dr.ntu.edu.sg. W. Khu… · Web viewSTRUCTURE AND ELECTROCHEMICAL BEHAVIOR OF NITROGEN DOPED...
Transcript of dr.ntu.edu.sg. W. Khu… · Web viewSTRUCTURE AND ELECTROCHEMICAL BEHAVIOR OF NITROGEN DOPED...
STRUCTURE AND ELECTROCHEMICAL BEHAVIOR OF NITROGEN
DOPED DIAMOND-LIKE CARBON THIN FILMS WITH OR WITHOUT
PLATINUM AND RUTHENIUM DOPING
KHUN NAY WIN
SCHOOL OF MECHANICAL AND AEROSPACE ENGINEERING
2011
STRUCTU
RE AND ELECTRO
CHEMICAL BEHAVIO
R OF N
ITROGEN
DOPED DIAM
ON
D-LIKE CARBON
THIN
FILMS W
ITH OR W
ITHOU
T PLATINU
M AN
D RUTHEN
IUM
DOPIN
G 2011
KHUN
NAY W
IN
STRUCTURE AND ELECTROCHEMICAL BEHAVIOR OF NITROGEN DOPED
DIAMOND-LIKE CARBON THIN FILMS WITH OR WITHOUT PLATINUM
AND RUTHENIUM DOPING
KHUN NAY WIN
School of Mechanical and Aerospace Engineering
A thesis submitted to the Nanyang Technological Universityin partial fulfillment of the requirement for the degree of
Doctor of Philosophy
2011
Abstract
Human activities have released toxic metals such as Zn, Pb, Cd, Cu and Hg, etc. into
the environment. Nowadays, the presence of toxic metals in the aquatic ecosystem implicates
directly to biota and indirectly to human beings. Therefore, fast detection and determination
of toxic metals in aqueous solutions are a tough challenge for analysts.
Diamond-like carbon (DLC) is a type of carbon, which consists of both sp2 (graphite
like) and sp3 (diamond like) bonds, and an environmentally friendly material. In addition,
DLC films can be produced at room temperature and achieve similar properties to those of
diamond films, so they have been explored as electrode materials for heavy metal tracing.
However, high electrical resistivity of DLC films has confined their electrochemical
applications. DLC films used as electrodes for electrochemical applications must be
conductive.
Nitrogen is an effective donor in DLC films because of its five valance electrons.
Therefore, nitrogen is used as a dopant for making conductive nitrogen doped diamond-like
carbon (N-DLC) films. However, an incorporation of nitrogen in DLC films lowers the
corrosion resistance of DLC films by degrading sp3-bonded cross-linking structure through
increased sp2 bonds though it can increase the electrical conductivity of the films. Poor
corrosion resistance of N-DLC films can affect the electrochemical performance of the films
such as sensitivity, long-time response stability, durability and repeatability.
It has been known that noble metals such as Pt and Ru are inherently corrosion
resistant and have outstanding catalytic properties. Pure Pt is much more expensive compared
to pure Ru, so alloying of Pt with Ru is an economical way to reduce the cost of the electrode.
Besides, Ru is also alloyed in Pt to keep good mechanical properties of the films due to its
i
ability to harden platinum. It has been reported that PtRu phases incorporated in an
amorphous carbon structure exist as nano aggregates. When these PtRu aggregates appear on
the surface of DLC films, they play a role of catalytic property which facilitates charge
transfer at electrode/solution interface. It is expected that incorporation of Pt/Ru/N in DLC
films may promote the electrical conductivity and electrochemical performance of the films
such as corrosion resistance and sensitivity for heavy metal tracing. Therefore, the
electrochemical behavior of nitrogen doped DLC (N-DLC) films with or without Pt and Ru
and its dependence on the film structure together with some other physical characteristics
were investigated in this research for practical application.
The corrosion behavior of the N-DLC films deposited on p-Si substrates using a
filtered cathodic vaccum arc (FCVA) deposition system by varying nitrogen flow rate was
investigated in 0.6 M NaCl solutions with both potentiodynamic polarization and immersion
tests at room temperature. The results revealed that the increased nitrogen content in the N-
DLC films degraded the corrosion resistance of the films. Potential windows of the N-DLC
films measured in 0.5 M HCl, 0.1 M KCl, 0.1 M NaCl, 0.1 M KOH, and 0.1 M NaOH were
2.4, 2.32, 3.2, 3.1 and 3.25 V, respectively. Though the N-DLC film electrodes offered (1)
wide potential windows in different types of solutions, (2) a very low and stable background
to improve the signal-to-background and signal-to-noise ratios, (3) the repeatability of
voltammograms, and (4) long-time response stability, their voltammograms were apparently
affected by their electrical conductivity, types of alkaline species and unbalanced H+ and OH-
ions. These film electrodes also provided a significant stripping response for determination of
single-element (Zn2+, Pb2+, Cu2+, and Hg2+) and multi-elements (Pb2+ + Cu2+ + Hg2+)
simultaneously in KCl solutions. It was observed that the sensitivity of the film electrodes to
the elements detected was apparently influenced by nitrogen concentration in the films,
ii
deposition time and potential, and concentration and type of elements. The simultaneous
analysis of heavy metals using linear sweep anodic stripping voltammetry (LSASV) produced
sharp and well-defined peaks with good peak separations. The novel N-DLC film electrodes
under development showed a great promise for the detection of heavy metals.
Pt/Ru/N doped DLC (PtRuN-DLC) films were deposited on p-Si substrates by a DC
magnetron sputtering deposition system by varying DC power applied to a Pt50Ru50 target.
The increased Pt and Ru incorporation in the PtRuN-DLC films improved the corrosion
resistance of the films in the NaCl solution at lower potential though more sp2 bonds were
formed in the films via metal-induced graphitization. However, the PtRuN-DLC films with
higher Pt/Ru content degraded early than the ones with lower Pt/Ru content at higher
potential. Noble metal incorporation appears to be a promising way to improve the corrosion
resistance of the N-DLC films.
Furthermore, N-DLC and PtRuN-DLC films were deposited on p-Si substrates using a
DC magnetron sputtering system under the same deposition conditions except Pt and Ru
doping for the PtRuN-DLC film in order to get more understanding of the effect of Pt and Ru
incorporation on the structural and electrochemical properties of the N-DLC films. Though
the N-DLC film electrodes showed wide potential windows in acidic solutions such as 0.1 M
H2SO4 and 0.1 M HCl and a neutral solution of 0.1 M KCl, it was found that the Pt and Ru
doping significantly narrowed down the potential windows of the N-DLC film electrodes in
these solutions due to their catalytic activities. The N-DLC film electrodes showed a good
electrocatalytic activity in Fe(CN)64-/Fe(CN)6
3- redox reactions. However, an increased kinetic
limitation upon the PtRuN-DLC film electrodes with Pt and Ru doping shifted the oxidization
peak to a more positive value and the reduction peak to a more negative value compared to
those obtained from the N-DLC film electrodes. It could be deduced that the introduction of
iii
Pt and Ru into the N-DLC films improved the corrosion resistance of the films but
significantly degraded the electrochemical behavior of the films.
iv
Acknowledgements
I would like to express my sincere appreciation and gratitude to my supervisor, Assoc.
Prof. Liu Erjia, for his invaluable advice and encouragement throughout the duration of this
project. He has not only provided enthusiasm and support but also imparted his personal
wisdom that will last forever. Without his help, the project would be impossible to accomplish
and the attainment would be much compromised. Co-Supervisor, Dr. Zeng Xianting, is also
gratefully acknowledged.
My sincere thanks would be extended to Research student, Wang Huili, for his
valuable discussion in the research. I would like to show my special thanks to the technicians
in Materials Lab 1 and Fuel cell Lab, School of Mechanical and Aerospace Engineering,
Nanyang Technological University, for their technical assistance and support.
I sincerely thank Prof. Sam Zhang and his students for their assistance for my sample
preparation and analysis. I would like to greatly thank Assoc. Prof. Jiang San Ping for his
PtRu alloy target and analytical instruments used for my research work.
I would like to thank my parents: Mr. Nay Win and Dr. Khin Myint Myint, for their
love and support to my research. My deep gratitude also goes to my parents-in-law: Mr. Zaw
Win and Madam Mya Mya Thin for their care and support to my family during my PhD
study. This thesis is dedicated to my wife Madam Zar Chi Win and my daughter Shin Thant
Mon.
Thank also goes to Nanyang Technological University, Singapore for providing a PhD
scholarship to this research. The financial support from the Environment & Water Industry
Development Council (EWI), Singapore is gratefully acknowledged.
v
Last but not least, I would like to thank my friends and many others who have in one
way or another contributed to the completion of the work.
vi
List of Publications
Journal papers
N. W. Khun, E. Liu, “Enhancement of adhesion strength and corrosion resistance of nitrogen
or platinum/ruthenium/nitrogen doped diamond-like carbon thin films by platinum/ruthenium
layer”, Diamond and Related Materials, 19 (2010) 1065.
N. W. Khun, E. Liu, “Effect of platinum and ruthenium incorporation on voltammetric
behavior of nitrogen doped diamond-like carbon thin films”, Electroanalysis, 21 (2009)
2590.
N. W. Khun, E. Liu, G. C. Yang, W. G. Ma, S. P. Jiang, “Structure and corrosion behavior of
platinum/ruthenium/nitrogen doped diamond-like carbon thin films”, Journal of Applied
Physics, 106 (2009) 013506.
N. W. Khun, E. Liu, X. T. Zeng, “Corrosion behavior of nitrogen doped diamond-like carbon
thin films in NaCl solutions”, Corrosion Science, 51 (2009) 2158.
N. W. Khun, E. Liu, “Linear sweep anodic stripping voltammetry of heavy metals from
nitrogenated tetrahedral amorphous carbon thin films”, Electrochimica Acta, 54 (2009)
2890.
N. W. Khun, E. Liu, “Cyclic voltammetric behavior of nitrogen doped tetrahedral amorphous
carbon films deposited by filtered cathodic vacuum arc”, Electroanalysis, 20 (2008) 1851.
vii
Patent
E. Liu, W. G. Ma, A. P. Liu, G. C. Yang, N. W. Khun, Z. M. Wang, “Diamond-like carbon
based microelectrode sensor and associated electrochemical cell for detection of trace heavy
and toxic metals in aqueous solutions”, US Patent (12/842,857).
viii
Table of Contents
Page
Abstract i
Acknowledgements v
List of Publications vii
Table of Contents ix
List of Nomenclatures xv
List of Figures xvi
List of Tables xxiii
Chapter 1 Introduction 1
1.1. Synopsis 1
1.2. Objective 7
1.3. Scope 7
1.4. Organization 8
Chapter 2 Literature review 10
2.1. Carbon 10
2.1.1. Diamond 11
2.1.2. Graphite 12
2.1.3. Diamond films 13
2.1.4. Diamond-like carbon (DLC) 13
2.2. Doping of DLC films 16
2.3. Deposition methods of DLC films 21
2.3.1. Magnetron sputtering deposition 22
2.3.2. Cathodic arc 24
ix
2.3.3. Pulsed laser deposition (PLD) 25
2.4. Surface morphological characteristics of DLC films 26
2.5. Adhesion strength of DLC films 28
2.6. Electrochemistry 31
2.6.1. Corrosion mechanisms 31
2.6.1.1. Linear polarization 34
2.6.1.2. Potentiodynamic polarization 35
2.6.1.3. Tafel plot 36
2.6.1.4. Electrochemical impedance spectroscopy 37
2.6.2. Corrosion properties of DLC films 41
2.7. Electrode materials for electroanalysis 44
2.8. Electrochemistry of DLC films 49
2.8.1. Cyclic voltammetry 52
2.8.2. Anodic stripping voltammetry 55
2.8.3. Stripping analysis of DLC films 56
2.8.4. Potential applications of DLC films as electrodes for electroanalysis 59
2.9. Summary 61
Chapter 3 Experimental details 63
3.1. Sample preparation 63
3.2. Characterization 65
3.2.1. Film structure 65
3.2.2. Film surface activity, morphology and topography 67
3.2.3. Adhesion strength of the film to p-Si substrate 67
3.3. Electrochemical measurement 68
x
3.3.1. Sample preparation 68
3.3.2. Electrochemical workstation 68
3.3.3. Setup of electrochemical cell 69
3.3.4. Potentiodynamic polarization test 70
3.3.5. Immersion test 70
3.3.6. Electrochemical impedance spectroscopy 71
3.3.7. Cyclic voltammetry 71
3.3.8. Anodic stripping voltammetry 72
3.4. Summary 72
Chapter 4 Structural and Electrochemical Properties of Nitrogen Doped Diamond-
like Carbon Thin Films 74
4.1. Introduction 74
4.2. Structural properties of N-DLC thin films 76
4.2.1. Chemical composition of N-DLC thin films 76
4.2.2. Raman results of N-DLC thin films 78
4.2.3. Surface morphology of N-DLC thin films 79
4.2.4. Adhesion strength of N-DLC thin films 80
4.3. Electrochemical performance of N-DLC thin films 83
4.3.1. Corrosion behavior of N-DLC thin films 83
4.3.1.1. Potentiodynamic polarization results of N-DLC thin films 83
4.3.1.2. Immersion results of N-DLC thin films 88
4.3.2. Linear sweep cyclic voltammetric behavior of N-DLC thin films 92
4.3.2.1. Cyclic voltammetry of N-DLC thin films in acidic solution 92
4.3.2.2. Cyclic voltammetry of N-DLC thin films in neutral solutions 94
xi
4.3.2.3. Cyclic voltammetry of N-DLC thin films in hydroxide solutions 96
4.3.2.4. Cyclic voltammetry of reversible couple (Ferricyanide) 98
4.3.3. Linear sweep anodic stripping voltammetric behavior of N-DLC thin films 99
4.3.3.1. Linear sweep anodic stripping voltammograms of Lead 99
4.3.3.2. Linear Sweep Anodic Stripping Voltammograms of Zinc and Lead 106
4.3.3.3. Linear Sweep Anodic Stripping Voltammograms of Copper 107
4.3.3.4. Linear Sweep Anodic Stripping Voltammograms of Mercury 109
4.3.3.5. Linear Sweep Anodic Stripping Voltammograms of Simultaneous Lead, Copper
and Mercury 111
4.3.3.6. Concentration effect for the ions traced by LSASV 113
4.4. Conclusions 115
Chapter 5 Structural and Electrochemical Properties of Platinum/ruthenium/nitrogen
Doped Diamond-like Carbon Thin Films 118
5.1. Introduction 118
5.2. Structural properties of PtRuN-DLC thin films 119
5.2.1. Chemical composition of PtRuN-DLC thin films 119
5.2.2. Microstructure of PtRuN-DLC thin films 121
5.2.3. XPS results of PtRuN-DLC thin films 122
5.2.4. Raman results of PtRuN-DLC thin films 126
5.2.5. Surface activity of PtRuN-DLC thin films 129
5.2.6. Surface morphology of PtRuN-DLC thin films 131
5.2.7. Adhesion strength of PtRuN-DLC thin films 131
5.3. Electrochemical performance of PtRuN-DLC thin films 133
5.3.1. Corrosion behavior of PtRuN-DLC thin films 133
xii
5.4. Conclusions 138
Chapter 6 Comparative Study of Structural and Electrochemical Properties of
Nitrogen-doped and Platinum/ruthenium/nitrogen-doped Diamond-like Carbon Thin
Films 140
6.1. Introduction 140
6.2. Structureal properties of N-DLC and PtRuN-DLC thin films 141
6.2.1. Chemical composition of N-DLC and PtRuN-DLC thin films 141
6.2.2. Microstructure of PtRuN-DLC thin films 142
6.2.3. Chemical bonding structure of N-DLC and PtRuN-DLC thin films measured by
XPS 142
6.2.4. Chemical bonding structure of N-DLC and PtRuN-DLC thin films measured by
Raman spectroscopy 144
6.2.5. Surface activity of N-DLC and PtRuN-DLC thin films 145
6.2.6. Surface morphology of N-DLC and PtRuN-DLC thin films 146
6.2.7. Adhesion strength of N-DLC and PtRuN-DLC thin films 147
6.3. Electrochemical performance of N-DLC and PtRuN-DLC thin films 149
6.3.1. Corrosion behavior of N-DLC and PtRuN-DLC thin films 149
6.3.2. Linear sweep cyclic voltammetric behavior of N-DLC and PtRuN-DLC thin
films 151
6.3.2.1. Cyclic voltammetry of N-DLC and PtRuN-DLC thin films 151
6.3.2.2. Cyclic voltammetry of reversible couple (Ferricyanide) 156
6.4. Conclusions 158
Chapter 7 Conclusions 160
7.1. Conclusions on N-DLC films prepared by a FCVA technique 160
xiii
7.2. Conclusions on PtRuN-DLC films prepared by a DC magnetron sputtering technique
162
7.3. Conclusions on N-DLC and PtRuN-DLC films prepared by a DC magnetron sputtering
Technique 163
7.4. Recommendations for future work 165
References 166
xiv
List of Nomenclatures
a-C Amorphous carbon
AFM Atomic force microscopy
ASV Anodic stripping voltammetry
BDD Boron doped diamond
CV Cyclic voltammetry
CVD Chemical vapor deposition
DC Direct current
DLC Diamond-like carbon
N-DLC Nitrogen doped diamond-like carbon
DPASV Differential pulse anodic stripping voltammetry
EAS Electrochemically active species
FCVA Filtered cathodic vacuum arc
FWHM Full width at half maximum
LSASV Linear sweep anodic stripping voltammetry
PtRuN-DLC Platinum/ruthenium/nitrogen doped diamond-like carbon
OCP Open circuit potential
PLD Pulsed laser deposition
PVD Physical vapor deposition
SCE Standard calomel electrode
SEM Scanning electron microscopy
ta-C Tetrahedral amorphous carbon
XPS X-ray photoelectron spectroscopy
xv
List of Figures
Page
Fig. 2.1: The sp3, sp2, sp1 hybridized bonding. 10
Fig. 2.2: Ideal diamond structure. 11
Fig. 2.3: Crystal structure of graphite. 12
Fig. 2.4: Structure of amorphous carbon. 14
Fig. 2.5: Ternary phase diagram of bonding in amorphous carbon-hydrogen alloys. 15
Fig. 2.6: Schematic DOS of a carbon showing σ and π states. 17
Fig. 2.7: Schematic of the levels of substitutional nitrogen in diamond and ta-C. 18
Fig. 2.8: Various nitrogen configurations in DLC, showing the doping configuration. One dot
means an unpaired electron. Two dots mean a lone pair (non-bonding). 19
Fig. 2.9: Diagram showing how nitrogen modifies bonding in graphitic layer, by introducing
five-membered rings and warping, or inter-layer bonding. 20
Fig. 2.10: General feature of sputter coater. 23
Fig. 2.11: FCVA source design. 24
Fig. 2.12: Configuration of a PLD chamber. 26
Fig. 2.13: SEM images of ta-C films deposited at various substrate biases: (a) 0 V, (b) 250 V,
(c) 300 V and (d) 400 V. 27
Fig. 2.14: Electrical double layer for electrode submerged in an electrolyte. 32
Fig. 2.15: Cyclic polarization curve. 33
Fig. 2.16: Linear polarization. 34
Fig. 2.17: Potentiodynamic polarization curve. 35
Fig. 2.18: Tafel plot. 37
xvi
Fig.2.19: (a) An equivalent circuit proposed for a porous free coating. (b) Nyquist plot and (c)
Bode magnitude and (d) phase plots of the coating. 39
Fig.2.20: (a) An equivalent circuit proposed for a coating with porosities. (b) Nyquist plot and
(c) Bode magnitude and (d) phase plots of the coating. 41
Fig. 2.21: Cyclic voltammetric i-E curves for boron doped diamond electrodes in (a) 1 M KCl
and (b) Fe(CN)6-3/-4 + 1 M KCl. 46
Fig. 2.22: Cyclic voltemmetric i-E curves obtained (a) N-DLC electrode (dash line) and
glassy carbon electrode (solid line) in 1 mM Fe(CN)6/0.1 M H2SO4 and (b) N-DLC electrode
in 0.5 M H2SO4. 48
Fig. 2.23: Bright-field transmission electron micrograph of a Pt-DLC composite film.
Platinum nanoparticles self-assemble into arrays (dark regions) within the DLC matrix (light
regions). 50
Fig. 2.24: Typical cyclic voltammogram. 53
Fig. 2.25: Typical DPASVs obtained at N-DLC film electrode in a 0.1 M KCl (pH 1.0)
solution. 57
Fig. 3.1: Schematic configuration of a FCVA deposition system. 63
Fig. 3.2: Schematic configuration of a magnetron sputtering system. 64
Fig. 3.3: Basic feature of a micro-scratch tester. 68
Fig. 3.4: (a) Size of a film-coated sample and (b) film-coated samples. 68
Fig. 3.5: (a) Schematic configuration and (b) outlook of a three-electrode electrochemical
cell. 69
Fig. 4.1: Fitted XPS C 1s spectra of N-DLC films deposited with nitrogen flow rates of (a) 0.5
and (b) 20 sccm. 77
xvii
Fig. 4.2: Raman spectra of N-DLC films deposited with different nitrogen flow rates. The
inset shows ID/IG and AD/AG as a function of nitrogen flow rate. 78
Fig. 4.3: (a) Ra values of N-DLC films versus nitrogen flow rate. AFM images of N-DLC
films deposited with nitrogen flow rates of (b) 0.5 and (c) 20 sccm. 80
Fig. 4.4: (a) Critical loads of N-DLC films with respect to nitrogen flow rate and (b) SEM
micrograph of a N-DLC film (0.5 sccm N2) scratch tested till a critical load of 456 mN. HP
and LP indicate high pressure and low pressure areas, respectively. The inset is the
progressive loading curve measured, from which the critical load is determined. 81
Fig. 4.5 Potentiodynamic polarization curves of N-DLC films measured in a 0.6 M NaCl
solution at room temperature. 84
Fig. 4.6: SEM micrographs showing surface morphologies of N-DLC films after
potentiodynamic polarization tests: (a) 0.5 sccm, Ecorr = -85.72 mV vs. SCE and (b) 3 sccm,
Ecorr = -57.41 mV vs. SCE, where the insets in the bottom right corners show their enlarged
views of locations A and B, respectively. 85
Fig. 4.7: Corrosion potentials (Ecorr) and polarization resistances (Rp) of N-DLC films as a
function of nitrogen flow rate. 86
Fig. 4.8: SEM micrographs of the corroded areas of N-DLC film coated samples after
immersion tests in 0.6 M NaCl solutions with different pH values: (a) pH 2, (b) pH 4.5 and (c)
pH 12 for the films deposited with 3 sccm N2 and (d) pH 4.5 for the film deposited with 20
sccm N2 for comparison. All the tests are conducted for 336 hr at room temperature and
ambient atmosphere. 89
Fig. 4.9: Cyclic voltammograms of N-DLC film electrodes with different nitrogen flow rates
measured in 0.5 M HCl solution at a scan rate of 100 mV/s. 93
xviii
Fig. 4.10: Cyclic voltammograms of N-DLC film electrodes with different nitrogen flow rates
measured in 0.1 M KCl solution at a scan rate of 100 mV/s. 94
Fig. 4.11: Cyclic voltammogram of N-DLC (20 sccm N2) film electrode measured in 0.1 M
NaCl solution at a scan rate of 100 mV/s. 95
Fig. 4.12: Cyclic voltammogram of N-DLC (20 sccm N2) film electrode measured in 0.1 M
KOH solution at a scan rate of 100 mV/s. 96
Fig. 4.13: Cyclic voltammogram of N-DLC (20 sccm N2) film electrode measured in 0.1 M
NaOH solution at a scan rate of 100 mV/s. 97
Fig. 4.14: Cyclic voltammograms of N-DLC (20 sccm N2) film electrode measured in 5 mM
K3Fe(CN)6 /0.1 M NaCl solution at different scan rates: (a) 30, (b) 50, (c) 70, (d) 90, (e) 110,
and (f) 130 mV/s. 98
Fig. 4.15: Stripping voltammograms obtained from N-DLC film electrodes deposited with (a)
3 and (b) 20 sccm N2 in a 1 × 10-3 M Pb2+ + 0.1 M KCl solution as a function of deposition
potential. Dependence of (c) stripping peak current and (d) stripping potential of Pb on
deposition potential. The deposition time and scan rate are 120 s and 36.36 mV/s,
respectively. 100
Fig. 4.16: A linear sweep cyclic voltammogram obtained from a N-DLC film electrode
deposited with 3 sccm N2 in the same solution as the one used for Fig. 4.13 with a scan rate of
36.36 mV/s. 101
Fig. 4.17: Stripping voltammograms obtained from a N-DLC film electrode (20 sccm N2) in a
1 × 10-6 M Pb2+ + 0.1 M KCl solution as a function of deposition potential. The deposition
time and scan rate are 120 s and 36.36 mV/s, respectively. 102
Fig. 4.18: Stripping voltammograms obtained from N-DLC film electrodes deposited with (a)
3 and (b) 20 sccm N2 in the same solution as the one used for Fig. 4.13 as a function of
xix
deposition time, dependence of (c) stripping peak current and (d) stripping potential of Pb2+ on
deposition time, and (e) relationship between stripping peak current and potential. The
deposition potential and scan rate are -1.2 V and 36.36 mV/s, respectively. 104
Fig. 4.19: Stripping voltammograms obtained from a N-DLC electrode (20 sccm N2) in two
different 0.1 M KCl solutions containing 1 × 10-2 M Zn2+ and 1 × 10-3 M Pb2+, respectively.
The scan rate, deposition time and deposition potential are 36.36 mV/s, 120 s and -1.2 V,
respectively. 106
Fig. 4.20: Stripping voltammograms obtained from a N-DLC film electrode (20 sccm N2) in a
2 × 10-5 M Cu2+ + 0.1 M KCl solution as functions of (a) deposition potential and (c)
deposition time. Dependence of stripping peak current and stripping potential on (b)
deposition potential and (d) deposition time, respectively. The deposition time is 120 s (a and
b), the deposition potential is -1.2 V (c and d), and the scan rate used for all the tests is 36.36
mV/s. 108
Fig. 4.21: Stripping voltammograms obtained from a N-DLC film electrode (20 sccm N2) in a
1.1 × 10-6 M Hg2+ + 0.1 M KCl solution as functions of (a) deposition potential and (b)
deposition time. Dependence of stripping peak current on (c) deposition potential and (d)
deposition time. The deposition time is 120 s (a and c), the deposition potential is -1.2 V (b
and d), and the scan rate for all the tests is 36.36 mV/s. 110
Fig. 4.22: (a) Stripping voltammograms obtained from a N-DLC film electrode (20 sccm N2)
in a 0.1 M KCl solution containing 8.9 × 10-6 M Pb2+ + 2.5 × 10-5 M Cu2+ + 9.2 × 10-6 M Hg2+
as a function of deposition time. Dependence of (b) stripping peak current and (c) stripping
potential on deposition time. The deposition potential and scan rate are -1.2 V and 36.36
mV/s, respectively. 112
xx
Fig. 5.1: N/(C+Ru+Pt+N), Pt/(C+Ru+Pt+N) and Ru/(C+Ru+Pt+N) atomic ratios with respect
to DC power applied to Pt50Ru50 target during film depositions. 120
Fig. 5.2: TEM micrograph of a PtRuN-DLC film deposited with a DC power of 30 W applied
to Pt50Ru50 target. 122
Fig. 5.3: XPS spectrum of a PtRuN-DLC film deposited with DC power of 30 W applied to
Pt50Ru50 target. 122
Fig. 5.4: Fitted XPS spectra of a PtRuN-DLC film deposited with DC power of 15 W applied
to Pt50Ru50 target: (a) C 1s + Ru 3d, (b) N 1s, (c) Pt 4f and (d) Ru 3p. The insets show the
relevant XPS spectra of PtRuN-DLC films as a function of DC power applied to Pt 50Ru50
target. 126
Fig. 5.5: Raman spectrum together with fitted G and D peaks of a PtRuN-DLC film deposited
with DC power of 15 W applied to Pt50Ru50 target. The inset shows the Raman spectra of
PtRuN-DLC films deposited with different DC powers applied to Pt50Ru50 target. 127
Fig. 5.6: Results from the fitted Raman spectra of PtRuN-DLC films as shown in the inset of
Fig. 5.5: (a) peak positions, (b) FWHMs and (c) ID/IG ratios of D and G peaks. 129
Fig. 5.7: Water contact angles of PtRuN-DLC films as a function of DC power applied to
Pt50Ru50 target. The insets show the water droplets on the surfaces of the films deposited with
DC powers of (a) 15 and (b) 30 W applied to Pt50Ru50 target. 130
Fig. 5.8: AFM images showing surface topographies of PtRuN-DLC films deposited with DC
powers of (a) 15 and (b) 30 W applied to Pt50Ru50 target. 131
Fig. 5.9: Critical loads of PtRuN-DLC films as a function of DC power applied to Pt50Ru50
target. 132
Fig. 5.10: (a) Potentiodynamic polarization curves, (b) corrosion current (Icorr) and
polarization resistance (Rp), and (c) corrosion potential (Ecorr) of PtRuN-DLC films as a
xxi
function of DC power applied to Pt50Ru50 target. The insets in (c) show SEM micrographs of
corroded areas of PtRuN-DLC films deposited with DC powers of 30 (top left) and 25 W
(bottom right) applied to Pt50Ru50 target. 135
Fig. 6.1: TEM image of PtRuN-DLC film. 142
Fig. 6.2: Fitted XPS C 1s and C 1s + Ru 3d spectra of (a) N-DLC and (b) PtRuN-DLC films,
respectively, and fitted XPS N 1s spectra of (c) N-DLC and (d) PtRuN-DLC films. 143
Fig. 6.3: Raman spectra of N-DLC and PtRuN-DLC films, where G and D represent fitted G
and D peaks, respectively. 145
Fig. 6.4: Water droplets on (a) N-DLC and (b) PtRuN-DLC film surfaces. 146
Fig. 6.5: AFM images showing surface morphologies of (a) N-DLC and (b) PtRuN-DLC
films. 147
Fig. 6.6: SEM micrographs showing surface morphologies of scratched (a) N-DLC and (b)
PtRuN-DLC films. The inset in (a) shows the scratch track of the N-DLC film. 147
Fig. 6.7: (a) Nyquist and (b) Bode plots of N-DLC and PtRuN-DLC films measured in 0.1 M
HCl solution. The frequency range is 105 – 10-2 Hz and the amplitude is 10 mV. The inset in
(a) shows an equivalent circuit for electrochemical reactions on N-DLC and PtRuN-DLC
coated samples. 150
Fig. 6.8: Cyclic voltammograms measured from N-DLC (solid line) and PtRuN-DLC (dash
line) film electrodes in (a) 0.1 M H2SO4 solution, (b) 0.1 M HCl solution and (c) 0.1 M KCl
solution, where scan rate is 100 mV/s. 153
Fig. 6.9: Cyclic voltammograms measured from (a) N-DLC film electrodes with different
scan rates and (b) N-DLC (solid line) and PtRuN-DLC (dash line) film electrodes in 1 mM
K3Fe(CN)6/0.1 M HCl solution, where scan rate is 10 mV/s. 157
xxii
List of Tables
Page
Table 1.1: Approximate potential ranges for platinum, mercury and carbon in aqueous and
non-aqueous electrolytes. 3
Table 3.1: Process parameters of N-DLC films. 63
Table 3.2: Process parameter of PtRuN-DLC films. 64
Table 3.3: Process parameter of N-DLC and PtRuN-DLC films. 65
Table 6.1: Chemical compositions and sp2/sp3 ratios of N-DLC and PtRuN-DLC films. 141
Table 6.2: Results determined from fitted Raman spectra as shown in Fig. 6.3. 145
Table 6.3: Water contact angles, surface roughnesses and critical loads of N-DLC and
PtRuN-DLC films. 146
Table 6.4: Results determined from EIS spectra based on the proposed equivalent circuit as
shown in the inset of Fig. 6.7a. 151
Table 7.1: Major findings from the three major work chapters (4 to 6). 164
xxiii
xxiv
1
Chapter 1 Introduction
1.1. Synopsis
Carbon is a unique and intriguing material with a diversity of technological
applications. With a 1s2 2s2 2p2 electronic ground state configuration, carbon naturally exists
in many allotropic forms such as graphite, diamond, bucky ball (C60) and so on. In the last
few decades an important scientific and technological breakthrough occurred with the
discovery that diamond thin films can be successfully grown by many chemical and physical
vapor deposition techniques, artificially.
Diamond-like carbon (DLC) is a type of carbon material that contains both sp2
(graphite-like) and sp3 (diamond-like) bonds. DLC has attracted researchers’ attention three
decades ago due to its outstanding properties similar to those of diamond. The unique
properties of DLC films and their modifications, together with the possibility to adjust the
properties by choosing the right deposition parameters, make them suitable for a variety of
applications. Some of the exploited properties of DLC materials are high hardness, high wear
resistance, low friction coefficient, high chemical inertness, good infrared transparency, high
electrical resistivity, and low dielectric constant.
DLC has found novel applications in many fields. For example, DLC films deposited
at low temperatures can be a suitable wear-protective layer on products made of plastics (e.g.
on sunglass lenses made of polycarbonate) [1]. DLC films have been widely used in hard disk
drives in which ultra thin DLC films are used for both magnetic disks and heads. Some other
tribological applications of DLC films are for cutting tools, bearings, gears and seals.
DLC is biocompatible so that DLC films can protect biological implants from
corrosion and serve as diffusion barriers due to their chemical inertness and dense structures
impermeable to biofluids. Depositions of DLC films on stainless steel and titanium alloys
1
used for the components of artificial heart valves also satisfy both mechanical and biological
requirements, thus resulting in the improved performance of these components [2].
Nowadays, human activity has released toxic metals into the environment. Pollutants
in water include a wide spectrum of chemicals, trace metals, pathogens, and chemical or
sensory changes due to the contamination, over-use and mismanagement of water resources
[3]. Many of the chemical substances are toxic, for example, pathogens can produce
waterborne diseases in human or animal hosts. Preventing of pollution is treated by
determination of wastewater discharge to the aquatic environment and water emission
limitations.
The presence of toxic metals such as mercury, lead, copper, etc. in the environment
has been a source of worry to environmentalists, government agencies and health practitioners
because these metals in the aquatic ecosystem have far-reaching implications directly to the
biota and indirectly to human beings. A concentration of Pb > 4.8 × 10 -7 M is detrimental to
foetuses and children with possible development of neurological problems [3]. For Cu, a
concentration of 7.87 × 10-5 M can give adverse chronic effects [3]. Mercury occurs in
deposits throughout the world. It is harmless in insoluble forms, such as mercuric sulfide, but
it is poisonous in soluble forms such as mercuric chloride or methylmercury. Mercury can
enter the environment from natural sources such as volcanoes, stationary combustion such as
power plants, productions such as gold, non-ferrous metals, mercury itself, etc., waste
disposal including municipal and hazardous water, crematoria, and the deposal of certain
products such as auto parts, batteries, etc. [4]. Case control studies have shown the effects of
trace mercury, such as tremors, impaired cognitive skills, and sleep disturbance in human
beings, with chronic exposure to mercury even at low concentrations in the range of 3.4 ×10 -9
– 2 × 10-7 M [5, 6]. A study has shown that acute exposure (4-8 hr) to calculated elemental
2
mercury level of 5.48 × 10-6 to 2.14 × 10-4 M can cause chest pain, dyspnea, cough,
hemoptysis, impairment of pulmonary function, and evidence of interstitial pneumonitis [7].
Although copper and zinc have been found to have low toxicity to human beings, prolonged
consumption of a large dose can result in some health complications. Therefore, fast detection
and determination of heavy metals are a tough challenge for electrochemical analysts.
Electroanalytical techniques are relatively simple, quick, cheap and easy to use in situ for
measurements in rivers or lakes [8, 9]. Anodic stripping voltammetry (ASV), adsorptive
stripping voltammetry (ASV) and cathodic stripping voltammetry (CSV) have allowed the
determination of a wide range of inorganic and organic species down to concentrations of an
order of 10-10 M under favorable conditions.
Table 1.1: Approximate potential ranges for platinum, mercury and carbon in aqueous and
non-aqueous electrolytes [9].
TBAP tetrabutylammonium perchlorate DMF dimethylformamide TBABF4 tetrabutylammonium tetrafluoroborate CAN acetonitrile
3
In the past, mercury was the first metal to be extensively used for electroanalytical
purposes in the form of the dropping mercury electrode because of its cyclic operation –
continual drop growth, release and renewal – avoiding many of the problems of electrode
poisoning in complex matrix. The drawback of the mercury electrodes is that the positive
parts of their electrochemical potential windows in aqueous solutions are very small as shown
in Table 1.1 [9]. Therefore, the detection limit can be too high and there are some
experimental manipulation difficulties because of its toxicity. There are some metallic ions
that cannot be determined at mercury (for example, Au and Ag). Therefore, solid electrode
materials have been developed in order to permit oxidation reactions to be studied but there
are no electrode materials being as good as mercury for studying reduction owing to its
extended negative limits. However, carbon has become common in the form of glassy
carbon, graphite, carbon paste and carbon fiber electrodes. Although glassy carbon is one of
the most widely used electrode materials in electroanalytical applications due to its
robustness, smooth surface and wide electrochemical window, irreproducible background
contribution and gradual loss of surface activity frequently affect its electroanalytical
performance. Noble metals like Pt and Au have been commonly used for microelectrodes.
However, in an aqueous solution, detection using such metal electrodes is often impossible at
negative potentials due to high currents produced by hydrogen evolution reactions on Pt [9].
Metal and graphite electrodes are limited by surface oxidation and reduction in terms of
sensitivity [10].
Boron doped diamond (BDD) film electrodes prepared by chemical vapor deposition
(CVD) can have remarkable properties such as wide potential window in aqueous and non-
aqueous media, low and stable background current, weak adsorption of polar molecules, high
resistance to electrode deactivation and fouling, long-time response stability and
4
superbmicrostructural and morphological stability at high temperatures [11, 12]. BDD
electrodes have been successful in stripping voltammetric analysis of Pb, Mn, Cd, Cu and Ag
[13-15]. However, fabrication of diamond films demands a high substrate temperature which
is impossible for microfabrication of semiconductors and microelectrode arrays [16]. In
addition, difficulties in controlling the deposition parameters limit the applications of
diamond films.
Recently, there has been much interest in the use of conductive DLC films for
electrochemical measurements [17]. DLC films can be deposited over a large surface area
with high uniformity of near-atomic level. A large variety of microstructures can be obtained,
ranging from graphite-like to diamond-like depending on the deposition methods and process
parameters employed, which has been reflected by the considerable variations of physical and
mechanical properties measured. The acceptance of dopants makes them electrically
conductive. Smoother surface topography, better adhesion to the underlying substrate, and
relative ease in fabrication compared to diamond films make DLC films a desirable choice in
certain applications. The unique combination of their chemical and mechanical properties
opens a possibility for their applications as electrode materials in electrochemistry. DLC films
also exhibit a low double-layer capacitance, large electrochemical window, and low
background density. In addition, compared to the traditionally used mercury coated graphite
electrode, DLC is environmentally friendly. When combined with the advantage of ready
deposition at ambient temperature and convenient masking for fabricating microelectrode
arrays and sensors, the properties of DLC films have suggested a valuable potential in
electroanalysis.
DLC films produced by physical vapor deposition (PVD) and chemical vapor
deposition (CVD) methods at room temperature have been explored as electrode materials for
5
heavy metal tracing [18-24]. However, high electrical resistivity and residual stress of DLC
films have confined their electrochemical applications. The electrical conductivity of DLC
films is very important for electrochemical analysis because it can abruptly affect the
sensitivity of the film electrodes. High residual stress in DLC films apparently reduces the
adhesion strength of the films. In a corrosion environment, poor adhesion strength of DLC
films allows undermining and delaminating of the films by attacking the interfacial bonds
between the films and their underlying substrates with electrochemically active species
permeated through the porosities in the films. The characteristics of DLC films can be altered
by incorporating different elements such as H, N, F, Si, Pt, Au, Ni, Ti, V, etc. in the films [25-
27]. It is known that an introduction of nitrogen into DLC films can lower both the electrical
resistivity and corrosion resistance of the films because of increased sp2 sites in the films [18].
However, the introduction of nitrogen into DLC films lowers the corrosion resistance of the
films due to the increased sp2 sites in the films. The poor corrosion resistance of N doped
DLC films can abruptly affect the electrochemical properties of the films such as sensitivity,
repeatability, long-time response stability and durability when the DLC films are used as film
electrodes for electrochemical purposes. Therefore, an improvement of the corrosion
resistance of N doped DLC (N-DLC) films becomes important for electrochemical
applications.
Pt and Ru possess high resistances to chemical attack, excellent high-temperature
characteristics, and stable electrical properties. Pt does not oxidize in air and is insoluble in
hydrochloric and nitric acids though it does dissolve in aqua regia [28]. Therefore, Pt/Ru/N
doped DLC (PtRuN-DLC) thin films to be developed as novel materials in this project are
expected to have the following potential applications in the areas of water quality technology,
reclaimed water technology and water technologies in industrial applications:
6
(a) qualitative and quantitative electrochemical analyses (e.g. detection of heavy
metals such
as Zn, Pb, Cu, Hg, etc) and
(b) water purification and disinfection.
The novelty of this project is:
• N-DLC thin films with or without Pt and Ru doping to be developed are novel
materials for electrochemical applications.
• Investigation of the structural and electrochemical properties of the novel thin films.
1.2. Objective
The main objective of this project is to develop N-DLC thin films with or without Pt and
Ru doping and to study their structural and electrochemical characteristics, which forms two
parts:
• Optimization of the deposition conditions of the N-DLC and PtRuN-DLC thin films
by controlling the doping levels of N, Pt and Ru in the films in order to improve the
electrical conductivity of the films and, at the same time, maintain the good
electrochemical performance of the films.
• Investigation of the structural and electrochemical properties of the N-DLC and
PtRuN-DLC films.
1.3. Scope
• The N-DLC and PtRuN-DLC thin films will be deposited on p-type Si substrates
using filtered cathodic vacuum arc (FCVA) and DC magnetron sputtering deposition
methods. The film deposition conditions, such as substrate bias, sputtering power
density, gas flow rate and deposition duration, will be optimized.
7
• The film chemical composition will be measured using X-ray photoelectron
spectroscopy (XPS).
• The film microstructure will be observed using transmission electron microscopy
(TEM).
• The film chemical structure will be diagnosed using XPS and micro-Raman
spectroscopy.
• The film surface activity, morphology and topography will be studied using contact
angle measurement, scanning electron microscopy (SEM) and atomic force
microscopy (AFM), respectively.
• The adhesion of the films to the substrates will be evaluated using micro scratch test.
• The corrosion resistance of the films in NaCl and HCl solutions will be studied using
potentiodynamic polarization test, immersion test and electrochemical impedance
spectroscopy (EIS).
• The cyclic voltammetric behavior of the films in different aqueous solutions, such as
HCl, H2SO4, KCl, NaCl, KOH, and NaOH, will be studied using linear sweep cyclic
voltammetry.
• Anodic stripping voltammetric behavior of the films is to be evaluated in terms of
trace heavy metals, such as Zn, Pb, Cu, Hg, etc., by using linear sweep anodic
stripping voltammetry.
• The relationship between the microstructure and electrochemical properties of the
films will be systematically investigated.
1.4. Organization
This thesis has seven chapters. Chapter 1 provides background, objectives and scope
of the dissertation. Chapter 2 summarizes the state-of-the-art on the development of DLC
8
films. Chapter 3 details the experimental methodologies used for the current research. Results
are systematically analyzed and discussed in Chapters 4 to 6. Based on the analysis of the
results, conclusions are drawn in the last chapter.
9
Chapter 2 Literature review
2.1. Carbon
Carbon is one of the commonest elements throughout the Universe. Carbon is a
chemical element that has the symbol C and atomic number 6. Carbon is the lightest element
of 4A group. The electronic configuration for carbon is written as 1s22s22p2. An abundant
nonmetallic, tetravalent element, carbon has several allotropic forms. As the free element it
forms allotropes from differing kinds of carbon-carbon bonds. Different forms include the
hardest naturally occurring substance (diamond) and one of the softest substances (graphite)
known. Depending on its allotropic form and the impurities, carbon can be an insulator,
conductor, or semiconductor.
Fig. 2.1: The sp3, sp2, sp1 hybridized bonding [29].
Carbon forms a great variety of crystalline and disordered structures because it is able
to exist in three hybridizations, sp3, sp2 and sp1 (Fig. 2.1). In the sp3 configuration, as in
diamond, a carbon atom’s four valance electrons are each assigned to a tetrahedrally directed
sp3 orbital, which makes a strong σ bond to an adjacent atom. In the three-fold coordinated sp2
configuration as in graphite, three of the four valance electrons enter trigonally directed sp2
orbitals, which form σ bonds in a plane. The fourth electron of the sp2 atom lies in a pπ
orbital, which lies normal to the σ bonding plane. This π orbital forms a weaker π bond with a
π orbital on one or more neighboring atoms. In the sp1 configuration, two of the four valence 10
electrons enter σ orbitals, each forming a σ bond directed along the ± x-axis, and the other two
electrons pπ orbitals in the y and z directions. Although it forms an incredible variety of
compounds, most forms of carbon are comparatively unreactive under normal conditions, e.g.,
it does not react with sulfuric acid, hydrochloric acid, chlorine or any alkalis.
2.1.1. Diamond
Diamond is composed of the single element carbon and the structure is shown in Fig.
2.2. The diamond structure is cubic and can be viewed as two interpenetrating FCC structure
displaced by (¼, ¼, ¼) ao along the body diagonal. The cube edge length a0 is 3.567 Å and the
nearest-neighbor carbon distance is 1.544 Å. Each of the carbon atoms in the structures is
tetrahedrally bonded to the four nearest carbon atoms at the corners of a regular tetrahedron
by strong covalent sp3 bonds.
Fig. 2.2: Ideal diamond structure [2].
Diamond is one of the best known allotropes of carbon, whose hardness and high
dispersion of light make it useful for industrial applications and jewelry. Most notable are its
extreme hardness, its high dispersion index, and high thermal conductivity, with a melting
point of 3820 K (3547 °C / 6420 °F) and a boiling point of 5100 K (4827 °C / 8720 °F).
Naturally occurring diamond has a density ranging from 3.15 to 3.53 g/cm³. Diamond is the
hardest natural material known to man - its hardness set to 10 (hardest) on Mohs scale of
11
mineral hardness and having an absolute hardness value of between 90, 167, and 231 GPa in
various tests, which makes it an excellent abrasive. Other specialized applications also exist or
are being developed, including use as semiconductors: Diamond with a band gap of 5.45 eV is
considered among the best electrical insulators. Through doping, diamond can be made into
an excellent p-type semiconductor.
2.1.2. Graphite
Graphite is one of the allotropes of carbon. Graphite holds the distinction of being the
most stable form of solid carbon ever discovered. Graphite is composed of the single element
carbon and the structure is shown in Fig. 2.3.
Fig. 2.3: Crystal structure of graphite [2].
Each carbon atom is covalently bonded to three other surrounding carbon atoms. The
flat sheets of carbon atoms are bonded into hexagonal structures. These exist in layers, which
are not covalently connected to the surrounding layers. Instead, different layers are connected
together by weak forces called van der Waals forces. The unit cell dimensions are a = b =
2.456 Å, c = 6.694 Å. The carbon-carbon bond length in the bulk form is 1.418 Å, and the
interlayer spacing is c/2 = 3.347 Å. Each carbon atom possesses a sp² orbital hybridization.
The π orbital electrons delocalized across the hexagonal atomic sheets of carbon contribute to
12
graphite's conductivity. In an oriented piece of graphite, conductivity parallel to these sheets
is greater than that perpendicular to these sheets. Graphite can conduct electricity due to the
vast electron delocalization within the carbon layers. These electrons are free to move, so are
able to conduct electricity. However, the electricity is only conducted within the plane of the
layers. The bond between the atoms within a layer is stronger than the bond of diamond, but
the force between two layers of graphite is weak. Therefore, layers of it can slip over each
other making it soft.
2.1.3. Diamond films
In the past two decades an important scientific and technological breakthrough
occurred with the discovery that diamond thin films can be successfully grown by a large
variety of chemical deposition technique. The feature common to all these methods is that
decomposed hydrocarbon radicals impinge upon a hot (> 900 ˚C) surface in the presence of
atomic hydrogen. Since the procedure enhances the formation of sp3 over sp2 bonds, diamond
can be grown. The films grown by these techniques are usually polycrystalline, consisting of
agglomerations of randomly oriented, small diamond crystallites (several micron in size), and
the films thus tend to have rough surface morphologies. Raman spectroscopic studies of
diamond films usually show a sharp peak at 1332 cm-1, typical for sp3 bonded carbon,
superimposed on a broad peak at about 1500 cm-1 which is due to graphitic sp2 bonding.
2.1.4. Diamond-like carbon (DLC)
Diamond-like carbon (DLC) is a metastable form of amorphous carbon (Fig. 2.4)
containing a significant fraction of sp3 bonds. DLC has some extreme properties similar to
diamond, such as hardness, elastic modulus and chemical inertness, but these are achieved in
an isotropic disordered thin film with no grain boundaries. Diamond-like carbon (DLC) films
13
lack any long-range order and contain a mixture of sp3, sp2 and sometimes even sp1
coordinated carbon atoms in a disordered network. The bond types have a considerable
influence on the material properties of amorphous carbon films. If the sp2 type is predominant,
the film will be softer, and if the sp3 type is predominant, the film will be harder.
Fig. 2.4: Structure of amorphous carbon [2].
The structure of DLC modeled by Robertson [30] is a random network of covalently
bonded carbon atoms in the different hybridizations, with a substantial degree of medium
range order on the 1 nm scale. DLC is a name attributed to a variety of amorphous carbon
materials with carbon atoms bonded in mainly sp3 and sp2 hybridizations, some containing up
to about 50 at% hydrogen (a-C:H), other containing less than 1 % hydrogen (a-C). The DLC
films contain a significant fraction of sp3 bonds, giving them attractive physical and
mechanical properties that are similar to those of diamond films. The properties of DLC films
are determined by the relative ratio of the two hybridizations (sp3 and sp2). The a-C:H films
typically contain sp3 fractions smaller than 50%, while the a-C films can contain up to 85%
sp3 bonds. A-C:H is commonly used to designate the hydrogenated form of diamond-like
carbon, while a-C is used to designate the non-hydrogenated carbon [31].
The compositions of the various forms of a-C:H alloys on a ternary phase diagram are
displayed in Fig. 2.5, as first used by Jacob et al. [32]. There are many a-Cs with disordered
14
graphitic ordering, such as glassy carbon and evaporated a-C, which lie in the lower left hand
corner. High degree of sp3 bonding in a-C can be extended by sputtering method. If the
fraction of sp3 bonding reaches a high degree, the a-C is denoted as tetrahedral amorphous
carbon (ta-C), to distinguish it from sp2 a-C [33]. Non-hydrogenated ta-C, which has small
fractions of sp2 bonds, has a very rigid network.
Fig. 2.5: Ternary phase diagram of bonding in amorphous carbon-hydrogen alloys [32].
Various materials derived from DLC films have been developed to change and
improve their properties. Such materials are similar in structure to DLC but containing Ti, Zr,
W, Nb, Cr, Ni, Fe, Mo, La2O3 or WC and non-metals such as Si, N, B, F and P in addition to
carbon and/or hydrogen [34-38]. Most modifications made to DLC are to reduce its internal
stresses, surface energy and friction coefficient, or to modify its electrical properties [39, 40].
Both the hydrogenated and non-hydrogenated DLC are metastable materials and their
structures will change towards graphite–like carbon by either thermal activation or irradiation
with energetic photons or particles [41, 42]. It has been reported that thermal activation can
induce changes in DLC films and cause the conversion of some sp3 carbon bonds to sp2 bonds
[42].
15
The properties of DLC films, such as hardness, elastic modulus and internal stress, are
directly correlated to the fraction of sp3 bonds in the films. The hardness of DLC films is in
the range of 10-30 GPa [43], with a corresponding Young’s modulus 6-10 times larger. The
hardness of ta-C films can reach higher values (in the range of 40-80 GPa) [44]. The high
internal stress limits the thickness of films that can be used for any application, often to less
than 1 µm thick. The stress can be reduced by incorporating N, Si, or metals in the films or by
building multilayered structures comprising soft and hard films, although the reduction in the
stress is often associated with a reduction in hardness and elastic modulus of the films [45,
46].
2.2. Doping of DLC films
DLC films with a high sp3 content appear more suitable for device fabrication, since
they are smooth, have no grain boundaries and a high electrical resistivity and band gap up to
about 3 eV[47]. The availability of a doping method is a necessary condition for their use as
active elements in semiconductor devices. Given the structural and chemical similarities of
diamond and silicon, it can be expected that the doping of DLC films can be achieved in the
same way as in the case of amorphous silicon (a-Si) films. However, no evidence of the
formation of even a simple p-n junction based on DLC on a Si substrate has been reported till
now.
Doping process is a way to both decrease the stress in DLC films and make them
electrically conductive. There are two processes for doping in DLC films: ion implantation
and in-situ doping in film deposition.
By implanting acceptor or donor ions, such as B+, BF+, P+, C+ and metal ions, the
conductivity of DLC films is found to increase once a certain threshold dose has been
reached. But the measurement has not revealed any significant difference between C and B
16
implantation despite the valence difference between C and B. It is therefore concluded from
the work that the electrical effect was due to implantation-induced damage and not to
chemical doping caused by the particular implanted ions. Inhomogeneous dopant
concentration, damage and graphitization make ion implantation for DLC films difficult in
practice.
In-situ doping, there are three ways to choose. First, a gas, containing the dopant
atoms, is introduced into a gas mixture during film growth (such as boron doping or nitrogen
doping). Second, dopant elements are impregnated into a solid carbon source. Third, dopant
ion beams, which are developed independently, are introduced into deposition processes.
Any semiconductor is only really useful when it can be doped both n- and p-types.
This is a particular problem in all wide-gap semiconductors. Carbon can exist in three
hybridizations, sp3, sp2 and sp1. The σ bonds of all carbon sites form occupied σ states in the
valance band and empty σ* states in the conduction band, separated by a wide σ-σ* gap (Fig.
2.6). The π bonds of sp2 and sp1 sites form filled π states and empty π* states, with a much
narrower π-π* gap [48, 49].
Fig. 2.6: Schematic DOS of a carbon showing σ and π states [48].
Semiconductor can only be doped to one polarity, either because of the dopant levels
that are too deep, or the low solubility, or because of auto-compensation. Solubility is not an
issue, as the flexibility of a random network allows atoms of any size to be incorporated. The
17
main problem in amorphous semiconductor is that network flexibility allows atoms to also
exert their chemically preferred valance [50] and also form a trivalent non-doping site [51].
Undoped ta-C is slightly p-type with EF lying just below mid gap [52]. Nitrogen has a similar
size to carbon, so nitrogen is soluble. Nitrogen is an obvious candidate as a donor in ta-C. It is
noted that ta-C has a narrower gap than diamond (Fig. 2.7), so that level like N which is deep
in diamond can be shallow in ta-C [53].
Generally, the doping efficiency of nitrogen is low in amorphous semiconductors
because N can adopt so many bonding configurations in a carbon network. Apart from the sp3
substitutional sites which are doping, these are the sp2 substitutional site (doping) and many
non-doping sites such as pyridine, pyrrole and nitrile. It is because it encourages carbon to
form sp2 bonding.
Fig. 2.7: Schematic of the levels of substitutional nitrogen in diamond and ta-C [53].
Bonding in carbon nitride systems is complex since both atoms; carbon and nitrogen
can exist in three hybridizations, sp3, sp2 and sp1. Therefore, at least nine different bonding
configurations are possible, as shown in Fig. 2.8.
18
Fig. 2.8: Various nitrogen configurations in DLC, showing the doping configuration. One dot means an unpaired electron. Two dots mean a lone pair (non-bonding) [54].
In its simple trivalent configuration (a) N30, nitrogen forms three σ bonds with the
remaining two electrons in a lone pair. Nitrogen in a four-fold coordinated substitutional site
(b) N4+, uses four electrons in σ bonds with the remaining unpaired electron available for
doping. A variant of this site is the (c) N4+ – C3- pair; which forms a trivalent carbon site and
the unpaired electron transfers to the carbon to give a positive N+ site/C- defect pair. The
remaining configurations correspond to π bonding. Nitrogen can substitute for carbon in a
benzene ring (d) pyridine and (e) doped pyridine-like. Another alternative, (f) pyrrole, is of
nitrogen bonded to three neighbors in a five-fold ring, here three electrons are in bonds and
the other two are used to complete the sextet of the aromatic ring. The other variants of π
bonding are; with nitrogen two-fold coordinated, where a double bond unit (g) has two
electrons in σ bonds and one in a π bond, leaving a non-bonding pair, and a double bond unit
(h) which uses three electrons in σ bonds, one in a π bond and the fifth in an antibonding π*
state, available for doping [54]. The last configuration (i) is the triple bond with an isolated
lone pair; as in cyanide. Substitutional N can dope graphite C by donating π electrons. The N
19
uses three of its valance electrons to form three σ bonds, and one electron to fill the π states
and its fifth electron enters the π* state, giving a ‘π-doping’.
Fig. 2.9: Diagram showing how nitrogen modifies bonding in graphitic layer, by introducing five-membered rings and warping, or inter-layer bonding [29].
New types of local bonding are found in sputtered N-DLC films. Nitrogen has one
more electron than carbon, so that nitrogen in a five-fold pyrrole ring now gives the six π
electrons needed for aromatic stability [29]. These rings can introduce warping in a graphite
layer (Fig. 2.9). Another effect of nitrogen is to favor inter-layer bonding. Substituting
nitrogen for C breaks a π bond and leaves an unpaired electron of the remaining C which is
available to make σ bond to a similar atom on an adjacent layer.
Bai et al. reported [55] that N incorporation increased the sp2 bonds which contributed
to lower resistivity of the N-DLC films. This is due to the three factors: (1) N could act as a
weak donor; the higher the donor density in the films is, the higher the electrical conductivity
of the films is, (2) N raises the Fermi level towards the conduction band and (3) N narrows the
band gap by graphitization of the bonding [55]. The last factor is dominant for the increase of
conductivity with annealing temperature and N concentration.
The wide-gap diamond-like carbon is not in itself electrochemically active; however,
it acquires electrochemical activity upon introducing platinum into the DLC bulk during the
film deposition. In Pt incorporated DLC films, Pt usually cannot be evenly doped [56]. The
20
effect of platinum is shown to be of a threshold nature: the electrochemical current appears at
approximately 3 at% and saturates at approximately 10 at% of Pt [56]. By contrast, the
differential capacitance increases continuously with increasing Pt content. The observed
effects are explained in terms of a model assuming non-uniform characters of both the
electrical conductance in the DLC bulk and the catalytic effect of Pt at the electrode/solution
interface. A DLC film is a sp3-hybridized carbon matrix, which contains areas of sp2-
hybridized states. The DLC film has a hopping type conductance, which is associated with
charge carrier hopping between the sp2-hybridized states. The metal may also be non-
uniformly distributed in the DLC matrix: atoms of metals form clusters. Isolated metal
clusters are electrically inactive in the film bulk. However, when appearing on the film
surface, the Pt clusters play the role of catalytic-active sites, which facilitate the charge
transfer at the electrode/solution interface. With increasing Pt concentration in the DLC film,
the distance between the Pt clusters decreases both in the film bulk and on its surface. When a
threshold value of the Pt concentration (approx. 3 at%) is reached, the distances between the
Pt clusters and the nearest sp2-hybridized states on the surface drop down to some critical
values which ensure their effective exchange with charge carriers. Therefore, effective current
flows occur both in the electrode bulk and through the electrode/solution interface.
2.3. Deposition methods of DLC films
In order to grow DLC film instead of graphitic films, one has to circumvent the natural
tendency of carbon to form stable sp2 graphite-like bonds as opposed to sp3 metastable
diamond-like bonds. In order to obtain the metastable structure of DLC, such films are
deposited by plasma assisted chemical vapor deposition or physical vapor deposition
techniques (sputtering or ion beams) using a variety of precursors. The deposition is
performed in hydrogen-containing environment to obtain DLC films containing 10-50 at%
21
hydrogen. The hydrogen is required for obtaining “diamond-like carbon” properties for these
materials. The deposition of hydrogen free DLC requires a carbon source and energy for the
carbon species. The superhard properties of DLC films are achieved by the high energies of
the impinging particles that form the films due to the films grown by subplantation [44]. The
required high energies of the depositing species are achieved by different variations of
cathodic arc discharges, such as filtered arc, pulsed arc, laser controlled arc, pulsed laser
depositions or mass selected ion beams.
2.3.1. Magnetron sputtering deposition
The most common industrial process for the deposition of DLC is sputtering [57]. The
most common form uses the DC or RF sputtering of a graphite target by Ar plasma. Because
of the low sputter yield of graphite, magnetron sputtering is often used to increase the
deposition rate.
In DC sputtering as shown in Fig. 2.10, the electrons that are ejected from the cathode
are accelerated away from the cathode and are not efficiently used for sustaining the
discharge. By the suitable application of a magnetic field, the electrons can be deflected to
stay near the target surface and by an appropriate arrangement of the magnets, the electrons
can be made to circulate in a closed path on the target surface. This high flux of electrons
creates a dense plasma near the cathode at low pressures so that ions can be accelerated from
the plasma to the cathode without loss of energy due to physical and charge-exchange
collisions. This allows a high sputtering rate with a lower potential on the target than with the
DC diode configuration. The most common magnetron source is the planar magnetron where
the sputter-erosion path is a closed circle on a flat surface. The magnetic field in magnetron
sputtering can be formed using permanent magnet or electromagnetic or a combination of the
two.
22
Fig. 2.10: General feature of sputter coater [57].
The disadvantage of the magnetron sputtering configurations is that the plasma is
confined near the cathode and is not available to activate reactive gases in plasma near the
substrate for reactive sputter deposition. This disadvantage can be overcome by applying an
unbalanced magnetron configuration where the magnetic field can be configured to pass
across to the substrate such that some electrons can escape from the cathode region. This
system helps the formation of sp3 bonding in carbon film.
Reactive sputter deposition from an elemental target relies on: (a) the reaction of the
depositing species with a gaseous species, such as nitrogen and (b) reaction with a co-
depositing species to form a compound. The reactive gas is in the molecular state (e.g., N2)
and “activated” by dissociation of molecular species to more chemically reactive radicals
(e.g., N2 + e-→ 2N0). Typically, the reactive gas has a low atomic mass (N=14) and is thus not
effective in sputtering. It is therefore desirable to have a heavier inert gas, such as argon, to
aid in sputtering. Mixing argon with the reactive gas also aids in activating the reactive gas by
the penning ionization processes. The appropriate gas composition and flow for reactive
sputter deposition can be established by monitoring the partial pressure of the reactive gas as a
function of reactive gas flow [57].
23
Scanning coil
Double coil
Grinder
StrikerCathode
Platform
2.3.2. Cathodic arc
A usual method for laboratory and industrial use is the cathodic arc [33]. An arc is
initiated in a high vacuum by touching the graphite cathode with a small carbon striker
electrode and withdrawing the striker. This produces an energetic plasma with a high ion
density. The power supply is a low voltage and high current supply. The cathode spot is
formed by an explosive emission process. This creates particulates as well as the desired
plasma.
Fig. 2.11: FCVA source design [57].
The particulates can be filtered by passing the plasma along a toroidal magnetic filter
duct. This is known as filtered cathodic vacuum arc (FCVA). The toroidal currents produce a
magnetic field of about 0.1 T along the axis of the filter. The electrons of the plasma spiral
around the magnetic field lines and so they follow them along the filter axis. This motion
produces an electrostatic field, which causes the positive ions to follow the electrons around
the filter axis. This motion produces a transport of the plasma around the filter. The plasma
beam is condensed onto a substrate to produce the ta-C. A DC or RF self-bias voltage applied
to the substrate is used to increase the incident ion energy. The particulates are generally
24
Focusing coil
Anode
submicron-size particles. These can still pass through the filter section by bouncing off the
walls. The filtering can be improved by a factor of 100, by adding baffles along the filter
section and by including a second bend to give a double bend or ‘S-bend’ filter [58].
The advantages of the FCVA are that it produces a highly ionised plasma with an
energetic species, a fairly narrow ion energy distribution, and high growth rates of 1 nm s -1 for
a low capital cost.
2.3.3. Pulsed laser deposition (PLD)
Pulse laser deposition, known well for over 2 decades, has gained prominence in the
deposition of a wide variety of thin film materials such as superconductors, semiconductors,
dielectrics, metals and biomaterials, among others. Pulsed laser deposition (PLD) (Fig. 2.12)
is a thin film deposition technique where a high power pulsed laser beam is focused inside a
vacuum chamber to strike a target of the desired composition. Pulsed excimer lasers such as
Ar give very short, intense energy pulses, which can be used to vaporize materials as an
intense plasma [59]. When the laser pulse is absorbed by the target, energy is first converted
to electronic excitation and then into thermal, chemical and mechanical energy resulting in
evaporation, ablation, plasma formation and even exfoliation.
The ejected species expand into the surrounding vacuum in the form of a plume
containing many energetic species including atoms, molecules, electrons, ions, clusters,
particulates and molten globules, before depositing on the typically hot substrate. The kinetic
energy of this expansion gives an ion energy analogous to the ion energy of the cathodic arc.
In this way, pulsed laser deposition produces ta-C films similar to those from the FCVA
methods [59-61]. In the pulsed laser depositions, the quality of the films is controlled by
several parameters such as laser parameters (laser fluence, laser energy, ionization degree of
ablated materials), surface temperature, substrate surface, background pressure, etc [59-61].
25
Fig. 2.12: Configuration of a PLD chamber [59].
2.4. Surface morphological characteristics of DLC films
Roughness evalution can be used for which a DLC film can be grown continuous and
pin-hole free. The knowledge of the surface evalution mechanism of DLC films allows one to
know if the loss of continuity in films is an intrinsic and unavoidable problem related to the
nature of the deposited film or if it is a technical problem, which could be improved with
better process condition.
Roughness is a measurement of the small-scale variation with the height of a physical
surface. Roughness is sometimes an undesirable property, as it may cause friction, wear, drag
and fatigue, but it is sometimes beneficial, as it allows surfaces to trap lubricants and prevents
them from welding together. Roughness evaluation studies can be conducted by atomic force
microscopy (AFM). Amplitude parameters, such as root mean square roughness value (Rq)
and arithmetic mean roughness value (Ra), are used in AFM studying of topography of DLC
films. Ra provides only information on the departure of the surface from the mean line and is
insensitive to extreme profile peaks and valleys. Although both Rq and Ra having the same
trend, the value of Rq is higher than that of Ra because Rq is more sensitive to peaks and
valleys due to the squaring function. In the studies of Zhang et al. [62] and Lifshitx et al. [63],
surface roughness changed with substrate bias, deposition time and C+ energy due to
bombarding effect. During the film deposition, increasing the target power or the pulse bias 26
leads to a higher electrical field, so that the carbon ions reaching the substrate surface possess
higher kinetic energy. When the kinetic energy reaches a certain value, the sp3 bonds
favorably formed result in the densification of DLC films and then a smooth surface can be
achieved. Therefore, the surface roughness (about 0.1 to 1.5 nm) of DLC films fabricated by a
FCVA system is lower than that (about 1 to 3 nm) of DLC films deposited by a sputtering
system because the FCVA system can produce the higher kinetic energy of the impinging
species. When the ion energy exceeds the optimum value, the surplus energy will cause
graphitization, so more sp2 bonds are formed in the form of graphic clusters spreading in the
sp3 matrix. The increased sp2 bonds in DLC films decrease the density of the films, leading to
the higher surface roughness. Zhang el al. [64] reported that average roughness increased with
annealing temperature due to heat treatment. Surface roughness of Si doped DLC films
decreases with increased Si content because the increased Si content decreases craters and
surface flaws distributed in the outer surface of the coating [65]. Maharizi et al. studied [66]
the dependence of the roughness of DLC films on the type of the substrates. The roughness of
the N doped DLC (N-DLC) films increases with increased N content. These features are
caused possibly by the nitrogen doping in the films or nitrogen normal bonding [67, 68].
Fig. 2.13: SEM images of ta-C films deposited at various substrate biases: (a) 0V, (b) 250 V, (c) 300 V and (d) 400 V [69].
27
Figure 2.13 shows the SEM pictures of ta-C films deposited at various substrate
biases. Carbon clusters are observed on the surfaces of ta-C films, and the sizes of carbon
clusters are varied with the substrate bias [69]. It is found that the lowest cluster size gives the
smoothest surface of the film.
The surface roughness of DLC films can affect the electrochemical properties of the
films because the low surface roughness (about 0.1 to 3 nm) of the films induces the low
double-layer capacitance at the film/electrolyte interfaces, resulting in an improvement of
signal-to-background ratio. The high surface roughness of the films can negatively affect the
corrosion resistance of the films because the high surface roughness may have a high
probability of defects or porosities that can lead to the dissolution of the underlying substrates
by allowing permeation of an electrolyte through the defects or porosities [29].
2.5. Adhesion strength of DLC films
Adhesion is an important property of the film/substrate system. Chemical bonding
between DLC films and their substrates is essential if the adhesion of the films is to be
adequate. Mechanical locking of the film to the surface of a rough substrate can play a part
but there is still a necessity for chemical bond formation if high degree of adhesion is to be
achieved. Generally, chemical bonding between the films and substrates is promoted by
deposition at high temperature whereas residual stress in the films is best controlled by
deposition at low temperature. The residual stress in DLC thin films consists of two main
components: thermal stress and intrinsic growth stress. Thermal stress is attributed to the
thermal expansion mismatch between films and substrates. The expansion of DLC is low
compared with many ceramics, even with Si and very low compared with most metals and
thus compressive thermal stress will be induced as the substrate temperature increases during
deposition. For many metallic alloys, where the expansion is too high, this leads to spallation
28
of the films due to the large compressive thermal stress unless the deposition temperature is
reduced. High temperature deposition also causes the DLC films to graphitization which in
turn reduces the thermal expansion because of reduced mechanical properties. There is thus a
compromise deposition temperature at which the adhesion of the films will be maximized;
this is clearly influenced by the intended operating temperature.
The intrinsic growth stress in DLC films is induced by enhanced cross-linkage and
bond distortions in the films caused by bombardment of energetic impinging species, which
abruptly affects the adhesion strength of the films. Therefore, the adhesion strength of DLC
films should be correlated to sputtering power density, chamber pressure and substrate bias.
The effects of the sputtering power, chamber pressure and substrate bias used during DLC
film depositions on the adhesion strength of the films can be explained in terms of kinetic
energy of the sputtered species which is proportional to the sputtering power density and bias
voltage and inversely proportional to the root square of the chamber pressure [70]. For
example, increasing the chamber pressure decreases mean free path of the ions; therefore
decreasing of the chamber pressure promotes the collision of ions onto the substrates which
leads to higher residual stress. The sputtering power and bias also impart more kinetic energy
to the ions [71, 72]. Not enough cleaning of substrate and landing of impurities from the
chamber walls and substrate holder on the substrate surface due to too high substrate bias are
also reasons degrading the adhesion strength of DLC films. As decreasing the kinetic energy
of the ions promotes the amount of sp2, an incorporation of nitrogen can also form new sp2
sites and encourages the sp2 sites to cluster [24]. As the sp2 bonds are shorter than the sp3
bonds, the increased sp2 sites would reduce the strain in the films. C=N bonds resulted from
nitrogen incorporation in the carbon matrix have the shortest bond length compared to C-C
and C=C [24]. It is reported [73] that depositing DLC films at above room temperature or
29
annealing them lower the residual stress in the films due to a conversion of sp3 bonded
configuration to sp2 bonded one. It is clear that the increased sp2 sites in DLC films improve
the adhesion of the films by reducing the residual stress in the films.
The scratch test is being increasingly used to qualitatively evaluate film adhesion. A
normal force at the moment of film detachment is called a critical load and gives a
comparative value of the film adhesion. In the study of Hedenqvist et al. [74], the maximum
normal force that the film-substrate system could sustain increased with increased film
thickness and substrate hardness. Hintermann et al. [75] proposed that the life of DLC coated
tools and their performance were considerably improved by high adhesive and cohesive
strengths of the DLC films because the bad adhesion leads to flaking while the poor cohesion
causes chipping. The effects of the interface topography, coating thickness and elastic
mismatch on the interfacial stress were investigated in the study of Wiklund et al. [76]. The
effect of the film thickness on the adhesion strength of DLC films was also investigated by
Perry et al. [77] and Sheeja et al. [78]. It was found that the adhesion strength of DLC films
was elevated by increasing the film thickness because the thicker films would need more
loads for the indenter to break through the films into the substrates. However, beyond a
critical film thickness, the critical load is reduced. This was likely due to weakening of the
film structures by the developed compressive stress in the films with increased film thickness.
The influence of deposition process parameters such as substrate temperature and bias, target
power and chamber pressure on the film adhesion was also found in the studies of Matinez et
al. [79], and Zhang et al. [71, 72]. In the study of Gupta et al. [80], the thinner film exhibited
instant damage when the normal load exceeded the critical load, whereas the thicker films
exhibited gradual damage through the formation of tensile cracks. The introduction of H, N,
30
F, Si and metals (e.g. Ti, V) into DLC films apparently reduces the residual stresses in the
films via degraded cross-linking structures [25-27].
It is difficult to quantitatively express the adherence because the critical load depends
on several parameters related to the testing and to the coating system. Both intrinsic
parameters, such as scratching speed, loading rate, diamond tip radius and shape, diamond
wear, and extrinsic parameters, such as substrate hardness, coating thickness, substrate and
coating roughness and friction coefficient, are considered in order to improve the
interpretation of the critical load results [81].
2.6. Electrochemistry
Introduction
Electrochemistry is the study of phenomena caused by charge separation. It deals with
the study of charge transfer processes at the electrode/solution interface, either in equilibrium
at the interface, or under partially or totally kinetic control. Most of the charge transfer
processes are transfer of electrons, which can be represented in the simplest case of oxidized
species, O, and reduced species, R, by
On+ + ne- ↔ R (2.1)
where O receives n electrons in order to be transformed into R.
2.6.1. Corrosion mechanisms
Generally, corrosion is a process that takes place when essential properties within a
given material begin to deteriorate, after exposure to elements that recur within the
environment. Corrosion can be separated into two types: (a) uniform or general corrosion, and
(b) localized corrosion. General corrosion uniformly occurs over the entire film surface.
Localized corrosion occurs at a small discrete location on a film surface and is usually
31
characterized by rapid, deep penetration through the film. Localized corrosion can be several
orders of magnitude faster than general corrosion. Corrosion reactions are often complex
heterogeneous reactions whose rates are typically determined by several factors such as (a)
usual kinetic considerations, (b) electrolyte chemical composition, (c) mass transfer between
electrolyte and material surface, and (d) various surface effects such as adsorption/deposition
and surface roughness.
Fig. 2.14: Electrical double layer for electrode submerged in an electrolyte [82].
The corrosion phenomenon can be explained by electrical double layer (EDL) model
as shown in Fig. 2.14. Electrode (film) ions leave their structure when an electrode is
submerged in an electrolyte. Water molecules surround the electrode ions as they leave and
the hydrated ions are free to diffuse away from the electrode. The negative charges (caused by
excess electrons) on the electrode surface attracts positively charged ions and a percentage of
them remains near the surface, instead of diffusing into the bulk electrolyte. The water layer
around the ions prevents most of them from making direct contact with the excess surface
electrons. The positive ions in the electrolyte are also attracted to the negatively charged 32
electrode surface. Consequently, the electrolyte layer adjacent to the electrode surface
contains water molecules and ions, and has a distinctly different chemical composition than
the bulk electrolyte. The negatively charged surface and the adjacent electrolyte layer are
collectively referred to as the electrical double layer (EDL) [82].
Fig. 2.15: Cyclic polarization curve [82].
A corrosion reaction can proceed in both the forward and reverse directions and
equilibrium is achieved when forward and reverse reaction rates are equal. There is no net
loss of electrode atoms when an EDL is at equilibrium, because electrode atoms are oxidized
at the same rate that electrode ions are reduced. When corrosion occurs, electrochemically
active species, which can be reduced by the excess electrons such as hydrogen ion reduction,
diffuse from the bulk electrolyte to the electrode surface and discharge the EDL at the point
on the electrode surface where electrons are removed, causing more electrode atoms to leave
in an effort to re-establish original EDL conditions.
The Nernst equation mathematically relates EDL composition to electrical potential [82]:
E=Eo- (RT/nF) ln(ao/ar) (2.2)
33
E: measured potential (V); Eo: open circuit potential when all species have unit activity; a:
concentration of chemical species; R: ideal gas constant (8.314 J/K.mol); T: temperature (K);
n: number of electrons in anodic half reaction; F: Faraday constant, (96485 C/mol) [82].
The cyclic polarization curve shown in Fig. 2.15 is used for a brief overview of DC
polarization corrosion measurement methods. The applied potentials are plotted as a function
of electrical current density.
2.6.1.1. Linear polarization
Electrode polarization causes certain processes to occur in an electrolyte that can limit
electrical current during electrochemical corrosion. The working electrode polarization is
controlled by a potentiostat supplying electrons to either the counter or working electrodes.
There are many factors other than potentiostat current that can control the magnitude of
electrode polarization. These are: (a) working electrode chemical composition, (b) working
and counter electrode surface condition, (c) working and counter electrode geometrical
shapes, (d) counter electrode size, (e) electrical double layer chemical composition, and (f)
electrolyte chemical composition.
Fig. 2.16: Linear polarization [82].
Working electrode electrical current is zero at open circuit potential (OCP) and
electrode potential polarity switches from cathodic to anodic as the scan proceeds past the
34
OCP. The slope for a linear polarization curve as shown in Fig. 2.16 is the change in potential
divided by the corresponding change in current density. This relation is written
mathematically as [82]:
Rp=∆E/∆i (Ωcm2) (2.3)
Rp is referred as the corrosion or polarization resistance.
2.6.1.2. Potentiodynamic polarization
A potentiodynamic polarization curve has a cathodic branch that is similar to that for a
Tafel plot. It also has an anodic branch, but the anodic branch extends over a wider potential
range and is often much more complex than the Tafel plot anodic branch.
Fig. 2.17: Potentiodynamic polarization curve [82].
Several quantities appear in the potentiodynamic polarization anodic branch as shown
in Fig.2.17. Epp is the primary passivation potential after which current either decreases, or
becomes essentially constant over a finite potential range. Eb is the breakdown potential where
current increases with increasing potential. The passive region is the portion of the curve
between Epp and Eb. The active region of the curve is the portion of the PDS curve where
35
potentials are less than Epp. The transpassive region of the curve is the portion of the curve
where potentials are greater than Eb.
Quantities like Epp, Eb and passive region width can be used to characterize the
corrosion behavior, and evaluate how a passive film effectively protects a film from
corrosion. General corrosion, and sometimes pitting, occurs in the active region; little or no
corrosion occurs in the passive region, and pitting corrosion can occur in the transpassive
region.
2.6.1.3. Tafel plot
A Tafel plot shown in Fig. 2.18 has anodic and cathodic branches, corresponding to
the anodic and cathodic half reactions for corrosion. Tafel plots are classified as activation
controlled when the corrosion rate is determined by how fast an electrode is capable of
transferring its electrons to electrochemically active species (EAS). A characteristic of
activation control is an increasing current density magnitude with increasing potential for both
branches. Tafel plots are classified as diffusion controlled when an EAS diffusion rate
determines the corrosion rate. Diffusion control theoretically causes cathodic current density
to become constant. The constant, or slowly changing, cathodic current is referred to as the
diffusion limited current. Diffusion can restrict access of the EAS to an electrode surface
when: (1) an electrolyte has a limited supply of EAS (eg. At pH 14, the hydrogen ion
concentration is 10-14 M), (2) EAS diffuses very slowly through the electrolyte to an electrode
surface, and (3) corrosion reaction products restrict EAS access to an electrode surface [82].
36
Fig. 2.18: Tafel plot [82].
Tafel equation [82]:
∆V = β × ln (Icorr / Io) (2.4)
where ∆V: the overpotential, (V); β: Tafel slope, (V/decade); Icorr: the current density, (A/cm2);
Io: the exchange current density, (A/cm2).
The corrosion current (Icorr) can be calculated using the following equation [82, 83]:
Icorr = [βa × βc] / [2.3 × Rp × (βa + βc)] (2.5)
where Rp: the polarization resistance (Ωcm2), Rp = [βa × βc] / [2.3 Icorr (βa + βc)]; βa: the anodic
Tafel slope, (V/decade); βc: the cathodic Tafel slope, (V/decade); (βa × βc) / (βa + βc) = Tafel
constant.
2.6.1.4. Electrochemical impedance spectroscopy
Diamond-like carbon (DLC) thin films can be deposited by physical vapor deposition
(PVD) and chemical vapor deposition (CVD) to get attractive properties for electronic,
optical, mechanical, biomedical and electrochemical applications. However, there are almost
always pores, even in good quality PVD or CVD DLC films [84]. Pores with dimensions of
approximately 0.5 nm are randomly distributed in the films, and water molecules with
dimensions of approximately 0.32 nm can easily permeate into the film [84]. DLC films are
37
usually electrochemically nobler than the substrates, so the presence of nanopores in the films
can rapidly lead to the electrochemical dissolution of the substrate. Although atomic force
microscopy (AFM) can measure such small sizes, it is not easy to locate the sites of nanopores
in a large area of measurement. Therefore, detecting the nanopores and assessing their effect
on the corrosion protection behavior of DLC films holds the key to fully understand the
electrochemical behavior of the films.
It is difficult to measure the pores on nanometer scales, even with a scanning electron
microscope (SEM). In the polarization method, a partially coated surface exhibits a mixed
potential when exposed to a corrosive electrolyte, depending on the number of pores and the
degree of polarization of each electrode component. In the case where a film has pores, the
polarization result is a combined output of the film and substrate. The polarization behavior of
DLC film itself is not clear so the polarization method is not suitable to measure the porosity
in DLC films. It is reported [84] that electrochemical impedance spectroscopy (EIS) is a
sensitive technique in detecting nanopores. Though the polarization study can provide data
about the corrosion processes occurring at the electrochemical interface, AC methods, such as
electrochemical impedance spectroscopy (EIS), offer potentially more information especially
regarding the performance of coatings with small defects on active substrates. EIS can be used
to quantify coating integrity and long term corrosion performance. Impedance spectroscopy is
based on applying a small sinusoidal potential modulation to the electrode system and
monitoring the amplitude and phase shift of current response over a range of frequencies. This
technique has been successfully applied to the evaluation of the protective character of
polymer on metal as well as dielectrics such as Al2O3 on steel [85-87].
38
Fig.2.19: (a) An equivalent circuit proposed for a solid coating. (b) Nyquist plot and (c) Bode magnitude and (d) phase plots of the coating [82].
Figure 2.19a shows an equivalent circuit of a perfect coating without defects or
porosities. The circuit includes a solution resistance (Rs), a charge transfer resistance (Rct) and
a double layer capacitance (Cdl). The total impedance magnitude for the circuit can be
mathematically expressed [88]:
Z = Rs + [Rct / (1 + ω2Rct2Cdl
2)] + [j(ωRct2Cdl) / ( 1 + ω2Rct
2Cdl2)] (2.6)
where Z is the total impedance in Ω, Rs and Rct are the solution and charge transfer resistances,
respectively, C is the capacitor capacitance in F, ω=2π (AC voltage frequency) and j is the
square root of -1 and referred to as an imaginary number. In equation 2.6, the second term,
[Rct / (1 + ω2Rct2Cdl
2)], is referred to as the real impedance and the third term, [j(ωRct2Cdl)/( 1 +
ω2Rct2Cdl
2)], is referred to as imaginary impedance. The imaginary impedance magnitude is
zero when the phase angle is zero. The imaginary impedance magnitude is infinity when the
phase angle is 90 degree.
39
In the Nyquist plot shown in Fig. 2.19b, the Rs can be found by reading the real axis
value at the high frequency intercept. This is the intercept near the origin of the plot. The real
axis value at the other (low frequency) intercept is the sum of the Rct and Rs. The diameter of
the semicircle is therefore equal to the Rct.
The Rs and the sum of the Rs + Rct can be read from the Bode magnitude plot shown in
Fig. 2.19c. Plot slope is zero when polarization is through resistances and the slope is less
than zero (negative) when capacitive reactance becomes part of the circuit response to a
polarization. The Rs, Rct and capacitive reactance are also noted in the Bode phase plot shown
in Fig. 2.19d. The inflection point frequency shown in Fig. 2.19d is equal to [88]:
Inflection point frequency = 1 / (1.77RctCdl) (2.7)
Comparing Bode magnitude and phase plots shows that phase plot inflection
corresponds to the area where the Bode magnitude slope is negative. Comparing phase and
magnitude plots also illustrates that magnitude plot resistances, Rs and Rct, correspond to
polarizing voltage frequencies where phase angle is zero. Organic coating capacitive
reactance and resistance properties can produce a second circuit, which contains bulk
resistivity (Rp) and capacitance (Cc) of the coating, in an equivalent circuit shown in Fig.
2.20a and appear a high frequency circle in the Nyquist plot shown in Fig. 2.20b. The bulk
conductivity of the coating can be attributed to its original conductivity and the electrolytic
conduction when the coating is composed of porosities that allow migration of ions to the
substrate [82]. Water and ions typically diffuse into an organic coating, after a coated sample
is submerged in an electrolyte, and change coating dielectric properties. Water and ions also
move inside a coating in response to AC polarizations. Therefore, water and ion movement
through a coating causes the electrolytic conduction in the bulk of the coating and is also
restricted by coating morphology, producing a coating, or pore resistance.
40
Fig. 2.20: (a) An equivalent circuit proposed for a coating with porosities. (b) Nyquist plot and (c) Bode magnitude and (d) phase plots of the coating [82].
Notice that there are two semicircles in the Nyquist plot shown in Fig. 2.20b; one
comes from the electrochemical reactions in interface region between the coating and
electrolyte at low frequencies and one is attributed to bulk electrical properties of the coating
at high frequencies. The associated magnitude plot has two slopes in Fig. 2.20c and the
associated phase plot in Fig. 2.20d has two inflection points.
2.6.2. Corrosion properties of DLC films
Diamond-like carbon (DLC) is an attractive candidate material for devices which need
to withstand exposure to a range of harsh environment. The unique combination of high
hardness, high wear resistance, low friction, electrical insulation, high corrosion resistance
and chemical inertness make DLC films ideal for protective coatings in corrosive
41
environment. Diamond is resistant to all acids, even at elevated temperatures although it can
be etched by fluxes of caustic alkalis, oxysalts, etc. The polycrystalline nature of diamond
films and small pores or pinholes within them are the major drawbacks of the diamond films.
Chemical attack at the grain boundaries or attack of the underlying substrate through pinholes
lead to unacceptable performance in many coating systems and diamond is not exceptional.
However, DLC is amorphous and has no grain boundaries compared to diamond, which is a
great benefit for the high corrosion resistance of DLC.
DLC films can potentially be used in applications requiring good corrosion resistance,
such as protection of metals in magnetic recording and micro electronics industries because of
their intrinsic stability under most aqueous conditions. Corrosion of metals with DLC
overcoats is usually initiated at microscopic pinholes that lead to the exposure of the metals to
the environment. Requirements for effective protection include chemical inertness together
with microstructural demands, such as high density, smooth surface without microporositites
and diffusion pathways, homogeneous stoichiometry, stress-free layers and good adhesion
between the film and the substrate material. The electrical conductivity and porosity level of
the carbon overcoat promote a galvanically-induced corrosion mechanism between the
overcoat and the substrate. To minimize the galvanic action between the substrate and the
overcoat, it is necessary to develop overcoats with high electrical resistance, in an ideal case
with dielectric properties. However, the electrodes used for electrochemical purposes, such as
tracing heavy metals, must need to have high electrical conductivity, clearly pointing out that
the best balance should be carefully optimized.
Morrison et al. [89] reported that DLC-metal composite films (Ag-DLC, Pt-DLC,
AgPt-DLC) exhibited low corrosion rate in phosphate buffered saline (PBS) electrolyte. The
ultra-thin ta-C films showed superior corrosion protective properties with respect to the
42
sputtered film when deposited on a Co-alloy disk (Co78Cr12Pt10) [90]. DLC films reduced the
corrosion rate of steel substrates in corrosive solution (3.5 wt% NaCl) by functioning as a
physical barrier and restraining from anodizing [91]. Liu et al. [92] showed that DLC films
could significantly improve anti-corrosion properties of AlTiC substrates when the film
thickness was more than a few tens of nanometers [80]. After potentiodynamic corrosion tests
of H-DLC and DLC films in 0.5 M NaCl aqueous solution were carried out, comparison
between the corrosion parameters of the DLC:H and DLC coated Ti-6Al-4V alloys showed
that DLC film presented higher corrosion potential and polarization resistance than those of
the H-DLC film [93]. The corrosion studies indicated that coating Ti (Ti-13Nb-13Zr) alloy
with DLC film could improve the corrosion resistance in simulated body fluid environment
[94]. Introduction of Si into DLC films led to a significant improvement in the corrosion
resistance of the films in 2 M HCl solution, as revealed by an increase in the charge transfer
resistance and a decrease in the anodic current [95]. An increase of bias voltage could
improve corrosion resistance of the films in a simulated body fluid (0.89 % NaCl solution).
This could be attributed to the formation of dense and low-porosity films which impeded
permeation of the solution [96]. Zeng et al. [84] proved that DLC films could serve as
protective layers over their underlying Si substrates in sulfuric acid solution because the films
with few nanopores could effectively prevent their underlying silicon substrates from
corrosion. Sharma et al. [97] found that increasing deposition time increased the corrosion
resistance of the films, showing that the increased thickness of the films improved the
corrosion resistance of the films by lessening the possible number of the porosities in the
films [97]. Liu et al. [98] reported that adhesion strength had a great influence on the
effectiveness of corrosion protection since the porosities in the films could allow the
electrolyte to permeate to the interfaces between the films and substrates and attacked the
43
interfacial bonds between them; the better the adhesion strength of the films was, the better
the corrosion performance of the films was. Papakonstantinou et al. [99] found that the
corrosion resistance of ultrathin DLC films in corrosive solution substantially increased with
immersion time due to filling of the pores with passivating materials and subsequently
stopping access of the solution to their substrates. It can be seen that the porosity density is
one of the most important parameters assessing the effectiveness of corrosion protection of
DLC films [25-27, 100, 101].
2.7. Electrode materials for electroanalysis
A number of experimental design factors have to be considered if it is decided to
perform an electrochemical analysis. These design factors depend on the technique employed,
electrode material, and electrode and cell configurations. The useful potential ranges of
electrode materials are determined by oxidation and reduction of a solvent, decomposition of
a supporting electrolyte, electrode dissolution or formation of a layer of
insulating/semiconducting substance on its surface. Electrode materials for voltammetry must
conduct electrons. Thus, their choice is limited to metals, other solids with metallic
conductivities and good semiconductors. Usually, it is also desired that the electrode material
is inert in the region of potential in which the electroanalytical determination is carried out.
The reliability and repeatability of experiments can be aided by assuring a constant
flux of electroactive species to the electrode. This is done by using controlled convective flow
over the electrode or by creating a sufficient high concentration gradient. The additional
advantage of this approach is that, because of the greater mass transport, sensitivity is
increased and detection limits are lowered.
Detection limits can also be affected by other electrode reactions which can occur in
the same potential range. The most prevalent of these is the reduction of oxygen since its
44
solubility in solutions open to the atmosphere is up to 10-4 M. Oxygen must be removed from
the solution by passage of an inert gas, prepurified nitrogen or argon, through the solution to
diminish the oxygen partial pressure to a very low value. The poisoning of electrode surfaces
has been one of the main limitations to the widespread use of electroanalysis by non-experts.
Mercury was the first metal to be extensively used, in the form of a dropping mercury
electrode; however, mercury’s useful potential range is limited by its oxidation which means
that, essentially, only reductions can be investigated. Therefore, solid electrode materials are
developed which permit oxidation reactions to be studied.
Sp2-carbon is a versatile material that has a wide range of applications in
electrochemistry. Because of its reasonable electrical conductivity and good corrosion
resistance, carbon has found widespread acceptance in electrodes.
Platinum possesses high resistance to chemical attack, excellent high-temperature
characteristics, and stable electrical properties. All these properties have been exploited for
industrial applications. Platinum does not oxidize in air at any temperature, but can be
corroded by cyanides, halogens, sulfur, and caustic alkalis. This metal is insoluble in
hydrochloric and nitric acid, but does dissolve in the mixture known as aqua regia. Pt
dissolution is more severe in phosphoric acid than in perchloric acid. Because of its chemical
inertness, it is chosen as electrodes for electroanalysis. However, in aqueous solutions,
detection of analyses is often not possible at negative potentials using such electrodes due to
the high Faradic currents produced by hydrogen evolution reaction. Mercury electrodes can
eliminate this problem of Faradic currents. However, the use of the mercury results in the
generation of hazardous waste.
Sp2-carbon, e.g. glassy carbon, is widely used for electroanalysis in aqueous media, as
it exhibits a relatively wide potential window. Glassy carbon has a number of properties,
45
including high conductivity, impermeability, and unreactivity, which make it an excellent
electrode material. However, this material has serious limitations, including high background
currents and deactivation via fouling. This is due to the irreversible absorption of product
formed in the electrochemical oxidation reaction. It is an inherent property of glassy carbon to
undergo deactivation upon exposure to the laboratory environment or working solution, which
is due to factors such as oxidation and adsorption of contaminants and reaction products. The
porous nature of the electrode may also complicate the voltammetric response since redox
sites which lie in deep pores will be subject to greater uncompensated solution resistance than
sites close to the outer surface of the electrode [102].
Fig. 2.21: Cyclic voltammetric i-E curves for boron doped diamond electrodes in (a) 1 M KCl and (b) Fe(CN)6
-3/-4 + 1 M KCl [11].
The attractive features of diamond films include a wide potential window in aqueous
media, very low capacitance, and high electrochemical stability. In addition, diamond
electrodes have recently been found to exhibit additional characteristics of interest for electro-
analysis. The diamond electrode material is stable with respect to platinum, gold, palladium,
and silver and shows a wide potential range for the water decomposition in various
electrolytes such as NaCl, H2SO4, HNO3, HCl, KOH, KNO3, and Na2SO4 [103]. Hupert et al.
[11] studied cyclic voltammograms obtained at diamond electrode in 1 M KCl and Fe(CN)6-3/-4
+ 1 M KCl solutions as found in Fig. 2.21.46
Boron-doped diamond thin-films possess a number of important and practical
electrochemical properties, unequivocally distinguishing them from other commonly used sp2-
bonded carbon electrodes, such as glassy carbon, pyrolytic graphite, and carbon paste. These
properties are (i) a low and stable background current, leading to improved signal-to-
background (SBR) and signal-to-noise (SNR) ratios; (ii) a wide working potential window in
aqueous and non-aqueous media; (iii) superb micro-structural and morphological stability at
high temperatures (e.g. 180 ˚C); (iv) good responsiveness for several aqueous and non
aqueous redox analytes without any conventional pretreatment; (v) weak adsorption of polar
molecules, leading to improved resistance to electrode deactivation and fouling; (vi) long-
term response stability [11, 104]. Electrically conductive diamond films yield film electrodes
of remarkable properties. The diamond conductivity basically comes from either the damage
generated sp2 (graphite) contents in case of ion-implanted diamond or the doped-boron in case
of CVD-diamond films [105]. The latter films have shown an exceptionally wide
electrochemical window, a low capacitance charging background current (3.7 to 7.1 µF/cm2)
for aqueous systems and remarkable stability comparable with the chemical resistance of pure
carbon diamond [106, 107]. Thus, diamond is becoming an interesting material to consider for
electrolysis. However, some restraints upon its applications arise, such as high substrate
temperature required (typically 1175 K), relatively small film area presently available, and
relative difficulty in deposition parameter control limit.
Compared to diamond, DLC materials are also chemically stable and can achieve a
wide range of excellent mechanical, electrical and chemical properties. Wide cyclic
voltammograms shown in Fig. 2.22 indicate that N-DLC films are promising electrodes for
electrical applications [97]. In the study of Zeng et al. [20], the repeatability of the
voltammogram showed the durability of N-DLC film to high anodic potential, which was
47
identified by scanning the voltammogram more than 20 times with essentially identical
profile. The wide electrochemical window (approximately -1.20 to +2.21 V vs. SCE) (Fig.
2.20b) of the N-DLC film electrode was comparable with that of diamond film electrode,
which is about -1.25 to +2.30 V (vs. SHE) [108]. For the graphite electrodes, the potential
window is significantly smaller. The potential window of about -0.30 to +1.80 V was reported
for glassy carbon electrode [109]. For a highly oriented pyrolytic graphite (HOPG) electrode,
it was about -0.40 to +1.60 V [109].
Fig. 2.22: Cyclic voltammetric i-E curves obtained (a) N-DLC electrode (dash line) and glassy carbon electrode (solid line) in 1 mM K3Fe(CN)6/0.1 M H2SO4 and (b) N-DLC electrode in 0.5 M H2SO4 [20].
Surface cleanliness is an important parameter influencing the responsiveness of all
electrodes. In a very general way, adsorbed contaminants can either block specific surface
sites, thus inhibiting surface sensitive redox reactions, or increase the electron-tunneling
distance for redox analytes, thereby lowering the probability of tunneling (i.e. the rate of
electron transfer). DLC films are not as susceptible to contamination (air or solution borne
contaminants), as other electrodes are, because of their hydrophobic surfaces. In addition,
DLC films can be deposited on various substrates like metals, ceramics, polymers and
semiconductors at low temperature and achieve the following characteristics:
48
• Conformal and well-adherent,
• Pinhole free,
• High corrosion resistance,
• Low coefficient of friction,
• High thermal conductivity,
• High hardness,
• Excellent biocompatibility,
• Wide electrochemical potential window and
• Controllable resistivity.
All processes required for production of DLC film electrodes are compatible with
semiconductor technologies. Thus, DLC film electrodes can directly be amenable to mass
production. These characteristics demonstrate a great promise of using conductive DLC as an
environmental-friendly disposable electrode material.
2.8. Electrochemistry of DLC films
Nowadays, special interest has been devoted to the preparation of optically transparent
carbon electrodes and the production of microelectrodes by PVD techniques such as electron
beam evaporation and sputtering [110, 111]. Sputtered carbon films at room temperature are
generally amorphous. Their surface smoothness can be obtained near atomic scale, which
would induce a low capacitance charging background current if the film is used for
electroanalysis. However, depending on the sputtering methods and the deposition
parameters, a large variety of microstructures can be obtained, ranging from graphite-like to
diamond-like carbon. This is reflected in the considerable variability of physical, mechanical
and electrochemical parameters. Diamond-like carbon has been used for several applications
49
such as mechanical, optical, electrical, tribological, biological and electrochemical
applications.
Furthermore, DLC can be thought of as a more universal application especially when
large surface area electrodes are required, because uniform and smooth DLC films can be
deposited over large surface areas of substrates without grain boundaries. Therefore, the
unique combination of variable electrical conductivity and diamond-like chemical and
mechanical properties opens the possibility for applications of these materials in several areas
such as microelectronics and electrochemical application, such as electrodepositing anode and
analytic electrode, etc.
Although there are many reports on the mechanical and electrical properties of DLC
films, limited reports have been published on their electrochemistry. It is said that DLC films
are chemically inert under biochemical conditions, and a good passivation coating for
engineering materials (Ti, Al/C and stainless steel) [112, 113]. Some successful applications
of DLC films as protective coatings for biomaterials have been reported [114].
Fig. 2.23: Bright-field transmission electron micrograph of a Pt-DLC composite film. Platinum nanoparticles self-assemble into arrays (dark regions) within the DLC matrix (light regions) [89].
In Pt incorporated DLC films (Fig. 2.23), Pt usually cannot be evenly doped, and does
not act as carriers [89, 115]. A wide-band gap diamond-like carbon is not in itself
50
electrochemically active; however, it acquires electrochemical activity upon introducing
platinum into the DLC bulk during the film deposition [116].
Schlesinger et al. [117] studied the potential limits of Ti-doped amorphous carbon thin
film electrodes with diamond-like character, deposited by sputtering, in several kinds of
aqueous solutions. They found that the films exhibited a low double-layer capacitance, a large
electrochemical window, and a relatively high activity toward ferricyanide reduction,
comparing with conventional carbon electrodes such as glassy carbon.
Nickel doped DLC (Ni-DLC) thin film coated samples were used as working
electrodes to electrocatalyze glucose oxidation in 0.1 M NaOH aqueous solutions [118]. It
was found that direct electrochemical response of glucose at the Ni-DLC thin film electrodes
gradually developed with increased glucose concentration in the electrolytic solutions. It was
deduced that the Ni nanoparticles distributed on the Ni-DLC film electrode surfaces played a
major role in promoting the glucose oxidation [118].
Yoo et al. [119] investigated electrodes made of N incorporated DLC (N-DLC) thin
films. They found that the electrodes exhibited excellent electrochemical characteristics. They
had (1) wide potential windows in aqueous system than those of boron doped diamond, (2)
reversible and excellent analytical behavior, (3) electron-transfer kinetics intermediate
between polished pyrolytic graphite and Boron doped diamond for the quinine/hydroquinone
couple and excellent analytical behavior, and (4) considerably higher catalytic activity for
Cl2/Cl- than that of boron doped diamond as well as durability to high anodic potentials. N-
DLC films showed wide potential windows in different aqueous solutions and had high
signals to heavy metals such as lead, copper and mercury at µM level [19, 20, 23, 24]. Lagrini
et al. [120] found that nitrogen partial pressure influenced potential windows of N-DLC films
in LiClO4 solution.
51
When these good characteristics of DLC films are combined with the advantage of
ready deposition at ambient temperatures, the properties of DLC suggest valuable additions to
the repertoire of electrochemistry. A more detailed examination of the preparative parameters
in relation to better understanding of electrode behavior is required before broad and practical
applications of DLC coated electrodes can be realized.
2.8.1. Cyclic voltammetry
Cyclic voltammetry is a type of potentiodynamic electrochemical measurement. In a
cyclic voltammetry experiment, a potential is applied to the system and the Faradic current
response is measured and plotted versus the applied potential to give the cyclic
voltammogram as shown in Fig. 2.24. The current response over a range of potentials (a
potential window) is measured, starting at an initial value and varying the potential in a linear
manner up to a pre-defined limiting value. At this potential (often referred to as a switching
potential), the direction of the potential scan is reversed, and the same potential window is
scanned in the opposite direction. This means that species formed by oxidation on the first
(forward) scan can be reduced on the second (reverse) scan. This technique is commonly
used, since it provides a fast and simple method for initial characterization of a redox-active
system. In addition to providing an estimate of the redox potential, it can also provide
information about the rate of electron transfer between the electrode and the analyte, and the
stability of the analyte in the electrolyzed oxidation states.
Voltammogram can be obtained by measurement of the current response as a function
of applied potential (or the potential response as a function of applied current); in other words,
they depend on the registering of current-potential profiles. Potential sweep methods consist
of scanning a chosen region of potential and measuring the current response arising from the
electron transfer and associated reactions that occur. They are widely used for the
52
investigation of electrode processes, which is a first step towards developing an
electroanalytical procedure. Naturally, they can also give quantitative information, since the
currents obtained are directly proportional to concentration.
Cyclic voltammetry (CV) is one of the most frequently used electrochemical methods
because of its relative simplicity and high information content. CV has a great advantage for
elucidating the mechanism of electrode reactions that are complicated by chemical (C)
reactions that either precede or follow the electron (E) transfer step. There are many possible
reaction mechanisms: EE, EC, CE, and ECE, etc. which are important to consider when
multiple peaks or strange-looking CV waves are experimentally encountered.
Fig. 2.24: Typical cyclic voltammogram [9].
In cyclic voltammetry, the basic response to potential is a peak-shaped curve. The
current begins to rise as potentials are reached where electrode reaction can occur. This
creates a concentration gradient which sucks in more electroactive species until depletion
effects set in and the current begins to fall again. The total current obtained is a sum of a
contribution from a Faradic current and a residual current. The Faradic current is associated
with the redox activity of the analyte. The residual current is primarily composed of four
components: (1) expected faradic current from electrochemical reactions, (2) charging
current, (3) variable current from electrochemical reactions of the electrode surface, and (4)
53
trace currents from electroactive impurities in solution. The first effect can be explained by
electrochemical rate theory which provides that finite current flows exist at potentials
considerably removed from the point where the sharp rise in oxidation or reduction current
occurs. The second effect comes from the charging or capacitive current resulting from
charging the double layer associated with the electrode-solution interface. The third residual
current results from extraneous processes primarily associated with the electrode surface.
Thus, according to the pass history of the electrode, the surface may have become oxidized or
covered with a layer of absorbed gas. Potentials applied to these electrodes may then give rise
to currents from dissolution of the oxides or gas films. In addition, no background electrolyte
is normally free of traces of oxidizable and reducible impurities. These substances contribute
to the overall residual current.
In cyclic voltammetry of reversible reactions, the product of the initial oxidation or
reduction is then reduced or oxidized, respectively, on reversing the scan direction. If a redox
system remains in equilibrium throughout the potential scan, the electrochemical reaction is
said to be reversible. In other words, equilibrium requires that the surface concentrations of O
and R are maintained at the values required by the Nernst Equation. For completely
irreversible reactions only the oxidation or reduction corresponding to the initial sweep
direction appears, since re-reduction or re-oxidation, respectively, cannot occur, there is no
reverse peak. The majority of redox couples fall between the two extremes and exhibit quasi-
reversible behavior. This means that the reverse peak appears but is smaller than the forward
peak.
In aqueous solution, usually, the ions except H+ and OH- are to be detected. It requires
that over potentials for hydrogen and oxygen evolutions are larger than that of expected ion
redox. The difference between the potentials for hydrogen and oxygen evolution is called
54
electrochemical potential window. The cyclic voltammetry is used to measure the potential
windows of electrodes: the wider the potential window is, more elements in solution can be
detected.
2.8.2. Anodic stripping voltammetry
Pollutants in water include a wide spectrum of chemicals, pathogens, and physical
chemistry or sensory changes due to the contamination, over-use and mismanagement of
water resources [25]. Many of the chemical substances are toxic. The presence of toxic metals
such as mercury, lead and copper in the environment has been a source of worry to
environmentalists, government agencies and health practitioners because these metals in the
aquatic ecosystem far-reaching implications directly to the biota and indirectly to human
beings. Tracing and determination of these poisonous metals in solutions are of major
importance in electrochemical analysis. Although electrochemical stripping voltammetric
measurement is a simple, quick and cheap way of tracing metals, their sensitivity changes
with electrode properties (electrical conductivity, surface roughness, surface cleanliness),
operation parameters (scan rate, deposition potential, deposition time, methods) and
environmental parameters (current flow, pH value of solution, metal concentration, metal
solubility).
Anodic stripping voltammetry (ASV) is a voltammetric method for quantitative
determination of specific ionic species. The ASV has been widely used for detection of heavy
metals in various solutions because of its remarkly low detection limits [121, 122]. Linear
sweep anodic stripping voltammetry (LSASV) is one of the anodic stripping voltammetric
methods to trace heavy metals. In linear sweep voltammetric (LSV) measurement, the current
response is plotted as a function of voltage rather than time, unlike potential step
measurement. The LSV technique shows better defined and selective peaks. This is due to
55
less interference of species that can be absorbed by the working electrode, contrasting with
differential pulse and square wave voltammetric techniques [123].
Anodic stripping voltammetry usually incorporates 3 or 4 steps. In the first step, the
potential is held at a lower potential that is enough to reduce the analyte and deposit it on the
electrode. Its efficiency depends on the rate of transport of the species to be accumulated onto
the electrode surface; a constant rate of transport will lead to better reproducibility and
repeatability and a linear dependence on accumulation. Thus, constant stirring is used in order
to increase sensitivity and decrease detection limit. After the first step, the electrode is kept at
the lower potential. The purpose of this second step is to allow the deposited material to
distribute more evenly in the mercury. If a solid inert electrode is used, this step is
unnecessary. The third step involves raising the working electrode to a higher potential
(anodic), and stripping (oxidizing) the analyte. As the analyte is oxidized, it gives off
electrons which are measured as a current. In this step, potential control used in stripping
voltammetry leads to a current peak whose height (and area) is proportional to the
concentration of the accumulated species. The last step is a cleaning step; in the cleaning step,
the potential is held at a more oxidizing potential than the analyte of interest for a period of
time in order to fully remove it from the electrode.
2.8.3. Stripping analysis of DLC films
Anodic stripping voltammetry (ASV) has been widely used for detection of heavy
metals in various solutions because of its remarkably low detection limits. Other advantages
of stripping voltammetry include the capability of simultaneous multi-element determination,
and relatively inexpensive instrumentation as compared with spectroscopic techniques used
for trace metal analysis. In addition, its low operating power makes them attractive as portable
and compact instruments for on-site monitoring of trace metals.
56
In the past, mercury-coated carbon-based electrodes such as graphite and glassy
carbon (GC) have been used for ASV extensively [124, 125]. The advantage of the mercury
electrode is to increase the negative potential range available, particularly important in the
case of the commonly analyzed zinc; the elements are dissolved in or form an amalgam with
the mercury. However, the major drawbacks of the mercury electrode are disposal problems
and cost. Moreover, mercury toxicity has been a great concern of environmentalists and has
motivated the sensor technology to develop mercury-free electrodes.
In recent years, boron-doped diamond film electrodes have been proposed as suitable
electrodes for the stripping analysis of heavy metals [126-128]. A diamond film electrode has
been developed. It consists of a thin film of carbon (with a preponderance of sp3 carbon
geometry) deposited on an inert substrate such as silicon by chemical vapor deposition
(CVD). This carbon is doped with boron, which has a high conductivity value. The boron
doped diamond (BDD) electrode has several qualities, such as a large electroactivity range in
water and a low background current (one magnitude lower than the one generated by glassy
carbon electrode). Furthermore, chemical and mechanical robustness of this electrode made it
suitable for working in corrosive media.
Fig. 2.25: Typical DPASVs obtained at N-DLC film electrode in an 0.1 M KCl (pH 1.0) solution [18].
57
DLC films have similar chemical characteristics to that of diamond, such as excellent
chemical inertness, high fouling resistance and large potential window. DLC can also be
doped into electrically conductive films and be deposited at room temperature. DLC’s
amorphous structure makes the DLC film surface much smoother than that of polycrystalline
diamond film, resulting in considerable decrease in residual current due to the lower
capacitance. The N-DLC film electrodes compare very favorably with conventional carbon
based electrodes such as glassy carbon, graphite and highly oriented pyrolytic graphite
(HOPG). The N-DLC film electrode exhibits a low double-layer capacitance, a wide
electrochemical potential window, and a relatively high electrochemical activity toward
ferricyanide reduction. In addition, the electrode exhibits a catalytic activity for Cl2/Cl-,
durability to high anodic potential and a high signal for the trace analysis of Pb2+. These
characteristics demonstrate great promise of the N-DLC film as a novel electrode material for
electrochemical analysis. Some researchers have started stripping analysis of heavy metals
using N-DLC electrodes as shown in Fig. 2.25 [18].
Conductive N-DLC film electrodes have been used to investigate the possibility of
detecting heavy metals such as lead, copper and cadmium by differential pulse anodic
stripping voltammetry (DPASV) in the absence of mercury film [18, 19]. Liu et al. [18] found
that N-DLC film could trace single metals such as Hg2+, Pb2+ and Cu2+ and multi-metals (Pb2+
+ Cu2+). Zeng et al. [19] also reported that N-DLC film deposited with DC magnetron
sputtering could detect multiple heavy-metals (Cd2+, Pb2+, Cu2+) at the same time. Recently,
Khun et al. [23, 129] proved that N-DLC films fabricated by FCVA had high potential to
detect single metals such as Zn2+, Pb2+, Cu2+ and Hg2+ and multi-metals such as Pb2+ + Cu2+,
Cu2+ + Hg2+ and Pb2+ + Cu2+ + Hg2+.
58
Rehacek et al. [130, 131] developed N-DLC electrodes as supports for bismuth
electroplating on highly conductive Si substrates. It was found that the electrodes had a
potential to replace toxic mercury used most frequently for determination of heavy metals
(Zn2+, Cd2+, Pb2+) by anodic stripping voltammetry.
2.8.4. Potential applications of DLC films as electrodes for electroanalysis
Microelectrodes for electrochemical analysis have been evaluated in recent years in
attempts to increase their detection sensitivity by minimizing the effect of analytes depletion
and double-layer charging current during the measurement of reversible redox species [132].
Microelectrodes are electrodes which have at least one dimension in the micrometer
range and thus have particular characteristics which are a direct function of their small size.
Microelectrodes can have many different forms. For example, the hemispherical diffusion
profile induces an improved mass transfer compared to that of macro-electrodes. This brings
much higher current density which translates into a high signal-to-noise ratio. Reduction of
size leads to lower capacitative contribution to the total current and the possibility of
attainment of steady-state currents within a short time. The very small surface areas of the
microelectrodes also mean that the double-layer capacitance is extremely small. This leads to
allowing a dramatic improvement to the ratio of faradic to capacitive current in analytic
measurement (high sensitivity and low responding time). The small size enables them to be
inserted in places where other electrodes are too large. Microelectrodes allow working in very
high resistive media without loss of sensitivity. This is due to the small currents and to the
facts that the ohmic drop is limited to a small area close to the electrode. The high current
density leads to good signal resolution and low detection limits. The sensing is done within
the diffusion layer and this induces a very low dependence on hydrodynamic conditions. With
such microelectrodes, the smaller the microelectrode diameters are, the lower the flow
59
dependence is. It is reported that the double layer capacitance of the DLC films is much
smaller than that of platinum electrode, which is an advantage of the DLC film electrodes
over the Pt electrodes [133].
Fouling can arise from interfering species or from the analyte itself or its electrode
reaction products. Surface modification can be done to prevent fouling. There is no doubt that
flow system to the electrode is an extremely important area of electroanalysis. Particular
advantages of continuous flow (stirring) to the electrode arise from the fact that fresh solution
is constantly being transported to the electrode surface so that there is no reagent depletion
and no build-up of products since they are continuously washed away. However, in flow
streams for on-line monitoring the fouling of the electrode can be a big problem since there is
continuous contact between the electrode and fresh contaminant.
At present, microelectrodes with submicro-meter dimensions are starting to be
fabricated. If they are to be used singly, the instrumentation becomes primarily important
since the currents registered are at the pico-ampere level. The use of arrays of microelectrodes
all with the same function can increase the total measured current whilst retaining the
particular properties of microelectrodes with respect to high concentration gradient. DLC
films would be a possible and novel material for fabrication of microelectrodes. In order to
study the electrochemical characteristics of the DLC films in detail and eventually apply them
to single microelectrode and array microelectrode sensors, further work will need to be
conducted.
Fabricating DLC films into micro-patterns is an essential step for producing single
microelectrodes and array microelectrodes in mass for electroanalysis. In theory, it may be
easier to deposit DLC films than diamond films on micro-fibers. However, the technology of
coating the conductive DLC films on metallic or non-metallic fibers is still in its infancy.
60
Although there are some publications about diamond film patterning and some other carbon
(mainly with sp2), there are very few reports on DLC films in this field [130, 131]. It clearly
points out that there is still a necessity to develop DLC films as array micro-film-electrodes
for electrochemical purposes.
2.9. Summary
Diamond-like carbon (DLC) films generally have a mixture of sp3 and sp2 bonding.
DLC films with a high sp3 content can possess properties close to those of diamond, including
high hardness, high wear resistance, low friction coefficient, high electrical resistivity,
excellent chemical inertness, and high optical transparency. As amorphous materials, they can
be deposited with near atomic smoothness, adding the relative ease of synthesis compared to
that of diamond films. DLC films are promising materials to replace diamond films in many
fields. There are many ways to synthesize DLC films with FCVA and magnetron sputtering
being among the most widely used methods.
DLC films are usually dielectric. In order to use them in electronics and
electrochemistry, it is necessary to dope them into a good electrical conductor, for example,
with nitrogen doping. Till now, only a few methods have been successful in doping DLC
films. A wide-gap diamond-like carbon is not in itself electrochemically active; however, it
acquires electrochemical activity upon introducing platinum into the DLC bulk during the
film deposition. The effect of platinum is shown to be of a threshold nature: the
electrochemical current appears at approximately 3 at% and saturates at approximately 10 at%
Pt.
DLC films can have a low double layer capacitance and a large potential window. The
electrochemical behavior of DLC films would be different according to their microstructure
61
and dopants employed. However, there have been rarely systematic studies on the
electrochemistry of DLC materials.
The literature review reveals that conductive DLC films can be a favorable novel
electrode material in the future. Fabrication of DLC films into fine patterns is an essential step
to promote the films for the applications in electronics and electrochemical analysis. In
principle, it is easier to pattern DLC films compared to diamond films. A detailed
examination of preparative parameters in relation to a better understanding of pure and doped
DLC film electrode behavior in electrochemistry is needed before the DLC film electrodes are
used for practical applications. Micro-fabrication of DLC films should be taken into
consideration in order to use DLC films as microelectrodes for future electrochemical
applications.
62
Chapter 3 Experimental details
3.1. Sample preparation
Doped diamond-like carbon (DLC) thin films were fabricated using two different
methods: filtered cathodic vacuum arc (FCVA) and DC magnetron sputtering deposition
methods.
Vacuum ar c supply
O ptical viewpor tS tr iker
F i lter field
Scanningmag netic field
Substratebias
G as
Vacuum chamber
C ar bon plasma
IonizedN itr og endopant
Substrate
discharg echamber
PBN
R F ion beam
Fig. 3.1: Schematic configuration of a FCVA deposition system.
The nitrogen doped DLC (N-DLC) thin films (used in Chapter 4) were deposited on
highly conductive p-Si (111) (1-6 10-3 cm) (Ra ~ 0.12 nm) substrates using a FCVA
system (nano films) shown in Fig. 3.1. A pure graphite (99.995%C) target was used as the
carbon source and nitrogen gas was introduced into the deposition chamber for doping. The
thickness of the films was about 100 nm. All the film depositions were carried out at room
temperature and the other deposition parameters are listed in Table 3.1.
Table 3.1: Process parameters of N-DLC films.
N2 flow rate (sccm) 0.5 3 5 10 20
Chamber pressure (Torr) 2.8×10-6 3.5×10-5 5.0×10-5 9.7×10-5 1.9×10-4
Bias (V) (pulse) 1500
63
1. Gas feed line 2. Magnetrons 3. Cryogenic pump 4. Rotary pump5. Butterfly valve 6. Substrate holder 7. Load lock 8. Gate valve
Substrate temperature Room temperature
Fig. 3.2: Schematic configuration of a magnetron sputtering system [62].
Table 3.2: Process parameters of PtRuN-DLC films.
DC power on Pt50Ru50 target (W) 15 20 25 30
DC power on C target (W) 650
Substrate bias (V) -30
Substrate rotation (rpm) 33
Substrate temperature Room temperature
Working pressure (mTorr) 3
Ar flow rate (sccm) 50
Deposition time (min) 120
N2 flow rate (sccm) 10
The platinum/ruthenium/nitrogen doped DLC (PtRuN-DLC) thin films (used in
Chapter 5) were deposited on highly conductive p-Si (100) (0.02–0.005 Ωcm) (Ra ~ 0.12 nm)
substrates with a size of 2 cm × 2 cm using a DC magnetron sputtering deposition system
(Penta Vacuum) (Fig. 3.2). Prior to the film depositions, the Si substrates were ultrasonically 64
cleaned with ethanol for 20 min followed by deionized water cleaning and air drying.
Thereafter, the Si substrates were pre-sputtered with Ar+ plasma at a substrate bias of -250 V
for 20 min in order to remove oxide layers and contaminations on the substrate surfaces.
High-purity graphite (99.999 %) and Pt50Ru50 (99.99 %) targets of 4 inch in diameter were co-
sputtered. The thickness of the PtRuN-DLC films increased from 220 to 300 nm with
increased DC power applied to the Pt50Ru50 target. The detailed deposition parameters are
summarized in Table 3.2.
The N-DLC and PtRuN-DLC thin films (used in Chapter 6) were deposited on p-Si
(100) (0.0035-0.001 Ωcm) (Ra ~ 0.12 nm) substrates with a size of 2 cm × 2 cm using a DC
magnetron sputtering deposition system (Penta vacuum) shown in Fig. 3.2. Prior to the film
depositions, the Si substrates were cleaned according to the procedure used in the preparation
of the PtRuN-DLC films as described above. The thicknesses of the N-DLC and PtRuN-DLC
films were 160 and 250 nm, respectively. The detailed deposition parameters are summarized
in Table 3.3.
Table 3.3: Process parameters of N-DLC and PtRuN-DLC films.
Samples N-DLC PtRuN-DLCDC power on Pt50Ru50 target (W) - 40
N2 flow rate (sccm) 15DC power on C target (W) 850
Substrate bias (V) -90Substrate rotation (rpm) 20Substrate temperature Room temperature
Working pressure (mTorr) 3Ar flow rate (sccm) 50
Deposition time (min) 30
3.2. Characterization
3.2.1. Film Structure
65
The microstructure of the films was investigated using transmission electron
microscopy (JEOL-JEM-2010) which was operated at an accelerating voltage of 200 kV. The
samples were prepared by dispersing a small amount of the scratched powder in ethanol with
ultrasonic treatment. A drop of the dispersion was taken by using a pipet and put on a holly-
carbon copper grid, followed by being dried at room temperature.
Since nitrogen is a weak dopant in carbon, techniques like core-level analysis are
required to establish the chemical bond of nitrogen with carbon. The chemical composition
and binding energy of the films were measured with X-ray photoelectron spectroscopy (XPS)
(Kratos Axis Ultra) with a hemispherical analyzer using a pass energy of 40 eV for C 1s, C 1s
+ Ru 3d, N 1s, Pt 4f, Ru 3p and O 1s core level spectra and 160 eV for the survey scans. The
energy resolutions were 1 eV for the survey scans and 0.1 eV for the narrow scans,
respectively. A monochromatic Al Kα X-ray radiation (hυ = 1486.71 eV) was employed for
the measurements. Fitting of the XPS spectra was performed by decomposing the spectra into
different components with a Gaussian line shape and by approximating the contribution of the
background with Shirley method. A calibration was done by C 1s (approximately 285 eV)
spectrum of a single crystal diamond.
The structure of the films was analyzed using micro Raman spectroscopy (Renishaw
S2000) with a He-Ne laser (632 nm) over the range of 800-2000 cm -1. An objective lens (×50)
was used for a better signal-to-noise ratio. Each Raman instrument used had a spectral
resolution of 1 cm-1 and a spatial resolution of 1 µm. The Raman spectrum acquired was
deconvoluted using Gaussian functions into two peaks: graphitic (G) and disordered (D)
peaks. Average values of the Raman parameters such as positions, full-widths-at-half-
maximum (FHWM) and intensity ratios (ID/IG) of D and G peaks were taken from five random
measurements on each sample.
66
3.2.2. Film surface activity, morphology and topography
The surface activity of the films was studied with water contact angle measurement (a
sessile liquid drop method, FTA 200). An average value of water contact angles was taken
from five random measurements per sample.
The surface roughness of the films was measured using atomic force microscopy
(AFM) (Digital Instruments, S-3000) with a tapping mode Si3N4 cantilever in the scan size of
1 µm × 1 µm. An average value of surface roughness (Ra) was determined from five
measurements per sample.
The film surface morphology was observed using scanning electron microscopy
(SEM) (JEOL-JSM-5600LV). Before SEM studies, each sample was coated with a gold layer
to avoid charging effect.
3.2.3. Adhesion strength of the film to the p-Si substrate
A microscratch tester (Shimadzu SST-101) (Fig. 3.3) was used to qualitatively
evaluate the film adhesive strength. During a scratch test, the spherical-shaped diamond
stylus, which had 15 µm in radius, was drawn over the sample surface under a normal force
that was progressively increased until the film was detached. The normal force at the moment
of the film detachment was measured as a critical load that was a comparative value of the
film adhesive strength. The scan amplitude and frequency, scratching rate, and down speed
for all the tests were set as 50 µm, 30 Hz, 10 µm/s, and 2 µm/s, respectively. Five
measurements were conducted on each sample and an average value of critical loads was
taken.
67
Fig. 3.3: Basic feature of a micro-scratch tester [62].
3.3. Electrochemical measurements
3.3.1. Sample preparation
For all the electrochemical tests, the DLC film coated silicon samples were cut into 2
cm × 2 cm square pieces as shown in Fig.3.4 and a gold layer was deposited on the backsides
of the Si substrates to get a good electrical contact during the tests.
(a) (b) Fig. 3.4: (a) Size of a film-coated sample and (b) film-coated samples.
3.3.2. Electrochemical workstation
The potentiodynamic polarization tests were carried out using a
potentiostat/galvanostat (EG&G 263A) having a potential resolution of 250 µV and a current
resolution of less than 2 pA. A bionano electrochemical workstation (LK6200) having a
potential resolution of 0.1 mV and a current resolution of less than 0.1 pA was applied for the
linear sweep cyclic and anodic stripping voltammetric measurements. The electrochemical 68
impedance spectroscopic (EIS) measurements were conducted using an Autolab Type II
potentiostat/galvanostat with GPES 4.9 software (Eco Chemie, Netherlands) having a
potential resolution of 150 µV and a current resolution of less than 30 fA.
3.3.3. Setup of Electrochemical cell
A flat cell kit (K0235, Princeton Applied Research) (Fig. 3.5) was used for all the
electrochemical measurements. The cylindrical shape of the flat cell and the placement of the
counter electrode directly opposite to the working electrode provide an optimum current
distribution over the surface of the working electrode. Offsetting the reference electrode and
using a capillary of 1.5 mm in outside diameter as a Luggin probe result in the minimum
shielding of the working electrode surface. Consequently, this device minimizes any iR drop
in the electrolyte associated with the passage of current in the electrochemical cell.
(a) (b)
Fig. 3.5: (a) Schematic configuration and (b) outlook of a three-electrode electrochemical cell.
First, a small piece of the film coated silicon sample was placed at the hole on the side
of the cell. It should be noted that the surface coated with film must face towards the
compartment of the electrochemical cell which will be filled with electrolytes as it is the area
for electrochemical testing. The film coated silicon sample was then tightened using the
clamp to seal off the hole completely. Secondly, the wires were connected to the three 69
electrodes, namely, the reference, counter and working electrodes of the electrochemical cell.
Thirdly, the cell compartment was filled up with different solutions depending on the
requirements of experiments under going. The electrochemical cell was then ready for all the
electrochemical tests. All the experiments were performed at room temperature.
The tested area on the films was a circle of 1 cm in diameter. The potentials were
measured with respect to a standard saturated calomel reference electrode (SCE) (244 mV
vs.SHE at 25 ˚C), and a platinum mesh counter electrode was used.
3.3.4. Potentiodynamic polarization test
Potentiodynamic polarization tests were conducted using a potentiostat/galvanostat
(EG&G 263A) to evaluate the corrosion performance of the DLC films in NaCl solutions with
different concentrations when the corrosion potential was nearly stable after immersion in the
solution for about 15 to 30 min. In this experiment, the potential scanning range was varied
according to experiment’s requirements and a scan rate of 0.8 mV/s was used. All the
corrosion parameters, such as corrosion potentials (Ecorr) and currents (Icorr), were taken using
Tafel’s technique.
3.3.5. Immersion test
Immersion tests were performed in NaCl solutions with different pH values which
were compensated with HCl (HCl → H+ + Cl-) and NaOH (NaOH → Na+ + OH-) for 366 hr at
room temperature. For the immersion tests, the samples were cut into 1 cm × 1 cm square
pieces and directly immersed and suspended in the corrosive media without any packaging of
the samples
.
70
3.3.6. Electrochemical impedance spectroscopy
Electrochemical impedance spectroscopy (EIS) was used to measure the
electrochemical properties of the films in a HCl solution. The EIS measurements were carried
out using an Autolab Type II potentiostat/galvanostat with GPES 4.9 software (Eco Chemie,
Netherlands). Bode and Nyquist plots of the film coated samples were acquired at the open
circuit potentials in the frequency range of 105 to 10-2 Hz with an AC excitation signal of 10
mV.
3.3.7. Cyclic voltammetry
Linear sweep voltammetry (LSV) was used for all the voltammetric measurements. In
the linear sweep voltammetric measurement, the current response was plotted as a function of
voltage. This method gave an exact form of voltammogram which could be rationalized by
voltage and mass transport effects. In addition, the characteristics of the linear sweep
voltammogram recorded depended on several factors including: (1) the rate of electron
transfer reaction, (2) the chemical and electrochemical reactivity of active species and (3) the
voltage scan rate.
Cyclic voltammetric experiments were carried out using a bionano electrochemical
workstation (LK6200). All the cyclic voltammograms were obtained with three electrodes
immersed in a flat cell kit containing an electrolytic solution at a scan rate of 100 mV/s. The
cyclic voltammetry was performed to measure the potential windows of the films. The
difference between the potentials for hydrogen and oxygen evolution was called a potential
window. The wider the potential window, the more elements in solution could be detected.
After every CV testing, a potential of 0.5 V vs. SCE was applied on the film electrode for 2
min for discharge cleaning, and then the electrode was washed in distilled water to further
clean the electrode surface.
71
3.3.8. Anodic stripping voltammetry
Linear sweep anodic stripping voltammetry (LSASV) was appled for tracing heavy
metals, and all the LSASV measurements were carried out using a bionano electrochemical
workstation (LK6200). Before the first electrochemical testing, the N-DLC film electrodes
were cleaned with acetone and distilled water. After each stripping analysis experiment, the
electrode was held at 0.5 V vs. SCE for 2 min to clean away the metals deposited on the
electrode surface. After that, the electrode was washed with distilled water. The pH value of
the electrolyte was adjusted with HCl and NaOH. With linear sweep voltammetry, the sweep
rate used was about 36.36 mV/s. All the chemicals employed were analytical reagent grade.
All the stripping voltammograms were compensated with the ohmic drop measured using EIS.
3.4. Summary
The N-DLC films were deposited on the p-Si substrates using a filtered cathodic
vacuum arc (FCVA) system by varying nitrogen flow rate from 0.5 to 20 sccm. The PtRuN-
DLC films were fabricated on the p-Si substrates using a DC magnetron sputtering system by
varying the DC power applied to the Pt50Ru50 target from 15 to 30 W. The N-DLC and
PtRuN-DLC films were deposited on the p-Si substrates using the DC magnetron sputtering
system under the same deposition conditions, except co-sputtering the Pt50Ru50 target.
The film chemical composition was measured using X-ray photoelectron spectroscopy
(XPS). The film microstructure was observed using transmission electron microscopy (TEM)
and the film chemical structure was diagnosed using XPS and micro-Raman spectroscopy.
The film surface activity, morphology and topography were studied using contact angle
measurement, scanning electron microscopy (SEM) and atomic force microscopy (AFM),
respectively. The adhesion strength of the films to the substrates was evaluated using micro
scratch test. The corrosion resistance of the films in aqueous solutions, such as NaCl and HCl,
72
was studied using potentiodynamic polarization test, immersion test and electrochemical
impedance spectroscopy (EIS). The cyclic voltammetric behavior of the films in different
aqueous solutions, such as HCl, H2SO4, KCl, NaCl, KOH, and NaOH, was studied using
linear sweep cyclic voltammetry. Anodic stripping voltammetric behavior of the films was
evaluated in terms of tracing heavy metals, such as Zn, Pb, Cu, Hg, etc., by using linear sweep
anodic stripping voltammetry.
73
Chapter 4 Structure and Electrochemical Properties of Nitrogen Doped
Diamond-like Carbon Thin Films
4.1. Introduction
Diamond-like carbon (DLC) is well-known as a suitable candidate material for use in
harsh environments. A unique combination of high hardness, low friction coefficient, high
wear resistance, and excellent chemical inertness of DLC materials makes them ideal
protective coatings in corrosive environments [29, 134].
Human activities have released toxic metals such as mercury (Hg), lead (Pb), copper
(Cu) and zinc (Zn), etc. into the environment. Nowadays, the presence of toxic metals in the
aquatic ecosystem implicates directly to biota and indirectly to human beings [3, 135]. Pb is
of great concern because of the high toxicity of its compounds and accumulation in various
organisms [3]. It was reported that a Pb concentration ≥ 0.4 µM in drinking water is
detrimental to fetuses and children with possible development of neurological problems [3].
Hg and its compounds are also highly toxic, which can accumulate in vital organs and tissues,
such as heart muscle, liver and brain, and cause kidney injury, central nerve system disorder,
intellectual deterioration, and even death [3]. Some other effects caused by Hg are tremors,
impaired cognitive skills, and sleep disturbance in human being with chronic exposure to
mercury even at low concentrations of µM [3, 135]. Although Cu and Zn have been found to
have relatively low toxicity to human being, prolonged consumption of a large dose can result
in some health complications [3, 135]. For Cu, a concentration ≥ 8×10−5M can adverse
chronic effects [3, 135]. Therefore, fast detection and determination of trace heavy metals are
a tough challenge for analysts. Hanging mercury drop electrodes (HMDE), noble metal
electrodes, and glassy carbon and graphite electrodes have been widely used for
electroanalysis, but the poisoning of Hg used in HMDE electrodes to the environment, and 74
surface oxidation and reduction of the metal and carbon electrodes limit their use for
electrochemical applications. Boron doped diamond (BDD) film electrodes have been
successfully introduced for stripping voltammetric analysis of Pb, Mn, Cd, Cu and Ag, but
fabrication of these films demands a high substrate temperature [18-22].
DLC films can be produced by physical vapor deposition (PVD) and chemical vapor
deposition (CVD) methods at room temperature and achieve similar properties to those of
diamond films, so they have been explored as electrode materials for heavy metal tracing [18-
24]. However, high electrical resistivity and residual stress of DLC films have confined their
electrochemical applications. The electrical conductivity of DLC films is very important for
electrochemical analysis because it can abruptly affect the sensitivity of the film electrodes. A
high residual stress in DLC films apparently reduces the adhesion strength of the films. In a
corrosion environment, poor adhesion strength of DLC films allows undermining of the films
by attacking the interfacial bonds between films and substrates with electrochemically active
species permeated through the porosities in the films. The characteristics of DLC films can be
altered by incorporating different elements such as H, N, F, Si, Pt, Au, Ni, Ti, V, etc. in the
films [25-27]. It was reported that the introduction of nitrogen into DLC films could lower the
electrical resistivity because of the increased sp2 sites in the films [18]. However, the
increased sp2 sites in DLC film with nitrogen incorporation lower the corrosion resistance of
the films due to increased prompt dissolution of the films. The poor corrosion resistance of
DLC films can abruptly affect the electrochemical properties of the films such as sensitivity,
repeatability, long-time response stability and durability when the DLC films are used as film
electrodes for electrochemical purposes. Therefore, an improvement of the corrosion
resistance of nitrogen doped diamond-like carbon (N-DLC) films becomes important for
electrochemical applications.
75
Many electroanalytical techniques are simple, quick, cheap, and easy to use for in situ
measurements in rivers or lakes [8]. Cyclic voltammetry (CV) is one of the most frequently
used electrochemical methods because of its relative simplicity and high information content.
CV has a great advantage for elucidating the mechanisms of electrode reactions. Linear sweep
voltammetry is a simple and useful technique in stripping analysis, which can produce well
defined and selective stripping peaks due to a lower interference of species that can be
adsorbed by working electrode, compared to differential pulse and square wave techniques
[136]. A voltammogram measured by this method and rationalized by the potential and
transport of metal ions allows a study on the kinetic behavior of the ions in solutions having a
wide range of metal ion concentrations.
This chapter investigates the influence of nitrogen concentration on the bonding
structure, surface morphology, adhesion strength and electrochemical performance of nitrogen
doped diamond-like carbon (N-DLC) thin films deposited on p-Si substrates using a filtered
cathodic vacuum arc (FCVA) deposition system in terms of nitrogen flow rate.
4.2. Structural properties of N-DLC thin films
4.2.1. Chemical composition of N-DLC thin films
The chemical composition of the N-DLC films deposited with different nitrogen flow
rates was measured using X-ray photoelectron microscopy (XPS). From the XPS results, it is
found that the N content on the N-DLC film surfaces significantly increases from about 0.98
to 6.48 at% when the nitrogen flow rate is increased from 0.5 to 20 sccm.
After plasma cleaning for 300 s at a chamber pressure of about 3.5 × 10 -8 Torr, the N
contents in the bulks of the N-DLC films deposited with the nitrogen flow rates of 0.5 and 20
sccm are about 0.52 and 4.34 at%, respectively.
76
Fig. 4.1 shows that the fitted XPS spectra of the N-DLC films deposited with different
nitrogen flow rates. Fitting of the C 1s spectra is performed by decomposing each of them into
four components with a Gaussian line shape and by approximating the contribution of the
background by Shirley method. The peak found about 284.1 eV is attributed to the C–C sp2
bonding, and the C–C sp3 bonding contributes to the peak observed at about 285 eV [18]. The
peak at about 286.1 eV is attributed to the C-O bonding formed at the surface of the films due
to their exposure to the air and the peak locating at about 288.5 eV comes from the C-N
bonding [18].
Fig. 4.1: Fitted XPS C 1s spectra of N-DLC films deposited with nitrogen flow rates of (a) 0.5 and (b) 20 sccm.
The sp3 carbon content is determined as a ratio of its corresponding peak over the total
C 1s peak area [18]. It is found that the sp3 content of the N-DLC films significantly decreases
from about 54.64 to 46.42 % with increased nitrogen concentration in the films probably due
the increased amount of the sp2 bonds in the films [54].
77
4.2.2. Raman results of N-DLC thin films
The chemical structure of the N-DLC films was characterized using micro-Raman
spectroscopy. Figure 4.2 shows the Raman spectra of the N-DLC films as a function of
nitrogen flow rate. The Raman spectra of the N-DLC films are fitted using a Gaussian
function for the G peaks and a Lorentzian function for the D peaks as shown in Fig. 4.2. The
FCVA process can generate high kinetic energy of impinging carbon species that in turn form
N-DLC films with a high fraction of sp3 bonds.
600 800 1000 1200 1400 1600 1800 2000 2200Raman shift (cm-1)
Inte
nsity
(a.u
. )
20 sccm
10 sccm
5 sccm
3 sccm
0.5 sccmD
G
Fig. 4.2: Raman spectra of N-DLC films deposited with different nitrogen flow rates. The inset shows ID/IG and AD/AG as a function of nitrogen flow rate.
As shown in Fig. 4.2, the Raman spectra of the films correspond to the sp2 bonded
carbon embedded in the sp3 networks. It is well known that a typical Raman spectrum of DLC
is composed of G and D peaks in which the G peak is due to the stretching vibrations of any
pairs of sp2 sites in chains or aromatic rings and the D peak comes from the breathing mode of
those sp2 sites only in aromatic rings [137]. Therefore, all the Raman spectra were fitted using
78
0.80.9
11.11.21.3
0 3 6 9 12151821N2 flow rate (sccm)
I D/I G
1.51.82.12.42.73
AD/A
GID/IG
AD/AG
ID/IG
AD/AG
Gaussian functions for both G and D peaks. Ferrari et al. [137] proposed that sp2 bonds could
exist not only as rings but also as chains in a dense matrix of DLC depending on its sp3
content. Usually, the introduction of nitrogen into a carbon network induces the formation of
new sp2 sites and encourages the sp2 sites to form clusters because of preferential π bonding of
nitrogen [54]. Therefore, the increased nitrogen content in the N-DLC films enhances the
amount and clustering of the sp2 sites. The increased intensity ratios (ID/IG) and integrated area
ratios (AD/AG) between the D and G peaks of the N-DLC films with increased nitrogen flow
rate (inset in Fig. 4.2) reveal that the increased nitrogen content in the N-DLC films promotes
the clustering of the aromatic rings, indicating an increase in graphitic phases [29, 138].
4.2.3. Surface morphology of N-DLC thin films
Figure 4.3a shows the Ra values of the N-DLC films as a function of nitrogen flow
rate. When the nitrogen flow rate changes from 0.5 to 20 sccm, the surface roughness (Ra) of
the N-DLC films slightly increases from about 0.12 to about 0.23 nm (91.7% increment) as
shown in Fig. 4.3a. It indicates that increasing the nitrogen concentration in the N-DLC films
increases the surface roughness of the films due to a reduced density of the films caused by
increased sp2 sites [24]. In addition, the aggregation of the nitrogen inclusions in the N-DLC
films also contributes to the surface roughness of the films due to the difference in
electronegativity values between carbon (~2.55, pauling scale) and nitrogen (~3.04) [24].
The N-DLC film deposited with 0.5 sccm N2 has fine asperities as shown in Fig. 4.3b,
which is due to a higher fraction of sp3 carbon bonding formed in the film. Larger surface
asperities can be seen on the N-DLC film deposited with 20 sccm N2 (Fig. 4.3c), which are
possibly caused by the increased number of sp2 sites and aggregation of nitrogen atoms [67,
68].
79
0.1
0.14
0.18
0.22
0.26
0 3 6 9 12 15 18 21N2 flow rate (sccm)
Ra (
nm)
(a)
(b) (c)
Fig. 4.3: (a) Ra values of N-DLC films versus nitrogen flow rate. AFM images of N-DLC films deposited with nitrogen flow rates of (b) 0.5 and (c) 20 sccm.
4.2.4. Adhesion strength of N-DLC thin films
In a scratch test, a critical load is determined when an abrupt change in tangential
force is observed, which comes from an instant failure between DLC film and Si substrate.
80
The critical loads for the N-DLC film coated samples increase from 445 to 477 mN (7.2%
increment) with increased concentration as shown in Fig. 4.4a.
430
440
450
460
470
480
490
0 3 6 9 12 15 18 21N2 flow rate (sccm)
Crit
ical
load
(mN
)
(a)
(b)
Fig. 4.4: (a) Critical loads of N-DLC films with respect to nitrogen flow rate and (b) SEM micrograph of a N-DLC film (0.5 sccm N2) scratch tested till a critical load of 456 mN. HP and LP indicate high pressure and low pressure areas, respectively. The inset is the progressive loading curve measured, from which the critical load is determined.
81
020406080
100
0 100 200 300 400 500Load (mN)C
artri
dge
outp
ut
sign
al
HPpP LP
LP
It is well known that the adhesive strength of a film is strongly influenced by its
residual stress. Sullivan and co-workers [73] proposed that the reduction in residual stress was
due to the conversion of a small fraction of sp3 sites to sp2 sites during the film deposition.
The sp2 bonds having shorter bond lengths than those of the sp3 bonds would reduce the strain
in the film plane. In addition, since the C=N bonds have shorter bond lengths compared to the
C-C and C=C bonds, the increased C=N bonds with increased nitrogen concentration give rise
to a higher critical load by reducing the strain in the film [24].
The inset in Fig. 4.4b shows the load-cartridge output signal from the scratch test of
the specimen with the N-DLC film deposited with 0.5 sccm N2 from which a critical load of
456 mN is determined. A cartridge output signal is a voltage ratio obtained from the motions
of the stylus and the cartridge that is also oscillating horizontally above the sample during the
scratch testing [139]. The cartridge output signals are an indication of tangential forces
experienced by the stylus during the scratch testing. No fluctuation can be seen on the
cartridge output signals with respect to normal load, except for an abrupt increase in the
critical load.
A stable linear relationship between tangential force and normal load indicates a
uniform film adhesion throughout the film-substrate interface. From the SEM micrograph
(Fig. 4.4b), no cohesive failure of the film can be found within the scratched path of the film,
while only the fractured area with flaking at the critical load can be viewed as a brittle
fracture. It implies that a high fraction of sp3 bonds can enhance the cohesive strength of the
N-DLC films so that the fracture occurs only at the critical load. This is in agreement with the
work by Gupta et al. in which a thinner film exhibited an instant damage when the normal
load exceeded the critical load [80]. When the cartridge vibrates, the stylus is pulled in the
direction of the vibration. The traction induced results in a nonuniform pressure distribution
82
that causes the shearing between the film and substrate, leading to the peeling-off and
breaking of the film.
4.3. Electrochemical performance of N-DLC thin films
4.3.1. Corrosion behavior of N-DLC thin films
4.3.1.1. Potentiodynamic polarization results of N-DLC thin films
Figure 4.5 presents the potentiodynamic polarization curves of the N-DLC films as a
function of nitrogen flow rate. From the curves, it is found that the cathodic branches do not
change with nitrogen concentration except for the sample deposited with 3 sccm N2.
However, the bonding structure of the N-DLC films, which varies with nitrogen
concentration, affects the anodic branches of their polarization curves. It shows that the N-
DLC film deposited with 3 sccm N2 has the highest corrosion resistance.
In the polarization curves of the N-DLC films deposited with 0.5 and 3 sccm N2, a
reduction in corrosion current is found at a potential above 500 mV vs. SCE. However, no
such a reduction in corrosion current is found from the rest curves, which can be explained
based on the observation of the surface morphologies of the corroded films. Figures 4.6a and
b shows the SEM micrographs of the corroded areas of the N-DLC films deposited with 0.5
and 3 sccm N2, respectively, after the polarization testing in the 0.6 M NaCl solution at room
temperature. Some localized solid products can be seen on the film surfaces, which usually
agglomerate around the defects or pores. These solid products would reduce the corrosion
currents by blocking the diffusion paths of the electrochemically active species from the
electrolytic solution to the film surfaces or to the underlying substrates. That the reduction of
the corrosion currents after the formation of these solid products takes place is consistent with
the decreased anodic currents in the anodic branches of the polarization curves of the N-DLC
films deposited with 0.5 and 3 sccm N2 (insets of Fig. 4.6).
83
-600
-400
-200
0
200
400
600
800
1000
1.00E-08 1.00E-07 1.00E-06 1.00E-05 1.00E-04
Current density (A/cm2)
Pol
ariz
atio
n po
tent
ial (
mV
) vs.
SC
E
0.5 sccm3 sccm5 sccm10 sccm20 sccm
10-8 10-7 10-6 10-5 10-4
Fig. 4.5: Potentiodynamic polarization curves of N-DLC films measured in a 0.6 M NaCl solution at room temperature.
In Fig. 4.7, the corrosion potentials (Ecorr) of the N-DLC films are determined by
fitting the potentiodynamic polarization curves measured at the fixed scan rate of 0.8 mV/s
(Fig. 4.5). The corrosion current (Icorr) and anodic (βa) and cathodic (βc) Tafel slopes are
determined at the same time. The polarization resistance (Rp) values in Fig. 4.7 are then
calculated using the following formula [83]:
Rp = βa × βc / 2.3 Icorr (βa + βc) (4.1)
where Rp is in kΩ cm2; βa and βc are in terms of V/I-decade; and Icorr is in µA/cm2.
84
(a)
(b)
Fig. 4.6: SEM micrographs showing surface morphologies of corroded N-DLC films after potentiodynamic polarization tests in 0.6 M NaCl solution: (a) 0.5 sccm, Ecorr = -85.72 mV vs. SCE and (b) 3 sccm, Ecorr = -57.41 mV vs. SCE, where the insets in the bottom right corners show their enlarged views of locations A and B, respectively.
It is found that the trend of Rp is similar to that of Ecorr as shown in Fig. 4.7. When the
nitrogen flow rate is increased from 0.5 to 3 sccm, the Ecorr shifts from -85.7 to -57.4 mV vs.
SCE and the Rp increases from 16.8 to 150.1 kΩ cm2. However, the Ecorr shifts to more
negative values from -57.4 to -97.6 mV vs. SCE and the Rp turns to decrease from 150.1 to
15.8 kΩ cm2 (89.5% decrement) with further increased nitrogen flow rate to 20 sccm. The
85
A
A
B
B
corrosion results point out that the N-DLC film deposited with 3 sccm N2 shows the highest
corrosion resistance among the films used in this study.
-100
-90
-80
-70
-60
-50
0 3 6 9 12 15 18 21N2 flow rate (sccm)
0
30
60
90
120
150
180
Ecorr
Rp
Cor
rosi
on p
oten
tial (
mV
) vs.
SC
E
Pol
ariz
atio
n re
sist
ance
(k Ω
cm
2 )
Fig. 4.7: Corrosion potentials (Ecorr) and polarization resistances (Rp) of N-DLC films as a function of nitrogen flow rate.
The decreased corrosion resistance of the N-DLC films with higher nitrogen flow rates
than 3 sccm is attributed to several factors such as the bonding structure, electrical resistivity,
surface roughness and porosity density of the films. An increase of nitrogen content in the
films promotes the graphitic phases as depicted by the increased ID/IG and AD/AG ratios (inset
of Fig. 4.2). The increased graphitic phases in the N-DLC films lead to an earlier dissolution
of the films because the increased sp2 bonds via the increased graphitization degraded the sp3-
bonded cross-linking structure which was responsible for preventing the anodic dissolution of
the films. Therefore, the decreased corrosion resistance of the N-DLC films with increased
nitrogen concentration is the reason decreasing the corrosion potentials (Ecorr) of the films.
In addition, the corrosion properties of the N-DLC films are related to the kinetics of
electrochemical reactions taking place at the film-solution interface. It is known that the
86
introduction of nitrogen into the films reduces the electrical resistivity of the films [18, 20,
23]. Another parameter which can determine the electrical conductivity of the DLC films is
sp2/sp3 ratio, i.e. the higher the sp2/sp3 ratio, the lower the electrical resistivity of the film is
[140]. The increased electrical conductivity of the N-DLC films with increased nitrogen
concentration promotes the electron transfer to or from the films during the corrosion testing,
thus accelerating the electrochemical reactions in the electrical double layers (EDL).
Moreover, due to the presence of many tiny anodic and cathodic sites on the surface of
the films caused by the aggregated nitrogen or the electrochemical potential difference
between the films and substrates [67, 98, 141], a galvanically-induced corrosion could occur
between them after the electrochemically active species access the film surface and permeate
into the substrate through the pores. Such corrosion becomes pronounced with increased
nitrogen flow rate.
Furthermore, the increased surface roughness of the N-DLC films may also contribute
to the decreased corrosion resistance of the films because of a larger exposed surface area to
the electrolyte during the corrosion testing.
These combined effects have resulted in the corrosion behavior of the N-DLC films as
shown in Fig. 4.7. A lower corrosion resistance of the N-DLC film deposited with 0.5 sccm
N2 than that of the film deposited with 3 sccm N2 may be attributed to more porosities formed
in the film deposited with 0.5 sccm N2, which probably results from an unbalance between the
kinetic energy of the sputtered C species arriving at the growing film surface and the surface
mobility of the adatoms on the growing surface [142, 143]. An increase of nitrogen gas in the
deposition chamber under the fixed argon flow rate reduces the mean free path of the
sputtered species, thus decelerating the kinetics of the sputtered species and lessening the
87
formation of porosities. Above combined factors can explain the higher corrosion resistance
of the N-DLC film deposited with 3 sccm N2 than that of the one deposited with 0.5 sccm N2.
4.3.1.2. Immersion results of N-DLC thin films
Immersion test offers a simple and cheap way of studying the corrosion behavior of
the N-DLC films immersed in the corrosive media for a longer time. In this investigation, the
N-DLC film coated samples (3 sccm N2) were immersed in the 0.6 M NaCl solutions with
different pH values compensated with HCl (HCl → H+ + Cl-) and NaOH (NaOH → Na+ +
OH-) for 336 hr. Fig. 4.8 shows the surface morphologies of the samples after the immersion
tests where no detachment of the films is found. The sample immersed in a solution of pH 2 is
more severely corroded than other two samples tested in the solutions of pH 4.5 and pH 12,
because of unbalanced hydrogen (H+) and hydroxide (OH-) ions in the solutions [141]. The
corrosion on the surface comes from the reaction of the film with the electrochemically active
species in the electrolyte.
When the N-DLC film atoms are dissolved as ions into the aqueous solution, the
electrons released will flow to the electrochemically active species in the solution where the
hydrogen ions resulting from the water dissociation (H2O → OH- + H+) gain these released
electrons in the cathodic reaction (H+ + e- → H) [90]. At pH < 7, the H+ ions mainly influence
the corrosion while at pH ≥ 7, the OH- ions deplete the H+ ions in the solution through
hydrolysis resulting in a reduced corrosion. Therefore, an apparent change in corrosion
characteristics on the surface of the N-DLC film with respect to pH value is observed in the
SEM micrographs (Fig. 4.8a–c).
At the same time, the effect of Cl- ions, which are increased by compensating the
solution with HCl to make the solution more acidic, on the corrosion resistance of the N-DLC
films should also be taken into account besides the effect of H+ ions. Since the catalytic
88
activity of the Cl- ions accelerates the corrosion process of the N-DLC films, it should be
noted that the increase in the concentration of the Cl - ions with the compensation of the
solution can give an additional effect on the decrease in the corrosion resistance of the films
with decreased pH as discussed above.
(a) (b)
(c) (d)
Fig. 4.8: SEM micrographs showing corroded areas of N-DLC film coated samples after immersion tests in 0.6 M NaCl solutions with different pH values: (a) pH 2, (b) pH 4.5 and (c) pH 12 for the films deposited with 3 sccm N2 and (d) pH 4.5 for the film deposited with 20 sccm N2 for comparison. All the tests are conducted for 336 hr at room temperature and ambient atmosphere.
The initiation of a pit on the film surface occurs when a small local site on the film is
exposed to the damaging species such as chloride and hydrogen ions. It is known that the pits
89
initiate at defects, surface compositional heterogeneities and porosities because these
imperfections degrade the cross-linking structure of the film. The DLC films may always
contain certain open pores that allow a direct access of the electrolyte to the underlying
substrate and some closed pores which are not absolutely open or which diameters are only
big enough to allow some specific ions or molecules in the electrolyte to gradually migrate to
the substrate surface [84]. When the electrolyte accesses the film surface, the pores become
the initiation sites for the corrosion. An increase in immersion time causes the closed pores to
become open and the open pores to grow, resulting in an increase of the porosity density in
the film. The pores serving as crevices allow the entrapment of the permeated
electrochemically active species inside them and a buildup of the positive hydrogen ions
through hydrolytic reactions. The buildup of the positive hydrogen ions increases the acidity
of the entrapped electrolyte inside the pores when the openings of the pores are covered with
the corrosion products resulted from the corrosion of the film. In addition, such a buildup of
the positive charges in the pores becomes a strong attractor to negative ions, e.g. chlorine,
which can be corrosive in their own right, resulting in a more severe corrosion inside the
pores than the surrounding. Eventually, the growth of the pores by connecting adjacent pores
forms the pits as shown in Fig. 4.8a. However, a collection of the pits (Fig. 4.8a), which
serves as an active site with respect to the surrounding (a cathodic site) and shows a more
severe corrosion than the surrounding, indicates that the localized occurrence of the pits
probably results from the aggregation of nitrogen in the amorphous carbon structure. The
aggregated nitrogen locally degrades the sp3 bonded cross-link structure through the increased
sp2 sites and the degraded cross-link structure can be easily attacked by the electrochemically
active species. As explained above, the insufficient repassivation of the film also accelerates
the localized dissolution of the film. When the pH value of the electrolyte is decreased, the
90
increased concentration of hydrogen ions increases the acidity of the electrolyte accessible to
the N-DLC film as well as the entrapped electrolyte inside the pores, thus accelerating the
growth of the pits. Therefore, after the immersion tests, a more apparent pitting of the N-DLC
sample (3 sccm N2) is found with the solution of pH 2 (Fig. 4.8a) than those tested with the
solutions of pH 4.5 (Fig. 4.8b) and pH 12 (Fig. 4.8c).
A DLC film is usually electrochemically nobler than a Si substrate, so a galvanically-
induced corrosion can occur between the film and the substrate when the electrolyte has
permeated into the substrate through some pores or defects of the film [84]. The increased
electrical conductivity of the film with nitrogen doping may facilitate the galvanically-
induced corrosion between the film and the substrate. With a prolonged immersion, more
oxygen existing in the electrolyte confined within the pores and undermining areas of the film
is consumed leading to an increase in concentration of H+ ions in the entrapped electrolyte.
The accumulated H+ ions can attract the negatively charged ions like chlorine ions in the
surrounding electrolyte. A highly acidic local environment with increased immersion time can
cause a substantial increase in corrosion rate, resulting in large undermining areas. The
increased electrolyte trapped inside the enlarged undermining areas leads to a higher surface
tension and eventually cracking and damaging happen to the film as revealed in Fig. 4.8b for
the N-DLC film (3 sccm N2). From the comparison between Fig. 4.8b and d, the N-DLC film
deposited with a higher N2 flow rate (20 sccm) has a lower undermining effect due to its
higher adhesion strength and at the same time, however, a higher pitting rate attributed to a
higher nitrogen doping level. Therefore, it can be deduced that the adhesive strength of the
films is an important parameter to lessen the undermining effect of the imperfections in the
films exposed to the corrosive media.
91
4.3.2. Linear sweep cyclic voltammetric behavior of N-DLC thin films
4.3.2.1. Cyclic voltammetry of N-DLC thin films in acidic solution
The cyclic voltammograms of the N-DLC film electrodes measured in a 0.5 M HCl
solution at a scan rate of 100 mV/s are illustrated in Fig. 4.9. A difference between the
potentials for hydrogen and oxygen evolutions in a cyclic voltammogram gives an
electrochemical potential window. The wider the potential window, the more the elements in
the solutions can be detected for metal tracing analysis. The potentials for hydrogen and
oxygen evolutions on the surface of the N-DLC film electrode (3 sccm N2) are about -1.25 V
and +1.15 V, respectively. The N-DLC film electrode (20 sccm N2) has a lower negative
potential value of about -0.8 V for hydrogen evolution. It is clearly seen that the potential
window of the films decreases with increasing nitrogen concentration, which can be explained
in terms of electrical resistivity of the films as nitrogen doping reduces the electrical
resistivity of the N-DLC film electrodes [55, 120, 144, 145]. The promoted electrical
conductivity reduces an electron transfer potential through the N-DLC film electrode and
results in early hydrogen evolution (2H+ + 2e- → H2↑) at a lower negative potential value in
the reduction.
From the cyclic voltammogram shown in Fig.4.9, no change in potential for the
oxygen evolution (4OH- → 2H2O + O2↑ + 4e-) is found in the oxidization half cycle at the
scan rate of 100 mV/s. It is clearly found that the potential window is mainly affected by the
potential for the hydrogen evolution in the acidic aqueous solution where the main influent
ions are H+ ions.
92
Fig. 4.9: Cyclic voltammograms of N-DLC film electrodes with different nitrogen flow rates measured in 0.5 M HCl solution at a scan rate of 100 mV/s.
A contribution of background current to the cyclic voltammogram of the N-DLC film
electrode (3 sccm N2) can be found. The background current at solid electrode results from
extraneous processes primarily associated with the electrode surface. Thus, according to the
past history of the electrode, the surface may have become oxidized. The potential applied to
the electrode may then give rise to a current from the dissolution of the oxides. An XPS
quantitative analysis of the stoichiometry of the N-DLC film electrode surfaces indicates that
the N-DLC film electrodes (3 sccm N2) have a higher surface oxygen fraction than the film
electrodes deposited with 20 sccm N2. This results from a higher affinity of oxygen with
carbon since there is a greater difference in electronegativities between oxygen (~3.44,
Pauling scale) and carbon (~2.55) with respect to nitrogen (~3.04). Thus, it can be deduced
that the background current is mainly attributed to the dissolution of the oxidized layers. The
occurrence of the peak observed in the reduction cycle may be due to the catalytic activity for
Cl2/Cl- [20, 146].
93
4.3.2.2. Cyclic voltammetry of N-DLC thin films in neutral solutions
Figure 4.10 shows the cyclic voltammograms of the N-DLCfilm electrodes tested in a
0.1 M KCl (pH 1) solution at a scan rate of 100 mV/s. The pH value of the solution was
adjusted by HCl (HCl → H+ + Cl-). The effect of nitrogen incorporation on the potential
windows of the N-DLC film electrodes in the solution can also be seen. The film electrodes
deposited with 3 and 20 sccm N2 have potential windows of approximately 2.32 and 1.9 V,
respectively.
Fig. 4.10: Cyclic voltammograms of N-DLC film electrodes with different nitrogen flow rates measured in 0.1 M KCl solution at a scan rate of 100 mV/s.
The potential window decreases with increased level of the incorporated nitrogen,
resulting from the increased electrical conductivity. This is in agreement with the result
mentioned in the preceding section where the contribution of the background current was
attributed to the dissolution of the oxidized layer on the N-DLC film electrode (3 sccm N 2). It
is noticed from this experiment that a higher nitrogen content in the N-DLC film electrodes
produces a lower background current leading to an improved signal-to-background ratio in the
electroanalytical measurement.
94
-4
-3
-2
-1
0
1
2
3
4
-3.5 -2.5 -1.5 -0.5 0.5 1.5 2.5E (V) vs. SCE
I (m
A)
Fig. 4.11: Cyclic voltammogram of N-DLC (20 sccm N2) film electrode measured in 0.1 M NaCl solution at a scan rate of 100 mV/s.
It has been shown that the nitrogen concentration apparently affects the potential
windows of the N-DLC film electrodes in the KCl and HCl solutions. The N-DLC film
electrodes (20 sccm N2) are used for further analysis of the electrochemical potential windows
in different solutions in order to investigate their outstanding electrochemical properties. The
cyclic voltammogram of the N-DLC film electrode (20 sccm N2) recorded in a 0.1 M NaCl
solution at the scan rate of 100 mV/s is shown in Fig. 4.11. The pH value of the NaCl solution
is not compensated by HCl. The potential window of the N-DLC film electrode (20 sccm N2)
in the NaCl solution is approximately 3.2 V, about 1.3 V higher than the potential window of
the same electrode measured in the KCl aqueous solution. The difference in potential window
can be explained by different pH values. The KCl solution compensated by HCl has a higher
concentration of active H+ ions than the NaCl aqueous solution uncompensated by HCl. Both
unbalanced hydrogen and hydroxide ions and different electroactive species existing in the
KCl and NaCl aqueous solutions can abruptly affect the cyclic voltammetric behavior of the
N-DLC films in the solutions. No peak occurring in the reduction half cycle measured in the
95
NaCl solution indicates a lower catalytic activity for Cl2/Cl- in the solution, which may be one
of the reasons why the cyclic voltammogram in the NaCl solution has a wider potential
window.
4.3.2.3. Cyclic voltammetry of N-DLC thin films in hydroxide solutions
Figures 4.12 and 4.13 show the cyclic voltammetric I–E curves of the N-DLC film
electrodes (20 sccm N2) recorded in 0.1 M KOH and 0.1 M NaOH solutions at a scan rate of
100 mV/s. The N-DLC film electrode tested in the NaOH solution has a wider potential
window of approximately 3.25 V compared to approximately 3.1 V measured in the KOH
solution. This may be attributed to different electroactive alkaline species of the solutions.
-4
-3
-2
-1
0
1
2
3
4
-3.5 -2.5 -1.5 -0.5 0.5 1.5 2.5E(V) vs. SCE
I(mA
)
Fig. 4.12: Cyclic voltammogram of N-DLC (20 sccm N2) film electrode measured in 0.1 M KOH solution at a scan rate of 100 mV/s.
The potential windows of the N-DLC film electrodes in the solutions containing a
higher concentration of OH- ions are wider, compared to those of the film electrodes in the
solutions containing a higher concentration of H+ ions. This can be explained by acid-base
reactions in which H+ ions and OH- ions are mainly involved. An alkaline hydroxide (KOH,
96
NaOH) dissociates into a cation and one or more hydroxide ions in water, making the solution
basic. These hydroxide ions react with hydrogen ions to form water (OH- + H+ → H2O),
resulting in a decrease in acidity of the solution.
-4.5-3.5-2.5-1.5-0.50.51.52.53.54.5
-3.5 -2.5 -1.5 -0.5 0.5 1.5 2.5E (V) vs. SCE
I (m
A)
Fig. 4.13: Cyclic voltammogram of N-DLC (20 sccm N2) film electrode measured in 0.1 M NaOH solution at a scan rate of 100 mV/s.
Hydrogen evolution lately occurs in a hydroxide solution containing a low
concentration of active H+ ions, resulting in a wider potential window. A background current
attributed to the charging effect of electrical double layers can be seen on both the cyclic
voltammograms of the N-DLC films measured in the KOH and NaOH aqueous solutions. The
OH- ions decomposed from KOH or NaOH near the electrode combine with hydronium ions
(H3O+) to form water molecules (OH- + H3O+ → 2H2O) which lead to an adsorbed water layer
on the electrode surface and a water hydration sheath surrounding the ions, resulting in the
separation of the charges. Because the charging effect is becoming greater with increasing the
concentration of the OH- ions, the cyclic voltammograms of the N-DLC film electrodes
measured in the hydroxide solutions have higher backgrounds attributed to the electrical
double layers.
97
4.3.2.4. Cyclic voltammetry of reversible couple (Ferricyanide)
Figure 4.14 shows the scan rate dependence of the cyclic voltammograms measured at
the N-DLC film electrodes (20 sccm N2) using the reversible ferri-ferrocyanide couple as a
redox system [120, 128]:
[Fe(CN)6]3- + e- ↔ [Fe(CN)6]4- (4.2)
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
-0.4 -0.2 0 0.2 0.4 0.6 0.8 1E (V) vs. SCE
I (m
A)
ab
dec
f
Fig. 4.14: Cyclic voltammograms of N-DLC (20 sccm N2) film electrode measured in 5 mM K3Fe(CN)6 /0.1 M NaCl solution at different scan rates: (a) 30, (b) 50, (c) 70, (d) 90, (e) 110, and (f) 130 mV/s.
The electrolyte used was 0.1 M NaCl with pH 1 compensated by H2SO4. It is found
that a peak-potential separation (∆Ep) value increases from 92 to 124 mV (34.8% increment)
when the scan rate is increased from 30 to 130 mV/s. This is due to the effect of the increased
kinetic limitation to shifting an oxidation to more positive potentials and a reduction to more
negative potentials [147]. As the scan rate is increased, the timescale of the experiment
becomes smaller so that, eventually, an equilibrium is not reached at the film electrode
surface and a kinetic effect begins to appear. It is found that the ratio between anodic and
cathodic current peaks (Ip,a/Ip,c) fluctuates between 1.02 and 1.06. An Ip,a/Ip,c ratio slightly
greater than unity indicates that the Fe(CN)64-/Fe(CN)6
3- redox reaction at the N-DLC film 98
electrode exhibits a quasi reversible behavior, which means that though the reverse peak
appears, it is slightly smaller than the forward one [20].
4.3.3. Linear sweep anodic stripping voltammetric behavior of N-DLC thin films
4.3.3.1. Linear sweep anodic stripping voltammograms of Lead
The stripping voltammograms of Pb obtained from the N-DLC film electrodes
deposited with 3 and 20 sccm N2 in a 1 × 10−3 M Pb2+ + 0.1 M KCl solution (pH 1) as a
function of deposition potential are shown in Fig. 4.15a and b. The deposition time and scan
rate used are 120 s and 36.36 mV/s, respectively. It can be clearly seen that the stripping
peaks of Pb increase with more negative deposition potentials for both nitrogen flow rates.
The increased negative deposition potentials cause the mobility of the Pb2+ ions to rise,
leading to an increase in the amount of reduced Pb atoms on the surfaces of the film
electrodes, which is confirmed by the increased stripping peak currents of Pb with the
increase of the negative deposition potentials as shown in Fig. 4.15c. However, the increase of
stripping peak current of Pb slows down beyond the deposition potential of −1.4 V, which
may be partially attributed to hydrogen evolution disrupting the reduction of the Pb2+ ions on
the surfaces of the film electrodes during accumulation [147].
The stripping peaks of Pb obtained from the N-DLC film electrode deposited with 20
sccm N2 are more pronounced than those from the film deposited with 3 sccm N2, which is
due to a decrease in the electrical resistivity of the film electrode with a higher nitrogen
content in it, leading to an increase in kinetics of electron transfer through the film electrode
[18, 52, 147]. The increased kinetics of the electron transfer results in more consumption of
the Pb2+ ions in the interface region between the film electrode and the solution. Therefore, a
greater concentration gradient created by the increased consumption of the ions promotes the
transport of the Pb2+ ions from the solution to the interface and leads to an increase of the
99
reduced Pb atoms on the surface of the film electrode [18]. Consequently, a less restriction in
the transport of the ions with a higher nitrogen content in the film electrode gives higher
stripping peak currents of Pb.
Fig. 4.15: Stripping voltammograms obtained from N-DLC film electrodes deposited with (a) 3 and (b) 20 sccm N2 in a 1 × 10-3 M Pb2+ + 0.1 M KCl solution as a function of deposition potential. Dependence of (c) stripping peak current and (d) stripping potential of Pb on deposition potential. The deposition time and scan rate are 120 s and 36.36 mV/s, respectively.
That the stripping potentials of Pb shifting to more positive values with more negative
deposition potentials for both nitrogen flow rates, as observed in Fig. 4.15d, indicate a quasi-
reversible reaction of Pb on the surface of the film electrode in the solution with 1 × 10 −3 M
100
Pb2+. It is well known that the linear sweep voltammetric technique gives a voltammogram
rationalized by the potential and transport of metal ions. The concentration of redox species at
the interface depends on the transport of these species from the bulk solution. When a flux of
metal ions to the electrode surface is slower than an electrode reaction, an equilibrium
between oxidized and reduced species involved in the electrode reaction is established at the
film electrode surface, which implies a reversible reaction corresponding to a case where the
electrode reaction is much faster than the transport of the metal ions. Therefore, it is expected
that there is a kinetic limit upon the electrode reaction compared to the transport of the Pb 2+
ions because the ion concentration is high enough to cause a high concentration gradient
between the solution and the interface region, so that the transport of the ions becomes faster
than the electrode reaction. Furthermore, the increased kinetic limitation with more negative
deposition potentials shifts the oxidation peak to more positive potentials and the reduction
peak to more negative potentials [147], which explains an opposite shift between the stripping
and deposition potentials of Pb as shown in Fig. 4.15d.
-0.4
0
0.4
0.8
1.2
-1.2 -1 -0.8 -0.6 -0.4 -0.2 0Applied potential (V) vs. SCE
Res
pond
ing
curre
nt (m
A
)
Fig. 4.16: A linear sweep cyclic voltammogram obtained from a N-DLC film electrode deposited with 3 sccm N2 in the same solution as the one used for Fig. 4.14 with a scan rate of 36.36 mV/s.
101
A linear sweep cyclic voltammogram obtained from the N-DLC film electrode (3
sccm N2) in the same solution confirms the quasi-reversible reaction of Pb as shown in Fig.
4.16, which illustrates a weaker reverse peak compared to the forward one and has a peak-
potential separation of 0.112 V and a ratio of 3.28 between anodic and cathodic current peaks.
-60
-40
-20
0
20
40
-0.65 -0.6 -0.55 -0.5 -0.45 -0.4Applied potential (V) vs. SCE
Res
pond
ing
curre
nt (µ
A )
-1.0 V-1.2 V
-1.4 V
-1.6 V
Fig. 4.17: Stripping voltammograms obtained from a N-DLC film electrode (20 sccm N2) in a 1 × 10-6 M Pb2+ + 0.1 M KCl solution as a function of deposition potential. The deposition time and scan rate are 120 s and 36.36 mV/s, respectively.
Figure 4.17 shows the effect of the deposition potential on the stripping response of
the N-DLC film electrode (20 sccm N2) in a 1 × 10−6 M Pb2+ + 0.1 M KCl solution (pH 1),
where the deposition time is 120 s and the scan rate is 36.36 mV/s. The stripping peaks of Pb
become stronger with more negative deposition potentials. It is clearly observed that the
increased negative deposition potentials result in a linear shift of the stripping potentials of Pb
to more negative values in Fig. 4.17, which indicates the reversible reaction of Pb on the
surface of the film electrode because a low concentration gradient between the bulk solution
and the interface region causes a slow transport of the Pb2+ ions so that the kinetics of the
electrode reaction becomes faster than the transport of the Pb2+ ions. It is clearly seen that the
102
metal ion concentrations at mM and µM levels induce different stripping voltammetric
behavior of Pb on the surfaces of the N-DLC film electrodes.
An influence of deposition time on the stripping voltammograms obtained from the N-
DLC film electrodes deposited at different nitrogen flow rates in the same solution as the one
used for Fig. 4.15 is shown in Fig. 4.18a and b. The deposition time is responsible for the
amount of analytes available on the surface of the film electrode at the stripping stage and,
therefore, for the sensitivity. Since an increase in deposition time leads to a proportional
increase in the amount of reduced Pb atoms on the film electrode surface, the stripping peaks
of Pb become stronger with a longer deposition time as revealed in Fig. 4.18a and b. The
replotted Pb anodic stripping peak currents with respect to deposition times (Fig. 4.18c) show
a near linear upward shift for both nitrogen flow rates [128]. It can be seen from Fig. 4.18c
that a higher nitrogen content in the N-DLC film electrode (20 sccm N2) apparently promotes
the sensitivity of the film electrode to the Pb2+ ions, resulting in a higher trend of the stripping
peak current versus deposition time. However, a slowdown in the increase of stripping peak
current of Pb is found beyond the deposition time of 120 s for the film deposited with 20 sccm
N2 (Fig. 4.18c). Since the concentration of the Pb2+ ions is fixed and the quantity of the Pb2+
ions is limited in the solution, the accumulation of the Pb2+ ions on the surface of the film
electrode eventually reaches a saturation when the time is prolonged. A more apparent
slowdown in the response of the N-DLC film electrode deposited with 20 sccm N2 observed
beyond the deposition time of 120 s compared to the one with 3 sccm N 2 is due to an earlier
saturation of the stripping peak current with the increased sensitivity of the film electrode.
103
Fig. 4.18: Stripping voltammograms obtained from N-DLC film electrodes deposited with (a) 3 and (b) 20 sccm N2 in the same solution as the one used for Fig. 4.14 as a function of deposition time, dependence of (c) stripping peak current and (d) stripping potential of Pb2+ on deposition time, and (e) relationship between stripping peak current and potential. The deposition potential and scan rate are -1.2 V and 36.36 mV/s, respectively.
104
A shift of the stripping potentials of Pb to more positive values with the increase of the
amount of reduced Pb atoms on the surface of the film electrode during the accumulation,
leading to a proportional increase in the concentration of the stripped Pb2+ ions during the
stripping, is found in Fig. 4.18d for both nitrogen flow rates when the deposition time is
increased from 20 to 140 s. The increased concentration of the stripped Pb2+ ions found in the
half oxidization reaction of Pb (Pb → Pb2+ + 2e-, E0) can be correlated to the increased
stripping potential according to the Nernst equation [147]:
E = E0 – (RT/nF) ln [Pb2+] (4.3)
where E0, R, T, n, and F are the standard half cell potential of Pb, the universal gas constant,
the absolute temperature, the number of electrons transferred in the half-reaction, and the
Faraday constant, respectively.
It is found that the increased sensitivity of the film electrode also causes a shift of the
stripping potential of Pb to more positive values, resulting from the increased concentration of
the stripped Pb2+ ions. The increased nitrogen content in the N-DLC film electrodes promotes
the sensitivity of the electrodes to the metal ions, which is related to the electrical
conductivity of the electrodes. It is found that the stripping peak currents of Pb measured from
the N-DLC film electrode deposited with 3 sccm N2 are consistently lower than those from
the film electrode deposited with 20 sccm N2 as shown in Fig. 4.18c. An ohmic drop caused
by a lower N doping level can be estimated by a stripping potential shift at a stripping peak
current of Pb according to Fig. 4.18e. It is clear that the N-DLC film electrode deposited with
a lower N2 flow rate (3 sccm) indicates a higher stripping resistance compared to the one
deposited with a higher N2 flow rate (20 sccm).
The measured standard deviation of stripping peak currents for five stripping cycles of
Pb2+ (N2 flow rate = 20 sccm, deposition potential = -1.2 V, deposition time = 120 s, scan rate
105
≈ 36.36 mV/s) is about 1.9 %, demonstrating the good measurement repeatability and stability
of the N-DLC film electrodes.
4.3.3.2. Linear Sweep Anodic Stripping Voltammograms of Zinc and Lead
Figure 4.19 shows the stripping voltammograms of Zn and Pb separately measured by
using a N-DLC film electrode (20 sccm N2) in two different 0.1 M KCl solutions (pH 1)
containing 1 × 10−2 M Zn2+ and 1 × 10−3 M Pb2+, respectively. The deposition potential,
deposition time and scan rate used for both metals are −1.2 V, 120 s and 36.36 mV/s,
respectively.
-0.8
-0.3
0.2
0.7
1.2
1.7
2.2
2.7
-1.3 -0.9 -0.5 -0.1 0.3Applied potential (V) vs. SCE
Res
pond
ing
curre
nt (m
A )
Zn
Pb
Fig. 4.19: Stripping voltammograms obtained from a N-DLC electrode (20 sccm N2) in two different 0.1 M KCl solutions containing 1 × 10-2 M Zn2+ and 1 × 10-3 M Pb2+, respectively. The scan rate, deposition time and deposition potential are 36.36 mV/s, 120 s and -1.2 V, respectively.
The stripping peaks of Zn and Pb are found at −0.861 and −0.354 V vs. SCE,
respectively. Though the concentration of the Zn2+ ions is higher than that of the Pb2+ ions, the
stripping peak current of Zn is much lower than that of Pb because of a smaller deposition
overpotential of Zn compared to that of Pb. It is well known that the larger the negative
106
deposition potential is, the closer the hydrogen evolution is. It is likely that a larger negative
deposition potential of Zn may also contribute to a lower stripping peak current because of the
presence of a closer hydrogen evolution [128]. The difference between the start and end
points of the Zn stripping band is attributed to the edge of the potential window though the N-
DLC film electrode is able to detect the Zn2+ ions due to its larger negative part of the
potential window. However, it is believed that the N-DLC film electrode used in this
experiment can have a higher efficiency for the detection of the Zn2+ ions with low
concentrations by lowering the disturbance of the potential window measured. The results for
Zn prove a low detection limit of the N-DLC film electrodes for tracing heavy metals.
4.3.3.3. Linear Sweep Anodic Stripping Voltammograms of Copper
Figure 4.20a shows that the Cu stripping peaks are significantly enhanced upon
applying more negative deposition potentials in the range of −0.8 to −1.4 V in a 2 × 10 −5 M
Cu2+ + 0.1 M KCl solution (pH 1) at a scan rate of 36.36 mV/s for 120 s. The stripping peaks
of Cu at lower negative deposition potentials may be affected by the Cu+ ions stabilized with
the Cl− ions through the following EC mechanism [148-150]:
Cu0 ↔ Cu+ + e- (4.4)
Cu+ + 2Cl- ↔ CuCl-2 (4.5)
For stabilization of the intermediate Cu(I) species, the deposition potentials required to
obtain a good efficiency during the accumulation are more negative. In addition, the
formation of copper oxides at the film electrode surface may affect the stripping process of Cu
by chemical kinetics [151]. It can be deduced that the effective negative deposition potentials
should be larger than −1.0 V with the concentration of Cu2+ ions used in this study. It is noted
that the stripping peak current of Cu first linearly increases with increased negative deposition
potential up to −1.2 V and then decreases with the further increase of the negative deposition
107
potential as shown in Fig. 4.20b, probably, because of a concomitant hydrogen evolution
during the accumulation [151].
Fig. 4.20: Stripping voltammograms obtained from a N-DLC film electrode (20 sccm N2) in a 2 × 10-5 M Cu2+ + 0.1 M KCl solution as functions of (a) deposition potential and (c) deposition time. Dependence of stripping peak current and stripping potential on (b) deposition potential and (d) deposition time, respectively. The deposition time is 120 s (a and b), the deposition potential is -1.2 V (c and d), and the scan rate used for all the tests is 36.36 mV/s.
As shown in Fig. 4.20b, a reduction of the Cu2+ ions at more negative deposition
potentials seems to have caused the stripping of Cu0 at more negative stripping potentials,
noting that a reversible reaction of Cu at the surface of the film electrode is attributed to a
faster electrode reaction compared to transport of the Cu species due to a low concentration of
108
Cu species. However, a non-linear relationship of the stripping potentials of Cu with the
deposition potentials reveals that the reaction of Cu at the surface of the N-DLC film
electrode is not perfectly reversible [151, 152].
Figure 4.20c shows the effect of the deposition time on the stripping voltammograms
of Cu in the same solution. All the stripping voltammograms of Cu obtained at the deposition
potential of −1.2 V produce sharp and well-defined peaks. As shown in Fig. 4.20d, the
stripping peak current of Cu first linearly increases with the increase of the deposition time up
to 200 s and then the increase of the current slows down that is due to the fixed concentration
of Cu2+ ions in the solution. A shift of the stripping potential of Cu to more positive values
with prolonged deposition time is found in Fig. 4.20d, resulting from an increased amount of
reduced Cu atoms on the surface of the film electrode.
4.3.3.4. Linear Sweep Anodic Stripping Voltammograms of Mercury
The stripping voltammograms recorded in a 0.1 M KCl solution (pH 1) containing
Hg2+ ions of 1.1 × 10−6 M as functions of deposition potential and time are shown in Fig.
4.21a and b. The deposition potential and scan rate used are −1.2 V and 36.36 mV/s,
respectively. The sharp stripping peaks of Hg indicate that there is almost no restriction to the
flow of the Hg2+ ions in the solution, which is supported by a monotonic increase of the
stripping peak current of Hg with not only more negative deposition potentials but also longer
deposition durations as shown in Fig. 4.21c and d. However, a deposition time in the range of
40 – 120 s cannot give an apparent increase of the stripping peak current with increased
deposition time, indicating that the deposition time should be greater than 120 s to effectively
detect the Hg2+ ions at µM levels.
109
Fig. 4.21: Stripping voltammograms obtained from a N-DLC film electrode (20 sccm N2) in a 1.1 × 10-6 M Hg2+ + 0.1 M KCl solution as functions of (a) deposition potential and (b) deposition time. Dependence of stripping peak current on (c) deposition potential and (d) deposition time. The deposition time is 120 s (a and c), the deposition potential is -1.2 V (b and d), and the scan rate for all the tests is 36.36 mV/s.
The stripping potential of Hg observed at around 0.106 V without a correlation with
the deposition potential and time reveals that the scan rate of 36.36 mV/s used in this study is
slow enough to correspond to the reaction kinetics of Hg0 in the solution containing the Cl−
ions during the stripping due to the oxidization of Hg itself, resulting in an immediate current
response without a delay to the applied voltage.
110
4.3.3.5. Linear Sweep Anodic Stripping Voltammograms of Simultaneous Lead, Copper
and Mercury
The linear sweep anodic stripping voltammograms (LSASV) obtained from the N-
DLC film electrode (20 sccm N2) in a solution (pH 1) containing 8.9 × 10−6 M Pb2+ + 2.5 ×
10−5 M Cu2+ + 9.2 × 10−6 M Hg2+ + 0.1 M KCl were measured over the deposition time of 40 –
280 s as shown in Fig. 4.22a. The deposition potential and scan rate used for the simultaneous
detection and determination of Pb2+, Cu2+ and Hg2+ are −1.2 V and 36.36 mV/s, respectively.
Both the unrestricted transport of the Hg2+ ions and the oxidization of the Hg atoms give rise
to more pronounced stripping peaks of Hg while a complicated oxidation process of Cu
during the stripping results in the broadest stripping peaks of Cu out of the three bands [153].
The current response of the N-DLC film electrode is evidently significant to differentiate all
the tested elements, which demonstrates that the three ions can easily be determined
simultaneously with good peak separations.
It is found that the stripping peak currents for Pb, Cu and Hg are linearly related to the
deposition time as shown in Fig. 14.22b. Pb, the last metal deposited, has the lowest stripping
peak currents out of the three because of not only its lowest concentration but also its smallest
deposition overpotential compared to the others. The correlations between the stripping
potentials of multispecies such as Pb2+, Cu2+ and Hg2+ and the deposition time are shown in
Fig. 4.22c. The stripping potential of Pb, the first metal stripped, is observed in the range of
−0.514 to −0.488 V vs. SCE depending on the deposition time. Cu comes next in the range of
−0.121 to −0.077 V vs. SCE, followed by Hg in the range of 0.13 to 0.148 V vs. SCE. The
differences in the stripping potentials of Pb, Cu and Hg are attributed to their different redox
potentials. It should be noted that all the stripping potentials of the three species shift to more
positive values with increased deposition time. Hg has the smallest variation in stripping
111
potential (0.018 V) compared to Cu (0.044 V) and Pb (0.026 V) within the range of deposition
time, owing to its sharpest stripping peaks resulting from the most unrestricted transport of the
Hg2+ ions and ready oxidation of the Hg atoms [147].
Fig. 4.22: (a) Stripping voltammograms obtained from a N-DLC film electrode (20 sccm N2) in a 0.1 M KCl solution containing 8.9 × 10-6 M Pb2+ + 2.5 × 10-5 M Cu2+ + 9.2 × 10-6 M Hg2+
as a function of deposition time. Dependence of (b) stripping peak current and (c) stripping potential on deposition time. The deposition potential and scan rate are -1.2 V and 36.36 mV/s, respectively.
112
4.3.3.6. Concentration effect for the ions traced by LSASV
In the plot of the stripping peak current of Pb against the deposition time, the stripping
peak current of Pb increases linearly up to 120 s and such an increase slows down beyond 120
s for the N-DLC film electrode (20 sccm N2), while such slowdown is not obvious for the N-
DLC film electrode (3 sccm N2) (Fig. 4.18c). The observed slowdown in the increase of the
stripping peak current of Pb at longer deposition time for the N-DLC film electrode (20 sccm
N2) implies the depletion of the Pb2+ ions in the solution, which has caused a small
concentration gradient between the interface region and the bulk solution, leading to a slower
transport of the Pb2+ ions. A similar stripping behavior is observed for Cu as depicted in Fig.
4.20d. The change in the stripping peak current of Hg with deposition time is small in the
range of 40 – 120 s and large from 120 to 280 s (Fig. 4.21d). The detection of small
concentrations of metal ions needs certain time to give rise to a linear relationship between
stripping peak current and deposition time. However, the stripping peak currents of the multi-
elements (Pb2+, Cu2+, and Hg2+) increase almost linearly as shown in Fig. 4.22b, which can be
explained by the effect of metal concentration. The concentrations of the Pb2+ ions used for
Fig. 4.18c and 4.22b are 1 × 10−3 M and 8.9 × 10−6 M, respectively. The mM level of the Pb2+
ions can produce a larger concentration gradient resulting in a faster depletion of the ions in
the solution even at a shorter deposition time compared to the µM level of the Pb2+ ions.
The stripping voltammetric measurements for Cu and Hg were conducted with very
small concentrations of 10−5 and 10−6 M, respectively. In the case of Cu, the stripping peak
current in Fig. 4.22b shows almost a linear relationship with the deposition time because the
concentration (2.5 × 10−5 M) of the Cu2+ ions used for Fig. 4.22b is a little higher than that (2
× 10−5 M) used for Fig. 4.20d. Thus, it may be a great favor for stripping analysis at very
small concentrations of Cu2+ ions. Besides, the presence of Hg during the simultaneous
113
multiple-element tracing can improve the sensitivity of the film electrode to the Cu2+ ions
[15]. It is clear that the stripping peak currents of the Hg2+ ions found in Fig. 4.22b show a
linear relationship with the deposition time due to a larger concentration (9.2 × 10−6 M Hg2+)
compared to that shown in Fig. 4.21d. A larger concentration of Hg2+ not only depresses the
depletion of the Hg2+ ions but also produces a larger concentration gradient which can provide
a faster transport of the Hg2+ ions during stripping analysis. In the plot of the stripping
potential of Pb versus deposition time, the increase of the stripping potential of Pb slows
down at the deposition time beyond 120 s (Fig. 4.18d). A similar phenomenon is also found
for Cu when the deposition time is beyond 200 s (Fig. 4.20d). However, the stripping
potentials obtained from the simultaneous tracing of the multi-elements have maintained a
linear relationship with the deposition time as shown in Fig. 4.22c. As the increased
concentration of the stripped ions can be correlated to the increased stripping potential
according to the Nernst equation, the trends observed for the stripping potentials in Figs.
4.18d, 4.20d and 4.22c are consistent with the trends for the stripping peak currents as shown
in Figs. 4.18c, 4.20d and 4.22b in terms of deposition time.
The detection ability of the N-DLC film electrodes to single elements (Pb2+, Cu2+ and
Hg2+) and three simultaneous elements (Pb2+ + Cu2+ + Hg2+) with respect to deposition
potential and time in the 0.1 M KCl solutions with the metal ion concentrations of around µM
levels indicates much lower detection limits of these elements compared to that reported in
the literature [18]. Such low detection limits are relatively reasonable to detect the levels of
Pb, Cu and Hg in water, which could be directly detrimental to the health of aquatic
ecosystems and indirectly to human beings.
114
4.4. Conclusions
The effect of nitrogen concentration on the structure and electrochemical performance
of the N-DLC films deposited on conductive p-Si substrates using a filtered cathodic vacuum
arc deposition system by varying nitrogen flow rate from 0.5 to 20 sccm was systematically
investigated.
The Raman results showed that the sp2 bonds in the films increased with nitrogen flow
rate. The increased sp2 sites promoted the surface roughness and adhesion strength of the
films. As found from the potentiodynamic polarization experiments, the corrosion resistance
of the N-DLC films decreased with increased nitrogen flow rate because the increased sp2
sites in the N-DLC films with increased nitrogen incorporation degraded the sp3-bonded
cross-linking structure that was main responsible for preventing prompt dissolution of the
films. The increased N content in the N-DLC films increased the electrical conductivity of the
films, which in turn accelerated the electron transfer kinetics affecting the corrosion rate of
the films. The corrosion results indicated that the nitrogen flow rate of 3 sccm (1.67 %N) used
during the film depositions gave rise to the highest corrosion resistance of the N-DLC films.
The immersion tests were employed to study the corrosion behavior of the N-DLC films in
similar media for long runs. It was observed that the pH value of the solutions affected the
anti-corrosion performance of the N-DLC films, i.e., the lower the pH value, the more severe
the corrosion of the film was.
Potential windows of the N-DLC films measured in solutions, such as 0.5 M HCl, 0.1
M KCl, 0.1 M NaCl, 0.1 M KOH, and 0.1 M NaOH, were about 2.4, 2.32, 3.2, 3.1, and 3.25
V, respectively. Although the N-DLC film electrodes offered i) wide potential windows with
different types of solutions, ii) very low and stable background currents to improve the signal-
to-background and signal-to-noise ratios, iii) repeatability of voltammograms, iv) durability of
115
the N-DLC film electrode to high anodic potential and v) long-time response stability, their
voltammograms were apparently affected by their electrical conductivity, type of alkaline
species and unbalanced H+ and OH- ions. Therefore, it was found that the lower nitrogen
content (3 sccm N2, 1.67 %N) in the N-DLC films resulted in the wider potential windows of
the films in the aqueous solutions due to the higher electrical resistivity of the films. In
addition, the background current was attributed to not only the concentration of OH- ions but
also the surface cleanliness (surface oxidized layer). The N-DLC films used in this study had
the desired voltammetric characteristics suitable for electrochemical analysis.
The N-DLC film coated samples were used as working electrodes to identify single
elements (Zn2+, Pb2+, Cu2+ and Hg2+) and simultaneous multi-elements (Pb2+ + Cu2+ + Hg2+) in
deaerated and unstirred 0.1 M KCl solutions (pH 1) using linear sweep anodic stripping
voltammetry (LSASV). The results showed that the current response of the N-DLC film
electrodes was significant to detect all the tested trace metal ions (Zn2+, Pb2+, Cu2+, and Hg2+)
and the three ions (Pb2+ + Cu2+ + Hg2+) could be simultaneously identified with good stripping
peak potential separations. It was found that the increase of nitrogen content in the N-DLC
film electrodes increased the sensitivity of the film electrodes to the trace metal ions so the
higher nitrogen content (20 sccm N2, 6.48 % N) in the N-DLC films gave rise to the higher
sensitivity of the films to the trace metals. It was also found that electrochemical deposition
parameters such as metal ion concentration, deposition potential and deposition time
systematically affected the stripping voltammetric behavior of the N-DLC film electrodes. It
was noted that the increased nitrogen content in the N-DLC film electrodes promoted the
sensitivity of the film electrodes to trace metals but apparently narrowed down the potential
windows of the film electrodes, pointing out that the nitrogen content in the N-DLC film
electrodes needed to be optimized between 1.67 and 6.48 % (between 3 and 20 sccm N2) to
116
get the best balance between the high sensitivity and the wide potential windows of the film
electrodes along with the high corrosion resistance. The degraded corrosion resistance of the
N-DLC film electrodes with N incorporation clearly pointed out that it was still a nessesity to
improve the corrosion resistance of the N-DLC film electrodes for electrochemical purposes.
Chapter 5 Structure and Electrochemical Properties of
Platinum/ruthenium/nitrogen Doped Diamond-like Carbon Thin Films
117
5.1. Introduction
A unique combination of high hardness, high wear resistance, low friction, high
corrosion resistance, excellent chemical inertness and electrical insulation makes diamond-
like carbon (DLC) films a favorable candidate for protective coatings [154, 155]. Nowadays,
DLC films are interested in electrochemical applications because the conductive films can be
successfully made by nitrogen doping [18-20, 23, 24]. In addition, an introduction of nitrogen
into DLC films enhances the adhesion strength of the films via the reduced residual stress in
the films [100]. However, the incorporation of nitrogen in DLC films degrades the corrosion
resistance of the films by increasing sp2 sites in the films [100].
It is well known that Pt and Ru are widely used materials in electrochemical
applications. Pt possesses high resistance to chemical and thermal attacks as well as stable
electrical properties. It is insoluble in hydrochloric and nitric acids, but can be corroded by
cyanides, halogens, sulfur, caustic alkalis and aqua regia [28]. Ruthenium is often used in
platinum alloys to enhance their wear resistance due to its ability to harden these materials
and improve their electrochemical performance because of their outstanding catalytic
properties coming from bimetallic functions. Ru can be attacked by halogens at high
temperatures. Besides, it can be dissolved in alkaline solutions, but it is stable in acidic
solutions [28, 156].
Magnetron sputtering deposition is one of the most interesting techniques employed to
deposit DLC films. This method can produce DLC films with lower residual stresses
compared to those produced by cathodic arc discharge or laser deposition methods. A reduced
residual stress in DLC films allows thicker films to be deposited, which promotes the
corrosion resistance of the films by lessening possible porosities in the films since the
presence of nanopores in the DLC films can rapidly lead to the electrochemical dissolution of
118
their underlying Si substrates due to permeation of water, environmental oxygen, and ions. It
is expected that noble metal and nitrogen doped DLC films fabricated by DC magnetron co-
sputtering can be a new type of film material interesting for electrochemical applications.
In this chapter, the effect of Pt/Ru/N incorporation in DLC films prepared by a DC
magnetron sputtering deposition system on the chemical composition, micro-structure,
bonding configuration, surface activity and morphology, adhesion strength, and corrosion
behavior of the films was systematically investigated.
5.2. Structural properties of PtRuN-DLC thin films
5.2.1. Chemical composition of PtRuN-DLC thin films
Figure 5.1 shows the N/(C+Ru+Pt+N), Pt/(C+Ru+Pt+N), and Ru/(C+Ru+Pt+N)
atomic ratios as a function of DC power applied to the Pt50Ru50 target, which are determined
from the integrated areas of the XPS N 1s, C 1s + Ru 3d, and Pt 4f peaks. It is found that the
DC power increased from 15 to 30 W gives rise to increases in the Pt/(C+Ru+Pt+N) from
0.022 to 0.042 and the Ru/(C+Ru+Pt+N) from 0.024 to 0.051 and a decrease in the
N/(C+Ru+Pt+N) from 0.127 to 0.085.
It is found from Fig. 5.1 that the Ru content in the PtRuN-DLC films is consistently
higher than the Pt one for all the DC powers applied to the Pt50Ru50 target. Higher Ru content
in the PtRuN-
DLC films is probably attributed to more Ru atoms located in the outermost surface
layers of the bimetallic Pt/Ru aggregates that appear on the film surfaces as shown in Fig. 5.2.
This may be due to a core-shell structure of Pt/Ru aggregates, as proposed by Liu et al. [157],
in which one element that is enriched in the core always exhibits a stronger preference for the
selection of homometallic bonding than the other element that existed in the shell. In addition,
Pt atoms can exhibit a preference of forming homometallic bonding, and thus there would be
119
more neighboring atoms around the cores of the Pt aggregate [157]. Therefore, a Pt–Ru core-
shell structure would be thought to possess an inner core enriched with Pt and an outer shell
enriched with Ru [22], which is in agreement with the finding by Babu et al. [158].
0.08
0.09
0.1
0.11
0.12
0.13
10 15 20 25 30 35DC power on Pt50Ru50 target (W)
N/(C
+Ru+
Pt+
N)
0.01
0.02
0.03
0.04
0.05
0.06
N/(C+Ru+Pt+N)Pt/(C+Ru+Pt+N)Ru/(C+Ru+Pt+N)
(Pt o
r Ru)
/(C+R
u+P
t+N
)
Fig. 5.1: N/(C+Ru+Pt+N), Pt/(C+Ru+Pt+N) and Ru/(C+Ru+Pt+N) atomic ratios with respect to DC power applied to Pt50Ru50 target during film depositions.
All the PtRuN-DLC films used in this study were deposited under the same process
parameters except the sputtering power applied to the Pt50Ru50 target. Therefore, it can be
deduced that the increased incorporation of Pt and Ru with higher DC power applied to the
Pt50Ru50 target mainly contributes to the decreased nitrogen content in the films, indicating the
difficult reactive nature of noble metals with nitrogen. Moreover, in the DC magnetron co-
sputtering processes, the increased DC power applied to the target increases the flux of the
sputtered species to the substrate surfaces, and their kinetic energy can be estimated from the
following relationship [159, 160]:
Uk ∞ (Dw × Vs) / Pg0.5 (5.1)
120
where Uk is the kinetic energy, Dw is the target power density, Vs is the substrate bias, and Pg
is the gas pressure. Since the substrate bias voltage and process pressure are fixed in all the
deposition processes, while Ar has a larger ionization rate than that of N2, varying DC power
applied to the Pt50Ru50 target eventually results in the differences in the N contents in the
PtRuN-DLC films [161].
In addition, an increase in surface oxygen percentage adsorbed on the PtRuN-DLC
film surfaces with increased DC power applied to the Pt50Ru50 target was observed during the
XPS compositional analysis. For example, the quantity of the surface oxygen with respect to
15 W is about 1.5 times larger than that corresponding to 30 W. The increase in surface
oxygen percentage results from a greater difference in electronegativities between O (~3.44,
pauling scale) and C (~2.55), Pt (~2.28), or Ru (~2.2) as compared to the one between C and
N (~3.04) [28]. Bewilogua et al. [162] proposed that weakly cross-linked metal doped DLC
films could react with oxygen much more promptly. It may be proposed that the increased
Pt/Ru aggregates on the surfaces of the films with increased DC power on the Pt 50Ru50 target
have resulted in more adsorbed oxygen because a nitrogen doped amorphous carbon network
surrounded by undissolved Pt/Ru aggregates is not strongly cross-linked.
5.2.2. Microstructure of PtRuN-DLC thin films
In Fig. 5.2, the dark regions correspond to the Pt/Ru aggregates and the light areas
correspond to the N-DLC matrices, where the sizes of the aggregates range between 2 and 5
nm. The TEM image in Fig. 5.2 demonstrates that the Pt/Ru aggregates are segregated within
the N-DLC matrices.
121
Fig. 5.2: TEM micrograph of a PtRuN-DLC film deposited with a DC power of 30 W applied to Pt50Ru50 target.
5.2.3. XPS results of PtRuN-DLC thin films
Figure 5.3 shows the XPS spectrum of the PtRuN-DLC film deposited with 30 W
applied to the Pt50Ru50 target, where several peaks such as C 1s, O 1s, N 1s, Pt 4s, Pt 4p, Pt 4d,
Pt 4f, Ru 3s, and Ru 3d can be clearly identified. The O 1s peak probably comes from surface
oxidation during the exposure of the sample to the atmosphere.
0 100 200 300 400 500 600 700Binding energy (eV)
Inte
nsity
(a.u
. ) (b) Pt 4f(c) Si 2p
(d) C 1s(a) Ru 4p
a
b
c
d
(e) Pt 4d
e(f) N 1s
f
(g) Ru 3p
g
(h) Pt 4p
i(i) O 1sh
(j) Ru 3s
j
(k) Pt 4p
k
Fig. 5.3: XPS spectrum of a PtRuN-DLC film deposited with a DC power of 30 W applied to Pt50Ru50 target.
122
The C 1s + Ru 3d spectrum of the PtRuN-DLC film deposited with 15 W applied to
the Pt50Ru50 target is shown in Fig. 5.4a from which it is observed that the C 1s peak entirely
covers Ru 3d3/2 and partially overlaps with Ru 3d5/2. By fitting the spectrum with a Gaussian
line shape and by approximating the contribution of the background with the Shirley method
[18], the peak at about 284.1 eV is attributed to the C–C sp2 bonding as the C–C sp3 bonding
contributes to the peak observed at about 285 eV [18, 163]. C–O bonds resulting from the
oxygen adsorbed after the sample’s exposure to the air contribute to the peak located at
around 286.2 eV [164, 165]. The peak at around 287.5 eV is attributed to C–N bonding
[164,165]. The two peaks at about 280.4 and 281.6 eV in the Ru 3d5/2 region are qualitatively
attributed to Ru0 and Ru–O bonds, respectively, resulting from their spin-orbit coupling effect
[166, 167]. The inset in Fig. 5.4a shows the C 1s + Ru 3d spectra of the PtRuN-DLC films
deposited with different DC powers applied to the Pt50Ru50 target. A slight shift in the C 1s
peaks to lower binding energies (285 to 284.7 eV) with an increase in the DC power indicates
that there are more sp2 bonds in the films, resulting from the increased Pt and Ru contents in
the films [140]. It is consistently found that the sp3 content of the PtRuN-DLC films increases
from about 34.11 to 21.36 % with an increase in the DC power. It is known that a higher
nitrogen incorporation in an amorphous carbon structure gives rise to a higher fraction of sp2
bonds. However, in this study, the increase in the sp2 bonds in the PtRuN-DLC films revealed
by the shift in the C 1s peak positions to lower binding energies cannot be correlated with the
decrease in the nitrogen content in the films with increased DC power applied to the Pt50Ru50
target. Therefore, it is proposed that the introduction of Pt and Ru into the films probably
increases the sp2 bonds in the nitrogen doped amorphous carbon structures, which is in
agreement with the literature [89, 168, 169]. The Ru 3d5/2 peaks also apparently shift to lower
binding energies with the increase in DC power applied to the Pt50Ru50 target, resulting from
123
an increase in neutral Ru as confirmed by the stronger Ru0 peaks than the Ru–O peaks. In
addition, such an increased Ru content in the films results in a decrease in Coulombic
interactions between photon-emitted electrons and ion cores, leading to an apparent shift in
the peaks to lower binding energies. The XPS analyses of Ru 3d, Pt 4f, and Ru 3p spectra
(Figs. 5.4a, c and d) show that their deconvolued components shift to higher binding energies
with their increased oxidization states due to the extra Coulombic interactions between the
photon-emitted electrons and the ion cores.
In Fig. 5.4b, the N 1s spectrum is deconvoluted into three components peaked at
approximately 398.8, 400.5, and 401.8 eV with Gaussian line shapes and a Shirley
background. It is found that the surface oxygen contributes to the broad peak centered at
about 401.8 eV, resulting in N–O bonds. As reported in the literature [170], the peaks at
around 398.8 and 400.5 eV can be attributed to nitrogen atoms each of which is bonded by
either two carbon atoms with one having π bond (N-sp2) or three C atoms all σ bonded (N-
sp3).
From the inset of Fig. 5.4b, a slight shift in the N 1s peaks to lower binding energies
(398.8 to 398.5 eV) with increased DC power applied to the Pt50Ru50 indicates that the
increased Pt and Ru contents with increased DC power have caused the configuration of
nitrogen incorporated amorphous carbon networks to change from N-sp3 to N-sp2, as
confirmed by an apparent increase in N-sp2/N-sp3 area ratio from about 2.39 to 3.04. It can be
further pondered that the occupancy of N in aromatic rings is increased by the increased Pt
and Ru contents in the films.
124
125
Fig. 5.4: Fitted XPS spectra of a PtRuN-DLC film deposited with DC power of 15 W applied to Pt50Ru50 target: (a) C 1s + Ru 3d, (b) N 1s, (c) Pt 4f and (d) Ru 3p. The insets show the relevant XPS spectra of PtRuN-DLC films as a function of DC power applied to Pt 50Ru50
target.
The Pt 4f region composed of three spin-orbit doublets is shown in Fig. 5.4c, where
the Pt 4f components located at about 71.1, 71.9, and 73.2 eV are attributed to Pt0, Pt2+, and
Pt4+, respectively, due to neutral Pt and its oxides [166, 167, 171]. The Pt 4f spectrum
illustrates that the PtRuN-DLC films used in this study consist of a significant amount of
oxidized Pt species. An increase in neutral Pt having lower binding energy and weaker extra
Coulombic interaction with increased DC power applied to the Pt50Ru50 target shifts the spin-
orbit coupling peaks of Pt 4f(7/2,5/2) to lower binding energies, as shown in the inset of Fig. 5.4c.
Two spin-orbit couplings of Ru 3p(3/2,1/2) recorded at about 461.9 and 463.4 eV are
shown in Fig. 5.4d, where the main peak at around 461.9 eV is attributed to Ru0 while 463.4
eV is related to Ru–O [166, 167, 171]. An apparent shift in the Ru 3p(3/2,1/2) peaks to lower
binding energies with increased DC power applied to the Pt50Ru50 target (inset in Fig. 5.4d) is
not found.
5.2.4. Raman results of PtRuN-DLC thin films
Figure 5.5 shows the deconvoluted Raman spectrum of the PtRuN-DLC film
deposited with DC power of 15 W applied to the Pt50Ru50 target. The spectrum is well fitted
with two Gaussian peaks for G (graphite) and D (disordered) bands lying on a linear
background. The inset in Fig. 5.5 shows the Raman spectra of the PtRuN-DLC films
deposited with different DC powers applied to the Pt50Ru50 target. The asymmetrical shapes of
the Raman spectra are changed to a symmetrical shape by developing the D peak (inset in Fig.
5.5) with the increased Pt and Ru contents in the films. Both G and D peaks shift from about
1548 to 1535 cm−1 and from about 1380 to 1374 cm−1, respectively, with increased DC power,
126
as shown in Fig. 5.6a. A redshift of both peak positions is mainly attributed to the introduction
of Pt and Ru into the nitrogen doped amorphous carbon structures, resulting in an increase in
the specific mass of the entire networks [64, 172, 173]. In addition, their inherent longer bond
vibrating at a lower frequency is also a fact of the redshift of both peaks [174, 175].
Fig. 5.5: Raman spectrum together with fitted G and D peaks of a PtRuN-DLC film deposited with DC power of 15 W applied to Pt50Ru50 target. The inset shows the Raman spectra of PtRuN-DLC films deposited with different DC powers applied to Pt50Ru50 target.
The Raman peak intensities of the PtRuN-DLC films decrease with increased Pt and
Ru contents in the films (inset in Fig. 5.5), which are similar to those of the metal doped DLC
films, as reported in the literature [176, 177], due to increased inactive phases. The inactive
phases result from the undissolved PtRu aggregates within the nitrogen doped amorphous
carbon matrices because of their larger atomic radii compared to that of C. In addition, similar
electronegativity values among Pt, Ru, and C are one of the reasons for the undissolved PtRu
aggregates in the carbon matrices [173, 178]. Furthermore, the increased PtRu aggregates
with increased Pt and Ru contents in the films result in a contraction effect on the surrounding
matrices, causing a reduction in the vibrational frequencies of the neighboring bonds around
them by absorbing their vibrational energies. This may also contribute to the downward shifts
of both peak positions.
127
The full-widths-at-half maximum (FWHMs) of both G and D Raman peaks decrease
with increased DC power from 15 to 30 W (Fig. 5.6b). Alternatively, such a decreased
FWHM for the G peak with increased Pt and Ru contents in the films can be correlated with a
decrease in residual stress in the films. Ting and Lee [179] and Morrison et al. [89] reported
that the reduction in G-band width with respect to metal incorporation could be attributed to a
decrease in residual stress within DLC films.
The information about the carbon bonding structure such as graphite clustering and
structure disordering can be obtained from an intensity ratio (ID/IG) between D and G peaks
[137, 93]. Tan and Cheng [180] reported that the tendency of nitrogen was to promote
clustered sp2 bonding, which in turn reduced the sp3 content with increased nitrogen in DLC
films. According to the XPS results, the ID/IG ratio would have downshifted with decreasing
nitrogen content in the films. However, such a correlation between the ID/ IG and nitrogen
content is not found in this study. Therefore, the observed upshift of the ID/IG ratio from about
1.81 to 2.43 with increased DC power (Fig. 5.6c) can be related to an increase in graphite-like
phases induced by the increased Pt and Ru contents in the films [169, 181].
It was reported in the literature [169, 182] that metal phases could enhance
graphitization of amorphous carbon around them because the metal phases could act as
catalysts due to a high sputtered carbon energy. A sputtering process can provide sufficient
energy to locally heat amorphous carbon on a metal surface according to a thermal spike
mode [183]. Moreover, the amorphous carbon contacting with the metal phases transforms
into graphite at a relatively low temperature [182-184]. Therefore, it can be depicted that a
local increase in sp2 bonds in the amorphous carbon matrices is induced by metal-induced
graphitization [182-185], which is also consistent with the work reported in the literature
128
[116, 186, 187], in which metal atoms in the carbon matrices catalyze the formation of sp2
sites.
Fig. 5.6: Results from the fitted Raman spectra of PtRuN-DLC films as shown in the inset of Fig. 5.5: (a) peak positions, (b) FWHMs and (c) ID/IG ratios of D and G peaks.
5.2.5. Surface activity of PtRuN-DLC thin films
Figure 5.7 shows the distilled water contact angles of the PtRuN-DLC films deposited
with different DC powers applied to the Pt50Ru50 target. It is well known that the wettability of
a solid surface is strongly affected by its surface characteristics. The increased water contact
angle from about 79° to 83° (5.1% increment) for the PtRuN-DLC films with increased DC
129
power applied to the Pt50Ru50 target indicates the decreased surface energy or wettability of
the films. A similar result was reported by Choi et al. [188] for their study on the effect of Ag
concentration on the surface energy of Ag doped DLC films. The formation of a surface oxide
on a DLC film surface due to the adsorption of atmospheric oxygen could reduce the water
contact angle by increasing the film surface energy via strong polarities induced by C–O, Pt–
O, and Ru–O bonds [189, 190]. However, the increased water contact angles of the PtRuN-
DLC films with increased Pt and Ru contents in the films (Fig. 5.7) illustrate that the surface
energies of these films are not apparently influenced by the adsorbed surface oxygen.
78
80
82
84
86
10 15 20 25 30 35DC power on Pt50Ru50 (W)
Wat
er c
onta
ct a
ngle
(°)
Fig. 5.7: Water contact angles of PtRuN-DLC films as a function of DC power applied to Pt50Ru50 target. The insets show the water droplets on the surfaces of the films deposited with DC powers of (a) 15 and (b) 30 W applied to Pt50Ru50 target.
Chen and Hong [190] and Grischke et al. [191] reported that the surface energy of
DLC films could be affected by the incorporated nitrogen on the films due to carbon-nitrogen
bonds that greatly enhanced the surface polarity and imposed attractive forces to react with
polar H2O molecules. The decreased nitrogen content in the PtRuNDLC films with increased
DC power applied to the Pt50Ru50 target (Fig. 5.1) probably increases the water contact angle
as the polarity of the films is weakened [190]. Besides, the PtRu aggregates on the PtRuN-
DLC film surfaces would also cause a decrease in surface polarity by decreasing the fraction 130
(a)
(b)
of carbon-nitrogen bonds. Therefore, the increased Pt and Ru contents in the PtRuN-DLC
films with increased DC power are one of the reasons for the increased water contact angle.
5.2.6. Surface morphology of PtRuN-DLC thin films
From the AFM images shown in Fig. 5.8, the film with higher Pt and Ru contents (Fig.
5.8b) has larger asperities than the film with lower Pt and Ru contents (Fig. 5.8a). The surface
roughness (Ra) of the PtRuN-DLC films increases from about 1.3 to 1.9 nm (46.2%
increment) with increased DC power applied to the Pt50Ru50 target from 15 to 30 W due to
increased sp2 sites and PtRu aggregates in the films [169, 192].
(a) (b)
Fig. 5.8: AFM images showing surface topographies of PtRuN-DLC films deposited with DC powers of (a) 15 and (b) 30 W applied to Pt50Ru50 target.
5.2.7. Adhesion strength of PtRuN-DLC thin films
An applied normal load corresponding to an abrupt change in tangential force, which
comes from an instant adhesive failure between the PtRuN-DLC film and the Si substrate, is
taken as a critical load [193]. It is found in Fig. 5.9 that the critical loads of the PtRuN-DLC
films increase from about 338 to 407 mN (20.4% increment), while the DC power applied to
131
the Pt50Ru50 target is increased from 15 to 30 W, indicating that the adhesion strength of the
films to the substrates increases with increased Pt and Ru incorporation [169].
330
350
370
390
410
430
10 15 20 25 30 35DC power on Pt50Ru50 target (W)
Crit
ical
load
(mN
)
Fig. 5.9: Critical loads of PtRuN-DLC films as a function of DC power applied to Pt 50Ru50
target.
It is well known that the adhesion strength of the films is strongly influenced by the
residual stress in the films, which results from the enhanced cross-linkage and bond distortion
in the films caused by the bombardment of the high energetic impinging ions during the film
deposition [194-196]. However, compressive residual stress can be decreased by increased
sp2/sp3 ratio in the films since the sp2 bonds are shorter than the sp3 bonds [73]. Therefore, the
increased sp2 sites in the PtRuN-DLC films with increased Pt and Ru contents have enhanced
the adhesion strength of the films to the Si substrates. In addition, the undissolved PtRu
aggregates could relax the rigidity of the amorphous carbon structures. The increased Pt and
Ru contents in the films with increased DC power lead to a proportional increase in the
critical load. Moreover, a nitrogen doped amorphous carbon structure can have an inherently
lower residual stress than a pure amorphous carbon structure because of shorter C=N bonds
compared to C–C or C=C bonds. Thus, several factors attributed to the noble metal
132
incorporation, such as low solubilities of noble metals in amorphous carbon structures,
formation of noble metal carbon nanocomposites, and graphitization of DLC structures,
apparently influence the adhesion of the DLC films to the substrates by causing the relaxation
of the stress in the films [197, 198].
It is found that the thickness of the PtRuN-DLC films increases from about 220 to 300
nm with increased DC power applied to the Pt50Ru50 target. Since the residual stress in DLC
films came along increased film thickness [29], the critical loads of the PtRuN-DLC films
should decrease with increased film thickness. However, the increased critical loads of the
PtRuN-DLC films with increased film thickness indicate that the influence of the film
thickness on the adhesion strength of the films in terms of residual stress is not significant. On
the other hand, a thicker film may need a higher load for the indenter to break through the
film [78]. Therefore, it may be supposed that the increased thickness of the PtRuN-DLC films
is one of the reasons increasing the critical loads of the films.
5.3. Electrochemical performance of PtRuN-DLC thin films
5.3.1. Corrosion behavior of PtRuN-DLC thin films
Figure 5.10a shows the potentiodynamic polarization curves of the PtRuN-DLC films
measured in 0.1 M NaCl solutions in atmospheric environment as a function of DC power
applied to the Pt50Ru50 target, which are analyzed using a Tafel technique to obtain corrosion
parameters such as corrosion potential (Ecorr) and current (Icorr). Thereafter, polarization
resistance (Rp) is calculated from the anodic (βa) and cathodic (βc) Tafel slopes and corrosion
current (Icorr) according to the equation 4.1.
From Fig. 5.10a, it is found that there are mainly active and transpassive regions in the
anodic branches of the polarization curves. The observed passivated region in the anodic
branch of the polarization curve of the PtRuN-DLC film deposited with 20 W applied to the
133
Pt50Ru50 target may come from the formation of some passivated PtRu aggregates (Pt–O
and/or Ru–O) by reacting with water through the following equation [73, 192, 197, 198]:
Me + H2O → Me(OH)3 → MeO(OH)2 → MeO2 (Me= Pt, Ru) (5.2)
since the carbon matrixes do not repassivate.
A decrease in Icorr from 3.05 to 0.42 µA and an increase in Rp from 44.29 to 162.1 kΩ
(266% increment) (Fig. 5.10b) with increased DC power applied to the Pt50Ru50 target indicate
that the introduction of Pt and Ru into the N-DLC films can induce an improved corrosion
resistance of the films. A shift in Ecorr to more positive values (17 – 149.5 mV vs. SCE) (see
Fig. 5.10c) reveals that the PtRuN-DLC films require a higher applied potential for the
dissolution of the films with increased Pt and Ru contents in the films. In addition, Pt and Ru
may be electrochemically more stable than C, which also contributes to the shift in Ecorr to
more positive values by promoting their cathodic protective behavior so that the polarization
occurs at more positive potentials.
The increased water contact angles of the films with increased DC power applied to
the Pt50Ru50 target reveal that a decreased film surface energy improves the corrosion
resistance of the films by lowering the surface activity of the films.
The above mentioned effects on the corrosion resistance of the films are so strong that
the increased sp2 sites in the films with the increase in the DC power applied to the Pt50Ru50
target cannot be simply correlated with the increased corrosion resistance of the films at lower
potentials. However, a significant increase in the anodic current observed beyond the applied
potential of 1 V can be related to the increased sp2 sites in the films because these sites have
reduced the rigidity of the sp3-bonded structures, resulting in an easy, prompt dissolution of
the films with increased sp2-bonded fraction at higher applied potentials [92].
134
Fig. 5.10: (a) Potentiodynamic polarization curves, (b) corrosion current (Icorr) and polarization resistance (Rp), and (c) corrosion potential (Ecorr) of PtRuN-DLC films as a function of DC power applied to Pt50Ru50 target. The insets in (c) show SEM micrographs of corroded areas of PtRuN-DLC films deposited with DC powers of 30 (top left) and 25 W (bottom right) applied to Pt50Ru50 target.
In addition, the increased anodic current of the films is also attributed to the increased
PtRu aggregates in the nitrogen doped amorphous carbon structures, resulting from the weak
135
interatomic interactions between the PtRu aggregates and the matrices because these weak
interfacial bonds can be easily attacked by electrochemically active species. In reality, the
film dissolution comes from the reaction of the film with the electrochemically active species
in the electrolyte. When the atoms of a film are ionized in an electrolytic solution, the
electrons released will flow through the path of the lowest resistance of the film to the places
where cathodic reactions occur. Therefore, the most important reactions come from the
reduction in dissolved oxygen, water molecules, and hydrogen ions in the electrolyte with the
released electrons via the following possible reactions [11, 28, 90, 99, 199]:
2H+ + 2e- → H2 (5.3)
2H2O + 2e- → H2 (g) + 2 OH- (5.4)
2H2O + O2 + 4e- → 4OH- (5.5)
C + 6OH- → CO32- + 3H2O + 4e- (5.6)
C + (O2-) → CO + 2e- (5.7)
CO + (O2-) → CO2 + 2e- (5.8)
where (O2-) represents a reduced state of oxygen in the form of H2O, OH-, and/or Me-O, etc.
The catalytic activity of Cl- ions causing the anodic dissolution of Pt and Ru from the
PtRuN-DLC films in the NaCl solution by formation of soluble chloride complexes
accelerates the corrosion process, probably through the following possible reactions [28, 200,
201]:
4OH- + 4NaCl → 4NaOH + 4Cl- (5.9)
NaCl → Na+ + Cl- (5.10)
Ru2+ + 5Cl- → RuCl52- + 1e- (5.11)
Pt + 4Cl- → PtCl42- + 2e- (5.12)
Pt + 6Cl- → PtCl62- + 4e- (5.13)
136
Furthermore, these electrochemically active species permeated into the interfaces
between the films and the underlying substrates through some pores or defects in the films
will attack the weak interfacial bonds between them. The attack becomes more severe with
higher Pt and Ru contents in the films because these aggregates create easier paths, allowing
the permeation of the electrochemically active species into the interfaces by developing
interfaces between the aggregates and the C matrix. Galvanically-induced corrosion, which
occurs between the films and the substrates due to different electrochemical potentials
between them, accelerates the attack to the interfaces. Consequently, the corrosion products
formed at the film–substrate interfaces may also force the films to delaminate from the
substrates. A higher applied potential during the anodic scanning can also promote the above
phenomena. Therefore, the PtRuN-DLC films having higher Pt and Ru contents (e.g.,
deposited with 25 and 30 W applied to the Pt50Ru50 target) show a more severe delamination
as revealed by the SEM micrographs (insets in Fig. 5.10c). When the delamination of the
films occurs, the corrosion current starts to increase. However, such delamination is not found
in the PtRuN-DLC films deposited with 15 and 20 W applied to the Pt50Ru50 target,
confirming that lower amounts of sp2 sites and PtRu aggregates contained in the films are
responsible for hindering a high anodic current flow from the films at higher applied
potentials.
The increased corrosion resistance of the PtRuN-DLC films may also be attributed to
the decreased surface energy of the films because the decreased attraction between the water
molecules in the aqueous solution and the films reduces the concentration of the water
molecules near the film surfaces and subsequently decelerates the cathodic reactions
(Equations from 5.3 to 5.5) which are responsible for the anodic dissolution of the films.
137
The increased surface roughness of the PtRuN-DLC films with increased DC power
applied to the Pt50Ru50 target may also contribute to the corrosion of the films because of a
larger exposed surface area to the corrosive medium, thus causing a more prompt dissolution.
It can be deduced that the corrosion resistance of the PtRuN-DLC films with increased Pt and
Ru contents in the NaCl solution increases at lower applied potentials but degrades at higher
applied potentials.
5.4. Conclusions
The chemical composition, microstructure, bonding structure, surface activity and
morphology, adhesion strength and corrosion resistance of the Pt/Ru/N incorporated DLC
films prepared with DC magnetron co-sputtering were investigated in terms of the DC
sputtering power applied to the Pt50Ru50 target. The nitrogen content was reduced with
increased Pt and Ru contents in the films. The XPS C 1s peaks of the films entirely
overlapped with Ru 3d3/2 and partially overlapped with Ru 3d5/2. According to the fitted N 1s
peaks, the area ratio between N-sp2 and N-sp3 bands increased from about 2.39 to 3.04 with
higher Pt and Ru contents in the films. The Pt 4f region was composed of three sets of spin-
orbit doublets, namely, Pt0, Pt2+, and Pt4+, peaked at about 71.1, 72.1, and 75.1 eV,
respectively, attributed to the pure Pt and its oxides. Two sets of spin-orbit couplings of Ru
3p(3/1,1/2) were recorded at about 461.9 and 463.4 eV that were attributed to Ru0 and Ru–O,
respectively. Both G and D Raman peak positions decreased while the ID/IG peak intensity
ratio increased with higher Pt and Ru contents in the films, whose trends were in agreement
with the XPS results. The surface roughness and adhesion strength of the PtRuN-DLC films
increased with increased Pt and Ru contents. It was found that the increased Pt and Ru
contents in the films also enhanced the water contact angles on the film surfaces. The
corrosion resistance of the PtRuN-DLC films increased with the increase in the Pt and Ru
138
contents in the films when the applied potential was less than 1 V. However, with higher
polarization potentials beyond 1 V, the films with higher Pt and Ru contents showed higher
anodic currents in the 0.1 M NaCl solution due to the delamination of the films. Therefore, the
Pt/(C+Ru+Pt+N) and Ru/(C+Ru+Pt+N) should be optimized between 0.033 (20 W) and 0.035
(25 W) and between 0.037 (20 W) and 0.041 (25 W), respectively, to lessen the anodic
dissolution of the PtRuN-DLC films.
139
Chapter 6 Comparative Study of Structural and Electrochemical Properties of
Nitrogen-doped and Platinum/ruthenium/nitrogen-doped Diamond-like Carbon Thin
Films
6.1. Introduction
Nitrogen doped diamond-like carbon (N-DLC) thin films have been developed for
electrochemical applications because of their wide electrochemical potential windows and
high signals to heavy metals at µM concentration [23, 24]. In addition, nitrogen doping
significantly improves the electrical conductivity and adhesion strength of DLC films [23, 24,
100]. However, the introduction of nitrogen into DLC films degrades the corrosion resistance
of the films [100].
It is well known that noble metals such as platinum (Pt) and ruthenium (Ru) possess
high corrosion resistance and excellent catalytic activity. It was reported in the previous
chapter (see Chapter 5) that increasing Pt and Ru contents in the N-DLC films improved the
corrosion resistance of the films and at the same time, their adhesion strength. Though the
influence of Pt/Ru concentration on the corrosion performance of the N-DLC films has been
investigated in the previous chapter, the effect of Pt/Ru doping on the structure and
electrochemical performance of the N-DLC films has not be reported yet.
In this chapter, nitrogen doped DLC films without (N-DLC) or with (PtRuN-DLC) Pt
and Ru doping were deposited on highly conductive p-Si substrates using a DC magnetron
sputtering deposition system to investigate the influence of Pt and Ru doping on the chemical
composition, bonding structure, micro-structure, surface activity and morphology, adhesion
strength, corrosion resistance and cyclic voltammetric behavior of these films.
140
6.2. Structural properties of N-DLC and PtRuN-DLC thin films
6.2.1. Chemical composition of N-DLC and PtRuN-DLC thin films
The atomic ratios of the N/(C+N), N/(C+Ru+Pt+N), Pt/(C+Ru+Pt+N) and
Ru/(C+Ru+Pt+N) are determined from the integrated areas of the XPS C 1s, N 1s, Pt 4f and
Ru 3d spectra and presented in Table 6.1. It is found that the N/(C+N) ratio of the N-DLC
film is about 0.207. The N/(C+Ru+Pt+N), Pt/(C+Ru+Pt+N) and Ru/(C+Ru+Pt+N) ratios for
the PtRuN-DLC film are about 0.181, 0.032 and 0.049, respectively. It is found that the Ru
content is higher than the Pt content for the PtRuN-DLC film, which may be attributed to a
core-shell structure of PtRu aggregates [101, 157, 158].
Table 6.1: Chemical compositions and sp2/sp3 ratios of N-DLC and PtRuN-DLC films.
Sample N/(C+N) N/(C+Ru+Pt+N) Pt/(C+Ru+Pt+N) Ru/(C+Ru+Pt+N) Csp2/Csp3 Nsp2/Nsp3
N-DLC 0.207 - - - 0.75 1.3
PtRuN-DLC - 0.181 0.032 0.049 1.42 2.21
Furthermore, the chemical compositions of the N-DLC and PtRuN-DLC films are
analyzed after Ar+ plasma cleaning for 5 min at a chamber pressure of 3.5 × 10 -8 Torr. It is
found that the N/(C+N) ratio in the N-DLC and N/(C+Ru+Pt+N) ratio in the PtRuN-DLC
films decrease to about 10% and 5%, respectively. However, it is found that the
Pt/(C+Ru+Pt+N) and Ru/(C+Ru+Pt+N) ratios in the PtRuN-DLC film increase to about 21%
and 29%, respectively, also implying that the Ru content is still higher than the Pt content in
the bulk of the film. The decreased N contents in the bulks of the films than those on the film
surfaces after 5 min long plasma cleaning are due to the existence of the adsorbed nitrogen
atoms on the film surfaces. It can be seen that the decreased adsorbance of nitrogen probably
increases the Pt and Ru contents in the bulk of the PtRuN-DLC film.
141
6.2.2. Microstructure of PtRuN-DLC thin films
Figure 6.1 shows a TEM micrograph of the PtRuN-DLC film, which clearly shows
that Pt and Ru exist as nano-aggregates embedded in the N-DLC matrix. The sizes of the
nano-aggregates are in a range of 2-5 nm.
Fig. 6.1: TEM image of PtRuN-DLC film.
6.2.3. Chemical bonding structure of N-DLC and PtRuN-DLC thin films measured by
XPS
Each C 1s spectrum of the N-DLC (Fig. 6.2a) and PtRuN-DLC (Fig. 6.2b) films is
mainly composed of four components located at around 285, 284.1, 286, and 287.3 eV
corresponding to C-C sp3, C-C sp2, C-O (adsorbed surface oxygen), and C-N bonds,
respectively. However, it is observed that the C 1s spectrum of the PtRuN-DLC film overlaps
with Ru 3d3/2 and partially overlaps with Ru 3d5/2, so additional spin-orbit doublets of Ru0 at
about 280.5 eV and Ru-O at about 281.7 eV are found in Fig. 6.2b [166, 167]. The existence
of the Ru-O component at the higher binding energy than the Ru0 is due to extra coulombic
interactions between photon-emitted electrons and ion cores.
142
It is found that the C-C sp2/C-C sp3 ratio, which is estimated from the relevant peak
areas, of the PtRuN-DLC film (about 1.42) is higher than that of the N-DLC film (about 0.75)
as shown in Table 6.1. The sp3 content (about 28.19%) of the PtRuN-DLC film is lower than
that that (about 42.5%) of the N-DLC film. All the DLC films used in this study were
deposited under the same process parameters except the sputtering power applied to the
Pt50Ru50 target during the deposition of the PtRuN-DLC film. It can be deduced that the
introduction of Pt and Ru mainly contributes to the increased C-C sp2/C-C sp3 ratio of the
PtRuN-DLC film due to metal-induced graphitization [101, 116, 182-186]. At the same time,
the effect of the DC sputtering power applied to the Pt50Ru50 target used during the deposition
of the PtRuN-DLC film should also be taken into account because it may be one of the
reasons promoting the metal-induced graphitization via producing the energetic metal species.
143
Fig. 6.2: Fitted XPS C 1s and C 1s + Ru 3d spectra of (a) N-DLC and (b) PtRuN-DLC films, respectively, and fitted XPS N 1s spectra of (c) N-DLC and (d) PtRuN-DLC films.
The N 1s spectra of the N-DLC and PtRuN-DLC films (Fig. 6.2c and d) are composed
of three components, i.e. N-sp2 at about 398.5 eV, N-sp3 at about 400.2 eV and broad N-O
component at about 401.4 eV [78]. The N-sp2/N-sp3 ratio of the PtRuN-DLC (about 2.21) is
higher than that (about 1.3) of the N-DLC film, indicating the preferential existence of N in
the sp2-bonded configuration.
6.2.4. Chemical bonding structure of N-DLC and PtRuN-DLC thin films measured by
Raman spectroscopy
Figure 6.3 shows the Raman spectra of the N-DLC and PtRuN-DLC films from which
it is clearly seen that the introduction of Pt and Ru into the N-DLC film apparently depresses
its Raman spectrum due to the inactive phases that result from the undissolved PtRu
aggregates. It is found that the PtRuN-DLC film causes a downshift of the G peak position
from about 1535 to 1525 cm-1 due to an increased specific mass of the entire network
compared to the N-DLC film [64, 172, 173]. However, the change in the D peak position is
not so significant. The introduced Pt and Ru can serve as catalysts for the formation of sp2
sites via graphitization of amorphous carbon around them [29, 116, 182, 184, 186, 187].
Therefore, clustering of the aromatic rings resulted from the metal-induced graphitization
144
promotes the density of the breathing vibration of the rings, which seems to cause an upward
shift of the D peak position from about 1360 to 1363 cm-1 instead of a downshift with
increased specific mass of the network.
The decreased full-widths-at-half-maximum (FWHMs) of both the G peak from about
204 to 184 cm-1 and the D peak from about 384 to 361 cm-1 with the introduction of Pt and Ru
are associated with decreases in bond and ring disorders [29]. An increase of intensity ratio
(ID/IG) between D and G peaks from about 1.3 to 1.59 with Pt and Ru doping indicates an
increase in cluster size [202]. The decreased FWHMs and increased ID/IG ratio clearly point
out that the introduction of Pt and Ru during the film deposition results in metal-induced
graphitization of the amorphous carbon structure.
-1000
30000
800 1100 1400 1700 2000Raman shift (cm-1)
Inte
nsity
(a.u
. )
N-DLC
PtRuN-DLC
G
D
Fig. 6.3: Raman spectra of N-DLC and PtRuN-DLC films, where G and D represent fitted G and D peaks, respectively.
Table 6.2: Results determined from fitted Raman spectra as shown in Fig. 6.3.
Sample G peak (cm-1) D peak (cm-1) FWHMG (cm-1) FWHMD (cm-1) ID/IG
N-DLC 1535 (± 1) 1360 (± 1) 204 (± 1) 384 (± 1) 1.3
PtRuN-DLC 1525 (± 1) 1363 (± 1) 184 (± 1) 361 (± 1) 1.59
145
6.2.5. Surface activity of N-DLC and PtRuN-DLC thin films
The water contact angles of the N-DLC and PtRuN-DLC films were determined with a
sessile water drop method. From Fig. 6.4 and Table 6.3, it is found that the water contact
angles of the N-DLC and PtRuN-DLC films are about 59 and 68.1°, respectively, pointing out
that the PtRuN-DLC film surface is more hydrophobic (15.4% increment) than the N-DLC
film surface. The incorporation of Pt and Ru in the N-DLC film weakens the polarity of the
film because the PtRu aggregates reduce polar carbon-nitrogen bonds that can promote the
surface energy of the film [190].
Fig. 6.4: Water droplets on (a) N-DLC and (b) PtRuN-DLC film surfaces.
Table 6.3: Water contact angles, surface roughnesses and critical loads of N-DLC and PtRuN-DLC films.
Sample Contact angle (º) Ra (nm) Critical load (mN)
N-DLC 59 (± 0.9) 0.81 (± 0.05 ) 359 (± 7)
PtRuN-DLC 68.1 (± 0.9) 1.18 (± 0.1 ) 393 (± 9)
6.2.6. Surface morphology of N-DLC and PtRuN-DLC thin films
146
Figure 6.5 shows the surface morphologies of the N-DLC and PtRuN-DLC films
measured with AFM. The small asperities found on the N-DLC film surface (Ra ≈ 0.8 nm,
Table 6.3) are attributed to the effect of nitrogen inclusions [24].
When Pt and Ru are introduced, together with nitrogen, into the amorphous carbon
structure, it is found that the surface roughness of the PtRuN-DLC film is increased to Ra ≈ 1.2
nm (50% increment) (Table 6.3) and the surface asperities become larger, resulting from the
increased sp2 sites and the undissolved PtRu aggregates in the film [101].
(a) (b)
Fig. 6.5: AFM images showing surface morphologies of (a) N-DLC and (b) PtRuN-DLC films.
6.2.7. Adhesion strength of N-DLC and PtRuN-DLC thin films
The influence of Pt and Ru doping on the adhesion strengths of the N-DLC and
PtRuN-DLC films is investigated by measuring a critical load at which a sudden change in
tangential force is observed [139].
147
(a) (b)
Fig. 6.6: SEM micrographs showing surface morphologies of scratched (a) N-DLC and (b) PtRuN-DLC films. The inset in (a) shows the scratch track of the N-DLC film.
It is observed that the critical load of the N-DLC film is about 359 mN as presented in
Table 6.3. The SEM micrograph in Fig. 6.6a shows the morphology of the scratched N-DLC
film, which is observed at the location where the indenter was stopped at the relevant critical
load. The inset in Fig. 6.6a shows full path of the scratch test conducted on the N-DLC film.
The observed brittle fracture of the N-DLC film deposited on the Si substrate, which occurs
by removing the film material as platelets from the region bounded by free surface and lateral
cracks developed during scratching, only at the critical load reveals a high cohesive strength
of the film and implies that the kinetic energy of the sputtered C species is sufficient to form
the rigid amorphous carbon network with a high enough sp3 fraction even with the
incorporation of N in the film.
It is found that the incorporation of Pt and Ru into the N-DLC film apparently
promotes the adhesion strength of the film so the critical load of the PtRuN-DLC film is even
higher (about 393 mN) (9.5% increment) than that of the N-DLC film as shown in Table 6.3
due to a higher degree of metal-induced-graphitization. The PtRu aggregates existing in the
N-DLC film also degrade the sp3-bonded cross-linking structure, leading to a lower residual
stress in the film. In addition, the PtRu aggregates could lower the crack density by blocking
148
Spallation
the crack propagation inside the film, which allows the removal of the scratched materials as
platelets as shown in Fig. 6.6b. The observed spallations in some scratched areas reveal that
the incorporation of Pt and Ru into the N-DLC film improves the adhesion strength of the
film so that the detachment of the film at the critical load comes from the bulk region of the Si
substrate instead of the interface between the film and the substrate.
It is found that the introduction of Pt and Ru into the N-DLC film increases the
thickness of the film, which is confirmed by the higher thickness (about 250 nm) of the
PtRuN-DLC film than that (about 160 nm) of the N-DLC film. Robertson [29] reported that
the residual stress in DLC films came along increased film thickness. However, the increased
critical load of the PtRuN-DLC film with increased film thickness compared to that of the N-
DLC film indicates that the influence of the film thickness on the adhesion strength of the
films in terms of residual stress is not significant. On the other hand, a thicker film may need
a higher load for the indenter to break through the film [78]. Therefore, it may be supposed
that the higher thickness of the PtRuN-DLC film is one of the reasons giving rise to the higher
critical load than that of the N-DLC film.
6.3. Electrochemical performance of N-DLC and PtRuN-DLC thin films
6.3.1. Corrosion behavior of N-DLC and PtRuN-DLC thin films
Figure 6.7 shows the Nyquist and Bode plots of the N-DLC and PtRuN-DLC films
measured in 0.1 M HCl solution over the frequency range of 105 - 10-2 Hz with an amplitude
of 10 mV. The inset in Fig. 6.7a shows an equivalent circuit to simulate the frequency
responses of the N-DLC and PtRuN-DLC films in the HCl solution.
In the circuit, R1 is the bulk resistance of the solution, R2 represents the charge transfer
resistance, R3 is the bulk resistance of the film, and Q2 and Q3 are the constant-phase-elements
(CPE) to replace the double-layer capacitance at the film/solution interface and the
149
capacitance of the film, respectively. The results deduced from the Nyquist and Bode plots
shown in Fig. 6.7a are summarized in Table 6.4. Q2 and Q3 in the circuit describe the
deviation of the actual electrochemical process from an ideal one with n = 1, where a CPE
resembles as a capacitor. Q2 and Q3 are attributed to several factors such as aggregations of N
and PtRu, non-uniform composition, and resultant non-uniform electron transport and
electrochemical reaction rate.
A higher R2 value of the PtRuN-DLC film than that of the N-DLC film indicates that
the introduction of Pt and Ru increases the charge transfer resistance of the PtRuN-DLC film,
meaning an enhanced anti-corrosion behavior of the film, which is in agreement with Ref.
[101].
150
Fig. 6.7: (a) Nyquist and (b) Bode plots of N-DLC and PtRuN-DLC films measured in 0.1 M HCl solution. The frequency range is 105 – 10-2 Hz and the amplitude is 10 mV. The inset in (a) shows an equivalent circuit for electrochemical reactions on N-DLC and PtRuN-DLC coated samples.
A higher R3 value of the PtRuN-DLC film than that of the N-DLC film reveals an
increase in the bulk resistance of the film with Pt and Ru doping. It was reported that the
electrical conductance of a DLC film electrode is associated with charge carriers hopping
between sp2-hybridized states [116, 203]. The introduction of N into the films promotes the
electrical conductivity of the films [20]. However, in the noble metal doped DLC film, the
isolated metal aggregates are electrically inactive in the film bulk.
An effective current can occur only when the gap distances between the metal
aggregates themselves, between sp2-hybridized states themselves, or between the metal
aggregates and the nearest sp2-hybridized states are reduced to certain critical values by
increasing metal concentration to ensure an effective exchange of the charge carriers [116,
203]. Therefore, it is supposed that the Pt and Ru contents in the PtRuN-DLC film used in this
study would be below critical value to give rise to sufficient conducting paths for the charge
carriers to pass through. Besides, the PtRu aggregates would non-uniformly distribute within
151
the DLC matrix when the film was deposited by co-sputtering the C and Pt50Ru50 targets.
Pleskov et al. [116] found that a noble metal was not an effective dopant in a DLC film as the
incorporated noble metal did not increase the concentration of free charge carriers.
Table 6.4: Results determined from EIS spectra based on the proposed equivalent circuit as shown in the inset of Fig. 6.7a.
Sample R1 (Ω.cm2) R2 (Ω.cm2) Q2
(sn/Ω.cm2) n2 R3 (Ω.cm2) Q3
(sn/Ω.cm2) n3
N-DLC 90 79 × 104 66.3 × 10-7 0.966 49 × 104 12.4 × 10-8 0.834
PtRuN-DLC 84.6 32 × 106 35.4 × 10-7 0.932 11 ×105 26 × 10-7 0.872
6.3.2. Linear sweep cyclic voltammetric behavior of N-DLC and PtRuN-DLC thin films
6.3.2.1. Cyclic voltammetry of N-DLC and PtRuN-DLC thin films
Figure 6.8a shows the cyclic voltammograms obtained from the N-DLC and PtRuN-
DLC film electrodes in a 0.1 M H2SO4 solution at a scan rate of 100 mV/s, where the N-DLC
film shows a broader potential window from approximately -0.65 to +2 V vs. SCE over which
water decomposition occurs. The overpotentials for hydrogen (2H+ + 2e- → H2↑) and oxygen
(4OH- → 2H2O + O2↑ + 4e-) evolutions for the N-DLC film are larger than those from glassy
carbon (-0.3 to +1.8 V) and highly oriented pyrolytic graphite (HOPG) (-0.4 to +1.6 V)
electrodes [204]. It is found that the introduction of Pt and Ru into the N-DLC film electrode
causes the hydrogen evolution to start at a higher potential of about 0.05 V vs. SCE, which is
probably due to the electrocatalytic effect of the incorporated Pt and Ru (M representing Pt
and Ru) that facilitates the hydrogen evolution on the film electrode in the acidic media based
on the following mechanism [205-209]:
Step 1. Primary discharge (Volmer Reaction):
M + H+ + e- ↔ MHad (6.1)
152
Step 2. Combined with either electrochemical desorption (Heyrovsky Reaction):
MHad + H+ + e- ↔ M + H2↑ (6.2)
or recombination desorption (Tafel Reaction):
MHad + MHad → 2M + H2↑ (6.3)
Furthermore, an earlier oxygen evolution is found at the PtRuN-DLC film electrode,
which is also attributed to the catalytic behavior of the Pt and Ru possibly through the
following process [164, 165]:
M + H2O → MOH + H+ + e- (6.4)
MOH + e- → M + OH- → 4OH- → 2H2O + O2↑ + 4e- (6.5)
It is observed that the oxygen evolution from the PtRuN-DLC film electrode begins at about
0.25 V lower than that from the N-DLC film electrode.
A similar catalytic effect of Pt and Ru doping is found in Fig. 6.8b, in which the cyclic
voltammograms obtained from the N-DLC and PtRuN-DLC film electrodes in a 0.1 M HCl
solution at a scan rate of 100 mV/s are presented. The negative part of the potential window
of the N-DLC film electrode in the HCl solution is about 0.25 V larger than that measured in
the H2SO4 solution (Fig. 6.8a), indicating a lower negative potential limit of the N-DLC film
electrode in the HCl solution due to a higher overpotential for the hydrogen evolution. It is
noted that H2SO4 (H2SO4 → 2H+ + SO42-) has two possible H+ ions to donate, making it twice
as acidic as HCl (HCl → H+ + Cl-) having the same concentration of 0.1 M, for an earlier
hydrogen evolution.
153
Fig. 6.8: Cyclic voltammograms measured from N-DLC (solid line) and PtRuN-DLC (dash line) film electrodes in (a) 0.1 M H2SO4 solution, (b) 0.1 M HCl solution and (c) 0.1 M KCl solution, where scan rate is 100 mV/s.
Though the H2SO4 solution can provide two H+ ions than the HCl solution, it does not
significantly influence the overpotential for the hydrogen evolution from the PtRuN-DLC film
electrode, which is confirmed by similar hydrogen evolution potentials in the both H2SO4 and
HCl solutions as shown in Fig. 6.8a and b. This may be due to a limited electrical
154
conductivity of the PtRuN-DLC film electrode which cannot provide sufficient electrons (e-)
to match the number of H+ ions for the hydrogen evolution according to Equation 6.1-6.3.
The experimental results from this study point out that the different acidic electrolytes
influence the negative parts of the cyclic voltammograms measured from the N-DLC film
electrode rather than the PtRuN-DLC film electrode. The earlier hydrogen evolution on the
PtRuN-DLC film electrode in the both H2SO4 and HCl solutions is assumed to be mainly
attributed to the purely catalytic effect of the Pt and Ru aggregates existing on the film surface
because these aggregates facilitate the charge transfer at the film/solution interface, which is
in agreement with the finding by Pleskov and coworkers [116] that the effect of a noble metal
is purely catalytic in acidic media. In Fig. 6.8b, the oxygen evolution on the PtRuN-DLC film
electrode in the HCl solution is found nearly at the same potential as that in the H2SO4
solution (Fig. 6.8a), indicating that there is no obvious influence from acidic solution type on
the positive portions of the cyclic voltammograms. It is clear that the catalytic effect of Pt/Ru
on the oxygen evolution according to Equations 6.4 and 6.5 is not influenced by the
concentration of H+ ions. The catalytic activity for Cl2/Cl- in the HCl solution causes the
oxygen evolution on the N-DLC film electrode starting at about 0.15 V lower than that
observed in the H2SO4 solution, which is in agreement with the report by Zeng et al. [20].
In Fig. 6.8c, the cyclic voltammogram measured from the N-DLC film electrode in a
KCl solution at a scan rate of 100 mV/s shows a rather larger negative part that is about 0.55
V wider than that measured in the H2SO4 solution (Fig. 6.8a) and about 0.3 V wider than that
measured in the HCl solution (Fig. 6.8b). The KCl solution cannot support as many H+ ions as
those from the H2SO4 and HCl acidic solutions for the hydrogen evolution, which explains
why the cyclic voltammogram from the N-DLC film electrode in the KCl solution has a larger
negative portion. The introduction of Pt and Ru into the N-DLC film electrode also results in
155
an earlier hydrogen evolution in the KCl solution possibly according to the following steps
[207]:
Step 1. Primary discharge:
M + H2O + e- ↔ MHad + OH- (6.6)
Step 2. Coupled with either electrochemical desorption:
MHad + H2O + e- ↔ H2↑+ OH- + M (6.7)
or H recombination desorption:
MHad + MHad ↔ H2↑ + 2M (6.8)
A larger negative part (about 0.8 V wider) of the cyclic voltammogram obtained from the
PtRuN-DLC film electrode than those measured in the both H2SO4 and HCl solutions also
reveals that a much lower concentration of H+ ions in the KCl solution than those in the
H2SO4 and HCl solutions is a main reason for the wider negative part of the potential window.
The potential for the oxygen evolution from the PtRuN-DLC film electrode in the KCl
solution appears to be slightly lower than those measured in the H2SO4 and HCl solutions,
which is probably due to a higher concentration of OH- ions in the KCl solution according to
Equations 6.4 and 6.5. The observed potential for the oxygen evolution from the N-DLC film
electrode in the KCl solution is similar to the one found in the HCl solution (Fig. 6.8b) and
slightly lower than that observed in the H2SO4 solution (Fig. 6.8a), indicating the effect of the
catalytic activity of Cl- ions on the oxygen evolution potential in the KCl solution. A
contribution of a background current to the cyclic voltammogram measured from the N-DLC
film electrode in the KCl solution is found. This is due to a charging effect of the double layer
since a high concentration of OH- ions near the electrode surface supports the formation of
water molecules that lead to an adsorbed water layer on the electrode surface, resulting in the
separations of the charges.
156
6.3.2.2. Cyclic voltammetry of reversible couple (Ferricyanide)
Figure 6.9a shows the cyclic voltammograms measured from the N-DLC film
electrode with two different scan rates using a reversible ferri-ferrocyanide couple as a redox
system [24, 120]:
[Fe (CN)6]3- + e- ↔ [Fe (CN)6]4- (6.9)
It is found that peak-potential separation ΔE and ratio of anodic to cathodic peak currents
(Ip,a/Ip,c) obtained from the N-DLC film electrode at the scan rate of 10 mV/s are about 73 mV
and 1.06, respectively. The N-DLC film electrode having a low ΔE indicates that the film
electrode can have a good electrocatalytic activity for the Fe(CN)64-/Fe(CN)6
3- redox reaction.
The Ip,a/Ip,c ratio greater than unity implies that the Fe(CN)64-/Fe(CN)6
3- redox reaction at the
N-DLC film electrode is quasi-reversible, meaning that though a reverse peak current appears,
it is slightly smaller than the forward one [20, 24]. When the scan rate is decreased to 5 mV/s,
the ΔE and Ip,a/Ip,c obtained from the N-DLC film are 65 mV (11% decrement) and 0.99,
respectively. As the scan rate is lowered, the time scale of the experiment becomes larger so
that an equilibrium condition is achieved at the electrode surface and the kinetic effect begins
to diminish. Therefore, the decreased kinetic limitation shifts oxidation to less positive
potential and a reduction to less negative potential [24].
The measurements of the cyclic voltammogram as shown in Fig. 6.9a were repeated
for more than 3 times with essentially identical profiles, revealing that a good repeatability of
the voltammograms.
157
Fig. 6.9: Cyclic voltammograms measured from (a) N-DLC film electrodes with different scan rates and (b) N-DLC (solid line) and PtRuN-DLC (dash line) film electrodes in 1 mM K3Fe(CN)6/0.1 M HCl solution, where scan rate is 10 mV/s.
It is clearly seen in Fig. 6.9b that the incorporation of Pt and Ru into the N-DLC film
electrode increases the ΔE and Ip,a/Ip,c to 200 mV (174% increment) and 1.15, respectively. It
is well known that linear sweep voltammetry measures a voltammogram rationalized by the
potential and transport of species. When a flux of species to the electrode surface is slower
than an electrode reaction, an equilibrium between oxidized and reduced species involved in
the electrode reaction is established at the film electrode surface, which implies a reversible
reaction corresponding to a case where the electrode reaction is much faster than the transport
of the species. When there is a kinetic limit upon the electrode reaction compared to the 158
transport of the species, the kinetic limitation shifts an oxidation peak to a more positive
potential and a reduction peak to a more negative potential. Therefore, more apparent shifts of
the redox peaks at the PtRuN-DLC film electrode to opposite potentials than those at the N-
DLC film electrode indicate an increased kinetic limitation upon the PtRuN-DLC film
electrode, which is attributed to the increased electrical resistivity of the film electrode by
doping Pt and Ru. The increased electrode kinetic limitation associated with the increased
electrical resistivity of the film electrode causes the Fe(CN)64-/Fe(CN)6
3- redox reaction at the
PtRuN-DLC film electrode to be more quasi-reversible, which is confirmed by the increased
Ip,a/Ip,c ratio.
6.4. Conclusions
Nitrogen doped DLC films without (N-DLC) or with (PtRuN-DLC) Pt and Ru doping
were deposited on highly conductive p-Si substrates using a DC magnetron sputtering
deposition system to investigate the influence of Pt and Ru doping on the chemical
composition, bonding structure, micro-structure, surface activity and morphology, adhesion
strength, corrosion resistance and cyclic voltammetric behavior of these films. The
introduction of Pt and Ru into the N-DLC films promoted the sp2 sites via metal-induced
graphitization, increased the surface roughness, and enhanced the adhesion strength and
corrosion resistance of the films. The enhancement of the corrosion resistance was confirmed
by the increased charge transfer resistance of the PtRuN-DLC film. The observed constant
phase element behavior of all the samples indicated that aggregations of N and PtRu in the
DLC films induced non-uniform currents and consequently non-uniform electrochemical
reaction rates. Though the N-DLC film electrodes showed wider potential windows in acidic
solutions such as H2SO4 and HCl and a neutral solution of KCl, the potential for hydrogen
evolution was significantly affected by the concentration of H+ ions in the solutions. It was
159
found that the Pt and Ru doping apparently narrowed down the potential windows of the N-
DLC film electrodes in these solutions due to their catalytic activities. The N-DLC film
electrodes showed a good electrocatalytic activity in Fe(CN)64-/Fe(CN)6
3- redox reactions.
However, an increased kinetic limitation upon the PtRuN-DLC film electrode shifted the
oxidization peak to a more positive value and the reduction peak to a more negative value
compared to those obtained from the N-DLC film electrode.
160
Chapter 7 Conclusions
7.1. Conclusions on N-DLC films prepared by a FCVA technique
In this study, high quality N-DLC thin films were deposited on conductive p-Si
substrates using a filtered cathodic vacuum arc (FCVA) deposition system by varying
nitrogen flow rate from 0.5 to 20 sccm. The chemical composition, bond structure, surface
morphology and adhesion strength of the N-DLC films were characterized using X-ray
photoelectron spectroscopy (XPS), micro-Raman spectroscopy, atomic force microscopy
(AFM) and scanning electron microscopy (SEM), and micro-scratch test, respectively. The
corrosion resistance of the N-DLC films was evaluated using potentiodyanmic polarization
and immersion tests. The cyclic voltammetric behavior and anodic stripping voltammetric
performance of the N-DLC films were investigated with linear sweep voltammetric method.
The increased nitrogen flow rate significantly increased the N content in the N-DLC
films. The Raman results indicated that the sp2 bonds in the N-DLC films increased with
increased nitrogen incorporation because nitrogen preferred π bonding. The increased sp2
bonds and nitrogen aggregation in the N-DLC films increased the surface roughness of the N-
DLC films. The increased critical loads of the N-DLC films with increased nitrogen flow rate
revealed that the increased nitrogen incorporation in the films promoted the adhesion strength
of the films due to the increased sp2 sites in the films.
It was found from the potentiodynamic polarization tests that the corrosion resistance
of the N-DLC films in the 0.6 M NaCl solution decreased with increased nitrogen
incorporation because the increased sp2 sites increased the prompt dissolution of the films by
degrading the sp3-bonded cross-linking structure. The corrosion results indicated that the
nitrogen flow rate of 3 sccm (1.67% N) used during the film depositions resulted in the
highest corrosion resistance of the N-DLC films. The immersion tests pointed out that the pH
161
value of the solution significantly affected the corrosion performance of the N-DLC films, i.e.
the lower the pH value, the more severe the corrosion was.
The N-DLC thin films showed excellent electrochemical behavior in different aqueous
solutions. The electrochemical potential windows of the N-DLC films measured in 0.5 M
HCl, 0.1 M KCl, 0.1 M NaCl, 0.1 M KOH and 0.1 M NaOH were about 2.4, 2.32, 3.2, 3.1
and 3.25 V, respectively. Although the N-DLC film electrodes offered (1) wide potential
windows in different types of solutions, (2) a very low and stable background to improve the
signal-to-background and signal-to-noise ratios, (3) repeatability of voltammograms, (4)
durability of the film electrodes to high anodic potential, and (5) long-time response stability,
their voltammograms were apparently affected by their electrical conductivity, type of
alkaline species and unbalanced H+ and OH- ions. It was found that the lower nitrogen content
(3 sccm N2, 1.67 %N) in the N-DLC films resulted in the wider potential windows of the
films in the aqueous solutions due to the higher electrical resistivity of the films..
The N-DLC films provided a significant stripping response for determination of
single-elements (Zn2+, Pb2+, Cu2+ and Hg2+) and multi-elements (Pb2+ + Cu2+ + Hg2+)
simultaneously in the KCl solution. However, the sensitivity of the N-DLC films was
significantly influenced by the nitrogen content in the films so the higher nitrogen content (20
sccm N2, 6.48%N) in the N-DLC films gave rise to the higher sensitivity of the films to the
trace metals. It was observed that the sensitivity of the film electrodes to the metal elements
was apparently influenced by deposition time and potential, concentration of elements in the
solution, pH value, and scan rate. The simultaneous detection of the heavy metals using linear
sweep anodic stripping voltammetry produced sharp and well-defined peaks with good peak
separations. It was noted that the increased nitrogen content in the N-DLC film electrodes
promoted the sensitivity of the film electrodes to the trace metals but apparently narrowed
162
down the potential windows of the film electrodes, pointing out that the nitrogen content in
the N-DLC film electrodes needed to be optimized between 1.67 and 6.48 % (between 3 and
20 sccm N2) to get the best balance between the high sensitivity and the wide potential
windows of the film electrodes along with the high corrosion resistance.
The novel N-DLC film electrodes under development showed a great promise for the
detection of trace metals at µM concentration. However, the degraded corrosion resistance of
the N-DLC films with nitrogen incorporation pointed out that it was important to improve the
corrosion resistance of the films because the poor corrosion resistance of the films could
affect the electrochemical performance of the films such as sensitivity, repeatability, long-
time response stability, durability, etc. Since noble metals, such as Pt and Ru, had high
corrosion resistance and bi-catalytic activities, it was expected that the incorporation of Pt and
Ru into the N-DLC films would improve the corrosion resistance of the films and at the same
time, their sensitivity.
7.2. Conclusions on PtRuN-DLC films prepared by a DC magnetron sputtering
technique
Platinum/ruthenium/nitrogen-doped DLC (PtRuN-DLC) thin films were deposited on
conductive p-Si substrates using a DC magnetron sputtering deposition system by varying DC
power applied to Pt50Ru50 target from 15 to 30 W. The influence of Pt and Ru on the chemical
composition, micro-structure, bonding structure, surface activity and morphology, adhesion
strength and corrosion resistance of the N-DLC films were investigated using X-ray
photoelectron microscopy (XPS), transmission electron microscopy (TEM), micro-Raman
spectroscopy, water contact angle measurement, atomic force microscopy (AFM), scanning
electron microscopy (SEM), micro-scratch test and potentiodynamic polarization test.
163
The Raman results indicated that the increased Pt and Ru contents in the PtRuN-DLC
films with increased DC power applied to the Pt50Ru50 target significantly increased
graphitization of the films. TEM micrograph showed that the incorporated Pt and Ru existed
as aggregates in the nitrogen doped amorphous carbon matrix. The increased metal-induced
graphitization and PtRu aggregation with increased Pt and Ru contents in the PtRuN-DLC
films increased the surface roughness and adhesion strength of the films. The increased water
contact angles on the PtRuN-DLC film surfaces revealed that the PtRuN-DLC film surfaces
became more hydrophobic with increased Pt and Ru contents.
The corrosion resistance of the PtRuN-DLC films increased with increased Pt and Ru
contents in the films when the applied potential was below 1 V. However, with higher
polarization potentials beyond 1 V, the films with higher Pt and Ru contents showed higher
anodic currents in the 0.1 M NaCl solution due to the delamination of the films. Therefore, the
Pt/(C+Ru+Pt+N) and Ru/(C+Ru+Pt+N) should be optimized between 0.033 (20 W) and 0.035
(25 W) and between 0.037 (20 W) and 0.041 (25 W), respectively, to prevent the anodic
dissolution of the PtRuN-DLC films.
7.3. Conclusions on N-DLC and PtRuN-DLC films prepared by a DC magnetron
sputtering technique
Furthermore, the N-DLC and PtRuN-DLC thin films deposited on p-Si substrates
using a DC magnetron sputtering deposition system were used to investigate the effect of Pt
and Ru doping on the chemical composition, micro-structure, bonding structure, surface
activity and morphology, adhesion strength, corrosion resistance and cyclic voltammetric
behavior of the films.
It was found that the PtRuN-DLC film had the higher degree of graphitization than the
N-DLC film, which was responsible for the higher surface roughness and adhesion strength of
164
the PtRuN-DLC film than those of the N-DLC film. The larger water contact angle on the
PtRuN-DLC film than that on the N-DLC film indicated that the PtRuN-DLC film surface
was more hydrophobic than the N-DLC film surface.
In EIS results, the higher R2 and R3 values of the PtRuN-DLC film than those of the N-
DLC film implied that the introduction of Pt and Ru into the N-DLC film significantly
increased the charge transfer resistance and bulk electrical resistivity of the film.
Table 7.1: Major findings from three major work chapters (4 to 6)
N-DLC films (FCVA technique) Chapter 4
PtRuN-DLC films (Magnetron sputtering technique) Chapter 5
N-DLC and PtRuN-DLC films (Magnetron sputtering technique) Chapter 6
Increasing N incorporation in N-DLC films1. Improved the adhesion strength of the films, but decreased the corrosion resistance of the films.2. Wide electrochemical potential windows in acidic, neutral and basic solutions.3. Decreased the electrochemical potential windows of the films (disadvantage)4. Increased the film electrodes’ sensitivity to toxic metals such as single elements (Pb2+, Cu2+ and Hg2+) and multi elements (Pb2+, Cu2+ and Hg2+).
Increasing Pt and Ru incorporation in PtRuN-DLC films1. Increased the adhesion strength and the corrosion resistance of the films at lower applied potentials (less than 1 V).2. Increased the anodic dissolution of the films at higher applied potentials.
Pt and Ru incorporation in N-DLC films1. Increased the adhesion strength and corrosion resistance of the films.2. Decreased the electrochemical potential windows and degraded the electrochemical performance (EC) of the films.
The linear sweep cyclic voltammetric measurements of the N-DLC and PtRuN-DLC
films clearly indicated that the introduction of Pt and Ru into the N-DLC film significantly
narrowed down the potential windows of the film because of their catalytic activities though
the N-DLC film showed wide potential windows in 0.1 M H2SO4, 0.1 M HCl and 0.1 M KCl
solutions of about 2.65, 2.9 and 3.2 V, respectively. The N-DLC film showed a good
electrocatalytic activity in Fe(CN)64-/Fe(CN)6
3- redox reactions. However, an increased kinetic
limitation upon the PtRuN-DLC film shifted the oxidization peak to a more positive value and
165
the reduction peak to a more negative value compared to those obtained from the N-DLC
film. Such the increased kinetic limitation upon the PtRuN-DLC film was attributed to the
increased electrical resistivity of the film with Pt and Ru incorporation, which was consistent
with the EIS results. It could be deduced that the introduction of Pt and Ru into the N-DLC
film improved the corrosion resistance of the film but apparently degraded the
electrochemical performance of the film. The major findings from the three major work
chapters (4 to 6) were summarized in Table 7.1.
7.4. Recommendations for future work
The experimental results obtained from this project have clearly revealed that N-DLC
films have a great promise for detection of heavy metals such as Pb2+, Cu2+ and Hg2+. A more
investigation of the detection ability of the N-DLC films, including the detection of these
heavy metals in drinking water or biofluid and the analysis of various other metals such as Cd,
Au, etc., should be carried out in future. Fabrication of N-DLC films into microelectrodes is
an essential step to promote the sensitivity of the films to trace metals because the
microelectrodes allow working in high resistive media without loss of sensitivity due to the
small currents and to the fact that the ohmic drop is limited to a small area close to the
electrode. In addition, the small microelectrode diameter lowers the flow dependence of the
microelectrode. Therefore, micro-fabrication of N-DLC films should be taken into
consideration in future in order to use the films as microelectrode for electroanalysis.
166
References
[1] F. M. Kimock, B. J. Knapp, Surf. Coat. Tech. 56 (1993) 273.
[2] C. J. McHargue, Y. Tzeng, M. Yoshikawa, M. Murakawa, A. Feldman, “Applications of
Diamond films and related materials”, Materials science monographs, Elsevier, 1991, New
York.
[3] O. R. Awofolu, Z. Mbolekwa, V. Mtshemla, O. S. Fatoki, Water SA. 31 (2005) 87.
[4] E. G. Pacyna, J. M. Pacyna, F. Steenhuisen, S. Wilson, Atoms. Environ. 40 (2006) 4048.
[5] C. H. Ngim, S. C. Foo, K. W. Boey, J. Keyaratnam, Brit. J. Indu. Medic. 49 (1992) 782.
[6] Y. X. Liang, R. K. Sun, Z. Q. Chen, L. H. Li, Envi. Resea. 60 (1992) 320.
[7] R. B. McFarland, H. Reigel. J Occup Med. 20 (1978) 532.
[8] C. Prado, S. J. Wilkins, F. Marken, R. G. Compton, Electroanalysis. 14 (2002) 262.
[9] C.M.A. Brett, A.M. Oliveira Brett, “Electroanalysis”, first edi, Oxford University Press,
2005.
[10] F. Bonet, C. Guery, D. Guyomard, R. H. Urbina, K.T. Elhsissen, .M.Tarascon, Solid
State Ionics. 126 (1999) 337.
[11] M. Hupert, A. Muck, J. Wang, J. Stotter, Z. Cvackova, S. Haymond, Y. Show, G. M.
Swain, Diam. Relat. Mater. 12 (2003) 1940.
[12] J. W. Strojek, M. C. Granger, G. M. Swain, T. Dallas, M. W. Holtz, Anal. Chem. 405
(1996) 9.
[13] A. J. Saterlay, C. A. Gutierrez, M. P. Taylor, F. Marken, R. G. Compton, Electroanalysis.
11 (1999) 1083.
[14] A. J. Saterlay, J. S. Foord, R. G. Compton, Analyst. 124 (1999) 1791.
[15] A. J. Saterlay, D. F. Tibbetts, R. G. Compton, Anal. Sci. 16 (2000) 1055.
167
[16] N. Mizuochi, M. Ogura, H. Watanabe, J. Isoya, H. Okushi, S. Yamasaki, Diam. Relat.
Mater. 11 (2004) 2096.
[17] A. Fujishima, C. Terashima, K. Honda, B. V. Sarada, T. N. Rao, New. Diam. Front.
Carbon. Tech. 12 (2002) 73.
[18] L. X. Liu, E. Liu, Surf. Coat. Tech. 198 (2005) 189.
[19] A. Zeng, E. Liu, S. N. Tan, S. Zhang, J. Gao, Electroanal. 14 (2002) 1294.
[20] A. Zeng, E. Liu, S. N. Tan, S. Zhang, J. Gao, Electroanal. 14 (2002) 1110.
[21] A. J. Saterlay, C. Agra-Gutierrez, M. P. Taylor, F. Marken, R. G. Compton, Electroanal.
11 (1999) 1083.
[22] A. J. Saterlay, J. S. Foord, R. G. Compton, Analyst. 124 (1999) 1791.
[23] N. W. Khun, E. Liu, Electrochim. Acta 54 (2009) 2890.
[24] N. W. Khun, E. Liu, H. W. Guo, Electroanal. 20 (2008) 1851.
[25] C. Du, X.W. Su, F.Z. Cui, X.D. Zhu, Biomater. 19 (1998) 651.
[26] G. Francz, A. Schroeder, R. Hauert, Surf. Interf. Anal. 28 (1999) 3.
[27] R. Hauert, U. Mueller, G. Francz, et al., Thin Solid Films 308-309 (1997) 191.
[28] J. V. Quagliano, “Chemistry”, 2nd Edi, 1963, Japan.
[29] J. Robertson, Mater. Sci. Eng, R 37 (2002) 129.
[30] J. Robertson, Adv. Phys. 35 (1986) 317.
[31] A. Grill, Diam. Relat. Mater. 8 (1999) 428.
[32] W. Jacob, W.Moller, Appl. Phys. Lett. 63 (1993) 1771.
[33] D. R. McKenzie, Rep. Prog. Phys. 59 (1996) 1611.
[34] T. I. T. Okpalugo, P. D. Maguire, A. A. Ogwu, J. A. D. McLaughlin, Diam. Relat. Mater.
13 (2004) 1549.
168
[35] C. Corbella, E. Pascual, G. Oncins, C. Canal, J. L. Andujar, E. Bertran, Thin Solid Films
482 (2005) 293.
[36] D. Y. Wang, Y. Y. Chang, C. L. Chang, Y. W. Huang, Surf. Coat. Tech. 200 (2005)
2175.
[37] G. K. Burkat, T. Fujimura, V. Y. Dolmatov, E. A. Orlova, M. V. Veretennikova, Diam.
Relat. Mater. 14 (2005) 1761.
[38] I. G. Gonzalez, J. D. Jesus, D. A. Tryk, G. Morell, C. R. Cabrera, Diam. Relat. Mater. 15
(2006) 221.
[39] A. Grill, Surf. Coat.Tech. 9495 (1997) 507.
[40] A. Grill, V. Patel, K. L. Saenger, C. Jahnes, S. A. Cohen, A. G. Schrott, D. C. Edelstein,
J. R. Paraszczak, Mater. Res. Soc. Symp. Proc.443 (1997) 155.
[41] A. Grill, V. Patel, C. Jahnes, J. Electrochem. Soc.145 (1998) 1649.
[42] M. Grishke, A. Hieke, F. Morgenweck, H. Dimigen, Diam. Relat. Mater. 7 (1998) 454
[43] J. Robertson, Surf. Coat. Tech. 50 (1992) 185.
[44] B. Schultrich, H. J. Schultrich, D. Drescher, H. Ziegele, Surf. Coat. Tech. 98 (1998)
1097.
[45] J. W. Ager, III. S. Anders, I. G. Brown, M. Nastasi, K. C. Walter, Surf. Coat. Tech. 91
(1997) 91.
[46] S. Anders, D. L. Callahan, G. M. Pharr, T. Y. Tsui, C. S. Bhaia, Surf. Coat. Tech. 9495
(1997) 189.
[47] B. Schultrich, H. J. Scheibe, G. Grandremy, D. Drescher, D. Schneider, Diam. Relat.
Mater. 5 (1996) 914.
[48] J. Robertson, E. P. O’Reilly, Phys. Rev. B 35 (1987) 2946.
[49] J. Robertson, Mater. Sci. Forum. 52 (1990) 125.
169
[50] N. G. Mott, Adv. Phys. 16 (1967) 49.
[51] R. A. Street, Phys. Rev. Lett. 49 (1982) 1187.
[52] X. Shi, H. Fu, J. R. Shi, L. K. Cheah, B. K. Tay, P. Hui, J. Phys. Condens. Mater. 10
(1998) 9293.
[53] J. Robertson, C. A. Davis, Diam. Relat. Mater. 4 (1995) 441.
[54] S. E. Rodil, S. Muhl, Diam. Relat. Mater. 13 (2004) 1521.
[55] H. L. Bai, E. Y. Jiang, Thin solid film. 353 (1999) 157.
[56] M. L. Morrison, R. A. Buchanan, P. K. Liaw, C. J. Berry, R. L. Brigmon, L. Riester, H.
Abernathy, C. Jin, R. J. Narayan, Diam. Relat. Mater. 15 (2006) 138.
[57] J. Schwan, S. Ulrich, H. Roth, H. Ehrhardt, S. R. P. Silva, J. Robertson, R. Samlenski, J.
Appl. Phys. 79 (1996) 1416.
[58] X. Shi, B. K. Tay, S. P. Lau, Int. J. Mod. Phys. 14 (2000) 154.
[59] A. A. Voevodin, M. S. Donley, Surf. Coat. Tech. 82 (1996) 199.
[60] B. Keiper, S. WeiBmantel, G. R.eiBe, S. Schulze, SPIE, 1 (1999) 403.
[61] J. Hu, P. Yang, C. M. Lieber, Appl. Surf. Sci. 127-129 (1998) 569.
[62] S. Zhang, X. L. Bui, Y. Fu, Surf. Coat. Tech. 167 (2003) 137.
[63] Y. Lifshitz, Diam. Rel. Mater. 8 (1999) 1659.
[64] S. Zhang, H. Du, S. E. Ong, K. N. Aung, H. C. Too, X. Miao, Thin Solid Films. 515
(2006) 66.
[65] H. G. Kim, S. H. Ahn, J. G. Kim, S. J. Park, K. R. Lee, Thin Solid Films. 482 (2005)
299.
[66] M. Maharizi, O. Segal, E. B. Jacob, Y. Rosenwaks, T. Meoded, N. Croitoru, A. Seidman,
Diam. Relt. Mater. 8 (1999) 1050.
170
[67] E. Liu, X. Shi, H. S. Tan, L. K.Cheah, Z. Sun, B. K. Tay, J. R. Shi, Surf. Coat. Tech.
120-121 (1999) 601.
[68] H. Ma, L. Zhang, N. Yao, B. Zhang, H. Hu, G. Wen, Diam. Rela. Mater. 9 (2000) 1608.
[69] C. Y. Cheng, F. C. H. N. Hong, Thin Solid Films. 498 (2006) 206.
[70] Y. Catherine, R.E.Clausing (Eds), “Diamond and Diamond-like Films and Coatings”,
Plenum, 1991, New York, 1991.
[71] S. Zhang, H. Xie, Surf. Coat. Technol. 113 (1999) 120.
[72] S. Zhang, H. Xie, P. Hing, Z. Mo, Surf. Eng. 15. (1999) 4.
[73] J. P. Sullivan, T. A. Friedmann, A.G. Baca, J. Electron. Mater. 26 (1997) 1021.
[74] P. Hedenqvist, M. Olsson, S. Macobson, S. Soderberg, Surf. Coat. Tech. 41 (1990) 31.
[75] H. E. Hintermann, Wear 100 (1984) 381.
[76] U. Wiklund, J. Gunnars, S. Hogmark, Wear 232 (1999) 262.
[77] A. J. Perry, Thin Solid Films 81 (1981) 357.
[78] D. Sheeja, B. K. Tay, K. W. Leong, C. H. Lee, Diam. Rela. Mater. 11 (2002) 1643.
[79] E. Martinez, M. C. Polo, E. Pascual, J. Esteve, Diam. Rel. Mater. 8 (1999) 563.
[80] B. K. Gupta, B. Bhushan, Thin Solid Films 170 (1995) 391.
[81] J. F. Lin, C. C. Wei, Y. K. Yung, C. F. Ai, Diam. Rela. Mater. 13 (2004) 1895.
[82] W. S. Tait, “An introduction to electrochemical corrosion testing for practicing engineers
and scientists”, 1994, USA.
[83] B. Bhushan, Tribology and Mechanics of Magnetic Storage Devices, 2nd Ed, Springer,
New York, 1996
[84] A. Zeng, E. Liu, I. F. Annergren, S. N. Tan, S. Zhang, P. Hing, J. Gao, Diam. Relat.
Mater. 11 (2002) 160.
[85] R. R. Dubin, K. D. Winn, L. P. Davis, R. A. Culter, J. Appl. Phys. 53 (1982) 2579.
171
[86] V. Novotny, G. Itnyre, A. Homola, L. Franco, IEEE Trans. Magn. 25 (5) (1987) 3645.
[87] D. Wang, W. D. Doyle, H. M. Saffarian, G. W. Warren, IEEE Trans. Magn. 31 (6)
(1995) 2761.
[88] O. E. Mark, T. Bernard, “Electrochemical impedance spectroscopy”, John Wiley & Sons,
2008, USA.
[89] M. L. Morrison, R. A. Buchanan, P. K. Liaw, C. J. Berry, R. L. Brigmon, L. Riester, H.
Abernathy, C. Jin, R. J. Narayan, Diam. Relat. Mater. 15 (2006) 138.
[90] B. Tomcik, S. C. Seng, B. Balakrisnam, J. Y. Lee, Diam. Relat. Mater. 11 (2002) 1409.
[91] G. F. Huang, Z. Lingping, H. Weiqing, Z. Lihua, L. Shaolu, L. Deyi, Diam. Relat. Mater.
12 (2003) 1406.
[92] Z. H. Liu, J. F. Zhao, J. McLaughlin, Diam. Relat. Mater. 8 (1999) 56.
[93] M. A. S. Oliveira, A. K. Vieira, M. Massi, Diam. Relat. Mater. 12 (2003) 2136.
[94] E. T. Uzumaki, C. S. Lambert, W. D. Belangero, C. M. A. Freire, C. A. C. Zavaglia,
Diam. Relat. Mater. 15 (2006) 982.
[95] P. Papakonstantinou, J. F. Zhao, P. Lemoine, E. T. McAdams, J. A. McLaughlin, Diam.
Relat. Mater. 11 (2002) 1074.
[96] H. G. Kim, S. H. Ahn, J. G. Kim, S. J. Park, K.R. Lee, Diam. Relat. Mater. 14 (2005) 53.
[97] R. Sharma, P. K. Barhai, N. Kumari, Thin Solid Films (2007) 7.
[98] C. L. Liu, D. Hu, J. Xu, D. Yang, M. Qi, Thin Solid Films 496 (2006) 457.
[99] P. Papakonstantinou, J. F. Zhao, A. Richardot, E. T. McAdams, J. A. McLaugnlin, Diam.
Relat. Mater. 11 (2002) 1124.
[100 N. W. Khun, E. Liu, X. T. Zeng, Corro. Sci. 51 (2009) 2158.
[101] N. W. Khun, E. Liu, G. C. Yang, W. G. Ma, S. P. Jiang, J. Appl. Phys. 106 (2009)
013506.
172
[102] J. H. Marsh, S. W. Orchard, Carbon 30 (1992) 895.
[103] R. Ramesham, M. F. Rose, Diam. Relat. Mater. 6 (1997) 17.
[104] T. A. Ivandini, R. Sato, Y. Makide, A. Fujishima, Y. Einaga, Diam. Relat. Mater. 14
(2005) 2133.
[105] W. Siangproh, N. Wangfuengkanagul, O. Chailapakul, Anal. Chimi. Acta. 499 (2003)
183.
[106] K. J. McKenzie, D. Asogan, F. Marken, Electroche. Communi. 4 (2002) 820.
[107] S. Miyagawa, S. Nakao, M. Ikeyama, Y. Miyagawa, Surf. Coat. Tech. 156 (2002) 322.
[108] H. B. Martin, A. A. Argoitia, U. Landau, A. B. Anderson, J. C. Angus, J. Electrochem.
Soc. 143 (1996) 133.
[109] J. P. Randin, in Encyclopedia of Electrochemistry of the Elements, Vol. VII (Ed: A. J.
Bard), Marcel Dekker, 1974, New York.
[110] J. S. Mattson, C. A. Smith, Analyt. Chemi. 47 (1995) 1122.
[111] J. Karpman, M. R. Fishman, J. Zahavi, D. Dhamelincourt, Diam. Relat. Mater. 4
(1994)10.
[112] D. H. Kim, H. E. Kim, K. R. Lee, C. N. Whang, I. S. Lee, Mater. Sci. Engi. 22 (2002) 9
[113] H. P. Feng, C. H. Hsu, J. K. Liu, Y. H. Shy, Mater. Sci. Engi. 347 (2003) 123.
[114] R. Hauert, Diam. Relat. Mater. 12 (2003) 583.
[115] Dement’ev, Diam. Relat. Mater. 8 (1999) 64
[116] Y. V. Pleskov, Y. E. Evstefeeva, A. M. Baranov, Diam. Relat. Mater. 11 (2002) 1518.
[117] R. Schlesinger, M. Bruns, H. J. Ache, J. Electrochem. Soci. 144 (1997) 6.
[118] G. Yang, E. Liu, N. W. Khun, S. P. Jiang, J. Electroanal.. Chem. 627 (2009) 51.
[119] K. Yoo, B. Miller, R. Kalish, X. Shi, Electroche. Solid-State. Lett. 2 (1999) 233.
173
[120] A. Lagrini, S. Charvet, M. Benlahsen, H. Cachet, C. Deslouis, Diam. Relat. Mater. 16
(2007) 1378.
[121] C. A. Gutierrez, J. L. Hardcastle, J. C. Ball, R. G. Compton, Analyst. 124 (1999) 236.
[122] J. Wang, Stripping Analysis, VCH, Deerfield Beach, FL, 1985, 119.
[123] H. B. P. Ferreira, T. A. F. de Lima, P. J. S. Barbeira, Electroanalysis. (2007)
(imprinted).
[124] M. Panigati, M. Piccone, G. D’Alfonso, M. Orioli, M. Carini, Talanta, 58 (2002) 481.
[125] J. Barek, A. G. Fogg, A. Muck, J. Zima, Crit. Review. Analy. Chem. 31 (2001) 291.
[126] D. At, A. Manivannan, A. Fujishima, Electrochem. Solid-State, Lett. 2 (1999) 455.
[127] M. C. Granger, J. Xu, J. W. Strojek, G. M. Swain, Anal. Chem. Acta. 297 (1999) 145.
[128] O. E. Tall, N. J. Renault, M. Sigaud, O. Vittori, Electroanaly. 19-11 (2007) 1152.
[129] N. W. Khun, E. Liu, Thin Solid Films (2010) (in press).
[130] V. Rehacek, I. Hotovy, M. Vojs, Sensor. Actuat. B-Chem 127 (2007) 193.
[131] V. Rehacek, I. Hotovy, M. Vojs, F. Mika, Microsyst. Technol. 14 (2008) 491.
[132] Y. Shao, M. B. Mirkin, Anal. Chemi. 69 (1997) 1627.
[133] M. J. Schining, G. Bu, F. Fablender, O. Gluk, H. Emons, G. Schmitt, J. W. Schutze, H.
Luth, Sensors and Actuators B. 65 (2000) 284
[134] H. Nakazawa, T. Mikami, Y. Enta, M. Suemitsu, M. Mashita, jpn, J. Appl. Phys. 42
(2003) L676.
[135] D. F. Flick, H. F. Kraybill, J. M. Dimittroff, Environ. Res. 4 (1971)71.
[136] H. B. P. Ferreira, T. A. F. de Lima, P. J. S. Barbeira, Electroanal. 20 (2007) 390.
[137] A. C. Ferrari, J. Robertson, Phys. Rev. B 20. 61 (2000) 14095.
[138] S. S. Roy, P. Papakonstantinou, R. McCann, G. Abbas, J. P. Quinn, J. A. McLaughlin,
Diam. Relat. Mater.13 (2004) 1459.
174
[139] N. Akita, Y. Konishi, S. Ogura, M. Imamura, Y. H. Hu, X. Shi, Diam. Relat. Mater. 10
(2001) 1017.
[140] W. Cai, J. H. Sui, Surf. Coat. Tech. 201 (2007) 5194.
[141] B. Tomcik, T. Osipowicz, J. Y. Lee, Thin Solid Films 360 (2000) 173.
[142] S. Mahieu, P. Ghekiere, D. Depla, R. De Gryse, Thin Solid Films 515 (2006) 1229.
[143] J. Seth, R. Padiyath, S. V. Babu, Diam. Relat. Mater. 3 (1994) 210.
[144] Y. V. Pleskov, M. D. Krotova, A. V. Saveliev, V. G. Ralchenko, Diam. Relat. Mater. 9
(2007) 1.
[145] B. Kleinsorge, A. C. Ferrari, J. Robertson, W. I. Milne, S. Waidmann, S. Hearne, Diam.
Relat. Mater. 9 (2000) 643.
[146] K. Yoo, B. Miller, R. Kalish, X. Shi, Electrochem. Solid-State. Lett. 2 (1999) 233.
[147] C. M. A. Brett, A. M. O. Brett, “Electroanalysis”, 1st ed., Oxford university press, 2005,
UK.
[148] K. Wikiel, J. Osteryoung, Electrochim. Acta. 15 (1993) 2291.
[149] M. L. Tercier, J. Buffle, Electroanalysis 5 (1993) 187.
[150] K. R. Wehmeyer, R. M. Wightman, Anal. Chem. 57 (1985) 1989.
[151] M. A. Baldo, C. Bragato, G. A. Mazzocchin, S. Daniele, Electrochim. Acta. 43 (1998)
3413.
[152] G. Gunawardena, G. Hills, I. Montenegro, J. Electroanal. Chem. 357 (1985) 184.
[153] T. Lee, K. C. Chung, J. Park, Electroanal. 14 (2002) 833.
[154] M. Schatter, Diam. Relat. Mater. 11 (2002) 1781.
[155] M. Neuhaeuser, H. Hilgers, P. Joeris, R. White, J. Windeln, Diam. Relat. Mater. 9
(2000) 1500.
[156] J. E. Mahan, “Physical vapor deposition of thin films”, 1st ed, 2000, USA.
175
[157] D. G. Liu, J. F. Lee, M. T. Tang, J. Molec. Cata. A. 240 (2005) 197.
[158] P. K. Babu, H. S. Kim, E. Oldfield, A. Wieckowski, J. Phys. Chem. B. 107 (2003)
7595.
[159] R. M. Bradley, J. M. E. Harper, D. A. Smith, J. Appl. Phys. 60 (12) (1986) 4160.
[160] S. E. Ong, S. Zhang, H. Du, D. Sun, Diam. Relat. Mater. 16 (2007) 1628.
[161] J. L. Jauberteau, D. Duchesne, C. Girault, J. Aubreton, A. Catherinot, Plasm. Chem.
Plasm. Process. 10 (1990) 589.
[162] K. Bewilogua, C. V. Cooper, C. Specht, J. Schroder, R. Wittorf, M. Grischke, Surf.
Coat. Tech. 132 (2000) 275.
[163] J. T. Titantah, D. Lamoen, Diam. Relat. Mater. 16 (2007) 581.
[164] J. Vlcek, M. Kormunda, J. Cizek, Z. Soukup, V. Perina, J. Zemek, Diam. Relat. Mater.
12 (2003) 1287.
[165] Y. S. Zou, Q. M. Wang, H. Du, G. H. Song, J. Q. Xiao, J. Gong, C. Sun, L. S. Wen,
Appl. Surf. Sci. 241 (2005) 295.
[166] R. K. Raman, A. K. Shukla, A. Gayen, M. S. Hegde, K. R. Priolkar, P. R. Sarode, S.
Emura, J. Pow. Sou. 157 (2006) 45.
[167] A. S. Arico, V. Baglio, A. D. Blasi, E. Modica, P. L. Antonucci, V. Antonucci, J.
Electroanal. Chem. 557 (2003) 167.
[168] G. D. Lian, E. C. Dickey, M. Ueno, M. K. Sunkara, Diam. Relat. Mater. 11 (2002) 1890
[169] B. K. Tay, P. Zhang, Thin Solid Films 420-421 (2002) 177.
[170] Y. Hayashi, K. M. Krishna, H. Ebisu, T. Soga, M. Umeno, T. Jimbo, Diam. Relat.
Mater. 10 (2001) 1002.
[171] C. Cattaneo, M. I. Sanchez de Pinto, H. Mishima, B. A. Lopez de Mishima, D. Lescano,
L. Cornaglia, J. Electroanal. Chem. 461 (1999) 32.
176
[172] G. Reisel, G. Irmer, B. Wielage, A. D. Reisel, Thin Solid Films 515 (2006) 1038.
[173] H. Daimon, Y. Kurobe, Catalysis Today 111 (2006) 182.
[174] S. Ulrich, T. Theel, J. Schwan, H. Ehrhardt, Surf. Coat. Tech. 97 (1997) 45.
[175] X. Shi, B. K. Tay, H. S. Tan, L. Zhong, Y. Q. Tu, S. P. R. Silva, W. I. Milne, J. Apply.
Phys. 79 (1996) 7239.
[176] A. Y. Wang, K. R. Lee, J. P. Han, Carbon 44 (2006) 1826.
[177] A. Y. Wang, H. S. Ahn, K. R. Lee, J. P. Ahn, Appl. Phys. Lett. 86 (2005) 111902.
[178] H. Du, B. Li, F. Kang, R. Fu, Y. Zeng, Carbon 45 (2007) 429.
[179] J. M. Ting, H. Lee, Diam. Relat. Mater. 11 (2002) 1119.
[180] A. H. Tan, Y. C. Cheng, Diam. Relat, Mater. 17 (2008) 36.
[181] A. D. Reisel, C. Schurer, G. Irmer, E. Muller, Surf. Coat. Tech. 177-178 (2004) 830.
[182] C. S. Lee, T. Y. Kim, K. R. Lee, K. H. Yoon, Thin Solid Films 447-448 (2004) 169.
[183] H. Hofsass, H. Feldermann,, R. Merk, M. Sebastian, C. Ronning, Appl. Phys. A 153
(1998).
[184] A. Oya, S. Otani, Carbon 17 (1979) 131.
[185] G. J. Qi, S. Zhang, T. T. Tang, J. F. Li, X. W.Sun, X. T Zeng, Surf. Coat. Tech. 198
(2005) 300.
[186] K. Bewilogua, R. Wittorf, H. Thomsen, M. Weber, Thin Solid Films 447-448 (2004)
142.
[187] C. Corbella, E. Bertran, M. C. Polo, E. Pascual, J. L. Andujar, Diam. Relat. Mater. 16
(2007) 1828.
[188] H. W. Choi, R. H. Dauskardt, S. C. Lee, K. R. Lee, K. H. Oh, Diam. Relat. Mater. 17
(2008) 252.
[189] H. K. Livingston, C. S. Swingley, Surf. Sci. 24 (1971) 625.
177
[190] L. Y. Chen, F. C. N. Hong, Diam. Relat. Mater. 12 (2003) 968.
[191] M. Grischke, K. Bewilogua, K. Trojan, H. Dimigen, Surf. Coat. Tech. 74-75 (1995)
739.
[192] Z. Zhang, X. Lu, J. Luo, Y. Liu, C. Zhang, Surf. Coat. Tech. 202 (2008) 1621.
[193] H. Wang, S. Zhang, Y. Li, and D. Sun, Thin Solid Films 516 (2008) 5419.
[194] J. C. Angus, F. Jansen, J. Vac. Sci. Tech. A. 6 (1988) 1778.
[195] D. Nir. Thin Solid Films 112 (1984) 41.
[196] J. Seth, R. Padiyath, S. V. Babu, J. Vac. Sci. Tech. A. 10 (1992) 284.
[197 C. Corbella, G. Oncins, M. A. Gomez, M. C. Polo, E. Pascual, J. Garcia-Cespedes, J. L.
Andujar, E. Bertran, Diam. Relat. Mater. 14 (2005) 1103.
[198] C. Strondl, N. M. Carvalho, J. T. M. DeHosson, G. J. VanderKolk, Surf. Coat. Tech.
162 (2003) 288.
[199] S. I. Pyun, Y. G. Ryu, S. H. Choi, Carbon 32 (1994) 161.
[200] A. N. Gavrilov, O. A. Petrii, A. A. Mukovnin, N. V. Smirnova, T. V. Levchenko, G. A.
Tsirlina, Electrochemi. Acta. 52 (2007) 2775.
[201] A. P. Yadav, A. Nishikata, T. Tsuru, Electrochemi. Acta. 52 (2007) 7444.
[202] E. Liu, H. W. Kwek, Thin Solid Films 2007, 516, 5201.
[203] K. I. Schiffmann, M. Fryda, G. Goerigk, R. Lauer, P. Hinze, A. Bulack, Thin Solid
Films 1999, 347, 60.
[204] J. P. Randin, Encyclopedia of Electrochemistry of Elements, Vol. VII (ed: A. J. Bard),
Marcel Dekker, 1974, New York.
[205] M. C. Tavares, S. A. S. Machado, L. H. Mazo, Electrochim. Acta 2001, 46, 4359.
[206] B. E. Conway, L. Bai, J. Electroanal. Chem. 1986, 198, 149.
[207] M. Elam, B. E. Conway, J. Appl. Electrochem. 1987, 17, 1002.
178
[208] K. J. Vetter, Electrochemical Kinetics-Theoretical and Experimental Aspects, Academic
Press, 1967,New York.
[209] M. W. Breiter, J. Phy. Chem. 1964, 68, 2249.
179