) and Re-distributed volume
(orange); b) Mass loss as a proportion of Mmech and Mchem, c) Total
mass loss vs. Cumulative energy. ..................................................... 233


at ±10 µm, ±25 µm±, 50 µm, ±150 µm displacement; a) Fretting
xxviii
fretting contacts tested in potentio-static conditions. ......................... 238
Figure 8-11 – Method 2: tribocorrosion mass loss; a) Wear volume; b)

Mass loss as a proportion of Mmech and Mchem. .................................. 238


at ±150 µm, ±50 µm, ±25 µm, ±10 µm displacement, a) Fretting

fretting contacts tested in potentio-static conditions. ......................... 242

Mass loss as a proportion of Mmech and Mchem. .................................. 243
Figure 8-16 – The sum of Mmech and Mchem for all displacements in the
three methods of assessment. The Mchem/Mmech ratio for each
method is also shown above. ............................................................ 247
Figure 9-1 – Outline of the fatigue experiments and analyses performed

in chapter 9. B, C and T represents Biolox, CoCrMo and Ti6Al4V
respectively. ...................................................................................... 249
Figure 9-2 – Fretting current vs. time plot for a complete fatigue loading
test. ................................................................................................... 252
Figure 9-3 – OCP of all combinations prior to potentio-static polarisation

of 0V vs. Ag/AgCl, a) OCP vs. Time plot b) average OCP and
applied over-potential. ....................................................................... 254
Figure 9-4 – Anodic current transient of all material combinations under

static conditions: a) CoCrMo – CoCrMo; b) CoCrMo – Ti6Al4V; c)
Biolox – CoCrMo; d) Biolox – Ti6Al4V; e) average current at the
time t = 7500 Sec and f) current at the time t = 0 Sec (Imax). ............. 256
xxix
Figure 9-5 – Fretting current (a, b, c, d) and Cumulative charge per
10000 cycles (e, f, g, h) of the material combinations; these were
subjected to ±1.5 kN cyclic load and mean load of 1.8 kN with varied
frequencies of 1, 2 and 5 Hz. ............................................................ 258
Figure 9-6 – Fretting current (a, b, c, d) and Cumulative charge (e, f, g,

h) of the material combinations; these were subjected to a cyclic
load of ±0.5 kN at 1 Hz with varied mean load of 0.8, 1.8 and 2.8
kN. ..................................................................................................... 259

h) of the material combinations; these were subjected to mean load
of 2.8 kN at 1 Hz with varied cyclic load of ±0.5, ±1.5 and ±2.5 kN. .. 261

h) of the material combinations; these were subjected to cyclic load
of ±1.5 kN and mean load of 1.8 kN for 5000 seconds at 1 Hz. ........ 263
Figure 9-9 – Cross-head displacement vs. Time graph. Each data point
for all material combinations during the dynamic stages represents
a point every 1000 seconds of the cross-head position. At the top
right corner, a schematic of the cross-head displacement is shown.
.......................................................................................................... 265
Figure 9-10 – Taper separation force for all material combinations ........... 266
Figure 9-11 – 3D reconstruction of CoCrMo – CoCrMo taper interface:

a) CoCrMo female taper, b) CoCrMo male taper. ............................. 267
Figure 9-12 – 3D reconstruction of CoCrMo – Ti6Al4V taper interface: a)

CoCrMo female taper, b) Ti6Al4V male taper. .................................. 268
Figure 9-13 – 3D reconstruction of Biolox – CoCrMo taper interface: a)

Biolox female taper, b) Image of the Biolox female taper, c) Full 3D
reconstruction of CoCrMo male taper d) Customised reconstruction
of CoCrMo male taper. ...................................................................... 269
xxx
Figure 9-14 – 3D reconstruction of Biolox – Ti6Al4V taper: a) Biolox
female taper, b) Image of the Biolox taper c) Full 3D reconstruction
of Ti6Al4V male taper, d) Customised reconstruction of Ti6Al4V
male taper. ........................................................................................ 270
Figure 9-15 – Fretting currents of Biolox – Ti6Al4V three repeats, a)

frequency variation, b) mean load variation c) cyclic load variation... 272
Figure 9-16 – 3D reconstruction of Biolox – Ti6Al4V taper – the case of

misalignment: a) BIOLOX female taper, b) Ti6Al4V male taper. ...... 272
Figure 10-1 – Fretting loop of gross slip regime; a) labelled fretting loop
b) de-convoluted fretting loop – displacement (solid-line) and
tangential force (dashed-line). ........................................................... 278
Figure 10-2 – Non-symmetrical fretting contact a) CoCrMo – CoCrMo

±50 µm fretting loop b) ball and flat wear surface .............................. 279
Figure 10-3 – Third-body grooving a) CoCrMo – CoCrMo, ±150 µm

fretting loop b) ball and flat wear surface .......................................... 280
Figure 10-4 – Influence of third-body a) Si3N4 – CoCrMo, ±50 µm fretting

loop b) BIOLOX – Ti6Al4V, ±50 µm fretting loop c) Si3N4 – CoCrMo
ball and flat wear surface d) BIOLOX – Ti6Al4V ball and flat wear
surface. ............................................................................................. 282
Figure 10-5 – The plunging effect (prior to fracture of cold-welded

interface): a) CoCrMo – Ti6Al4V, ±50 µm fretting loop, b)
schematics of the plunging effect [221]. ............................................ 283
Figure 10-6 – Slip within localised fretting grooves (after the fracture of
cold-welded interface): a) CoCrMo – Ti6Al4V, ±50 µm fretting loop,
b) 3D surface of the ruptured welded interface. ................................ 284
Figure 10-7 – Mechanism of cold-welding at the CoCrMo – Ti6Al4V

interface ............................................................................................ 288
Figure 10-8 – Mechanism of pseudo-amorphous structure of CoCrMo

and amorphisation of Ti6Al4V alloys. ................................................ 289
Figure 10-9 – Mechanism of stiction at the Si3N4 – Ti6Al4V interface ....... 292
xxxi
Figure 10-10 – Schematic diagram of different fretting corrosion products
at specific regions within the composite contact of self-mated
CoCrMo, ±50 µm. .............................................................................. 294
Figures 11-1 – Composition of FBS [231] ................................................. 334
Figures 11-2 – Composition of PBS .......................................................... 335
Figures 11-3 – Quasi-stick fretting loop of (a) CoCrMo – CoCrMo and (b)
CoCrMo – Ti6Al4V tribo-couples ....................................................... 337
xxxii
List of Tables
Table 2-1 – Chemical composition (in wt%) and recommendation of
mechanical properties for wrought CoCrMo alloys according to the
American Society for Testing and Materials (ASTM)............................. 7

mechanical properties for unalloyed Titanium and Extra Low
Interstitial (ELI) Ti6Al4V alloy according to the ASTM ........................... 9

mechanical properties for 316L stainless steel according to ASTM.
............................................................................................................ 10
Table 3-1 – Criteria for corrosion and fretting scores [67] ........................... 39
Table 4-1 Fretting contact parameters ........................................................ 74
Table 5-1 – TEM – EDS chemical composition of CoCrMo ......................... 99
Table 5-2 – TEM – EDS chemical composition of Ti6Al4V.......................... 99
Table 6-1 – Attribution of Cr and O species from CoCrMo – CoCrMo

interface at ±50 µm ........................................................................... 133
Table 6-2 – Attribution of Cr, Co and O species from CoCrMo – Ti6Al4V

interface at ±50 µm. Creviced region – Region ‘C’ ........................... 135
Table 6-3 – Attribution of Ti and O species from CoCrMo – Ti6Al4V

interface at ±50 µm. Region of mechanically induced crack – region
‘D’ ...................................................................................................... 138
Table 7-1 – Attribution of Si, N and O species from Si3N4 – CoCrMo at

±50 μm. Si&O-rich third-body product on CoCrMo. ........................... 191
Table 7-2 – Attribution of Si, Ti, Al, N and O species from Si3N4 – Ti6Al4V
at ±50 μm. Mechanically mixed layer on Ti6Al4V – Region ‘C’ ......... 194
Table 8-1 – Breakdown of static current, Faradaic mass loss for both the
entire surface sample surface and the fretting surface...................... 234
Table 10-1 – Electrode potentials of Si, Cr, Ti and Al vs. SHE [2, 21,
230]. The electrons transferred in the half-cell reaction are omitted
from the left column for illustrative purposes. .................................... 297
xxxiii
Nomenclature
Terms Definition Units
𝜶𝑪𝒘 𝒂𝒏𝒅 𝜷𝑪𝒘 Unknown proportions of C
corrosion enhanced wear
a Hertzian contact width µm
𝜶 Head – Neck angle °
A Sliding – Displacement ratio
𝜷 Adduction angle °
C Pure corrosion C
𝜹𝒔 Interfacial slip µm
𝜹𝒙 Displacement amplitude µm
𝜹𝒆 Elastic displacement µm
e Fretting/reciprocating criteria
𝑬𝒑𝒑 Passivating potential V
ε-Co HCP Co phase
γ-Co FCC Co phase
F Faraday CMol-1
𝑭𝒕 /𝑭𝒙 Tangential force N
h depth nm
𝑰𝒂 Anodic current Amps
𝑰𝒄 Cathodic current Amps
𝑰𝒑𝒂𝒔𝒔 Passivating current Amps
𝑰𝒃𝒂𝒔𝒆 Baseline current Amps
𝑰𝒑𝒆𝒂𝒌 Peak current Amps
𝑰𝒄𝒓𝒊𝒕 Critical current Amps
𝑰𝒎𝒂𝒙 Maximum current Amps
M Molar mass gmol-1
𝑴𝒕𝒐𝒕𝒂𝒍 Total mass loss µg
𝑴𝒄𝒉𝒆𝒎 Chemical mass loss µg
𝑴𝒎𝒆𝒄𝒉 Mechanical mass loss µg
n Number of electrons
P load N
𝑷𝒎𝒂𝒙 Maximum contact pressure GPa
𝒒𝒙 Traction N
Q Charge C
𝑹𝒂 Roughness profile nm/ µm
𝑹𝒑 Profile peak nm/ µm
𝑹𝒗 Profile valley nm/ µm
𝑹𝒛 Average maximum height nm/ µm
t time s
At% Atomic percent
Wt% Weight percent
xxxiv
Chapter 1 Introduction
Total hip arthroplasty (THA) is known to be one of the most successful surgical
interventions of the 21st century. In most cases, the prosthetic device has
restored function to the elderly and young by alleviating pain and discomfort
brought about by arthritis and other diseases. Across a time period of 12
years, there has been over 800 000 primary hip operations performed in the
UK alone; and this is expected to increase on a yearly basis due to an aging
population [1].
The hip prosthesis has undergone several stages of evolution since the
1950s; one of which is the incorporation of modularity. The feature comes with
tremendous advantages to both patients and surgeons as it offers a broad
range of design and material options. For example, a broader range of head
and taper sizes, neck extensions, several material combinations can be all
explored. A number of these benefits contribute to ensuring the patient’s
specific anatomy can be matched as closely as possible thus achieving the
satisfaction of both patient and surgeon.
On the flip side of this intervention, the consequence of introducing modularity

is realised. It is widely known amongst ‘tribologists’ and surface engineers that
with every additional interface introduced into a tribological system, the risk of
wear and corrosion is largely increased. This is specifically the case in hip
prosthesis; micromotions (fretting) at the modular taper interface, in the
physiological media of the human body result in wear and corrosion of the
implant in-vivo. This has been the nemesis of modularity in THA since its
inception in the 1970s. Of over 77,000 records of revision procedures held by
the National Joint Registry (NJR), over 15% were due to implant wear, fracture
and adverse local tissue reactions (ALTR). Therefore, it is of utmost
importance to understand the in-vivo degradation mechanisms that leads to
wear particles and ion release [1].
1
So far, it is known that corrosion, wear and the synergistic interaction of both
are the overarching degradation mechanisms leading to the failure of these
implants. Therefore, the motive of this research is to better understand the
various damage modes and the mechanisms manifested at the interface of
metallic and ceramic biomaterials used in modular hip prosthesis. This
research was therefore conducted within the remits of fretting corrosion of the
modular taper interface.
1.1 Research objectives
The aim of this research is to develop a better understanding of the

mechanisms involved at the fretting interfaces of metal – metal and ceramic –
metal systems. This covers a wide scope of experimental and analytical
techniques such as: assessment of tribological parameters such as interfacial
energy and wear; monitoring of electrochemical activities during in-situ
tribocorrosion; surface topography; assessment of subsurface metallurgical
transformations; and characterisation of tribochemical products. The
objectives of this study are as follows:
1. Characterisation of CoCrMo and Ti6Al4V alloys (Chapter 5)

 Employ the use of spectroscopy techniques to characterise
passive films on both alloys.
 Utilise advance microscopy techniques to characterise
metallurgical structures of a polished sample
 Perform nano-indentation hardness on a polished sample
2. Assessment of fretting corrosion in metal – metal and ceramic –

metal systems (Tribometer based study – Chapters 6 and 7)
 Assess various material combinations for varied displacement
amplitudes
 Determine the wear volume and conduct surface profilometry
 Utilise advance microscopy and spectroscopy techniques to
characterise emanating fretting corrosion products and the
resultant metallurgical transformations
2
3. Investigate the effect of surface history due to wear and oxidation
on the evolution of fretting corrosion current and the
tribocorrrosion degradation mechanism (Tribometer based study
– Chapter 8)
 Conduct fretting tribometer study under potentio-static
conditions; modify surface history by imposing fretting
displacement amplitudes in series (both increasing and
decreasing order).
4. Assessment of metal – metal and ceramic – metal systems using

realistic taper geometry (Chapter 9)
 Conduct fatigue loading experiments of the material
combinations under various loading conditions and with in-situ
corrosion measurements.
1.2 Thesis outline
 Chapter Two: Relevant background and fundamental theories

related to this research
 Chapter Three: Literature review of implant retrieval studies, in-

vitro studies and tribocorrosion
 Chapter Four: Experimental methodology and analytical

techniques
 Chapter Five: Results obtain from material characterisation of

CoCrMo and Ti6Al4V alloys
 Chapter Six: Fretting corrosion assessment of metal – metal

systems (Tribometer based study)
3
 Chapter Seven: Fretting corrosion assessment of ceramic –
metal systems (Tribometer based study)
 Chapter Eight: Quantification of fretting corrosion current using

self-mated CoCrMo systems (Tribometer based study)
 Chapter Nine: Comparison of fretting corrosion current from

metal – metal and ceramic – metal systems (Realistic
components study)
 Chapter Ten: An overall discussion of all five result chapters
 Chapter Eleven: Main conclusions and recommendation for

future work
It is worth the mention that each of the result chapters end with a brief
‘Discussion and summary’ section. This allows the reader to appreciate the
findings of each chapter distinctively due to most of the result chapters being
quite extensive. Subsequently, the ‘Overall discussion’ chapter (Chapter 10)
is where the overarching themes and mechanisms observed across the whole
thesis are discussed in detail.
4
Chapter 2 Background to biomaterials, corrosion and
tribology
2.1 Introduction
In this chapter, the background to metallic and ceramic biomaterials used in

THA are explored. And also, the fundamental sciences of tribology, fretting
and electrochemistry of corrosion are reviewed.
2.2 Metallic biomaterials
Metallic alloys such as Cobalt (Co) based, Titanium (Ti) based and Iron (Fe)
based alloys are commonly used as biomaterials in biomedical applications.
This is due to their desirable mechanical properties, corrosion resistance and
biocompatibility [2]. The main metallic content are generally alloyed with
various other elements such as: Chromium (Cr), Nickel (Ni), Tantalum (Ta),
Niobium (Nb), Molybdenum (Mo) and Tungsten (W). Each of these alloying
elements play a key role in the optimisation of the material in order to achieve
the required properties specific to the material’s application[3].
2.2.1 Cobalt-based alloys
Co-based alloys have been used in the medical industry since the 1940s [4].
They fall into two basic categories namely: that castable Cobalt-Chrome-
Molybdenum (CoCrMo) alloy and the wrought Co-based alloy composed
mainly of Co, Cr, Ni, W and Mo. The castable and wrought CoCrMo alloys are
known to have subtle differences in their mechanical properties, however,
both are commonly used as a bearing component in hip prosthesis due to their
high abrasion resistant property [5].
CoCrMo alloys forms a solid solution of up to 60% Co which materializes into

a Co-rich matrix of Face Centred Cubic (FCC) structure that is stable at room
temperature. Its high stiffness arises from the precipitation of carbides at the
grain boundaries and can be modified by altering the concentration of the
5
carbon content. The approved wrought CoCrMo alloy for prosthetic implant
applications are commercially available as wrought low carbon CoCrMo
(ASTM F 1537 – 08 type 1) and high carbon CoCrMo (ASTM F 1537 – 08 type
2). The former contains less than 0.14 wt% carbon max. and the latter, 0.15 –
0.35 wt% max. [6]. Their full chemical composition and mechanical properties
are outlined in Table 2-1. Furthermore, due to the low stacking fault energy of
CoCrMo alloys, the FCC structure is susceptible to a strain-induced
transformation to a Hexagonal Close Packed (HCP) structure. This is also
evidenced by subsurface twinning through which the alloy work-hardens [6].
The role of Cr and Mo in the alloy is mainly for surface passivation. In this
process, surface oxides are formed spontaneously on the surface of the alloy
thus protecting it from corrosion – hence the corrosion resistant property of
CoCrMo alloys. It is also known that the inclusion of Mo aids the formation of
finer grains in the alloy matrix, thus further improves the strength of the alloy
[3]. More recently, the bio-compatibility of CoCr alloys are in question due to
the possible release of Cr (VI) (a supposedly carcinogenic specie) which forms
bio-complex molecules in the physiological media of the human body [7, 8].
6
mechanical properties for wrought CoCrMo alloys according to the
American Society for Testing and Materials (ASTM)
Wrought
Specification Co Cr Mo Si Fe Ni Mn C N
CoCrMo
Low F1537 – 08 Bal. 26.0 5.0 1.0 0.75 1.0 1.0 0.14 0.25
Carbon 30.0 7.0 max max max max max max
High F1537 – 08 Bal. 26.0 5.0 1.0 0.75 1.0 1.0 0.15 0.25
Carbon 30.0 7.0 max max max max 0.35 max
Wrought Ultimate tensile Yield strength % Elongation Hardness

CoCrMo strength (MPa) (MPa) (per min) (HRC)
Low
897 – 1172 517 – 827 12 – 20 25 – 30
Carbon
High
897 – 1172 517 – 827 12 – 20 25 – 30
Carbon
2.2.2 Titanium and titanium-based alloys
Commercially pure titanium (CP Ti) and extra low interstitial Titanium-
Aluminium-Vanadium (ELI – Ti6Al4V) are the two most common Ti-based
biomaterials used for biomedical applications. They are known to be highly
biocompatible and do not promote adverse reactions (within considerable
volume of wear) in the human body. This is due to a ~10 nm thick TiO2 surface
oxide film which spontaneously formed on both surfaces [3, 9]. CP Ti is very
light with a density of 4.5 g/cm³ and it is regarded as pure within the range
98.9 – 99.6 wt% Ti. The remainder of the wt% are interstitial elements such
as O, C and N. Nitrogen specifically, is known to have a greater influence on
Ti as it has double the hardening effect (per atom) than C or O. Table 2-2
shows the elemental constituents of both unalloyed and alloyed Ti as well as
specific mechanical properties.
7
Regarding phases and structure, Ti has a HCP α-phase structure below
882°C and a β-phase structure above this temperature. In Titanium-
Aluminium-Vanadium (Ti6Al4V) alloy, the Al plays the role of stabilizing the α-
phase which forms an HCP structure by heightening the temperature required
for the transformation from α to β phase. On the other hand, the V is included
to stabilize the β-phase which has a Body Centred Cubic (BCC) structure by
lowering the temperature required for the same transformation thus forming a
two-phase system [9]. The α-phase has excellent strength characteristics and
is capable of producing localised strain field that can absorb deformation
energy. CP Ti and Ti6Al4V have a relatively low stiffness of 110 GPa
compared to CoCrMo (240 GPa). A lower stiffness is desirable for a
‘cementless’ hip prosthetic procedure in order to reduce the effect of stress
shielding. An even lower modulus (80 GPa) alloy called Titanium-
12Molydenum-6Zirconium-2Iron (TMZF), also known as the beta-Titanium
was developed to further improve osteo-integration at the Ti-Bone interface
[10]. On a less desirable note, Ti alloys are known to having poor shear
strength and hence low wear resistance relative to CoCrMo alloy. Therefore,
Ti alloys are typically not suitable for bearing applications in hip prosthesis [3,
9].
8
mechanical properties for unalloyed Titanium and Extra Low Interstitial
(ELI) Ti6Al4V alloy according to the ASTM
Titanium Specification Ti Al V Fe Ni C N O H
0.5 0.05 0.08 0.4 0.015

Unalloyed F67 – 06 Bal. - - -
max max Max max max
5.5 3.5 0.25 0.08 0.05 0.13 0.012

ELI Ti6Al4V F136 – 08 Bal. -
6.5 4.5 max max max max max
Ultimate tensile Yield strength % Elongation

Titanium Hardness (HRC)
strength (MPa) (MPa) (per min)
Unalloyed 500 483 15 –
ELI Ti6Al4V 825 – 860 760 – 795 8 – 10 –
2.2.3 Stainless steel
Type 302 was the first stainless steel to be fabricated for medical implants.
Shortly after, a relatively small amount of Mo (2 – 4 wt%) was introduced into
the matrix to further improve its corrosion resistance for in-vivo biomedical;
applications. Thus the type 302 became the common type: 316 stainless steel.
The inclusion of Mo specifically plays the role of enhancing resistance to
pitting corrosion in salt water. The type 316 was later modified in the 1950s
and became known as 316L when the carbon content was reduced from 0.08
wt% to a maximum of 0.03 wt%. The aim of this was to minimize the effect of
sensitization whereby specific alloying elements such as Cr is preferentially
precipitated from the bulk to the grain boundary as Metal Carbide. Ni is also
largely present in both commercial 316 and 316L alloys with a content range
of 12 – 14 wt%; its role in the alloy is mainly to stabilize the FCC phase at
room temperature.
As for Cr content, the minimum composition of Cr to effectively impart

corrosion resistance through passivation in these alloys is said to be 11 wt%
9
however as its wt% composition also affects the stability of the FCC austenitic
phase, the ASTM F138 prescription falls between 17 – 19 wt% as outlined in
Table 2-3. On the negative side, 316L alloy is known to be susceptible to
crevice corrosion in the harsh crevice environment of the human body where
limited access to oxygen inhibits its ability to adequately restore its protective
surface oxides [3].

mechanical properties for 316L stainless steel according to ASTM.
Stainless Steel Spec. Fe Cr Ni Mo Mn Si Cu C N
17.0 13.0 2.25 2.0 0.75 0.5 0.03 0.25

316L F138 – 08 Bal.
19.0 15.0 3.0 max max max max max
Ultimate
tensile Yield strength
Stainless Steel % Elongation (min) Hardness (HRC)
strength (MPa)
(MPa)
316L 490 – 1350 190 – 690 12 – 40 25
2.3 Ceramic biomaterials
The use of ceramics have been incorporated into various prostheses for
application in hip, knee, shoulder and spine. Ceramics used in biomedical
applications possess desirable properties like: high hardness and strength,
electrochemical inertness, excellent biocompatibility and wear resistance [11].
The following are the commonly used ceramics in various prosthetic
applications today: alumina, zirconia, zirconia-toughened alumina (ZTA),
alumina matrix composites (AMC), alumina-toughened zirconia (ATZ), Silicon
Nitride (Si3N4) and hydroxyapatite (HAp) [10]. Further details on ZTA and
Si3N4 ceramics (the two relevant ceramics to this study) are provided in the
subsequent sections.
10
2.3.1 Alumina (Al2O3) and alumina composite
Alumina has been the most widely used structural ceramic for about 5
decades now. Pierre Boutin in 1970 implanted the first hip prosthesis with
Alumina ceramic bearings and its success led to the current hard-on-hard and
hard-on-soft bearing combinations such as the: Ceramic – on – Ceramic
(CoC), Ceramic – on – Metal (CoM) and Ceramic – on – Polymer (CoP). It has
good biocompatibility, high hardness, low coefficient of friction and hence
excellent wear resistance. However, its inadequate fracture strength and
toughness led to numerous cases of in-vivo fracture; thus leading to failure
rates of up to 13.4% being reported in the 1970s [11, 12].
Optimisation of manufacturing processes has led to generations of improved

alumina ceramic with excellent mechanical properties as shown in Figure 2-1
[13, 14]. For example, the third generation alumina ceramic (still
commercialised today) benefitted from the use of hot-isostatic pressing (HIP)
leading to a significant reduction in grain size from ~3.2 µm to ~1.8 µm. The
most recent generation of alumina ceramics used specifically in orthopaedic
applications is the ZTA ceramic commercially known as Biolox®Delta. It’s a
composite ceramic of 82% alumina and 17% zirconia nano-particles by
volume. In Figure 2-1 it is observable that the strength of the composite is
almost double that of the third generation Alumina. Figure 2-2 depicts the two
mechanisms attributed for its excellent toughness. The first is the presence of
zirconia nano-particles that are used to for arresting propagating cracks. The
second are large platelets of elongated oxide crystals that form in-situ. To
compensate for the lower hardness of the Zirconia particles in the matrix,
hardness is restored to the material by adding chromium oxide (Cr2O3) as a
solid solution in the alumina matrix [13, 15]. The recent failure rates of these
toughened ceramics now range from 0.004 – 0.05% which is a significant
improvement [11].
11
Figure 2-1 – Generations of alumina ceramic [13]
Figure 2-2 Mechanism of improved toughness in composite alumina

ceramic [15]
12
2.3.2 Silicon Nitride (Si3N4)
Silicon nitride is used in a wide range of applications including: automotive,

aerospace and more recently biomedical applications [14]. The material had
been synthetically developed as early as 1859 however, commercial interest
in the material was not until the 1950s when the properties of the material was
better understood. The material is known for its high fracture toughness,
strength and desirable wear properties. With respect to its biomedical
applications, silicon nitride is known for its biocompatibility and ability to
dissolve in physiological media thus reducing the effect of particle induced
osteolysis [16, 17]. However, it wasn’t until 2006 that silicon nitride was
officially approved by the food and drug administration (FDA) as a biomaterial
[18]. Since 2008, it is being used as spinal fusion cage implants which has
proven to be a successful venture with ~25000 implants with only few adverse
reported events [14].
The two typical processing methods for manufacturing silicon nitride are:
reaction-bonding and HIP. The reaction-bonded silicon nitride produces a
relatively low density, high porosity and low strength material. On the other
hand, while the HIP method incurs higher manufacturing cost, it was
developed to address the listed poor properties of the reaction-bonded
method [19]. However, the two technologies can be combined by
manufacturing the silicon nitride through an initial stage of sintering and
subsequently reaction-bonding. This results in a higher strength material for
the manufacturing cost that is lower than the HIPed system. Silicon nitride
has a composition of up to 10 wt% Yttria (Y2O3) and alumina as an additive
[19]. Its typical mechanical properties are listed as follows:
 Bending strength – 600 MPa
 Vicker’s Hardness ~ 12 – 13 GPa
 Youngs Modulus – 299 GPa
 Poisson ratio – 0.27
 Typical grain size ~ 0.59 µm
Silicon nitride forms a solid-solution of two phase structure: equixed αSi3N4
and acicular βSi3N4. The toughness of the ceramic can be controlled by the
13
engineering both the grain size and the transformation of the equiaxed αSi3N4
to the acicular βSi3N4 phases [14]. Quite importantly, Si3N4 also has a strong
network of SiO2 in its grain boundaries which has been shown to play a
significant role in the formation of thin oxidised-Si tribofilms and the corrosion
of the ceramic [20].
2.4 Corrosion
This section covers fundamental principles of corrosion while also drawing on

specific electrochemical behaviour of CoCrMo and Ti6Al4V alloys. The
synovial fluid in the joint capsule consist of Cl and phosphate ions and other
electro-active species (EAS) that contributes to creating an aggressive
physiological media for the metallic components used in hip prosthesis. The
metallic implants are susceptible to various forms of corrosion mechanisms,
namely: fretting corrosion, pitting, crevice corrosion, galvanic corrosion,
intergranular corrosion and others.
2.4.1 Electrochemistry of corrosion
Corrosion of metals is mostly electrochemical in nature as well as underpinned

on thermodynamic principles. Electrochemical reactions is simply
characterised by two half-cell reactions in aqueous solution: Anodic
(Oxidation) reaction and Cathodic (Reduction) reaction. The anodic reaction
releases electrons while the cathodic reaction receives the electrons.
Equation 2-1 shows a standard anodic half-cell equation of a metal ‘M’ with
charge valence ‘n’ which corresponds to ‘n’ number of electrons.
𝑀 → 𝑀𝑛+ + 𝑛𝑒 − 2-1
An example of a cathodic half-cell reaction is one of oxygen reduction within

a base or alkali solution shown in Equation 2-2 [21].
𝑂2 + 2𝐻2 𝑂 + 4𝑒 − → 4𝑂𝐻 − 2-2
A metal within an electrolyte is an array of structured positive cations with a

shroud of free electrons, each of which belongs to the valence of an atom in
14
the structure. The free electrons is the main determinant of the metal’s
conductivity [22]. In an aqueous solution, water gets adsorbed onto the
metallic surface creating an interface between the metallic substrate and
hydrated metallic cations ions, which migrated from the surface of the metallic
substrate. This leaves the metallic substrate negatively charged and the
hydrated layer, populous of positive ion. The two parallel layers can be
described as two planes of separate charges hence forming an
electrochemical capacitance with a potential difference across the plane, this
is known as the electrical double layer (EDL), a model initially introduced by
Helmholtz in 1850 [23]. Figure 2-3a shows a schematic of the surface as well
as the linear increase in potential (left of the outer Helmholtz plane (OHP))
and the non-linear increase in potential (right of the OHP).
The rate of electrochemical half-cell reactions that occurs on a metallic surface
is highly depended on the anodic/cathodic surface area ratio. As a metal
corrodes, electrons released at the anodic site flows through the metal to react
at the anodic site as depicted in Figure 2-3b for Zinc metal in acidic solution.
The flow of electron is the measure of current, therefore, the current due to
the anodic reactions at the anodic site can be expressed as a function of the
cathodic current and area as shown in Equation 2-3. 𝐼𝑎𝑛𝑜𝑑𝑖𝑐/𝑐𝑎𝑡ℎ𝑜𝑑𝑖𝑐 are both
the total anodic and cathodic currents measured over the respective surface
areas 𝑆𝑎 and 𝑆𝑐 while 𝑖𝑎 and 𝑖𝑐 represents the currents measured at the local
sites [21].
15
(a) (b)
Figure 2-3 – Electrical interactions at the metallic surface a) Schematic

diagram of electric double layer [23] b) Anodic and Cathodic sites during
corrosion [21].
𝐼𝑎𝑛𝑜𝑑𝑖𝑐 = 𝑖𝑎 𝑥 𝑆𝑎 = 𝐼𝑐𝑎𝑡ℎ𝑜𝑑𝑖𝑐 = 𝑖𝑐 𝑥 𝑆𝑐
𝑆𝑎 2-3
∴ 𝑟𝑒𝑎𝑟𝑟𝑎𝑛𝑔𝑖𝑛𝑔, 𝑖𝑎 = 𝑖𝑐 ( )
𝑆𝑐
2.4.2 Types of corrosion effective at the modular junction of THA
2.4.2.1 General corrosion
General corrosion, also known as uniform corrosion acts uniformly over the
entire surface area of a metal when exposed to an electrolyte. Air can also be
an electrolyte for corrosion if the critical relative humidity of the material is
reached or surpassed. This form of corrosion is specifically characterised by
numerous individual electrochemical reactions occurring simultaneously over
a metallic surface. It is the most common form of corrosion and can be
prevented through the use of coating or by solution controls i.e. modifying the
pH, temperature etc. [24, 25].
16
2.4.2.2 Intergranular corrosion
Intergranular corrosion occurs in metals and alloys where there is a different

electrochemical interaction between the grain boundaries and the grains. In
such cases, the grain boundary is prone to localized attack while the
remaining parts of the metal remain unaffected. This eventually damages the
mechanical integrity of the material. Figure 2-4 shows a micrograph of
intergranular corrosion attack on CoCrMo alloy. Causes of this form of
corrosion are as follows: impurities at the grain boundary, enrichment or
depletion of a particular element in an alloy – like in the case of sensitisation.
It has been found that in a Cl rich media, precipitate phases of the alloy can
induce intergranular corrosion and so does the precipitation of carbides at the
boundaries [24]. CoCrMo alloy is known to be more susceptible to this form of
corrosion than Ti6Al4V [26, 27].
Figure 2-4 – Intergranular attack on CoCrMo alloy [27]
2.4.2.3 Pitting corrosion
Pitting corrosion is a highly localised corrosion mechanism that proves to be

highly destructive to mechanical systems. The process can be very rapid with
no appreciable mass loss from the surface under attack. For this reason,
17
pitting can be very difficult to predict or even detect on a surface. For passive
metals and alloys, pitting occurs by locally destroying the oxide layer which
leaves the metal or alloy exposed; hence encouraging further pitting of the
surface. Titanium and CoCr based alloys used in orthopaedic implants can
release metallic ions during an anodic reaction that causes metal dissolution
[24, 25]. Release of the metal ions in the presence of EAS and oxygen can
make the pitting more aggressive. Pitting corrosion has been detected as a
degradation mechanism for CoCrMo in both like and similar material
combinations [27].
Figure 2-5 – Pitting corrosion attack on CoCrMo alloy [27].
2.4.2.4 Crevice corrosion
Crevice corrosion occurs in cracks and crevices of materials that are either
similar or dissimilar. This form of corrosion usually occurs in passive metals
and is influenced by the following conditions: geometry of crevice, alloy
composition, passive film characteristics, bulk solution composition and
temperature. Fontana and Greene [28] explains three conditions required to
be met in order for crevice corrosion to be initiated. Firstly, there has to be an
electrical connection between the crevice and the external surfaces.
Secondly, the fluid in the crevice should be trapped such that the dissolved
oxygen in the solution is used up in a reaction with the metal surface to form
18
oxides within the crevice. The stagnancy of the fluid also means that diffusion
of oxygen into trapped fluid is restricted resulting in depleted oxygen.
The third condition is satisfied if the material outside the crevice remains
passive. This causes a differential aeration cell between the external material
and the microenvironment inside the crevice. The lack of oxygen as well as
the migration of Cl ions in the environment also inhibits the normal
repassivation that would occur on the metallic surface. Therefore, a
permanent breakdown of the passive film also occurs at the crevice [29]. One
other key implication of crevice corrosion is that the products of corrosion often
result in a tighter crevice as depicted in Figure 2-6. Reactions from the anodic
site releases H+ ions thus creating a highly acidic environment where pH could
drop to as low as 0, thus resulting in faster metallic dissolution [29].
Figure 2-6 – A simplified 3 stage depiction of crevice corrosion: a)

conducting electrolyte in contact with crevice. b) cathodic and anodic
sites are formed, c) corrosion products create a severer crevice [21].
2.4.2.5 Galvanic corrosion
Galvanic corrosion in its simplest manifestation occurs as a result of

differences in the standard potential of two metals or alloys that are in a
conductive corrosive environment. The factors that significantly affect galvanic
19
corrosion are: potential difference between two coupled materials; the
electrolyte both are exposed to, the area, distance and geometry of the
materials when in close proximity of one another or in contact. Furthermore,
in galvanic coupling, the surface area is a very important consideration
because one of the metals may act as the cathode while the other behaves
as the anode. This behaviour is a function of the standard potential of each
metal and the electrolyte composition [24, 25]. In the context of modularity in
THA, the coupling of CoCrMo alloy against Ti6Al4V is one that had sparked a
debate as to whether galvanic corrosion plays a relevant role in the
degradation mechanism [27, 30].
2.4.3 Passivity of biomaterials
Metallic biomaterials are generally described as passive alloys due to the thin
(nano-scale) oxide layer that spontaneously form on the surface of the metal.
The corrosion resistance of metallic biomaterials are largely due to their
protective oxide layer. This layer limits the rate of corrosion by becoming a
physical barrier against metallic ionic dissolution. The absence of such
protective layer would make the metal surface prone corrosion attack by EAS
in the body [2]. However, metallic biomaterials can behave like active metals
in the absence of the passive layer or while the passive film is unstable.
Figure 2-7 shows a theoretical plot of a potentio-dynamic scan for a passive

alloy. The scan begins in an anodic direction from a potential where the
corresponding current is minimum i.e. zero. This potential is known as the
open circuit potential (OCP); it is the point where the rate of anodic reaction
equals the rate of cathodic reaction thus yielding zero net current. From this
point, the metal behaves like an active metal as the corrosion rate rapidly rises
up to a critical point termed icrit. This critical point occurs at the passivation
potential Epp subsequent to which the current falls to a constant rate termed
ipass. This current ipass remains theoretically constant as potential increases; it
is the current required (net anodic reaction current) to form the oxide layer.
This region of constant current ipass is known as the passive region and the
larger the potential at this region is, the more stable the oxide layer is. Beyond
the passive region is the transpassive region; the region is characterised by
20
an accelerated corrosion current caused by break down in the oxide layer.
The metal’s ability to repassivate as well as its resistivity to pitting and crevice
corrosion are assessed at this potential [31].
Figure 2-7 – Theoretical plot of a potentio-dynamic scan in the anodic

direction for a passive alloy [32]
2.4.4 Faraday’s law
Faraday’s law on electrolysis converts the current generated from an anodic

reaction into an equivalent mass loss. This law relates the current measured
and the number of moles of electrons involved in the reaction by the
conversion of 1mol of electrons i.e. 1 Avogadro’s number of electrons (6.022
x 1023) into charge. The product of the number of electrons and the charge of
a single electron (1.6 x 10-19) gives 1 F which equals to 96485 Cmol-1 of
electrons [21].
The mass loss (m) from the electrochemical reaction is therefore calculated
using Faraday’s equation in Equation 2-4. The electrical charge (Q) with unit
C is determined as the sum of the anodic current generated across a time
period (t) as shown in Equation 2-5. ‘M’ with unit gmol-1 is the molar mass of
the reacting species and ‘n’ is the number of electrons in the electrochemical
reaction.
21
𝑄𝑀 2-4
𝑚=( )
𝐹𝑛
𝑡
2-5
𝑄 = ∫ 𝐼 𝑑𝑡
0
2.5 Tribology
Tribology is the study of friction, lubrication and wear between two contacting
surfaces in relative motion. Friction is the main route through which energy is
dissipated between two contacting materials. The energy may be dissipated
through wear, structural transformation of the material, heat etc. Therefore,
improved control of friction can prove beneficial while poor management can
prove costly. Lubrication is one of such ways in which friction between two
moving surfaces can be controlled thus minimising interfacial damage [33].
The following subsections focuses mainly on wear and the most commonly
observed wear mechanisms.
2.5.1 Wear
Wear generally occurs as a result of mechanical action between two or more

solid bodies in contact. A mechanical wear process undergoes two stages:
the running-in stage and the steady-state wear rate. After a long period of
time, the steady-state wear rate could transition into fatigue behaviour which
further reduces the wear rate [34]. Wear is manifested through different
mechanisms namely: adhesive wear, corrosive wear, fatigue wear, abrasive
wear and others. The subsequent sections address four main wear
mechanisms that are effective in biomaterials. Fretting wear and fatigue
mechanisms are addressed in Section 2.6.
2.5.2 Adhesive wear
Adhesive wear is also known as scoring, galling or seizing together of two

contacting surfaces. Materials are prone to this mechanism during relative
22
motion of two contacting surfaces that are subjected to high contact
pressures. The process can also be described as the welding together of the
two surface, in which case, the asperities at the contact are plastically
deformed and during strong adhesion of metallic surfaces, electrons can be
transferred between the bodies in contact. Figure 2-8 provides a depiction of
adhesive wear mechanism. Increasing the hardness of surfaces may prevent
against this wear mechanism [33].
Figure 2-8 – Adhesive wear mechanism [35]
2.5.3 Abrasive wear
Removal of material due to abrasive wear mechanism is the process where

particles or hard asperities are indented into a softer surface with subsequent
relative motion. This mechanism, in some cases, is preceded by the adhesive
wear in which case, the material detachment came from the welded interface
thus creating a much harder particle. This mechanism can also lead to micro-
cutting, fracture, as well as ploughing of a surface and hence may contribute
to both the initiation and propagation stages of cracks in a contact subject to
fatigue stresses. The bearing surface of CoM is one interface susceptible to
this mechanism as the hard particulate debris of the ceramic component may
abrade the softer metallic surface as depicted in Figure 2-9.
23
Figure 2-9 – Abrasive wear mechanism [35]
2.5.4 Fatigue Wear
Cyclic application of stress at the surface of a solid body relative to another,

results in fatigue wear. Fatigue wear mechanism generates much less wear
compared to other wear mechanisms. The pre-requisite for crack initiation on
one of both of the bodies are high normal load and/or high surface compliance
of one or both of the bodies. From Figure 2-10, it can be seen that two surfaces
that are in contact are loaded against each other and if the tangential forces
are insufficient to overcome the static frictional force, the surface elastically
deforms repetitively leading to a cyclic fatigue motion. The partial slip regime
of a fretting contact often create such contact condition [36].
Figure 2-10 – Fatigue wear mechanism [35]
24
2.5.5 Corrosive wear
This wear mechanism occurs in a corrosive environment and the interactions

between tribology and oxidation reactions occurring at the interface of the
contacting materials results in the formation of wear and oxidised products
(corrosion). Air at room temperature contains humidity and may also contain
other industrial vapors, therefore, oxides and hydroxides of metals are
common products formed during corrosive wear [37]. Oxides from the metal
surface has the ability to reduce the rate of corrosion given that it satisfies
desirable mechanical properties such as wear resistivity, low friction and
relatively thick for corrosion resistivity. In some other cases the film formed on
the surface is too thin and therefore the film might be too weak to withstand
sliding hence the film has a short life span which results in an overall increase
in wear rate.
The described process also corresponds to the corrosion enhanced wear (Cw)
synergy factor in a tribocorrosion system as discussed in Section 3.5.
Corrosive wear product can be both beneficial and adversely effectual on
biomaterials and this is dependent on the characteristics of the oxide film that
is formed on the material surface [33]. Figure 2-11 shows a depiction of
corrosion and wear products on the wearing surface.
Figure 2-11 – Corrosive wear mechanism [35]
25
2.6 Fretting
Fretting is defined as an oscillatory displacement at low amplitude between

two contacting surfaces. Although the phrase ‘low amplitude’ has no formally
defined boundary, the generally accepted upper limit is 300 µm above which
the contact transitions into reciprocating sliding [38]. However, Fouvry et al
[39] expressed fretting independently of an upper limit by expressing it as a
characteristic of the Hertzian contact radius ‘a’ as shown in Figure 2-12. Figure
2-12a shows a schematic representation of a gross slip fretting contact on the
left and the representation of a reciprocating sliding contact on the right.
The parameter ‘e’ is the ratio between the displacement amplitude ‘δx’ and the
hertzian contact radius ‘a’. Through the ratio ‘e’, a contact can be distinguished
as to whether it is in fretting (e < 1) or reciprocating sliding contact (e > 1). For
the fretting contact, an unexposed region exist at the point where the two
hertzian contacts overlap. The damage manifested on the contacting surfaces
during fretting are generally: minimal; fatigue dominant or wear dominant as
characterised by Vingsbo and Soderberg [38]. Furthermore, these three
damage modes are generally linked with three fretting regimes which were
first described by Mindlin in 1949 [40]. These are outlined in detail in the
subsequent sections.
Experimental efforts to characterise fretting damage in realistic systems are

often conducted using tribometers where various contact configurations
namely: flat-on-flat, cylinder-on-flat, sphere-on-flat and others are established
with selectively chosen contact pressures and displacements to represent the
realistic conditions as close as possible.
26
Figure 2-12 – Difference between a fretting contact (left) and a
reciprocating sliding contact (right). Adapted from [39]
2.6.1 Fretting regimes
The fretting regimes described by Mindlin are namely: Stick, Partial-slip (or
mixed stick-slip) and gross slip. These were represented by a tangential force
(Ft) – Displacement (δx) plots also known as fretting loops, using a cross
cylinder configuration by Vingsbo and Soderberg [38]. Figure 2-13 shows the
schematic representation for each fretting regime. In the same figure, the ratio
of interfacial slip (δs) to the displacement amplitude (δx) is denoted ‘A’ and the
area encompassed by the fretting loop corresponds to the dissipated energy
(Ed).
2.6.1.1 Stick regime
The stick regime is characterised by a tightly closed fretting loop. Wear

generated at its interface and fretting damage are minimal as only surface
asperities undergo plastic deformation [38]. The damage mechanism is
fatigue dominant due to continuous cyclic elastic deformation of the contacting
surfaces.
27
Figure 2-13 – Fretting hysteresis loops from a plot of tangential force
and fretting displacement amplitude [41], beneath the fretting regimes
are the characteristic surface damage profiles [38].
2.6.1.2 Partial-slip regime
The Partial-slip regime is characterised by an elliptical shape signifying the

introduction of slip at the interface which corresponds to a transition from a
dominant elastic shear to plastic shearing of the surface. The surface damage
identified in Figure 2-13 (above) shows a central stick region surrounded by
an annular of slip area [39, 41]. The damage in the central region is indicative
of plastic deformation of the asperity-asperity contacts which contributes little
to the overall wear of the interface. The stresses effective at the worn annulus
of this regime is more dominant in fatigue crack formation especially at the
boundary between the central stick region and the worn region [38, 39].
However, the further away from the boundary and the closer towards the edge
of the Hertzian contact, the regime becomes increasingly wear dominated.
This is illustrated in Figure 2-14a where a yield annulus is introduced.
The yield annulus corresponds to the point of maximum tangential force prior
to the transition from static friction to kinetic (or dynamic) friction; it can also
28
be referred to as the traction forces (qx) [42]. Intersecting horizontal cracks
formed in both yield and slip annular often result in the removal of larger
delamination wear product, especially in the case where the contact had been
oxidised making the surface harder and brittle compared to rest of the bulk
[38].
Figure 2-14 – Fretting regimes: a) characterisation of the fretting

contact of a partial-slip regime in terms of tangential force (Ft ) and
traction force (qx) [42].
2.6.1.3 Gross slip regime
The gross slip regime is generally characterised by a parallelogram shape as

shown in Figure 2-13 (above). The levelling off of the tangential force in the
fretting loop is indicative of a complete transition from static friction to dynamic
friction from the micrograph displayed in Figure 2-13 (above), it can be seen
that majority of the central region had undergone a plastic shear and across
the entire contact, all asperity contacts are broken off [38]. In general, the
regime is characterised by extensive wear through delamination and shear of
adhesive junctions (especially from the central region) [38, 39, 43]. Wear and
oxidation products are usually trapped between the contacting surfaces,
resulting in possible shift in wear mechanism to an abrasive dominated
mechanism. It is also common for the two bodies subjected to fretting wear to
form layers of wear debris and oxidised surfaces that thickens over time and
accelerates degradation of the two surfaces through abrasion [44].
29
2.6.1.4 Mixed regime
The term ‘Mixed regime’ is often used to describe the partial-slip regime,
however, Fouvry et al [39] introduced the ‘mixed slip regime’ which represents
the case whereby the contact condition changes during a fretting experiment.
This may be manifested in fretting regime transition from any of the three other
regimes to each other. For example, the partial-slip may transition to gross
slip during an on-going test and vice-versa. Such fretting behaviour is
prominent at a contact characterised with high loads where the imposed
displacement exceed the threshold for transition from the partial-slip regime
[39].
2.6.1.5 Mindlin’s hypothesis
Mindlin’s hypothesis as described in the reference [45] predicts that for two
materials subject to a single point contact in a purely elastic condition,
interfacial slip will always begin at the outer annular as depicted in Figure 2-
15. This description becomes relevant at the end of a fretting test when the
contact is reviewed. The presence of a central stick region or in the case of
gross slip regime, central partial-stick region gives an indication that the
wearing mechanism at the fretting interface is still affected by elastic contact
compliance. The alternative, which is the more common case, is the influence
of third-body product at the fretting contact.
Figure 2-15 – Schematic representation of Mindlin’s hypothesis
30
2.6.2 Fretting corrosion, wear and fatigue
The term fretting wear was initially described as fretting corrosion because
oxidation plays a large role in its wear mechanism [46] and fretting wear is
also often referred to as fretting fatigue due to the delamination processes in
its wear mechanism [38]. However, fretting corrosion is more commonly
known today as a specific process in a tribocorrosion system [47]. On the other
hand, fretting wear and fretting fatigue modes of damage in their right
definitions have been compiled from a number of fretting experiments by
Vingsbo and Soderberg [38] as a factor of displacement amplitude (Figure 2-
16a) and load (Figure 2-16b).
With respect to the wear and fatigue modes of damage in (Figure 2-16a) it is
observable that the wear rate rises sharply with a small increase in
displacement during the stick regime and the fatigue life is highest at this point
whereas as the contact transitions into partial/mixed slip regime, wear rate
remains almost constant for a relative large range of displacement while
fatigue life drops drastically. On transition into gross slip, a directly
proportionality exist between displacement and wear rate. However, a sharp
rise in the fatigue life plateaus before reciprocating sliding is reached.
From Figure 2-16b it is observable that at low loads, fatigue life is high and
wear rate also but the fatigue life drops sharply at still relatively low loads. At
much higher loads, fatigue life drops significantly along with wear rate. And at
very high loads, the contact is in a stick regime, fatigue life rises and wear rate
drops, providing the least damaging condition.
31
(a) (b)
Figure 2-16 – Fretting damage modes: a) wear and fatigue modes as a

function of displacement, b) wear and fatigue modes as a function of
load [38].
32
Chapter 3 Literature review
3.1 Total hip arthroplasty
The earliest record of THA was by Glϋck in 1891; although it was not until the
1950s that THA became a long term solution for arthritis and diseases of the
joints through Sir John Charnley’s low friction arthroplasty [48]. His design
consisted of three main parts; a mono-block metallic femoral stem, a
polyethylene acetabular component and acrylic bone cement [49]. Presently,
advances in the design of polyethylene through the development of
crosslinked ultra-high-molecular-weight-polyethylene (UHMWPE) in the late
1990s has kept Metal-on-Polymer (MoP) bearing the most widely used
combination in cemented primary THA (see Figure 3-1a) [1, 49, 50].
On the contrary, excessive metal ion release into the body led to the down fall
of the thriving Metal-on-Metal (MoM) articulating bearings. Consequentially,
the excessive metal ion release into the body led to inflammatory responses,
pseudo-tumours and other adverse reactions. This resulted in a significant
increase of revision rates which led to the recall of some implant designs like
the DePuy articular surface replacement (ASR) as well as the Zimmer Durom
hip resurfacing system. The impact of this action is reflected in the sharp
decline of MoM THA use since 2007 as represented in Figure 3-1b [1].
As aforementioned, alternative bearing combinations became an option in the

early 1970s when stem modularity was introduced to the design of THA. This
introduced the use of alumina ceramics as a femoral head component. The
idea of CoC bearing was first introduced by Pierre Boutin in 1970. And as
shown in Figure 3-1b, the use of CoC gained ample attention in uncemented
primary THA procedures in the early 2010s. Nevertheless, the hard-on-soft
alternative (CoP) is seen as a more popular option in recent times than CoC
as shown in the same (Figure 3-1b). CoC bearing combination is today
prescribed the best bearing combination for young active patients [51] while
the CoP combination is consistently reported as the material combination with
the lowest revision risk among all the other bearing combinations [1]. In
33
general, the number of registered primary procedures in the UK had increased
steadily to over 80,000 in 2014 and 2015 and the value is expected to rise due
to an aging population [1].
(a)
(b)
Figure 3-1 – Percentage of bearing combinations a) cemented b)

uncemented [1].
3.1.1 Rationale for modularity in THA
The head-neck modularity also known as the Morse taper was introduced to
THR in the early 1970s with the purpose of achieving a durable fixation
between ceramic femoral ball and a metal stem hence creating an alternative
bearing to Cobalt-Chromium-Molybdenum (CoCrMo) alloys [52]. Other
advantages of modularity in THA design are: increasing option of intra-
34
operative customisation of the implant device thus the natural biomechanics
of a patient’s hip can be better reproduced after surgery; complications in
revision surgery and costly inventory can be greatly reduced [53].
Furthermore, modularity allows titanium alloy to be used for the stem
component of THA whilst a more wear resistant alloy is used as the bearing
component thus the effect of stress shielding is reduced [54].
In the early 1990s, a second modularity was introduced to the femoral stem
component of hip prosthesis. This is commonly known as the bi-modular
design; it is depicted in Figure 3-2. Subsequently, other forms of modularity
were introduced such as the tri-modular design – a design particularly used in
the treatment of peri-prosthetic femoral fracture. The proximal shoulder,
anterior/posterior pads, modular pads and stem-sleeve junction are also
examples of other forms of modularity in THA [55]. Figure 3-2 shows a collage
of various implant designs depicting the evolution of the modularity in THA
from Charnley’s monobloc design in the 1960s to the more recent bi-modular
and tri-modular THA design.
It is noteworthy that with every new interface introduced to the hip prosthesis,
new challenges related to wear and corrosion arises inevitably. More so, one
study has estimated that modular femoral stem components can cost between
15% to 25% more than standard femoral stem components for primary or
revision THA thus, the extra cost is also worth considering [54].
Figure 3-2 – Evolution of modularity in THA
35
3.1.2 The Morse taper (head-neck taper)
The Morse taper is ubiquitous in the design of THA devices as a means of

securing the male taper unto the female taper throughout the in-vivo lifetime
of the implant. Impaction force by the surgeon provides frictional resistance
against moment or axial forces incurred at the taper interface during daily
living activities. Nevertheless, this doesn’t eradicate micro-motions at the
modular interface hence fretting and corrosion damage are always found at
these interfaces after some time in-vivo. There are three main taper designs
commonly used today namely: 12/14 ‘Euro’-taper, the V-40 and the ‘Type 1’
taper. Other taper designs exist such as the: C-taper, 14/16 taper, and the
11/13 taper although these have become less common in the orthopaedic
market. For example, the larger 14/16 taper was replaced to reduce the risk
of dislocation and the narrower trunnion (also known as male taper) of the
11/13 provides less surface area for interference fit thus increasing fretting
damage [56].
Figure 3-3 is a depiction of increasing head sizes from left to right and
decreasing taper sizes in the same direction. The Euro – 12/14 is the most
widely used amongst various companies while the other two are specific to a
company; its dimensions are 12 mm proximal diameter and 14 mm distal
diameter. The tolerance and taper surface morphology of these taper differs
between companies; some models have smooth finish while others have
grooves of varying pitches and depths. For this reason, mixing and matching
of components from different manufacturers is generally advised against [57,
58]. A recent study [59], found that the choice between smooth or grooved
male taper affects the taper strength as their study reported a higher taper
strength when smooth tapers are used compared to grooved tapers. Although,
the grooved tapers are necessary for securing ceramic bearings unto the male
taper as the grooves deform against the ceramic when the taper is assembled
with an impaction force thus creating an interference fit. The multifactorial
challenges encountered at the head-neck modular interface are discussed in
greater detail later in this chapter.
36
Figure 3-3 – Increasing bearing sizes from left to right and decreasing
taper sizes in the same direction [57].
3.1.3 The neck–stem taper
Most femoral stem taper designs are elliptical in shape although few have a
circular profile as later shown in Figure 3-8 and Figure 3-9 [54]. The neck
trunnion and stem bore both have are tapered profile with a smooth surface
finish so that the neck engages with the bore and forms a tight fit with the
necessary surgical impaction force through the head component. This
modular interface is subject to both axial and bending load possibly leading to
micromotions. However, modularity at the neck-stem have the advantage of
allowing adjustable leg length independently of both horizontal and femoral
offsets. In addition, this extra modularity is said to have improved the range of
movement in several patients with complex anatomies [55].
3.1.4 Assembly of modular tapers
In the uni-modular system, the Morse taper at the head-neck interface is

simply assembled, firstly by twisting the femoral head unto the male taper (for
grooved trunnions) or simply seating the head bore unto the trunnion (for
smooth trunnions), subsequently, the taper is secured by applying an
impaction force unto the femoral head using a head impactor and a mallet as
shown in Figure 3-4. Larger heads (>28 mm) are said to require larger
impaction forces to ensure tight fit due to the larger friction forces they transmit
to the modular interface [54, 60]. Furthermore, during assembly it is essential
that the tapered surfaces are dry and free from contamination before
37
impaction as this may lead to accelerated fretting and possibly reduce the
fracture toughness of ceramic bearings [11, 61].
Bi-modular tapers such as the Zimmer M/L Taper with Kinetiv® Technology
prescribes only a single strike using the head impactor to secure both head-
neck and neck-stem tapers [62] however multiple impaction may be necessary
to ensure the tapers are well fixed. Indeed, a number of studies are advocating
the need for larger impaction forces to reduce fretting at the taper interface
[59, 63, 64]. In a study by Rehmer et al [65] which focused on the Morse taper
connection, it was demonstrated that the taper strength increases linearly with
increase in assembly force while multiple impaction force had no effect.
However, the ability to separate the tapers by the application of turn-off
moments varied correspondingly to material combination and assembly force
[59, 65]. In the case of the tri-modular tapers, Figure 3-2 shows how the design
allows for various lengths of the stem to be fastened using a screw and once
secure, the femoral head is subsequently impacted unto the neck component.
Figure 3-4 – Fixation of both head–neck and neck–stem [62]
3.2 Challenges with modularity
The challenges with modularity ultimately culminates in patient discomfort.

Examples of these are in-vivo fracture, excessive wear and voluminous metal
ion release and others. Retrieval analysis of the modular taper interface dates
back as far as the early 1970s when modularity was initially introduced [66].
Through the use of electron microscopy, energy dispersed x-ray spectroscopy
38
(EDS), x-ray powder diffraction (XRD) as well as analysis of urine and serum
samples, the extent and effect of fretting corrosion can be quantified and
qualitatively characterised. Insight into the various mechanisms that are
effective in the degradation processes of the modular interfaces can prove
useful in creating hypothesis that shapes the design of in-vitro experiments.
This section reviews the failure mechanisms of retrieved explants as well as
trends in the nature of metal ions released into the body.
3.2.1 Fretting and corrosion scores in retrieval studies
In the early 1990s, Gilbert et al [27] introduced a qualitative system for

characterising corrosion at the modular taper junction. It worked by
categorising explants into 4 groups: no visible corrosion, mild corrosion,
moderate corrosion and severe corrosion. In the study by Goldberg et al [67]
numbers were allocated to the listed categories as well as a criteria through
which the specific categories are qualified hence becoming an official scoring
system commonly known as the ‘Goldberg score’ (see Table 3-1). The same
scoring system was adopted and modified to charactering the extent of fretting
in other components within hip implants [68, 69]. It had also been adopted for
characterising corrosion products transferred unto ceramic heads [70].
Table 3-1 – Criteria for corrosion and fretting scores [67]
39
3.2.2 Review of retrieval studies – head-neck interface
Studies conducted on retrieved implants have led to the identification of

various fretting and corrosion mechanisms at the head-neck interface. The
large retrieval analyses of 148 modular prosthesis conducted by Gilbert et al
[27] identified mechanisms such as: preferential dissolution of Co from the
CoCrMo alloy, fretting, pitting, formation of an interfacial phases (shown in
Figure 3-5a), interdendritic corrosion and intergranular corrosion. Figure 3-5b
depicts wear and corrosion products trapped within the taper interface which
would influence both the wearing mechanism and the local chemistry at the
modular interface.
The modular prostheses analysed in their study were from similar (CoCrMo –
CoCrMo) and mixed (CoCrMo – Ti6Al4V) material combinations. Their results
reveal that a higher percentage of CoCrMo head taper in the mixed category
experienced significant corrosion damage compared to the similar category.
This agrees with the observation of Goldberg et al [67] in whose study, 231
cases were analysed. Further observations from these studies [27, 67] and
another [71] also shows that the CoCrMo head taper is generally measured
with a higher corrosion score than the trunnion taper.
In cases where ceramic bearings were used instead of CoCrMo bearings, the
trunnion tapers coupled with CoCrMo heads were observed with a higher
degree of fretting corrosion damage than the ceramic – metal trunnion tapers
[72]. The study by Kurtz et al [72] also identified that fretting corrosion scores
were lower when the ceramic bearings were coupled against titanium alloys
(Ti6Al4V and TMZF) compared to CoCr heads. More recently, a study solely
comparing the material losses of a single CoCrMo stem design of 48 retrieved
explants from CoCrMo heads and ceramic heads concluded that there were
no significant difference in the material losses from both male tapers [73].
40
(a) (b)
Figure 3-5 – Fretting corrosion at the head-neck interface: a) evidence

of Ti-Cr-Mo interfacial layer [27] b) evidence of corrosion products on
the trunnion surface [67].
3.2.2.1 Factors influencing fretting corrosion at the head-neck interface
Other trends related to fretting corrosion of the head-neck interface have been
explored for example, Goldberg et al [67] assessed the effect of implantation
time, metallurgy and stiffness. Their study observed a higher degree of
corrosion at the head and neck tapers with longer implantation time while
fretting alone was found to be unaffected by implantation time. With respect
to metallurgy, corrosion scores for both wrought CoCrMo and wrought Ti6Al4V
necks were not different however, their fretting scores differ with Ti6Al4V alloy
scoring higher. The cast CoCrMo heads measured slightly higher corrosion
scores than the wrought CoCrMo heads while the vice versa was the case
(even more significantly) when considering their fretting scores. In addition, it
was also observed that the stiffer the neck i.e. using CoCrMo instead of
Ti6Al4V, the lower the probability of corrosion of both head and neck
components.
3.2.2.2 Rare cases of trunnion fracture at the head-neck interface
Fracture of the neck at the head-neck interface of THA is rarely observed and
hence rarely reported in literature however in a rare case of poorly
41
manufactured CoCrMo alloy, intergranular corrosion led to the fracture of two
CoCrMo stem components [26]. More recently, in a specific case of 3 patients,
Morlock et al [74] analysed a M2A Magnum CoCr head with a Ti6Al4V ball
head insert and a Ti6Al4V stem which cracked at the head-neck interface due
to corrosion initiated at the interface as shown in Figure 3-6. This describes a
corrosion fatigue failure mechanism in Ti6Al4V.
(c)
Figure 3-6 – Fracture of the Ti6Al4V trunnion resulting from corrosion

[74]
3.2.2.3 Wear pattern
Wear at the modular head-neck taper interface often follow a particular pattern
which is influenced by the contact area, micro-motion, load, taper angle
difference etc. Bishop et al [75] identified two distinctive wear pattern
commonly observed at the interface namely: asymmetric and axisymmetric.
These are depicted in Figure 3-7; the asymmetric wear pattern was described
to have resulted from a toggling effect due to turning-moments generated in-
vivo. Radial separation at the head-neck taper interface may result in uneven
stress distribution and thus manifested as toggling. This generally leads to
significant wearing of a local region at the proximal and/or the distal ends of
the taper. On the other hand, a form of uniform wearing of the tapers occur
42
with the axisymmetric wear pattern. In one example it was observed that the
softer titanium alloy trunnion taper wore the harder CoCrMo head bore and
migrated up to 1 mm into the CoCrMo alloy further raising interest in the
postulation that hard oxides of the titanium alloy causes the wearing of the
CoCrMo alloy [76].
Figure 3-7 – Wear pattern commonly observed at the head-neck taper

interface: Assymetric and Axisymmetric [75].
3.2.3 Review of retrieval studies – the neck-stem interface
The neck-stem taper interface is generally regarded a more severe contact in

comparison to the head-neck interface. This is partly due to higher contact
stresses experienced at the neck-stem interface as demonstrated in the finite
element (FE) diagram shown in Figure 3-8a [77]. In addition to wear and
corrosion mechanisms experienced at this interface, in-vivo fracture and cold
welding is also prevalent at these contacts causing higher levels of failure
rates [55]. From a retrieval analysis of 27 failed Rejuvenate® modular hip
arthroplasties (Stryker) by Buente et al [78], an observed characteristic wear
pattern at the region of material loss on the neck component is said to be
indicative of the failure mechanism. The failure was attributed to two main
factors: the implant design and the combination of materials used thus
43
highlighting the possible contribution of engineers and manufacturers to the
failure of modular implants.
Figure 3-8b shows an image of corrosion at the neck surface of the Adaptor
GHE/s Short Stem Modular femoral component (Eska Implants AG) which
were said to be experiencing much in-vivo corrosion due to high stresses at
the modular junction [79]. Grupp et al [80] also reported a 1.4% failure rate in
the Metha® short stem necks in a time period of 2 years from implantation
time. Other assessments in the same study showed that an improvement in
the device was observed when the material used was changed from Ti6Al4V
to CoCrMo. Thus it is observable that the stiffness of alloy (CoCrMo in this
case) plays a significant role in reducing fretting whilst enhancing the fatigue
strength of the neck-stem modular interface [55].
(a) (b)
Figure 3-8 – Neck components of bi-modular tapers a) distribution of

stresses across the neck component at the head-neck and neck-stem
interfaces [77] b) Corrosion at the neck-stem interface of a circular
neck design [79].
44
3.2.3.1 In-vivo cases of fracture at the neck-stem interface
In-vivo fracture of the neck component had been a serious problem observed
in at neck-stem interface. This is due to various causes but the most important
is the use of titanium alloy as neck component [54]. Dangles and Altstetter
[81] reported a fracture of the modular neck in an obese middle-aged man,
Figure 3-9 shows the fractured neck with the other half still fixed within the
stem bore. In their study, the cause of fracture was proposed to be a result of
increase in offset and crevice corrosion, other fractures reported by Skendzel
[82] were suggested to be related to the use of long varus style modular neck
which is said to increase bending moment by 32.7% compared to short varus
stems. The study also concluded that the stem design played a significant role
in fracture of the neck. Ti6Al4V alloy necks have been shown to be more prone
to fracture than CoCrMo alloy and a study by Nganbe et al [83] gives a good
reason through their result which shows that CoCrMo alloy has a 38% higher
load-bearing capacity and 1000 times longer fatigue life than Ti6Al4V. This
explains why CoCrMo is a better material choice for the neck component.
However, considering the deleterious consequence of adding another

CoCrMo component to the hip prosthesis i.e. the risk of increased Co and Cr
ions release into the peri-prosthetic tissues; such compromise is a rather
difficult one to justify. Nevertheless, it proves a better decision than risking in-
vivo fracture. In addition, it has been reported that Ti6Al4V alloy is more
susceptible to a cold welding phenomena – a solid state joining process
occurring without heat at the interface. This has led to several reported cases
of damage of the neck component and difficulty in disengaging the neck from
the stem component [55, 84].
45
Figure 3-9 – Fracture of the neck component at the neck-stem interface
[81]
3.2.4 Corrosion products and metal ion releases
3.2.4.1 Solid corrosion product
The key implication of corrosion at modular taper interfaces are the metal ion
and corrosion products released into the peri-prosthetic tissues and blood
stream. Consequentially, this causes adverse local tissue reactions (ALTR)
that ultimately results in tissue necrosis, bone resorption, inflammation and
pain at the joint [8, 54, 85]. High levels of circulating metallic ions have been
linked to other serious health problems such as thyroiditis, auditory
disturbance, and granulomatous lesions [86]. Various form of solid corrosion
products have been found at different regions of the modular taper interface
which highlights the transportability of corrosion products in-vivo.
For example, studies conducted in the 1990s, by Jacobs et al [85, 87],

revealed that solid corrosion products observed within the taper interface were
a mixed crystalline metal-oxides of Cr, Mo and Ti (when titanium alloy is a
stem component). In addition, Co, Cr and Mo chlorides were also observed in
the same region – this is identified as region ‘A’ in Figure 3-10. However,
outside the interface, at the rim of the CoCrMo head were crystals of
46
chromium-orthophosphate hydrate-rich and in 2 unusual cases of the 10
explants assessed, a cobalt-phosphorus oxygen-rich compound was
identified – this is signified as region ‘B’ in Figure 3-10. These products have
been categorically linked with the modular interface through peri-prosthetic
tissue isolation in both mixed and similar material combinations [54].
Figure 3-10 – Solid corrosion products in retrieval. The region ‘A’ is

typically mixed metal-oxides and chlorides while the region ‘B’ is
typically metal-phosphates [87].
Studies have shown that the dominant corrosion product emanating from the
modular taper interface involving CoCrMo alloy is the trivalent Cr (3+) in the
form of Chromium-orthophosphate although it is still unclear if the hexavalent
Cr (6+) compound also forms in-vivo as a result of corrosion [8, 77, 85, 88].
Particles of corrosion products observed were as large as 500 µm in size and
it has been found that these corrosion products can migrate to the bearing
interface as a third-body, causing increase wear of the softer Poly-Ethylene in
a MoP combination [87, 88].
Several studies have also shown that the products obtained from the titanium
based alloys are not precipitates of corrosion products as observed in CoCrMo
and Stainless steel, but rather, examinations of tissues around the stem reveal
that titanium, aluminium and vanadium were measured in similar ratios to their
47
alloy constituents thereby suggesting that the products are wear particles of
the titanium based alloy [89-91]. Nevertheless, it is worth noting that
subsequent oxidation of the particles are possible due to increased surface
area which encourages further oxidation thus forming oxides and other
corrosion products. Physical and histological inspection of the peri-prosthetic
tissue have revealed discolouration and titanium particles within macrophages
and fibrocytes of densely collagenous membranes but these have not be
correlated to the presence of inflammatory cell or necrotic debris at tissue-
implant interface [90, 91]. These findings all point to CoCrMo alloy as the focal
problem at the modular taper interface.
3.2.4.2 Metal ion release
Metal ion release from implant corrosion are generally measured by analysing
urine and serum samples. It has been shown that even in well-functioning hips
metal ion levels in serum can be upto 3-folds higher of titanium and 5-folds
higher of CoCr than in control samples [92]. An elevated measure of Co ions
beyond the range expected to be found naturally in the body serves as one of
the first indications of a failing implant. In the study by Jacob et al [85] urine
and serum analysis showed that Co ions were elevated in the serum while Cr
in the urine samples from patients with moderate and severe corrosion in
comparison to those with ‘none’ or mild corrosion. Comparison between
patients with implants and those without revealed significantly higher urine
concentration of Co and Cr for those with implants. A recent study by Scharf
et al [8] characterising the extent of Cr and Co-related toxicity concluded that
Cr (3+) and Co (2+) were responsible for damages to proteins via an oxidation
mechanism and Fenton reaction and they observed a positive correlation
between tissue oxidative damage and the concentration of Cr (3+) and Co
(2+) ions present in the tissue. Co toxicity has also being linked to damage in
the skin and respiratory systems [93].
48
3.2.5 Metallurgical transformations in explants
Although metallurgical transformation of CoCrMo alloy have been well studied

at the bearing interface [94, 95]. Only very few studies have examined the
subsurface metallurgy of CoCrMo and titanium alloy explants from the
modular taper interface. Zeng et al [96] focused their study on the oxide film
at the surfaces of CoCrMo head and male tapers but no detailed discussion
on metallurgical transformations were made. Gilbert et al [97] examined the
fracture of a Ti6Al4V neck and described the corrosion mechanism as an
“oxide induced stress corrosion cracking” (OISCC) mechanism but the
discussion on metallurgical transformation was also insufficient. However, the
study by Bryant et al [98] examined both female and male taper of CoCrMo
and TMZF respectively. Their study provided a good insight into how worn and
unworn regions within the taper explant translate to subsurface
transformations. In the CoCrMo alloy, loss of nano-crystalline layers were
observed at regions that had experienced mostly gross sliding regime
whereas, an unworn patch which had experienced a partial slip regime was
observed with grain refinement ~300 nm in its subsurface from the top surface
as well as evidence of directionality 1 µm deep into the alloy. This is shown in
Figure 3-11a. On the other hand, the subsurface of the TMZF at a region which
had experienced gross sliding revealed mechanical mixing and material flow
5 – 10 µm deep into the alloy while the retained unworn patch was observed
with over 2 µm deep of polycrystalline structure. Precipitation and oxidation
of elemental species were seen from the surface migrating into the subsurface
through a near-surface crack. This is also depicted in Figure 3-11b.
This study therefore strongly highlights the importance of assessing

metallurgical transformations of explants and subsequently attempting to
replicate the same under in-vitro fretting study.
49
(a) (b)
Figure 3-11 – Subsurface transformation in the partial-slip region of a)

CoCrMo and b) TMZF [98]
3.3 In-vitro studies of fretting corrosion
As aforementioned, retrieval studies are pre-requisites for a well-designed in-

vitro experiment. The challenge of fretting corrosion at the modular taper
interface has so far been shown to be multi-factorial in nature. Some examples
are: patient weight, implant design, material combination, surgical impaction
force, taper sizes and morphology, minor and major misalignment,
contamination of the interface during surgical procedures, bearing size etc.
This section reports on the two main in-vitro methods used for assessing
fretting corrosion at the modular junction of THA.
3.3.1 Common in-vitro methods for assessing fretting corrosion

in THA
Several in-vitro studies were purposed to assess the metallurgy and

electrochemical characteristics of CoCrMo and Ti6Al4V alloys in different
physiological solutions and electrochemical potentials. Subsequently, in-vitro
studies of these alloys under fretting conditions in similar, mixed-metal and
ceramic-metal material combinations are undertaken. Examples of
50
controllable parameters in these studies are: physiological solution, pH, load,
displacement, electrochemical potential etc. Elucidating the effect of each
individual parameter or their combined effect is achievable using a tribometer.
This is a cost effective way of understanding relevant biomedical materials in
different environments. For this approach, the relevant material combinations
are set-up either in a flat-on-flat, pin-on-plate or ball-on-flat configuration. The
tribometer can be instrumented with a 3 – electrode cell set-up so as to
monitor in-situ electrochemistry as shown in Figure 3-12. Post-test analysis of
the interface typically include surface and chemical analytical techniques such
as electron microscopy, interferometry and spectroscopy. These are expected
to yield the following results: wear volume, wear mechanisms, corrosion
current, ion release, corrosion products etc.
Figure 3-12 – Fretting tribometer instrumented with in-situ

electrochemistry for the assessment of fretting corrosion [99]
However, a more realistic level of assessing fretting corrosion is one whereby

implant materials are assessed using geometrical and biomechanically
relevant taper interfaces. ASTM F 1875 – 98 [100] is the specific standard that
prescribes the loading condition and test set-up for fretting corrosion
assessment. Studies can either test material combinations to this standard in
its entirety or partly in accordance to it. In-vitro assessment of fretting
corrosion at this level offers a more relevant quantitative information regarding
wear and corrosion than the tribometer studies. However, this comes at a
51
higher financial cost than the tribometer studies. For this configuration, fatigue
loading cycles are imposed onto the femoral head in a physiological media
and factors such as: surgical impaction force, frequency, taper size and
morphology, misalignments, head size etc. can be investigated either in
isolation or in combination. Figure 3-13 shows an example of the set-up with
incorporated 3-electrode cell for in-situ electrochemistry.
Figure 3-13 – Fatigue assembly of the Morse taper interface

instrumented with in-situ electrochemistry [101].
3.3.2 Review of tribometer studies
3.3.2.1 Contact conditions
Studies conducted using a fretting tribometer are generally assessed under a

broad range of load and fretting parameters. The contact pressures used in
several studies ranges between 10 MPa to 4.2 GPa while for fretting
displacement amplitudes, the range varies between 1 to 200 µm [99, 102-
105]. Such a broad range is partly due to various studies attempting to
understand fretting corrosion in the context of fretting regimes – the realised
52
fretting regimes at an interface is determined by the combination of load and
fretting displacement imposed onto the contact. Although contact pressures
and fretting displacements is not known to ever reach as high as 4.2 GPa and
200 µm respectively in-vivo, diverse daily living activities (DLA) ranging from
running, to stumbling do occur in-vivo thus manifested as varying contact
pressures and fretting regimes at the real modular taper contact in-vivo.
Furthermore, a study by Donaldson et al [106] using FE analysis had shown

that contact pressures at the modular taper may increase by 85 MPa for every
0.1° of angular misalignment. This highlights the impact of surgical errors on
interfacial contact pressures. Their study also showed that, fretting at the
interface may increase by 1 µm for every extra 10 kg of patient weight. Lanting
et al [77], using FE analysis showed that the contact stresses experienced by
the titanium stem can be as high as 1.6 GPa on average at the neck-stem
interface of the modular taper junction while in another study, Zhang et al [107]
simulated peak contact pressures of 65 MPa at the head-neck interface. Thus
the contact pressures and fretting displacements varies with different taper
contacts and under different contact conditions.
3.3.2.2 Fretting corrosion of various material combination
The fretting contact of three main material combinations namely CoCrMo –

CoCrMo, CoCrMo – Ti6Al4V and Ti6Al4V – Ti6Al4V were assessed by
Swaminathan and Gilbert [99, 108]. Their study examined the effect of
electrochemical potential, frequency and load. It was observed in both studies
that the Ti6Al4V – Ti6Al4V combination was more susceptible to fretting
corrosion than CoCrMo – CoCrMo and CoCrMo – Ti6Al4V. The two contacts
involving CoCrMo, was found to have similar coefficient of friction (cof) of 0.3
and Ti6Al4V – Ti6Al4V had friction of ~0.7 during gross slip sliding at a contact
pressures of ~73 MPa.
The fretting current measured for Ti6Al4V – Ti6Al4V contact (3 mA/cm²),

superseded that of CoCrMo – CoCrMo contact (1.2 mA/cm²) which was found
to generate double the fretting current of CoCrMo – Ti6Al4V (0.6 mA/cm²) for
the same contact pressures. The lesser currents measured at the CoCrMo –
53
Ti6Al4V contact was due to higher interfacial compliance which reduces the
overall amount of fretting displacement incurred at the interface in comparison
to the CoCrMo – CoCrMo contact [99, 109]. It was also observed that the cof
varied with potential for all three combinations however, the Ti6Al4V – Ti6Al4V
was the most influenced. With respect to effects of frequency, this was
observed to increase linearly with fretting current density only at regions of
stable passive film [108].
3.3.2.3 Effect of solution chemistry on fretting corrosion
From fundamental basics of crevice corrosion, the crevice at the modular

taper junction ultimately becomes de-aerated and pH drops to acidic levels
[27]. Studies have therefore investigated the effect of pH on fretting corrosion
and in one study by Royhman et al [103], increased fretting corrosion currents
were measured at all displacements amplitudes ranging from 25 - 200 µm and
all at the pH of 3.0 and pH of 7.6. Another study which focused on the effect
of Molybdate ions identified that Molybdate ions promoted the formation of
tribofilm on CoCrMo surface at low pH (4.5) compared to pH 7.6 – both studies
were conducted on CoCrMo – Ti6Al4V combination [110].
The role of proteins at the tribological interface of metal have been well
documented. Studies by Wimmer et al [111, 112], Martin et al [113] and
Matthew et al [95], all of whom are from the same group, have demonstrated
that tribofilms consisting predominantly of denatured proteins form at a metal-
on-metal sliding interface contributing to low wear, friction and corrosion
resistance. However, as it was understood in their findings that the Molybdate
ion Mo(VI) enhances the formation of these tribofilms by interacting with
proteins in physiological solution, a recent study [110] from the group applied
this to the fretting contact of CoCrMo – Ti6Al4V and observed that addition of
Molybdate ions decreased the release of Co ions by 32% whilst increasing the
release of Ti ions. This can be regarded as a more favourable exchange.
54
3.3.3 Review of in-vitro studies using realistic taper components
One of the earliest in-vitro studies aimed at assessing parameters such as the
effect of material combination on fretting corrosion was conducted by Cook et
al [114] in the early 1990s. Their study found that all material combinations
which included: wrought CoCrMo – Ti6Al4V (smooth, rough and nitrogen
implanted tapers); CoCrMo (10mm neck length increase) – Ti6Al4V (smooth)
and Zirconia – Ti6Al4V (smooth and rough tapers) generated wear debris in
saline solution at a uniform range of 0.255 – 2.306 µm. The volume of particles
were found to be greatest in the first 1 million cycles and it was only the
roughened and nitrogen-implanted surfaces that had lesser number of
particles than the rest. In the study by Gilbert et al [115], Cook et al [114] and
others [101, 116, 117], higher offsets and increase in neck length showed
greater fretting and corrosion damage as well as increased particle generation
compared to components with 0 mm offset. This indicates the role of
increased moment arm in the increase of wear. In addition, dimensional
mismatch especially at the metal – metal interface were found to be a factor
which led to significant rise in particle generation [114].
In the same study by Gilbert et al [115] Stainless steel (SS) – CoCrMo couple
were observed to be more susceptible to fretting corrosion than CoCrMo –
CoCrMo combination as the corrosion currents were much larger in SS –
CoCrMo than for CoCrMo – CoCrMo at the end of 1 Million cycles of fatigue
loading. The rationale for the SS – CoCrMo material combination is mostly
aligned with economic reasons because the combination is widely used
outside the United States in the interest of cost reduction – an approximate
cost of manufacturing high-nitrogen SS stem is 50% less than the CoCrMo
alloy equivalent [54].
3.3.3.1 ASTM F 1875 – 98 (2014)
For the assessment of fretting corrosion mechanisms in THA, fretting

corrosion tests are conducted in part reference to ASTM F 1875 – 98 [100].
The following are the relevant prescription from the standard:
55
 Assembly load – Apply a single static load to 2000 N under dry
conditions.
 Cyclic load – Apply a cyclic load of 3 kN with minimum load of 0.3 kN
and a maximum of 3.3 kN.
 Frequency – For the 10 million cycles fretting corrosion damage
assessment, 5 Hz should be used. For other studies involving
electrochemical assessment, 1 Hz is recommended. To assess peak
to peak current, 2 Hz is suitable.
 Solution – Volume of solution can range between 5 and 100 mL.
Electrolyte solution can be 0.9% NaCl in distilled water. Proteinaceous
solution should consist of 10% solution of calf serum.
3.3.3.2 Effect of impaction and fretting load
It had been hypothesised that impaction force and dynamic loading at the
modular interface has an effect on the degree of corrosion observed. Witt et
al [63] examined the effect of assembly load on contact area and observed
that at 500 N ~9% of the titanium alloy stem taper surface ridges were in
contact with the CoCrMo head and by 8000 N of assembly load, 100% of the
ridges were in contact. This corresponded to an overall surface area increase
of ~0.6% and ~16% respectively. Mroczkowski et al [118] in their short-term
study of 18,900 cycles showed that increasing the magnitude of fretting load
led to an increase in fretting current during dynamic loading for both CoCrMo
– CoCrMo and CoCrMo – Ti6Al4V combinations. The study, which uses
realistic components with realistic taper geometry observed higher fretting
corrosion current from the mixed material combination compared to the similar
material combination thus agreeing with observations from in-vivo explant
analysis [27, 67] but disagreeing with tribometer observations [99]. This
highlights the relevance of tribocorrosion assessment using realistic
components and configurations for quantitative assessment of corrosion. A
recent study, featuring a long-term 5 million cycles test of the same CoCrMo
– CoCrMo and CoCrMo – Ti6Al4V combinations demonstrated that greater
impaction force resulted in less fretting corrosion currents regardless of
engagement length, surface roughness or material combination [119].
56
3.3.3.3 Effect of frictional torque and bending moments
Poor lubrication of large femoral heads may result in increased frictional forces
that are transmitted to the modular taper junction. In-vivo edge loading and
misalignment have also been linked to high torque generated at the bearing
interface [120, 121]. The study by Bergmann et al [122] has shown through
in-vivo load measurement of DLA that average torsional moments acting on
the hip may reach up to 17.1 Nm for activities such as stair climbing and as
high as 70.5 Nm in extreme cases of stumbling. In-silico approach by Farhoudi
et al [123] has shown that maximum torque acting at the head-neck interface
ranges from 2.3 – 5.7 Nm and bending moment ranges from 7 – 21.6 Nm.
Subsequently, in-vitro studies evaluating how changes in torque level and

bending moment may effect fretting and corrosion at the modular taper have
been a recent focus. Jauch et al [124] identified for CoCrMo – Ti6Al4V
combination that torques as low as 3.9 Nm was enough to remove the passive
layer at the modular interface for an impaction force of 4.5 kN. However, at a
higher impaction force of 6 kN, the torque required was almost double at 7.2
Nm. Furthermore, the study by Panagiotidou et al [101] showed that for
CoCrMo – CoCrMo, CoCrMo – Ti6Al4V, and BIOLOX – CoCrMo
combinations, increase of bending moment through increase in offsets was
found to increase fretting corrosion currents at the taper interface [101]. These
studies highlight the effect of torsional and bending loads on fretting corrosion
however, the degree of this contribution to fretting and corrosion relative to the
axial load is unclear.
3.3.3.4 Effect of contact area and topography
The use of shorter male tapers were one of the solutions introduced to
increase range of motion and hence prevent impingement in hip prosthesis.
The result of this was a smaller surface area of engagement at the head-neck
taper interface. Panagiotidou et al [125, 126] found that the reduction in
surface area led to higher fretting corrosion currents due to concentrated
bending loads at the reduced contact region. It was also observed in their
57
study that the roughened surface finish also enhanced fretting corrosion at the
interface through significant increase in surface roughness.
3.3.3.5 Micromotions
Micromotions at the modular taper interface of head-neck and neck-stem

taper junctions is a pre-requisite for mechanically assisted crevice corrosion
(MACC – discussed in section 3.5.2 ). Several in-vitro studies have identified
factors that influences micromotions at the modular interface. An in-silico
study by Dyrkacz et al [127] has demonstrated that micromotions at the head-
neck interface may vary significantly depending on: head size, material
combination, assembly load, taper size, offsets and angle of loading. They
measured the highest micromotion of 49 µm when a 500 N assembly load was
applied. It was observed that in-vitro measurements of micromotion at the
head-neck interface is scarce amongst studies assessing realistic
components whereas, the contrary is the case for the neck-stem interface.
In the study by Jauch et al [128] micromotion at the neck-stem is shown to be

variable for Ti6Al4V – Ti6Al4V combination depending on the taper design.
However, independent of the design, the study also revealed that stems made
of titanium alloy resulted in higher interface micromotion (fretting) than those
made of CoCrMo alloy. Similar observations were made by the same author
in a separate study where loads were varied from 2.3 kN (walking contact
forces) to 5.3 kN (stumbling loads) [129]. Micromotions has also been shown
to be directly influenced by the seating distances of the modular taper which
in itself is directly influenced by the magnitude of the assembly load [64, 128]
and micromotion decreases with larger negative taper angle difference
between male and female tapers [64].
3.4 Effect of proteins on corrosion of biomaterials
The synovial fluid in the human joint consists of an ultra-filtrate of blood

plasma, hyaluronic acid and glycol-proteins. The composition of albumin in
the synovial fluid can be as high as 70% for arthritic patient whereas this can
58
be up to 20% lower when bovine calf serum (BCS) is used to represent the
synovial fluid [130]. It is important to highlight such a significant difference
because several studies have shown that protein concentration has a big
influence on both tribological and electrochemical behaviour of biomaterials
in-vivo [94, 111-113, 131, 132]. One of the most influential mechanism through
which protein effect the surface of the biomaterial is adsorption. This
interaction is generally governed by chemical composition, surface energy,
roughness and the topography of the surface in contact with the proteinaceous
fluid. In general, the role of protein on the corrosion of biomaterials is one that
is dependent on pH, the bulk alloy material and the concentration of EAS in
the solution [132-134].
3.5 Tribocorrosion
Tribocorrosion at the interface of two materials in relatively motion is a

degradative process involving the tribology; electrochemistry and the
synergistic effect of both processes. The damages inflicted by tribocorrosion
processes in hip prostheses are more commonly studied at the sliding
interface of the bearing components; the fretting interfaces of the head-neck,
neck-stem and stem-cement contacts. The conductive physiological solution
in the joint space is the electrolyte that completes the electrochemical system.
Tribocorrosion of conductive biomedical materials in relative motion can

generally be characterised by assessing the total mass loss ‘T’ of the materials
in contact. Equation 3-1 expresses ‘T’ as the sum of: ‘W’ – the pure wearing
process in the absence of oxidational effects; ‘C’ – the pure corrosion process
or dissolution of the metal in an electrolyte in the absence of mechanical wear
effects; and ‘S’ – the synergies of both oxidation and mechanical processes.
The synergy can be further expressed as the sum of two components namely:
‘W c’ that is, wear enhanced corrosion – oxidational processes enhanced as a
result of wear at the contact and ‘Cw’ that is, corrosion enhanced wear – wear
processes enhanced as a result of oxidation at the interface [135]. Equation
3-2 shows another general denotation of the expression in Equation 3-1. In

this denotation, total mass loss ‘T’ is denoted ‘𝑀𝑡𝑜𝑡𝑎𝑙 ’ and both wear and
59
oxidational processes are categorised as ‘𝑀𝑐ℎ𝑒𝑚 ’ the sum of C and 𝑊𝑐 and
‘𝑀𝑚𝑒𝑐ℎ ’ the sum of W and 𝐶𝑤 [136].
𝑇 =𝑊+𝐶+𝑆
3-1
𝑆 = 𝑊𝑐 + 𝐶𝑤
𝑀𝑡𝑜𝑡𝑎𝑙 = 𝑀𝑐ℎ𝑒𝑚 + 𝑀𝑚𝑒𝑐ℎ
𝑀𝑐ℎ𝑒𝑚 = 𝐶 + 𝑊𝑐 3-2
𝑀𝑚𝑒𝑐ℎ = 𝑊 + 𝐶𝑤
Mathew et al [137] identified four key areas that is to be considered and

monitored when addressing a tribocorrosion system (see Figure 3-14). The
four areas are namely: 1.) Mechanical factors such as normal force, velocity
etc.; 2.) Electrochemical factors such as potential, repassivation kinetics etc.;
3.) Material properties such as hardness, microstructure etc. and 4.) Solution
chemistry such as pH, conductivity and so on. In this vein, it is deducible that
tribocorrosion is a complex and multifactorial subject to which several factors
are to be controlled in order to obtain any meaningful information from a single
factor.
Figure 3-14 – Considerations of a tribocorrosion systems [137]
60
Mischler et al [138] described a model tribocorrosion system in which three
distinctive zones experience different degradation mechanisms. Figure 3-15
shows a model wear surface representing a sphere-on-flat reciprocating
sliding contact. Zone ‘I’ is characterised by deformation wear of the contact
area – both elastic and plastic deformation of the material occurs in that region
leading to the depassivation of the oxide layer. Zone ‘II’ is a worn region
whereby the oxide layer is removed and exposure of the unprotected bulk
metal encourages anodic dissolution of the region. The two main damage
mechanisms at this region are: detachment of metallic wear particles
instigated by crack of oxide layers and accelerated wear enhanced corrosion.
Zone ‘III’ is the mechanically unaffected region. Pure corrosion is the dominant
damage mechanism at this region. It is postulated that this region accelerates
the pure corrosion occurring in zone II as a result of a galvanic coupling
established between the noble zone III and worn zone II. A specific case of
tribocorrosion i.e. fretting corrosion in which third-body products have a great
effect on the corrosion currents and wear mechanism is discussed in the
following sub-section.
Figure 3-15 – A model of tribocorrosion system in a sliding contact

[138]
Stack and Chi [136] developed a criteria by which the degradation mode of a
tribocouple in a tribocorrosion system can be predicted based on the ratio of
𝑀𝑐ℎ𝑒𝑚 to 𝑀𝑚𝑒𝑐ℎ (see Figure 3-16). Although this criteria was developed using
a sliding contact, a recent study by Bryant and Neville [139] adopted the
61
criteria for a fretting interface using a self-mated CoCrMo contact. The study
further highlighted the importance of breaking down the contribution of
corrosive damage 𝐶𝑤 which is generally attributed to 𝑀𝑚𝑒𝑐ℎ only into both its
mechanical and oxidation components.
Figure 3-16 – Stack and Chi’s criteria of degradation mechanism in a

tribocorrosion system [136].
3.5.1 Tribocorrosion in fretting
Third-body wear and corrosion products form at the interface of material

couples in a tribocorrosion system. Often times in a fretting contact, these
remain within the interface and thus influence the subsequent tribocorrosion
characteristics of interface. In the specific case of fretting corrosion, the
influence of third-body products are significant. For example, a material couple
in a gross slip fretting regime with relatively large displacement amplitudes
can behave quite similar to a reciprocating sliding contact. Therefore, its
tribocorrosion characteristics may agree in part with the model shown in
Figure 3-15 [104]. However, for a fretting contact, ‘zone I’ in Figure 3-15 would
cover a significant proportion of the entire wear scar as depicted earlier in
Figure 2-12. Therefore, accumulation of third-body products within the wear
scar may significantly reduce the exposed areas (anodic sites) which would
have undergone oxidational processes.
The material generated at the fretting contact that are detained at the interface
has been found in some studies to influence the electrochemical activity at the
interface by reducing fretting corrosion current [110, 140]. In another study
which employed electrochemical impedance spectroscopy (EIS), their result
showed that polarisation resistance of the fretting interface upon fretting
62
ceasing was increased thus indicating that the fretting contact is more
corrosion resistance after fretting than prior [103].
On the contrary, the formation of third-body objects at the fretting interface

may also result in accelerated corrosion current as observed in the study by
Landolt and Mischler et al [47]. The study highlighted that the third-body
product formed at the interface of their contact was a large particle that was
removed from one of the metal surface through an adhesive wear mechanism.
This created the exposure of a large anodic site in which oxidation reactions
were accelerated. Furthermore, the study highlighted that the accumulation
of third-body wear (if unejected) may modify the pressure distribution at the
fretting contact thus reducing the rate of mechanical wear as well as reducing
the area of unprotected bare metallic surface corresponding to lesser
corrosion currents. This shows that the role of third-body products at the
fretting interface may be both beneficial and adverse within a tribocorrosion
system.
3.5.2 Mechanically assisted crevice corrosion
Mechanically assisted crevice corrosion is the widely accepted mechanism

that accelerates the degradation of the modular device in THA. The
mechanism was officially named by Gilbert et al [27] although the process had
been previously described in part by Collier et al [30]; it agrees with the
fundamental processes of crevice corrosion described in section 2.4.2.4.
MACC in the context of modular taper junctions of hip prosthesis is initiated
by micromotions at the taper interface. The process is described as follows:
- At the initiation of fretting within the modular interface, shear stresses
cause the protective oxide films to fracture.
- Oxygen-rich solution migrated into the interface interacts with the
freshly exposed bulk metal (which is highly reactive) to restore the
oxide layer at the site.
- The repetition of the above process ultimately result in the depletion of
oxygen within the crevice while increasing the concentration of free
positive metal ions. Depleted oxygen also mean that the region is
63
unable to repassivate thus metallic dissolution in the region is
accelerated.
- Excess of positive metal ions attract EAS such as Cl- to the crevice thus
forming metal-chlorides etc.
- Subsequently, the metal-chlorides react with water to form metal-
hydroxide and hydrochloric acids thus lowering the pH significantly in
the crevice.
3.5.3 Repassivation characteristics of CoCrMo and Ti6Al4V
Studies by Gilbert et al [141], Goldberg and Gilbert [142] characterised the

passive layer on Ti6Al4V and CoCrMo alloys respectively, using a scratch test
method. Their study observe that for the Ti6Al4V alloy, aeration had little effect
on the repassivation mechanism of passive layer once scratched and that
higher peak currents were associated with low pH and higher electrochemical
potential. The time taken for passive film restoration were also observed to
vary with pH and potential. In a similar manner to Ti6Al4V, aeration was found
to have no significant effect on the repassivation process of the CoCrMo alloy
and this was said to be due to the availability of oxygen through the hydrolysis
of water. It was observed for CoCrMo that above a certain potential (300 mV);
high peak currents and lower time constants were measured for lower pH of
2 compared to pH 7 thus highlighting the joint effects of lower pH and potential
on the repassivation behaviour of CoCrMo.
A general assessment to determine which oxide is restored faster was

conducted in 0.9% saline for both Ti6Al4V and CoCrMo and their finding
revealed that Ti6Al4V had a ~35% faster surface oxide regeneration time
compared to CoCrMo [143, 144]. Two crucial observations from both Gilbert
and Goldberg studies highlighted key differences in the passivation behaviour
of both alloys in phosphate buffered saline (PBS): a) CoCrMo displayed a
transpassive behaviour at 500 mV while no transpassive behaviour was
observed for Ti6Al4V b) peak currents were measured for the Ti6Al4V alloy
from as early as -700 mV while for CoCrMo this only began between -400 mV
and -300mV. The authors concluded that this suggests that corrosion current
64
is generated from the point of oxide-substrate shear failure for Ti6Al4V
whereas, permanent deformation (abrasion) of the oxide from the substrate is
required for corrosion current generation in CoCrMo.
3.6 Summary
This chapter has briefly reviewed, from the inception of modularity in THA, the
design and challenges associated with the modular taper junction. Analysis of
implant retrieval have proven to be a suitable approach for semi-quantitative
assessment of the extent of fretting corrosion damage at the modular taper
interface. The approach had been useful for identifying several factors and
trends related to fretting corrosion i.e. material combinations, implantation
time, stiffness and so on. However, retrieval studies are limited, as the insight
it offers into the outcomes of several mechanical and electrochemical
processes occurring in-vivo may be regarded inadequate. Thus, it is from this
point that in-vitro studies may prove necessary for identifying in-situ wear and
corrosion mechanisms. The studies would be specifically designed to
elucidate the role of individual factors that contribute to the multi-factorial
problem of fretting corrosion. Although the review of both tribometer and
realistic taper component assessment of fretting corrosion had highlighted the
following: effect of pH, contact loads, surface topography, impaction force and
many more, there are yet still several gaps to be filled through an in-vitro
approach. The subsequent section details the gaps that are still present in
literature.
3.6.1 Gaps in literature
 A study conducted using a fretting tribometer will prove much cheaper

than those conducted at component level. However, to the best of this
review, no study using a tribometer has been able to replicate the
composite contact condition of the modular taper in order to generate
the diverse forms of corrosion products seen on explants. Successfully
65
achieving this may prove economically beneficial for assessing THA
wear and corrosion mechanism.
 Retrieval studies have highlighted (to an extent), mechanical and

corrosion contributions that may lead to the in-vivo fracture of titanium
alloys. However, a mechanistic in-vitro study assessing various fretting
regimes and contact conditions of materials couples involving titanium
alloy are not well studied. Such study will help to better understand the
mechanical contributions and the corresponding electrochemical
response to localised corrosion which may be contributing to titanium
fracture.
 No study has been found that has correlated fretting regime and
interfacial wear mechanisms to the metallurgical transformations
occurring in the subsurface of CoCrMo and Ti6Al4V. Moreso, only a
few studies have assessed the subsurface degradation of both alloys
from a retrieval studies.
 The principle of tribocorrosion and the mechanisms associated such as

the effect of third-body have been studied for several materials.
However these are commonly in sliding conditions. Exploring the
synergistic effect of wear and corrosion as well as understanding the
contact conditions that optimises synergistic behaviour or antagonistic
effects within a fretting interface is very much lacking in literature.
 Several studies have applied the Goldberg fretting score to semi-

quantitatively characterise the degree of fretting corrosion damage in
metal – metal and ceramic – metal systems. Other techniques such as
volumetric loss have also used for comparison purposes. However,
there is still a need for an in-vitro assessment of the tribocorrosion
characteristics of these two systems under various loading conditions
which represent DLA. Perhaps, the effect of replacing a metallic
bearing with a ceramic one on fretting corrosion can be elucidated.
66
Chapter 4 Experimental methodology and surface analytical
techniques
4.1 Introduction
This chapter outlines the experimental methods and analytical techniques

employed in this research. It provides detailed protocol and rationale for the
tribocorrosion experiments, surface and subsurface analyses undertaken in
this study. The fundamental theories pertaining to each technique are also
summarised in this section.
4.2 Experimental materials
The two metallic alloys studied in this thesis are namely: wrought low carbon
Co28Cr6Mo and Ti6Al4V. These are abbreviated to CoCrMo and Ti6Al4V
respectively. The ceramic materials are namely: Biolox®delta (Biolox) and
Si3N4. All tests were conducted in protein containing physiological solution as
outlined in the sub-section below.
4.2.1 Metallic and ceramic sample preparation
CoCrMo and Ti6Al4V alloys used for ball and flat components were wrought
manufactured according to the specifications in ASTM F 1537 and ASTM F
136 outlined in Table 2-1 and Table 2-2 respectively. The materials were
supplied as a 25 mm diameter bar which were cut into plates (flats) of 6 mm
thickness and the CoCrMo femoral heads were 28 and 36 mm diameter. For
the assessments involving realistic taper components, the alloys were
supplied as 12/14 taper spigots with 0 mm offset by Aesculap, Germany. The
manufactured CoCrMo heads were highly polished to a roughness (Ra) < 0.02
µm. The surface of the 6 mm flats after being cut were firstly grounded using
Silicon carbide grinding paper then polished to a mirror finish (Ra < 0.02 µm)
using a 3 µm diamond paste and polishing paper by Bruker Corporations, UK.
67
Commercially available G28 Si3N4 ball bearing with a maximum Ra of 0.05 µm
and diameter of 25.4 mm (1 inch) was purchased from R&M bearings, UK.
The polished Si3N4 plate had a maximum Ra of 0.08 µm and was purchased
from Keranova AB, Sweden. Femoral Biolox heads of 28 and 36 mm diameter,
manufactured to an ultra-smooth Ra of 0.002 µm were purchased from
Aesculap, Germany. The 28 mm diameter CoCrMo and Biolox heads were
only used for the tribometer studies and the 36 mm heads for the taper cyclic
loading tests.
4.2.2 Test environments and lubricants
All tests were conducted in 25% (v/v) diluted Foetal Bovine Serum (FBS),
Seralab, UK, in part reference to BS ISO 14243-3:2004. 0.03% (v/v) Sodium
Azide, G-Biosciences, USA, was added to the solution in order to slow down
the growth of bacteria during the long-term cyclic fatigue tests. The solution
was balanced with isotonic phosphate buffer saline (PBS), Sigma Aldrich, UK
to represent the osmolality and ion concentration of the human joint fluid. The
pH of the solution was 7.6. A list of ionic and protein contents can be found
Appendix A.
4.3 Electrochemical technique
Throughout this study, in-situ electrochemical measurements were utilised to

characterise the corrosion mechanisms of fretting interfaces in both metal –
metal and ceramic – metal fretting contacts. The Autolab PGSTAT101,
Metrohm, Switzerland and the CompactStat, Ivium technologies, Netherlands
potentio-stats were used in two separate studies. The potentio-stat connects
to a 3-electrode cell which comprises of: the working electrode (WE) – where
connections were taken from the metallic alloys (CoCrMo and Ti6Al4V) in the
experimental set-up; the reference electrode (RE) – Ag/AgCl (+ 0.196V vs.
Standard Hydrogen Electrode (SHE)); and the counter electrode (CE) –
platinum plate. In this study, the RE and CE exist as a single device – the
Thermo Fisher Scientific Sureflow REDOX electrode. Experiments performed
68
under OCP conditions require the WE and RE only whereas, all three
electrodes are required for the potentio-static tests. The Autolab potentio-stat
was controlled using the licensed software Nova 10.1 and the licenced Ivium
soft was used for the CompactStat.
4.3.1 Open circuit potential measurement
The value of OCP is often used as a qualitative indication of the corrosion

regime that a metal is in that is, either it is in an active state or a passive state.
In the context of tribocorrosion, the value of OCP is indicative of the galvanic
coupling between the passive (unworn) site and the depassivated (worn) site
of the metallic interfaces involved it the contact [145]. The passive region is
mainly dominated by cathodic reactions while the depassivated region is
dominantly anodic reactions. Furthermore, the value of OCP for a mixed-metal
contact i.e. CoCrMo and Ti6Al4V alloys rubbing against each other represents
the sum of galvanic coupling occurring between the worn and unworn regions
of both alloys. As each alloy have a different electrode potential, the OCP is a
composite of several galvanic interactions. The OCP as an electrochemical
technique does not provide information regarding the reaction kinetics.
However the following are theoretical parameters that affect the OCP
measured in a tribocorrosion system [146].
 Characteristic corrosion potential of worn and unworn surfaces

 Ratio and position of the worn area relative to the unworn
 Kinetics of electrochemical reactions
4.3.2 Potentio-static polarisation
In a potentio-static test, a fixed potential (anodic or cathodic) is imposed unto

a material using the 3-electrode cell. The tribocorrosion characteristics of a
material can be assessed under anodic or cathodic conditions. For the anodic
condition, an anodic current is generated by applying an over-potential
(positive potential above the OCP) to the metallic sample prior to rubbing. On
the other hand, a cathodic current is generated by applying a negative
69
potential below the OCP. At very high cathodic potentials, the material’s ability
to repassivate is suppressed thus allowing the wear characteristics of the
material to be assessed. In this study, fretting tribometer and cyclic fatigue
loading tests were conducted under anodic polarisation of 0V vs Ag/AgCl.
Potentio-dynamic studies [147, 148] have shown that both CoCrMo and
Ti6Al4V may exist in their passive region in serum containing solutions at 0V.
Thus, the potential was deemed suitable for the assessment of fretting
corrosion currents in this study.
Anodic current measured as a function of time was used to monitor the

mechanical depassivation and repassivation characteristics of the various
material combinations. Figure 4-1 shows a typical Current vs. Time plot during
a tribocorrosion test. At the point of potentio-static polarisation, the current
rapidly rises before decreasing exponentially. The peak of the anodic current
is the ‘Imax’. Subsequently, the current tends to zero (𝐼 → 0) as the time tend
to infinity (𝑡 → ∞). The area shaded in blue corresponds to the ‘Pure corrosion’
C. At the initiation of fretting, anodic reactions become dominant for as long
as fretting persists. The area encapsulated between the pure corrosion current
and the fretting current is cumulative charge representing the wear enhanced
corrosion (W c) + an unknown proportion of corrosion enhanced wear denoted
‘αCw’ (see section 8.6 for detailed discussion).
Figure 4-1 – Schematic representation of a tribocorrosion current –

time plot.
70
4.4 Experimental set-up
This section details the tribocorrosion tests set-up for both tribometer studies
and assessment of realistic taper component. A sample size of n = 3 was
conducted across all studies except otherwise stated. Error bars where used,
represent standard error of the mean.
4.4.1 Fretting tribometer with in-situ electrochemistry
The tribometer used in this study was a custom built fretting rig based in the
School of Mechanical Engineering, University of Leeds, UK. A schematic
representation of the tribometer is shown in Figure 4-2; the rig is controlled
using LabView programming software, National Instruments, UK. The
required displacement is manually entered into the program and so also, the
normal load is applied manually. Fitted electro-mechanical actuator from Data
Physics Corporations, UK, applies the desired displacement and the axial
displacement of the cantilever is measured through a metal plate positioned
in front of the fibre optic displacement sensor which is directly connected to
the load cell and shaft of the actuator. The displacement sensor is located on
a linear stage and the information collected through it is fed back to ensure
the actuator reaches a steady state at the desired displacement. The actuator
is capable of applying reciprocal motion at frequencies ranging from 1 – 30
Hz. The load cell by Kistler, USA was used to monitor the horizontal tangential
forces (Ft) during reciprocal fretting.
Figure 4-2 – Schematic diagram of the fretting tribometer

71
Figure 4-3 shows an image of the tribocorrosion test set-up. A delrin bath
holds the physiological solution – a volume of 12 – 14 ml which ensures the
ball and flat point contact is totally immersed in the solution. The solution is
kept at a temperature of 37 ± 2°C throughout all tribometer studies. A
peristaltic pump is used to circulate hot water from the beaker of water on a
thermocouple controlled hotplate. The thermocouple is positioned inside the
solution to ensure the solution is constantly at the desired temperature.
The flat metallic/ceramic sample is centred into the bath; flatness is then
verified using a spirit level. Due to the unpolished sides of the sample, and in
order to ensure that there are no electrochemical activities between the
unpolished sides of the sample and the polished surface, a VariDur 10,
Buehler, UK acrylic resin was allowed to firmly set around the sides of the
sample leaving only the flat surface of the sample exposed. Moreover, the
hard resin formed a circumferentially secured fixation of the flat component,
thus eliminating relative motion of the flat component during fretting. Once the
resin is fully set, the surface of the flat components were wiped with acetone;
this was to ensure that part of the polished metallic surface was not stained
with dried acrylic resin.
Figure 4-3 – Image of the fretting corrosion tribometer set-up
72
4.4.2 Tribometer fretting corrosion studies (test protocol)
The two electrochemical techniques employed for the tribometer studies are
OCP and potentio-static. In both conditions, a ball-on-flat configuration was
used to model a point contact. A maximum contact pressure (Pmax) of 1 GPa
was used across all tribometer studies except in one instance where Pmax of
0.77 GPa was used specifically to attain a gross slip fretting regime. The use
of a relatively high Pmax to model the modular taper were for the following
reasons: a) to establish a large asperity-asperity contact which would
generate to a reasonable volume of wear in the short-term 6000 cycles test;
b) to ensure sufficient pressures which will facilitate a crevice environment by
trapping cumulated wear and oxidised products. The crevice environment is
expected to establish a subsequent chemical transformation of metal-oxides.
c) encourage in-vitro subsurface transformations in the metallic alloy that may
take years of fatigue loading in-vivo. Other studies have taken a similar
approach for a ball-on-flat configuration and applied pressures up to 4.2 GPa
[105]. The contact half-width ‘a’ established from the least compliant material
combinations to the most compliant combinations ranged between 120 – 250
µm.
The displacement amplitudes assessed were: ±10 µm, ±25 µm, ±50 µm and
in the case of CoCrMo – CoCrMo and CoCrMo – Ti6Al4V, an extra
displacement (±150 µm) was tested. Through the diverse displacement
amplitude, all material combinations were assessed under various fretting
regimes, including transitional regimes such as the ‘mixed regime’. The largest
displacement amplitude (±150 µm) was specifically tested in order to
understand the fretting corrosion behaviour of the metal – metal couple at very
large displacements. In the case of CoCrMo – CoCrMo, ±150 µm was very
close to the limits of it Hertzian contact half-width ‘a’; in other words, it was on
the verge of transitioning into a reciprocal sliding contact where the value of
‘e’ is close to 1. Table 4-1 presents an outline of the fretting contact
parameters used in all the tribometer studies. Determination of cumulative
energy, slip ratio and transition criteria are based on fundamental work by
Fouvry et al [149].
73
Table 4-1 Fretting contact parameters
Fretting contact parameters
Pmax for all 1 GPa (except for the case of BIOLOX – Ti6Al4V where 0.77 GPa was
combinations used to attain a gross slip regime)
Load (N) ‘a’ (µm)

CoCrMo – CoCrMo 50 160
CoCrMo – Ti6Al4V 120 250
Corresponding loads
Si3N4 – CoCrMo 35 130
and contact half-
Si3N4 – Ti6Al4V 85 200
width for each
Si3N4 – Si3N4 30 120
material combination
Biolox – CoCrMo 40 140
Biolox – Ti6Al4V
110 / 50 230 / 180
(1 GPa / 0.77 GPa)
Fretting ±10 µm, ±25 µm, ±50 µm and ±150 µm (±150 µm was only
displacements tested for CoCrMo – CoCrMo and CoCrMo – Ti6Al4V)
Frequency 1 Hz
No. of Cycles 6000 (OCP) and 2000 (potentio-static)
Repetitions n = 3 (except for Biolox fretting contacts which had n = 1)
4.4.2.1 Fretting corrosion protocol (OCP conditions)
The tribocorrosion experiments reported in chapters 6 and 7 of this thesis

were carried out under OCP conditions. Figure 4-4 shows the schematic of a
typical OCP trend for a large displacement tribocorrosion system. The protocol
used for this study is incorporated in the schematic diagram and is
represented as stages. OCP was monitored throughout the three stages of
the test at an acquisition rate of 1 point per second. In ‘stage 1’ the ‘Initial
Surface Potential’ may rise (ennoblement) or decline. However the direction,
the purpose of this stage is to permit the loaded contact to equilibrate.
74
Once fretting has started, a sharp potential shift in the negative direction
(cathodic shift) may or may not occur. This depends on the amplitude of the
displacement and the applied load. The cathodic shift shown in Figure 4-4 is
given the term ‘Depassivation potential’. The potential measured subsequent
to the depassivation potential is termed the ‘Fretting potential’ i.e. ‘Stage 2’.
This may behave either of four ways: rise (ennoble), remain constant, decline
or a consecutive occurrence of the previous three. As soon as fretting ceases
in ‘Stage 3’, the recovery stage may be manifested in two ways: a sharp rise
in the ‘Repassivation potential’ followed by a slower surface ennoblement
termed the ‘Final surface potential’ or depending on the rate of the
‘Repassivation potential’, the two parts of the recovery stage may occur as
one whole recovery stage. The pattern observed in the OCP across the three
stages is often indicative of the corrosion regime of the material couple. The
maximum (Max.) cathodic shift is determined as the magnitude of the most
negative potential during fretting relative to the noblest potential from the
initiation of fretting.
Figure 4-4 – Schematic diagram of fretting corrosion protocol under

OCP conditions
75
4.4.2.2 Fretting corrosion protocol (potentio-static conditions)
Tribocorrosion tests under potentio-static conditions were conducted for better

understanding of the effect of surface history on the fretting corrosion
behaviour of THA in-vivo. The tests were conducted on self-mated CoCrMo
combination only and the results are presented in Chapter 8 of this thesis. An
anodic polarisation potential of 0V vs. Ag/AgCl was applied and the data
acquisition rate for the fretting current was 1 current point per second.
Figure 4-5 shows a schematic diagram of the fretting corrosion experiments

conducted. The same contact conditions for CoCrMo – CoCrMo outlined in
Table 4-1 were also applied in this test. However, the overall protocol were
carried out using three methods:
 In ‘Method 1’ 2000 cycles of fretting (1 Hz) were carried out for each
displacement amplitude (±10 µm, ±25 µm, ±50 µm and ±150 µm). Prior
to each test, 4000 seconds of equilibrating period was permitted in
order to allow the current transient to be as close to ‘0 amps’ as
possible.
 In ‘Method 2’, the tests were conducted for the same displacements as
in method 1. However, in this case, the tests were performed in series,
in the increasing order of ±10, ±25 µm, ±50 µm, ±150 µm with an
incorporated rest period of 500 seconds after each displacement
amplitude.
 ‘Method 3’ follows the same process as method 2, however, in this case
the tests were conducted in a decreasing order (±150, ±50 µm, ±25 µm,
±10 µm).
76
Figure 4-5 – Schematic diagram of potentio-static fretting corrosion
protocol
4.4.2.3 Calculations
Tribocorrosion mass loss was determined using Faradays equation

expressed in Equation 2-4. The charge ‘Q’ was determined by integrating the
area underneath the fretting current transient as depicted in Figure 4-1. The
molar mass ‘M’ was determined as the stoichiometric average wt%
composition for CoCrMo alloy and the valence ‘n’, thus giving a value of 55.7
gmol-1 for M and 2.4 for n respectively. These values are similar to those used
in others studies for CoCrMo [108, 150].
The contribution of mass loss due to pure corrosion ‘C’ was determined by
firstly establishing an estimated total area of the ball and flat components that
were in contact with the solution, thus contributing to the pure corrosion. The
estimated area was 11.9 cm2. However, as the total mass loss was only
measured from the fretting wear scar, the largest fretted area from both ball
and flat components was for the largest displacement amplitude which was
±150 µm; this was estimated to 7 x 10-3 cm2. Note, the pure corrosion
77
contribution to the total mass loss is estimated using the fretted area rather
than the total exposed metallic surface area. The result of this calculation are
tabulated in Chapter 8.
4.4.3 Fatigue loading set – up with in-situ electrochemistry
4.4.3.1 Component assembly
For this study, the ElectroPulsTM E10000, Instron, UK was used to apply static
assembly force and fatigue cyclic load unto a geometrically relevant head-
neck taper material combination. The machine is capable of applying ±7 kN of
static force and ±10 kN of dynamic force at a frequency up to 100 Hz. A
DynacellTM load cell is connected to the crosshead of the machine and is
capable of measuring load of ±10 kN and ±100 Nm.
In this set-up, spigots of CoCrMo and Ti6Al4V alloys (see Figure 4-6a) were
used to represent the neck of the femoral stem component in hip prosthesis.
The spigot has a 12/14 micro-grooved taper whereby Rz (Rp + Rv) is ~ 12 µm.
The femoral heads were 36 mm sized ceramic and metal head namely: Biolox
and CoCrMo respectively. The tapers were assembled with a quasi-static
assembly load of 2 kN according to ASTM F 1875 – 98, at a rate of 1 kN per
second. An image of the static loading procedure is shown in Figure 4-6b. In
order to minimise the possibility of misalignment, a spirit level is used to
ensure flatness of the spigot prior to a 50 N preload. The purpose of the
preload was to further ensure an aligned engagement of the taper is
established before the 2 kN load is applied. The heads were also pulled off
using the same set up; the dowel (shown in the picture) was used to fasten
the spigot down while the pull-off force is being applied. Figure 4-6c shows an
example of how the taper separation force (pull-off force) is determined.
78
(a) (b)
(c)
Figure 4-6 – Taper assembly: a) Dimension of spigot and surface

profile of the grooved taper, b) assembly and pull-off method, c) an
example of the taper separation force.
4.4.3.2 Component tribocorrosion cell
Figure 4-7 shows the set-up used for the fatigue study. The orientation of the
spigot holder was designed according to ISO 7206-4:2010 which prescribes a
10° adduction and 9° flexion (β). The head – neck angle (α) used was 135°
based on other tribocorrosion studies with similar fatigue set-up [101, 119].
79
The assembled taper interface was exposed to 85 mL solution which covered
the femoral head up to the line denoted ‘J’. The plastic-metal interface
identified as ‘E’ was isolated from the physiological solution by the use of
silicon sealant. The test was conducted at room temperature (20±2°C) which
minimised the amount of solution evaporated over the 2.5 days of dynamic
loading. The WE for the electrochemical connection was taken from the
bottom of the male spigot ‘C’ and both RE and CE were suspended in the
solution in close proximity to the taper interface ‘B’ thus ensuring accurate
electrochemical measurements.
Figure 4-7 – 3D representation of the fatigue set-up for assessing

fretting corrosion of realistic taper geometry.
4.4.3.3 Fatigue loading (test protocol)
The fatigue loading tests were performed under potentio-static conditions for
four material combinations namely: CoCrMo – CoCrMo, CoCrMo – Ti6Al4V,
Biolox – CoCrMo and Biolox – Ti6Al4V. Thus, both metal – metal and ceramic
80
– metal realistic taper interfaces are studied. Various loading conditions were
performed in a series of stages as shown in the schematic diagram of Figure
4-8. In ‘Stage 1’ the system was allowed to equilibrate for 1500 seconds whilst
under a static load of 0.3 kN and during which OCP was monitored.
Subsequently, anodic potential of 0V vs. Ag/AgCl was imposed unto the
system and allowed to equilibrate for at least, another 7500 seconds. In this
time, the current drops exponentially and theoretically reaches zero at time (t)
= ∞. ‘Stage 2’ is the beginning of the dynamic loading; three frequencies (1, 2
and 5 Hz) were evaluated consecutively at a constant compressive load of
1.8 kN and cyclic load of 3 kN (standard ASTM F 1875 – 98 profile). However,
as the loading condition changes from one profile to the next, 2000 seconds
of rest period is permitted in order to ensure adequate repassivation before
the next loading profile is actuated.
In-between the stages, a standard ASTM F 1875 – 98 loading profile is applied

for 5000 cycles at a frequency of 1 Hz in order to monitor possible systemic
effects at the taper interface. Such effects may influence the measured fretting
current for the subsequent loading profiles within the series. Nevertheless, the
systemic effects were minimised by testing the highest cyclic loads (which is
expected to inflict the greatest fretting micromotion) last in ‘Stage 4’. In ‘Stage
3’ fretting corrosion current for an increasing mean compressive load (0.8, 1.8,
2.8 kN) with minimal effect of cyclic loads is assessed – ±0.5 kN cyclic load
was used for this stage of assessment. In ‘Stage 4’ the effect of cyclic loading
(±0.5, ±1.5 and ±2.5 kN) at a mean compressive load of 2.8 kN at 1 Hz was
assessed. Fretting current sampling rate were 1 point per 5 seconds for n = 2
test and a higher resolution of 10 points per second for n = 1 of the tests.
81
Figure 4-8 – Protocol for fatigue loading tests on realistic taper
components
82
4.5 Surface, sub-surface and chemical analysis
4.5.1 Optical light microscopy
The Leica DM6000 M optical light microscope was used to capture a cross-
sectional image of the Resin-Redistributed wear-Ti6Al4V interface prior to
nano-indentation tests of the redistributed wear piled-up in the wear surface.
The 2D image stacking feature was used to obtain a good contrast for
depressed regions within the surfaces of the interfacial structure.
4.5.2 White light interferometry (WLI)
3D optical profiler NPFLEX by Bruker Corporations UK is a non-contact

surface measurement technique which uses white light interferometry
technology. The white light is split into two beams; the beam slitter directs one
beam towards the reference mirror while the other beam is directed towards
the sample being visualised. Subsequently, the beams are recombined to
form an interference pattern which are analysed to examine the optical path
difference (in length) of the two beams. The technique was used in this study
to obtain 2D quantitative measure of wear scar depth, 3D surface topography
and the quantitative measure of volume losses. Analysis were carried out
using the licenced Vision 64 software by Bruker.
The wear volume from the fretting contact of the ball and flat components were
determined according to the methods outlined by Fouvry et al [151]. Two
procedures (outlined below) were used for the estimation of volume loss in
this study. Procedure 1 considers ‘wear’ as the loss of matter from the wear
surface and thus, this procedure was used for all the material combinations
whereby redistributed wear is minimal. All material combinations tested in this
study are aligned with this description except for the material couples involving
Ti6Al4V. For these, wear volume was estimated using Procedure 2. In most
cases, wear at the interface of material couples involving Ti6Al4V are
redistributed within the contact thus giving a near-zero ‘loss of matter’.
Therefore, for these interfaces, the sum of the negative volume (V-) is the
83
estimated wear. This assumes that the material redistributed (V+) in the
contact originated from the wear depths i.e. the volume below the zero-axis
(V-) as depicted in Figure 4-9. However, the limitation of Procedure 2 is that,
the redistributed volume below the zero-line is not accounted for as shown in
Figure 4-9.
Procedure 1: Wear volume = (V- - V+)ball + (V- - V+)flat
Procedure 2: Wear volume = (V-)ball + (V-)flat
Figure 4-9 – Schematic for determining wear volume
4.5.3 Scanning electron microscopy (SEM)
The Carl Zeiss EVO MA15 scanning electron microscope (SEM) was used for
surface analysis in this study. High resolution micrographs and surface
characteristics to the nano-meter (nm) scale of worn metallic surfaces were
obtained. SEM offers a higher resolution than conventional light microscopes
because the wavelengths of electrons bombarded at the sample can be
several orders of magnitude smaller in SEM. Images may be obtained in
secondary electron (SE) mode where electrons fired at atoms on the
specimen surface knocks out an electron in the atom which is then scanned
to create an image. The other mode is the back scattered electron (BSE)
mode, in this technique, the fired electrons that experienced elastic collision
with atoms on the sample surface are scanned to obtain a back scattered
image. The BSE technique is also capable of producing a contrasted image
84
which gives an indication of the relative molecular weight of compounds
present on surfaces. For example, on a BSE micrograph image, objects on a
surface that have a lower molecular weight appear much darker than those
with larger molecular weight. All SEM images in this study were captured in
BSE mode at 20 k eV.
4.5.4 Focused ion beam (FIB)
Focused ion beam (FIB) is a sample preparation technique used in this study
for further electron microscope characterisation using the transmission
electron microscope (TEM). Tribocorrosion products and metallurgical
transformations of the metallic samples were characterised using the TEM. It
was anticipated that subsurface metallurgical transformation of the alloys
would be site-specific considering that while certain regions of a fretting
contact may experience a stick behaviour, another region may experience
gross sliding. Thus, considering the composite contact phenomenon in
fretting, it is important to obtain samples from a consistent region across all
material combinations and displacement amplitudes as much as possible.
Figure 4-10 shows the stress distribution for an elastic contact of a ball-on-flat
configuration plotted by Vingsbo and Schon [42]. The material used to
generate the plot in their study was Copper and it is observable Max. traction
force is ~63% away from the centre of the Hertzian contact. While the region
of Max. traction force is expected to vary across various material
combinations, 3D surface profilometry of the flat component may be used to
estimate the region. The region of Max. traction force is expected to be the
deepest region in the worn annulus of a fretting contact. Therefore, in this
study, 3D profilometry was used as a guide for selecting the region FIB
samples were obtained from, thus allowing for consistency. In some special
cases, samples were directly obtained from regions of interests. These are:
areas covered in redistributed tribocorrosion wear products and areas with
surface cracks.
85
Figure 4-10 – Elastic contact stress pattern for a ball-on- flat contact
[42]
An FEI Nova200 Nanolab dual beam FIBSEM equipped with a Kleindiek

micromanipulator was used for the sample preparation. The key steps taken
in the process are outlined as follows:
1. The area of interest of dimension ~ 15 x 2 µm is coated with 300 nm of

electron beam deposited Pt so as to protect the interested area from
ion beam damage. A further 1.5 µm thick ion beam deposited Pt
follows. This stage is depicted in Figure 4-11a. All FIB samples were
obtained across the wear scar i.e. orthogonal to the direction of fretting.
2. Trenches were removed from either side of the coated region using the
ion beam at 30kV, leaving a lamella as shown in
Figure 4-11b.
3. The lamella was thinned to 1 µm thickness, cut free and removed in-
situ using a Kleindiek micromanipulator, shown in Figure 4-11c.
4. The sample was then attached via Pt deposition to a standard 3 post
Cu TEM support (Ominprobe), shown in Figure 4-11d.
5. The sample was thereby finally thinned to electron transparency of
about 10 nm using Ga+ ions at 5kV, shown in Figure 4-11e
86
(a) (b) (c)
(d) (e)
Figure 4-11 – FIB SEM micrographs of TEM sample preparation stages

(a) Pt deposited on area of interest (b) removal of material from either
side of Pt deposit (c) removal of lamella from bulk material (d)
attachment of lamella to TEM post (e) final thinning of sample to
electron transparency
4.5.5 Transmission electron microscope (TEM)
In TEM, electron beams are transmitted through an electron transparent

sample prepared using a FIB. Digital micrographs are obtained as the electron
passes through the specimen. When compared to the light microscope, the
TEM is capable of imaging at resolutions up to a x1000 higher. This is due to
wavelength of ~ 0.2 nm for electron compared to wavelengths of ~ 200 nm in
light microscopes. In this study, samples were analysed using a FEI Tecnai
F20 FEGTEM (200kV) fitted with a Gatan Orius SC600 CCD camera. Bright-
field and dark-field images were captured using the digital camera, the former
is generated by passing a single direct beam through an objective aperture
and for the dark field, a diffracted beam is imaged instead of the direct beam.
87
The TEM is also capable of generating selective area electron diffraction
(SAED) images which holds information about the orientation of atomic planes
within a solid crystal. When electrons beams are fired at voltages as high as
200 kV, they begin to behave like waves hence travelling at de Broglie’s
wavelengths of ~ 2.96 pm. This is 100x smaller than the lattice parameters of
crystals which is about ~ 0.3 nm, thus allowing the electrons to be diffracted
between the ‘gratings’ created by atoms. Electrons undergo an elastic
scattering process when they hit the atom that is, they lose no energy and its
wavelength remains unchanged [152]. Therefore, ‘Spot’ electron patterns and
ringed patterns are generated which represents single-crystals and nano-
crystalline structures respectively.
4.5.6 X-Ray powder diffraction (XRD) analysis
X-ray diffraction provides crystallographic information of a single crystal by

providing information related to the ‘d-spacing’ between atomic planes. A
schematic diagram of the atomic planes, incidence and diffraction of beams is
shown Figure 4-12. The expression relating the d-spacing to the wavelength
of X-ray ‘λ’ and the angle of incidence theta (θ) is knowns as Bragg’s Law of
diffraction as shown in Figure 4-12. In this study, XRD analysis of polished
CoCrMo and Ti6Al4V surfaces were conducted using the Cu Kα radiation over
an area of 15 x 15 mm across a 2θ scan between 30° and 60°.
Figure 4-12 – Schematic diagram of atoms in planes separated by a

space ‘d’ as incident X-ray of wavelength λ diffracts at angle ‘θ’.
88
4.5.7 Energy disperse X-Ray spectroscopy (EDS)
The SEM and TEM used in this study were both equipped with EDS. This
technique was used mostly as a post-analysis method to accurately determine
the elements present at the worn surfaces of the fretting contact. It was also
used with the TEM to identify the distribution of elements in the subsurface of
the metallic alloys. The technique works by detecting characteristic X-rays
released; thus elements whose characteristic X-rays were detected can be
easily identified. When the surface of a specimen is bombarded by electrons
during electron microscopy, the incident electrons often displaces another
electron from within an atomic shell of the specimen. In order to preserve
equilibrium, one of the electrons from a higher energy orbital drops to a lower
orbital. Through this process, the characteristic X-rays of the specific element
is released.
It is worth noting that certain elements have energies that overlap such as Mo
and S; P and Pt depending on the characteristic X-ray energy used. Therefore,
when generating a map using the licensed software, Aztec, Oxford
Instruments, UK, the ‘TruMap’ function was used; which gives a more
accurate quantification. The intensity of the EDS maps depicted in the various
chapters of this thesis is indicative of the relative abundance of the element.
However, there are other limitations associated with the EDS technique:
inaccurate quantification of light elements such as N, C etc. Therefore these
may appear as background noise in the map and are generally deemed
absent unless they are observed to be more concentrated in a specific region
relative to another within a single map.
4.5.8 X-Ray photoelectron spectroscopy (XPS)
X-Ray Photoelectron Spectroscopy (XPS) data was obtained in this study

using the K-AlphaTM + XPS, Thermo Fisher Scienctific, USA. This is a surface
technique which was employed to further characterise the tribocorrosion
products identified using EDS. In XPS, atoms within compounds or bulk
materials when exposed to X-ray photons (hv) become ionised as electrons
89
from their inner core is ejected. Thus, the binding energy (BE) of the electron
is equivalent to the energy required to remove it from its shell.
Elements identified on the surface of a sample, either through EDS or XPS

survey scans can be better quantified using high resolution scanning. X-rays
can penetrate the specimen surface to a depth of several micro-meters,
however, the photoelectrons analysed are generally from several nano-
meters of the immediate subsurface of the material. This makes the
characterisation of the passive film (generally few nano-meters thick) and the
bulk alloy possible. For this study, to assess the variation in atomic
composition of high resolution scanned elements as a function of depth, the
surfaces were exposed to Argon-ion monoatomic beam at 4 keV across an
area of 1 mm x 2 mm (when characterising passive films on CoCrMo and
Ti6Al4V polished surface) and 200 μm x 400 μm (when analysing post –
fretting surfaces). Licensed CasaXPS software was used to fit the XPS peaks
obtained for each element. A Shirley background remove feature were used
in fitting the peaks and the Aliphatic C 1s at 284.8 eV were used for the charge
correction of the peaks. The fitting parameters namely: full width half
maximum (FWHM), BE and peak shapes were obtained from various sources
in literature and respectable webpage repositories [153-165].
4.5.8.1 Estimations of etching rate
A very rough estimation of the etching rate for CoCrMo and Ti6Al4V protective
layer was calculated using the average measured thickness (obtained from
TEM micrograph) of the protective film. The average known thickness of the
protective layer from each alloy is then divided by the maximum etching time
(400 sec) analysed before the bulk metal is reached. These rates were
estimated to 0.0094 nm/sec and 0.019 nm/sec for CoCrMo and Ti6Al4V
respectively.
90
4.5.9 Nano-indentation hardness
Hardness is a key consideration of the mechanical properties of a material in

the context of tribological interactions. In this study, nano-indentation
hardness of polished CoCrMo and Ti6Al4V and tribocorrosion products were
characterised within the top 40 – 60 nm of the metallic surfaces. The Nano-
Indentor – NanoTest NTX, Micromaterials, UK was used for the study. A
calibrated Berkovich diamond indenter with tip-end radius of ~200 nm radius
was used. Hardness is automatically determined by the device using the
Oliver and Pharr [166] method. The apparatus obtain hardness
measurements by measuring penetration depth as a function of load during a
load-unload cycle – hardness is the peak load divided by the projected area
at peak load. A minimum of 20 indentations from different sites of the sample
surfaces were conducted for statistical reliability. Further details are provided
in the result chapter for each case.
4.5.10 Coordinate measurement machine (CMM)
A 3D reconstruction of the male and female taper surfaces were generated by

mapping the surface using the coordinate measurement machine (CMM),
Legex 322, Mitutoyo, Japan. The Redlux sphere profiler, UK is a commercially
available software used to generate the 3D reconstruction. Deviations from a
cone reference in the software gives an indication of preferentially worn region
thus, the wear pattern from the tapers can be identified. This method was
employed mainly to obtain a qualitative assessment of wear pattern on the
taper surface for each material combination.
The samples were wiped with acetone then wiped thoroughly with de-ionised
water to ensure the stylus is not in contact with acetone. Subsequently, the
samples were dried with Nitrogen air-gun. The samples were positioned in a
fixture and flatness was ensured using a spirit level. A calibrated M3, 1.5 mm
diameter ruby ball, 30 mm length styli was used to generate points 0.2 mm
apart (horizontally and vertically). In every case, the measurement begins
from the uppermost surface of the taper, to the lowest point.
91
Chapter 5 Metallurgy and surface chemistry of CoCrMo and
Ti6Al4V
5.1 Introduction
In the subsequent chapters of this thesis, results from fretting experiments will
be presented. This purpose of this chapter is therefore to characterise a
polished CoCrMo and Ti6Al4V sample: examining the crystallinity of the alloy,
surface chemistry and hardness in a dry condition. The subsurface
characterisation of the polished alloys in this chapter will thus act as a
reference for the subsurface transformations observed in the subsequent
chapters. An outline of the analysis carried out in this chapter is shown in
Figure 5-1.
Figure 5-1 – Outline of analyses performed in chapter 5
92
5.2 Crystalline structure of CoCrMo and Ti6Al4V
XRD was used to characterise the crystalline structure of both CoCrMo and
Ti6Al4V using the Cu Kα radiation. The 2θ scans between 30° and 60° shown
in Figure 5-2 and Figure 5-3 distinctively identified the HCP and FCC phases
of CoCrMo alloy as well as the HCP, BCC (α, β) phases of the Ti6Al4V alloy.
Figure 5-2 shows the co-existing HCP and FCC phases. The FCC (111) peak
is the most prominent and is indicative of a metal (M) carbide (M23C6). The
FCC (200) peak is the γ-Co phase. The peaks at (101̅0) and (101̅1) represent
the HCP ε-Co phases as confirmed in other studies [167-169].
Figure 5-2 – Crystalline structure of polished CoCrMo alloy
The 2θ plot for polished Ti6Al4V alloy in Figure 5-3 shows two prominent HCP
(101̅1) and (101̅1) peaks and two other less prominent HCP peaks (0002) and
(101̅2). Two BCC (110) and (200) peaks are also evident in Figure 5-3 thus
confirming the co-existing HCP and BCC phases. The XRD peak agrees with
reference [170]. Ti and Ti–Al are both α phases with an HCP structure as
identified from the XRD database [171, 172]. The BCC peaks corresponds to
the Ti–V β phase [173].
93
Figure 5-3 – Crystalline structure of polished Ti6Al4V alloy
5.3 Subsurface microscopy and spectroscopy
TEM samples of the polished alloys were obtained using the FIB preparation
method outlined in section 4.5.4. High resolution micrographs of the
subsurface metallurgy in both alloys was thereby acquired. Incorporated with
the TEM was SAED through which the crystallinity of a localised region in the
subsurface structure was identified. In addition, EDS incorporated with the
TEM was used to map the distribution of elements in the grain structure within
a selected subsurface region.
5.3.1 TEM and SAED of polished CoCrMo and Ti6Al4V
Figure 5-4a shows the bright-field micrograph of the polished CoCrMo alloy.
Directly beneath the two layers of platinum, a protective layer of thickness
ranging from 2.5 to 5 nm is observed on the CoCrMo alloy. This is within the
expected thickness for CoCrMo alloy [174-176]. The dark-field micrograph of
the same region is presented in Figure 5-4b. The micrograph offers a better
contrast between the different crystalline structures thus making it possible to
distinguish between the nano-crystalline region and the single crystal beneath
94
it. The nano-crystalline structure has a depth of 130 nm from the uppermost
surface. The single crystal structure is seen directly beneath the nano-
crystalline structure. It can be seen that the single crystal has experienced
some degree of twinning which may have been a result of strain-induced
transformations during mechanical polishing of the surface[177].
For further clarification, SAED was conducted at regions A and B. The ringed
pattern in Figure 5-4c is an indication of a nano-crystalline structure and
Figure 5-4d shows evidence of crystal planes as expected in a single crystal
structure [98, 150] thus confirming the observations from the micrographs.
Figure 5-5 shows the TEM and SAED micrographs of Ti6Al4V. The protective
layer on the Ti6Al4V is seen more distinctively in the bright-field image of
Figure 5-5a and its thickness ranges from 5 to 10 nm as others have also
observed [174, 178]. However, thicker surface oxide layer (~15 nm) have
been reported for a mechanically polished surface [179]. From both Figure 5-
5a and b, the nano-crystalline structure beneath the protective layer does not
appear to be of a consistent depth unlike in CoCrMo, although the depth is
equally less than 200 nm. SAED pattern in Figure 5-5c confirms the structure
in region A to be nano-crystalline and the diffraction pattern of region B in
Figure 5-5d confirms the structure to be a single crystal.
95
(a) (b)
[A] [B]
(c) (d)
Figure 5-4 – TEM and SAED of polished CoCrMo: a) Bright-field

region ‘A’ d) SAED of single crystal ‘B’
96
(a) (b)
(c) (d)
Figure 5-5 – TEM and SAED of Polished Ti6Al4V: a) Bright-field

region ‘A’ d) SAED of single crystal ‘B’
5.3.2 TEM – EDS of polished CoCrMo and Ti6Al4V
TEM-EDS maps of CoCrMo and Ti6Al4V are shown in Figure 5-6 and Figure
5-7 respectively. The most abundant elements for both alloys Co, Cr, Mo and
Ti, Al, V and O (enriched in the protective layer although not clearly visible)
are visible. It is observable that in CoCrMo alloy the three elements are equally
distributed at varying intensities across the material subsurface as expected.
97
On the other hand, distinctive grains corresponding to the α and β phases of
Ti6Al4V is seen Figure 5-7. The β phase is distinguished by the higher
intensity of V and lesser Al content.
Figure 5-6 – TEM–EDS map of polished CoCrMo showing Co, Cr, Mo, O.
Figure 5-7 – TEM–EDS map of polished Ti6Al4V showing Ti, Al, V, O.
98
The chemical composition of the alloys were quantified using TEM-EDS.
Selected regions from Figure 5-6 and Figure 5-7 were a minimum area of 500
by 200 nm. The results outlined in Table 5-1 for CoCrMo and Table 5-2 for
Ti6Al4V are presented in wt% and at%. The proportion of all the heavier
elements from CoCrMo observed were within the range presented in the
ASTM standard. However in the Ti6Al4V, the V content in the α phase was 4x
higher than the ASTM prescription and 10x the V content in the β phase. This
simply highlights the difference in the chemical composition of the two phases
whereas ASTM represents an overall α – β phase alloy. The C, O and N
content appear exaggerated in both alloys due to the known limitation of EDS
technique – light elements are not accurately quantified by EDS.
Table 5-1 – TEM – EDS chemical composition of CoCrMo
ASTM F 1537-08 – Chemical Composition (%wt)
LC Co Cr Mo Mn C N O Si
CoCrMo
Wt% Bal. 26.0 5.0 1.0 0.14 0.25 1.0

-
30.0 7.0 max max max max
EDS – Chemical composition
LC Co Cr Mo Mn C N O Si
CoCrMo
Wt% 64.2 27.6 6.0 0.9 0.4 0.2 0.1 0.5
At% 61.4 30.0 3.5 0.9 1.8 0.8 0.5 1.1
99
Table 5-2 – TEM – EDS chemical composition of Ti6Al4V
ASTM F 136-08 – Chemical composition (%wt)
Ti Al V Fe C N O
5.5 3.5 0.25 0.08 0.05 0.13

Bal. -
Ti6Al4V 6.5 4.5 max max max max
EDS – Chemical composition
α Ti6Al4V Ti Al V Fe C N O -
Wt% 88.3 6.3 2.3 0.0 0.4 0.3 2.3
At% 79.4 10.1 2.0 0.0 1.4 1.0 6.2
β Ti6Al4V Ti Al V Fe C N O -
Wt% 71.2 2.2 21.8 2.1 0.1 0.0 2.6
At% 67.5 3.7 19.4 1.7 0.3 0.0 7.4
5.4 Surface chemistry
The corrosion resistance of CoCrMo and Ti6Al4V alloys are mainly due to the
protective layer that spontaneously form on their surfaces. In this section, XPS
is employed to characterise the layer that are identified on both alloys. High
resolution scans of C, O, N, Co, Cr Mo were carried out for the CoCrMo and
C, O, N, Ti, Al, V for the Ti6Al4V. The elements were examined through a
depth profile of five levels (100 s per level). An estimated average etching rate
were determined to be 0.0094 nm/sec and 0.019 nm/sec for CoCrMo and
Ti6Al4V respectively as outlined in section 4.5.8.
5.4.1 XPS assessment of CoCrMo passive film
The concentration in at% of the elements assessed for CoCrMo are shown in
Figure 5-8. It is observable that the uppermost surface is mainly organic
contaminants composing of C, O and N as described in other references [178,
100
179]. A sharp increase in the O content between the first and second levels
corresponds to the increase in Cr and Mo, this is indicative of the two main
metallic elements that contribute to the protective layer. However, the
abundance of Cr compared to Mo was seen to be roughly 8:1 ratio across the
depths beyond the first level confirming that Cr is the dominant metallic
element in the protective layer [180]. Furthermore, no evidence of Co was
observed throughout the 5 levels of the depth profile analysis which suggests
that the bulk surface of the alloy was not reached. More so, the absence of
any oxides of Co is likely due to an observation made in the study referenced
already [180] where it is noted that Co-oxide exist deeper within the protective
layer – interfacing with the bulk metal.
Figure 5-8 – XPS Depth profile analysis of CoCrMo passive film
In Figure 5-9, Cr, Mo and O are presented in the percentage proportion of

their species as observed through the high resolution scans. Three typical O
peaks were identified as shown in Figure 5-9a. These are broadly termed: MO
(Metal-oxide) which mainly represents the O2- at binding energy (BE) ~529.7
eV; NBO (non-bridging oxygen) which represents the OH- at BE ~531.4 eV
101
and the BO (bridging oxygen) which mainly represents surface adsorbed
water (H2O) at BE ~533.3 eV [158, 180].
From Figure 5-9b it is observable that the Cr (3+) which generally exist within
the protective layer in the form of Cr2O3 and Cr2(OH)3 [181] decreases rapidly
in proportion to Cr (0) metal from the uppermost surface towards a constant
proportion at the last examined depth. On the other hand, a fairly constant
proportion of all Mo species were observed in Figure 5-9d beyond the
uppermost surface. The peak at BE 228.1 eV corresponds to Molybdenum
sub-oxide (Mo(2+/x) – oxide) which had been identified previously in literature
as a peak of Mo (2+)/non-stoichiometric Mo – oxide [182].
(a)
(b) (c)
Figure 5-9 – Proportion of species in CoCrMo protective layer: a) O

species, b) Cr species, c) Mo species
102
5.4.2 XPS assessment of Ti6Al4V passive film
Figure 5-10 shows the concentration of the elements assessed on the TI6Al4V
surface. Similar to the CoCrMo surface, the uppermost surface of the Ti6Al4V
alloy also composed of C, O and N and a small proportion of the most reactive
metal in the alloy – Cr in the case of CoCrMo and Ti in Ti6Al4V alloy. The
proportion of O, C and Ti on the Ti6Al4V follow a similar pattern with O, C and
Cr contents as described for the CoCrMo. However, the proportion of Ti in the
Ti6Al4V protective layer is seen to be higher than that of Cr in the CoCrMo
layer for all levels. Interestingly, the concentration of V was higher in
proportion to Al for all levels beyond the uppermost surface.
Figure 5-10 – XPS Depth profile analysis of Ti6Al4V passive film
The proportion of species for key elements in Ti6Al4V such as Ti, V, Al and O
are shown in Figure 5-11. The O species in Figure 5-11a are observed in a
similar proportion compared to the CoCrMo alloy and in Figure 5-11b, it is
found that Ti exist in the protective layer in multiple oxidation states. Ti (4+)
which generally forms TiO2 was the most abundant at the uppermost layer as
similarly described by Hanawa et al [143]. Interestingly, although Ti (2+) was
103
the most abundant specie across the depths beyond the uppermost surface,
it is known to be thermodynamically less favourable at the uppermost surface
which explains why only Ti (4+) and Ti (3+) were found at the uppermost
surface [183]. Oxidised Al (3+) and its metallic component both contributed to
the layer beyond the uppermost surface as shown in Figure 5-11c. At the
second level where Al (3+) was highest, a non-stoichiometric oxide, V (1/2+)
at BE ~513.4 eV was also present. Whereas V doesn’t structurally exist in its
1+ oxidation state, it is said that the V (3+) specie commonly exist in the form
of V (1+) and V (2+) hence the notation V (1/2+) [157]. Beyond the second
level, only metallic specie of V was detected.
(a) (b)
(c) (d)
Figure 5-11 – Proportion of species in Ti6Al4V protective layer: a) O

species, b) Ti species, c) Al species, d) V species
104
5.5 Nano–indentation hardness of polished surface
CoCrMo alloy is the main metallic material used for the bearing component of
hip prosthesis; this is partly due to its relatively high bulk hardness. On the
other hand, titanium alloys are preferred to CoCrMo for the design of stem
components due to its much lower stiffness (~50% that of CoCrMo). Nano-
indentation hardness of the polished surfaces of both alloys were performed
to further characterise the hardness of the two materials. These are shown in
Figure 5-12a and b for CoCrMo and Ti6Al4V respectively. In Figure 5-12c an
average of n = 40 indentations at a maximum depth of 40 nm in CoCrMo and
67 nm in Ti6Al4V yielded hardness values of 5.8 GPa and 3.6 GPa
respectively. The values measured for Ti6Al4V agrees well with values
reported in another study [184]. However, values quoted for CoCrMo in other
studies range from 4.5 – 8 GPa [17, 67, 150, 185].
(a) (b)
105
(c)
Figure 5-12 – Nano-indentation hardness of CoCrMo and Ti6Al4V a) P

vs. h plot for CoCrMo b) P vs. h plot for Ti6Al4V c) calculated hardness
of CoCrMo and Ti6Al4V.
5.6 Discussion and summary
This chapter has briefly characterised both CoCrMo and Ti6Al4V alloys – the
two main materials of interest in this research. XRD analysis of the crystalline
structure for both alloys agree with the findings from other studies in literature.
However, while only the polished surface of the alloys were analysed in this
study, there are other factors such as: surface treatment and roughness that
may influence the intensities and/or the existence of a peak observed through
XRD that other studies have shown [150, 170].
The key differences identified between CoCrMo and Ti6Al4V alloys are the
following: a prominent metal-carbide peak in CoCrMo; segregated α – β
microstructure in Ti6Al4V; at least 60% higher hardness in CoCrMo and a less
uniform subsurface nano-crystalline structure in the Ti6Al4V alloy.
TEM analysis of CoCrMo and Ti6Al4V surfaces reveal the common feature of
a protective thin layer. Both layers had a non-uniform thickness ranging from
2.5 to 5 nm and 5 to 10 nm respectively. Further, depth profile analysis
identified metallic species as a contributing constituent of the protective thin
106
layer. In addition, in both alloys, the depth profile analysis did not reach the
bulk material. Therefore, the existence of metallic species in the layer before
the bulk is reached suggests that the metallic surface can be represented in
three layers, similar to the description offered in a technical report on
passivation [186].
Figure 5-13 shows a schematic diagram of the three layers with the chemical
compositions that represents the surface of the two alloys. The three layers
are namely: passive layer; transition area and base alloy (bulk material). The
levels assessed during the depth profile analysis in this study can be
apportioned to the three layers. The passive layer corresponds to level 1
although surface contaminants from the topmost surface dwarfs the
concentration of the most reactive metals i.e. Cr in the case of CoCrMo and
Ti in the case of Ti6Al4V. This is due to the concentrations being normalised
to 100%. The transition area consist of both oxides and metallic species and
thus, it corresponds to levels 2 – 5. The bulk, which was not reached in this
study is expected to be dominantly metallic constituents with only a small
concentration of oxygen.
For the CoCrMo surface, the transition area was enriched in Cr and Mo
metallic species. Mo is typically added to Cobalt-based alloys for a similar
reason as it is for Iron-based alloys. Mo has a role in the improvement of
passivation; it is said to facilitate passivation by encouraging Cr enrichment in
the passive layer thus protecting against pitting corrosion and general forms
of corrosion [3, 186, 187]. Other studies have also demonstrated that Mo at
the surface plays a key role in protein adsorption. This may encourage the
formation of a proteinaceous tribofilm layer above the passive layer thus may
influence friction at the tribological interface of CoCrMo [113].
On the other hand, the transition area on the Ti6Al4V surface was observed
to be enriched in Ti and V. Contrary to the case of CoCrMo where the Mo had
a specific role to play in the passivation of the alloy, it is not clear in literature
if V has such role in Ti6Al4V. Nevertheless, it is clear that its oxides also
contribute to the TiO2 – rich passive layer as observed in other studies [179].
However, what proves to be more common in literature about V are concerns
107
about its toxicity in the human body. Several research effort have been
dedicated to search out other options to replace V. So far, studies have been
conducted on other alloying elements such as Mo, Zr and Nb [188-191]. It is
therefore highlighted through the findings of this study that the enrichment of
V at the nearest surface of Ti6Al4V alloy, may prove more of a negative
attribute of its protective layer than a positive one.
Figure 5-13 – Schematic diagram of a three-layer surface structure of

CoCrMo and Ti6Al4V. Surface contaminants are not included in the
three layers.
108
Chapter 6 Fretting wear mechanism of metal – metal material
combinations
6.1 Introduction
The two most commonly used metal – metal combinations at the modular
taper junction are: CoCrMo – CoCrMo and CoCrMo – Ti6Al4V. The two
combinations are established at the head-neck and neck-stem modular taper
interface. During the active life of the implant in-vivo, the material
combinations are susceptible to fretting corrosion. This is manifested through
various mechanisms of fretting wear and fatigue, substantiated by the
penetration of corrosive body fluids into the taper junction as the modular taper
experiences micromotions at the interface.
The purpose of this chapter is to experimentally simulate and systematically

investigate fretting corrosion mechanisms at the modular taper interface. This
is achieved by establishing a point contact using a ball-on-flat configuration;
the fretting displacement amplitude is varied whilst keeping the initial contact
pressure the same (see section 4.4.1). This allows for several material
combinations to be assessed under different fretting regimes and contact
conditions. In-situ OCP was monitored throughout the static and fretting
stages of the test for each combination. Post-test surface analytical
techniques such as: WLI, SEM-EDS, XPS and TEM-EDS were utilised for the
characterisation of the tribocorrosion products and subsurface metallurgical
transformations. Nano-indentation was also conducted on the product formed
at the interface of CoCrMo – Ti6Al4V in order to quantify the hardness of the
material. A schematic outline of the experiments and analysis performed in
this chapter are shown in Figure 6-1.
109
in chapter 6. C and T represents CoCrMo and Ti6Al4V respectively. The
letters that appear in red represent the samples analysed.
110
6.2 Tribological assessment of metal – metal fretting
contacts
The two metal – metal combinations were subjected to fretting displacement

amplitudes ‘δx’ of ±10 µm, ±25 µm, ±50 µm and ±150 µm. an additional
displacement of ±150 µm. Figure 6-2 shows the slip ratio for both material
combinations at their individual fretting displacements assessed. It is
observable that CoCrMo – CoCrMo falls within the gross slip regime for all
displacements except at ±10 µm which is in a partial-slip regime according to
the gross slip criteria defined in reference [149]. On the other hand, the contact
exists in a stick fretting regime for both ±10 µm and ±25 µm displacements for
the CoCrMo – Ti6Al4V material combination. There is no specific slip ratio
criteria that makes a distinction between the stick and partial-slip regime,
instead, these are generally distinguished through the degree of surface
damage at the end of the experiment. The fretting regime at ±50 µm
displacement for CoCrMo – Ti6Al4V is better described as a mixed regime as
later shown in Figure 6-4b through the transitions in the dissipated energy. A
full gross slip regime is realised at ±150 µm displacement.
Figure 6-2 – Slip ratio of CoCrMo – CoCrMo and CoCrMo – Ti6Al4V

under OCP conditions from ±10 to ±150 µm displacement.
111
6.2.1 Fretting loop analysis
Figure 6-3a shows a graphical representation of the fretting regimes of the

CoCrMo – CoCrMo combination. These appear as expected for each regime,
the partial-slip regime is represented as a quasi-stick profile while the gross
slip regime at ±25 µm appear in a parallelogram-like shape as expected for a
gross slip contact. However, for ±50 µm and ±150 µm displacement, subtle
deviations from the standard gross slip shape are indicative of the wear
mechanisms effective at the interface.
The energy dissipated (𝐸𝑑 ), is a function of the tangential force (Ft) and relative
slip at the fretting interface (δs) by the expression 𝐸𝑑 = 𝑓(𝐹𝑡 , 𝛿𝑠 ). The tangential
force is also known as the ‘frictional force’ at the gross slip regime. Also, the
energy dissipated per cycle in a gross slip regime is a representation of the
interfacial friction as derived in the reference [192]. Thus in this study, energy
dissipated is interchangeably used with the terms ‘frictional energy’ and
‘interfacial friction’ while the term ‘frictional force’ is also interchangeably used
with the tangential force. The energy dissipated per cycle is presented in a log
plot in Figure 6-3b due to the significant difference in the magnitude of energy
from the transition between partial-slip regime to gross slip regime and also
among the gross slip regimes.
112
(a)
(b)
Figure 6-3 – Fretting of CoCrMo – CoCrMo: a) Fretting loop b) Energy

dissipated per unit cycle.
The 2D fretting loop for CoCrMo – Ti6Al4V in Figure 6-4a confirms the fretting
contacts at ±10 µm and ±25 µm are in a stick regime. While only the gross slip
fretting loop is shown for ±50 µm, the graph of energy dissipated in Figure 6-
4b shows that the contact was in a lower energy regime for few hundred cycles
before it gradually transitioned to a gross slip regime at ~1000 cycles. The
113
fretting loop also shows that the gross slip regime at ±150 µm occurred at a
much higher frictional force than that of the CoCrMo – CoCrMo combination
in Figure 6-3a.
(a)
(b)
Figure 6-4 – Fretting of CoCrMo – Ti6Al4V: a) Fretting loop b) Energy

dissipated per unit cycle.
114
6.3 Open circuit potential measurement
A representation of the OCP measurement for the CoCrMo – CoCrMo

combination is shown in Figure 6-5a. During the pre-fretting (static) stage,
most of the contact experienced ennoblement which is generally attributed to
a build of protective oxide at the interface [193].
At the initiation of fretting, no cathodic shift was measured at ±10 µm which

was similarly observed in reference [139]. However, for all the gross slip
displacements i.e. ±25 µm, ±50 µm and ±150 µm, a cathodic shift was
observed. Interestingly, during the fretting cycles of ±150 µm, a gradual
positive rise in potential which lasted until the end of the fretting cycles was
observed. At the end of the fretting cycles for the gross slip regime, a two-
stage potential recovery was observed; the repassivation stage and
ennoblement stage. The former occurred at a faster rate than the latter. For
all gross slip regime, the final OCP recovered to a more noble potential than
at the start of the fretting.
The maximum cathodic shift was determined by obtaining the potential

difference from the point immediately prior to the initiation of fretting and the
most negative point during the fretting potential. Although, for partial-slip
contacts where a general ennobling behaviour is observed, maximum
cathodic shift is determined by the potential difference between the potential
before and after the transient cathodic shift. Figure 6-5b shows no negative
potential was observed at ±10 µm, however, the maximum cathodic shift
increased from ±25 µm to ±50 µm and dropped for ±150 µm to similar values
with that of ±25 µm.
115
(a)
(b)
Figure 6-5 – Open circuit potential of CoCrMo – CoCrMo: a) OCP plot

for ±10 µm, ±25 µm, ±50 µm and ±150 µm. b) Max. cathodic shift of all
displacements.
The OCP curves for CoCrMo – Ti6Al4V displacements are shown in Figure 6-
6a. The results show a gradual rise in potential throughout the fretting cycles
116
for both ±10 µm and ±25 µm. Although, at ±25 µm a slower rate and minimal
cathodic shift is observed at the initiation of fretting. A much larger drop in
potential was observed for the two gross slip displacements i.e. ±50 µm and
±150 µm. The rate and magnitude of the potential drop was seen to be higher
for the larger of the two gross slip contacts. Once fretting ceased, no potential
recovery was observed for the two sticking contacts, a two-stage recovery
process was observed for both gross-slip displacements. The first stage was
a rapid repassivation process followed by a much slower ennoblement
process in the second stage. It is notable that the total potential recovery after
fretting ceased were ~150 mV lesser than the potential before fretting started.
Figure 6-6b shows the maximum cathodic shift for the four displacements. It
is observable that the potential shift increased with increasing displacement.
However, on average, the maximum cathodic shift for ±50 µm and ±150 µm
were observed to be similar.
(a)
117
(b)
Figure 6-6 – Open circuit potential of CoCrMo – Ti6Al4V: a) OCP plot for
±10 µm, ±25 µm, ±50 µm and ±150 µm. b) Max. cathodic shift of all
displacements.
6.4 Wear volume
Figure 6-7 shows the total wear volume and redistributed volume of all four
displacements for CoCrMo – CoCrMo fretting contact. The column chart
represent the sum of both ball and flat components. The redistributed volume
(defined in section 4.5.2) were only measurable at the gross slip
displacements and were all of similar magnitude. On the other hand, the wear
volume was similar for both ±10 µm and ±25 µm and much higher at ±50 µm
and ±150 µm displacement as would be expected. However, the proportion of
the wear volume between ±50 µm and ±150 µm were not proportional to the
three-times multiple of the displacement amplitude thus indicating a transition
in wearing mode.
118
for CoCrMo – CoCrMo fretting contact.
The total wear volume for CoCrMo – Ti6Al4V material combination is shown
in Figure 6-8. The result shows a general increase in the redistributed volume
for all displacements. The wear volume for ±10 µm and ±25 µm (stick regime
contacts) were near zero as expected. Interestingly, at ±50 µm, the wear
volume and redistributed volume were very similar thus describing the case
whereby the worn material is almost entirely redistributed within the interface.
The wear volume at ±150 µm was significantly larger than at ±50 µm and the
error bar was also much higher at this displacement. This is attributed to
erratic contact behaviour involving ejection and detainment of third-body
product – this is later confirmed through surface analysis in Figure 6-10. In
comparison to the CoCrMo – CoCrMo contact, wear volume and redistributed
volume for CoCrMo – Ti6Al4V are atleast double at the gross slip regimes.
119
for CoCrMo – Ti6Al4V fretting contact.
6.5 Surface Analysis
In this section, post-test surface analysis using the WLI was used to obtain 3D
and 2D surface profilometry for both ball and flat components at all
displacements. Subsequently, BS-SEM was used to obtain high resolution
surface micrographs for CoCrMo and Ti6Al4V (flat components) at ±50 µm
displacement for characterisation of third-body products generated at the
fretting interface.
6.5.1 3D and 2D surface profilometry of fretted contacts
The 3D and 2D surface profilometry of CoCrMo – CoCrMo ball and flat

surfaces are shown in Figure 6-9a and Figure 6-9b. The 2D cross-sections
were taken from the vertical y-axis across the centre in most cases. However
in some other cases, the 2D cross section that most describes the wear
120
surface based on the 3D surface profile is used instead. The wear depth is
estimated using the 2D cross sectional profile. The key observations from
each displacement are outlined below:
 ±10 µm – Both ball and flat components have a worn annulus with wear
depth ~350 nm and a partially worn central partial-stick region.
 ±25 µm – The central partial-stick region on the ball is still present,
however, the worn annulus is relatively wider as expected, due to the
transition into gross slip regime. The wear depth is ~1 µm on the ball
component and the wear product are redistributed across the surfaces.
 ±50 µm – The proportion of the central partial-stick region on the ball
decreased further with increase in displacement whilst the wear depth
remains ~1 µm. On the flat component, an increased worn annulus is
observed however, around the lightly abraded central region are third-
body product.
 ±150 µm – The central region at this displacement is seen to be totally
worn for the ball component resulting in a wear depth of ~1 µm.
Interestingly, it appears as though third-body products are redistributed
within the ball wear surface which inflicted a deeper groove (wear depth
of ~1.4 µm) into the flat surface. The final length of the flat component
is more than double the initial Hertzian contact width (2a = 320 µm)
thus suggesting a transition to reciprocating sliding.
In general, the presence of the central partial-stick region from ±10 to ±50 µm
shows that the fretting contacts are still influenced by elastic contact
compliance whilst the absence of the central partial-stick region at ±150 µm,
is further indication of a third-body controlled wear mechanism at this
displacement [39].
121
(a)
(b)
Figure 6-9 – Surface profilometry of CoCrMo ball and CoCrMo flat for
all four displacements under OCP conditions: a) 3D surface profile, b)
2D cross-section of the vertical axis of the wear surface.
122
The 3D and 2D surface profiles of the CoCrMo – Ti6Al4V combination are
shown in Figure 6-10a and Figure 6-10b The key observations are outlined
below:
 ±10 µm – The stick regime fretting contact at this displacement show

no visible signs of wear on either CoCrMo or Ti6Al4V.
 ±25 µm – Although this fretting contact is also in a stick regime, a
relatively thin worn annulus developed on the CoCrMo ball The
maximum wear depth of the annulus is ~ 300 nm and across the entire
central partial-stick region, third-body materials likely transferred from
the worn annulus were observed. The third-body materials appear in
needle-like peaks of height ~1 µm. On the Ti6Al4V flat however, there
is no evidence of wear other than material transferred unto the flat.
 ±50 µm – The fretting contact at this displacement experienced a gross-
slip regime for most of the fretting cycles. The CoCrMo ball appear to
have experienced surface ruptures whereby heaps of fragments with
heights up to 3 µm above the unworn surface is visible. A layer of third-
body material up to 2 µm thick can be observed on the Ti6Al4V surface.
Areas within the layer where cracks of maximum depth ~5 µm exist can
be seen to fit well with the area where fragmented heaps of material
exist on the CoCrMo. This therefore suggests that the layer fractured
from the CoCrMo surface.
 ±150 µm – The surface profile shows a much larger wear surface for
both the ball and flat components compared with the surfaces at ±50
µm; heaps of third-body material with height ~ 6 µm were observed on
the CoCrMo surface. The Ti6Al4V surface is mainly characterised with
large wear depths with maximum depths reaching up to ~11 µm.
In general, the size of the wear contact surface for both ball and flat
components of the stick regimes were similar to the initial Hertzian contact
width (2a = 500 µm) and wear began at the edge of the Hertzian contact
boundaries as expected. Furthermore, the CoCrMo ball is shown to have worn
more than the Ti6Al4V for both ±25 µm and ±50 µm displacements.
123
(a)
(b)
Figure 6-10 – Surface profilometry of CoCrMo ball and Ti6Al4V flat for
all four displacements under OCP conditions: a) 3D surface profile, b)
2D cross-section of the vertical axis of the wear surface
124
6.5.2 Back scatter SEM of CoCrMo and Ti6Al4V surface
The surface and chemical analysis in this section is focused on characterising

the interfacial wear and corrosion products generated at the fretting contacts
of CoCrMo – CoCrMo and CoCrMo – Ti6Al4V. The flat components of ±50 µm
in both material combinations withhold a substantial amount of these products
hence, surface analysis were conducted only on the flat components.
BS-SEM micrograph of the CoCrMo flat from the CoCrMo – CoCrMo interface
is shown in Figure 6-11. There are three regions of interest: Region ‘A’
identifies the worn annulus with evidence of third-body products present in the
region. Region ‘B’ shows the boundary between the worn annulus and the
lightly worn central region as also observed in the 3D surface profilometry of
Figure 6-9a, ±50 µm. Region ‘C’ illustrates evidence of the third-body products
generated at the interface. The back scatter technique provides an additional
information which helps distinguish third-body products from the base material
in the wear surface. Often times, the third-body products are oxidised products
of the alloying elements therefore have a lower atomic mass than the bulk
alloy hence they appear darker than the base alloy in BS-SEM. Interestingly,
the micrograph also shows that majority of the third-body product are located
near the boundary between the worn annulus and the central partial-stick
region.
125
A
Figure 6-11 – Back scatter-SEM of CoCrMo flat from the CoCrMo –

CoCrMo interface at ±50 µm under OCP conditions.
The BS-SEM micrograph of Ti6Al4V surface in Figure 6-12 reveal a rather

complicated wear surface. Four main region of interests were labelled ‘A – D’.
Region ‘A’ is the Ti6Al4V surface external to the fretting contact, the visible
scratches observed in the region is a result of the jerking effect from the
fretting machine as it returns to zero position at the end of a test. Region ‘B’
points out the thick layer that ruptured from the CoCrMo surface as described
in Figure 6-10a, ±50 µm. Evidence of wear tracks seen across the surface of
region B is an indication that the CoCrMo alloy was fretting against this layer
instead of Ti6Al4V.
Region ‘C’ is seen to be almost entirely surrounded by the layer of region B.

The height of the material above the plain surface is measured in Figure 6-
10b is up to 2 µm. This indicates that region C exists at a lower surface level
than region B and thus a likely creviced region. Within region C and all the
other regions like it, were found extensive cracks to the Ti6Al4V surface.
Furthermore, lighter appearing materials of various shapes and sizes were
seen within the region. Considering that the material with the heaviest atomic
mass appears the brightest with BS-SEM, the material is likely CoCrMo bulk
alloy which has a higher atomic mass than Ti6Al4V (as spectroscopy
126
techniques later confirms). Region ‘D’ is seen with surface cracks inflicted unto
the Ti6Al4V surface which appear different from those observed in region ‘C’.
While the cracks in region ‘C’ appear like cracks formed through surface
tension between the elevated layer and the enclosed Ti6Al4V surface, the
crack in region D appears to be mechanically induced, that is, due to direct
contact with the fragments of material rupture from the CoCrMo as previously
shown in Figure 6-10a, ±50 µm.
Figure 6-12 – Back scatter-SEM of Ti6Al4V flat from the CoCrMo –

Ti6Al4V interface at ±50 µm under OCP conditions.

metal – metal interface
In this section, SEM-EDS and XPS techniques were used to characterise the
chemical composition of the third-body products at the interfaces of CoCrMo
– CoCrMo and CoCrMo – Ti6Al4V material combinations. SEM-EDS maps
are presented to highlight the key elements of interest on the surface. The
elements that are not present on the EDS maps are either absent or their
concentration is too low to be detected.
127
The flat surfaces characterised in the previous sections are the same surfaces
analysed here. For the high resolution XPS scan, the tribocorrosion products
in region ‘C’ of the CoCrMo flat (Figure 6-11) was scanned for the following
key elements: C, O, Co, Cr, Mo, P, S and Si. As for the Ti6Al4V flat, regions
C and D of Figure 6-12 were scanned for C, O, Ti, Al, V, Co Cr, Mo, P, S and
Si. S was also scanned at high resolution but no peak was detected for all
three analysed areas on the CoCrMo and Ti6Al4V flats.
6.6.1 EDS analysis of CoCrMo and Ti6Al4V surfaces (metal –

metal)
EDS map of the CoCrMo flat shown in Figure 6-13 identified the presence of
Co, Cr and Mo across the entire surface of the alloy as expected. However,
Co appears with a lower concentration across regions covered in third-body
products. The table inset in Figure 6-13 shows that Co has a concentration
about three-times less than its usual bulk concentration which is ~60 at%. O
on the other hand, is seen to be mostly abundant across the third-body
product. C and N are both identified but have are relatively weak.
Interestingly, Cl is observed to be most abundant within the third-body

products that surrounds the boundary between the worn annulus and the
central region whereas, P is seen to be concentrated within the product that
is present in the worn annulus. The inset table which analysed the P-rich
region shows evidence of P however no evidence of Ca. The presence of P
in a tribocorrosion region is generally in the form of Ca-phosphate or as Cr-
phosphate. Thus the absence of Ca may suggest the presence of Cr-
phosphate (as confirmed through XPS). Si and Mo were also detected in the
background as would be expected, considering that they are also constituents
of the bulk alloy (see section 5.3.2).
128
Figure 6-13 – SEM-EDS of CoCrMo flat at ±50 µm from CoCrMo –
CoCrMo interface under OCP conditions. The inset table outlines the
composition of a specific third-body region.
129
The EDS map of Ti6Al4V flat shown in Figure 6-14 shows evidence of Co, Cr
and Mo – the main chemical composition of CoCrMo alloy across the wear
surface. These were most concentrated in region B (the areas covered by the
fractured layer) as indicated in Figure 6-12 and illustrated in Figure 6-14. It is
observable that O was also more concentrated across the areas covered by
the layer, thus indicating the material composes mainly of mixed metal-oxides
of CoCrMo. The regions identified as regions, A, C and D in Figure 6-12
appear with relatively higher concentrations of Ti, Al and V – the main
chemical composition of Ti6Al4V. N and Cl appear to be more prominent in
region B although at relatively low concentrations.
Figure 6-14 – SEM-EDS of Ti6Al4V flat at ±50 µm from CoCrMo –

Ti6Al4V interface under OCP conditions.
130
6.6.2 XPS analysis of CoCrMo and Ti6Al4V surfaces (metal –
metal)
The analysis of the CoCrMo – CoCrMo interface is shown in Figure 6-15(a –

g). The specific region scanned is the third-body product illustrated in Figure
6-11 as region ‘C’. In this region, a typical adventitious C peak and others
generally attributed to contamination and denatured protein is seen in Figure
6-15a. Also, the three typical O peaks namely MO, NBO and BO can be seen
in Figure 6-15b. The peak from Co2p in Figure 6-15c indicates the presence
of oxidised/metal species of Co as EDS identified in Figure 6-13, however, the
intensity of the peak is too weak relative to the background to reliably compute.
A similar case is seen for the Mo3d peak in Figure 6-15g. Figure 6-15d shows
both metallic and oxidised species of Cr2p. In Figure 6-15e, evidence of the
phosphate P (5+) peak is seen as well as Si2p in variable oxidation states in
Figure 6-15f.
In Table 6-1 the Cr and O species are attributed mainly to Cr2O3 and CrPO4.
The significant proportion of the high energy Cr (3+) relative to the lower
energy is an indication that CrPO4 is more abundant in the third-body product
than Cr2O3. The presence of other metallic species (although at much lower
intensities) suggests the presence of mixed hydroxide species of Cr, Co and
Mo. Although high resolution scanning of Cl was not conducted using XPS,
EDS confirms it is present within the analysed third-body product (see Figure
6-13). Thus, it is expected that an unknown proportion of the lower binding
energy of Cr (3+) would contribute to the formation of CrCl3.
131
C1s Scan
3
x 10
10
C-C, C-H OH-/NBO O2-/MO
8
Org–O/H2O/BO
CPS
6 C-OH,
4
C-O-C
O-C=O C=O
2
291 290 289 288 287 286 285 284 283 282
Binding Energy (e V)
(a) (b)
Cr2p Scan
Co2p Scan
2
2 x 10
x 10
40
26 Cr (3+) Cr (3+)–Oxide
38
24
36
Cr 2p3/2
CPS
22
CPS
34
20
32
18
30
582 579 576 573 570
786 783 780 777 774
(c) (d)
P2p Scan Si2p Scan
1 1
x 10 x 10
60 60
P (5+) – 2p3/2 Si–Org.
55
55 P (5+) – 2p1/2 Si (4+) – Silicates
50
Oxide
CPS
CPS
50 45
40
45
35
40 30
108 105 102 99 96

138 135 132 129 Binding Energy (e V)
(e) (f)
M o3d Scan
1
x 10
75
70
CPS
65
60
55
237 234 231 228 225 222

(g)
Figure 6-15 – XPS analysis of wear and corrosion product from the
wear surface of CoCrMo – CoCrMo combination at ±50 µm under OCP
conditions.
132
Table 6-1 – Attribution of Cr and O species from CoCrMo – CoCrMo
interface at ±50 µm
Peak Position (eV) Attributions [154, 158] (%)

Cr 2p 574.2 Cr(0) – Metal 16
576.0 Cr(3+) – Oxide/Chloride 22
577.1 Cr(3+) – Hydroxide & 62

Phosphate
O 1s 530.5 MO O2- 44
531.3 NBO OH-/PO4- 40
532.6 BO H2O/Si-O/C-O 16
As previously mentioned, two regions of the Ti6Al4V flat from the CoCrMo –
Ti6Al4V material combinations were analysed. Analysis of region C (the
creviced area) of Figure 6-12 are shown in Figure 6-16(a – j). Figure 6-16a
shows the C 1s scan. The typical adventitious C peaks were not present on
this surface, but three distinctive C species were identified: 281.9 eV – which
corresponds to metal-carbide peak; the Aliphatic C specie at ~284.8 eV and
an organic C-O specie at ~285.7 eV. The absence of the adventitious C
species is likely because the part of the surface analysed in this region is
located below the uppermost surface of the wear scar which is where
adventitious C is typically found.
Figure 6-16b shows the three standard O peaks. Figure 6-16c – e shows
evidence of bulk metal, non-oxidised Ti, Al and V species respectively. Mo, P
and Si species were not detected on this surface as shown in Figure 6-16f –
h respectively. Interestingly, in Figure 6-16i, only the metallic specie of Co at
778.3 eV was detected and in Figure 6-16j, three distinctive peaks of Cr were
observed including the metallic specie of Cr at ~ 574.0 eV. Cr (3+) – oxide on
this surface was fitted with five peaks starting from ~575.5 eV. This is a
detailed way of fitting Cr (3+)-oxide as documented in reference [154] although
the single peak approach (used in Figure 6-15) is also commonly used [174].
The third Cr specie observed in this creviced region was the highly debated
133
Cr (6+) peak at 578.7 eV. Considering that the proportion of the MO specie is
double the NBO specie as shown in Table 6-2, the Cr (6+) specie is likely in
its oxide form of CrO3. It is worth noting that, a proposed rule for ensuring the
peak being fitted is actually Cr (6+) as supposed to multi-plets of Cr (3+) oxide
is that the peak should encompass an area greater than 15% of the Cr2p scan
[154, 159]. The area enclosed by the Cr (6+) in this analysis was up to 25%
and all five multi-plet peaks of the Cr (3+) oxide are also fitted.
C1s Scan V2pO1s Scan

1 2
x 10 x 10
95
90 75
85 Org. C-O C-C, C-H 70
O2-/MO
80
Carbide 65
OH-/NBO
CPS
75
CPS
60
70 55 Org. O/H2O/BO
65 50
60
45
55
40
288 286 284 282 280 540 537 534 531 528
Binding Energy (e V) Binding Energy (e V)
(a) (b)
Ti2p Scan Al2p Scan
3 1
x 10 x 10
18
100
16 Ti 2p3/2
14
95
Al 2p1/2 Al 2p3/2
90
12 Ti 2p1/2
85
CPS
CPS
10
8 80
6 75
4 70
2 65
468 464 460 456 452 78 75 72 69 66

(c) (d)
V2pO1s Scan M o3d Scan
2 1
x 10 85 x 10
75 80
70
75
65
70
CPS
CPS
60
55 65
V 2p1/2 V 2p3/2
50 60
45 55
40
524 520 516 512 508 237 234 231 228 225 222
(e) (f)
134
P2p Scan Si2p Scan
1 1
x 10 x 10
85 85
80 80
75
75
CPS
CPS
70
70
65
65
60
60
55
144 140 136 132 128 108 105 102 99 96

(g) (h)
Co2p Scan
Cr2p Scan
2
54 x 10
66 x 10
2 Cr (3+)–oxide
52
Cr 2p3/2
64 Cr (6+)
50
Co 2p3/2 62
60
CPS
48
CPS
46 58
56
44
54
42
582 579 576 573 570
786 783 780 777 774 Binding Energy (e V)
(i) (j)
wear surface of CoCrMo – Ti6Al4V combination at ±50 µm. Creviced
region – Region ‘C’
Table 6-2 – Attribution of Cr, Co and O species from CoCrMo – Ti6Al4V

interface at ±50 µm. Creviced region – Region ‘C’
Peak Position (eV) Attributions [154, 158] (%)

Cr 2p 574.0 Cr(0) – Metal 20
575.5 Cr(3+) – oxide 55
578.7 Cr(6+) – oxide 25
Co 2p 778.3 Co(0) – Metal 100
O 1s 530.5 MO O2- 63
531.3 NBO OH- 31
532.6 BO C-O 6
135
Analysis of the second region on the Ti6Al4V surface of the CoCrMo – Ti6Al4V
interface is shown in Figure 6-17(a – j). This corresponds to region D from the
BS-SEM micrograph of Figure 6-12; cracks within this region was described
to have been mechanically inflicted by the CoCrMo counterpart. In addition,
unlike region C, this region is not locally enclosed and is further away from the
centre of wear surface. In Figure 6-17a, it can be seen that the adventitious C
is present on this surface and so are the three O peaks in Figure 6-17b. shows
No metallic species of Ti were found on this surface as shown in Figure 6-17c

however, two oxidised species of Ti (4+) at 459.1 eV and Ti (3+) at 457.6 eV
were present. The abundance ratio of Ti (4+) to Ti (3+) are about 8:2 as shown
in Table 6-3. Figure 6-17d – f shows that neither oxidised nor the metallic
species of Al, V and Mo respectively were detected on the surface and the Si,
Co and Cr species in Figure 6-17h – j respectively may either be absent or too
weak to compute.
Evidence of a prominent phosphate P (5+) specie was also found in Figure 6-

17g. This explains the significant proportion of the NBO specie compared to
the MO specie. Further analysis of this region using TEM-EDS in section 6.7.2
may explain the absence of the species pertaining to the CoCrMo.
C1s Scan
3
x 10
10
C-C, C-H OH-/PO43- /NBO
8
C-OH,
CPS
6
C-O-C
4
Org–O/H2O/BO
O-C=O C=O O2-/MO
2
291 290 289 288 287 286 285 284 283 282
(a) (b)
136
Ti2p Scan Al2p Scan
1 1
x 10 x 10
160
Ti (4+) – 30
150
Ti (4+) – 2p3/2 28
2p1/2 Ti (3+) –
140 26
2p1/2 24
CPS
CPS 130
Ti (3+) – 22
120 2p3/2 20
110 18
16
100
468 464 460 456 452 78 75 72 69 66

(c) (d)
V2pO1s Scan M o3d Scan
2 1
x 10 52 x 10
60
50
55
48
50
46
45
44
40
CPS
CPS
42
35
40
30
38
25 36
20 34
15 32
524 520 516 512 508 237 234 231 228 225 222
(e) (f)
P2p Scan Si2p Scan
1 1
x 10 x 10
40
45
40 P (5+) – 2p1/2 P (5+) – 2p3/2 35

CPS
30
CPS
35
25
30
25 20
138 135 132 129 108 105 102 99 96

(g) (h)
Co2p Scan Cr2p Scan
1 1
x 10 270 x 10
330
260
320
250
310
240
300
CPS
CPS
230
290
220
280
210
270
786 783 780 777 774 582 579 576 573 570
(i) (j)
wear surface of CoCrMo – Ti6Al4V combination at ±50 µm under OCP
conditions. Region of mechanically induced crack – region ‘D’.
137
Table 6-3 – Attribution of Ti and O species from CoCrMo – Ti6Al4V
interface at ±50 µm. Region of mechanically induced crack – region ‘D’
Peak Position (eV) Attributions [153, 159] At (%)

Ti 2p 459.2 Ti(4+) – Oxide 84
457.7 Ti(3+) 16
O 1s 530.3 MO O2- 10
531.8 NBO OH-/PO4- 74.5
533.3 BO H2O/C-O 15.5
6.7 Characterisation of metallurgical transformations (metal

– metal)
It is well understood that the fretting regime of a material combination

influences the form of surface and subsurface transformations the energy
(work done) expended at the interface would yield [39]. Furthermore, it is
expected that increasing energy input by increasing fretting displacements
would yield different forms of subsurface transformation in CoCrMo and
Ti6Al4V alloys due to their different mechanical properties. For this reason, a
cross-section from the CoCrMo flat of the CoCrMo – CoCrMo combination and
the Ti6Al4V flat of the CoCrMo – Ti6Al4V combination was prepared with the
FIB. Samples were obtained from the worn region of each individual
displacement amplitude. From the 3D surface profile of Figure 6-9 and Figure
6-10, it can be seen that the fretting wear surface is highly varied and therefore
it is expected that transformation in the metallurgical structures would be site-
specific in most cases (see section 4.5.4 for further detail on how specific
region for analysis were selected).
6.7.1 TEM of Fretted CoCrMo and Ti6Al4V alloy
The following at the key observations made from the cross section of CoCrMo
flat components:
138
±10 µm – The bright and dark field images are shown in Figure 6-18a and b
respectively. A non-uniform ~500 nm deep nano-crystalline region is seen in
the subsurface. A worn groove is observed at the uppermost surface to be
filled with tribochemical products and numerous wear particles.
±25 µm – The micrographs of Figure 6-18c and d, shows a surface with

tribochemical deposits of ~200 nm thick. There were no nano-crystalline
structure observed beneath the tribochemical material. However, from the
dark field images, strain-induced subsurface twinning of width ~500 nm as
indicated with red-dotted lines is visible in the material.
±50 µm – Figure 6-18e and f reveal no nano-crystalline structures however a

higher degree of strain-induced twinning, beyond those seen at ±25 µm is
observed. The smaller sized twins are indicated with red dotted lines.
±150 µm – A broad view of the subsurface structure at this displacement is

shown in Figure 6-18g and h. It is evident that the strain-induced twinning at
this interface is less severe compared to the two previous gross slip
displacements. Interestingly, a material deposit on the surface appears with
various tribochemical structures as illustrated in the dark field micrograph.
Further analysis of this structure and the characterisation of the tribochemical
materials on the surface are evaluated in section 6.7.2.
139
Pt
(a) (b)
Pt
(c) (d)
Pt
(e) (f)
140
Pt
(g) (h)
Figure 6-18 – TEM bright and dark field micrographs of CoCrMo flat
from CoCrMo – CoCrMo interface at ±10 µm (a & b), ±25 µm (c & d), ±50
µm (e & f), ±150 µm (g & h).
The following are the key observations made from the TEM cross-section of
the Ti6Al4V flat component:
±10 µm – The regime at this displacement was a stick fretting regime. Figure
6-19a and b covers a subsurface area of over 4 µm x 4 µm however, a typical
grain structure was not found within this area. Rather, nano-crystalline
structures of ~500 nm deep from the uppermost surface were seen. The
contrasts provided by the dark field image of Figure 6-19b corresponds to the
different orientations of crystals within alloy. The absence of a grain-like
structure like one observed in the polished reference for Ti6Al4V in section
5.3.1 is a strong indication that a strain-induced subsurface transformation is
effective at this displacement. Evidence of directionality downward into the
alloy is further illustrated with red-dotted lines. Discretely located passive
layers are also visible at the uppermost surface.
±25 µm – The fretting regime at this displacement also experienced a stick

regime. Figure 6-19c and d shows a larger cross-sectional area of the
subsurface with dimension of ~6 µm x 8 µm. Similarly, no distinctive grain-like
structure were observed. The subsurface transformations in this displacement
were very similar to the transformations at ±10 µm.
141
±50 µm – Transitioning into a gross slip regime, Figure 6-19e and f captures
both the Co, Cr and O – rich layer that ruptured from the CoCrMo alloy. The
transformed structure of the underlying Ti6Al4V alloy is observed to be nano-
crystalline. Detailed analysis of this structure is provided in section 6.7.2.
±150 µm – Figure 6-19g and h, shows at least 4 µm deep Ti6Al4V structure

that has been transformed into a nano-crystalline structure from the
uppermost downward. In the bright field image, the white arrows are
illustrating regions of subsurface cracks which have propagated up to 500 nm
downwards from the uppermost surface. The dark field image aided the visual
identification of mechanical mixing up to 1 µm deep into the subsurface.
Pt
(a) (b)
Pt
(c) (d)
142
Pt
(e) (f)
Pt
(g) (h)
Figure 6-19 – TEM bright and dark field micrographs of Ti6Al4V flat
component from CoCrMo – Ti6Al4V interface at ±10 µm (a & b), ±25 µm
(c & d), ±50 µm (e & f), ±150 µm (g & h).
6.7.2 TEM-EDS and SAED analysis
In this section, subsurface TEM, EDS spectroscopy and SAED are combined
together in order to in-depthly characterise the third-body wear and corrosion
products. The specific samples analysed in this section are outlined below:
 CoCrMo flat from the CoCrMo – CoCrMo combination at ±150 µm:

Third-body product deposited on the flat component under
143
reciprocating sliding conduction following a transition from a gross slip
regime.
 Ti6Al4V flat from the CoCrMo – Ti6A4V combination at ±50 µm: The
region analysed in this section is ‘B’ from Figure 6-12; the ruptured
material which appear as a layer on the Ti6Al4V flat.
 Ti6Al4V flat from the CoCrMo – Ti6A4V combination at ±50 µm: The
region analysed in this section is ‘D’ from Figure 6-12; the mechanically
induced cracked region within the Ti6Al4V flat.
 CoCrMo ball from the CoCrMo – Ti6A4V combination at ±50 µm: the
CoCrMo ball component is analysed.
6.7.2.1 Analysis of third-body deposit at the CoCrMo – CoCrMo

interface
TEM and SAED analysis of CoCrMo flat component from the fretting
displacement at ±150 µm is shown in Figure 6-20a. Three distinctive
structures are visible: the first structure is denoted ‘1’; it is a tribochemical
third-body material (~300 nm thick) located within the wear track. The second
structure, denoted ‘2’, is better observed in the higher resolution micrograph
of Figure 6-20b. It has a honeycomb appearance with a nano-crystalline
structure. The bulk alloy, directly beneath both structures 1 and 2 is denoted
‘3’. In this region, the structure visually appear to be a single crystal.
SAED of structure 1 in Figure 6-20c appears with a shroud of bright spots as

well as distinctive spots, thus an indication of an amalgamated amorphous
and nano-crystalline structures. The diffraction pattern for structure 2 in Figure
6-20d and structure 3 in Figure 6-20e both agrees with the TEM micrograph
observations that both are nano-crystalline and single crystal structures
respectively.
144
Pt
(a) (b)
[311]
(c) (d) (e)
Figure 6-20 – TEM and SAED of CoCrMo – CoCrMo at ±150 µm: a)

Bright field TEM micrograph of CoCrMo subsurface b) Close-up image
of tribochemical products on surface c) SAED of structure 1 d) SAED of
structure 2 e) SAED of CoCrMo bulk.
EDS map of the third-body product shown in Figure 6-21 reveals that the
chemical composition of structure 1 is dominantly Cr, Co and O. Cl can also
be seen faintly in the region nearest to structure 2. In structure 2, it is
observable that Cr is absent in the region leaving the honeycomb-like
structure of mainly Co and Mo. In addition, a layer of Ca with evidence of P
can be seen at the uppermost surface of structure 1. This may suggest the
formation of Ca-phosphate at the interface as other studies have found for a
sliding contact [194].
145
Figure 6-21 – TEM-EDS map of CoCrMo – CoCrMo flat surface at ±150
µm
146
In order to better understand the processes involved in the formation of the
tribochemical product shown in Figure 6-20, a collage of micrographs from
interferometry, SEM, TEM and site-specific EDS analysis is shown in Figure
6-22. From this figure, it is better observed that the honeycomb-like structure
from which Cr is depleted, has a maximum thickness of ~400 nm.
Interestingly, it can also be seen that a passive layer of maximum thickness
16 nm formed both across the bulk alloy and the crystalline-like mixed Co and
Cr-Oxide material (i.e. structure 1). Furthermore, it is also interesting that at
the thickest region of structure 1 – specifically the region where the Co and
Mo honeycomb-like structure is lacking – the passive oxide layer is not
observed. Rather, the uppermost region of the area is abundant in Ca, P, O
and Cl. This perhaps highlight the crucial role of Mo in the formation of Cr-rich
passive layers.
147
Figure 6-22 – TEM and EDS analysis of the transferred third-body at the
interface of CoCrMo – CoCrMo, ±150 µm.
148
6.7.2.2 Analysis of mechanically mixed layers in at the CoCrMo –
Ti6Al4V interface
Figure 6-23a shows the micrographs of mechanically mixed layers with varied
thicknesses across the surface and subsurface of Ti6Al4V alloy. The
uppermost layer (light grey) has a maximum thickness of ~2.5 µm in the
thickest region. Beneath this layer is a transformed amorphous-like structure
which a darker grey appearance. Directly beneath both layers, a nano-
crystalline structure of Ti6Al4V alloy is observed. At the interface where the
amorphous-like structure meets the nano-crystalline Ti6Al4V structure, a
crack can be seen. The existence of a crack at this point suggests that the
mechanical properties of the three structures are likely very different thus
causing an internal tension.
Figure 6-23b shows a high resolution micrograph of Figure 6-23a depicting all
three layers. Structure 1 appears to have nano-scale sized materials from the
amorphous-like structure 2. Similarly, particles from the nano-crystalline
Ti6Al4V structure (structure 3) appears to be suspending within structure 2.
The diffraction pattern in Figure 6-23c corresponding to structure ‘1’ is
confirmed to be a nano-crystalline structure. Although from the TEM
micrograph, the nano-scale crystals appear to be suspending within an
amorphous-like ‘mesh’.
Figure 6-23d shows the diffraction pattern of structure ‘2’; the micrograph
reveals a stronger evidence of an amorphous structure although the presence
of distinctive rings likely represents the nano-sized particles of Ti6Al4V. In
Figure 6-23e i.e. structure ‘3’, a more definitive nano-crystalline rings of
Ti6Al4V polycrystalline structure is observed.
149
1
(a) (b)
1 2 3
(c) (d) (e)
Figure 6-23 – TEM and SAED of mechanically mixed layers on Ti6Al4V

alloy: a) TEM micrograph of Ti6Al4V subsurface b) high-resolution
micrographs of the subsurface structures c) SAED of the mechanically
mixed layer d) SAED of mechanically mixed Ti6Al4V structures e)
SEAD of Ti6Al4V nano-crystalline structure.
EDS map of the TEM cross-section is presented in Figure 6-24. It can be seen
that chemical composition of structure 1 is largely a mix of of Co, Cr, Mo, Ti
and high concentration of O. Cl is also evident across the structure. It is thus
deducible that structure 1 exists as mixed oxides and chlorides of the above
metals where Cr is the dominant metal-oxide. It is worth noting that the
elements tabulated as insets in Figure 6-24 are a normalised quantification of
the most concentrated elements in the structure. For example, the table does
not include Cl, Al and V however, all three are identified in the TEM-EDS map.
150
The map also clarifies the white appearing particulates suspending in both
structures 1 and 2 are Co-rich particles.
The EDS map also confirms structure 2 as an amorphized Ti6Al4V mixed with
the physiological solution as evidenced by the presence of O, N and Cl. EDS
also confirms the chemical composition of the nano-crystalline Ti6Al4V
structure to be Ti, Al and V as expected. This highlights the role that the
physiological solution played in the transformation of nano-crystalline Ti6Al4V.
C in the map is mainly from the background signal as it exist uniformly across
all three structures.
Figure 6-24 – TEM-EDS map of the mechanically mixed layer and

transformed Ti6Al4V alloy from CoCrMo – Ti6Al4V interface.
151
6.7.2.3 Analysis of mechanically induced crack at the CoCrMo –
Ti6Al4V interface
TEM cross-section from the region in Ti6Al4V flat where surface cracks were
identified to be mechanically induced (region D of Figure 6-12) was also
analysed. In Figure 6-25a the structure appears much coarse and less
cohesive in comparison to the structure 1 observed in Figure 6-23b. Large
particulates (appears to be grey and are blending into the coarse structure)
are seen to be suspending within the coarse structure as indicated with the
red arrows. The micrograph also shows numerous horizontal and diagonal
cracks across the coarse structure from the uppermost surface deep down
into the alloy. Other particles/structures which appear darker within the
structure were also seen dispersed across the micrograph. Figure 6-25b
presents a high resolution micrograph of such ‘darker’ appearing particles.
From the micrograph, there are evidence of several of these particles
suspending in the coarse structure with sizes ranging from 10s of nm to sizes
of over 200 nm. The diffraction patterns in structures ‘1’ and ‘2’ are both
indicated in Figure 6-25c and Figure 6-25d respectively. These consists of
nano-crystalline ringed patterns of multiple structures.
Pt
1
(a) (b)
152
1 2
[222]
[0002] [0002]
[300]
(c) (d)
Figure 6-25 – TEM and SAED of Ti6Al4V surface (mechanically cracked

region): a) TEM micrograph of Ti6Al4V subsurface b) high resolution
image of the region in red dotted square c) SAED of the nano-
crystalline mesh d) SAED of a single large particle.
EDS map of the TEM cross-section of the coarse structure is shown in Figure
6-26. The map confirms the ‘dark’ particles within the coarse structure, in
Figure 6-25b are Co-rich particles. The map also confirms the coarse structure
consists mainly of mixed oxides and chlorides of both CoCrMo and Ti6Al4V
alloys. A Cr-chloride structure is particularly illustrated with red dashed-
circular lines.
It is worth pointing out that, this EDS map, along with others presented
previously were unable to distinguish/identify reliably the presence of proteins
using EDS. This mainly due to overlapping energies of S and Mo. Reliably
detecting S would require other techniques such as electron energy loss
spectroscopy (EELS) which was not employed in this study. However, the C
map (illustrated with red arrows) in Figure 6-26 identifies a pathways through
which physiological solution migrate deeper into the alloy and thus possibly
fostering localised corrosion deep inside the alloy.
153
Figure 6-26 – TEM-EDS map of the Ti6Al4V flat at the mechanically
induced crack region of the CoCrMo – Ti6Al4V interface.
6.7.2.4 Analysis of CoCrMo ball from the CoCrMo – Ti6Al4V interface
The subsurface transformations of CoCrMo under partial-slip and gross slip

regimes have been observed through the TEM micrographs of Figure 6-18.
However, considering that a mixed regime was realised at the CoCrMo –
Ti6Al4V interface at ±50 µm, TEM and SAED analysis of the CoCrMo ball was
analysed. It had been deduced from previous surface analysis in this chapter
that the evidence of Co, Cr and Mo – rich structure from Figure 6-12 is
indicative of CoCrMo ball separating from the shared interface with Ti6Al4V
through a rupture of the interfacial wear and corrosion products formed. This
was denoted ‘structure 1’ in Figure 6-23b. Figure 6-27a shows the dark field
micrograph of the subsurface structure of the CoCrMo alloy. The image reveal
evidence of severe strain-induced twining of large grains upto depts of 7 µm
into the alloy. The red dashed-line in the figure is illustrative of the boundary
between a twinned crystal (below) and a nano-crystalline transformed grain
(above).
154
The white arrows are illustrative of cracks which may have propagated along
the nano-crystalline structures and cracks formed at the boundary of twinned
structures. Figure 6-27b shows the bright field image; evidence of material
deposits of ~500 nm thickness was found on the surface; it was denoted
structure ‘1’ and the bulk alloy beneath it was denoted structure ‘2’. In Figure
6-27c, the diffraction pattern shows a nano-crystalline structure for structure
‘1’. In structure ‘2’, Figure 6-27d, the diffraction pattern appear like a strained
single crystal structure as expected.
Pt
(a)
155
1 2
1 Pt
2
2 µm
(b) (c) (d)
Figure 6-27 – TEM and SAED of CoCrMo ball surface: a) Dark field TEM
micrograph of CoCrMo subsurface b) Bright field image c) SAED of the
transferred material d) SAED of near-surface CoCrMo bulk.
EDS examination of structure ‘1’ from Figure 6-27b is shown below in Figure
6-28. The map shows that the material is a mixture of Co-rich particle, oxides
of Cr and Ti, as well as constituents of the physiological solution namely: K
and C. S, Cl and Mo are also present.
Figure 6-28 – TEM-EDS analysis of the CoCrMo ball surface from the
CoCrMo – Ti6Al4V interface.
156
6.8 Nano-indentation
Analysis of surface and subsurface chemistry has so far characterised the

tribocorrosion products at the interface of both CoCrMo – CoCrMo and
CoCrMo – Ti6Al4V material combinations. These were observed to be
mixtures of metal-oxides, phosphates, chlorides and hydroxides. However,
the layer characterised on the Ti6Al4V flat component from section 6.7.2.2 is
sufficiently thick to be characterised using nano-indentation. Figure 6-29a
shows the area indented sandwiched between resin and Ti6Al4V alloy. The
Figure 6-29b shows the unprocessed P vs. h indentation data. In Figure 6-29c
the mean average hardness of the layer is 25.6 GPa. This is more than 3x the
bulk hardness of CoCrMo alloy and more than 6x the nano-indentation
hardness of Ti6Al4V alloy.
(a)
(b) (b)
Figure 6-29 – Nano-indentation of layer formed at the CoCrMo –

Ti6Al4V interface: a) image of the indented cross-section b) P vs. h
graph of indentations c) average hardness of the layer.
157
In this chapter, two metal – metal material combinations were assessed under
varied fretting displacement amplitudes. The characteristic surface and
subsurface damage modes pertaining to each material combination (CoCrMo
– CoCrMo and CoCrMo – Ti6Al4V) were assessed. Furthermore, the
assessment in chapter has presented in-depth characterisation of
tribocorrosion and tribochemical products formed at the interface of each
material couple.
For the same initial contact pressure of 1 GPa, the initial Hertzian contact
width ‘a’ for CoCrMo – CoCrMo is calculated to be 60% that of CoCrMo –
Ti6Al4V. Thus, the interfacial compliance experienced at the CoCrMo –
CoCrMo contact is less than CoCrMo – Ti6Al4V. This explains why self-mated
CoCrMo transitions into gross slip at a much lower fretting displacement than
CoCrMo – Ti6Al4V. In general, the displacements assessed in this chapter
reveal CoCrMo – CoCrMo contact transition from a partial-slip regime to gross
slip and to the transition stages of reciprocating sliding condition at ±150 µm.
Examination of the wear pattern at the interface showed the influence of

elastic compliance up until reciprocating sliding condition where influence of
third-body products at the interface appear to be more prominent. As for the
CoCrMo – Ti6Al4V combination, the contact progressed from stick regime to
mixed regime then to gross slip regime. Mindlin’s description for an elastic
contact (see section 2.6.1.5) was observable at the stick regime contacts. The
mixed regime was characterised with rupture of what appears to have been a
cold-welded interface. Upon transition from mixed regime to a gross slip, a
significant increase in the wear volume generated was observed and a large
proportion of the wear volume was redistributed at the interface. The main
difference in the wearing modes of both material combinations is thus the
adhesive wear characteristics of the Ti6Al4V alloy which facilitated
redistributed wear at the interface.
In-situ OCP measurement in both material combinations highlighted some

similarities and differences. As expected, both material combinations
158
responded to the gross slip regimes with a cathodic shift and subsequent
repassivation as soon as fretting stopped. The following are the observed
characteristic behaviour of each individual material combinations:
1. The magnitude of cathodic shift at all displacements were much larger

for the CoCrMo – Ti6Al4V contact than the CoCrMo – CoCrMo contact
as compared in Figure 6-30. This is indicative of a more
electrochemically active interface when using a mixed metal
combination compared to the self-mated alternative. This may explain
why corrosion damage in mixed-metal combination is often more
severe than in self-mated material combinations.
Figure 6-30 – Comparing Max. cathodic shift of the two metal – metal
material combinations.
2. No cathodic shift was measured at the partial-slip regime (±10 µm) of

CoCrMo – CoCrMo despite evidence of wear. This suggests that during
the partial-slip regime, the reactive (bare) metal was not exposed; likely
covered by the wear products. Secondly, the OCP curve was seen to
be rising during on-going fretting for the transitioned ±150 µm
displacement. This is either indicative of galvanic interaction within the
wear interface due to a rearranged contact condition with increase in
159
third-body influence. Or, the formation of a tribochemical film which acts
as a charge barrier thus causing the potential to rise.
3. The depassivation rate and the magnitude of cathodic shift from OCP
observations shows that CoCrMo – Ti6Al4V under fretting conditions is
more electrochemically responsive to the magnitude of displacement
at the interface than the self-mated CoCrMo combination. This
behaviour can be attributed to the Ti6Al4V alloy which is known to be
more reactive than CoCrMo. A fretting study by Baxmann et al [102]
whereby the Ti6Al4V alloy was the only conductive component
observed a similarly large cathodic shift upon fretting.
4. The potential during the recovery stages of CoCrMo – Ti6Al4V after
gross slip fretting was seen to be partial relative to the potential prior to
fretting. This may be due irreversible plastic damage at the fretting
interface thus reducing the corrosion resistant properties of the alloys
as reported for Ti6Al4V in previous studies. Other possible reasons are
the establishment of sites for localised corrosion and substantial
changes to the local solution chemistry such that repassivation at the
interface is inhibited [195, 196]. Each explanation is equally feasible in
this study because there was evidence of cracks within the Ti6Al4V
wear surface (see Figure 6-14) thus confirming the establishment of
sites for localised corrosion.
Subsurface metallurgical transformations of the CoCrMo alloy at the CoCrMo

– CoCrMo interface were observed at various fretting regimes. The partial-slip
regime was characterised by material removal (wear) and preservation of
nano-crystalline structures at the nearest subsurface. However, with increase
in the interfacial slip at the gross slip regimes (±25 µm and ±50 µm) loss of the
nano-crystalline structures at the nearest subsurface was evident. This
observation agrees with the explant analysis performed by Bryant et al [98]
who examined TEM images of adjacent unworn (partial-slip) and worn (gross
slip) regions of a fretting interface of CoCrMo modular neck.
Furthermore, it was also evident that the severity of subsurface twinning

increased with the interfacial energy during the gross slip regimes. However,
at the transition gross slip regime of ±150 µm, twinning was observed to be
160
less severe than at ±50 µm. Loss of nano-crystalline structures were also
evident but on the uppermost surface, tribochemical products transferred unto
the flat component was observed. This observation also agrees with several
in-vitro studies and explant analysis of self-mated CoCrMo alloy subjected to
a reciprocating sliding condition. It is commonly observed that the wear
mechanism at this interface is highly influenced by tribochemical reactions;
especially in a proteinaceous media [94, 95, 177, 197, 198].
Electron microscopy and spectroscopy techniques used to analyse third-body

products at the self-mated CoCrMo interface observed wear and corrosion
products that are chemically identical to the products observed from retrieval
studies [85, 199, 200]. Altogether, at the self-mated CoCrMo these were:
mixed-metal oxides and chlorides, hydrated Cr-phosphate and possibly Ca-
phosphate from the transition gross slip interface. Furthermore, through high
resolution imaging, electrochemical mechanisms such as Cr-depletion a
nano-crystalline structured wear product was observed. This finding is a novel
visual evidence of the electrochemical activity of wear and corrosion products
generated from the tribological interfaces of biomedical materials.
Dissipated energy at the interface of a material couple is manifested in several

ways as outlined by Fouvry et al [39]. In this study, subsurface metallurgical
transformation of Ti6Al4V at the CoCrMo – Ti6Al4V interface was observed to
be progressively fatigue oriented unlike the self-mated CoCrMo couple which
displayed a preferential damage mode of wear generation. For example,
subsurface analysis of the stick regime reveal evidence of strain-induced
subsurface transformations which were manifested in directionality and re-
orientation of crystallographic planes. Furthermore, it is interesting to note that
during the stick regimes of CoCrMo – Ti6Al4V, the Ti6Al4V experienced
subsurface transformations and little to no interfacial wear whilst the wear
annulus of the CoCrMo increased in size with increasing displacement thus
further highlighting the preferential damage modes of the two individual
materials.
At the mixed fretting regime of CoCrMo – Ti6Al4V (±50 µm) where both partial-
slip and gross-slip regimes occurred consecutively, cross-sectional TEM
161
micrograph showed a material with a ‘pseudo-amorphous’ structure which
until now was described as a ‘mechanically mixed layer’ on the Ti6Al4V
surface. The term ‘pseudo-amorphous’ is used here because the TEM
micrograph shows a material with solid particulates suspending within an
amorphous-like ‘mesh’. On the other hand, the SAED micrographs describes
a nano-crystalline structure which likely corresponds to the crystalline
structure of the several particulates of Ti6Al4V wear and Co as revealed in the
TEM-EDS (see Figure 6-24). TEM-EDS also revealed that the non-particulate
amorphous-like material composes mainly of mixed-metal oxides and
chlorides of Cr, Co, Mo and Ti.
From 3D surface analysis, it was clear that the pseudo-amorphous material

ruptured away from the CoCrMo counter-body. It was therefore deducible that
the rupture corresponded to the regime transition from partial-slip to gross slip
as discussed further in the overall discussion chapter (section 10.3.1). This is
to say that prior to the sudden rupture, the wear and corrosion products at the
interface during the partial-slip regime is likely welding the CoCrMo and
Ti6Al4V alloys together through the process of cold-welding. This may also
explains why the chemical composition of the corrosion product mainly
constituents of CoCrMo and Ti6Al4V alloys both in oxides and particulate
forms. Nano-indentation hardness of the corrosion material was found to be
more than triple that of CoCrMo alloy and upto six-time that of Ti6Al4V alloy.
It is anticipated that such a hard interfacial material would largely influence the
wear phenomena at the interface. From the transition to gross slip, the softer
Ti6Al4V is no longer in contact with the CoCrMo counter-body. Rather, the
CoCrMo is in direct fretting contact with the interfacial layer several times
harder than itself, thus it is expected that CoCrMo would suffer more wear in
such condition whilst Ti6Al4V would experience no wear. If such hard phases
form in-vivo, it proposes a more plausible reason than those given by
Moharrami et al [76] as to why CoCrMo wears more than its softer Ti6Al4V
counter-part [27].
Underneath the CoCrMo-rich pseudo-amorphous material a Ti6Al4V

amorphous material was observed. From this point in the writing, the pseudo-
162
amorphous ‘weld’ is called ‘pseudo-amorphous CoCrMo or shortened to p-a
CoCrMo’ while the mechanically transformed amorphous-like Ti6Al4V
material will be referred to as ‘amorphous Ti6Al4V’. Section 10.3.1 details the
mechanism leading to the formation of these two structures. Furthermore,
Ti6Al4V alloy in a gross slip regime is mainly characterised with subsurface
cracks, mechanically mixing of oxides and deep recrystallization to form nano-
crystalline structures.
In general, the corrosion products which are commonly observed in retrieval

studies of the mixed-metal couple were also observed at the CoCrMo –
Ti6Al4V interface of this study. More so, it was observed that the nature of
corrosion products found at the interface was linked to region of the composite
contact the product was found. For example, corrosion products located within
crevices were abundant in Cl ions whilst those outside of crevices were
phosphates and metal-oxides. Figure 6-31 shows a schematic diagram
depicting the chemical species observed at specific regions of the fretting
interface. SEM micrographs of both CoCrMo and Ti6Al4V is also included in
the schematic to show evidence of pitting mechanism on the CoCrMo and the
products of pitting deposited on the Ti6Al4V.
The most isolated areas to fresh physiological solution i.e. the most severe
crevice of the CoCrMo – Ti6Al4V was located at the centre of the wear
surface. The crevice were further isolated because of the pseudo-amorphous
CoCrMo (weld) material that was built up within the interface. In this region,
evidence of bulk CoCrMo deposits, through a pitting corrosion mechanism
was observed. A similarly observation was made by Gilbert et al [27] but the
corrosion mechanism was said to be intergranular. In this study, Cr6+ was also
observed in the same region. Cr6+ is said to be a carcinogenic specie of
oxidised Cr. It is a strong indication that a severe crevice such as can facilitate
pitting corrosion mechanism can also result to the formation of Cr6+ in-vivo [8].
On the other hand, at the non-creviced region i.e. the region closest to the
edges of the Hertzian contact, evidence of physiological solution which
penetrated deep down into worn and oxidised structures of the macro-scale
cracked region of Ti6Al4V was observed. The implication of this may be the
163
electrochemical process of oxide-induced stress corrosion cracking as
previously observed in a retrieval study of Ti6Al4V [97].
Figure 6-31 – Schematic diagram of fretting corrosion products and

SEM micrographs of CoCrMo and Ti6Al4V from the CoCrMo – Ti6Al4V
interface at ±50 µm displacement.
164
Chapter 7 Fretting wear mechanism of ceramic – metal
material combinations
7.1 Introduction
In the previous chapter, fretting wear mechanism of metal – metal material

combinations were assessed. This chapter is focused on the ceramic – metal
material combinations with the aim of understanding the fretting wear and
corrosion mechanism of the ceramic – metal systems. Ceramic bearings are
increasingly being used as an alternative bearing to CoCrMo alloy at the head-
neck taper interface. More so, recent studies have shown evidence of
reduction in fretting corrosion degradation of the modular taper through the
use of ceramic bearings [72, 201].
The fretting test parameter, test conditions and analyses conducted in this
chapter are very similar to those conducted in chapter 6. Si3N4 ceramic was
strategically chosen for this study because it is a material increasingly gaining
research attention for applications in hip and knee replacements both as bulk
material and as a coating[14]. In this vein, characterising the fretting corrosion
products and tribochemical interactions at the interface of Si3N4 – CoCrMo and
Si3N4 – Ti6Al4V combinations is a topical assessment. However, in order to
elucidate the role of Si3N4 in the wear mechanism of a ceramic – metal
combination, Si3N4 – Si3N4 (a ceramic – ceramic couple) was also assessed.
Biolox ceramic is the most commonly used commercially available ceramic for
hip prosthesis. Therefore, fretting wear mechanism of Biolox – CoCrMo and
Biolox – Ti6Al4V was also conducted in this chapter. A schematic outline of
the experiments and analysis performed in this chapter are shown in Figure
7-1 and Figure 7-2.
165
in chapter 7 for Si3N4 ceramic. C and T represents CoCrMo and Ti6Al4V
respectively. The letters that appear in red represent the samples
analysed. The green ticks represent the analysis performed on self-
mated Si3N4.
166
in chapter 7 for Biolox ceramic.
7.2 Tribological assessment of Ceramic – Metal and Ceramic

– Ceramic Fretting Contacts
The fretting displacements assessed for Si3N4 – CoCrMo, Si3N4 – Ti6Al4V and
Si3N4 – Si3N4 were ±10 μm, ±25 μm and ±50 μm. All the material combination
reached a gross-slip fretting regime by ±50 μm as shown in Figure 7-3;
although the fretting regime at the Si3N4 – Ti6Al4V interface was not stable in
gross slip regime. At ±10 μm all contacts were in a partial-slip or stick regime
and at ±25 μm, only Si3N4 – Ti6Al4V was still in a stick regime.
167
Figure 7-3 – Slip ratio of Si3N4 – CoCrMo, Si3N4 – Ti6Al4V and Si3N4 –
Si3N4 under OCP conditions for ±10, ±25 ±50 µm displacements.
The fretting behaviour of the three Si3N4-related combinations appeared to be

less stable at ±50 μm displacement in comparison to the metal – metal
combinations in the previous chapter. For this reason, an additional 2D fretting
loop was plotted at ±50 μm displacement in order to portray two extreme in
the variation of the contact behaviour. These are presented in green dashed-
lines in Figure 7-4a, Figure 7-5a and Figure 7-6a. The green solid-lines
represents one of the final few cycles of the fretting contact.
The fretting loop analysis for Si3N4 – CoCrMo is shown in Figure 7-4, Figure
7-4a confirms the regimes at this contact as partial-slip at ±10 μm and gross
slip for both ±25 μm and ±50 μm displacements. The energy dissipated per
cycle throughout the duration of the fretting tests are shown in Figure 7-4b.
These appear to be steady after few 100s of cycles for ±10 μm and ±25 μm.
However, at ±50 μm, a steady rise in the energy dissipated can be observed
which directly corresponds to a gradual increase in interfacial friction. This is
168
evidenced in the differences between the tangential forces of the fretting loops
at the 500th cycle and 6000th cycle in Figure 7-4a. The fretting loop at the
6000th cycle is also seen to be less stable than at the 500th cycle thus
indicating a change in the contact condition and wearing mechanism.
(a)
(b)
Figure 7-4 – Fretting contact analysis for Si3N4 – CoCrMo: a) Fretting

loop b) Energy dissipated per cycle.
169
The fretting behaviour at the interface of Si3N4 – Ti6Al4V is displayed in Figure
7-5. Both ±10 μm and ±25 μm displacements are in a stick regime as shown
in the fretting loops of Figure 7-5a. Despite this, there is a clear difference in
the magnitude of their energies as shown in Figure 7-5b. This highlights the
effect of higher tangential forces in the energy experienced at the contacting
interface. From the fretting loop, it is observable that the tangential force at
±25 μm is more than double that of ±10 μm which is in proportion to the
magnitude of their displacement amplitudes.
The two fretting loops displayed at ±50 μm displacement appears significantly

different, although both displace a gross slip shape. The loop at the 500th cycle
has a quasi-static feature at its corners. This feature is known as ‘the plunging
effect’ and it is commonly observed at the fretting contacts involving Ti6Al4V
alloy especially at high contact pressures [202]. For the same fretting loop, the
displacement (δx) can also be seen to be quite close to the applied ±50 μm
displacement whereas, for the fretting loop representing the 4750th cycle, δx
is seen to have overshot to as high as ±90 μm due to the highly unstable
fretting interface. The tangential force is also seen to be much higher and
unsteady.
The energy dissipated plot of Figure 7-5b show the degree of instability at the
fretting interface of ±50 μm. A continuous fluctuation between high tangential
force (stick regime energy) to large overshot (gross slip) fretting displacement
energy is observed. This behaviour is better described as an intermittent
stiction behaviour.
170
(a)
(b)
Figure 7-5 – Fretting contact analysis for Si3N4 – Ti6Al4V: a) Fretting

loop b) Energy dissipated per cycle.
171
In Figure 7-6a, the fretting loop for Si3N4 – Si3N4 shows a stick regime at ±10
μm and gross slip regimes for both ±25 μm and ±50 μm. In both gross-slip
cases, a distorted, irregular, non-symmetrical loop was observed. This is
further illustrated in Figure 7-6b where the energy dissipated per cycle plots
appear highly unstable after about 2500 cycles through to the end of the
fretting cycles. This is indicative of a change in the wearing mechanism at the
interface for both contacts.
(a)
172
(b)
Figure 7-6 – Fretting contact analysis for Si3N4 – Si3N4: a) Fretting loop
b) Energy dissipated per cycle.
7.3 Open circuit potential measurement
The OCP of the ceramic – metal combinations were monitored in the same
manner as in the previous chapter. Figure 7-7a shows the OCP behaviour of
Si3N4 – CoCrMo combination; it can be observed that depassivation potential
i.e. cathodic shift was only observed at ±50 μm. An ennoblement was
observed at ±10 μm as similarly seen with the CoCrMo – CoCrMo combination
in the previous chapter. For this specific test at ±10 μm, a rapid potential drop
followed by an immediate recovery was observed. This may be due to an
abrupt change in the fretting surface i.e. a momentary exposure of a
depassivated CoCrMo surface to the bulk solution. Interestingly, for the ±25
μm displacement, a general rise in potential with no cathodic shift was
observed despite the contact being in a gross slip regime.
As for the potential at ±50 μm, the cathodic shift continued gradually through
to the end of the fretting cycles as was also observed for the CoCrMo –
CoCrMo contact at the same displacement. The two gross-slip contacts
173
displayed a two-stage potential recovery pattern. The potential rises faster in
the first stage than in the second thus, in the same manner as observed for
the CoCrMo – CoCrMo contact. In addition, the repassivation potential was
seen to rise at a faster rate for the ±50 μm than at ±25 μm. This is likely related
to the surface area exposed to the bulk solution. Figure 7-7b shows the
maximum cathodic shift measured across both static and fretting stages for all
the combinations. The errors were large for ±10 μm displacement because
the cathodic shift was only observed once and it was due to the transient seen
in Figure 7-7a. The maximum cathodic shift at ±25 μm is seen to be minimal
relative to ±50 μm despite being in a gross-slip fretting regime. The largest
cathodic shift was observed for ±50 μm at an average value of ~60 mV.
(a)
174
(b)
Figure 7-7 – Open circuit potential of Si3N4 – CoCrMo: a) OCP plot for ±10
μm, ±25 μm, ±50 μm. b) Max. cathodic shift for all displacements.
The OCP measurements for the Si3N4 – Ti6Al4V combinations are shown in
Figure 7-8a. It is observable that at the stick fretting contacts (±10 μm and ±25
μm) a general rise in potential and no evidence of repassivation at the end of
the fretting cycles were observed. However, for the gross slip contact at ±50
μm, an instant cathodic shift to a large negative potential was seen. During
fretting, the stiction behaviour identified from the fretting loops was seen to be
represented in the OCP plot as a continuous repassivation (sticking) and de-
passivation (gross-slip) process. As soon as the fretting ceased, a single
recovery stage was identified. In addition, the potential was seen with upto
250 mV of potential difference between the initial potential and at the end of
the test – the same observation was made for the CoCrMo – Ti6Al4V interface.
In Figure 7-8b, maximum cathodic shift is seen to be minimal at ±10 μm and
±25 μm compared to the cathodic shift measured at ±50 μm.
175
(a)
(b)
Figure 7-8 – Open circuit potential of Si3N4 – Ti6Al4V: a) OCP plot for ±10
μm, ±25 μm, ±50 μm. b) Max. cathodic shift for all displacements.
176
7.4 Wear volume
The sum of wear volume for ball and flat components of the Si3N4 – CoCrMo
combination are shown in Figure 7-9. The sum of redistributed volume is also
shown in the same figure. Across all three displacements, the redistributed
volume measured on the surface is minimal. However, while the wear volume
at the partial-slip is minimal (as expected), the wear volume for both gross slip
displacements are very similar despite the difference in the displacement
amplitude by a factor of two. This is an indication of changes to the wearing
mechanism at ±50 μm displacement.
Figure 7-9 – Wear and redistributed volume from the Si3N4 – CoCrMo
fretting contact. A summation of both ball and flat components.
The wear volume measured for the Si3N4 – Ti6Al4V combination is shown in
Figure 7-10. The result shows that the wear volume was minimal for the two
contacts in a stick regimes (±10 μm and ±25 μm) as expected. On the other
hand, the redistributed volume at ±25 μm was observed to be higher than the
wear volume thus indicating the possibility of material gained at the interface.
This may be due to interactions between interfacial wear products and the
physiological solution thus resulting in a net material gain. For the mixed
fretting regime at ±50 μm displacement of Si3N4 – Ti6Al4V, the wear volume
177
is seen to be about three-times higher than those at gross slip regime for Si3N4
– CoCrMo combination. So also is the redistributed volume.
Figure 7-10 – Wear and redistributed volume from the Si3N4 – Ti6Al4V
In Figure 7-11, it is observable that for Si3N4 – Si3N4 fretting contact, the
measured redistributed volume is negligible and no wear volume was
measured at the stick regime. The wear volume measured at the gross slip
contacts are seen to be proportional to the displacement amplitude.
Figure 7-11 – Wear and redistributed volume from the Si3N4 – Si3N4
178
7.5 Surface analysis
In this section, 3D and 2D surface profiles were obtained following a similar

outline to the previous chapter. BS-SEM was used to obtain high resolution
micrographs of the metallic wear surfaces from the ceramic – metal
combinations. The samples assessed are CoCrMo and Ti6Al4V flat
components at ±50 μm displacement.
7.5.1 3D and 2D surface profilometry of fretted wear surface
Figure 7-12a shows the 3D profile of the wear surface while Figure 7-12b
shows the 2D cross-sectional profile of both Si3N4 ball and CoCrMo flat. The
following are the key observations across the displacement amplitudes
consecutively:
±10 μm – Fretting at this displacement was a partial-slip regime, evidence of

a worn annular on the Si3N4 ball of ~120 nm in depth was observed. At the
boundary of the worn annulus and the unworn partial-stick region, a heap of
third-body material with heights reaching up to ~850 nm is can be seen. The
2D surface profile depicts this clearer. From Figure 7-12b, it is also clear that
the worn area on the CoCrMo flat corresponds to the third-body product
observed on the Si3N4 ball. Thus, it can be deduced that third-body products
at this interface may inflict wear mechanisms such as micro-cutting on to the
softer component which in this case was the CoCrMo alloy. The depth of the
cut measures up to 600 nm.
±25 μm – At the fretting interface of this gross-slip regime, the Si3N4 ball is
seen to have experienced uniform wear of depths ~260 nm. However, the
centre of the ball experienced the most severe wear with depths reaching up
to 1 μm. The CoCrMo flat on the other hand, appear to have worn more than
the Si3N4 ball at the region where Si3N4 ball experienced the least. And at the
centre where the CoCrMo wore the least, the Si3N4 ball wore more
significantly. The phenomena is better depicted in Figure 7-12b.
179
±50 μm – At this fretting interface, the Si3N4 ball experienced wear of uniform
depth of ~1.3 μm with no evidence of a preserved central region. On the other
hand, wear depth across the CoCrMo flat component was surprisingly low at
a maximum depth of ~1 μm. Third-body material was seen to be agglomerated
to the centre of the CoCrMo wear surface. It is anticipated that the third-body
material which is later characterised led to the harder Si3N4 ball wear much
greater than the softer and less stiff CoCrMo counter-part.
(a)
180
(b)
Figure 7-12 – Surface profilometry of Si3N4 ball and CoCrMo flat

surfaces: a) 3D surface profile of all displacements, b) 2D cross-section of
the vertical axis of the wear surface.
Figure 7-13a and b shows the 3D and 2D surface profilometry of Si3N4 ball
and Ti6Al4V flat.
±10 μm – The contact at this fretting displacement experienced a stick regime.

The surface profile of the Si3N4 ball shows no evidence of wear but an
agglomeration of third-body products which appear ‘needle-like’ and have
heights up to ~850 nm. The third-body material mainly originated from the
Ti6Al4V surface where evidence of plastic deformation at the edges of the
Hertzian contact are visible.
±25 μm – Although the fretting regime at this displacement was also a stick
regime, it is observable that a worn annulus of maximum wear depth ~250 nm
developed on the Si3N4 ball at the edge of the Hertzian contact. Similarly to
the ±10 μm displacement, third-body products populated the central region
having the same ‘needle-like’ appearance. At this interface, the height of the
needle-like peaks reach up to 1.1 μm. On the Ti6Al4V flat however, a similar
feature of dented central region and material pile up around the wear surface
is observed.
181
±50 μm – Fretting at this contact is characterised by intermittent sticking and
slipping i.e. stiction. The scale of the y-axis in the 2D plot of Figure 7-13b was
changed enlarged in order to accommodate the depth of the cracks present
within the Ti6Al4V alloy. In addition, the 2D cross-section of Ti6Al4V at this
displacement was obtained from the horizontal axis instead of the standard
vertical axis. This allowed for both the height of the developed interfacial layer
and the crack to be quantified within the same cross-section.
A wear depth of ~2 μm was measured within the Si3N4 ball, however, some
peaks (likely unworn areas) within the wear surface are also visible. This may
be indicative of brittle fracture at the interface. The layer present on the wear
surface of Ti6Al4V appear similar to that of the ‘mechanically mixed layer’
observed at the CoCrMo – Ti6Al4V interface from the previous chapter. The
layer had peaks reaching upto 2 μm above the zero-line. The 3D topography
of the surface also depicts evidence of directionality and plastic flow along the
axis of the reciprocating fretting displacement at the surface. This
phenomenon was repeatable in all n = 3 tests and in other tests, the layer
reached up to 4 μm. Furthermore, a relatively wide and deep crack of depths
reaching upto 13 μm is observed in the Ti6Al4V alloy.
182
(a)
(b)
Figure 7-13 – Surface profilometry of Si3N4 ball and Ti6Al4V flat surfaces:
a) 3D surface profile of all displacements, b) 2D cross-section of the
vertical axis of the wear surface.
183
Figure 7-14 shows the 3D and 2D surface profile of the Si3N4 ball and flat
components.
±10 μm – A worn annulus of wear depth ~100 nm is visible at the surface of

the ball while still in a stick regime. However, no such pattern is observable
on the surface of the flat component.
±25 μm – A gross slip regime was realised at this fretting contact. The wear
depth for both ball and flat components were ~1 μm and it is observable from
the flat component that wear was maximum at the centre. In addition, the 3D
topography of the wear surfaces depicts a contact dominated by brittle fracture
as would be expected for a self-mated Si3N4 ceramic couple [203].
±50 μm – At this displacement, the wear depth for the flat component
increased to ~2.5 μm while the width of the ball’s wear surface increased
relative to the width at ±25 μm displacement. Interestingly, the 3D surface
profile of the ball shows evidence of brittle fracture at the centre of the wear
surface while a uniformly abraded annulus is observe adjacent to the central
region. This suggests that as the displacement amplitude increases, the wear
mode transitions from a fracture-based wear mechanism to an abrasive wear
mechanism. Thus at this displacement, a composite contact of both
mechanism was observable.
184
(a)
(b)
Figure 7-14 – Surface profilometry of Si3N4 ball and Si3N4 flat surfaces: a)
3D surface profile of all displacements, b) 2D cross-section of the vertical
axis of the wear surface.
185
7.5.2 BS-SEM analysis
In a similar fashion to the analysis of metal – metal material combinations in

the previous chapter, the third-body products at the ceramic – metal interface
of this chapter were also assessed on at ±50 μm displacement where it is most
prominent. Analyses were only performed on the metallic flat components for
all cases.
The BS–SEM micrograph of CoCrMo surface from the Si3N4 – CoCrMo

material combination at ±50 μm displacement is shown in Figure 7-15. As
already identified through the 3D surface topography, third-body products can
be seen across the centre of the wear surface. The red dashed-line is drawn
around the real wear surface so as to make a distinction between the confines
of the wear surface and the additional area scratched upon the cessation of
fretting.
Figure 7-15 – Back scatter-SEM micrograph of CoCrMo flat from the

fretting against Si3N4 ball at ±50 μm.
186
BS–SEM micrograph of the Ti6Al4V surface from the Si3N4 – Ti6Al4V material
combination at ±50 μm displacement is shown in Figure 7-16. This figure
identifies four regions of interest. Region ‘A’ shows the wear tracks on the
Ti6Al4V surface. Region ‘B’ is a material deposited unto the Ti6Al4V surface.
The material in region ‘C’ appears smooth and blended into the Ti6Al4V
surface despite being a different material. The material in region C
corresponds to the ‘mechanically mixed layer’ described in the 3D surface
profile of Figure 7-13. In region ‘D’ the surface of the Ti6Al4V looks plastically
deformed but no wear tracks could be seen in the region. The is because,
region D was once covered with the layer identified in region C however,
during the polishing procedure preceding the nano-indentation analysis of the
mechanically mixed layer, most of interfacial layer ‘flaked’ off from the surface,
leaving the underlying Ti6Al4V surface, exposed. This proved a positive
occurrence as evidence of subsurface plastic deformation occurring in the
Ti6Al4V when the Si3N4 ball frets relative to the layer could now be imaged.
A
D
Figure 7-16 – Back scatter-SEM micrograph of Ti6Al4V flat fretted against

Si3N4 ball at ±50 μm.
187
ceramic – metal interface
In this section, EDS and XPS spectroscopy techniques is used to further

characterise the surfaces imaged with the BS-SEM. High resolution XPS
analysis of the third-body products in the centre of the CoCrMo wear surface was
scanned for the following key elements: C, O, N, Co, Cr, Mo, P, S and Si. As for
the Ti6Al4V surface, high resolution XPS analysis were conducted at two regions
namely: regions C and D from Figure 7-13. The key elements scanned in both
regions were: C, O, N, Ti, Al, V, P, S and Si. Across all three regions analysed, S
was not detected as was the also the case during the analysis of the metal –
metal material combinations.
7.6.1 EDS analysis of CoCrMo and Ti6Al4V surfaces (ceramic –

metal)
The EDS map of the CoCrMo flat surface in Figure 7-17 shows that Co, Cr
and Mo were present at the background as expected being the most abundant
elements in the bulk alloy. On the other hand, Si and O can be seen to be only
highly concentrated across the third-body deposit. C, Cl, S and Ca are present
as background signal. This SEM-EDS map strongly indicates that the third-
body material present at the Si3N4 – CoCrMo interface is mainly oxidised Si.
Figure 7-17 – SEM-EDS of CoCrMo flat fretted against Si3N4 ball at ±50 μm.
188
The EDS map of the Ti6Al4V flat surface in Figure 7-18 shows the Ti, Al and
V at the background as expected for Ti6Al4V alloy. However, the material
deposit in region B of Figure 7-16 is seen to constitute mainly of C, Ca, S and
Cl; these elements could only have originated from the physiological solution
in which the fretting occurred. The presence of S is also indicative of proteins
thus suggesting the deposit is a cluster of denatured protein and other organic
species. The ‘mechanically mixed layer’ in region C of Figure 7-16 was
identified to constitute mainly of Ti, Si, O and Al. This suggests that the
material is made up of mixed oxidised compounds similar to the observation
from CoCrMo – Ti6Al4V in the previous chapter. Region D is seen to be
covered in C, O and Ca which suggests that the physiological solution
migrated beyond the uppermost where fretting occurred and thus might have
been trapped at the interface between of the ‘mechanically mixed layer and
bulk Ti6Al4V.
Figure 7-18 – SEM-EDS of Ti6Al4V flat fretted against Si3N4 ball at ±50 μm.
189
7.6.2 XPS analysis of CoCrMo and Ti6Al4V surface (ceramic –
metal)
Analysis of the Si3N4 – CoCrMo interface is shown in Figure 7-19(a – h). The
region scanned is the Si and O-rich third-body product agglomerated onto the
CoCrMo flat component. In Figure 7-19a, the series of peaks pertaining to the
adventitious C is observed. Three peaks of oxygen was observed in Figure 7-
19b however none of these corresponds to the MO specie. Rather, an NBO
specie is observed at 531.4 eV followed by a BO specie at 532.1 eV which is
attributable to organic species. The other BO specie at 533.4 eV is attributable
to SiO2 [160]. The main species of interest at this interface are outlined in
Table 7-1. From Figure 7-19c – e it can be seen that Co and Cr may be present
in the region however, their intensity relative to the background signal appear
too low to be computed. As for Mo, there is no evidence it is present within the
structure. In Figure 7-19f, evidence of organic N; likely from the serum solution
was present. No P peak were identified in Figure 7-19g and both high energy
Si (4+) and the organic Si specie were observed for the Si scan in Figure 7-
19g. The main product at this interface is therefore SiO2 or an hydrated specie
of silica generally denoted SiOx-OHy as also observed in previous studies [17].
C1s Scan O1s Scan

3 2
x 10 x 10
45
10 40
C-C, C-H 35 OH-/NBO
8
OH-/NBO
30
CPS
CPS
6
C-OH, 25 Org–O/H2O/BO
4 C-O-C
O-C=O C=O 20
2 15
291 290 289 288 287 286 285 284 283 282 537 534 531 528
(a) (b)
Co2p Scan Cr2p Scan
1 1
x 10 x 10
290
200
280 195
270 190
185
260
180
CPS
CPS
250 175
240 170
165
230
160
220 155
786 783 780 777 774 582 579 576 573 570
(c) (d)
190
M o3d Scan N1s Scan
1 1
x 10 155 x 10
66 150
64 145 Org.–N
62
140
60
135
CPS
CPS
58
56 130
54 125
52 120
50 115
237 234 231 228 225 222 405 402 399 396
(e) (f)
P2p Scan Si2p Scan
1 1
x 10 x 10
45
50
45
40 Si (4+) Org.–Si
35
CPS
CPS
40
30
35
25
30
138 135 132 129 108 105 102 99 96
(g) (h)
surface of Si3N4 – CoCrMo combination at ±50 μm. Si&O-rich third-body
product on CoCrMo.
Table 7-1 – Attribution of Si, N and O species from Si3N4 – CoCrMo at

±50 μm. Si&O-rich third-body product on CoCrMo.
Element Position (eV) Attributions [17, 160, 161, 204] (%)

Si 2p 100.9 Org.–Si: Si-C 33
102.4 SiO2/SiOx-OHy 67
N 1s 399.6 Serum – Organic N 100
O 1s 531.4 NBO OH- 38
532.1 BO H2O/Si-O/C-O/N-O 44
533.4 BO SiO2 18
The curves in Figure 7-20(a – h) are the scans obtained from region C of
Figure 7-16 which corresponds to the surface of the Si, Ti, Al and O-rich layer.
For the C 1s scan in Figure 7-20a, an extra peak, in addition to the adventitious
191
C peak was fitted to BE ~284.3 eV which corresponds to an organic specie of
Si-C [155]. The three O peaks in Figure 7-20b are the standard MO, NBO and
BO peaks. From Table 7-2 it can be seen that the MO and NBO species are
close in proportion. The MO species corresponds mainly to Ti-oxide whilst the
NBO specie corresponds mainly to Ti and Al-silicates. Thus Figure 7-20c and
Figure 7-20d show oxidised species of both Ti as Ti (4+) and Ti (3+) and Al as
Al (3+) with no indication of their metallic species. No evidence of the V specie
was found both in its oxidised and metallic form as shown in Figure 7-20e.
Two N species were found in Figure 7-20f: the first corresponds to the N-H
bond which is likely ammonia – a by-product of Si3N4 tribochemical reactions
[205] and the other specie corresponds to the organic-N. From Figure 7-20g
no specie of P was found on this surface. The Si peaks are of two oxidised Si
species: the Si (4+) and the organic-Si.

2 2
x 10 x 10
70
50 C-C, C-H Ti-O-Si/NBO
60 O2-/MO
40
C-OH, 50
Si-C
CPS
CPS
30 C-O-C
C=O 40 Org–O/H2O/BO
20 O-C=O
30
10
20
290 288 286 284 282 280 540 537 534 531 528
(a) (b)
Ti2p Scan Al2p Scan

2 1
35
x 10
Ti (4+) – 40
x 10
2p3/2 38 Al (3+)
30
Ti (4+) – 36
Ti (3+) – 34
25 2p1/2 32
2p1/2
CPS
CPS
20 Ti (3+) – 30
28
15
2p3/2 26
24
10 22
20
468 464 460 456 452 78 75 72 69 66

(c) (d)
192
V2pO1s Scan N1s Scan
2 1
x 10 x 10
100
70 95
60 90 Org-N
50 85 N-H
CPS
CPS
80
40
75
30
70
20
65
524 520 516 512 405 402 399 396

(e) (f)
P2p Scan Si2p Scan
1 1
x 10 x 10
44 70
42 65 Si (4+) Si–Organic
40 60
38 55
CPS
CPS
36 50
45
34
40
32
35
30
30
28 25
138 135 132 129
108 105 102 99 96
(g) (h)
surface of Si3N4 – Ti6Al4V combination at ±50 μm. Mechanically mixed
layer on Ti6Al4V – Region ‘C’.
193
Table 7-2 – Attribution of Si, Ti, Al, N and O species from Si3N4 –
Ti6Al4V at ±50 μm. Mechanically mixed layer on Ti6Al4V – Region ‘C’
Element Position Attribution [156, 161, 204] (%)

Si 2p 102.8 Ti-Si-O/Al-Si-O 41
101.8 C-Si-O/Si3N4 59
N 1s 399.9 Serum: Organic N 80
401.8 Ammonia: N-H 20
Ti 2p 458.8 Ti4+: Ti-Si-O/Ti-O 92
457.3 Ti3+: Ti-O 8
Al2p 74.4 Al-Si-O/Al-O 100
O1s 530.4 Ti-O 42
532.0 Al-Si-O/ Ti-Si-O 48
533.3 H2O/Si-O/C-O/N-O 10
The second surface scanned on the Ti6Al4V flat corresponds to region ‘D’ of
Figure 7-16. Figure 7-21a shows that the typical adventitious C peaks were
not present in this area as similarly observed in the previous chapter. As
aforementioned, the absence of the adventitious C confirms that the region
being analysed is located at depth below the uppermost surface. Thus
indicative of less surface contamination at surfaces below the uppermost.
Similarly to the observations from the previous chapter, the Aliphatic C-C at
BE ~284.8 eV and an organic C-O at BE ~ 285.8 eV are both present on the
surface. The three standard O species are observed in Figure 7-21b and in
Figure 7-21c – e Ti, Al and V were present in their non-oxidised metallic state.
From Figure 7-21f – h, N, P and Si species were not detected on the surface.
Thus confirming that this region had little to no interactions with the Si 3N4
counter-body as expected considering that the surface was unveiled
unexpectedly from beneath an interfacial layer.
194
1 2
x 10 x 10
80
100
95 75 OH-/NBO
90
Organic O2-/MO
CO bond C-C, C-H 70
85
CPS
65
CPS 80 Org–O/H2O/BO
60
75
70 55
65 50
291 288 285 282 540 537 534 531 528

(a) (b)
Ti2p Scan Al2p Scan
3 1
25 x 10 x 10
130
Ti 2p3/2
20 120
Ti 2p1/2 110 Al 2p1/2 Al 2p3/2

15
CPS
CPS
100
10
90
5
80
468 464 460 456 452 78 75 72 69 66

(c) (d)
V2pO1s Scan N1s Scan
2 1
x 10 x 10
80
75
75
70 70
CPS
CPS
65
65
60
55
60
50
524 520 516 512 508 405 402 399 396

(e) (f)
P2p Scan Si2p Scan
1 1
x 10 x 10
90 90
85 85
80
CPS
CPS
80
75
75
70
70
138 135 132 129 108 105 102 99 96

(g) (h)
surface of Si3N4 – Ti6Al4V combination at ±50 μm. Ti6Al4V surface which
was once covered by the mechanically mixed layer – Region ‘D’
195
7.7 Characterisation of metallurgical transformations
(ceramic – metal)
This section employs TEM, SAED and TEM-EDS to characterise the

subsurface tribocorrosion and metallurgical transformations of CoCrMo and
Ti6Al4V alloys which were fretted against Si3N4 at ±50 μm displacement. The
fretting regimes realised at the metal – metal and ceramic – metal interfaces at
±10 μm and ±25 μm fretting displacments were found to be similar although
occurring at a higher dissipated energy for the ceramic – metal fretting contacts.
For ths reason, subsurface analyses of the metal were not conducted at those
fretting displacements as they can be expected to be similar to those reported in
section 6.7.1. TEM samples in this section were selected mainly from the visual
observation of the region of interests using BS-SEM micrographs. The Si and
O-rich third-body region was selected from the CoCrMo surface of the Si3N4 –
CoCrMo material combination. However for the Ti6Al4V sample, two
subsurface samples were obtained: the first was obtained on the region
covered with the ‘mechanically mixed layer’ and the second was obtained from
the that was not covered with a layer. Both samples were obtained from the
same Ti6Al4V surface. As shown in the 3D topography in Figure 7-13a the
Ti6Al4V surface has three levels, the top of the mechanically mixed layer, the
Ti6Al4V surface and the crack within the Ti6Al4V. Only the first two levels
were assessed in study.
The micrographs in Figure 7-22 shows the subsurface cross-sectional view of

the CoCrMo flat component from the Si3N4 – CoCrMo material combination.
Figure 7-22a shows the Si and O-rich third-body material on the bulk CoCrMo
alloy which has a thickness of ~500 nm in the thickest region. Suspending
within the third-body material are fragments of CoCrMo particles as later
confirmed with EDS in the following section. The size of the particles range
from 10s of nm to as large as ~130 x 250 nm. Figure 7-22b shows a high
resolution view of the uppermost region. Figure 7-22a also reveals ~7 μm
depth of twinned CoCrMo bulk alloy beneath the Si and O-rich film.
Further analysis of the Si and O-rich product using SAED confirmed the
presence of nano-crystalline structures as shown in Figure 7-22c. TEM
196
micrograph also indicate that the Si and O-rich material has an amorphous
structure although not fully verified. Directly beneath the nano-crystalline
region, the SAED structure in Figure 7-22d shows a single crystal structure.
Pt
Pt
Pt
1 Pt Pt
(a)
Pt
1
[210]
2
̅2]
[20𝟐
(b) (c) (d)
Figure 7-22 – TEM and SAED of CoCrMo from the Si3N4 – CoCrMo
combination: a) TEM micrograph of CoCrMo subsurface b) high
resolution image of the uppermost subsurface c) SAED of Si and O-rich
material d) SAED of CoCrMo single-crystal.
197
TEM-EDS map of the CoCrMo flat component in Figure 7-23 shows evidence
of particles suspending in the Si and O-rich structure which appear to be
concentrated in Co and Cr thus confirming the presence of CoCrMo
particulates in the layer. The purpose of the white-line drawn is to create a
demarcation between the Pt deposits and the Si and O-rich layer. N is also
evident in the film as identified also through XPS, although the specie
identified was the organic-N specie.
Figure 7-23 – TEM-EDS map of the CoCrMo subsurface from the Si3N4 –
CoCrMo material combination. Above the white-line are deposited Pt
and empty space.
198
TEM cross-section of the Ti6Al4V alloy from the Si3N4 – Ti6Al4V material
combination is shown in Figure 7-24, this figure represents the region covered
by the ‘mechanically mixed layer’ denoted region ‘C’ in Figure 7-16. Three
region of interests are identified in the first micrograph in Figure 7-24a. The
first, denoted structure ‘1’ represents the ‘mechanically mixed layer’ that exists
with heights over 2 μm above the plain Ti6Al4V surface (see Figure 7-13,
Ti6Al4V at ±50 μm). The material visually appear highly densified and non-
crystalline with some horizontal cracks across it. In addition, nanoscale
particles are suspending within the structure.
The second, denoted structure ‘2’ is separated by a clear demarcation which

indicates that the ‘mechanically mixed layer’ moves relative to the plain
surface thus explaining why elements from the physiological solution was
present across the surface where the mechanically mixed layer had ‘flaked
off’ from in Figure 7-18. It also confirms the evidence of plastic flow and
directionality observed in Figure 7-13. Figure 7-24b shows a higher
magnification view of the ‘mechanically mixed layer’. In structure 2, a large
particle-like material of dimension ~2.5 x 3.0 μm suspending in the non-
crystalline, ‘mechanically mixed material’ of similar appearance to that of
structure1 is shown. Further down in the subsurface cross-section, the third,
denoted structure ‘3’ is a much larger Ti6Al4V bulk. It can be seen that the
structure had suffered fracture into smaller segments due to further ingress of
the non-crystalline ‘mechanically mixed layers’ i.e. structures 1 and 2. This is
depicted clearer in Figure 7-24c.The fracture is seen to have propagated
deeper into the alloy beyond the scope of the TEM micrograph.
SAED analyses were conducted on the three areas and are shown in Figure
7-24d – f consecutively. Quite interestingly, SAED reveals that all three
structures are amorphous in structure. This suggests that Ti6Al4V underwent
a full transformation to an amorphous structure at this interface as similarly
observed with Ti6Al4V at the CoCrMo – Ti6Al4V interface. However, as it is
evident that the energy experienced at the ceramic – metal interface is larger
than the metal – metal interface, what appears like a brittle fracture of
amorphous Ti6Al4V structure is observed under the ceramic – metal
condition.
199
1
2
(b)
(a) (c)
1 2 3
(d) (e) (f)
combination (region of mechanically mixed layer): a) TEM micrograph of
the Ti6Al4V subsurface b) high resolution image of the mechanically
mixed layer c) high resolution image of the transformed amorphous
Ti6Al4V structure d) SAED of the mechanically mixed layer e) SAED of the
suspending large particle f) SAED of fractured region.
200
The TEM-EDS of the same cross-section shown in Figure 7-25. The
observation complements the species observed using XPS. The
‘mechanically mixed layer’ was identified as Ti and Al silicates by the XPS and
the map in this figure shows the layer to be mainly Ti, Al, Si and O-rich. It is
also evident that the layer migrated deeper beyond the uppermost surface. Si
is identifiable in the large particle-like material denoted structure ‘2’ from
Figure 7-24a thus further confirming that the particle is no longer a crystalline
Ti6Al4V bulk material. Rather, it has undergone a complete transformation
from crystalline to nano-crystalline and finally to an amorphous structure. N
is equally dispersed across the structure hence it is arguably absent within the
structure although XPS identified organic-N species at the uppermost surface.
Cl seems to be faintly present specifically in the region where the amorphous
Ti6Al4V particle-like materials are. This highlights the possible role of the
physiological solution in the full transformation of crystalline bulk Ti6Al4V to
an amorphous structure. One cannot confidently say C and S are present in
any specific region.
Figure 7-25 – TEM-EDS map of the Ti6Al4V subsurface from the Si3N4 –
Ti6Al4V material combination – region of mechanically mixed layer.
201
Figure 7-26 shows the TEM and SAED analysis of the region on Ti6Al4V
surface where ‘mechanically mixed layers’ were not formed. These regions
are generally located in-between ‘mechanically mixed layers’. Figure 7-26a
shows a totally transformed structure and the micrograph also depicts what
appears to be subsurface rolling of nano-crystalline Ti6Al4V structures. A
significant ‘bust’ of energy most likely inflicted during the stiction mechanism
is attributed to be the cause of such a transformation.
There are three main regions of interest: In structure ‘1’, horizontal cracks run
across the uppermost region. Structure ‘2’ appears to be a densely packed
amorphous structure of over 4 μm in depth. Structure ‘3’ is a region of densely
packed nano-crystalline structures with evidence of directionality. A high
resolution image in Figure 7-26b reveals the process through which crystalline
structures reach its smallest size in nano-meters closest to the top surface
before it gets ‘swamped into the sea’ of amorphous Ti6Al4V structures.
SAED analyses of the three main area reveal an amorphous structures for
structures 1 and 2 as expected and a nano-crystalline structure in structure 3
as expected.
3
2
(a) (b)
202
1 2 3
(c) (d) (e)
combination (region absent of mechanically mixed layer): a) TEM
micrograph of CoCrMo subsurface b) high resolution image of the nano-
crystalline region c) SAED of the uppermost region d) SAED of the totally
transformed region e) SAED of nano-crystalline area.
The TEM-EDS cross-section of Ti6Al4V on which no ‘mechanically mixed

layers’ are formed is shown in Figure 7-27. It is observable that Ti and Al are
less abundant in the amorphous Ti6Al4V region as compared to the nano-
crystalline region. On the other hand, O and Cl can be seen to be more
abundant in the amorphous region compared to the nano-crystalline region. V
is seen to be equally distributed across both amorphous and nano-crystalline
structures however, some V-rich region is seen within the nano-crystalline
structure thus indicating the β-phase structure within the nano-crystalline
Ti6Al4V alloy. Si, C and S seem to be present as background signal – this is
generally attribute to all elements. N may be present in the structure although,
only in trace amounts that is generally present in bulk Ti6Al4V alloy.
203
Figure 7-27 – TEM-EDS map of the Ti6Al4V subsurface from the Si3N4 –
Ti6Al4V material combination – region absent of mechanically mixed
layer.
7.8 Nano-indentation
Nano-indentation hardness of the Si and O-rich ‘mechanically mixed layer’

was performed following the same procedure outlined in section 4.5.9. Figure
7-28a shows the raw P vs. h data and in Figure 7-28b, the average hardness
of the layer is 22.7 GPa. The hardness measured here is a little less than the
hardness measured at the CoCrMo – Ti6Al4V interface. Nevertheless, the
hardness of the layer is almost double the general bulk hardness of bulk Si3N4
(~13GPa) and about five-times the hardness of Ti6Al4V alloy.
204
(a) (b)
Figure 7-28 – Nano-indentation of layer formed at the Si3N4– Ti6Al4V

interface: a) P vs. h graph of indentations b) average hardness of the
layer.
7.9 Fretting of Biolox Ceramic on CoCrMo and Ti6Al4V alloy
In this section, Biolox was assessed against CoCrMo and Ti6Al4V alloys
under the same parameters and conditions as with all other material
combinations. However, as it was not the main focus of this study, only the
±50 μm displacement was assessed with a sample size of n=1. The aim of
assessing the ceramic – metal combinations only at this displacement was to
observe the third-body interactions that would occur at the Biolox – metal
interface during a gross slip regime. It was observed that a gross-slip regime
for Biolox – Ti6Al4V was not achievable at the initial contact pressure of 1
GPa, therefore, a second Biolox – Ti6Al4V test was conducted at a lighter load
with initial contact pressures of 0.77 GPa. The three tests were
complemented with OCP measurement, 3D and 2D surface profilometry and
SEM-EDS surface analysis.
205
7.9.1 Fretting wear mechanism of Biolox – CoCrMo
In Figure 7-29a, the fretting loop shows a gross slip fretting regime established
at the Biolox – CoCrMo contact. The figure shows three super-imposed
fretting loops: one in the early stage of the fretting cycles (500th) then another
in the middle (3000th) and the fretting loop towards the end of the fretting
cyclea (5750th). It can be seen that the tangential force increased from the
middle of the fretting cycles through to the end. The slight increase in the
tangential force was reflected in the energy dissipated curve of Figure 7-29b.
The OCP in Figure 7-29c shows a cathodic shift at the initiation of fretting and
the cathodic shift continued gradually during fretting as did CoCrMo – CoCrMo
and Si3N4 – CoCrMo at ±50 μm displacement. However after ~3000 cycles, a
sharp rise in potential was observed during ongoing fretting at the contact. At
the end of fretting, a sharp repassivation stage was observed followed by a
slower ennoblement stage.
(a)
206
(b) (c)
Figure 7-29 – Fretting of Biolox – CoCrMo material combination at ±50

µm, Pmax = 1 GPa: a) 2D fretting loop b) Energy dissipated per cycle c)
OCP measurement.
The 3D and 2D surface profilometry for both ball and flat components are
shown in Figure 7-30. No evidence of wear was observed on the Biolox
surface, rather, third-body products are observed of heights reaching up to 3
µm. The CoCrMo flat on the other hand has a wear scar with wear depths
reaching up to 4 µm. This is more than double the wear depths measured at
any of the CoCrMo – CoCrMo/Si3N4 – CoCrMo material combinations. This
indicates that the mechanism involved at the Biolox – CoCrMo interface is
more degradative than other materials coupled with CoCrCo. Evidence of
redistributed third-body products within the CoCrMo wear scar is seen. It is
worth noting that the region outside the wear surface of the flat component is
a visual effect which depicts material gained outside of the wear surface as
though they were material loss.
207
Figure 7-30 – 3D and 2D surface profilometry of Biolox – CoCrMo at ±50
µm displacement.
BS-SEM micrographs of CoCrMo flat surface in Figure 7-31a depicts a wear

surface that takes on the smoothness of the ultra-smooth Biolox counter-body.
However, a few wear track marks at the centre of the wear surface is still
visible. In the centre of the wear scar also lies third-body products and outside
the wear scar are precipitates as indicated in the figure.
SEM-EDS in Figure 7-31b confirms the chemical composition of the third-body

products within the wear surface to be mainly of Cr and O as well as C
although appearing very faintly. The precipitate outside the wear surface on
the other hand is seen to be rich in O, Ca, P, N and C but lacking any of the
metallic elements. Therefore, the precipitation outside the wear surface may
be electrochemically linked with the oxidation reactions occurring within the
fretting contact during on-going fretting. The absence of Al is further evidence
that the Biolox ball did not experience wear at this interface.
208
Third-body Precipitate
(a)
(b)
Figure 7-31 – SEM-EDS map of Biolox – CoCrMo flat surface at ±50 µm

a) BS-SEM micrograph b) SEM – EDS
209
7.9.2 Fretting wear mechanism of Biolox – Ti6Al4V (mixed regime)
The Biolox – Ti6Al4V material combination assessed at 1 GPa is shown in

Figure 7-32. The fretting loop in Figure 7-32a shows a gross slip behaviour in
the initial stages (500th cycle) of the fretting cycles. In addition, it is observable
that the interfacial slip during this stage was ~30 µm at a relatively low
tangential force of ~20 N. However by the 3000th cycle, the fretting regime had
transitioned to quasi-static fretting loop describing a partial-slip behaviour. The
point of transition is identified with the red arrow in Figure 7-32b.
The transition also corresponds to a drop in the energy and surface slip but a
significant increase in tangential force reaching up to ~65 N. As fretting at the
interface began with a gross slip regime, electrochemical response was that
of a cathodic shift of ~200 mV as expected. In Figure 7-32c the potential is
seen to rise gradually to a plateau. Interestingly, as soon as fretting ceased,
there was no repassivation stages observed thus indicating that repassivation
at the interface may be hindered.
(a)
210
(b) (c)

µm (Pmax = 1 GPa): a) 2D fretting loop b) Energy dissipated per cycle
c) OCP measurement.
The 3D and 2D surface profilometry of both ball and flat components show no
evidence of wear on the Biolox ceramic. However, its surface is seen to be
covered in third – body material as also observed against CoCrMo. These can
be seen to reach heights of up to ~3 µm. On the other hand, the Ti6Al4V flat
component is seen with third – body products agglomerated on one side of
the wear surface with heights reaching up to ~1 µm. Cracks with depths of up
to ~4 µm were observable on the Ti6Al4V surface also, particularly in the
region where the third – body material agglomerated. This suggests that the
heap of third – body material on the Biolox surface, located at the centre is a
material pulled out from within the Ti6Al4V considering the surface
profilometry indicates a shared interface in both. This is suggestive of a
localised corrosion phenomena.
211
µm displacement (Pmax = 1 GPa).
The BS-SEM micrograph of the Ti6Al4V surface is shown in Figure 7-34a. The
fretting surface is contained within the red-dotted circle. Outside the dotted
red circles are darker appearing material deposits (precipitates) agglomerated
around the wear surface. However, within the wear surface are found a region
with evidence of galling while in a deferent region a similar shade of grey with
the Ti6Al4V background is observed; 3D surface analysis identified this
region as the area with agglomerated third-body product. The region within
the red-dotted square indicates the area where TEM sample was obtained
using the FIB (the data from this analysis is not presented).
The SEM-EDS maps in Figure 7-34b offers insight into the complexities of this
particular mixed fretting contact. Ti, Al and V are observed abundantly across
the entire surface as expected for a Ti6Al4V bulk. O appears to present across
the wear surface except for the region where evidence of galling in the Ti6Al4V
wear surface was identified. C is seen to be highly concentrated in the region
outside wear surface. N is seen as background signal thus its presence cannot
be verified using EDS. Ca is seen to form outside the wear surface as similarly
observed for the Biolox – CoCrMo interface. P is seen in the same region as
Ca however it appears with a higher intensity. Cl on the other hand is seen to
be highly concentrated within the wear scar, and specifically located within in
the region filled with third – body products. Ti is seen to be present across
both regions rich in Cl and P while the Al and V were less represented in that
212
region. This strongly suggests that the tribocorrosion products generated at
the Ti6Al4V mixed regime fretting contact can be identified as two main
products: the third – body product trapped in the fretting contact is mixed Ti-
oxide and Ti-chloride where Ti-chloride is the more dominantly present specie.
Outside of the wear surface, compounds of Ca-phosphate or ionic species of
Ca and P are precipitated.
(a)
213
(b)

(1 GPa) a) BS-SEM micrograph b) SEM – EDS
7.9.3 Fretting wear mechanism of Biolox – Ti6Al4V (gross slip

regime)
In Figure 7-35a, the fretting loop captured a rather unstable fretting contact;
one similar to the observations made at the Si3N4 - Ti6Al4V contact at ±50 µm
displacement. This contact also experienced an overshot gross slip
displacement of δx = 85 µm. The energy dissipated per cycle in Figure 7-35b
describes the instability of the fretting contact; a stiction-like mechanism is
also observed at this interface. Stability at the contact was established at
4500th cycle till fretting ceased. Figure 7-35c shows how the initial large gross
slip at the interface led to a sharp cathodic shift of ~550 mV and a maximum
cathodic shift of 630 mV. At the end of fretting displacements, a single sharp
repassivation process occurred. Quite importantly, it was noticeable that the
OCP was fully recovered unlike in other cases of mixed/gross slip fretting
regimes of CoCrMo – Ti6Al4V and Si3N4 - Ti6Al4V.
214
(a)
(b) (c)

µm (0.77 GPa): a) 2D fretting loop b) Energy dissipated per cycle c)
OCP measurement.
215
Figure 7-36 shows the 3D and 2D analysis of the wear surface. No evidence
of wear was observed at the surface of the Biolox, rather, a large heap of third
– body material with heights up to ~4 µm was seen at the edge of the fretting
contact. The fretting surface of the Ti6Al4V is firstly seen to be elevated above
the plain Ti6Al4V surface. Secondly, the wear surface appear to have wear
tracks at its edges while the centre of the wear surface appear smooth and at
lower depths than its surroundings. The 2D wear depth profile reveals a large
crack within the surface of depth ~7 µm. The depth is located where the larger
material deposit on the Biolox is also located thereby suggesting that the heap
of material was removed directly from the Ti6Al4V bulk subsurface thus it isn’t
a material transfer.

µm displacement (Pmax = 0.77 GPa).
BS-SEM micrograph of Figure 7-37a shows the wear surface of this gross slip
contact. Evidence of wear track is identifiable beneath the smooth third – body
material that lies at the centre of the wear surface. The smoothness of the
surface confirms that the Biolox counterpart had not experienced wear at this
interface. The SEM-EDS of Figure 7-37b shows that Ti is almost equally
abundant across the entire wear surface while Al appears to be more
concentrated within the wear track which appears to be devoid of O. This
suggests that Al is less concentrated in the third-body product. V appears only
to be faintly present across the entire surface. O appears to be more
216
concentrated in the third – body material as expected. C, Ca and P appears
to be absent from the wear surface but are precipitated outside the wear
surface. Cl appear more concentrated within the third – body region and the
presence of N cannot be verified using the EDS map.
Wear track
Precipitate
Third-body
(a)
217
(b)

(0.77 GPa) a) BS-SEM micrograph b) SEM – EDS
In this chapter, ceramic – metal and ceramic – ceramic material combinations

were assessed under varied displacement amplitudes. The two ceramic
materials assessed against CoCrMo and Ti6Al4V were Si3N4 and Biolox.
These offered two different perspectives of fretting corrosion mechanisms that
can be expected at the ceramic – metal interface of modular taper junction in
hip implants. The Si3N4 ceramic represents the mechanisms involved at the
fretting interface whereby the ceramic material is engaged both in wear and
oxidation processes. On the other hand, the Biolox ceramic represents the
mechanisms involved at the fretting interface whereby the ceramic material is
inert both chemically and mechanically. In both cases, an extensive surface
and subsurface analysis of the ceramic – metal interfaces were carried out.
218
Amongst the material combinations assessed in this chapter, the least
compliant contact was the self-mated Si3N4 material combination meaning that
there is less elastic contact compliance to overcome before slip is initiated at
the interface. However, interestingly, self-mated Si3N4 was in a stick fretting
regime at ±10 µm displacement whilst Si3N4 – CoCrMo; a combination with
more elastic compliance was in a partial-slip regime at the same
displacement. This highlights the first fundamental difference between a
ceramic – metal and a ceramic – ceramic system. For slip to be initiated at the
ceramic – ceramic interface, the traction forces must be sufficient enough to
induce fracture at the asperity – asperity level. Whereas, when considering
other material combinations such as the self-mated CoCrMo and Si3N4 –
CoCrMo, less tangential force is required for the transition from static friction
to kinetic friction i.e. stick regime to partial-slip regime to occur.
On the opposite extreme, a significant interfacial compliance occurs at the

Si3N4 – Ti6Al4V interface hence the contact remains in a stick regime even at
±25 µm. 3D surface topography of the two stick regimes at the Si3N4 – Ti6Al4V
interface reveal evidence of a growing worn annulus in the Si3N4 and a
plastically deformed Hertzian contact boundary in the Ti6Al4V alloy (Figure 7-
13). Thus, it is the case that the multiple times harder Si3N4 ceramic wears
relative to the Ti6Al4V which on the other hand experiences plastic
deformation at the edges of the Hertzian contact. The wear and corrosion
products generated within the stick regime fretting contact are generally
trapped at the interface.
In the absence of an in-situ imaging technique, the mechanical structure of

the third-body products trapped at the interface cannot be characterised.
Nevertheless, the examination of 3D surface topography of the third-body
products at these regimes after fretting lead us to be that the third-body
material displays visco-elastic properties. This is evident by the ‘needle-like’
structures observed in both Figure 7-13 (±10 µm and ±25 µm) and Figure 6-
10 (±10 µm) and by the plastic flow observed in Figure 7-13 (±50 µm).
Furthermore, assuming the ‘needle-like’ structures form upon separation of
the Si3N4 ball from the Ti6Al4V flat after the fretting test is completed; then the
3D surface topography images of the Si3N4 – Ti6Al4V sticking fretting contact
219
may be offering a depiction of the real (in-situ) contact area at the tribological
interface. Such an observation may prove useful for computational modelling
of real interfacial contacts.
Across all the material combinations, gross slip/mixed regimes were realised
at ±50 µm displacement, hence the contact is less governed by elastic
interfacial compliance. Therefore, the characteristic fretting wear mechanism
for each material combination was described mainly on their characteristic
behaviour at ±50 µm displacement. These are outlined as follows:
Si3N4 – Si3N4: This combination was mainly governed by a fracture-induced

wear mechanism at low displacement amplitudes while it progressively
transitioned towards an abrasive wear mechanism as displacement increases
(see Figure 7-14). This observation agrees with a previous study [203].
Si3N4 – CoCrMo: The wear mechanism at the interface of this material

combination was dominantly governed by tribochemical interactions. This was
also the case in a previous study which assessed a Si3N4 – Steel system [206].
Interestingly, in the referenced study, Fe from the steel contributed to the
tribochemical material formed at the interface but not Cr. And in the present
study, it was observed that Cr-oxide – the natural oxide formed during
repassivation in a tribocorrosion process of CoCrMo was suppressed from
forming. The tribochemical product observed at this interface was SiOx -OHy.
3D and 2D surface analytical techniques revealed that the SiOx -OHy
tribochemical product largely influence wear at the interface by inhibiting the
severe wearing of CoCrMo. Rather, the Si3N4 counter-body suffered greater
wear-depths than CoCrMo at ±50 µm displacement. However, as energy is
not consumed, the interfacial energy at the Si3N4 – SiOx -OHy interface,
culminated into subsurface fatigue transformations in the CoCrMo alloy. This
was manifested in subsurface twinning up to 7 µm deep into the alloy.
Si3N4 – Ti6Al4V: Tribochemical reactions were also highly influential in the

manifested wear mechanisms and fretting regime at this interface. Unlike the
Si3N4 – CoCrMo interface where only Si was involved in the tribochemical
product formed, both material couples were engage in the oxidation processes
at this interface. This led to the formation of mixed Ti and Al silicates as well
220
as oxides of Ti and Si. The material property of the product formed at the
interface is suggested to play a significant role in the intermittent stiction
mechanism experienced at the contact during fretting at ±50 µm displacement.
The proposed mechanism of stiction is addressed in further detail in section
10.3.2. The implication of having a fretting couple with relatively large
differences in their hardness and stiffness is that, energy is not easily
dissipated through wear. If the fretting interface is not subjected to a large
enough displacement amplitude to facilitate gross sliding behaviour, the less
stiff material would experience a significant metallurgical transformation which
may involve subsurface cracking while the much harder ceramic material
experiences fracture-induced wear. This can be catastrophic for a ceramic
with low fracture toughness.
TEM-EDS analysis showed no evidence of Si3N4 wear particle in any of the

structures analysed in this chapter. This suggests that particulates of Si 3N4
and SiO2 from the bulk material were likely crushed prior to being mechanical
mixed into oxides of Ti and Al thus engaging in oxidational process under high
contact pressures at the interface to form a new tribochemical product (Ti and
Al silicates). A similar process is described in a previous study [207].
Biolox – CoCrMo and Biolox – Ti6Al4V (gross slip): 3D surface

topography of Figure 7-30, Figure 7-33 and Figure 7-36 all reveal an interfacial
mechanism whereby the ultra-smooth Biolox ball remained chemically and
mechanically inert throughout the fretting stages. Interestingly, it appears to
play the role of using third-body wear and corrosion products of CoCrMo and
Ti6Al4V alloys to wear against themselves. For the Biolox – CoCrMo contact,
an abrasive mechanism is prevalent and for the Biolox – Ti6Al4V material
combination, a galling mechanism is prevalent. The galling mechanism which
is generally a result of interfacial adhesion can also be characterised by a
highly fluctuant dissipated energy curve as shown in Figure 7-35. This is a
typical mechanism observed for contacts where titanium alloy is dominant at
the interface [208-210].
Figure 7-38 is a summary plot of the cathodic shift measurement for all the
material combinations assessed in this chapter at various fretting
221
displacements. It is can be seen that the cathodic shift experienced at the stick
and partial-slip regimes of both Si3N4 – CoCrMo and Si3N4 – Ti6Al4V are
minimal. This is because at low fretting displacement amplitudes, a relatively
small area is depassivated i.e. exposed bare active metal [146]. Furthermore,
in most cases, as majority of the fretting contact remain unexposed to the
electrolyte, OCP curve displays contact ennoblement despite on-going
fretting. However, across the gross slip/mixed regime fretting contacts where
large area of depassivation occurs i.e. at ±50 µm displacement, cathodic shifts
were observed as expected.
Similarly as observed for metal – metal combinations, the maximum cathodic

shift measured at the fretting contacts involving Ti6Al4V are significantly larger
than those of CoCrMo. In the case of ceramic – metal combinations involving
Ti6Al4V, the magnitude of Max. cathodic shift and the repassivation kinetics
(from OCP observation) was observed to be mainly influenced by the fretting
regime. As for the Biolox – Ti6Al4V material combination subjected to a
partial-slip regime, third-body products accumulated at the interface were
unable to escape the contact. Although the magnitude of cathodic shift was
relatively small compared to the cathodic shift at Biolox – Ti6Al4V under gross
slip regime, the repassivation kinetics at the partial-slip contact was indicative
of active localised corrosion mechanisms. This is evidenced by the MACC
process identified at the interface. Excess metal ions as O is depleted in the
partial-slip contact attracted Cl ions into the region as depicted in the SEM-
EDS micrograph of Figure 7-34. This may explain why repassivation process
at the interface was inhibited; it is well known that Cl ions play a significant
role in the inhibition of passive oxide restoration thus subsequently triggering
localised corrosion mechanisms [27, 211].
In the case of Si3N4 – Ti6Al4V where a stiction mechanism was dominant,

repassivation kinetics were largely influenced by the mechanically induced
cracks within the Ti6Al4V surface as evidenced in the 3D surface topography
of Figure 7-13, ±50 µm displacement. The same observation was made for at
the CoCrMo – Ti6Al4V interface in the previous chapter. The Biolox – Ti6Al4V
contact at gross slip regime experienced the largest cathodic shift of all the
combinations. However, as there was adequate exchange in the physiological
222
solution during on-going fretting thus restoring access to O at the interface,
the third-body product observed was mainly Ti-oxide. In the scarcity of O
supply, like in the partial-slip fretting contact, the main corrosion product would
be metal-chlorides.
Maximum cathodic shift in both Biolox – CoCrMo and Si3N4 – CoCrMo at ±50
µm were very similar despite forming two different oxidation products at their
interfaces. This confirms that the cathodic potential measured at both ceramic
– metal interfaces is more representative of the exposed active site in the
CoCrMo alloy and less about the reaction kinetics. The main third-body
product in the case of the former is Cr-oxide while in the latter was SiOx -OHy
compound.
Figure 7-38 – Comparing Max. cathodic shift in ceramic – metal

combinations.
223
Chapter 8 Quantification of fretting corrosion current using
self-mated CoCrMo material combination
8.1 Introduction
So far, fretting of metal – metal and ceramic – metal combinations have been
assessed under OCP conditions. However, the quantification of fretting
corrosion current and Faradaic mass loss as a function of displacement
amplitude is yet to be addressed. Thus, the purpose of the experiments in this
chapter is to quantify corrosion current and assess the effect of surface history
on the evolution of fretting current and interfacial energy. By the phrase ‘effect
of surface history’ it is anticipated that the fretting interface of the modular
taper, experiences diverse forms of oxidation processes and wear
mechanisms which may alter the contact conditions and fretting regime in-vivo
[199]. In the context of tribocorrosion, changes to the contact condition or film
formation can strongly influence the synergies of a fretting contact.
The study was carried out on self-mated CoCrMo using three methods
schematically outlined in section 4.4.2.2: ‘Method 1’ represents fretting tests
conducted for individual displacements; ‘Method 2’ represents fretting tests
conducted in series with increasing displacement amplitudes and ‘Method 3’
represents fretting tests conducted in series with decreasing displacement
amplitude. A schematic outline of the experiments and analysis performed in
this chapter are shown in Figure 8-1.
224
Figure 8-1 – Outline of the fretting experiments performed in chapter 8
Figure 8-2 shows the slip ratio for the four displacements assessed across the
three methods. In comparison to the gross slip regime criteria, it can be seen
that all the displacements amplitudes ≥ ±25 µm experienced gross slip fretting
regime. Whereas, all displacements at ±10 µm experienced a partial-slip
regime except in method 3 where the slip ratio is observed to have exceeded
the gross slip criteria.
225
Figure 8-2 – Slip ratio of CoCrMo – CoCrMo under potentio-static
conditions for ±10 – ±150 µm in ‘Methods 1 – 3’.
8.2 Anodic current in static condition
Following the potentio-static polarisation of the loaded ball-on-flat contact,

from OCP to 0V vs. Ag/AgCl, the fretting current dropped exponentially as
expected; this is generally indicative of surface coverage of anodic oxide film
[212]. The curve follows a power relationship as shown in Figure 8-3. The time
‘t’ = 4000 sec corresponds to an end of the equilibrating stage and the initiation
of fretting displacements to the contact. The mean average and standard error
of n = 14 samples at this point is 714 ± 160 nA. This value corresponds to the
pure corrosion current ‘C’ effective at the contact under static conditions.
226
Figure 8-3 – Anodic current transient of loaded CoCrMo – CoCrMo
contact subjected to 0V vs. Ag/AgCl of potentio-static polarisation for
4000 seconds in static conditions.
8.3 Method 1
Figure 8-4a shows the fretting loop of the four displacement amplitudes.
These are seen to be very similar to the CoCrMo – CoCrMo contact under
OCP condition in (section 6.2.1). Each fretting loop represents the fretting
contact at the 1000th cycle. This cycle corresponds to the fretting loop at
middle of the total length of cycle (2000 cycles). The non-symmetrical loop at
the ±150 µm is indicative of an asymmetric wear surface (see section Figure
10-2 for further details).
Figure 8-4b shows the average plot of energy dissipated per cycle for each
displacement variable. The energy dissipated is seen to be higher with higher
displacement amplitudes as expected. However, it is observable that the
energy drops suddenly at ±25 µm to a value close to that of the partial-slip
regime at ±10 µm. This is indicative of a momentary transition from gross-slip
to partial-slip and then back to gross slip again. An inset image is included to
227
highlight the sudden drop in the energy dissipated at ±150 µm which could not
be observed properly in the log plot.
(a)
(b)

conditions. a) Fretting loop b) Energy dissipated per cycle.
228
8.3.2 Electrochemical results
Figure 8-5a is a representation of the in-situ fretting current transient as a

result of oxidation processes of each fretting displacement. The result shows
that fretting current transient at the partial-slip regime is near zero as
expected. However at the gross slip regime, a distinctive fretting corrosion
current transient is observed. At ±25 µm, the sudden transition between
partial-slip to gross slip regime indicated in the energy dissipated is fully
reflected in the fretting current. The fretting current at ±50 µm was not
significantly higher than at ±25 µm, however, in comparison to ±150 µm, the
magnitude is almost double.
In Figure 8-5b the time-dependent fretting current is represented as

cumulative charge which corresponds to the area encompassed within the ‘I
vs. t’ plot of Figure 8-5a. Thus as expected, the difference in magnitude of
cumulative charge for the various displacements in Figure 8-5b are seen to
be very comparable with that of the fretting current in Figure 8-5a.
(a)
229
(b)

at ±10 µm, ±25 µm±, 50 µm, ±150 µm displacement, a) Fretting current
transient b) Cumulative charge
8.3.3 Surface analysis
3D surface profilometry of both CoCrMo ball and flat surfaces are shown in
Figure 8-6. The general trends observed are as expected: a) the size of the
fretting wear surface increased with increase in fretting displacement; b) the
size of central partial-stick region reduced in proportion to the worn annulus
as displacement increased. From observation of the ball component at ±50
µm and ±150 µm, it is observable that the main difference between the two is
that the central region had worn significantly (although not entirely) in depth
at ±150 µm.
230
Third-body
Central
partial-stick
region
Worn
annulus

fretting interface assessed under potentio-static conditions.
8.3.4 Tribocorrosion mass loss
Figure 8-7a shows the wear volume from all the displacements. Interestingly,
the pattern of increase from the lowest fretting displacement to the highest
corresponds to the cumulative charge increase in Figure 8-5b. In general, the
result shows that the wear volume increases with increasing displacement.
However, amongst the gross slip fretting displacement increase in wear was
not proportional to the increase in displacement amplitude. The wear volume
at the gross slip contact of ±25 µm is about 6 times higher than at ±10 µm thus
231
showing the effect of regime transition from partial-slip to gross slip. The re-
distributed volume is considered the volume of material that sits within the
wear surface but above the zero-line (see section 4.5.2). These were
negligible at all fretting displacements.
In Figure 8-7b, the Faradaic mass loss is reported as a percentage ratio of

Mmech and Mchem. The result shows that at the partial-slip regime, the mass
loss is completely a result of chemical oxidation and for the gross slip contacts,
the ratio of Mmech increased with increasing displacement and so does the
overall mass loss. However for all displacements, Mchem is the largest
contributor to total mass loss.
In Figure 8-7c, the total tribocorrosion mass loss which incorporates both
Mchem and Mmech is plotted against the cumulative energy corresponding to
each displacement. The graph shows a linear correlation between ±10 µm and
±50 µm (inclusive). These are the same displacements according to Figure 8-
6 where the partial central stick region are still present thus fretting at these
displacements are still highly influenced by elastic contact compliance.
However, from ±50 µm to ±150 µm, a less steep gradient is observed thus
highlighting the influence of third-body interaction at the interface [39].
(a) (b)
232
(c)
Figure 8-7 – Method 1: the sum of tribocorrosion mass loss from ball
and flat components; a) Wear volume (blue) and Re-distributed volume
(orange); b) Mass loss as a proportion of Mmech and Mchem, c) Total
mass loss vs. Cumulative energy.
The contribution of pure corrosion (C) to the total mass loss for the three
methods are shown in Table 8-1. In method 1, the maximum mass loss
contribution from pure corrosion was calculated from the sum of the largest
worn areas of ball and flat components i.e. the ball and flat surfaces of ±150
µm displacement (see section 4.4.2.3 for the calculation). The total
contribution was ~0.2 ng. As for methods 2 and 3 whereby all four
displacements were performed in series, the contribution of C was estimated
to 0.4 ng. Thus, in all three methods, the contribution of pure corrosion was
found to be negligible compared to the Faradaic mass loss in Figure 8-7b.
Nevertheless, these values were included in the calculations of Mchem.
233
Table 8-1 – Breakdown of static current, Faradaic mass loss for both
the entire surface sample surface and the fretting surface.
Static current Faradaic Mass Faradaic Mass

loss loss
[Entire surface]
[Entire surface] [Fretted surface]
@T = 4000 sec
(µg) (µg)
(nA)
Method 1
0.343 ± 0.077 0.0002 ± 0.000045
(Max. per
710.4 ± 160
displacement)
Methods 2 and 3 0.705 ± 0.167 0.0004 ± 0.000039
8.4 Method 2
Figure 8-8a shows the fretting loop generated from method 2. The fretting
loops observed from this method are the same regimes in every displacement
as identified in method 1, although, with subtle differences in the gross slip
shapes.
Figure 8-8b shows a super-imposed time-dependent plot of slip ratio and

energy dissipated for the four displacements tested in series to each other.
The main points of observation are the drop in energy seen at ±25 µm and
±150 µm. When considering the slip ratio corresponding to the ±25 µm, the
drop in energy may be correlated with the stabilisation of slip at the interface.
However for the ±150 µm, the energy dropped significantly despite the slip
ratio remaining constant; this implies the drop in energy is directly correlated
to the interfacial friction.
234
(a)
(b)

235
Figure 8-9a shows the time-dependent fretting current transient for each
displacement in an increasing order. As expected, the fretting current at ±10
µm is minimal and as the contact transitions into gross slip regime, the fretting
current increased significantly. Unlike in method 1, the fretting current at ±50
µm was almost double the current at ±25 µm which also corresponds to the
ratio between the two displacements. Interestingly, the fretting current was not
significantly higher at ±150 µm and after just a few hundred cycles, the fretting
current dropped rapidly in a manner that corresponds to the drop in the energy
dissipated plot in Figure 8-8c. These highlights the effect of surface history as
later discussed in the summary section of this chapter. Figure 8-9b shows the
cumulative charge from each displacement and as expected, the magnitude
of the charge differences from each fretting displacement corresponds with
the fretting current transient.
(a)
236
(b)

at ±10 µm, ±25 µm±, 50 µm, ±150 µm displacement; a) Fretting current
transient b) Cumulative charge
The wear surface from method 2 is shown in Figure 8-10. Evidence of surface
grooving can be seen from the 3D surface profilometry. It appears as though
third-body product on the ball had inflicted excessive wear at a localised
central region of the flat component. And across the entire flat wear surface
evidence of grooving is also noticeable due to the presence of thin grooves
which are aligned with those on the ball component.
237
Ball Flat
Evidence of
grooving

fretting contacts tested in potentio-static conditions.
Figure 8-11a shows that the wear volume for method 2 which totals to 520000
± 43000 µm³. Figure 8-11b shows the mass loss as a ratio of Mmech and Mchem
for method 2. The magnitude of Mchem measured in this contact is very similar
with method 1; it contributes to 85% of the total mass loss. The total
contribution of Mmech in this method is seen to be roughly half that of the total
contribution in method 1.
(a) (b)

Mass loss as a proportion of Mmech and Mchem.
238
8.5 Method 3
Figure 8-12a shows the fretting loop generated from method 3. It is worth
noting that the slip ratio in this method (see Figure 8-2) was similar with the
other methods for all displacements except the displacement at ±10 µm where
the slip ratio was 3x higher relative to the other methods. Figure 8-12b shows
a super-imposed time-dependent plots of both slip ratio and energy dissipated
for all displacements, in decreasing order.
It is observable that the energy dropped suddenly and significantly during the
initial fretting at ±150 µm. It is also notable that the point where the energy
dissipated dropped corresponds to a small increase in the slip ratio, thus
indicating that the drop in energy is a direct result of changes to the interfacial
friction. Subsequently, the transitions to ±50 µm and ±25 µm resulted a stable
slip ratio and energies dissipated, while slip at the partial-slip regime appear
less stable.
(a)
239
(b)

Figure 8-13a shows the corresponding fretting current transient in the

decreasing order of the fretting displacements. The point where a significant
drop in energy was observed for the ±150 µm is reflected in the fretting current
transient as an intermittent reduction in current. Subsequently, while the
energy remained low as seen in Figure 8-12b, the fretting current transient
was seen to rise to a peak before steadily dropping. This observation was
contrary to method 2 in which case, a continuous drop in energy dissipated
resulted in a continuous drop in fretting current. The fretting currents transient
at ±50 µm and ±25 µm were also relatively higher than in the other methods.
The larger slip ratio at the partial-slip regime in this method resulted in a higher
fretting current compared to the two previous methods as expected.
The cumulative charge for all the displacements are shown in Figure 8-13b.
The values are seen to be higher than all the other methods within the margins
240
of error and the cumulative charge at ±25 µm in this method is seen to be
double the value in method 2.
(a)
(b)
Figure 8-13 – Method 3: fretting current from CoCrMo – CoCrMo

contact at ±150 µm, ±50 µm, ±25 µm, ±10 µm displacement, a) Fretting
current transient b) Cumulative charge
241
The wear surface from method 3 is shown in Figure 8-14. The wear surface
reveal an interesting profile, similar to one already observed at the CoCrMo –
CoCrMo, ±150 µm fretting contact, under OCP conditions (section 6.5.1,
Figure 6-9). The two heaps of material, present on either side of the central
region is a re-distributed third-body product that has become aligned in the
reciprocating fretting direction. The material preserved at the centre of the
surface is what is left of the central partial-stick region. Evidence of third-body
inflicted grooving is observed within the flat wear surface.
Ball Flat
Evidence of
grooving

fretting contacts tested in potentio-static conditions.
Figure 8-15a shows that the wear volume for method 3 is 670000 ± 53000 µm³
which is higher than the wear volume of both methods 1 and 2. Figure 8-15b
shows the mass loss as a ratio of Mmech and Mchem for method 3. The Mchem
measured in this method is much larger than in methods 1 and 2; it contributes
to 93% of the total mass loss. Thus the contribution of Mmech in this method is
seen to be much lower than the two previous methods.
242
(a) (b)

Mass loss as a proportion of Mmech and Mchem.
In this chapter, self-mated CoCrMo was used to assess fretting corrosion

currents for varied fretting displacement amplitudes. Through this study, it was
confirmed that fretting corrosion currents at a partial-slip regime is minimal in
comparison to the gross slip regime. It is importance to highlight this so that a
link can be established between fretting currents and the corresponding
fretting regime. Therefore in the subsequent chapter where realistic taper
components are subjected to fatigue loads, fretting current can be translated
to fretting regimes in the qualitative sense.
The synergistic interactions of wear and corrosion is a topic that has been
widely discussed across several decades [136, 139, 145, 213, 214]. Uhlig’s
mechanistic approach to tribocorrosion mass loss is expressed in Equation
8-1. Equations 8-2 and 8-3 is a widely regarded expression of the contributing
mechanical, chemical and synergistic factors contributing to Mmech and Mchem.
The study by Bryant and Neville [139] critiqued this expression and highlighted
243
that the synergistic factor ‘Cw’ which is solely placed under the broad umbrella
of Mmech is unfittingly positioned considering that it represents aspects of
oxidational processes.
𝑀𝑡𝑜𝑡𝑎𝑙 = 𝑀𝑐ℎ𝑒𝑚 + 𝑀𝑚𝑒𝑐ℎ 8-1
𝑀𝑐ℎ𝑒𝑚 = 𝐶 + 𝑊𝑐 8-2
𝑀𝑚𝑒𝑐ℎ = 𝑊 + 𝐶𝑤 8-3
The general approach for determining Cw is shown in Equation 8-4; where ‘W’
is obtain by suppressing corrosion through cathodic polarisation. W c is
determined by integrating the area encompassed within the tribocorrosion
current transient like those presented in Figure 8-5a. C is determined by
integrating the area encompassed within the period of fretting from zero
current to the measured current whilst the contact was in static conditions (see
Figure 4-1). However, determining Cw in this manner undermines its
magnitude, be it positive (synergistic) or negative (antagonistic). It is meant
that, the unknown proportion of electrochemical contributions of Cw in the
tribocorrosion current transient is not considered when Equation 8-4 is used.
For this reason, Equations 8-2 and 8-3 are redefined as Equations 8-5 and
8-6 where ‘α’ and ‘β’ are unknown proportions of Cw. Therefore, the area
encompassed by the fretting current transients is interpreted as 𝑊𝑐 + 𝛼𝐶𝑤 in
this study (see schematic diagram shown in section 4.3.2).
𝐶𝑤 = 𝑀𝑡𝑜𝑡𝑎𝑙 − 𝑊 − 𝑊𝑐 − 𝐶 8-4
𝑀𝑐ℎ𝑒𝑚 = 𝐶 + 𝑊𝑐 + 𝛼𝐶𝑤 8-5
𝑀𝑚𝑒𝑐ℎ = 𝑊 + 𝛽𝐶𝑤 8-6
Positive +𝛼𝐶𝑤 represents a synergistic contribution to tribocorrosion mass

loss through oxidational processed thus, fretting current is increased.
Negative −𝛼𝐶𝑤 represents an antagonistic contribution to tribocorrosion mass
loss through oxidational process leading to a reduction in the fretting current.
244
Positive +𝛽𝐶𝑤 represents a synergistic contribution to tribocorrosion mass
loss through mechanical wear processes thus wear is increased at the
interface. Negative −𝛽𝐶𝑤 represents an antagonistic contribution to
tribocorrosion mass loss through mechanical wear processes leading to a
reduction in wear. Several mechanisms such as tribochemical film formation,
third-body agglomeration and many more can be occurring simultaneously
within a tribocorrosion interface. The unknown (but predictable) summation of
these interactions determines whether Cw in the system is synergistic or
antagonistic.
The contribution of pure corrosion (C) to the total mass loss in this study was
found to be negligible thus Mchem was only 𝑊𝑐 + 𝛼𝐶𝑤 . Pure wear (W) was not
determined in the present study because the ratio of Mchem and Mwear was
deemed sufficient for determining the degradation mechanisms of the three
methods assessed using Stack and Chi’s [136] criteria (outlined in section
3.5.1).
It was observed in Method 1 that Mchem was the main contributor to the total
tribocorrosion mass loss for all displacements. This agrees with Bryant and
Neville’s result which found that Mmech was more dominant at low initial contact
pressures (0.4 GPa) whereas at high initial contact pressures i.e. 1 GPa, the
contribution of Mchem becomes more significant. Stack and Chi’s criteria states
that for 1 < Mchem/Mmech ≤ 10, the degradation mechanism is one of corrosion
enhanced wear. Using this criteria, it is observable that the fretting contacts in
Method 1 is in a corrosion enhanced wear regime for all the displacements as
shown in Figure 8-16. This therefore suggests that, the magnitude of +𝛼𝐶𝑤 in
the expression of Mchem may exceed 𝑊𝑐 hence making the contact a corrosion
enhanced wear dominated. This further highlighting the importance of splitting
𝐶𝑤 into its sub-factors i.e. 𝛼𝐶𝑤 and 𝛽𝐶𝑤 . Furthermore, it was observed that
with increase in fretting displacement amplitude, the degradation mechanism
increasingly transitions towards a wear enhanced corrosion mechanism in
which the contribution of 𝑊𝑐 increases relative to 𝛼𝐶𝑤 . This may be linked to
reduction in contact pressures with larger fretting displacements.
245
The purpose of methods 2 and 3 was to identify the effect of surface history
on the evolution of fretting corrosion so the displacements were performed in
series. One of the key observations from method 2 in which the displacement
amplitude increases was the gradual drop in both fretting current and energy
dissipated (which represents interfacial friction) at the last displacement ±150
µm. The gradual drop in energy is likely due to tribochemical interactions at
the interface which is triggered at a specific charge density as described by
Yan et al [194, 215]. In both their studies, they proposed that the tribochemical
film forms a physical barrier at the tribological interface thus significantly
reducing charge transfer hence the reduction in current. Thus, this behaviour
is characterised with an antagonistic interaction whereby 𝛼𝐶𝑤 is negative
(−𝛼𝐶𝑤 ). Assuming the friction modifying film reduced overall wear, the contact
would also be characterised with a negative −𝛽𝐶𝑤 .
Contrary to method 2, in method 3 where fretting started at the ±150 µm and

decreased, fretting current dropped suddenly in response to a sudden drop in
the interfacial friction however, the drop in current was short-lived despite the
sharply reduced interfacial friction. This behaviour can be characterised partly
by +𝛼𝐶𝑤 as the fretting current was seen to rapidly rose back up. At the least
displacement (±10 µm) where both methods 1 and 2 experienced a partial-slip
regime, interfacial slip (δs) increased 3x higher in method 3 thus rising above
the gross slip criteria and generating higher fretting currents.
The transition to gross slip regime for ±10 µm in method 3 is an indication of

how the contact conditions of a surface with history of wear and corrosion can
significantly change. In this case, the central partial-stick region had been
worn away from the initial large gross slip regime of ±150 µm hence there was
no elastic compliance at the interface by the time the displacement reached
±10 µm. More so, the 3D surface analysis in Figure 8-14 reveal evidence of a
micro-grooved pathway created through third-body re-alignment at the
interface.
Furthermore, Figure 8-16 shows the Mchem/Mmech ratio for the summed up total
mass losses of all the displacements assessed using the three methods.
Using Stack and Chi’s criteria, it is observable that both methods 1 and 2 are
246
in a corrosion enhanced wear regime whilst method 3 is in a corrosion
dominated regime. This observation is clinically relevant in the context of re-
using well-fixed implanted stems during a THA revision procedure. In light of
this result, it is of utmost importance for surgeons to carefully inspect the stem
for signs of wear and corrosion as the effect of such surface history could
accelerate the interface into a corrosion degradation mechanism.
Figure 8-16 – The sum of Mmech and Mchem for all displacements in the
three methods of assessment. The Mchem/Mmech ratio for each method is
also shown above.
247
Chapter 9 Assessment of fretting current from the head-neck
metal-metal and ceramic-metal taper interface
9.1 Introduction
The previous three chapters have assessed fretting corrosion mechanisms of

metal – metal and ceramic – metal material combinations using a fretting
tribometer. However, a rather complex and composite tribological interface is
established at the real modular taper junction. This is exemplified through
various wear patterns and damage modes which are a result of multi-factorial
loading profiles, toggling motion, surface topography and several other factors
as described in literature [216].
In-vitro fatigue configurations such as described in section 3.3.3 is the

standard approach through which short-term and long-term tribocorrosion
tests are carried out on modular tapers. Majority of fatigue assessment aim to
resolve the role of individual factors or combinations of factors on the evolution
of fretting corrosion. The aim of the study presented in this chapter is to assess
the effect of frequency variation, mean load variation and cyclic load variation
for four relevant material combinations namely: CoCrMo – CoCrMo, CoCrMo
– Ti6Al4V, Biolox – CoCrMo and Biolox – Ti6Al4V. A schematic outline of the
loading profile and analysis performed in this chapter are shown in Figure 9-
1.
248
Figure 9-1 – Outline of the fatigue experiments and analyses performed
in chapter 9. B, C and T represents Biolox, CoCrMo and Ti6Al4V
respectively.
249
Clinically relevant contexts are provided for the fatigue cyclic loading profiles
assessed in this chapter. These are outlined below:
Frequency variation – ASTM F 1875 – 98 prescribes a frequency of 5 Hz for

assessing fretting corrosion damage, and it also recommends 1 Hz and/or 2
Hz for experiments that incorporate electrochemical assessment. Several
other studies have conducted fatigue loading experiments using frequencies
ranging from 0.66 Hz to 15 Hz [115, 118, 125, 217, 218]. However, to date,
only the study by Brown et al [116] had assessed the effect of frequency on
fretting corrosion using realistic head-neck taper components and their study
was only conducted on CoCrMo – Ti6Al4V combination. Therefore this section
aim to assess the effect of frequency variation (1, 2 and 5 Hz) on fretting
corrosion current for the four material combinations.
Mean load variation – The study by Bergmann et al [122] highlighted the

significant role that patient weight plays on the peak loads experienced at the
hip. Their study showed that for the DLA of going down the stairs, peaks loads
were more than double from 2 kN for a 75 kg patient to 4.2 kN for a 100 kg
patient. Thus, a 25 kg weight increase yielded an additional 2 kN to the peak
load experienced at the hip. Therefore in this section, the effect of mean load
variation (0.8, 1.8 and 2.8 kN) were examined. By ‘mean load variation’ it is
meant that, the effect of patients’ weight that yields an additional 1 kN load on
an implant is examined for the four material combinations. This was done in
isolation of cyclic load as much as possible by keeping the cyclic load at ±0.5
kN.
Cyclic load variation – Bergmann et al’s study also reported the loading
profile of diverse DLAs, all of which occur at various amplitudes of cyclic
loading. For example, while standing on one leg may result in amplitudes of
~1 kN, stumbling loads can reach amplitudes of up to 9 kN [122]. This section
therefore aims to examine the fretting current of the four material combinations
whilst subjected to varied cyclic loads of ±0.5, ±1.5 and ±2.5 kN.
Systemic effects – Fatigue studies like this have previously being conducted
in an incremental order where parameters are tested in series to each other
250
[118]. Similarly, in this study, all three aspects of the study outlined above
were conducted in series to each other but not without caution. It was
anticipated that systemic effects such as: lengthy anodic polarisation of the
components; migration of the trunnion taper into the female bore over time
and plastic deformation of the materials may occur. Therefore, a series of
checks (partly in accordance with ASTM F 1875 – 98) were implemented after
every section. Thus the systemic effect were monitored.
An example of what a full experiment looks like is shown in Figure 9-2. All
measurements are an average of n = 3 tests except for Biolox – Ti6Al4V which
is an average of n = 2 data. The error bars in this case represent the difference
of two tests because observations from CMM surface analysis reveal
evidence of misalignment. Therefore the results from the third test are
reported separately in section 9.5.
251
Figure 9-2 – Fretting current vs. time plot for a complete fatigue loading
test.
252
9.2 Electrochemical results
In this section, all electrochemical results pertaining to the static and dynamic
stages of the tests are reported. During the static stages, OCP and anodic
current transients from the potentio-static polarisation were monitored
respectively. Subsequently, in the dynamic stages, the fretting current and
cumulative charge of the variable loading profiles are reported. It is worth
noting that in most cases, the scale of the y-axis for both fretting current and
cumulative charge of Biolox – Ti6Al4V are much lower than the other
combinations. This is because a much smaller fretting current emanated from
the contact in comparison to the others.
9.2.1 Open circuit potential
The results in Figure 9-3a shows the OCP profile for each material
combinations. All combination are seen to ennoble over 1500 seconds after
which the anodic polarisation to 0V vs. Ag/AgCl is imposed. Figure 9-3b shows
the average potential for each material combination at the point of polarisation.
The magnitude of the values in Figure 9-3b also corresponds to the average
over-potential applied to each material combination. On average, a similar
over-potential of <200 mV is applied for all the material combinations involving
CoCrMo. It was only for Biolox –Ti6Al4V combination that an average over-
potential of 250 mV was applied.
253
(a)
(b)
Figure 9-3 – OCP of all combinations prior to potentio-static

polarisation of 0V vs. Ag/AgCl, a) OCP vs. Time plot b) average OCP
and applied over-potential.
254
9.2.2 Static anodic current transient
Following the anodic polarisation, the anodic current transient was monitored
for each material combination. Figure 9-4 (a – d) shows the exponential drop
in current for CoCrMo – CoCrMo, CoCrMo – Ti6Al4V, Biolox – CoCrMo and
Biolox – Ti6Al4V respectively. The denotation CC, CT, BC and BT are also
used sometimes to represent the four material combinations respectively. An
exponential curve with the expression 𝑌 = 𝑚𝑋 𝑏 was fitted to each current
transient. As already described in the methodology section (section 4.3.2) ‘m’
represents the peak current (Ipeak) which is the difference between the
maximum current Imax and the base line current at t = ∞ [175].
The result in Figure 9-4 a & b represents the two metal – metal combination
(CC and CT respectively) and the result in Figure 9-4c & d represent the two
ceramic – metal combinations (BC and BT respectively). It is observable that
the peak current for the metal – metal combination is an order of magnitude
higher than the ceramic – metal combination. The same is the case for the
average base line current at t = 7500 seconds in Figure 9-4e which
corresponds to the current immediately prior to dynamic loading. In order to
determine Imax accurately, a higher resolution data sampling rate at 10 points
per sec was used for n = 1 of each combination. As expected, Imax in Figure
9-4f is also measured to be an order of magnitude different in the currents.
These results highlight the effect of having two conductive metal – metal
components vs. a single conductive surface in the ceramic – metal
combinations.
(a) (b)
255
(c) (d)
(e) (f)
Figure 9-4 – Anodic current transient of all material combinations

under static conditions: a) CoCrMo – CoCrMo; b) CoCrMo – Ti6Al4V; c)
Biolox – CoCrMo; d) Biolox – Ti6Al4V; e) average current at the time t =
7500 Sec and f) current at the time t = 0 Sec (Imax).
9.2.3 Frequency variation
The fretting current vs. time graphs for each frequency (1, 2 and 5 Hz) and all
four material combinations are outlined along a single column in Figure 9-5.
They appear in the same order (a – d) of CoCrMo – CoCrMo, CoCrMo –
Ti6Al4V, Biolox – CoCrMo and Biolox – Ti6Al4V respectively. The column on
the right (e – h) corresponds to the average cumulative charge derived from
the fretting current transient. Therefore, consider that the cumulative charge
represents the quantitative average of the fretting currents emanated over a
time period, the cumulative charge (e – h) is used from this point onwards as
the main method of describing fretting current of the material combinations.
256
It is observable from Figure 9-5e that within the margins of error, there is no
significant difference in the cumulative charge for all three frequencies of
CoCrMo – CoCrMo. On the other hand, Figure 9-5f shows a much lower
cumulative charge per 10000 cycles at 5 Hz frequency for the CoCrMo –
Ti6Al4V combination despite a slightly higher fretting current at 5 Hz than at
the other frequencies (Figure 9-5b). This is therefore indicating that the
increase in fretting current was not proportional to the number of cycles of
fretting per sec. As for both ceramic – metal combinations in Figure 9-5g and
h, the cumulative charge decreases as the frequency increases.
(a) (e)
(b) (f)
(c) (g)
257
(d) (h)
Figure 9-5 – Fretting current (a, b, c, d) and Cumulative charge per

10000 cycles (e, f, g, h) of the material combinations; these were
subjected to ±1.5 kN cyclic load and mean load of 1.8 kN with varied
frequencies of 1, 2 and 5 Hz.
9.2.4 Mean load variation
The fretting currents in Figure 9-6a – d are multiple times lower than that of
the frequency variation for all material combinations likely due to the low cyclic
loads used for the assessment of mean load variation. This is due to the
minimal cyclic load used for this assessment. In Figure 9-6e it can be observed
that the cumulative charge for CoCrMo – CoCrMo is higher for mean loads
above 0.8 kN. A similar pattern is seen for CoCrMo – Ti6Al4V in Figure 9-6f.
However, in Figure 9-6g, cumulative charge was about double both the lower
and higher mean loads for Biolox – CoCrMo at 1.8 kN. For Biolox – Ti6Al4V
in Figure 9-6h it is observable that the cumulative charge at loads lower than
2.8 kN are negligible and considering the fretting current plot in Figure 9-6d, it
is observable that the cumulative charge measured at 2.8 kN was not due to
fretting current but only a result of the anodic current. It is important to note
that the anodic current is main constituted by the current generated through
ionic dissolution and the current consumed in the process of film formation.
258
(a) (e)
(b) (f)
(c) (g)
(d) (h)

h) of the material combinations; these were subjected to a cyclic load
of ±0.5 kN at 1 Hz with varied mean load of 0.8, 1.8 and 2.8 kN.
259
9.2.5 Cyclic load variation
From Figure 9-7 it is observable that all material combinations displayed an

increase in fretting current (a – d) and cumulative charge (e – h) as cyclic load
amplitude increased. Therefore, a linear correlation was fitted to all of the
combinations in e – h. It is observable that the rate of cumulative charge
increase for CoCrMo – CoCrMo and Biolox – CoCrMo were similar although
the ceramic – metal combination was higher. On the other hand, the rate of
cumulative charge increased in CoCrMo – Ti6Al4V and Biolox – Ti6Al4V were
much similar. Interestingly, the rate of cumulative charge with increasing cyclic
loads was found to be nearly double for the metal – metal and ceramic – metal
systems with CoCrMo stem relative to those with Ti6Al4V stems.
When the fretting currents in Figure 9-7 (a – d) is examined, it is observable

that for cyclic loads of ±1.5 and ±2.5 kN in all combinations except CoCrMo –
Ti6Al4V at ±1.5 kN an anodic current is observed upon initiation of cyclic loads
at the contact. The anodic current was quite significant for Biolox – CoCrMo
and even more so for Biolox – Ti6Al4V. Interestingly, it was further observed
that all combinations involving CoCrMo experienced a quick transition from
anodic current to fretting current while Biolox – Ti6Al4V experienced a lengthy
transition from anodic current to fretting current. In addition, it appears that the
higher the cyclic load, the longer it take for a complete transition to fretting
current thus indicating a characteristic mechanism of Ti6Al4V alloy upon
abrasion.
(a) (e)
260
(b) (f)
(c) (g)
(d) (h)

h) of the material combinations; these were subjected to mean load of
2.8 kN at 1 Hz with varied cyclic load of ±0.5, ±1.5 and ±2.5 kN.
9.2.6 Systemic checks
For the systemic checks, Figure 9-8e shows that there are no significant
differences in the cumulative charge for CoCrMo – CoCrMo at the end of each
loading profiles. Although Figure 9-8a shows an example of one of the test
where systemic effects were reflected in the fretting currents after the largest
cyclic load of fretting were applied. For the CoCrMo – Ti6Al4V, Figure 9-8f
reveals that the cumulative charge is highest post-mean loads. This is
261
reflected in the magnitude of the fretting current in Figure 9-8b and for this
particular test only, it was observed that the fretting current did not return back
to zero when load was removed. For both ceramic – metal combinations in
Figure 9-8 g and h, the cumulative charge were greatest for the post-cyclic
load variation.
(a) (e)
(b) (f)
(c) (g)
262
(d) (h)

h) of the material combinations; these were subjected to cyclic load of
±1.5 kN and mean load of 1.8 kN for 5000 seconds at 1 Hz.
9.3 Mechanical results
It was not possible in this study to obtain information regarding the relative
displacement occurring at the taper interface during fatigue loading. However,
by tracking the digital position of the cross-head through which the fatigue load
is applied, it is possible to obtain information regarding the displacement of
the cross head in response to static and dynamic loads. Figure 9-9 shows the
graph of cross-head displacement against time from the fatigue test of all
material combinations. A schematic image describing the cross-head
displacement is also shown as an inset within Figure 9-9.
The schematic diagram represents how femoral head migration can be a

contributing factor to the total cross-head displacement measured. The other
main factor that may contribute to the cross-head displacement would be the
elastic deformation of the spigot during compressive loads. The deformation
is expected to be greater for Ti6Al4V spigot being the less stiffer component.
At the beginning of the equilibrating stages, 0.3 kN of static compressive load

is applied. From Figure 9-9 it is observable that the load resulted in cross-
head displacement for all materials combinations except CoCrMo – CoCrMo.
On the other hand, the displacement measured for the Biolox – CoCrMo is
seen to be double those of the combinations with Ti6Al4V spigot – both of
263
which were very similar. Considering that there was no cross-head
displacement measured for the CoCrMo – CoCrMo combination, this suggests
that no measurable elastic deformation of the CoCrMo spigot occurred during
the application of static load. Therefore, the cross-head displacement
measured for Biolox – CoCrMo is most likely entirely due to the femoral head
migration. On the other hand, it is difficult to know how much of the cross-head
displacement for both combinations with Ti6Al4V spigots is due to femoral
head migration however, considering the relatively low static load, it is likely
that elastic deformation would also be negligible for the Ti6Al4V spigot.
A compressive load of 1.8 kN was applied prior to the initiation of dynamic

loading. It was observed that the CoCrMo – CoCrMo combination experienced
the least cross-head displacement while the displacement in Biolox – CoCrMo
is double that of CoCrMo – CoCrMo. The combinations with Ti6Al4V spigots
were both very similar, however, the displacement for CoCrMo – Ti6Al4V was
slightly larger than Biolox – Ti6Al4V. More so, throughout the various dynamic
loading stages, all the material combinations appear to have a similar rate of
increase in cross-head displacement which suggest similar rates of femoral
head migration.
264
Figure 9-9 – Cross-head displacement vs. Time graph. Each data point
for all material combinations during the dynamic stages represents a
point every 1000 seconds of the cross-head position. At the top right
corner, a schematic of the cross-head displacement is shown.
At the end of each test, the force required to separate the femoral head from
the male taper is shown in Figure 9-10. The result shows that the separation
forces for CoCrMo – Ti6Al4V and Biolox – Ti6Al4V were very similar whereas
the Biolox – CoCrMo differs from the CoCrMo – CoCrMo by almost 1 kN. Quite
expectedly, the required taper separation force was highest for the Biolox –
CoCrMo combination which in Figure 9-9, experienced the largest femoral
head migration. Similarly, the CoCrMo – CoCrMo combination which
experienced the least femoral head migration required the least separation
force. Both CoCrMo –Ti6Al4V and Biolox – Ti6Al4V combinations which were
in the middle for femoral head migration, also experienced very similar
separation loads that fall in the middle of the greatest and the least taper
separation forces.
265
Figure 9-10 – Taper separation force for all material combinations
9.4 Surface Analysis
The use of CMM in this chapter was solely to identify the wearing pattern and
taper engagement pattern for both male and female tapers across the four
material combinations. In most cases, between Figure 9-11 to Figure 9-15, a
Customised 3D reconstruction of the tapers was presented. By ‘Customised’
it is meant that, regions with excessive deviation from the conical reference,
either through material gain or material loss appear in bright colour while the
other regions are masked and therefore appear in black. Other exceptions to
this are described in each specific figure. All material combinations analysed
in this study had a positive taper (cone) angle difference. A positive taper
angle difference means that the female taper angle is greater than the male
taper angle therefore, their initial contact occurs at the proximal-end [219].
9.4.1 CoCrMo – CoCrMo
From Figure 9-11a it is observable from the female taper that the taper
engagement is circumferential however, from the male taper in Figure 9-11b,
the wearing pattern can be described as a coup-countercoup pattern as
266
described in reference [216]. This wear pattern is generally an indication of a
toggling motion.
(a) (b)
Figure 9-11 – 3D reconstruction of CoCrMo – CoCrMo taper interface:

a) CoCrMo female taper, b) CoCrMo male taper.
9.4.2 CoCrMo – Ti6Al4V
Figure 9-12a shows that the CoCrMo female taper was engaged on two
opposing sides and from Figure 9-12b, the fretted region on the Ti6Al4V male
taper is seen to be present on two opposing sides. However, the wear pattern
observable shows that fretting was focused at the proximal and near-distal
ends of one side as well as distal and near-proximal ends of the opposing
side. Thus, a toggling (shimmying) motion is also described at this interface.
267
(a) (b)
Figure 9-12 – 3D reconstruction of CoCrMo – Ti6Al4V taper interface: a)

CoCrMo female taper, b) Ti6Al4V male taper.
9.4.3 Biolox – CoCrMo
The pattern identified on the Biolox femoral taper in Figure 9-13a appears
similar to a coup-countercoup pattern. However, the pink colour of the Biolox
head in the image shown in Figure 9-13b helps to physically examine the
female taper. The coup-countercoup pattern is seen to be mainly material
transfer from the metal component unto the ceramic as often observed at the
ceramic – metal taper interface [72, 219].
On the other hand, in Figure 9-13c the unmodified ‘Full’ 3D reconstruction of

the CoCrMo male taper displayed an alternating colour effect. This effect is
an indication that no preferential wearing pattern was observable across the
entire taper surface. The ‘interchanging colour’ which only occurs on grooved
male tapers is a result of the interchanging contact of the measuring stylus
between the troughs and peaks of the grooved taper. This is to suggest that
wearing occurred uniformly across the peaks of the grooved male taper, thus
suggesting a circumferential engagement with the female taper. Therefore,
the wear from any particular region on the taper surface was not significant
enough to create a substantial deviation from the conical reference. This is
contrary to the observations made for the metal – metal combinations.
268
Figure 9-13d shows the attempt to create a ‘Customised’ reconstruction of the
male taper in order to confirm the absence of any preferentially worn region.
As expected, no preferentially fretted or worn area was identifiable. However,
Figure 9-13a gives an indication that a preferential transfer of material exist
hence the displayed coup-countercoup pattern.
(a) (c)
(b) (e)
Figure 9-13 – 3D reconstruction of Biolox – CoCrMo taper interface: a)

Biolox female taper, b) Image of the Biolox female taper, c) Full 3D
reconstruction of CoCrMo male taper d) Customised reconstruction of
CoCrMo male taper.
269
9.4.4 Biolox – Ti6Al4V
A similar observation to the Biolox – CoCrMo ceramic – metal combination is

also seen for the Biolox – Ti6Al4V. Figure 9-14a shows a pattern at the distal
and proximal ends of the Biolox female taper. The pattern is confirmed in
Figure 9-14b as material transferred from the Ti6Al4V. It appears more
prominent at this interface than the Biolox – CoCrMo interface. The same
‘alternating colour’ effect is observed for the Ti6Al4V in Figure 9-14c and
Figure 9-14d shows no evidence of preferential wearing pattern. Similarly, this
is an indication of uniform engagement of the Ti6Al4V male taper into the
ceramic female taper.
(a) (c)
(b) (d)
Figure 9-14 – 3D reconstruction of Biolox – Ti6Al4V taper: a) Biolox

female taper, b) Image of the Biolox taper c) Full 3D reconstruction of
Ti6Al4V male taper, d) Customised reconstruction of Ti6Al4V male taper.
270
9.5 The case of misalignment
A total of three test were conducted for Biolox – Ti6Al4V combination. For one
of the repeats, the fretting current and anodic current was found to be
significantly higher than the other repeats. The differences are most
substantial at the highest frequency of 5 Hz (see Figure 9-15a); the highest
mean load of 2.8 kN (see Figure 9-15b) and the fretting currents peaked at a
magnitude ten-times higher for the cyclic load at ±2.5 kN in Figure 9-15c.
The 3D reconstruction of both male and female taper was analysed. Figure 9-
16a revealed that the Biolox was abraded from the near-distal entrance of the
female taper down towards the proximal end by the Ti6Al4V metal. This could
only mean that the male taper was misaligned with respect to the Biolox
female taper during the application of 2 kN quasi-static load. From Figure 9-
16b it is observable that the male Ti6Al4V taper experienced a toggling motion
during the fatigue loading stages likely as a result of misalignment. A coup-
countercoup wear pattern is identified.
(a)
271
(b)
(C)
Figure 9-15 – Fretting currents of Biolox – Ti6Al4V three repeats, a)

frequency variation, b) mean load variation c) cyclic load variation.
(a) (b)
Figure 9-16 – 3D reconstruction of Biolox – Ti6Al4V taper – the case of

misalignment: a) BIOLOX female taper, b) Ti6Al4V male taper.
272
In this chapter, the fretting current and wear pattern from four material
combinations with realistic modular taper interfaces were assessed under
various loading conditions. For the metal – metal combinations, a large
metallic surface area (from the bearing) and the taper interface was exposed
to the physiological solution during the fatigue loading tests. Thus, a realistic
level of exposure comparable to the in-vivo condition is established. As a
result of this exposure, the peak current was observed to be an order of
magnitude higher than the ceramic – metal combinations.
Assessing the effect of frequency across the four material combinations

revealed that in most cases, higher frequencies do result in slightly higher
fretting currents as observed by Brown et al [116]. However, when the
cumulative charge is considered, it is the case that the higher fretting current
is not proportional to the increase in fretting cycles. Therefore, fatigue loading
5 Hz frequency was seen to undermine the degree of fretting corrosion quite
significantly in most material combinations. For ceramic – metal combinations
specifically, a clear declining trend in the cumulative charge with increase in
frequency was observed.
As for the mean load variation, increasing the mean load without an keeping
a minimal constant cyclic load is synonymous to creating a high contact
pressure stick regime at the fretting interface. Therefore the fretting current
generated from the mean load assessment was relatively minimal. And from
all the material combinations, CoCrMo – CoCrMo generated the most
cumulative charge, especially at the higher loads. Interestingly, for the Biolox
– CoCrMo, the lowest and highest mean loads generated minimal fretting
currents. This suggests that at low mean loads, a large proportion of
contacting areas experienced minimal interfacial micromotion that would
cause oxide abrasion. Whereas at the highest mean load, interfacial
compliance (sticking regime) which also reduces micromotions, limits the
areas in contact at the taper interface that could experience oxide abrasion.
No fretting current was measured for Biolox – Ti6Al4V which is also a result
of minimal interfacial micromotions. However, at the highest mean load,
273
fracture of the protective oxide led to a transient of anodic current that was
immediately restore to zero.
The individual fretting corrosion characteristics of CoCrMo and Ti6Al4V within

a metal – metal and ceramic – metal interface was identified during the cyclic
load variation assessment. The metal – metal and ceramic – metal
combinations that had CoCrMo spigots yielded a higher fretting current than
those with Ti6Al4V (except for the case of misalignment). This result could be
expected from an electrochemical point of view if one considers Ti6Al4V’s
ability to repassivate faster than CoCrMo in a non-creviced environment [143].
However, a contrary argument is that the low stiffness of titanium alloy is
generally known to cause increase in fretting at the modular taper and thus
result in higher fretting corrosion [80, 220]. Therefore, it is worth understanding
examining the mechanical influences that could result in larger fretting
currents generated for both CoCrMo – CoCrMo and Biolox – CoCrMo fretting
contacts.
It was identified that CoCrMo – CoCrMo experienced the least femoral head
migration of the four material combinations. This would comparably limit its
taper strength because migration of femoral head increases the areas in
interference fit within the taper interface [63]. Thus the coup-countercoup wear
pattern observed at the CoCrMo – CoCrMo interface is explained; this is
indicative of a toggling form of micromotion. This explains why fretting currents
were high at the CoCrMo – CoCrMo interface.
On the other hand, Biolox – CoCrMo is the least compliant contact of all four
material combinations (determined from the reduced modulus calculation in
Appendix B). Therefore, femoral head migration was found to be the largest
at this interface as expected. This also meant that the largest taper separation
force was required for disassembly thus indicating that it had the highest taper
strength [63]. A greater femoral head migration means a large surface area of
CoCrMo in direct contact with the Biolox ceramic as evidenced in the 3D
reconstruction of the CoCrMo trunnion. Such a large contacting surface area,
against a much harder counter-part would result in abrasion across a larger
surface area upon large micromotions at higher cyclic loads. Therefore, it is
274
deducible that fretting current was at Biolox – CoCrMo contact as a result of
greater interference fit between the micro-grooves on CoCrMo and Biolox.
The good balance between taper strength and how Ti6Al4V engages the
CoCrMo is suggested to be the reason for its relatively modest fretting current.
Although it is expected that the Biolox – Ti6Al4V contact would generate less
fretting current than CoCrMo – Ti6Al4V, 3D reconstruction of the Ti6Al4V male
taper in Figure 9-14 also reveals an interference fit most Ti6Al4V micro-
grooves hence the currents measured are comparable with those of CoCrMo
– Ti6Al4V. The large anodic currents observed for the Biolox – Ti6Al4V couple
upon the initiation of cyclic loads contributed significantly to the total
cumulative charge measured. This suggests that a significant proportion of
the current emanated at this fretting interface is a result of film restoration at
the interface once depassivation as occured.
The single case of misalignment in this chapter highlighted how deleterious

such cases can be in reality. Unfortunately, misalignment of taper assembly
during surgical procedure is not improbable and may result in early revision of
the implant. The misalignment observed in this study affected the taper
engagement thus creating a coup-countercoup wear pattern at the interface
and as a result, a significantly higher fretting corrosion current was generated.
275
Chapter 10 Overall discussion
10.1 Introduction
Fretting corrosion and tribocorrosion mechanisms of metal – metal and

ceramic – metal material combinations have been assessed using both
tribometer point-contact configuration and fatigue loading of realistic taper
component. Mechanical and electrochemical data were complimented with in-
depth surface and sub-surface analyses. Each result chapter in this thesis
were summarised with a brief discussion and summary section in order to
consolidate the findings from the various material combinations and
parameters tested within the chapter. The purpose of this discussion chapter
is therefore to discuss the overarching themes and proposed mechanisms.
This discussion is split into the following key sections:
 Linking fretting loops with the fretting interface

 Aspects of fretting and corrosion mechanisms
 Characteristics of biomedical materials at the modular taper junction of
hip implants
 Relevance to clinical application
10.2 Linking fretting loops with the fretting interface
The use of fretting loops to monitor wear regime transformation during fretting
tribometer studies is well established in the fretting field. Such fretting loops
were presented for all displacement amplitudes and material combinations
assessed in this research. The vast number of fretting experiments conducted
in this research has made it possible to interpret subtle deviations in the shape
of a typical gross slip fretting loop. This deviation is thereby correlated to in-
situ wear mechanisms. A standard gross slip regime shape has a
parallelogram-like shape as depicted in Figure 10-1a. The loop shown was
obtained from the gross slip fretting loop of self-mated CoCrMo at ±25 µm
displacement. The fretting loop is labelled and correlated to the de-convoluted
276
graphs of displacement (solid line) and tangential force (dashed line) in Figure
10-1b. The points begin at A1 and is described in the clockwise direction to
end at A2.
At A1, the tangential force is relatively constant as the ball displaces towards
the first corner (C1) of the fretting contact limit. At the side ‘B’ of the fretting
loop, the tangential force transitions from kinetic friction to zero while the
cantilever of the tribometer experience few micro-meter of displacement which
corresponds to the ‘system compliance’. The tangential force increases from
zero at the same rate in the opposite direction (negative y-axis) while still at
C1. This stage represents the static increase in tangential forces (static
friction) until a peak value is reached – the value corresponds to the transition
force require for the contact to overcome static friction and kinetic friction is
introduced. The point of transition is denoted ‘corner 1 reverse’ (C1R).
At side C, the ball displaces from C1 to the second corner (C2). Subsequently,
stage B is repeated at C2 labelled ‘D’ and once again, the tangential force
generates enough traction to translate into kinetic friction thus the ball slips in
the direction of C1 again; this is denoted A2.
There are several interfacial mechanisms at the fretting interface arising from
the accumulation of third-body products such as surface ploughing, micro-
cutting/grooving, film formation, interfacial layers and so on. Some of these
can be identified through the deviations observed in fretting loops. The
deviations often appear at specific parts of the fretting loop like the sides A1 &
2 and C as well as corners C1 and C2. Some of these mechanisms are outlined
in the following subsections:
277
(a) (b)
Figure 10-1 – Fretting loop of gross slip regime; a) labelled fretting loop
b) de-convoluted fretting loop – displacement (solid-line) and
tangential force (dashed-line).
10.2.1 Fretting loop deviations at the self-mated CoCrMo interface
The deviation shown in Figure 10-2a was only observed at the self-mated
CoCrMo fretting contacts (Figure 6-3a and Figure 8-4a). The red dashed-lines
represents the expected gross-slip shape for a symmetrical fretting interface.
From the fretting loop shape (plotted in black), it can be seen that the
tangential force at the positive x-axis of the fretting loop decreases gradually
towards the fretting contact limit i.e. C1 and increases as it approaches C2.
This is a result of differences in wear depth about the centre-line of the fretting
contact as illustrated in the 3D surface profile of the flat component in Figure
10-2b. The cause of this phenomenon is attributed to possible irregular
pressure distribution at the interface due to the presence of third-body
products. This is a common process observed at fretting contacts [39, 47].
The fretting loops corresponding to the early cycles of fretting for this interface
were verified in order to ensure this phenomenon is not a result of contact
misalignment.
278
(a) (b)
Figure 10-2 – Non-symmetrical fretting contact a) CoCrMo – CoCrMo

±50 µm fretting loop b) ball and flat wear surface
Other interesting deviations corresponding to third-body build-up at self-mated

CoCrMo interface is presented in Figure 10-3. The behaviour was typically
observed at ±150 µm both in OCP and potentio-static conditions. In this
specific case, maximum tangential force in Figure 10-3a is seen at the centre
of the contact which corresponds to the location of a redistributed third-body
product at the centre of the moving ball in Figure 10-3b. Correspondingly, the
largest micro-cutting/grooved region of the flat component is observed at the
centre of the wear surface.
279
C1 C2
(a) (b)
Figure 10-3 – Third-body grooving a) CoCrMo – CoCrMo, ±150 µm

fretting loop b) ball and flat wear surface
10.2.2 Fretting loop deviations at ceramic – metal interface
The fretting loops shown in Figure 10-4a and Figure 10-4b includes a feature
indicated with the green arrows. This feature corresponds to a quasi-static rise
in tangential force during gross slip fretting. Specifically in Figure 10-4a,
another deviation from the standard gross slip shape can be seen. The
tangential force during kinetic friction is observed with a curved profile. The
two figures represent Si3N4 – CoCrMo and Biolox – Ti6Al4V contacts at ±50
µm respectively.
Figure 10-4c shows the surface profile for both Si3N4 ball and CoCrMo flat.
The agglomerated third-body product on the CoCrMo component is seen at
the centre of the fretting contact post-test. However, the fretting loop describes
a non-symmetrical fretting behaviour at the interface. As the Si 3N4 ball
translates towards C1 the third-body product builds-up near the centre of the
wear contact and thus obstruct relative motion momentarily – this is the cause
280
for the quasi-static rise in tangential force. However, as the ball translates
back from C1 to C2, the quasi-static rise is not present because, the third-body
product is redistributed across the CoCrMo surface thus abrading the Si 3N4
ball rather than obstructing its relative motion. The curved tangential force
profile during kinetic friction corresponds to the wear surface of the CoCrMo
flat where the centre of the wear surface is the lowest point during fretting.
The same quasi-static feature is observed in Figure 10-4b. However, for this
contact, the quasi-static rise in tangential force were also present during the
return from C1 to C2 and twice in this case. It has already been established
through 3D surface profilometry and BS-SEM (Figure 7-36 and Figure 7-37)
that the BIOLOX ball experienced no measurable material loss. Rather, wear
and corrosion products of Ti6Al4V alloy which became trapped at the interface
formed a smooth-appearing third-body material on one side (the side in
contact with the Biolox – Figure 10-4d) and on the flip side, the third-body is
in direct contact with Ti6Al4V. The wear track identified beneath the third-body
product is most likely a result of a galling mechanism which is typical of
adhesive contacts and has been well reported for Ti6Al4V – Ti6Al4V contact
[208-210] In this contact, the quasi-static rise in tangential force is likely an
indication of adhesion at the third-body – Ti6Al4V interface. The feature was
also identified at the Si3N4 – Ti6Al4V interface in Figure 7-5a.
(a) (b)
281
(c) (d)
Figure 10-4 – Influence of third-body a) Si3N4 – CoCrMo, ±50 µm fretting

loop b) BIOLOX – Ti6Al4V, ±50 µm fretting loop c) Si3N4 – CoCrMo ball
and flat wear surface d) BIOLOX – Ti6Al4V ball and flat wear surface.
10.2.3 Fretting loop deviations at CoCrMo – Ti6Al4V interface
Figure 10-5a shows the fretting loop of CoCrMo – Ti6Al4V at ±50 µm

displacement. For this interface, the quasi-static rise in tangential force was
observed at the fretting contact limits (C1 and C2). This is indicative of some
degree of deformation was accommodated at the Ti6Al4V contact edges
resulting in the quasi – static rise in tangential force. The phenomenon is
commonly observed for Ti6Al4V as shown in the fretting loop of other authors
[99, 221]. However, it was termed ‘The plunging effect’ in the study by Fouvry
et al [221], Figure 10-5b shows a schematic diagram of the feature obtained
from their study.
282
(a) (b)
Figure 10-5 – The plunging effect (prior to fracture of cold-welded

interface): a) CoCrMo – Ti6Al4V, ±50 µm fretting loop, b) schematics of
the plunging effect [221].
Other interesting deviations observed for the CoCrMo – Ti6Al4V at ±50 µm

displacement is the fretting loop of the ruptured welded interface during gross
slip regime. The fretting loop shown in Figure 10-6a corresponds to the gross
slip fretting behaviour subsequent to the fracture of the cold-welded interface.
The 3D surface profile in Figure 10-6b shows the interface with well aligned
surface cracks and peaks as well as the contact boundary. The main deviation
of this fretting loop shape is observable at corners C1R and C2R in which case,
a curved tangential force profile is observed. The non-definitive transition into
kinetic friction and the non-linear increase in tangential force may be indicative
of surface adhesion between multiple intertwined grooves established in the
ruptured interface. That is, relative slip is no longer just occurring between a
point contact, rather, a frictional contact is also established along the sides of
the fragmented peaks of the CoCrMo and the surface cracks in the Ti6Al4V.
283
(a) (b)
Figure 10-6 – Slip within localised fretting grooves (after the fracture of
cold-welded interface): a) CoCrMo – Ti6Al4V, ±50 µm fretting loop, b)
3D surface of the ruptured welded interface.
10.2.4 Summary
Deviations in the fretting loop shapes relative to the standard gross slip shape
were correlated to their corresponding in-situ interfacial mechanisms. Through
the identification of these characteristic contact features, in-situ wear
mechanisms and changes to the contact conditions can be detected in the
absence of any in-situ surface analytical techniques. It serves as a preliminary
stage to characterising in-situ wear in a tribological system.
10.3 Aspects of fretting and corrosion mechanisms
10.3.1 Mechanism of cold-welding and pseudo-amorphous

structures at CoCrMo – Ti6Al4V interface
Cold-welding is a solid-state welding process whereby an interfaces adheres

together (under high contact pressures) without the use of heat. The cold-
welding process involves the oxidation of the two materials in contact. Several
cases of cold-welding in modular implants have been reported in literature [80,
284
84, 222-225]. In all the cases, cold-welding was only prevalent at the self-
mated Ti alloy interface. The welding occurs subsequent to thick Ti-oxide
build-up resulting from MACC. Consequentially, the welded components are
inseparable by the surgeon during a revision procedure, leading to the whole
modular stem having to be removed.
To the best of our knowledge, no retrieval study or in-vitro study has yet
reported the observation of cold-welding at the CoCrMo – Ti6Al4V interface.
There are two main reasons that may explain this: either that cold-welding at
this interface doesn’t occur in-vivo (an option deemed unlikely) or it does
occur, but it is not as resistant to separation force as the thick, cold-weld
formed at the self-mated Ti interface. The latter is considered the more likely
reason because while in the present study, the cold-weld (pseudo-amorphous
material) observed at the CoCrMo – Ti6Al4V interface had a maximum
thickness of 2 µm, the study by Grupp et al [80] reported cold-weld with
thickness up to 30 µm at the self-mated Ti interface.
A very recent study by Hall et al [199] described evidence of material transfer

at the CoCrMo – Ti6Al4V interface with “related surface break-outs” that is,
the two metallic materials were once bonded together. This strengthens the
argument that cold-welds do form in-vivo of CoCrMo – Ti6Al4V interface but
may be susceptible to fracture whilst still assembled or may fracture during
surgical revision as a result of low thickness or brittleness. A schematic of the
proposed in-situ mechanism of how the ‘cold-weld’ with a pseudo-amorphous
structure is formed at the CoCrMo – Ti6Al4V interface is shown in Figure 10-
7. The 3D fretting loop confirms the point where a regime transition from
partial-slip to gross slip occurred. It is believed that this transition stage
corresponds to point for which the pseudo-amorphous material ruptured,
leaving most of the material still ‘bonded’ to the more ductile of the two alloys
– Ti6Al4V.
A proposed mechanism to explain how the pseudo-amorphous structure

forms from an initially crystalline structure of the two contributing alloys is
shown in Figure 10-8. The mechanism also address the processes involved
in the formation of amorphised Ti6Al4V. The interfacial and subsurface
285
transformations observed from both CoCrMo and Ti6Al4V alloys are
categorised into four stages. In this mechanism, ‘Stage 1’ represents the
deepest subsurface of the alloy and ‘Stage 4’ represents the uppermost
subsurface/interface of the alloy.
In stage 1 – the grain sizes of the base alloy are generally few microns in size
for both CoCrMo and Ti6Al4V. For stage 2 – smaller grain sizes in the sub-
micron range begin to emerge. The low stacking fault energy of CoCrMo leads
to strain-induced twinning hence the recrystallization of CoCrMo alloy at this
level [6]. Whereas in Ti6Al4V, crystallographic reorientation becomes evident
through the presence of directionality (see Figure 6-19d). So also, smaller
grain sizes are caused by recrystallization from internal dislocations initiated
by plastic straining [226]. In stage 3 – grains are few nano-meters in size and
are generally observed near the upper most surface. These nano-crystalline
structures in non-lubricated fretting contacts of metallic alloys are known as
tribological transformed structures (TTS) [226]. It has also been shown that
TTS structures can be multiple times harder than the bulk alloy itself; they are
brittle and susceptible to fracture [226].
The CoCrMo and Ti6Al4V interface in this study is both lubricated and
subjected to a partial-slip regime prior to fracture of the pseudo-amorphous
material, thus ‘Stage 4’ exist at this interface. Under normal gross slip
conditions, the transformation of the alloy would reach ‘Stage 3’ and the
material is subsequently removed from the interface as wear and oxidation
product. However, the high contact compliance of the Ti6Al4V alloy relative to
the CoCrMo facilitates the detainment of third-body products. And thus, in the
presence of physiological solution, mechanical mixing and oxidation
processes lead to the formation of ‘Stage 4’ which is the pseudo-amorphous
material (or the cold-weld).
The transformation of bulk crystalline Ti6Al4V into an amorphous structure

also follows a similar process to the formation of the pseudo-amorphous
structure. However, as shown in Figure 10-8, it is proposed that the stages at
which each structure formed differ. Although it cannot be fully verified, it is
proposed that subsequent to the rupture of the pseudo-amorphous structure,
286
gross slip fretting is initiated at the interface of CoCrMo and the pseudo-
amorphous structure. Interfacial shear stresses from the interface transmitted
through the pseudo-amorphous structure effects subsurface recrystallization
in the underlying Ti6Al4V. As the thickness of the pseudo-amorphous
structure is not uniform, subsurface mechanical mixing with access to the
physiological solution results result in the formation of amorphous Ti6Al4V.
287
Figure 10-7 – Mechanism of cold-welding at the CoCrMo – Ti6Al4V
interface
288
Figure 10-8 – Mechanism of pseudo-amorphous structure of CoCrMo
and amorphisation of Ti6Al4V alloys.
289
10.3.2 Mechanism of stiction at the Si3N4 – Ti6Al4V interface
The word ‘stiction’ originates from two words ‘static’ and ‘friction’. It is more
commonly used to describe the process in diesel engines where static friction
is created as a result of build-up of hydrocarbons and is needed to be
overcome. The word is used in this context to describe the intermittent
alternations between static friction (stick regime) and kinetic friction (gross
slip) at the Si3N4 – Ti6Al4V interface.
The mechanism observed at this contact have some similarities with those of
CoCrMo – Ti6Al4V interface in the previous section (10.3.1). The two main
similarities are: formation of a pseudo-amorphous structures and formation of
amorphous Ti6Al4V structure. However, there are several significant
differences between the two material combinations. For examples, it had been
discussed previously that one of the dominant wear mechanisms for Si3N4
when in contact with Ti6Al4V is fracture-induced wear. The energy dissipated
graph (see Figure 7-5) reveal that a significant ‘burst’ of energy is measured
during the intermittent transitions from stick regime to the gross slip regime. It
is suggested that this sudden regime translation is always as a consequence
of micro-fracture of Si3N4 at the interface which subsequently leads to
interfacial slip. As a consequence of such changes in energy, significant
subsurface damage is inflicted to the Ti6Al4V alloy beyond those observed in
the Ti6Al4V of the CoCrMo – Ti6Al4V couple. Figure 10-9 shows a schematics
diagram of the proposed stiction mechanism. While in the previous section,
four stages of Ti6Al4V was outlined (see Figure 10-8), an additional
degradation stage (Stage 5) was introduced for Ti6Al4V in this figure.
In a similar manner to the CoCrMo – Ti6Al4V interface, both Si3N4 and Ti6Al4V
contributed to the pseudo-amorphous material. A fusion of both Si-oxide, Ti-
oxides and Al-oxide under a stick fretting regime is the proposed mechanism
leading to the formation of the pseudo-amorphous structure. TEM-EDS and
XPS was used to identify the material as possibly Ti and Al silicates.
The energy involved at this interface is such excessive that the regions that
are not covered by the pseudo-amorphous material (as shown in phase ‘D’ of
290
Figure 10-9) also experienced a full (Stage 1 – 4) subsurface transformation.
On the other hand, the regions that were covered by the pseudo-amorphous
material (as shown in phase ‘E’ of Figure 10-9) experienced the newly
introduced ‘Stage 5’ transformation. Stage 5 transformation in Ti6Al4V occurs
likely during the sudden transition process from stick regime to gross slip.
In summary, it can be deduced that the stiction mechanism occurs when the
Ti6Al4V becomes more compliant due to subsurface fracture (stage 5) and
migration of the Ti and Al silicate pseudo-amorphous material. This
compliance leads the fretting contact back into a stick regime. During the stick
regime, micro-fractures of Si3N4 at the interface piles up and oxidises to form
a new layer of the pseudo-amorphous material. Once enough traction force is
gained at the interface, the contact is translated back to a gross slip regime
and the process is repeated.
291
Figure 10-9 – Mechanism of stiction at the Si3N4 – Ti6Al4V interface
292
10.4 Characteristics of biomedical materials at the modular
taper junction of hip implants
This section brings together the key findings from both tribometer and realistic
taper component assessment. It further highlights the characteristic role of
each biomedical material at the fretting interface of the modular taper junction.
10.4.1 CoCrMo at the modular taper interface
In Chapter 5 the role of Mo in the passivation process of a polished CoCrMo

surface was highlighted. This was further emphasised in Chapter 6 (see
Figure 6-21 and Figure 6-22) where a thick passive oxide is formed above a
Co-Mo structures. In the region absent of the Co-Mo structure, the passive
oxide did not form thus highlighting the role of Mo in enhancing surface
passivation. Evidence of a Cr-depleted region is generally associated with
grain-boundary corrosion or sensitization as its better known in literature [174,
227]. However, the observation from this research found that in the case of
micro-cutting or grooving, active wear particles can be generated thus the
wear product is not only Cr-oxide and phosphate species. The ‘honey-comb’
Co-Mo structure was formed as a result of electrochemical process involving
preferential depletion of Cr to form oxides. Subsequent fracture of the ‘honey-
comb’ fracture may result in release of tiny Co-rich particles thus increasing
the risk of Co-induced toxic effects and other forms of ALTR [93].
Self-mated CoCrMo alloy subjected to fatigue loading had the least taper
strength as it was observed to experience the least taper migration. The
consequence of this is increase in toggling micromotion which may lead to
accelerated fretting corrosion (especially at high cyclic loads). It is such
toggling motions that manifest in ploughing or grooving of the modular taper
as observed in retrieval analyses [199]. From monitoring the fretting current,
it was often the case that the fretting current at the self-mated CoCrMo
interface drops after a few thousand cycles. This is attributed to possible
electrochemically induced film formation or agglomeration of third-body
293
products at the interface thus creating a physical barrier against charge
transfer [215]. In order words, an antagonistic tribocorrosion reaction.
From Figure 10-10, the presence of a central partial-stick region is indicative

of an ‘unexposed’ region in the gross slip fretting contact. This region is
typically poorly aerated hence a type of crevice. The ‘unexposed’ region in a
ball-on-flat configuration is representative of the crevice established at the
proximal-end of the modular taper. On the other hand, the slip annulus in the
gross slip ball-on-flat configuration represents a well aerated environment and
thus representative of the distal-end of the modular taper. As such, the
composition of the third-body wear and corrosion products formed in this
research are the same with those reported in previous retrieval analyses [85].
Figure 10-10 – Schematic diagram of different fretting corrosion

products at specific regions within the composite contact of self-mated
CoCrMo, ±50 µm.
The coupling of CoCrMo against Ti6Al4V was observed to have led to the
establishment of severe crevice conditions. This is evidenced by pitting
corrosion of the CoCrMo; the observation of pitting suggests that the pH of the
crevice depicted in Figure 6-31 may be lower than 3.5 – the point where the
passive Cr2O3 layer becomes unstable according to Pourbaix diagram [228].
294
10.4.2 Ti6Al4V at the modular taper interface
From retrieval studies, it is common that the mixed-metal combination of

CoCrMo – Ti6Al4V experienced greater corrosion damage than the self-mated
CoCrMo material combination [27, 67]. While there are several factors
contributing to this, it is believed that long-term accumulation of third-body
wear and corrosion product at the mixed-metal interface is a key contributing
factor. Adhesive wear mechanism of Ti6Al4V at the mixed-metal interface was
seen in this study to facilitate a significant volume of redistributed wear at the
interface unlike self-mated CoCrMo. From what is known about MACC,
significant agglomeration of wear and corrosion product can rapidly change to
the local chemistry at the interface which would inevitably accelerate corrosion
in both CoCrMo and Ti6Al4V alloy [27].
Titanium alloys have been the most reported material to undergo in-vivo
fracture all the metallic biomaterials used in THA [54]. Most fractures
reportedly occur at the neck-stem interface largely because of the greater
moment arm relative to the to the head-neck interface [82]. Fretting and
corrosion mechanisms depending on the contact conditions are also major
causes of failure and fracture in-vivo. The level of metallurgical
recrystallization, subsurface oxidation and crevice corrosion mechanisms
experienced by Ti6Al4V in this study further emphasises the greater risk of
Ti6Al4V failure relative to the other materials. To a large extent, several in-
vivo failure mechanisms, be it mechanical, corrosion or tribocorrosion have
been successfully replicated through the use of a tribometer in this research.
10.4.3 Si3N4 at the modular taper interface
Si3N4 ceramic was not assessed to the realistic taper component level in this
research. However, there are on-going research efforts (led by Amedica
Corporation) to develop a bulk Si3N4 femoral head component for application
in THA. In addition, an EU funded research project – Life Long Joint is also
actively developing a silicon nitride (SiNx) coating for applications in THA. This
is to say that, the following material combinations at the modular taper may
295
be commercially available in the near furture: Si3N4 – CoCrMo, Si3N4 –
Ti6Al4V, CoCrMo – SiNx etc. In this vein, there are a few characteristic of
Si3N4, observed in this study that should be considered when it is coupled
against CoCrMo and Ti6Al4V in a fretting contact:
1. Si3N4 forms a tribochemical material (SiOx -OHy) against CoCrMo alloy

2. Oxidation processes that forms Cr2O3 is suppressed by the reactivity
of Si in Si3N4 (see Table 10-1).
3. Si3N4 forms a tribochemical material with TiAl4V under sticking and
slipping (stiction) contact condition. The ‘surge’ of energy involved in
this process may only ever be experienced at the modular taper
interface during a stumbling load where peak loads can rise up to
eleven-times the body weight within a single cycle [122].
4. Si3N4 when subjected to a partial-slip regime (as commonly realised
within the modular taper) or when rubbing against its own oxide, may
wear significantly more relative to CoCrMo and Ti6Al4V alloys.
5. Si3N4 experiences a fracture-induce wear mechanism against Ti6Al4V
during a stiction behaviour.
Self-mated Si3N4 is well known for its ultra-low friction (< 0.002) due to
tribochemical film that it forms in water-based lubricants [229]. However, the
tribochemical material formed against the CoCrMo and Ti6Al4V alloys in this
study did not function as a solid lubricant rather, the opposite was the case,
thus agreeing with observations of another study [17]. Table 10-1 shows the
corresponding electrode potentials for the reduction reactions of several
element involved with Si at the tribological interface. The more negative the
electrode potential of a reaction is, the higher the tendency for the reaction to
occur. It can be seen that Si has a higher tendency to oxidise than Cr whereas
most species of Ti and Al have a significantly higher tendency to react than
Si. This may explain why there is a preferential formation of SiOx -OHy instead
of Cr2O3. Considering the in-vivo toxicity of Cr, the suppression of Cr3+
formation may prove to be a positive attribute of a Si3N4 – CoCrMo
combination.
296
Table 10-1 – Electrode potentials of Si, Cr, Ti and Al vs. SHE [2, 21,
230]. The electrons transferred in the half-cell reaction are omitted from
the left column for illustrative purposes.
Reduction reactions Electrode potential ‘Eo’ vs. SHE

Si ⇌ Si4+ -0.91
Cr ⇌ Cr3+ -0.74
Ti ⇌ Ti2+, Ti3+, Ti4+ -1.63, -1.23, -0.88
Al ⇌ Al3+ -1.66
10.4.4 Biolox at the modular taper interface
It is a reasonable hypothesis that by replacing the metallic head component

of hip prosthesis with a ceramic (an inert material) alternative, the fretting
current at the fretting interface of the modular taper will be halved. Several
recent studies conclude that metal corrosion and wear when coupled against
ceramic is less or not different than in metal – metal combinations [72,73, 201].
However, when comparing the fretting currents measured during the cyclic
loading from the metal – metal and ceramic – metal combination in this study,
it is observable that the ceramic made no difference in reducing fretting
corrosion current. Rather, the current was highest in the Biolox – CoCrMo
combination thus highlighting that although the ceramic remains inert, fretting
corrosion is influenced by how the metallic component engages it.
Other interesting characteristics of the Biolox is that wear and corrosion

products are typically attached to it at the taper further preventing it from
abrasive wear. The accumulation of these interfacial products may also act as
a protective barrier against charge transfer thus reducing fretting corrosion
current in the long-term. This may also modify the contact condition at the
interface in a manner that reduces wear. Nevertheless, long-term
accumulation of third-body products at the ceramic – metal interface may
activate localised corrosion dominated mechanisms the kind of which current
techniques such as volume loss measurements or the semi-quantitative
Goldberg fretting score might not detect.
297
10.5 Relevance to clinical application
The purpose of this section is not to exalt one material or material combination
while condemning the other out-rightly; this study is not considered exhaustive
enough to take such a stand. However, the stance taken in this study is that,
every material and material combination has an advantage it brings to the
prostheses used in THA. Having said this, this study hopes to enlighten the
orthopaedic community i.e. the clinical practitioners, manufacturers,
engineers and regulatory bodies about the possible failure modes and
damage mechanisms that each material and material combination is
susceptible to. Across all the in-vitro assessments conducted in this research,
the findings reveal that a heavy and active patient poses the greatest risks to
implant failure. And in addition, extreme loading conditions like stumbling and
surgical errors such as misalignments and component mismatch equally pose
a great risk to implant failure.
Micro-cutting or grooving is one of the wearing mechanism identified at the

metal – metal interface of a modular taper junction [199]. The mechanism is
expected for a modular taper contact where the trunnion is micro-grooved and
the contact pressures are disproportionately distributed across the taper,
leading to contact pressures bring concentrated at a local region. From the
findings of this research, high resolution imaging analysis of wear generated
through grooving at a self-mated CoCrMo interface showed evidence of
electrochemical activity in the wear product. The implications of such active
CoCrMo particles in comparison to oxidised particles i.e. oxides and
phosphate species retrieved from the peri-prosthetic tissue are not fully
known. Perhaps, it may enhance the formation of complex ionic species and
adverse local tissue reactions. On the other hand, for titanium alloys, the
adhesive wear mechanism and electrochemical activity of its particles makes
it less likely to be released into the tissue in a highly electrochemically active
form. Except in the case of implant fracture or severe wearing at the distal-
end of the modular taper junction where large volume of particles are expected
to be released.
298
The evidence provided in this study also shows the possible implication of
applying new metallic sleeves onto worn and corroded trunnions. Although
only the self-mated CoCrMo combination was assessed, it showed how the
damage mode may be accelerated to a corrosion dominated mechanism
when fretting occurs at an interface that has a history of wear and corrosion.
In addition, the possibility of an antagonistic tribocorrosion interaction where
interfacial friction and corrosion currents are reduce for self-mated CoCrMo
combination suggests it CoCrMo sleeve may be a better option when coupling
it against a CoCrMo trunnion. One of the primary aim when using a sleeve is
to ensure interference fit at the sleeve – trunnion interface thus preventing any
secondary fretting. However, the findings in this research reveal that if a
crevice is established at the same interface for a mixed-metal material
combination, the contact can rapidly become a severe crevice so much so that
repassivation is inhibited thus activating pitting corrosion of the CoCrMo and
possibly causing the release Cr6+. These mechanisms were identified at the
fretting interface of CoCrMo – Ti6Al4V in this research and the fretting test
lasting just under two hours. Therefore, one can imagine such condition in-
vivo where the implant is fitted for the longer term.
The selection of ceramic coupled with CoCrMo and Ti6Al4V ought to be done
with caution when considering a heavy and active patient. From the fretting
corrosion assessment of the ceramic – metal couples in this study, the
combination of high mean-load (heavy patient) and large cyclic loads (active
patient) resulted in greater fretting currents than the metal – metal couples in
some cases. This was a counter-intuitive observation and it was attributed to
the fact that a larger metallic surface area of the grooved metallic trunnion is
in contact with a much harder ceramic counter-part. Therefore, upon
application of large cyclic loads as expected for a heavy and active patient,
fretting occurs across a large surface area thus generating a relative higher
fretting corrosion current.
On another note, the use of titanium alloy trunnion with ceramic head for an
average patient may benefit from the surgeon applying a high assembly loads.
This should ensure better load distribution across the taper, it is important
because, the excellent toughness and wear resistance of ceramic means that
299
energy at the ceramic – metal interface is not dissipated through the wearing
of both ceramic and metal components. Rather, most of the loads and
interfacial shearing at the ceramic – metal interface is absorbed by the less
stiff titanium alloy. This makes the titanium alloy more susceptible to high
levels of metallurgical recrystallization and risk of crack propagation if load is
inadequately distributed.
Ultimately, the application of a durable, corrosion resistant and biocompatible

coating at the trunnion may prove highly effective in the mitigation of fretting
corrosion of the modular taper junction. Such a coating should serve to
increase taper strength while reducing interfacial fretting. And in the case
where fretting occurs, the coating should be wear resistant to minimise wear
and stiff enough to resist subsurface recrystallization of the underlying metal.
And if the coating experiences wear and oxidation, the coating and its product
ought to be soluble in physiological media.
300
Chapter 11 Conclusions and Future work
The thesis has explored the fretting interface of various relevant biomedical
material combinations. Numerous findings and novel research themes have
been established. Listed below are some key findings which are split into
subsections:
11.1 Conclusions
11.1.1 Material characterisation, metal – metal and ceramic –

metal systems (Chapters 5 – 7)
 Metallic species contribute to the protective layers formed on CoCrMo

and Ti6Al4V alloys. Cr and Mo were enriched in the transition area of
CoCrMo while Ti and V was enriched in the transition area of Ti6Al4V.
 Mo was seen to affect the restoration of a passive oxide layer on

CoCrMo worn surface.
 Contact compliance of the material combinations within metal – metal

and ceramic – metal systems largely influenced the corresponding
fretting regimes realised at the interface. All combinations involving
Ti6Al4V experienced greater contact compliance.
 Cathodic shift in both metal – metal and ceramic – metal systems were
influenced by the fretting regime. Cathodic shift was minimal for both
systems during a stick/partial-slip regime while at gross slip regime, the
magnitude of cathodic shift was multiple times higher for contacts
involving Ti6Al4V than those without Ti6Al4V.
 The wear mechanism at the CoCrMo – CoCrMo contact becomes

increasingly third-body influenced at the gross slip regime. On the other
hand, significant proportion of the volume generated at the CoCrMo –
Ti6Al4V interface are redistributed within the interface due to an
adhesive wear mechanism.
301
 A cold-weld is formed at the interface of both mixed regime contacts of
CoCrMo – Ti6Al4V and Si3N4 – Ti6Al4V. The hardness of the welds
were 25.6 GPa and 22.7 GPa respectively. This corresponds to
interfacial materials with hardness 3x of CoCrMo and 6x of Ti6Al4V in
the metal – metal systems; almost 2x the hardness of Si3N4 and 5x
Ti6Al4V in the ceramic – metal system.
 Subsurface degradation of CoCrMo in a CoCrMo – CoCrMo interface

was characterised with: formation of nano-crystalline structures during
partial-slip regime; loss of nano-crystalline structures and increased
severity of strain-induced twinning during gross slip regime; reduced
severity of twinning with influence of interfacial tribochemical
interactions at the transition to reciprocating sliding.
 Subsurface degradation of Ti6Al4V in a CoCrMo – Ti6Al4V interface

was characterised by strain-induced subsurface transformations
including directionality and recrystallization at both stick and partial-slip
regimes. At the gross-slip regimes, mechanically mixed oxides with
nano-crystalline structures. Amorphization of Ti6Al4V was also
observed.
 Cold-welds in both metal – metal and ceramic – metal systems had a

pseudo-amorphous structure and their main compostion was mixed
oxides and chlorides where mainly particles of Ti6Al4V were
suspending within the amorphous structure.
 At both contacts with cold-weld, amorphization of Ti6Al4V occurred

beneath the pseudo-amorphous material.
 The types of corrosion products formed at the interface of the metal –

metal system and the Biolox – metal contacts was dependent on the
contact conditions. In most cases, metal – oxides/chlorides were
formed at the creviced contacts. However, non-creviced regions were
characterised with Phosphates and various forms of metal – oxides.
 Cr6+ specie were detected at the severest region of the CoCrMo –

Ti6Al4V contact.
302
 Si3N4 was engaged both in wear and oxidation at the ceramic – metal
interface whereas the Biolox ceramic was inert in both cases.
 Si3N4 suppressed the formation of Cr3+ at the Si3N4 – CoCrMo interface

by forming a SiOx -OHy material whereas, at the Si3N4 – Ti6Al4V
interface, the pseudo-amorphous material composed of Ti and Al
silicates.
11.1.2 Quantification of fretting corrosion current on self-mated

CoCrMo (Chapter 8)
 The contribution of chemical oxidation to total mass loss was much

higher than the mechanical contributions for the fretting displacements
assessed.
 Fretting of individual displacements was found to be in corrosion

enhanced wear degradation regime. However, the higher the fretting
displacement, the more the degradation mechanism transitions to a
wear enhanced corrosion mechanism.
 The unknown contribution of corrosion enhanced wear in the fretting

current transient could make up a relatively large proportion of fretting
current transients during fatigue loading in comparison to the wear
enhanced corrosion.
 Effect of surface history when fretting displacement occurs in an

incremental order of displacement didn’t change the degradation
mechanism from corrosion enhanced wear dominated. However, at the
transition to reciprocating sliding, tribochemical interactions at the
interface led to a significant drop in the interfacial energy and fretting
current.
 Effect of surface history when fretting displacement occurs in a

decreasing order was found in a corrosion dominated degradation
mechanism and the absence of elastic compliance by the least
displacement resulted to a transition from partial-slip to gross slip and
thus increases fretting current.
303
11.1.3 Component testing of metal – metal and ceramic – metal
tapers (Chapter 9)
The term ‘fretting current’ is used in this summary rather than the cumulative
charge as the cumulative charge was the quantitative means through which
the fretting current was averaged.
 The use of 5 Hz frequency in component testing undermines the

degree of fretting currents for both metal – metal and ceramic – metal
systems except in CoCrMo – CoCrMo which showed no significant
difference across 1Hz, 2Hz and 5Hz.
 The magnitude of cyclic load had a greater effect on the evolution of

fretting current than the increase in mean load for both metal – metal
and ceramic – metal systems.
 CoCrMo – CoCrMo was most susceptible to fretting of all four

combinations during the mean load assessment. While CoCrMo –
Ti6Al4V experienced the most systemic effect after the mean load
assessment.
 In general, the Biolox – Ti6Al4V combination generated the least

fretting current apart from the cyclic load variation where very similar
values were measured with that of CoCrMo – Ti6Al4V. This was found
to be a result of larger interference fit at the Biolox – Ti6Al4V contact.
 The single case of misalignment in the Biolox – Ti6Al4V combination

resulted in a significant increase in fretting current. Specifically at the 5
Hz frequency, the largest mean load and the largest cyclic load.
 There was a strong correlation between the extent of femoral head

migration and taper separation force.
 The metal – metal combinations both experienced a toggling type of

motion while the taper strength which is linked to the contacting area
at the taper interface was highest for the ceramic – metal combinations.
304
11.2 Future work
Characterising the surface chemistry of tribofilms during/after fretting –

this begins by characterising a surface that has undergone the process of film
formation resulting from fretting a contact that has a history of surface wear
and corrosion. XPS analysis of this film can thereby be compared to the
chemistry of the films identified by other studies in literature.
Quantify the hardness of the third-body product formed at the Biolox –

Ti6Al4V fretting interface – this can be done by replicating the smooth third-
body product formed at the Biolox – Ti6Al4V surface and subsequently
performing nano-indentation hardness to quantify its hardness.
Employ an in-situ technique to observe the formation of amorphous

Ti6Al4V – this may be attempted by in-situ imaging and characterisation of a
Ti6Al4V alloy rubbing against a transparent counter-body.
Perform a component level assessment of bulk Si3N4 – the purpose of this

is to see if the findings observed in this study through a 2D fretting contact
would apply at a 3D fretting interface.
305
References
[1] NJR. (2016, 29-05-2017). National Joint Registry for England and
Wales: National Joint Registry 13th Annual Reports. Available:
http://www.njrreports.org.uk/Portals/0/PDFdownloads/NJR%2013th%
20Annual%20Report%202016.pdf
[2] D. J. T. M J Yaszemski, K Lewandrowski, V Hasirci, D Altobelli, D L

Wise, Biomaterials in Orthopedics Marcel Dekkar, 2004.
[3] J. B. Park, Kon Kim, Y. , Metallic Biomaterials: Boca Raton: CRC Press
LLC, 2000.
[4] L. Shi, Derek O. Northwood, and Zhengwang Cao, "The properties of

a wrought biomedical cobalt-chromium alloy," Journal of materials
science, vol. 29, pp. 1233-1238, 1994.
[5] G. K. Smith, "Orthopaedic biomaterials," 1985.
[6] T. J. Band, "Materials and metallurgy," in Modern Hip Resurfacing, ed

London: Springer 2009, pp. 43-63.
[7] Y. Hedberg, and Inger Odnevall Wallinder, "Metal release and

speciation of released chromium from a biomedical CoCrMo alloy into
simulated physiologically relevant solutions," Journal of Biomedical
Materials Research Part B, vol. Applied Biomaterials 102, pp. 693-699,
2014.
[8] B. Scharf, Clement, C.C., Zolla, V., Perino, G., Yan, B., Elci, S.G.,
Purdue, E., Goldring, S., Macaluso, F., Cobelli, N. and Vachet, R.W.,
"Molecular analysis of chromium and cobalt-related toxicity," Scientific
reports, vol. 4, p. 5729, 2014.
[9] C. Oldani, Alejandro Dominguez, and T. Eli., "Titanium as a Biomaterial

for Implants," in Recent Advances in Arthroplasty, S. K. Fokter, Ed., ed:
InTech, 2012.
306
[10] J. V. Bono, Revision total hip arthroplasty: Springer Science &
Business Media, 1999.
[11] B. Weisse, Ch Affolter, G. P. Terrasi, G. Piskoty, and S. Köbel, "Failure

analysis of in vivo fractured ceramic femoral heads," Engineering
Failure Analysis, vol. 16, pp. 1188-1194, 2009.
[12] R. Elke, "Particle disease: status and today’s solutions," presented at

the 7th international BIOLOX symposium, 2002.
[13] B. MASSON, "Emergence of the alumina matrix composite in total hip

arthroplasty," International Orthopaedics, pp. 33, 359-363., 2009.
[14] B. J. McEntire, B. S. Bal, M. N. Rahaman, J. Chevalier, and G. Pezzotti,

"Ceramics and ceramic coatings in orthopaedics," Journal of the
European Ceramic Society vol. 35, pp. 4327-4369, 2015.
[15] CeramTec. (2015). BIOLOX(R)delta. Available:

http://www.ceramtec.co.uk/files/mt_biolox_delta_en.pdf
[16] S. Lal, Lisa Allinson, Richard M. Hall, and Joanne L. Tipper, "A
comprehensive assessment of biological responses to silicon nitride
nanoparticles and cobalt chromium wear debris from total hip
replacements," In Front. Bioeng. Biotechnol. , vol. 4, 2016.
[17] J. Olofsson, T. Mikael Grehk, Torun Berlind, Cecilia Persson, Staffan

Jacobson, and Håkan Engqvist, "Evaluation of silicon nitride as a wear
resistant and resorbable alternative for total hip joint replacement,"
Biomatter vol. 2, pp. 94-102, 2012.
[18] Arx Spinal System, US Food and Drug Adiministration (FDA), 2006.
[19] B. S. Bal, and Mohamed Rahaman, "The Rationale for Silicon Nitride
Bearings in Orthopaedic Applications," Advances in Ceramics-Electric
and Magnetic Ceramics, Bioceramics, Ceramics and Environment,
InTech, Bioceramics, Ceramics and Environment, 2011.
307
[20] M. Herrmann, "Corrosion of silicon nitride materials in aqueous
solutions," Journal of the American Ceramic Society, vol. 96, pp. 3009-
3022, 2013.
[21] P. R. Roberge, Corrosion engineering: principles and practice: New

York: McGraw-Hill, 2008.
[22] R. B. E. Stansbury, "Fundamentals of Electrochemical Corrosion,"

Ohio: ASM International, 2000.
[23] U. o. Cambridge. (2017, 13-06-2017). The Electrical Double Layer.

Available: http://www.ceb.cam.ac.uk/research/groups/rg-
eme/teaching-notes/the-electrical-double-layer
[24] INSA-LYON-LABORATOIRE. (2014, 23-7-14). The Multimedia

Corrosion Guide. Available:
http://www.cdcorrosion.com/mode_corrosion/corrosion_uniform.htm
[25] Z. Ahmad, Principles of corrosion engineering and corrosion control:

Butterworth-Heinemann, 2006.
[26] J. L. Gilbert, Christine A. Buckley, Joshua J. Jacobs, Kim C. Bertin, and

Michael R. Zernich, "Intergranular corrosion-fatigue failure of cobalt-
alloy femoral stems. A failure analysis of two implants," J Bone Joint
Surg Am, vol. 76, pp. 110-115, 1994.
[27] J. L. Gilbert, Christine A. Buckley, and Joshua J. Jacobs., " In vivo

corrosion of modular hip prosthesis components in mixed and similar
metal combinations. The effect of crevice, stress, motion, and alloy
coupling.," Journal of biomedical materials research, vol. 27, pp. 1533-
1544, 1993.
[28] M. G. Fontana, and N. G. Greene, Corrosion Engineering: New York:

Mc Graw Hill, 1978.
[29] J. W. Oldfield, and W. H. Sutton, "Crevice Corrosion of Stainless Steel

I – A Mathematical Model," British corrosion journal, vol. 13, pp. 13-22,
1978.
308
[30] J. P. Collier, Surprenant, V. A., Jensen, R. E., & Mayor, M. B.,
"Corrosion at the interface of cobalt-alloy heads on titanium-alloy
stems," Clinical orthopaedics and related research, vol. 271, 1991.
[31] R. W. Revie, and Herbert H. Uhlig, "Thermodynamics: Corrosion

Tendency and Electrode Potentials," in Corrosion and Corrosion
Control: An Introduction to Corrosion Science and Engineering, 4th ed:
John Wiley and Sons, New York, 2000, pp. 21-41.
[32] No-Authors-Listed. (2011, 17/8/2014). How to explain

Potentiodynamic. Available: http://srizam-
expro.blogspot.co.uk/2011/03/how-to-explain-potentiodynamic.html
[33] G. Stachowiak, and Andrew W. Batchelor, Engineering Tribology, 3rd

ed.: Butterworth-Heinemann, 2005.
[34] J.Williams, Engineering Tribology, 2 ed.: Cambridge University Press,

1994.
[35] V. Cuffaro, "Prediction Method for the Surface Damage in Splined

Couplings," PhD, 2013.
[36] J. A. Williams, Engineering Tribology. Oxford: Oxford University Press,

1998.
[37] J. T. Burwell, "Survey of possible wear mechanisms," Wear, vol. 1, pp.

119-141, 1957.
[38] O. Vingsbo, and Staffan Söderberg, "On fretting maps," Wear, vol. 126,
pp. 131–147, 1988.
[39] S. Fouvry, Philippe Kapsa, and Leo Vincent, "Quantification of fretting

damage," Wear, vol. 200, pp. 186-205, 1996.
[40] R.D.Mindlin, "Compliance of elastic bodies in contact," J. Applied

Mechanics, vol. 16, pp. 259 - 268, 1949.
309
[41] B. Tritschler, Bernard Forest, and Jean Rieu., "Fretting corrosion of
materials for orthopaedic implants: a study of a metal/polymer contact
in an artificial physiological medium.," Tribology international, vol. 32,
1999.
[42] O. Vingsbo, and Joakim Schön, "Gross slip criteria in fretting," Wear,
vol. 162, pp. 347-356, 1993.
[43] P. Schaaff, "The role of fretting damage in total hip arthroplasty with
modular design hip joints-evaluation of retrieval studies and
experimental simulation methods," Journal of applied biomaterials &
biomechanics: JABB, vol. 2, pp. 121-135, 2003.
[44] I.-M. Feng, and B. G. Rightmire, "An experimental study of fretting,"

Proceedings of the Institution of Mechanical Engineers, vol. 170, pp.
1055-1064, 1956.
[45] U. Bryggman, and Staffan Söderberg, "Contact conditions in fretting,"

Wear, vol. 110, pp. 1-17, 1986.
[46] G. A. Tomlinson, "The rusting of steel surfaces in contact," in

Proceedings of the Royal Society of London, 1927, pp. 472-483.
[47] D. Landolt, S. Mischler, M. Stemp, and S. Barril, "Third body effects

and material fluxes in tribocorrosion systems involving a sliding
contact," Wear, vol. 256, pp. 517-524, 2004.
[48] I. D. Learmonth, C. Young, and C. Rorabeck, "The operation of the

century: total hip replacement," The Lancet, vol. 370, pp. 1508-1519,
2007.
[49] S. R. Knight, Randeep Aujla, and Satya Prasad Biswas, "Total Hip
Arthroplasty-over 100 years of operative history," Orthopedic reviews,
vol. 3, 2011.
[50] H. McKellop, Fu‐wen Shen, Bin Lu, Patricia Campbell, and Ronald
Salovey, "Development of an extremely wear‐resistant ultra high
310
molecular weight polythylene for total hip replacements," Journal of
Orthopaedic Research, vol. 17, pp. 157-167, 1999.
[51] U. Sentuerk, P. von Roth, and C. Perka, "Ceramic on ceramic

arthroplasty of the hip," Bone Joint J, vol. 98, pp. 14-17, 2016.
[52] S. U. Leslie F. Scheuber, Florence Petkow. (2014, 02-06-2017). The

Neck Taper in Hip Arthroplasty - What does the surgeon have to
consider? Available:
https://www.ceramtec.com/files/mt_taper_and_compatibility.pdf
[53] T. McTighe, Declan Brazil, Louis Keppler, John Keggi, and Edward
McPherson, "Metallic Modular Taper Junctions in Total Hip
Arthroplasty," Reconstructive Review, vol. 5, 2015.
[54] A. Srinivasan, Edward Jung, and Brett Russell Levine, "Modularity of

the femoral component in total hip arthroplasty," Journal of the
American Academy of Orthopaedic Surgeons, vol. 20, pp. 214-222,
2012.
[55] H. Krishnan, S. P. Krishnan, G. Blunn, J. A. Skinner, and A. J. Hart,

"Modular neck femoral stems," Bone Joint J, vol. 95, pp. 1011-1021,
2013.
[56] S. Hussenbocus, Kosuge, D., Solomon, L. B., Howie, D. W., & Oskouei,
R. H., "Head-neck taper corrosion in hip arthroplasty," BioMed research
international, 2015.
[57] M. Morlock, Dennis Bünte, Julian Gührs, and Nicholas Bishop,

"Corrosion of the Head-Stem Taper Junction—Are We on the Verge of
an Epidemic?," HSS Journal®, pp. 1-8.
[58] R. Chana, C. Esposito, P. Campbell, W. Walter, and W. Walter, "Mixing

and matching causing taper wear," J Bone Joint Surg Br, vol. 94, pp.
281-286, 2012.
[59] S. Y. Jauch-Matt, A. W. Miles, and H. S. Gill, "Effect of trunnion

roughness and length on the modular taper junction strength under
311
typical intraoperative assembly forces," Medical Engineering &
Physics, vol. 39, pp. 94-101, 2017.
[60] N. Bishop, F. Waldow, and M. Morlock, "Friction moments of large

metal-on-metal hip joint bearings and other modern designs," Medical
engineering & physics, vol. 30, pp. 1057-1064, 2008.
[61] W. H. Wuttke V, Kempf K, Oberbach T, Delfosse D, "Influence of

various types of damage on the fracture strength of ceramic femoral
heads," Biomed Tech (Berl), vol. 56, pp. 333–9, 2011.
[62] Zimmer. (2014). Zimmer M/L Taper Hip Prosthesis with Kinectiv
Technology Surgical Technique. Available:
http://www.zimmer.com/medical-professionals/products/hip/ml-taper-
kinectiv-technology.html
[63] F. Witt, Julian Gührs, Michael M. Morlock, and Nicholas E. Bishop,

"Quantification of the contact area at the head-stem taper interface of
modular hip prostheses," PloS one, vol. 10, 2015.
[64] H. Haschke, S. Y. Jauch-Matt, K. Sellenschloh, G. Huber, and M. M.

Morlock, "Assembly force and taper angle difference influence the
relative motion at the stem–neck interface of bi-modular hip
prostheses," Proceedings of the Institution of Mechanical Engineers,
Part H: Journal of Engineering in Medicine, vol. 230, pp. 690-699, 2016.
[65] A. Rehmer, N. E. Bishop, and M. M. Morlock, "Influence of assembly

procedure and material combination on the strength of the taper
connection at the head–neck junction of modular hip endoprostheses,"
Clinical Biomechanics, vol. 27, pp. 77-83, 2012.
[66] L. C. Lucas, R. A. Buchanan, and J. E. Lemons, "Investigations on the

galvanic corrosion of multialloy total hip prostheses," Journal of
Biomedical Materials Research Part A, vol. 15, pp. 731-747, 1981.
[67] J. R. Goldberg, Gilbert, J. L., Jacobs, J. J., Bauer, T. W., Paprosky, W.,
& Leurgans, S., "A multicenter retrieval study of the taper interfaces of
312
modular hip prostheses," Clinical orthopaedics and related research,
vol. 401, pp. 149-161, 2002.
[68] H. J. Higgs G, MacDonald D, Kane W, Day J, Klein G, Parvizi J, Mont

M, Kraay M, Martell J, Gilbert J, "In Metal-On-Metal Total Hip
Replacement Devices," in Method of characterizing fretting and
corrosion at the various taper connections of retrieved modular
components from metal-on-metal total hip arthroplasty, ed. ASTM
International, 2013.
[69] S. Munir, Michael B. Cross, Christina Esposito, Anna Sokolova, and

William L. Walter, "Corrosion in modular total hip replacements: An
analysis of the head–neck and stem–sleeve taper connections," In
Seminars in Arthroplasty, vol. 24, pp. 240-245, 2013.
[70] R. M. Urban, Joshua J. Jacobs, Jeremy L. Gilbert, Stephen B. Rice,

Murali Jasty, Charles R. Bragdon, and Jorge O. Galante,
"Characterization of solid products of corrosion generated by modular-
head femoral stems of different designs and materials," In Modularity
of Orthopedic Implants, ASTM International, 1997.
[71] G. B. Higgs, Josa A. Hanzlik, Daniel W. MacDonald, Jeremy L. Gilbert,

Clare M. Rimnac, Steven M. Kurtz, and Implant Research Center
Writing Committee. ": a retrieval study.", "Is increased modularity
associated with increased fretting and corrosion damage in metal-on-
metal total hip arthroplasty devices?," The Journal of arthroplasty, vol.
28, pp. 2-6, 2013.
[72] S. M. Kurtz, Sevi B. Kocagöz, Josa A. Hanzlik, Richard J. Underwood,

Jeremy L. Gilbert, Daniel W. MacDonald, Gwo-Chin Lee et al, "Do
ceramic femoral heads reduce taper fretting corrosion in hip
arthroplasty? A retrieval study.," Clinical Orthopaedics and Related
Research®, vol. 471, pp. 3270-3282, 2013.
[73] A. Di Laura, H. Hothi, J. Henckel, I. Swiatkowska, M. H. L. Liow, Y. M.

Kwon, J. A. Skinner, and A. J. Hart, "Retrieval analysis of metal and
313
ceramic femoral heads on a single CoCr stem design," Bone and Joint
Research, vol. 6, pp. 345-350, 2017.
[74] M. M. Morlock, Dennis Bünte, Harmen Ettema, Cees C. Verheyen, Åke

Hamberg, and Jeremy Gilbert, "Primary hip replacement stem taper
fracture due to corrosion in 3 patients," Acta orthopaedica, vol. 87, pp.
189-192, 2016.
[75] N. Bishop, F. Witt, R. Pourzal, A. Fischer, M. Rutschi, M. Michel, et al.,

"Wear patterns of taper connections in retrieved large diameter metal-
on-metal bearings," J Orthop Res, vol. 31, pp. 1116-22, Jul 2013.
[76] N. Moharrami, D. J. Langton, O. Sayginer, and S. J. Bull, "Why does

titanium alloy wear cobalt chrome alloy despite lower bulk hardness: A
nanoindentation study?," Thin Solid Films, vol. 549 pp. 79-86, 2013.
[77] B. A. Lanting, Teeter, M. G., Vasarhelyi, E. M., Ivanov, T. G., Howard,

J. L., & Naudie, "Correlation of corrosion and biomechanics in the
retrieval of a single modular neck total hip arthroplasty design: modular
neck total hip arthroplasty system," The Journal of arthroplasty, vol. 30,
pp. 135-140, 2015.
[78] D. Buente, Huber, G., Bishop, N., & Morlock, M., "Quantification of
material loss from the neck piece taper junctions of a bimodular primary
hip prosthesis. A retrieval study from 27 failed Rejuvenate bimodular
hip arthroplasties," Bone Joint J, vol. 97, pp. 1350-1357, 2015.
[79] I. P. S. Gill, J. Webb, K. Sloan, and R. J. Beaver, "Corrosion at the

neck-stem junction as a cause of metal ion release and pseudotumour
formation," Journal of Bone & Joint Surgery, vol. British Volume 94, pp.
895-900, 2012.
[80] T. M. Grupp, Thomas Weik, Wilhelm Bloemer, and Hanns-Peter

Knaebel "Modular titanium alloy neck adapter failures in hip
replacement - failure mode analysis and influence of implant material,"
BMC musculoskeletal disorders 11, vol. 1, 2010.
314
[81] C. J. Dangles and C. J. Altstetter, "Failure of the modular neck in a total
hip arthroplasty," J Arthroplasty, vol. 25, pp. 1169 e5-7, Oct 2010.
[82] J. G. Skendzel, J. D. Blaha, and A. G. Urquhart, "Total hip arthroplasty

modular neck failure," J Arthroplasty, vol. 26, pp. 338 e1-4, Feb 2011.
[83] M. Nganbe, Usman Khan, Hakim Louati, Andrew Speirs, and Paul E.
Beaulé, "In vitro assessment of strength, fatigue durability, and
disassembly of Ti6Al4V and CoCrMo necks in modular total hip
replacements," Journal of Biomedical Materials Research Part B:
Applied Biomaterials, vol. 97, pp. 132-138, 2011.
[84] A. M. Kop, C. Keogh, and E. Swarts, "Proximal component modularity

in THA—at what cost?: an implant retrieval study," Clinical
Orthopaedics and Related Research®, vol. 470, pp. 1885-1894, 2012.
[85] J. J. Jacobs, Robert M. Urban, Jeremy L. Gilbert, Anastasia K. Skipor,

Jonathan Black, Murali Jasty, and Jorge O. Galante, "Local and distant
products from modularity," Clinical orthopaedics and related research,
vol. 319, pp. 94-105, 1995.
[86] M. Oldenburg, Ralf Wegner, and Xaver Baur, "Severe cobalt

intoxication due to prosthesis wear in repeated total hip arthroplasty,"
The Journal of arthroplasty, vol. 24, pp. 825-e15, 2009.
[87] J. J. Jacobs, Jeremy L. Gilbert, and Robert M. Urban, "Current

concepts review-corrosion of metal orthopaedic implants," J Bone Joint
Surg Am, vol. 80, pp. 268-82, 1998.
[88] R. M. Urban, Joshua J. Jacobs, Jeremy L. Gilbert, and Jorge O.

Galante., "Migration of corrosion products from modular hip
prostheses. Particle microanalysis and histopathological findings," J
Bone Joint Surg Am, vol. 76, pp. 1345-1359, 1994.
[89] H. J. Agins, N. W. Alcock, Manjula Bansal, E. A. Salvati, P. D. Wilson,

P. M. Pellicci, and P. G. Bullough, "Metallic wear in failed titanium-alloy
315
total hip replacements. A histological and quantitative analysis," J Bone
Joint Surg Am, vol. 70, pp. 347-356, 1988.
[90] J. Black, H. Sherk, J. Bonini, W. R. Rostoker, F. Schajowicz, and J. O.

Galante, "Metallosis associated with a stable titanium-alloy femoral
component in total hip replacement," JBJS, vol. 72, pp. 126-130, 1990.
[91] G. Meachim, and D. F. Williams, "Changes in nonosseous tissue

adjacent to titanium implants," Journal of Biomedical Materials
Research Part A, vol. 7, pp. 555-572, 1973.
[92] J. J. Jacobs, Anastasia K. Skipor, Leslie M. Patterson, Nadim J. Hallab,

Wayne G. Paprosky, Jonathan Black, and Jorge O. Galante, "Metal
release in patients who have had a primary total hip arthroplasty. A
prospective, controlled, longitudinal study," J Bone Joint Surg Am, vol.
80, pp. 1447-58, 1998.
[93] L. Leyssens, Bart Vinck, Catherine Van Der Straeten, Floris Wuyts, and
Leen Maes, "Cobalt toxicity in humans. A review of the potential
sources and systemic health effects," Toxicology, 2017.
[94] M. A. Wimmer, Alfons Fischer, Robin Büscher, Robin Pourzal,

Christoph Sprecher, Roland Hauert, and Joshua J. Jacobs, "Wear
mechanisms in metal‐on‐metal bearings: The importance of
tribochemical reaction layers," Journal of Orthopaedic Research, vol.
28, pp. 436-443, 2010.
[95] M. T. Mathew, C. Nagelli, R. Pourzal, A. Fischer, M. P. Laurent, J. J.

Jacobs, and M. A. Wimmer, "Tribolayer formation in a metal-on-metal
(MoM) hip joint: an electrochemical investigation," Journal of the
mechanical behavior of biomedical materials, vol. 29, pp. 199-212,
2014.
[96] P. Zeng, W. M. Rainforth, and R. B. Cook, "Characterisation of the

oxide film on the taper interface from retrieved large diameter metal on
polymer modular total hip replacements," Tribology International, vol.
89, pp. 86-96, 2015.
316
[97] J. L. Gilbert, Mali, S., Urban, R. M., Silverton, C. D., & Jacobs, J. J., "In
vivo oxide‐induced stress corrosion cracking of Ti‐6Al‐4V in a neck–
stem modular taper: Emergent behavior in a new mechanism of in vivo
corrosion," Journal of Biomedical Materials Research Part B: Applied
Biomaterials, vol. 100, pp. 584-594, 2012.
[98] M. G. Bryant, D. Buente, A. Oladokun, M. Ward, G. Huber, M. Morlock,

and A. Neville, "Surface and subsurface changes as a result of
tribocorrosion at the stem-neck interface of bi-modular prosthesis,"
Biotribology, 2017.
[99] V. Swaminathan, and Jeremy L. Gilbert, "Fretting corrosion of CoCrMo

and Ti6Al4V interfaces," Biomaterials, vol. 33, pp. 5487-5503, 2012.
[100] "ASTM F1875-98(2014), Standard Practice for Fretting Corrosion

Testing of Modular Implant Interfaces: Hip Femoral Head-Bore and
Cone Taper Interface," ed. West Conshohocken, PA: ASTM
International, 2014.
[101] A. Panagiotidou, J. Meswania, K. Osman, B. Bolland, J. Latham, J.

Skinner, F. S. Haddad, A. Hart, and G. Blunn, "The effect of frictional
torque and bending moment on corrosion at the taper interface," Bone
Joint J, vol. 97, pp. 463-472, 2015.
[102] M. J. Baxmann, S. Y. Schilling, C. Blomer, W. Grupp, T. M. Morlock,

M. M., "The influence of contact conditions and micromotions on the
fretting behavior of modular titanium alloy taper connections," Med Eng
Phys, vol. 35, pp. 676-83; discussion 676, May 2013.
[103] D. Royhman, Patel, M., Runa, M. J., Wimmer, M. A., Jacobs, J. J.,
Hallab, N. J., & Mathew, M. T., "Fretting-Corrosion Behavior in Hip
Implant Modular Junctions: The Influence of Friction Energy and pH
Variation. ," Journal of the Mechanical Behavior of Biomedical
Materials, 2016.
317
[104] S. Barril, S. Mischler, and D. Landolt, "Influence of fretting regimes on
the tribocorrosion behaviour of Ti6Al4V in 0.9 wt.% sodium chloride
solution," Wear, vol. 256, pp. 963-972, 2004.
[105] A. Bazzoni, S. Mischler, and N. Espallargas, "Tribocorrosion of pulsed

plasma-nitrided CoCrMo implant alloy," Tribology Letters, vol. 49, pp.
157-167, 2013.
[106] F. E. Donaldson, James C. Coburn, and Karen Lohmann Siegel, "Total

hip arthroplasty head–neck contact mechanics: A stochastic
investigation of key parameters," Journal of biomechanics, vol. 47, pp.
1634-1641, 2014.
[107] T. Zhang, N. M. Harrison, P. F. McDonnell, P. E. McHugh, and S. B.

Leen, "A finite element methodology for wear–fatigue analysis for
modular hip implants," Tribology International, vol. 65, pp. 113-127,
2013.
[108] V. Swaminathan, and Jeremy L. Gilbert, "Potential and frequency

effects on fretting corrosion of Ti6Al4V and CoCrMo surfaces," Journal
of Biomedical Materials Research Part A, vol. 101, pp. 2602-2612,
2013.
[109] A. Oladokun, M. Pettersson, M. Bryant, H. Engqvist, C. Persson, R.

Hall, and A. Neville, "Fretting of CoCrMo and Ti6Al4V alloys in modular
prostheses," Tribology - Materials, Surfaces & Interfaces, vol. 9, pp.
165 - 173, 2015.
[110] J. Dufils, Markus A. Wimmer, Joachim Kunze, Mathew T. Mathew, and

Michel P. Laurent, "Influence of molybdate ion and pH on the fretting
corrosion of a CoCrMo–Titanium alloy couple," Biotribology, 2017.
[111] M. Wimmer, Mathew Mathew, Michel Laurent, Christopher Nagelli,

Yifeng Liao, Laurence Marks, Robin Pourzal, Alfons Fischer, and
Joshua Jacobs, "Tribochemical reactions in metal-on-metal hip joints
influence wear and corrosion," In Metal-On-Metal Total Hip
Replacement Devices. ASTM International., pp. 1–18, 2013.
318
[112] M. A. Wimmer, C. Sprecher, R. Hauert, G. Täger, and A. Fischer,
"Tribochemical reaction on metal-on-metal hip joint bearings: a
comparison between in-vitro and in-vivo results," Wear, vol. 255, pp.
1007-1014, 2003.
[113] E. J. Martin, Robin Pourzal, Mathew T. Mathew, and Kenneth R. Shull,

"Dominant role of molybdenum in the electrochemical deposition of
biological macromolecules on metallic surfaces," Langmuir, vol. 29, pp.
4813-4822, 2013.
[114] S. D. Cook, Robert L. Barrack, Gregory C. Baffes, Alastair JT Clemow,

Paul Serekian, Nick Dong, and Mark A. Kester, "Wear and corrosion of
modular interfaces in total hip replacements," Clinical orthopaedics and
related research vol. 298, pp. 80-88, 1994.
[115] J. L. Gilbert, M. Mehta, and B. Pinder, "Fretting crevice corrosion of

stainless steel stem-CoCr femoral head connections: comparisons of
materials, initial moisture, and offset length," J Biomed Mater Res B
Appl Biomater, vol. 88, pp. 162-73, Jan 2009.
[116] S. A. Brown, Alula Abera, Mark D'Onofrio, and Curt Flemming, "Effects
of neck extension, coverage, and frequency on the fretting corrosion of
modular THR bore and cone interface," in Modularity of Orthopedic
Implants, ed: ASTM International, 1997.
[117] A. Traynor, Amy Kinbrum, Jonathan Housden, and Simon Collins. "." ,
no. SUPP (): . "Fretting and Corrosion of Tapered Head-Neck
Junctions of Modular Hip Components," Bone Joint J, vol. 95, pp. 270-
270, 2013.
[118] M. L. Mroczkowski, J. S. Hertzler, S. M. Humphrey, T. Johnson, and C.

R. Blanchard, "Effect of impact assembly on the fretting corrosion of
modular hip tapers," J Orthop Res, vol. 24, pp. 271-9, Feb 2006.
[119] A. Panagiotidou, Timothy Cobb, Jay Meswania, John Skinner, Alister

Hart, Fares Haddad, and Gordon Blunn, "The Effect Of Impact
Assembly On The Interface Deformation And Fretting Corrosion Of
319
Modular Hip Tapers: An In Vitro Study," Journal of Orthopaedic
Research, 2017.
[120] M. M. Morlock, Nick Bishop, Jozef Zustin, Michael Hahn, Wolfgang

Rüther, and Michael Amling, "Modes of implant failure after hip
resurfacing: morphological and wear analysis of 267 retrieval
specimens," The Journal of Bone & Joint Surgery, vol. 90,, pp. 89-95,
2008.
[121] M. Lavigne, Etienne L. Belzile, Alain Roy, François Morin, Traian

Amzica, and Pascal-André Vendittoli, "Comparison of whole-blood
metal ion levels in four types of metal-on-metal large-diameter femoral
head total hip arthroplasty: the potential influence of the adapter
sleeve," J Bone Joint Surg Am, vol. 93, pp. 128-136, 2011.
[122] G. Bergmann, F. Graichen, A. Rohlmann, A. Bender, B. Heinlein, G. N.

Duda, et al., "Realistic loads for testing hip implants," Biomed Mater
Eng, vol. 20, pp. 65-75, 2010.
[123] H. Farhoudi, Reza H. Oskouei, Ali A. Pasha Zanoosi, Claire F. Jones,

and Mark Taylor, "An Analytical Calculation of Frictional and Bending
Moments at the Head-Neck Interface of Hip Joint Implants during
Different Physiological Activities," Materials, vol. 9, 2016.
[124] S. Jauch, L. Coles, L. Ng, A. Miles, and H. Gill, "Low torque levels can
initiate a removal of the passivation layer and cause fretting in modular
hip stems," Medical engineering & physics, vol. 36, pp. 1140-1146,
2014.
[125] A. Panagiotidou, J. Meswania, J. Hua, S. Muirhead‐Allwood, A. Hart,

and G. Blunn, "Enhanced wear and corrosion in modular tapers in total
hip replacement is associated with the contact area and surface
topography," Journal of orthopaedic research, vol. 31, pp. 2032-2039,
2013.
[126] A. Panagiotidou, Jay Meswania, Jia Hua, Sarah K. Muirhead-Allwood,

John A. Skinner, Alister Hart, and Gordon Blunn, "Do Surface
320
Topography and Contact Area Effect Fretting Corrosion Behaviour of
the Modular Taper Interface in Total Hip Replacements?," Bone Joint
J, vol. 95, pp. 474-474, 2013.
[127] R. M. R. Dyrkacz, J. M. Brandt, J. B. Morrison, S. T. O’Brien, O. A. Ojo,

T. R. Turgeon, and U. P. Wyss, "Finite element analysis of the head–
neck taper interface of modular hip prostheses.," Tribology
International, vol. 91, pp. 206-213, 2015.
[128] S. Y. Jauch, G. Huber, H. Haschke, K. Sellenschloh, and M. M.

Morlock, "Design parameters and the material coupling are decisive for
the micromotion magnitude at the stem-neck interface of bi-modular hip
implants," Med Eng Phys, Dec 11 2013.
[129] S. Y. Jauch, G. Huber, K. Sellenschloh, H. Haschke, M. Baxmann, T.

M. Grupp, et al., "Micromotions at the taper interface between stem and
neck adapter of a bimodular hip prosthesis during activities of daily
living," Journal of Orthopaedic Research, vol. 31, pp. 1165-1171, 2013.
[130] B. L. Brandt JM, Marr J, MacDonald SJ, Bourne RB, Medley JB,
"Biochemical Comparisons of Osteoarthritic Human Synovial Fluid with
Calf Sera Used in Knee Simulator Wear Testing," Journal of Biomedical
Materials Research Part A, vol. 94A, pp. 961-971, 2010.
[131] S. A. Brown, Paula J. Hughes, and Katharine Merritt, "In vitro studies
of fretting corrosion of orthopaedic materials," Journal of orthopaedic
research, vol. 6, pp. 572-579, 1988.
[132] A. C. Lewis, M. R. Kilburn, I. Papageorgiou, G. C. Allen, and C. P.

Case, "Effect of synovial fluid, phosphate‐buffered saline solution, and
water on the dissolution and corrosion properties of CoCrMo alloys as
used in orthopedic implants," Journal of Biomedical Materials
Research, vol. Part A 73, pp. 456-467, 2005.
[133] T. Hanawa and S. Hiromoto, " Surface Modification of metals and

Alloys in a Human Body Environment," Corrosion Engineering, pp. 895-
907, 1998.
321
[134] J. P. Glusker, "Structural aspects of metal liganding to functional
groups of proteins," Advances in protein 1991.
[135] Y.Yan., "Corrosion and Tribo-corrosion behaviour of metallic

orthopaedic implant material," PhD: , University of Leeds, 2006.
[136] M. M. Stack, and K. Chi, "Mapping sliding wear of steels in aqueous

conditions," Wear, vol. 255, pp. 456-465, 2003.
[137] M. T. Mathew, P. Srinivasa Pai, R. Pourzal, A. Fischer, and M. A.

Wimmer, "Significance of tribocorrosion in biomedical applications:
overview and current status," Advances in tribology 2009, 2010.
[138] N. E. Stefano Mischler, Anna Igual Muñoz, "Bio-Tribocorrosion

Fundamentals to Orthopaedic Surgeon and Researcher," presented at
the ORS Annual Meeting, 2014.
[139] M. Bryant, and Anne Neville, "Fretting corrosion of CoCr alloy: Effect of
load and displacement on the degradation mechanisms," Proceedings
of the Institution of Mechanical Engineers, Part H: Journal of
Engineering in Medicine, 2016.
[140] B. J. Smith, and Paul Ducheyne, "Transitional Behavior in Ti-6AL-4V

fretting corrosion," in In Medical Applications of Titanium and Its Alloys:
The Material and Biological Issues, ed. ASTM International, 1996.
[141] J. L. Gilbert, Christine A. Buckley, and Eugene P. Lautenschlager,

"Titanium oxide film fracture and repassivation: the effect of potential,
pH and aeration," in Medical applications of titanium and its alloys, ed.
The material and biological issues: ASTM International, 1996.
[142] G. J. Goldberg JR, "Electrochemical Response of Cocrmo to High-

Speed Fracture of Its Metal Oxide Using an Electrochemical Scratch
Test Method. ," John Wiley & Sons, pp. 421-431., 1997.
[143] T. Hanawa, "Metal ion release from metal implants," Materials Science
and Engineering: C, vol. 24, pp. 745-752, 2004.
322
[144] N. N. M. Morita, Y. Tsukamoto, T. Sasada, Japanese Society for
Biomaterials, vol. 10, p. 209, 1992.
[145] S. Mischler, "Triboelectrochemical techniques and interpretation

methods in tribocorrosion: a comparative evaluation," Tribology
International, vol. 41, pp. 573-583, 2008.
[146] F. W. P Ponthiaux, D Drees, JP Celis "Electrochemical techniques for

studying tribocorrosion processes," Wear, vol. 256, pp. 459-468, 2004.
[147] R. W.-W. Hsu, Chun-Chen Yang, Ching-An Huang, and Yi-Sui Chen,
"Electrochemical corrosion properties of Ti–6Al–4V implant alloy in the
biological environment," Materials Science and Engineering, vol. 380,
pp. 100-109, 2004.
[148] B. Alemón, M. Flores, W. Ramírez, J. C. Huegel, and Esteban

Broitman, "Tribocorrosion behavior and ions release of CoCrMo alloy
coated with a TiAlVCN/CNx multilayer in simulated body fluid plus
bovine serum albumin," Tribology International, vol. 81, pp. 159-168,
2015.
[149] S. Fouvry, Ph Kapsa, and L. Vincent., "Analysis of sliding behaviour for

fretting loadings: determination of transition criteria," Wear, vol. 185,
pp. 35-46, 1995.
[150] M. Bryant, "Fretting-crevice corrosion of cemented metal on metal total

hip replacements," PhD PhD Thesis, University of Leeds, 2013.
[151] K. Elleuch, and S. Fouvry, "Wear analysis of A357 aluminium alloy

under fretting," Wear, vol. 253, pp. 662-672, 2002.
[152] C. Hammond, Basics of crystallography and diffraction vol. 214: Oxford

University Press, 2009.
[153] M. C. Biesinger, Brad P. Payne, Andrew P. Grosvenor, Leo WM Lau,

Andrea R. Gerson, and Roger St C. Smart, "Resolving surface
chemical states in XPS analysis of first row transition metals, oxides
323
and hydroxides: Sc, Ti, V, Cu and Zn," Applied Surface Science, vol.
257, pp. 887-898, 2010.
[154] M. C. Biesinger, Brad P. Payne, Andrew P. Grosvenor, Leo WM Lau,

Andrea R. Gerson, and Roger St C. Smart, "Resolving surface
chemical states in XPS analysis of first row transition metals, oxides
and hydroxides: Cr, Mn, Fe, Co and Ni," Applied Surface Science, vol.
257, pp. 2717-2730, 2011.
[155] A. K.-V. Alexander V. Naumkin, Stephen W. Gaarenstroom, and Cedric

J. Powel NIST X-ray Photoelectron Spectroscopy Database [Online].
Available: https://srdata.nist.gov/xps/Default.aspx
[156] G. Beamson, and David Briggs, "High resolution XPS of organic

polymers," 1992.
[157] G. D. Silversmit, Diederik Poelman Hilde, Marin Guy B. De Gryse,

Roger, "Determination of the V2p XPS binding energies for different
vanadium oxidation states (V5+ to V0+)," Journal of Electron
Spectroscopy and Related Phenomena, vol. 135, pp. 167-175, 4//
2004.
[158] I. Milošev, and Maja Remškar, "In vivo production of nanosized metal
wear debris formed by tribochemical reaction as confirmed by high‐
resolution TEM and XPS analyses," Journal of Biomedical Materials
Research Part A vol. 91.4, pp. 1100-1110, 2009.
[159] M. Biesinger. ([27-06-2017]). X-ray photoelectron spectroscopy.

Available: http://www.xpsfitting.com
[160] W. F. S. John F. Moulder, Peter E. Sobol, Kenneth D. Bomben,

Handbook of X-ray Photoelectron Spectroscopy: Perkin-Elmer
Corporation, Physical Electronics Division, 6509 Flying Cloud Drive,
Eden Prairie, Minnesota 55344, United States of America, 1992.
[161] No-Authors-Listed. (2017, [27-06-17). XPS Knowledge Base.

Available: http://xpssimplified.com/periodictable.php
324
[162] L. V. Duong, Barry J. Wood, and J. Theo Kloprogge, "XPS study of
basic aluminum sulphate and basic aluminium nitrate," Materials
Letters, vol. 59, pp. 1932-1936, 2005.
[163] A. I. Muñoz, and S. Mischler, "Interactive effects of albumin and

phosphate ions on the corrosion of CoCrMo implant alloy," Journal of
the Electrochemical Society, vol. 154, pp. C562-C570, 2007.
[164] Y. Fu, Hejun Du, Sam Zhang, and Weimin Huang, "XPS
characterization of surface and interfacial structure of sputtered TiNi
films on Si substrate," Materials Science and Engineering: C, vol. A
403, pp. 25-31, 2005.
[165] F. Cordier, and E. Ollivier, "X‐ray photoelectron spectroscopy study of

aluminium surfaces prepared by anodizing processes," Surface and
interface analysis, vol. 23, pp. 601-608, 1995.
[166] W. C. Oliver, and George Mathews Pharr, "An improved technique for
determining hardness and elastic modulus using load and
displacement sensing indentation experiments," Journal of materials
research, vol. 7, pp. 1564-1583, 1992.
[167] Z. Guo, Xiaolu Pang, Yu Yan, Kewei Gao, Alex A. Volinsky, and Tong-
Yi Zhang, "CoCrMo alloy for orthopedic implant application enhanced
corrosion and tribocorrosion properties by nitrogen ion implantation,"
Applied surface science, vol. 347, pp. 23-34, 2015.
[168] M. Bryant, R. Farrar, R. Freeman, K. Brummitt, and A. Neville, "The

role of surface pre-treatment on the microstructure, corrosion and
fretting corrosion of cemented femoral stems," Biotribology, vol. 5, pp.
1-15, 2016.
[169] S. Kurosu, Naoyuki Nomura, and Akihiko Chiba, "Effect of sigma phase
in Co-29Cr-6Mo alloy on corrosion behavior in saline solution,"
Materials transactions, vol. 47, pp. 1961-1964, 2006.
325
[170] T. Zhu, and Miaoquan Li, "Effect of 0.770 wt% H addition on the
microstructure of Ti–6Al–4V alloy and mechanism of δ hydride
formation," Journal of Alloys and Compounds, vol. 481, pp. 480-485,
2009.
[171] A. S. E. J. McHargue C.J., Hammond J.P, "Aluminum Titanium - ICS

Database," Trans. Am. Inst. Min. Metall. Pet. Eng, vol. 197, 1953.
[172] R. J. Wasilewski, "Titanium - ICS Database," Trans. Met. Soc. AIME,

vol. 221,, 1961.
[173] S. Suwarno, Solberg, J.K., Maehlen, J.P., Krogh, B., Yartys, V.A.,
"Titanium Vanadium - ICS Database," Int. J. Hydrogen Energy, vol. 37,
2012.
[174] M. Jenko, Matevž Gorenšek, Matjaž Godec, Maxinne Hodnik, Barbara

Šetina Batič, Črtomir Donik, John T. Grant, and Drago Dolinar, "Surface
chemistry and microstructure of metallic biomaterials for hip and knee
endoprostheses," Applied Surface Science, 2017.
[175] J. R. Goldberg, and Jeremy L. Gilbert, "Electrochemical response of

CoCrMo to high-speed fracture of its metal oxide using an
electrochemical scratch test method," Journal of biomedical materials
research, vol. 37, 1997.
[176] G. Manivasagam, Durgalakshmi Dhinasekaran, and Asokamani

Rajamanickam, "Biomedical implants: Corrosion and its prevention-a
review," Recent patents on corrosion science, vol. 2, pp. 40-54, 2010.
[177] R. Büscher, and A. Fischer "The pathways of dynamic recrystallization

in all-metal hip joints," Wear vol. 259, pp. 887-897, 2005.
[178] M. Textor, Caroline Sittig, Vinzenz Frauchiger, Samuele Tosatti, and

Donald M. Brunette, "Properties and biological significance of natural
oxide films on titanium and its alloys," in Titanium in medicine, ed.
Berlin Heidelberg: Springer 2001, pp. 171-230.
326
[179] J. Vaithilingam, Elisabetta Prina, Ruth D. Goodridge, Richard JM
Hague, Steve Edmondson, Felicity RAJ Rose, and Steven DR Christie,
"Surface chemistry of Ti6Al4V components fabricated using selective
laser melting for biomedical applications," Materials Science and
Engineering, vol. C 67, pp. 294-303, 2016.
[180] I. Milošev, and H-H. Strehblow, "The composition of the surface

passive film formed on CoCrMo alloy in simulated physiological
solution," Electrochimica Acta, vol. 48, pp. 2767-2774, 2003.
[181] M. Metikoš‐Huković, and Mihajlo Ceraj‐Cerić, "p‐Type and n‐Type

Behavior of Chromium Oxide as a Function of the Applied Potential,"
Journal of the Electrochemical Society, vol. 134, pp. 2193-2197, 1987.
[182] N. McIntyre, D. Johnston, L. Coatsworth, R. Davidson, and J. Brown,

"X‐ray photoelectron spectroscopic studies of thin film oxides of cobalt
and molybdenum," Surface and interface analysis, vol. 15, pp. 265-
272, 1990.
[183] T. R. Beck, "Electrochemistry of freshly-generated titanium surfaces—

I. Scraped-rotating-disk experiments," Electrochimica Acta, vol. 18, pp.
807-814, 1973.
[184] J. R. Tavera, DY Peña Ballesteros, and HA Estupiñán Duran,

"Evaluation of the stiffness and friction of Ti6Al4V ELI treated by glow
discharge nitriding," In Journal of Physics: Conference Series, vol. 687,
p. 012021, 2016.
[185] R. Liu, Xiaoying Li, Xiao Hu, and Hanshan Dong, "Surface modification
of a medical grade Co‐Cr‐Mo alloy by low-temperature plasma surface
alloying with nitrogen and carbon," Surface and Coatings Technology,
vol. 232, pp. 906-911, 2013.
[186] D. L. Roll, "Passivation and the Passive Layer," ASTRO PAK.
327
[187] K. Hashimoto, K. Asami, M. Naka, and T. Masumoto, "The role of
alloying elements in improving the corrosion resistance of amorphous
iron base alloys," Corrosion Science, vol. 19, pp. 857-867, 1979.
[188] M. Metikos-Huković, Ana Kwokal, and Jasenka Piljac, "The influence

of niobium and vanadium on passivity of titanium-based implants in
physiological solution," Biomaterials, vol. 24, pp. 3765-3775, 2003.
[189] S. M. Bhola, and Brajendra Mishra, "Effect of pH on the

Electrochemical Properties of Oxides formed over β–Ti-15Mo and
Mixed Ti-6Al-4V Alloys," Int. J. Electrochem Sci, vol. 8, pp. 7075-7087,
2013.
[190] D. Mareci, Daniel Sutiman, Adrian Cailean, and Igor Creţescu,

"EFFECT OF VANADIUM REPLACEMENT BY ZIRCONIUM ON THE
ELECTROCHEMICAL BEHAVIOR OF Ti6Al4V ALLOY IN RINGERS
SOLUTION," Environmental Engineering and Management Journal,
vol. 7, pp. 701-706, 2008.
[191] N. J. Hallab, Shelley Anderson, Marco Caicedo, Amee Brasher, Katalin

Mikecz, and Joshua J. Jacobs, "Effects of soluble metals on human
peri‐implant cells," Journal of Biomedical Materials Research Part A,
vol. 74, pp. 124-140, 2005.
[192] S. Fouvry, C. Paulin, and T. Liskiewicz, "Application of an energy wear

approach to quantify fretting contact durability: Introduction of a wear
energy capacity concept," Tribology International, vol. 40, pp. 1428-
1440, 2007.
[193] M. Bryant, R. Farrar, K. Brummitt, R. Freeman, and A. Neville, "Fretting

corrosion of fully cemented polished collarless tapered stems: The
influence of PMMA bone cement," Wear, vol. 301, pp. 290-299, 2013.
[194] Y. Yan, A. Neville, and D. Dowson, "Biotribocorrosion—an appraisal of

the time dependence of wear and corrosion interactions: II. Surface
analysis," Journal of Physics D: applied physics, vol. 39, p. 3206, 2006.
328
[195] A. C. Fernandes, F. Vaz, E. Ariza, L. A. Rocha, A. R. L. Ribeiro, A. C.
Vieira, J. P. Rivière, and L. Pichon. , "Tribocorrosion behaviour of
plasma nitrided and plasma nitrided+ oxidised Ti6Al4V alloy," Surface
and Coatings Technology, vol. 200, pp. 6218-6224, 2006.
[196] M. T. Mathew, Valentim A. Barão, Judy Chia-Chun Yuan, Wirley G.

Assunção, Cortino Sukotjo, and Markus A. Wimmer, "What is the role
of lipopolysaccharide on the tribocorrosive behavior of titanium?,"
Journal of the mechanical behavior of biomedical materials, vol. 8, pp.
71-85, 2012.
[197] P. Zeng, J. Sharp, A. Rana, R. Thompson, W. M. Rainforth, and R. B.

Cook, "Sub-surface characterisation of tribological contact zone of
metal hip prostheses," In Journal of Physics: Conference Series, vol.
644, p. 012029, 2015.
[198] R. Pourzal, R. Theissmann, M. Morlock, and A. Fischer, "Micro-

structural alterations within different areas of articulating surfaces of a
metal-on-metal hip resurfacing system," 267, vol. Wear, pp. 689-694,
2009.
[199] D. J. Hall, Robin Pourzal, Hannah J. Lundberg, Mathew T. Mathew,

Joshua J. Jacobs, and Robert M. Urban, "Mechanical, chemical and
biological damage modes within head‐neck tapers of CoCrMo and
Ti6Al4V contemporary hip replacements," Journal of Biomedical
Materials Research Part B, vol. Applied Biomaterials, 2017.
[200] A. Di Laura, Quinn, P. D., Panagiotopoulou, V. C., Hothi, H. S.,

Henckel, J., Powell, J. J., … Hart, A. J. , "The Chemical Form of Metal
Species Released from Corroded Taper Junctions of Hip Implants:
Synchrotron Analysis of Patient Tissue," Scientific Reports, vol. 7,
2017.
[201] S. B. Kocagoz, Underwood, R. J., MacDonald, D. W., Gilbert, J. L., &

Kurtz, S. M., "Ceramic Heads Decrease Metal Release Caused by
329
Head-taper Fretting and Corrosion," Clinical Orthopaedics and Related
Research, vol. 474, pp. 985-994, 2016.
[202] S. R. Pearson, P. H. Shipway, J. O. Abere, and R. A. A. Hewitt, "The

effect of temperature on wear and friction of a high strength steel in
fretting," Wear, vol. 303, pp. 622-631, 2013.
[203] J. Vižintin, M. Kalin, S. Novak, G. Dražič, L. K. Ives, and M. B. Peterson,

"Effect of slip amplitude on the fretting wear of silicon nitride against
silicon nitride," Wear, vol. 192, pp. 11-20, 1996.
[204] M. R. Alexander, R. D. Short, F. R. Jones, W. Michaeli, and C. J.

Blomfield, "A study of HMDSO/O 2 plasma deposits using a high-
sensitivity and-energy resolution XPS instrument: curve fitting of the Si
2p core level," Applied Surface Science, vol. 137, pp. 179-183, 1999.
[205] B. J. McEntire, Ramaswamy Lakshminarayanan, Darin A. Ray, Ian C.

Clarke, Leonardo Puppulin, and Giuseppe Pezzotti, "Silicon Nitride
Bearings for Total Joint Arthroplasty," Lubricants, vol. 4, p. 35, 2016.
[206] S. Novak, G. Dražič, M. Kalin, and J. Vižintin, "Interactions in silicon

nitride ceramics vs. steel contact under fretting conditions," Wear, vol.
225, pp. 1276-1283, 1999.
[207] J. Qu, Peter J. Blau, Thomas R. Watkins, Odis B. Cavin, and Nagraj S.
Kulkarni, "Friction and wear of titanium alloys sliding against metal,
polymer, and ceramic counterfaces," Wear, vol. 258, pp. 1348-1356,
2005.
[208] C. H. Hager Jr, J. H. Sanders, and S. Sharma, "Characterization of

mixed and gross slip fretting wear regimes in Ti6Al4V interfaces at
room temperature," Wear, vol. 257, pp. 167-180, 2004.
[209] K. G. Budinski, "Tribological properties of titanium alloys," Wear, vol.

151, pp. 203-217, 1991.
[210] J. S. Kawalec, Stanley A. Brown, Joe H. Payer, and Katharine Merritt,

"Mixed‐metal fretting corrosion of Ti6Al4V and wrought cobalt alloy,"
330
Journal of Biomedical Materials Research Part A, vol. 29, pp. 867-873,
1995.
[211] Titanium in medicine: material science, surface science, engineering,

biological responses and medical applications: Springer Science &
Business Media, 2012.
[212] J. C. Scully, Ed., Corrosion: Aqueous Processes and Passive Films

(Treatise on Materials Science and Technology. 1983, p.^pp. Pages.
[213] S. W. Watson, F. J. Friedersdorf, B. W. Madsen, and S. D. Cramer,

"Methods of measuring wear-corrosion synergism," Wear, vol. 181, pp.
476-484, 1995.
[214] H. Uhlig, "Mechanism of fretting corrosion," J Appl Mech, vol. 21, pp.
401–7, 1954.
[215] Y. Yan, Anne Neville, and Duncan Dowson, "Tribo-corrosion properties

of cobalt-based medical implant alloys in simulated biological
environments," Wear, vol. 263, pp. 1105-1111, 2007.
[216] H. S. Hothi, Andreas C. Panagiotopoulos, Robert K. Whittaker, Paul J.

Bills, Rebecca A. McMillan, John A. Skinner, and Alister J. Hart,
"Damage patterns at the head-stem taper junction helps understand
the mechanisms of material loss," The Journal of arthroplasty, vol. 32,
pp. 291-295, 2017.
[217] J. M. M. Anna Panagiotidou, Osman Khabab, Alister Hart, John

Skinner, Fares Haddad and Gordon Blunn, "The effect of diamond like
carbon coating on fretting corrosion and wear at the modular head-
stem junction of metal on metal total hip replacements," presented at
the 10th World Biomaterials Congress, Montréal, Canada, 2016.
[218] J. P. Kretzer, Eike Jakubowitz, Michael Krachler, Marc Thomsen, and

Christian Heisel, "Metal release and corrosion effects of modular neck
total hip arthroplasty," International orthopaedics, vol. 33, p. 1531, Dec
2009.
331
[219] S. B. Kocagöz, Richard J. Underwood, Shiril Sivan, Jeremy L. Gilbert,
Daniel W. MacDonald, Judd S. Day, and Steven M. Kurtz, "Does taper
angle clearance influence fretting and corrosion damage at the head–
stem interface? A matched cohort retrieval study," Seminars in
arthroplasty, vol. 24, pp. 246-254, 2013.
[220] S. Y. Jauch, G. Huber, E. Hoenig, M. Baxmann, T. M. Grupp, and M.

M. Morlock, "Influence of material coupling and assembly condition on
the magnitude of micromotion at the stem–neck interface of a modular
hip endoprosthesis," Journal of biomechanics, vol. 44, pp. 1747-1751,
2011.
[221] S. Fouvry, P. Duo, and Ph Perruchaut, "A quantitative approach of Ti–

6Al–4V fretting damage: friction, wear and crack nucleation," Wear, vol.
257, pp. 916-929, 2004.
[222] M. Kiran, & Boscainos, "Adverse reactions to metal debris in metal-on-

polyethylene total hip arthroplasty using a titanium-molybdenum-
zirconium-iron alloy stem," The Journal of arthroplasty, vol. 30, pp. 277-
281, 2015.
[223] R. K. Whittaker, Ahmed M. Zaghloul, Harry S. Hothi, Imran A. Siddiqui,

Gordon W. Blunn, John A. Skinner, and Alister J. Hart, "Clinical Cold
Welding of the Modular Total Hip Arthroplasty Prosthesis," The Journal
of arthroplasty, vol. 32, pp. 610-615, 2017.
[224] C. R. Fraitzl, Luis E. Moya, Lorenzo Castellani, Timothy M. Wright, and

Robert L. Buly, "Corrosion at the stem-sleeve interface of a modular
titanium alloy femoral component as a reason for impaired
disengagement," The Journal of arthroplasty, vol. 26, pp. 113-119,
2011.
[225] N. A. Nassif, Danyal H. Nawabi, Kirsten Stoner, Marcella Elpers,

Timothy Wright, and Douglas E. Padgett, "Taper design affects failure
of large-head metal-on-metal total hip replacements," Clinical
Orthopaedics and Related Research®, vol. 472, pp. 564-571, 2014.
332
[226] E. Sauger, S. Fouvry, L. Ponsonnet, Ph Kapsa, J. M. Martin, and L.
Vincent., "Tribologically transformed structure in fretting," Wear, vol.
245, pp. 39-52, 2000.
[227] E. E. Hoffman, Alex Lin, Yifeng Liao, and Laurence D. Marks, "Grain
Boundary Assisted Crevice Corrosion in CoCrMo Alloys," Corrosion,
vol. 72, pp. 1445-1461, 2016.
[228] J. M. West, "Basic corrosion and oxidation," 1986.
[229] H. Tomizawa, and T. E. Fischer, "Friction and wear of silicon nitride

and silicon carbide in water: hydrodynamic lubrication at low sliding
speed obtained by tribochemical wear," ASLE transactions, vol. 30, pp.
41-46, 1987.
[230] G. Milazzo, Sergio Caroli, and Robert D. Braun, "Tables of standard

electrode potentials," Journal of The Electrochemical Society, vol. 125,
pp. 261C-261C, 1978.
[231] G. Gstraunthaler, and Toni Lindl. Zell-und Gewebekultur, allgemeine

Grundlagen und spezielle Anwendungen: Springer-Verlag, 2013.
333
Appendix A– Composition of FBS and PBS
Figures 11-1 – Composition of FBS [231]
334
Figures 11-2 – Composition of PBS
335
Appendix B– Determination of contact compliance
For a fretting displacement amplitude δx imposed upon a tribo-couple, a
proportion of this displacement is not translated into slip at the contact, rather,
some of the displacement is expended in two main form of compliances
namely: Contact compliance (Cc) and System compliance (Cs). While Cc may
differ with varying material combinations, Cs varies in depending on the value
of Cc as both parameters together make up the total recorded compliance (Cr).
For example, Cc is a function of the contact stiffness ‘K’ and the Hertzian
contact radius ‘a’ of the tribo-couple as expressed by Fouvry et al [149] and
shown in Equation 11-1.
2𝐾 11-1
𝐶𝑐 (𝐹𝑡 ) =
3𝑎
Cs on the other hand is a parameter related to the cantilever deformation of
the fretting tribometer. It is determined only by firstly measuring the recorded
compliance Cr which is obtainable directly from the quasi-stick fretting loops
of the material combinations in subject. Figures 11-3a and b are the fretting
loops at δx = ±10 µm for CoCrMo – CoCrMo and CoCrMo – Ti6Al4V
respectively. Errors in the determination of Cr can be minimised by using
quasi-stick fretting loop because the maximum tangential force ‘𝐹𝑡 ’ occurs at
the maximum displacement δx. In addition, δx ≅ δe where δe is the elastic
(reversible) part of the displacement and δs is the irreversible, dissipative slip
incurred at the tribo-couple interface. Cr is determined using Equation 11-2.
∆𝜕𝑥
𝐶𝑟 (𝐹𝑡 ) = 11-2
∆𝑄
Cs is determined from the difference between Cr and Cc as expressed in

Equation 11-3.
𝐶𝑠 = 𝐶𝑟 (𝐹𝑡 ) − 𝐶𝑐 (𝐹𝑡 ) 11-3
The calculated values are outlined in Table 4-1 along with the Hertzian contact
radius and the reduced modulus. The results show that Hertzian contact
336
radius of CoCrMo – Ti6Al4V is ~ 36% larger than it is for CoCrMo – CoCrMo
while the reduced modulus values also differ by ~32%, and CoCrMo –
CoCrMo is the larger value. The contact compliance at the contact of both
combinations were ~ 30 nm/N and relatively similar system compliances with
values of 96.1% and 97.6% for CoCrMo – CoCrMo and CoCrMo – Ti6Al4V
respectively. The system compliance, reported as a percentage ratio of the
total compliance was measured at the quasi-stick fretting loop of both material
combinations.
(Ft (N), δx (µm)) (𝐹𝑡 (N), δx(µm))

(12.5, 10)
(15.8, 10)
δe
Displacemen
t due to slip
(δs)
CoCrMo – CoCrMo CoCrMo – Ti6Al4V

(a) (b)
Figures 11-3 – Quasi-stick fretting loop of (a) CoCrMo – CoCrMo and (b)
CoCrMo – Ti6Al4V tribo-couples
337
Table B 1 – Tribological parameters of the Metal – Metal fretting contact
CoCrMo – CoCrMo CoCrMo – Ti6Al4V
Hertzian Contact
160 250
Radius ‘a’ (µm)
Reduced Modulus
123 84
(GPa)
Contact Compliance
0.0315 0.0293
‘Cc’ (µm/N)
System Compliance ‘Cs’

96.1 97.6
(%)
Table B 2– Tribological parameters of the Ceramic (Si3N4) – Metal and

Ceramic (Si3N4) – Ceramic (Si3N4) fretting contact
Si3N4 – CoCrMo Si3N4 – Ti6Al4V Si3N4 – Si3N4
Hertzian Contact Radius

130 200 120
‘a’ (µm)
Reduced Modulus (GPa) 143 94 171
Contact Compliance ‘Cc’

0.0325 0.0325 0.0285
(µm/N)

96.6 95.0 95.0
(%)
338
Table B 3 – Tribological parameters of the Ceramic (Biolox) – Metal
fretting contact
Biolox – Biolox – Ti6Al4V Biolox – Ti6Al4V

CoCrMo
(1 GPa) (0.77 GPa)
(1 GPa)
Hertzian Contact
140 230 180
Radius ‘a’ (µm)
Reduced Modulus
147 97 97
(GPa)
Contact Compliance
0.0291 0.0276 0.0353
‘Cc’ (µm/N)

97.0 96.5 93.7
(%)
339
Appendix C – Interactions between energy and wear
Figure11-4– Wear-Energy plot of metal – metal and ceramic – metal

material combinations.
Figure 11-5 – Average wear volume of metal – metal and ceramic metal
material combinations. The contribution of the ball component is
shown above and the contribution of the flat component beneath.
340
Appendix D – List of oral presentations
1. Fretting corrosion regime and wear mechanism of CoCrMo and

Ti6Al4V
• {Leeds – Lyon Tribology Conference, Leeds, UK} 2014
2. In-vitro studies of fretting corrosion at the modular THR taper junction:

st
The effect of mixed metal couples 1 Prize (poster) Award!
• {Corrosion Science Symposium, Manchester, UK} 2014
3. Fretting of CoCrMo and Ti6Al4V alloys in modular prostheses

• {Micro-tribology Symposium, Warsaw, Poland} 2014
4. Subsurface investigation of fretted CoCrMo and Ti6Al4V alloys

st
1 Prize (oral) Award!
• {International Society of Technology and Arthroplasty, Vienna,
Austria} 2015
• {TriboUK Conference, Leeds, UK} 2016
5. Effect of tribocorrosion products in metal – metal modular fretting

interfaces
• {Leeds – Lyon Tribology Conference, Leeds, UK} 2016
6. The effect of cyclic load on the evolution of fretting current at the

interface of metal-metal and ceramic – metal taper junction of hip
prosthesis
• {International Society of Technology and Arthroplasty, Boston,
USA} 2016
7. The influence of surface history on the evolution of fretting corrosion

current
• {International Conference on Biotribology, London, UK} 2016
8. Subsurface transformation of Ti6Al4V alloy under fretting conditions

IMechE Research Prize!
• {Mission of Tribology, London, UK} 2017
341

AbimbolaOladokun PhDThesis

Uploaded by

Copyright:

Available Formats

AbimbolaOladokun PhDThesis

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

AbimbolaOladokun PhDThesis

Uploaded by

Copyright:

Available Formats

MECHANISM OF FRETTING CORROSION AT THE MODULAR

TAPER INTERFACE OF HIP PROSTHESIS

Abimbola Oluwawemimo Oladokun

Submitted in accordance with the requirements for the degree of

The University of Leeds

© 2017 The University of Leeds and Abimbola Oluwawemimo Oladokun

 A. Oladokun, M. Pettersson, M. Bryant, H. Engqvist, C. Persson, R.

 Oladokun, A., M. Pettersson, M. Bryant, R. Hall, and A. Neville. "SUB-

 Oladokun, A., M. Bryant, R. Hall, and A. Neville. "The effect of cyclic

Special thanks to Dr. Todd Stewart who supervised my bachelors’ research

I am truly grateful to the iFS technicians that supported my research work:

Special thanks to my Pastors: Paul and Natasha Okhuoya who supported me

I am so grateful to my parents who sacrificed their jobs and settlement in

Subsequent to the failure of the implant, analysis of retrieved explants

The results showed that CoCrMo and Ti6Al4V display independent

The subsurface transformation in CoCrMo alloy when subjected to fretting was

It was also observed that the chemical composition of fretting corrosion

List of Figures ........................................................................................ xviii

List of Tables ........................................................................................ xxxiii

Nomenclature ....................................................................................... xxxiv

Chapter 1 Introduction ................................................................................ 1

1.1 Research objectives ...................................................................... 2

1.2 Thesis outline ................................................................................ 3

Chapter 2 Background to biomaterials, corrosion and tribology ........... 5

2.1 Introduction ................................................................................... 5

2.2 Metallic biomaterials ...................................................................... 5

2.2.1 Cobalt based Alloys.............................................................. 5

2.2.2 Titanium and Titanium based alloys ..................................... 7

2.2.3 Stainless steel ...................................................................... 9

2.3 Ceramic biomaterials .................................................................. 10

2.3.1 Alumina (Al2O3) and alumina composite ............................ 11

2.3.2 Silicon Nitride (Si3N4) ........................................................ 13

2.4.1Electrochemistry of corrosion .............................................. 14

2.4.2 Types of corrosion effective at the modular junction of

2.4.3Passivity of biomaterials ...................................................... 20

2.4.4 Faraday’s law ..................................................................... 21

2.5 Tribology ..................................................................................... 22

2.5.3 Abrasive wear..................................................................... 23

2.5.4 Fatigue Wear ...................................................................... 24

2.5.5 Corrosive wear ................................................................... 25

2.6 Fretting ........................................................................................ 26

2.6.1 Fretting regimes ................................................................. 27

2.6.2 Fretting corrosion, wear and fatigue ................................... 31

Chapter 3 Literature review ...................................................................... 33

3.1 Total hip arthroplasty ................................................................... 33

3.1.1 Rationale for modularity in THA.......................................... 34

3.1.2 The Morse taper (head-neck taper) .................................... 36

3.1.3 The neck–stem taper .......................................................... 37

3.1.4 Assembly of modular tapers ............................................... 37

3.2 Challenges with modularity ......................................................... 38

3.2.1 Fretting and corrosion scores in retrieval studies ............... 39

3.2.2 Review of retrieval studies – head-neck interface .............. 40

3.2.3 Review of retrieval studies – the neck-stem interface ........ 43

3.2.4 Corrosion products and metal ion releases ........................ 46

3.2.5 Metallurgical transformations in explants ........................... 49

3.3 In-vitro studies of fretting corrosion ............................................. 50

3.3.1 Common in-vitro methods for assessing fretting

3.3.2 Review of tribometer studies .............................................. 52

3.3.3 Review of in-vitro studies using realistic taper