design development and applications
TRANSCRIPT
RUHR UNIVERSITÄT BOCHUM
Integrated Scanning Kelvin probe – Scanning
Electrochemical Microscopy system: design,
development and applications
Dissertation
zur Erlangung eines Doktors
der Naturwissenschaften
der Fakultät für Chemie und Biochemie
an der Ruhr-Universität Bochum
vorgelegt von
Artjom Maljusch
Bochum, Juli 2012
This work was carried out in the period from October 2007 to July 2012 at the faculty of
Chemistry and Biochemistry, AG Elektronanalytik & Sensorik under the supervision of Prof. Dr.
Wolfgang Schuhmann, Ruhr Universität Bochum, Germany.
Eingereicht am: 06.08.2012
Vorsitz: Prof. Dr. rer. nat. M. Wark
Referent: Prof. Dr. rer. nat. W. Schuhmann
Koreferent: Prof. Dr. rer. nat. M. Stratmann
III
Acknowledgement
I think that it would not be possible to do all this work and to finalize this doctoral thesis without the help of many people. To all of them I am very thankful.
First of all, I would like to thank Prof. Dr. Wolfgang Schuhmann who gave me a chance to go my own way, who showed me how the world of science functions, who supported me during my scientific career and development of my personality, who helped me to spread the way of thinking, who taught me how to communicate with people, who was for me more than just a supervisor, more than just a mentor, who was and is a very good friend. I will never forget that short talk in the middle of 2005 when he asked me about my plans for the future and they were ambitious, but support was missing. He just sad “I will help you” and he did.
Prof. Dr. Martin Stratmann for kindly taking over the co-referat of my doctoral thesis.
Bettina Stetzka for her always friendly smile, her support for all administrative stuff and applications, her very profound and detailed explanations of basic principles of the German social system and organizations, and especially for many small talks about things which I was never thinking about, and also for challenging but very lovely minutes during social evenings by playing Jungle Jam.
Bernd Schönberger for all his knowledge and experience which he so kindly shared with me, his always very detailed and profound explanations of the function of any electrical component used in the integrated SKP-SECM system and especially for his enormous input by designing the SKP-SECM system and also development of the Kelvin current amplifier and the integrator implemented in the developed system.
Armin Lindner for all his time, energy and effort by designing mechanical components used in the integrated SKP-SECM system and also to the team of the mechanical workshop of the faculty of chemistry and biochemistry of Ruhr University Bochum for building all of these components.
Dr. Michael Rohwerder for his interesting ideas at the stage of designing the integrated SKP-
SECM system and for valuable scientific discussions about the theory of the SKP and interesting interpretations of experimental results.
Dr. Ceylan Senöz for her great cooperation and synthesis of Al2CuMg sample.
M.Sc. Kirill Sliozberg and Dr. Dmitrii Guschin for their willingness to help and support, their understanding, interesting discussions about science and other topics, both of them for being
IV
very good friends. I am very grateful to Kirill for evenings at his place. Without him many photographs presented in this work would never be so nice.
Dr. Aliaksandr Bandarenka for introducing me to the field of electrocatalysis on well-ordered surfaces, for interesting and valuable discussions about science and life, as well as his great cooperation and very friendly attitude. It was a big pleasure for me to work with him and members of his research group as also to publish joint articles.
M.Sc. Aleksandar Zeradjanin for sharing with me his fundamental knowledge in electrochemistry and his broad understanding of political, economic and social processes going on all around the world. I enjoyed very much our discussions and especially his outstanding sense of humor.
Dr. John Henry and M.Sc. Justus Masa for their enormous effort to improve the linguistic aspects of this doctoral thesis and also all published articles to such a high level. Especially to John I am very grateful for his great cooperation and joint scientific work.
Dr. Michaela Nebel for interesting and valuable scientific discussions about electrical noise reduction, some practical aspects of SECM application and also for very kindly sharing with me her pulled ultramicroelectrodes.
Dr. Kathrin Eckhard for introducing me to the SECM world and for very kind sharing with me all her knowledge and experience in this field.
Dr. Thomas Erichsen and Dr. Dominik Schäfer for programming and troubleshooting the software modules for control of the integrated SKP-SECM system, for taking care of the local network and also solving any software and hardware related issues.
M.Sc. Alan Savan and Dr. Sigurd Thienhaus for fabrication of high-quality test samples.
All members of the Elektroanalytik & Sensorik research group namely Edgar Ventosa, Jorge Eduardo Yanez Heras, Yvonne Beyl, Julia Hebel, Giorgia Zampardi, Raoudha Haddad, Magdalena Gebala and Andrea Puschhof for nice discussions during coffee breaks and social evenings in a nice and warm atmosphere.
The International Max-Plank Research School SurMat for financial support during the three years of this work and especially to Dr. Christoph Somsen, Dr. Rebekka Loschen and Elke Gattermann who guided me through my fellowship time in SurMat and whose company I enjoyed
very much during many social activities.
V
Finally, I would like to say how deeply grateful I am to my family.
I am very thankful to my grandmother who gave me the desire to write a PhD thesis as she did in the early 1970s. I will never forget this feeling of literally “touching the science” when I was preparing for examinations during study at my home university while sitting in front of her working table full of hand written lectures in economics.
I am deeply thankful to my parents for their love, patience, time they invested in me, my education and my personal development, for their unlimited trust and infinite support, for being always present when I needed their advice, for being always honest with me, for giving me all that they had and doing all that they could.
Without my family I would never be the person who I am now.
Дорогие, мама и папа,
я безгранично благодарен Вам за вашу любовь, терпение, время, которые Вы уделяли мне, которые Вы инвестировали в моё образование и моё становление как личности. Огромное Вам спасибо за ваше безграничное доверие и поддержку, за то, что Вы всегда рядом, когда мне нужен Ваш совет, за то, что вы всегда честны со мной, за то, что Вы дали мне всё то, что у Вас было и сделали всё, что могли.
Без Вас я бы не стал тем, кто я сейчас.
“The loss of imagination is death in itself”
Eleanor Roosevelt
VI
Acknowledgement .......................................................................................................................... III
1. Introduction ............................................................................................................................... 1
2. State of the art ........................................................................................................................... 3
2.1. Scanning Kelvin Probe (SKP) ............................................................................................ 3
2.1.1. The Work Function ..................................................................................................... 3
2.1.2. The Contact Potential Difference (CPD) .................................................................... 5
2.1.3. The scanning Kelvin probe (SKP) .............................................................................. 6
2.1.4. The absolute electrode potential (Eabs) ........................................................................ 8
2.1.5. Practical aspects of application of the SKP ................................................................ 9
2.1.5.1. Calibration of the system ..................................................................................... 9
2.1.5.2. Sample preparation ............................................................................................ 10
2.1.5.3. Tip preparation .................................................................................................. 10
2.1.5.4. Lateral resolution ............................................................................................... 11
2.1.6. Scanning Kelvin Probe Force Microscopy (SKPFM) .............................................. 12
2.1.7. Application of SKP ................................................................................................... 15
2.1.7.1. Determination of corrosion potential ................................................................. 15
2.1.7.2. Atmospheric corrosion ...................................................................................... 18
2.1.7.3. Delamination experiments ................................................................................. 20
2.2 Scanning Electrochemical Microscopy (SECM) ............................................................. 23
2.2.1. Practical aspects of the application of SECM ........................................................... 24
2.2.1.1. General aspects .................................................................................................. 24
2.2.1.2. Fabrication of SECM tips .................................................................................. 25
2.2.1.3. Lateral resolution ............................................................................................... 26
2.2.1.4. Control of the tip-to-sample distance ................................................................ 27
2.2.2. Operation modes of the SECM ................................................................................. 29
2.2.2.1. Feedback mode (FB) ......................................................................................... 29
2.2.2.2. Direct mode ....................................................................................................... 33
2.2.2.3. Generation-collection mode (GC) ..................................................................... 35
2.2.2.4. Alternating current mode (AC-SECM) ............................................................. 39
2.2.2.5. Redox-competition mode (RC-SECM) ............................................................. 44
3. Problem identification ............................................................................................................. 49
4. Results and Discussion ........................................................................................................... 51
VII
4.1. Preliminary investigations ................................................................................................ 51
4.1.1. Combined SKP, SKPFM and SECM measurements ................................................ 51
4.1.2. Conclusion ................................................................................................................ 56
4.2. Development of the integrated SKP-SECM system ........................................................ 57
4.2.1. Design of the integrated SKP-SECM system ........................................................... 57
4.2.2. Mechanical components of the integrated SKP-SECM system ................................ 61
4.2.2.1. Design and development of the measurement head .......................................... 61
4.2.2.2. Assembly procedure for the SKP-SECM measurement head ........................... 63
4.2.2.3. Design and development of the sample holder unit .......................................... 64
4.2.2.4. Design of the vibration damping unit ................................................................ 66
4.2.3. Electrical components of the integrated SKP-SECM system ................................... 67
4.2.3.1. The design of the Kelvin current amplifier ....................................................... 67
4.2.3.2. The design of the integrator ............................................................................... 69
4.2.3.3. The high-frequency power supply ..................................................................... 71
4.2.3.4. The lock-in amplifier ......................................................................................... 72
4.2.3.5. Peripheral components ...................................................................................... 74
4.3. Performance of the integrated SKP-SECM system.......................................................... 75
4.3.1. Settings and first tests of the developed SKP-SECM system ................................... 75
4.3.2. Calibration of the combined SKP-SECM system ..................................................... 85
4.3.3. Off-null measurement of the Volta potential ............................................................ 86
4.3.4. Test measurements with 125 µm / 25 µm Pt SKP-SECM electrodes ....................... 89
4.3.5. Compensation of the electrostatic charging of the Pt SKP-SECM electrodes.......... 93
4.3.6. Application of water saturated air ............................................................................. 94
4.3.6.1. New design of the sample holder ...................................................................... 94
4.3.6.2. Influence of H2O saturated air on the SKP-SECM system operation ............... 95
4.3.6.3. Lateral resolution of the modified SKP-SECM system................................... 100
4.3.7. Application of the external compensation voltage.................................................. 100
4.4. Optimization of the integrated SKP-SECM system ....................................................... 102
4.4.1. Optimization of mechanical components................................................................ 102
4.4.2. Optimization of electrical components ................................................................... 102
4.5. Development of “glass free” Pt SKP-SECM electrodes ................................................ 103
4.5.1. The concept of “glass free” Pt SKP-SECM electrodes ........................................... 103
4.5.2. Etching of Pt SKP-SECM electrodes in concentrated HF ...................................... 104
VIII
4.5.3. Reduction of the outer diameter of Pt SKP-SECM electrodes ............................... 106
4.5.4. Insulation of Pt SKP-SECM electrodes .................................................................. 108
4.6. Optimization of working parameters.............................................................................. 113
4.7. Investigation of the optimized SKP-SECM system performance .................................. 114
4.7.1. Operation of the integrated SKP-SECM system using “glass free” electrodes ...... 115
4.7.2. Constant distance operation of the integrated SKP-SECM system ........................ 119
4.7.2.1. Software based tilt correction procedure ......................................................... 119
4.7.2.2. Software based feedback controller ................................................................. 120
4.7.3. Lateral resolution of the optimized SKP-SECM system ........................................ 123
4.7.4. Vertical resolution of the optimized SKP-SECM system ....................................... 124
4.8. Application of the SKP-SECM system .......................................................................... 128
4.8.1. Investigation of high-performance catalyst for oxygen reduction reaction ............ 128
4.8.1.1. State of the art in ORR electrocatalysis ........................................................... 128
4.8.1.2. Concept of research ......................................................................................... 136
4.8.1.3. Preparation of Pt(111)-like thin films .............................................................. 137
4.8.1.4. Preparation of PtCu NSA on Pt(111)-like thin films ....................................... 145
4.8.1.5. Investigation of Cu-Pt(111) NSA using the integrated SKP-SECM system ... 154
4.8.2. Combined high-resolution SKP-SECM investigations of local corrosion ............. 162
4.8.2.1. State of the art in research on Al alloys ........................................................... 162
4.8.2.2. Synthesis of bulk Al2CuMg single crystals ..................................................... 165
4.8.2.3. Characterization of bulk Al2CuMg single crystals .......................................... 166
4.8.2.4. SKP-SECM measurements on synthesized Al2CuMg single crystals ............. 169
4.8.2.5. Conclusion ....................................................................................................... 176
5. Conclusions ........................................................................................................................... 177
6. Outlook ................................................................................................................................. 184
7. Experimental part .................................................................................................................. 185
7.1. Materials ......................................................................................................................... 185
7.1.1. Analytical reagents.................................................................................................. 185
7.1.2. Solutions ................................................................................................................. 185
7.1.3. Consumables ........................................................................................................... 186
7.1.4. Instrumentation ....................................................................................................... 187
7.1.5. Software .................................................................................................................. 190
7.2. Fabrication of electrodes ................................................................................................ 191
IX
7.2.1. Glass insulated Pt-disk microelectrodes ................................................................. 191
7.2.2. Glass insulated Pt SKP-SECM electrodes .............................................................. 191
7.2.3. “Glass free” Pt SKP-SECM electrodes ................................................................... 191
7.2.4. Insulated “glass free” Pt SKP-SECM electrodes .................................................... 192
7.2.5. Pt wire electrodes .................................................................................................... 192
7.2.6. Miniaturized reference electrodes ........................................................................... 192
7.2.7. Miniaturized Pt counter electrodes ......................................................................... 193
7.2.8. Deposition of a Ag layer on the glass surface of the Pt SKP-SECM tip ................ 193
7.3. Fabrication and preparation of samples ......................................................................... 194
7.3.1. Single Cu islands deposited on an Al layer (Type A test sample) .......................... 194
7.3.2. Pt layer partially deposited on a W layer (Type B test sample) .............................. 194
7.3.3. Pt grid structure deposited on a W layer (Type C test sample) .............................. 194
7.3.4. Polycrystalline Pt thin film sample ......................................................................... 195
7.3.5. Bulk Pt(111) single crystal...................................................................................... 195
7.3.6. Pt(111)-like thin film sample .................................................................................. 196
7.3.7. Cu-Pt near-surface alloy on Pt(111)-thin film sample ............................................ 196
7.3.8. Al2CuMg single crystals in solid solution Al matrix .............................................. 197
8. References ............................................................................................................................. 198
9. Appendix ............................................................................................................................... 215
9.1. Technical drawings ........................................................................................................ 215
9.2. Electronic plans .............................................................................................................. 232
9.3. List of abbreviations ....................................................................................................... 237
9.4. List of symbols ............................................................................................................... 239
9.5. List of publications ......................................................................................................... 241
9.6. Oral and poster presentations ......................................................................................... 243
9.7. Curriculum Vitae ............................................................................................................ 245
Introduction
1
1. Introduction
Atmospheric corrosion continuously causes significant damage to modern materials which
results in substantial financial costs to the society [1]. This has lead fundamental research to focus
on understanding corrosion processes and investigate corrosion inhibition mechanisms. However,
a reliable prediction of long-term corrosion performance is still unavailable.
Over the last 30 years many new techniques have been developed to probe corrosion processes.
However, in general, the examination of corrosion is mainly performed using macroscopic
techniques such as the mass loss method, recording polarization curves and electrochemical
impedance spectroscopy [2]. The spatial resolution of these techniques is rather limited. This leads
to the investigation of corrosion processes as uniform phenomenon over the whole surface. On the
other hand, microscopic aspects such as the presence of chemical or crystallographic
inhomogeneities, contaminations, defects and inclusions at the solid / liquid interface were shown
to have significant influence on the corrosion behavior and may remarkably increase the
susceptibility of passive films to localized breakdown which leads to enhanced dissolution rates of
the underlying metal [3].
To investigate the influence of microscopic features on corrosion various electrochemical
techniques with improved spatial resolution have been developed; such as the scanning reference
electrode technique (SRET) [4], the scanning vibrating electrode technique (SVET) [5] and
localized electrochemical impedance spectroscopy (LEIS) [6]. Despite gaining knowledge about
some aspects of corrosion processes, all of these techniques still suffer from insufficient spatial
resolution. This makes it impossible to obtain highly localized information about the initiation of
corrosion sites. As the signal detected in any near field microscopy technique depends on the tip-
to-sample separation, only techniques that can be applied in constant distance modes of operation
can deliver reliable information while not being influenced by topographic features. Two such
techniques are the Scanning Kelvin Probe (SKP) [7] and Scanning Electrochemical Microscopy
(SECM) [8].
The scanning Kelvin probe is a non-contact technique which measures the contact potential
difference (CPD) between an electrically conductive vibrating reference probe (Kelvin probe) and
an electrically conductive sample [7]. Due to its extremely high surface sensitivity this method has
been used in many different areas of modern chemistry to study surface chemistry [9], surface
reactivity [10], semiconductor doping [11], organic semiconductors [12], organic mono-
layers [13], [14], heterogeneous catalysis [15] and corrosion [16 - 18]. In addition, SKP is a
Introduction
2
powerful tool for in situ measurements of CPD at interfaces between organic coatings and metal
sample [19 - 21].
Scanning electrochemical microscopy is a scanning probe technique that records faradaic cur-
rent changes during the movement of a microelectrode across the surface of a sample. Over the
last 20 years many different modes of operation were reported [22 - 24]. The redox-competition
mode of SECM (RC-SECM) enables in situ visualization of the local catalytic activity towards the
oxygen reduction reaction (ORR) [25]. It was successfully applied to visualize the local ORR
activity of objects such as biofuel cell cathodes [26], microstructured metal hexacyanofer-
rates [27] and different metalloporphyrins [28]. More recently, RC-SECM was used to investigate
novel Pt-based catalysts [29] and was adapted to visualize the local electrocatalytic activity of
ORR catalysts in concentrated HCl solution [30].
Pure untreated Al is not sufficiently strong for most engineering applications. Therefore, it
requires strengthening with different pre-treatment procedures such as alloying to produce high-
performance materials. For example, commercial Al alloys of the 2XXX family are alloyed with
Cu and Mg. They are known to be heat treatable and gain significant strengthening by precipi-
tation hardening. This increases the solubility of the species with both increasing temperature and
time [31]. However, this strengthening process leads to precipitates, so-called intermetallic
particles (IMPs), which increase the susceptibility of the system to corrosion, especially in envi-
ronments containing chloride ions. The electronic conductivity of the oxide which covers most
IMPs is a key factor in the corrosion behavior of Al alloys, as the cathodic activity on the Al
matrix is nearly completely suppressed by the insulating character of the surface that covers the
alumina layer. To optimize the corrosion protection of these alloys a better understanding of the
role and interaction of localized galvanic activities is necessary. However, due to the small size of
such IMPs a direct study by in-situ electrochemical methods is difficult. As the potentials
measured with a Kelvin probe are of electrochemical nature [32], the Scanning Kelvin probe
technique is a promising tool to obtain high-resolution maps of the distribution of possible
galvanic elements. On the other hand, localized investigation of the oxygen reduction reaction on
IMPs and the subsequent effect on the surrounding matrix can provide an insight into the
fundamental corrosion mechanism of this type of Al alloys.
For this reason, a dedicated approach that combines SKP, on the one hand, with an in-situ
high-resolution electrochemical technique such as SECM, on the other hand, to localize and
subsequently spatially resolve localized faradaic processes during the active corrosion event
would be highly desirable. Hence, the sequential deployment of SKP and SECM is a logical
approach to gain an increased knowledge of the underlying mechanisms of corrosion processes.
State of the art
3
2. State of the art
In this chapter, the theoretical background, as well as some applications of the techniques used
in this work will be presented and discussed in detail. The main focus of the discussion is on
Scanning Kelvin Probe and Scanning Electrochemical Microscopy as both of them were
integrated into the developed SKP-SECM system.
2.1. Scanning Kelvin Probe (SKP)
2.1.1. The Work Function
The work function (Φ) is defined as the minimum work needed to extract electrons from the
Fermi level of a solid carrying no net charge [33]. It is assumed that the electron is removed to a
position just outside the sample (end position) where the interaction between the solid and the
electron is no longer existent, i.e. the contribution from the image forces is eliminated.
Additionally, the dimension of the surface is large compared to the distance between the surface
of the solid and the end point.
Three main factors are contributing to the work function. The first one is the structure of the
solid. The exchange energies of electrons at the Fermi level inside the solid determine the bulk
contribution to the main work function. The second one is the work required for the electron to
pass the barrier at the surface of the solid which is the dipole layer in the case of conductors and
the band bending in the case of semiconductors. The final one is the image force acting outside
the surface of the solid. Since the electrons inside the bulk of any conductor can move freely and
there are always enough free electrons present on the surface of the solid, the work function is
mainly a “surface property”. In conductors, the work function is primarily defined by the work
needed to overcome the dipole layer at the surface and to bring the electron to the end point. On
the other hand, in semiconductors, the position of the Fermi level is between the valence and
conduction bands. Thus, additional work is required to move the electron from the Fermi level to
the conduction band.
A more common description of the work function is its presentation as the sum of the chemical
work to transfer the electron from just outside the sample (bulk level) to the Fermi level (µe,
chemical potential) and the electrostatic work to move an electron trough the dipole layer (χ,
dipole or surface potential). Originally, the chemical potential is positive as it describes the work
required to transfer an electron from a point at infinity to the sample. Thus, for the description of
the work function the minus sign should be used since definition of work function is based on the
State of the art
4
transfer of the electron from the Fermi level to the position just outside the sample, i.e. the other
way round (Eq. 2.1). The surface potential describes the potential drop between a point just inside
the bulk and one just outside the sample.
μ (2.1)
Equation 2.1. Description of the work function as the sum of the chemical potential (µe) and the surface potential (χ).
As the bulk is positively charged because of the electron cloud over the surface, usually, the
surface potential is also positive (Fig. 2.1). Factors such as the crystal face orientation and the step
density can significantly influence the surface potential.
Figure 2.1. Schematic representation of the work function in the case of uncharged and charged samples (Φ – the work function, µe – the chemical potential, χ – the surface potential).
In the case of an uncharged sample, the energy levels are equivalent in their energy to the vacuum
level and the work function is equal to the difference in energy between the vacuum and Fermi
levels. However, if the sample is charged, two additional energy levels should be taken into
consideration, the so-called “just inside the bulk” and the absolute vacuum level. In this case,
additional work (eΨ, where Ψ is the Volta potential or the outer potential) should be done to
transfer the electron from the vacuum level to the absolute vacuum level (infinitely far away from
the surface of the sample). The sum of the Volta potential and the surface potential give the
Galvani potential (φ) which is also called the inner potential, i.e. the potential drop between the
absolute vacuum level infinitely far away from the surface of the sample and the bulk level.
The main methods for measurement of the work function can be divided into direct methods
and indirect methods (Chapter 2.1.2). Direct methods rely on the measurement of the current as a
State of the art
5
consequence of the emission of electrons from the surface of the solid caused by heating
(thermionic method [34], [35]), light illumination (photoelectric method [36]) or electric field
(field emission method [37]). However, all direct methods are mainly applicable to conductors
while the interpretation of measurement results in the case of semiconductors provides no
meaningful values of the work function. Additionally, such requirements as the uniformity of the
sample surface and negligible contribution of the applied field should always be fulfilled. It
should be noted that values of the work function determined by different methods are not the same
and can be influenced by the experimental conditions. Thus, it is recommended to calculate the
main value of the work function instead of direct comparison data obtained using different
methods.
2.1.2. The Contact Potential Difference (CPD)
Connection of two solids with different work functions ΦA and ΦB will lead to a flow of
electrons until a uniform redistribution of electrons at the Fermi levels Eand E
of both solids is
reached (Fig. 2.2). As the work functions of both solids are different, the equilibration of Fermi
levels will cause charging of one solid relative to the other. This will lead to the observation of a
Volta potential difference ∆Ψwhich is equivalent to the difference in work functions of the
two solids. As the work function of the solid is an intrinsic parameter, it remains constant after
connection of two different solids while the electrochemical potential of the electrons within both
phases become identical.
Figure 2.2. Schematic representation of the relationship between work function (Φ) and contact potential difference (∆Ψ
).
State of the art
6
The determination of the contact potential difference (CPD) is an indirect method to measure
the work function of a material. If the work function of one of the solids (reference) is known,
then the determination of the contact potential difference can provide the value of the work
function of the sample (Fig. 2.2).
Various methods have been reported for measurement of the CPD and they include:
• the Kelvin method [38 - 40]
• the static capacitor method [41]
• Oatley´s magnetron method [42]
• the breakdown field method [43]
• the saturated diode method [44]
• Anderson´s electron beam method [45]
• the space-charge-limited diode method [46]
Among those, the Kelvin method is one of the most commonly used. Further discussion will be
solely focused on the theoretical background of the Kevin method and its technical realization in
modern conventional scanning Kelvin probe (SKP) set-ups.
2.1.3. The scanning Kelvin probe (SKP)
The SKP is a non-contact, non-destructive vibrating capacitor technique, which measures the
contact potential difference between an electrically conductive vibrating reference probe, also
called the Kelvin probe, and an electrically conductive sample [39]. The main principles and some
fundamental aspects of the SKP have been discussed in detailed reviews [47], [48]. A
conventional SKP consists of a plane-ended cylindrical electrode vibrating perpendicular to a
stationary sample with a frequency ω (Fig. 2.3). In this way, both electrodes form a planar
capacitor. If an external electrical contact between the two capacitor plates is formed, their Fermi
levels start to equalize and the resulting charge flow causes a potential gradient (UCPD). Periodic
modulation of the distance between the electrodes, d = d0 + ∆d*sin(ωt), leads to changes in the
capacitance, (dC / dt), thereby causing a displacement current or so-called Kelvin current (IKelvin)
to flow through the external circuit (Eq. 2.2). In the conventional nulling technique, a variable
backing potential (U0) is adjusted until the Kelvin current vanishes. When U0 is equal to -UCPD,
the electric field between the capacitor plates is compensated and a zero output signal is recorded
State of the art
7
(Eq. 2.3). In this case, the Volta potential difference between the Kelvin probe and the sample is
equal to the applied compensation voltage.
Figure 2.3. Schematic representation of the set-up for the measurement of the contact potential difference (∆Ψ
– contact potential difference between the Kelvin probe and the sample, d0 - main tip-to-sample distance, ∆d - tip oscillation amplitude,
U0 - variable backing potential, IAC – Kelvin current).
∆ ∗ ∗ ∗ ∗ ∆ ∗
∆ (2.2)
Equation 2.2. Kelvin equation (IKelvin – Kelvin current, ∆Ψ – contact potential difference
between the Kelvin probe and the sample, ε - dielectric constant of the medium, ε0 - electric field constant, A - active tip area, ∆d - tip oscillation amplitude, ω - tip oscillation
frequency, d0 - main tip-to-sample distance).
∆ ∗
, ∆
(2.3)
Equation 2.3. Compensation of the Kelvin current with an external backing potential (IKelvin – Kelvin current, ∆Ψ
– contact potential difference between the Kelvin probe and the sample, U0 - variable backing potential, dC / dt – changing of the capacitance between
the vibrating Kelvin probe and the sample).
The accuracy of the determination of the contact potential difference using the scanning Kelvin
probe is determined by its sensitivity S (Eq. 2.4).
State of the art
8
≈ ∗∗∗∆∗
(2.4)
Equation 2.4. The sensitivity (S) of the measurement of the contact potential
difference using the scanning Kelvin probe.
The sensitivity of the SKP set-up is directly proportional to the active area of the Kelvin probe (A)
and reciprocal to the tip-to-sample distance. Thus, to achieve high lateral resolution, tips with
small active area should be used while keeping the tip-to-sample distance at very small values.
2.1.4. The absolute electrode potential (Eabs)
The physical meaning and the concept of an “absolute electrode potential” was developed and
elaborated in detail by Bockris [49 - 51] and Gileady [52] in the 1970s. However, the modern
concept of an absolute electrode potential was suggested by Trasatti [53 - 55]. If a sample is
covered with a liquid layer of an electrolyte, the absolute electrode potential can be defined as the
minimum work required to transfer the electron from the Fermi level of the sample through the
solid / liquid interface, through the liquid layer and through the surface layer of the liquid to a
position just outside the liquid [55]. In this way, the definition of the absolute electrode potential
is very similar to that of the work function. If the Kelvin probe is connected to a sample covered
with a thin layer of an electrolyte, then the determination of the contact potential difference can
provide the absolute potential of the immersed sample (Eq. 2.5).
=
+ ( − ) (2.5)
Equation 2.5. Description of the absolute electrode potential (EAbs) based on the determination
of the Volta potential difference (ΨKP - ΨSample) using the scanning Kelvin probe (ΦKP – work function of the Kelvin probe).
One of the requirements for the determination of the absolute potential of the sample is that the
absolute value of the work function of the Kelvin probe is known. However, this it is not always
the case. Instead of the absolute value of the work function, the Kelvin probe can be calibrated vs.
a reference electrode, e.g. standard hydrogen electrode (SHE). In this way, measurement of the
contact potential difference can provide the value of the absolute electrode potential of the sample
State of the art
9
referenced to a given reference electrode which can also be described as the work function of the
sample with regard to the electrode potential scale.
It is important to note that the SKP does not measure the distribution of the Volta potential, as
reported in many publications, but it provides the distribution of the Volta potential difference or
the contact potential difference between the reference electrode (Kelvin probe) and the sample. If
the absolute value of the work function of the Kelvin probe is known or the Kelvin probe was
calibrated against a reference electrode, then the SKP can provide a map of the local distribution
of the work function all over the sample.
2.1.5. Practical aspects of application of the SKP
2.1.5.1. Calibration of the system
The potential of a corroding material immersed in an electrolyte is commonly designated in
corrosion science as the corrosion potential (Ecorr). This term is also used for the open circuit
potential of a passive surface when the surface remains intact or only partially corroding.
In practice, the electrode potential (Ecorr) measured using the SKP technique is referenced
against a standard reference electrode (Eq. 2.6):
= ∆ +
−
= ∆ + (2.6)
Equation 2.6. Description of the corrosion potential (Ecorr) based on measurement of the Volta potential difference (ΨKP - ΨSample) using the scanning Kelvin probe (ΦKP – work function of the Kelvin
probe, E - potential of the reference electrode on the absolute electrode potential scale).
Since the absolute value of the work function of the Kelvin probe is not always known, one can
still obtain the corrosion potential of the sample if the Kelvin probe was calibrated against a given
standard reference electrode. The calibration of the Kelvin probe can be done by direct calibration
with a chosen reference electrode. For this, the Kelvin probe needs to be positioned over the
surface of an electrolyte droplet (1 M KCl) placed on the surface of a conductive sample which is
connected to a potentiostat while the micro reference electrode and a small counter electrode are
dipped in the electrolyte droplet. The potential of the sample is adjusted to different potentials
while the corresponding CPD is measured using the SKP. A plot of the measured CPD values as a
function of the potential applied to the sample provides a linear relation, while the crossing point
with the ordinate axis gives the value of the constant for equation 2.6. An alternative way to
State of the art
10
calibrate the Kelvin probe is to measure the CPD over a metal electrode which is exposed to an
electrolyte containing cations of the metal at a predefined concentration (reference system with
known corrosion potential). One of the reference systems widely used in corrosion science is a
drop of saturated CuSO4 solution placed in a small Cu pot [56]. However, as the calibration
constant contains the value of the work function of the Kelvin probe, it strongly depends on the
properties of the Kelvin probe. Thus, it is recommended to perform the calibration of the Kelvin
probe every time the Kelvin probe is exchanged or after contact between the tip and the sample
surface.
2.1.5.2. Sample preparation
Investigation of the uncovered metallic sample does not require any special sample preparation
procedures. Little precaution mainly focused on removing any contaminants and dust particles
from the sample surface are necessary. However, if the thin layer of electrolyte is placed on the
sample surface, corrosion conditions during measurement have to be well defined (clean
electrolyte solution with known concentration etc.).
In the case of samples coated with a thin polymer layer, the polymer surface must not be
electrically charged as this will lead to huge potential fluctuations during mapping of the Volta
potential differences. To avoid charging of the polymer surface or to discharge the sample surface,
saturation of the atmosphere with almost 100% relative air humidity is widely used. To obtain
reliable results, long-term exposure of the sample to such conditions prior to measurement of the
CPD is often required.
Investigation of delamination processes of organic coatings in the presence of coating defects is
one of the main applications of the SKP technique. To obtain reliable results very strict require-
ments such as absence of any galvanic coupling between the coated surface and the defect should
be fulfilled. This is possible if the surface resistance is much larger than the resistance of the
polymer perpendicular to the surface especially at high humidity. Usually, such conditions can be
reached in the case of extremely clean polymer surfaces.
2.1.5.3. Tip preparation
One of the most important requirements of the Kelvin probe used in corrosion science is the
stability of the work function of the tip over long periods of time. The most widely used material
for preparation of reference electrodes for conventional SKP measurements are Cr / Ni alloys.
State of the art
11
Usually, 500 µm wires are used for direct fabrication of tips, while electrochemical etching is
necessary if higher lateral resolution is required. Additionally, great care has to be taken to obtain
a planar surface at the tip as this strongly influences the performance of the SKP set-up. Conically
etched tips are characterized by low lateral resolution and strong stray capacitance effects. Thus,
to obtain a flat surface at the top of the etched Cr / Ni wire, the wire is embedded into a resin
directly after electrochemical etching and subsequently polished to a planar and smooth finish.
Following this approach reference tips with Dout of about 100 µm could be obtained.
2.1.5.4. Lateral resolution
The lateral resolution of the conventional SKP set-up is strongly dependent on the shape of the
Kelvin probe and the mean distance between the reference electrode and the sample surface. It is
very important to ensure planarity at the front surface of the tip. To minimize stray capacitance
effects, the side walls of the tip should be perpendicular to the sample surface leading to an
optimal shape of the Kelvin probe being in the form of a cylinder. To additionally lower the effect
of stray capacitances, the side walls of the tip can be electrically insulated, e.g. using a thin layer
of epoxy resin. The thin layer of conductive material subsequently deposited on top of the
insulating layer should be grounded. A very detailed evaluation of factors influencing the stray
capacitance can be found in [57]. It is important to keep the vibration amplitude of the Kelvin
probe rather small and the tip-to-sample distance as small as possible. To significantly improve
the signal-to-noise ratio while simultaneously avoiding excessive distortion, Johnson and co-
workers [58] proposed to use a ratio between the main tip-to-sample distance (d0) and the tip
oscillation amplitude (∆d) of about 1:3. To ensure a homogeneous distribution of the field bet-
ween the Kelvin probe and the sample surface, Baumgärtner and Liess [59] suggested to operate
the SKP system with the ratio between the outer needle diameter (Dout) and d0 of at least 5:1.
Under these conditions the assumed model of parallel plate capacitor geometry is still valid.
A very comprehensive investigation of the ability of the SKP to resolve lateral differences in
potential of the specimen surface was reported by McMurray and Williams [60]. The authors used
electrostatic calculations to obtain the maximum lateral resolution which is possible with a “point
probe'' with negligible small dimensions. The theoretical calculations were analyzed as a function
of the tip-to-sample distance. Additionally, the lateral resolution of a conventional SKP system
was experimentally investigated by scanning across a sharp edge between two continuous
coplanar areas of dissimilar metals at varying heights from 90 µm to 240 µm.
State of the art
12
Based on electrostatic calculations, McMurray and Williams reported that in the case of a point
probe vibrating with an infinitely small amplitude orthogonal to the sample surface, one half of
the Kelvin current signal originates from a circular sample area with a diameter 1.532*d0 (d0 is the
mean tip-to-sample distance) placed exactly below the tip. If a vibrating point probe is “scanned”
across a linear edge separating two dissimilar metals, the lateral response width (∆L50)
corresponding to a change of the measured Volta potential difference between 25% and 75% of
the overall change in Kelvin potential was calculated to be 0.884*d0. In the case of a sample with
more complex geometry like a circular area of diameter Ds with a Volta potential V1 surrounded
by an infinitely large area with a Volta potential V2, underestimation of the measured value of
V1-V2 is highly probable. A Kelvin probe positioned exactly above a given circular area will
underestimate the quantity of V1-V2 by 50% when d0 = 0.65*Ds. Further increase of the d0/Ds
ratio will lead to much faster increase of the degree of underestimation. Additionally, as the model
used for electrostatic calculations was based on a point probe with negligible small size, the
reported dependence should be seen as estimation of minimum errors. During real SKP
measurements effects such as the stray capacitance and the non-ideal geometry of the reference
electrode will lead to additional lateral spreading of the electric field influencing the performance
of the SKP system and increasing the degree of underestimation of V1-V2.
Experimental evaluation of the lateral resolution of a conventional SKP system was performed
using cylindrical plane-ended gold probes of varying diameters from 125 µm to 250 µm. As a test
sample, a polished plate of pure Zn partially covered with a 5 nm thin layer of gold was used. The
experimental data obtained during scanning across the linear edge between Zn and Au was fitted
using a semi empirical relationship according to ∆L50 = 0.884*d0 + 0.4*D for d0 < 140 µm (D
describes the area of overlap between the metal regions and the plane-circular Kelvin probe). For
d0 larger than 140 µm, the ∆L50 tends towards ~1.6 which is probably caused by increased
influence of electrostatic contributions from the Kelvin probe sides. Thus, one should always keep
in mind that the distribution of the Volta potential differences resolved by SKP as well as the
dimensions of the visualized inhomogeneous areas on the sample surface are influenced by the
physical size of the Kelvin probe and the tip-to-sample distance.
2.1.6. Scanning Kelvin Probe Force Microscopy (SKPFM)
Since its introduction in corrosion science by Stratmann et al. [16] for the study of samples
covered by an ultra-thin layer of electrolyte, SKP has been increasingly used. However, because
of the comparatively low lateral resolution and the considerable effort for reliable operation, it has
State of the art
13
not found very broad application yet. Integration of the SKP into an AFM leading to scanning
Kelvin probe force microscopy (SKPFM) provides a resolution in the nm range and can thus be
used for resolving the CPD of intermetallic compounds in different alloys whose dimensions are
too small to be resolved by the conventional SKP set-up.
The principle and technical realization of the SKPFM are discussed and described in detail in
Nonnemacher et al. [61], [62] and also in Jacobs et al. [63], [64]. The SKPFM is a two-pass
technique based on a line scan in the AFM tapping mode to obtain information about the
topography of the sample. Using the recorded topographic information the metal-coated or doped
silicon cantilever is lifted 50 nm to 100 nm and the line is rescanned in the so-called ”lift mode”.
During the rescan, the tapping mode of the AFM operation is switched off and an alternating
voltage (Uac*sin (ωt)) is applied to the cantilever. As the cantilever and the sample are electrically
connected and their Fermi levels equilibrated, both of them carry a charge which in combination
with an applied alternating voltage causes oscillations of the cantilever. Thus, the capacitance of
the capacitor formed by the cantilever and the sample changes periodically and the Volta potential
difference between them can be measured using the same compensation principle as described for
SKP. Application of an external compensation voltage leads to changes in the magnitude of the
cantilever oscillations which is controlled by the AFM detection scheme. Thus, nulling of the
induced charge will cancel the oscillation of the cantilever and the value of the applied
compensation voltage will provide the Volta potential difference.
Usually, the Kelvin probe used in conventional SKPFM set-ups is an AFM cantilever covered
with a thin metal layer. Opposite to the metal wires used as the Kelvin probes in conventional
SKP set-ups, such a design of the Kelvin probe leads to a strong dependence of the measured
Volta potential difference on the tip-to-sample distance. Since it is not possible to separate the
contribution of the cantilever itself from the contribution of the tip alone, the visualized
distribution of the Volta potential differences on a sample surface is not always representative of
the real potential differences. The effect of stray capacitances is especially pronounced in the case
of samples with remarkably inhomogeneous topographic and compositional features leading to the
measured Kelvin signal being strongly influenced by the contribution of the cantilever. For
instance, as it was reported by Blucher and co-workers [65], the potential difference between Al
inclusions and the Mg matrix obtained with SKPFM for model Mg alloys was 0.07 V, while the
Volta potential difference measured using the SKP was found to be about 0.6 V. A similar
observation was reported by Rohwerder and co-workers [66]. The authors found that the Volta
potential difference observed between delaminated and intact areas on the sample surface derived
with SKPFM was about 4 - 5 times smaller than the values measured by SKP. In some cases,
State of the art
14
depending on the type of the used cantilever, even an inversion of the contrast was
observed [67], [68].
The influence of stray capacitances is not the only single effect which influences the Volta
potential difference resolved by SKPFM. Topographic artifacts, particularly those in the form of
steps, are responsible for faulty contrast which leads to wrong values of the resolved Volta
potential differences. The main reason for the appearance of these topographic artifacts is the
effect of the gradient of the cantilever contribution which is especially large in the presence of
protrusions on the sample surface. If the cantilever is placed above one of these protrusions, the
contribution of the tip to the overall value of the measured CPD is much smaller than that of the
cantilever itself as the cantilever will “sense” not only the sample surface but also the side walls of
the protrusion. One of the possible solutions is to polish the sample to obtain as flat as possible
surface with minimal mechanical defects. However, in the case of samples from corrosion
experiments, which are the main objects for SKPFM measurements, corrosion of the sample
sometimes leads to deposition of the corrosion products and / or significant topographic changes
on the sample surface.
The size effect is an additional effect characteristic for both SKP and SKPFM techniques.
McMurray and co-workers [60] reported theoretical calculations and results of experimental
evolution of the distance over the sample required to obtain full contrast (distribution of the Volta
potential difference) between two materials with different work functions. The full contrast could
be achieved over a distance which was about a factor of 5 larger than the tip-to-sample distance,
even in the case of an ideal point tip. Thus, taking into consideration a distance of 100 nm which
is characteristic for SKPFM experiments, it becomes clear that the full contrast would to be
expected for features larger than 500 nm. However, as the effect of stray capacitance is always
present, the size of the features on the sample surface necessary to obtain a reliable distribution of
the Volta potential difference is even larger than 500 nm (up to sub-µm range). This effect
becomes very pronounced when the relative difference in the work functions of the features is
considerably small.
One of the most significant differences between SKP and SKPFM is that in the case of the
SKPFM the space between the tip and the sample is not free of any electric field. Application of
an alternating voltage needed for proper operation of SKPFM generates a very strong electric field
in a range of about 5 V per 10 nm distance. This electric field might have a significant influence
on the measurement of the Volta potential distribution for semiconducting samples. SKPFM
experiments performed by Sommerhalter and co-workers [69] on p-type WSe2 samples revealed
an increase of the measured Volta potential difference with decreasing alternating voltage. The
State of the art
15
observed effect was attributed to the band bending induced by the electric field. On the other
hand, experiments performed on highly ordered pyrolytic graphite (HOPG) did not reveal the
same effect thus confirming that it occurs only on semiconducting samples. The existence of such
an effect is important to keep in mind, as surfaces of most metals and alloys are covered with
semiconducting oxide layers. This effect can play a significant role in the case of SKPFM
experiments on substantially corroded samples since an oxide layer of a certain thickness can be
present on the metal surface.
2.1.7. Application of SKP
The scanning Kelvin probe is often applied as a standard surface analysis technique due to its
extremely high surface sensitivity. It has been widely used for the study of semiconductor
doping [70], organic monolayers [71], corrosion processes [72] and for in situ measurements of
Volta potential differences at the interface between an organic coating and a metal sample [73],
[74]. A literature survey focused on the application of the SKP in corrosion science revealed three
main topics: determination of corrosion potentials (Chapter 2.1.7.1), investigation of atmospheric
corrosion (Chapter 2.1.7.2) and delamination experiments on samples covered with a layer of an
organic coating (Chapter 2.1.7.3).
2.1.7.1. Determination of corrosion potential
As already discussed in chapter 2.1.5.1, the corrosion potential (Ecorr) of a dry sample can be
directly measured using the SKP technique if the Kelvin probe is calibrated against a suitable
reference electrode or against any reference system with well-known corrosion potential prior to
the determination of the Volta potential difference. If the sample surface is covered with a thin
layer of an electrolyte, the measured Volta potential difference will be modified by an additional
potential drop across the interface between the metal and the electrolyte (Eq. 2.7):
∆ =
∗ ( + ∆
− ) (2.7)
Equation 2.7. Volta potential difference over a metal surface covered with a thin layer of an electrolyte measured using the SKP technique (∆φ
is the Galvani potential difference between the metal and the electrolyte).
State of the art
16
If the metal surface is covered with a thick electrolyte layer or a layer of low resistance, SKP
will provide an electrode potential similar to that of a conventional reference electrode (open
circuit potential). Under such conditions, the potential coupling between different surface sites
will lead to the detection of an average electrode potential analogous to any other reference
electrode. On the other hand, if the metal surface is covered with an insulating layer of an oxide or
an organic coating, the Galvani potential difference between the metal surface and the electrolyte
(∆φ ) is composed of three terms (Eq. 2.8).
∆ = ∆
!" + ∆!" + ∆!" (2.8)
Equation 2.8. Galvani potential difference between the metal surface covered with an insulating layer of an oxide or an organic coating and the electrolyte.
The presence of an insulating layer requires taking into account potential drops occurring at both
interfaces and across the insulating phase itself while the correlation between the measured Volta
potential difference, Galvani potential difference and the difference in work functions described in
equation 2.7 remains the same.
Deposition of a thin layer of an organic coating on the metal surface in combination with a
defect in a given coating covered by a thin electrolyte layer is a common test system for the
investigation of the delamination process of the coating. Such a sample will have areas which are
already delaminated, delaminating and still intact interface between the coating and the metal
surface. Scanning across these areas will reveal substantial differences in the measured Volta
potential differences and in the corresponding corrosion potentials if the Kelvin probe was
calibrated against a reference system. The Volta potential difference measured over the defect site
covered with a thin electrolyte layer can be described using equation 2.7 as discussed earlier.
However, over the delaminated area, additional interfaces between the thin electrolyte layer
directly on the sample surface and a delaminated organic coating are present. Thus, the value of
the Volta potential difference as measured by SKP will be modified by a dipole layer at the
sample / electrolyte interface (∆φ ), the potential drop at the electrolyte / coating interface
(∆φ
) and the surface potential of the organic coating (χC) (Eq. 2.9). The potential drop at
the electrolyte / coating interface is determined by the so-called Donnan potential (∆).
State of the art
17
∆ =
∗ (− + ∆
− ∆# + ) (2.9)
Equation 2.9. Volta potential difference measured over a delaminated coating area on the metal surface using SKP (∆$ is the Donnan potential, χC is the
surface potential of the organic coating).
Equation 2.10 describes the Donnan potential which depends on both the concentration of ions
in the electrolyte and the concentration of fixed charged groups in the coating [75].
∆ =
∗
+
≈
∗ ≪ ≈
∗ ≫ (2.10)
Equation 2.10. Donnan potential (∆$- Donnan potential, z - valence of ions in a given electrolyte, R – universal gas constant, T – temperature of the test system, F – Faraday constant,
x – ratio between the fixed charge carrier density in the organic coating and the electrolyte concentration).
Application of SKP for measurement of the corrosion potential on a metal sample with an
organic coating placed on the surface of a thin electrolyte layer located on top of the sample
surface is usually performed under potential control. For this, the electrode potential of the sample
is controlled by a potentiostat using a standard reference electrode while measurement of the
Volta potential difference is performed by the conventional SKP set-up. This approach is very
convenient for the investigation of the susceptibly of a given organic coating to develop the
Donnan potential indicating the presence of a certain amount of fixed charges. Leng and co-
workers [56] applied SKP to investigate the properties of an epoxy ester containing –OH, -COOH,
-CONH and –NH functional groups to induce a significant Donnan potential. It was found that in
the case of an intact polymer film covering the metal sample placed in an electrolyte of pH = 9,
the value of the Donnan potential was very small indicating much larger concentration of the
electrolyte than the density of fixed charge carriers (x<<1). On the other hand, if the delamination
process around the defect site in the coating proceeded for a while, the local pH became very
alkaline (up to pH = 14) as a result of the oxygen reduction reaction. Under these conditions, the
oxidative attack of OH* radicals caused degradation of the coating and formation of charged
functional groups such as -COO-. Thus, if the concentration of the electrolyte is rather low, a
Donnan potential of up to 80 mV was observed indicating a large increase in the fixed charge
carriers [56]. However, no substantial contribution of the Donnan potential to the measured value
of the corrosion potential was reported for commonly used coating systems. As the surface
State of the art
18
potential of a polymer is usually about 50 mV [76], measurement of the Volta potential difference
between the coating layer and the Kelvin probe will provide the corrosion potential at the buried
interface.
2.1.7.2. Atmospheric corrosion
To investigate the electrochemical mechanism of atmospheric corrosion of iron and steel under
wet / dry conditions Stratmann and Streckel [16], [77], [78] developed the so-called Kelvin probe
potentiostat. The authors placed the Kelvin probe in a closed volume and the partial pressure of O2
was measured with respect to a reference chamber where the oxygen pressure was kept constant.
The oxygen reduction rate was calculated based on changes of the partial pressure of O2.
Determination of the corrosion potential by the Kelvin probe was performed simultaneously with
measurement of the rate of metal dissolution using a magnetic balance technique. During the
cyclic corrosion experiments clear differentiation between wetting of the sample surface, wet
surface and drying was possible based on the metal dissolution rates. The highest metal
dissolution rates of iron were observed during drying of the metal surface after wetting. It was
found that anodic dissolution of iron is compensated by the cathodic reduction of oxygen.
Interestingly, the metal dissolution during the first drying cycle is significantly larger than the
corresponding oxygen reduction. This effect was attributed to reduction of iron oxide taking place
parallel to oxygen reduction [16].
Figure 2.4 shows calculated corrosion rates of a pure iron sample and of a model alloy (Fe-
3.4 % Cu), both contaminated with 1 g / cm2 of SO2, plotted against the corrosion potential
measured by Kelvin probe during a few wet / dry cycles [77]. Combination of the metal corrosion
rates with corresponding corrosion potentials provides “polarization curves” which describe the
electrochemical corrosion process during wet / dry corrosion cycles. The corrosion potential of the
wet iron surface has much lower values than that of the dry surface. Drying of the metal surface
leads to a steep increase of the corrosion rate which reaches a maximum followed by rapid decay
to almost zero. This effect is denoted as the active / passive transition. The relative difference
between corrosion potentials measured for wet and dry surfaces remains constant over many
wet / dry cycles while the absolute values of the corrosion potential are shifted in positive
direction. Every additional wet / dry cycle leads to reduction of the corrosion rate since every
drying process leads to an anodic shift of the electrode potential. However, in the case of the Fe-
3% Cu model alloy, disappearance of the active / passive transition was observed with increased
number of wet / dry cycles. Thus, the lack of activation during wetting is the main reason for the
State of the art
19
A B
observed superior corrosion resistance of this model alloy as the active / passive transition
characterized with the highest corrosion rates is not present.
Figure 2.4. Corrosion rates of pure iron sample (A) and Fe-3 % Cu model alloy (B) both contaminated with
1 g / cm2 of SO2, plotted as a function of the corrosion potentials measured by the Kelvin probe potentiostat during a few wet / dry transitions [from 77].
Investigation of corrosion inhibition during atmospheric corrosion of iron using the Kelvin
probe potentiostat was reported by Leng and Stratmann [79] in the early 1990s. Ammonium
benzoate was used as the corrosion inhibitor and the influence of the inhibitor concentration and
thickness of the electrolyte layer were investigated in detail. It was found that the dissolution rate
of the passive surface depends on the concentration of the inhibitor within the thin electrolyte
layer. The boundary concentration of the inhibitor was deduced from the comparison of the rate of
the anodic (surface dissolution) and cathodic (oxygen reduction) partial reactions. At
concentrations of the inhibitor below 0.025 M, the dissolution rate (passive current density) was
larger than the rate of the oxygen reduction reaction leading to activation of the surface, while at
concentrations higher than this critical level, the reduction of oxygen was faster than the
dissolution of the passive metal. Therefore, the surface remained passive.
To investigate the influence of the electrolyte layer thickness on the corrosion behavior of iron,
the rate of oxygen reduction and the metal dissolution rate were monitored as a function of the
electrolyte layer thickness. It was found that in bulk electrolytes, the oxygen reduction reaction is
not influenced by the inhibitor and a metastable passive surface is the result of the slow transport
of oxygen to the metal / electrolyte interface. However, the diffusion limited rate of oxygen
reduction depends on the electrolyte layer thickness and for very thin layers very high rates of
oxygen reduction were observed. An increased rate of the cathodic partial reaction should lead to
State of the art
20
anodic shift of the corrosion potential and to the corresponding increase of the corrosion rates.
However, it was found that, if a sufficient amount of the inhibitor is present in the electrolyte, the
rate of the cathodic oxygen reduction increased with decreasing thickness of the electrolyte,
whereas the rate of dissolution of the metal was drastically decreased. In the case of a 10 µm thin
electrolyte layer, no anodic currents were observed and the surface remained passive. The
influence of the layer thickness is only due to the transport of oxygen to the metal / electrolyte
interface but not due to the kinetics of the charge transfer reaction. Thus, a decrease of the film
thickness leads to an increase of the diffusion limited current density. However, the cathodic
current in the potential range of iron passivation showed a rather small dependence on the
thickness of the electrolyte layer which was predominantly controlled by the rate of charge
transfer. On the other hand, the anodic current density strongly depends on the electrolyte
thickness leading to a decrease in the iron corrosion rates with decreasing electrolyte thickness.
Additionally, a decrease of the electrolyte thickness leads to an increase of the inhibitor
concentration. Thus, ammonium benzoate offers very good performance as a vapor-phase
inhibitor if a very thin layer of the electrolyte or even only few electrolyte layers are present on
the metal surface.
2.1.7.3. Delamination experiments
Protection of reactive metals against corrosion by organic coatings and lacquers is a widely
used practice. Many examples could be found such as the protection of car bodies against
atmospheric corrosion, pipelines against corrosion in humid ground, ships against corrosive attack
in sea water [80]. Originally, it was believed that organic coatings offer a good corrosion
protection because of their barrier properties limiting the penetration of water and oxygen to the
metal / polymer interface [81]. However, many of the very efficient protective coatings are highly
permeable for water and oxygen which is contradictory to the original assumption. Thus, not the
barrier properties of organic coatings are responsible for corrosion protection, but the
electrochemical conditions at the metal / polymer interface originating from the properties of the
organic coating as shown by Stratmann and Feser [82]. Most probably, formation of an extended
diffuse double layer takes place. The presence of defects (holes, pins etc.) can significantly lower
the barrier properties of the coating leading to formation of galvanic elements which determine
the delamination rate of a given coating [83]. Usually, the defect site acts as a local anode (metal
oxidation) while the oxygen reduction reaction takes place at the delamination frontier (local
cathode). This process is referred as cathodic delamination. On the other hand, if the delamination
State of the art
21
frontier acts as the local anode while some areas on the sample surface are responsible for oxygen
reduction, as it is the case for many Al-based alloys, then this process is called anodic
delamination.
Detailed investigation of the cathodic delamination process was reported by Leng and co-
workers [56]. If the metal surface is exposed to the electrolyte and oxygen is present, the metal is
oxidized (anodic process) while oxygen is reduced at the sample surface (cathodic process). This
leads to the adjustment of the electrode potential to a value at which the sum of the anodic and
cathodic currents equals zero and the overall corrosion rate is limited by the diffusion of oxygen
through the electrolyte layer to the metal surface. However, if the metal is completely insulated by
an organic coating, the rate of metal oxidation becomes very slow and no electrons are available
for the oxygen reduction reaction. If the coating gets damaged via a defect or a pinhole, the area
of metal exposed to the electrolyte starts to operate as a local anode and electrons can flow to the
intact metal / coating interface, where these electrons can be used for the reduction of oxygen.
Oxygen reduction takes place at the previously intact interface which promotes further dissolution
of the metal at the defect site. The change of the reactivity at the previously intact interface leads
to shifting of the electrode potential in a negative direction. This effect can be nicely visualized
using the SKP technique. To compensate the flow of electrons from the defect site to the intact
interface, the transport of ions alongside the interface transforms exclusively into transport of
cations while the transport of anions becomes inhibited by the electric field. One of the
intermediate products of the oxygen reduction process is the OH* radical. It is a very reactive
group which attacks the metal / coating bonds leading to degradation of the coating, diminishing
of the coating adhesion on the metal surface and creation of the delamination frontier at the
interface [84]. Additionally, formation of the main product of oxygen reduction such as OH-
causes a significant increase of the pH value at the interface. This further promotes degradation of
the coating layer. However, as iron is passive at high pH values, it does not further dissolve at the
interface but only at the defect site itself [85].
Opposite to steel or galvanized steel samples, homogeneous delamination of an organic coating
deposited on the sample surface is not observed on Al alloys. In this case the delamination is
localized and this process is called “filiform corrosion”. Detailed investigation of the filiform
corrosion on the surface of the AA2024-T3 Al alloy was reported by Schmidt and Stratmann [20].
The electrochemical mechanism of filiform corrosion is based on the development of a corrosion
filament which consists of an “active head” and a tail. Low oxygen concentration was found to be
present at the filiform head while direct connection to the ambient air is established by porous
corrosion products at the tail. This effect is known as a differential aeration cell. Application of
State of the art
22
SKP for the investigation of filiform corrosion on AA2024-T3 alloy revealed a more negative
corrosion potential at the front of the filament while a more positive corrosion potential was
observed at the tail. Thus, during filiform corrosion, the head acts as the local anode while the tail
acts as the local cathode, which is opposite to the cathodic delamination process on pure iron
surfaces.
State of the art
23
2.2 Scanning Electrochemical Microscopy (SECM)
Scanning electrochemical microscopy was introduced by Bard et al. [8] as a scanning probe
technique for the characterization of electrochemical processes and structural features at a sample
surface by monitoring changes in the Faradaic current during movement of a Pt or Au disk
ultramicroelectrode (UME) across the surface of a sample. Additionally, movement of the UME,
also called SECM tip, normal to the sample surface can probe the diffusion layer by detecting
sample-generated species at the UME [86]. The current measured at the SECM tip is caused by
electrochemical processes at both the tip and sample. It is controlled by electron transfer kinetics
at both interfaces and the mass transport of electrochemically active species in a given electrolyte.
The SECM tip is polarized at a potential which is sufficient to invoke oxidation or reduction of
electrochemically active species at the electrode surface. The current, measured at the UME is
characterized by a transient where the duration is a function of the tip diameter. The decay of the
transient results in a steady-state current (iT,∞) described by equation 2.11.
%,& = !"#$ (2.11)
Equation 2.11. Steady-state current measured at a SECM tip in a given electrolyte containing an electrochemically active species (n – number of transferred electrons, F – Faraday constant, C – concentration of the active species converted at the SECM tip, D – diffusion coefficient
of the active species, a –diameter of the active area of the SECM tip [87].
Equation 2.11 describes the steady-state current when the SECM tip is positioned far away
from the surface of the sample. However, if the tip approaches the sample surface to a certain tip-
to-sample distance (≈ equal to the radius of the SECM tip) the value of the steady-state current (iT)
starts to change in dependence on the properties of the sample surface. In the case of an
electrochemically active sample iT will be larger than iT,∞ (positive feedback) while on electro-
chemically inactive surfaces iT < iT,∞ (negative feedback). A detailed discussion about these two
effects can be found in chapter 2.2.2.1. A common representation of the recorded data is a
dimensionless plot of iT / iT,∞ as a function of d / rtip (where d is the tip-to-sample distance and rtip
is the radius of the active tip area) or a color-coded two-dimensional plot of iT as a function of the
x- and y-positions of the SECM tip. An alternative way to represent the data is a three-
dimensional plot with iT values plotted as the z-axes while keeping the x- and y-axes the same as
for the 2D plots.
Generally, the SECM set-up consists of a positioning unit that scans an UME as a local probe
in close distance to the sample surface. A typical electrochemical cell for the SECM consists of an
State of the art
24
UME as working electrode, a reference electrode and an auxiliary electrode. By means of a
bipotentiostat, the sample can be connected as a second working electrode and thus polarized at a
user-defined potential.
2.2.1. Practical aspects of the application of SECM
2.2.1.1. General aspects
The choice of the scanning speed has a direct influence on the quality of the obtained SECM
images. A scanning speed that is too fast will lead to operation of the SECM in a non-steady state
regime. This will limit the applicability of the obtained data in a quantitative analysis using the
developed theoretical models and also lower the lateral resolution of the system. On the other
hand, scanning speed that is too low will require measurement times that are too long to be
practical. Thus, to define the optimal scanning speed one should take into account the time needed
for the microelectrode to reach a steady state, and also the size of the tip electrode and the tip-to-
sample distance. The smaller the tip size is and the closer the tip is positioned to the sample
surface the more rapidly the steady state regime will be reached. To estimate the waiting time
after each movement step of the tip required for equilibration of the hemispherical diffusion field
in front of the tip one can use equation 2.12.
= '∗
(#
(2.12)
Equation 2.12. Time required to reach the steady state regime (tss is the needed time,
τ is the RC cell time constant, rtip is the radius of the active tip area, D0 is the diffusion coefficient of the mediator) [88].
Calculation of tss for the commonly used 10 µm SECM tip reveals an estimated waiting time of
0.5 s for [Ru(NH3)6]3+ and 0.6 s for [Fe(CN)6]
3- which are both used as mediators during
experiments in the feedback mode of the SECM.
The choice of a suitable redox mediator for experiments in the feedback mode of SECM should
be made with regard to the chemical nature of the SECM tip and the sample. As the study of
corrosion processes is highly sensitive to strong oxidants and reductants, it is important to select a
redox couple that will not adversely react with the sample.
State of the art
25
2.2.1.2. Fabrication of SECM tips
The size and the nature of the SECM tip are the two main parameters which determine the
spatial resolution of the SECM measurement and the information that can be obtained from it. The
first reported SECM tips [8] were based on ultramicroelectrodes produced by sealing a Pt wire of
the desired radius in a Pyrex glass capillary. The sealed end of capillary was polished with
sandpaper until a disk-shaped surface of Pt was exposed. Subsequently, the exposed surface was
polished using diamond paste with successively decreasing size of diamond particles. The size of
the glass sheath at the very top of the tip was decreased by conical sharpening on emery paper
until the overall diameter of the glass insulation surrounding the Pt disk was less than 100 µm.
Finally, the unsealed end of the Pt wire was electrically connected to a thicker Cu wire using
silver paint. To obtain a more reproducible shape of tips and to avoid the need to sharpen the glass
sheath, conically pulled glass capillaries were used for the fabrication of SECM tips [89]. Prior to
sealing the metal wire into glass capillary, the outer diameter of one side of the capillary was
drastically reduced by melting the glass with a heating coil in combination with simultaneous
pulling of the capillary. Nowadays, Au and Pt SECM tips with diameters of 10 µm to 100 µm are
fabricated following these procedures with some custom modifications.
To obtain microelectrodes with a diameter of the active electrode surface below the diameter of
commercially available metal wires a range of electrochemical etching procedures have been
reported [90 - 92]. Etching of metal wires leads to the formation of protruding conical tips where
the side walls of the wire can be insulated using Apiezon wax [93], electrophoretic paint [94] or
by sealing the etched wire in glass [95]. The successful fabrication of disk-shaped nanoelectrodes
by means of a commercially available laser puller was reported by Shao et al. [96] using
borosilicate capillaries and also by Zhang et al. [97] using Teflon tubes. However, no detailed
description of the fabrication procedure was revealed which limited the availability of this type of
UME to other research groups. A comprehensive and very detailed description of the fabrication
procedure for disk-shaped Pt nanoelectrodes was published by Katemann and Schuhmann [98]. A
short piece of 25 µm Pt wire was placed in the middle of a quartz glass capillary while a focused
laser beam was used to locally heat the capillary. As soon as a particular viscosity of the quartz-
glass was reached, a single hard pull was automatically initiated which led to the formation of two
nanoelectrodes with diameters of sealed Pt wire down to 10 nm. Usually, the quality of the
fabricated nanoelectrodes is characterized by cyclic voltammetry, SEM and SECM.
As an alternative to the application of metal wires, Bauermann and co-workers [99] used thin
carbon fibers (≈ 7 – 8 µm) to fabricate flexible carbon fiber (CF) microelectrodes with vibration
State of the art
26
characteristics suitable for an optical shearforce-based feedback adjustment of the tip-to-sample
distance. If the desired diameter of the tip is below that of the original carbon fiber,
electrochemical etching in 10 mM NaOH with a periodic square wave potential of 3.9 V and a
frequency of 45 Hz can be applied to uniformly decrease the diameter of protruding carbon
fiber [100]. Insulation of the exposed carbon fiber surface using an anodic electrophoretic paint
and subsequent cutting of the very top of the tip with a sharp scalpel to expose a fresh disk-shaped
carbon surface allows fabrication of high-quality CF electrodes with a RG value ≈ 1
(RG = Dout / 2rtip, where Dout is the overall outer diameter of the tip while rtip is the radius of the
active disk-shaped electrode area) [101]. If the etching procedure was performed until the current
dropped to zero, than conically shaped tips with a nanometer-sized end were obtained. To
fabricate CF nanoelectrodes with nanometer tip curvature Hussien and co-workers [102] used
shearforce-based constant-distance SECM for the controlled approach of conically etched CF tips
to a film of a silicone elastomer until soft contact was made. The authors insulated the exposed CF
surface using electrochemically induced polymer deposition while the foremost end of the tip was
not insulated as it was protected by the polymer film. Following this procedure CF tips with
effective radii down to 46 nm were obtained. A detailed description of different applications of
CF micro- and nanoelectrodes in the investigation of individual cultured living cells can be found
in the comprehensive review by Schulte et al. [103].
2.2.1.3. Lateral resolution
The lateral resolution of the SECM set-up is mainly limited by the diameter of the active area
of the scanning tip (rtip) and the tip-to-sample distance (d). Thus, to reach high lateral resolution a
SECM tip with a very small diameter should be moved as close as possible to the sample surface
(≈ d / rtip = 0.1). Scanning at ≈ 25 nm above the sample surface becomes difficult as stray
vibrations and irregularities in the sample surface can cause a collision between the tip and sample
which can lead to damage of the SECM tip. Thus, for high-resolution measurements the tip-to-
sample distance must be controlled very precisely. One could control iT by keeping it at a constant
value adjusting the tip-to-sample distance with a feedback loop, as is done in STM. This approach
can be applied only if the sample is completely uniform with respect to its electrochemical
properties. However, in the case of samples with locally varying electrochemical activity, this
approach does not work and other methods that can “recognize” the nature of the sample or
methods that are completely independent of the sample properties are necessary. Wipf and co-
workers [104], [105] proposed modulation of the tip in z-direction while simultaneously recording
State of the art
27
the sign of diT / dz as a method to identify the nature of the sample. Alternative approaches to
control the tip-to-sample distance are discussed later (chapter 2.2.1.4). Additionally, the lateral
resolution of the SECM system can be improved ex situ by the application of image processing
techniques that efficiently lower the diffusional broadening by implementation of a range of
mathematical functions and filters, for example, a Gaussian filter [106].
2.2.1.4. Control of the tip-to-sample distance
The analytical signal recorded at the SECM tip is affected by the nature of the sample and also
by the topographic variations on the sample surface. Thus, precise control of the tip-to-sample
distance to exclude the influence of topographic changes on the measured signal is of great
importance. Additionally, to obtain high lateral resolution the SECM tip must be positioned in
very close proximity to the sample surface (d ≈ rtip). One of the first approaches reported for
imaging in the feedback mode of SECM was developed by Wipf and Bard [104]. This tip position
modulation (TPM) approach was based on the sinusoidal oscillation of a tip normal to the sample
surface with amplitude of about 10 % of the tip radius. An alternating current, which is
demodulated by a lock-in amplifier at the reference frequency of the tip oscillation, is used as the
imaging signal. An increase in the distance between the tip and the sample causes a decrease in
the tip current over a conductor but causes an increase over an insulating surface. Therefore, tip
currents measured over conductive and insulating surfaces are 180° out-of-phase. This change in
phase allows a reliable determination of surface properties. Detection of the in-phase component
of the alternating current allows the conductive nature of the sample surface to be identified. The
DC component of the measured current (Faradaic conversion of the mediator) was acquired
simultaneously to the AC component which allowed direct comparison of both signals. The
reported SECM images acquired with the in-phase TPM signal over the surface of an
interdigitated array (IDA) showed increased resolution as a result of the increased sensitivity to
the insulating material.
The TPM approach was used by Wipf and co-workers [105] for the constant-current imaging
of sample surfaces that contained both electronically insulating and conducting areas. The authors
reported an SECM system which combined an automatically switchable closed-loop servo system
to maintain a constant tip current by varying the distance between the tip and sample while the
TPM signal obtained information about the nature of the sample surface. Mapping of the sample
topography was performed by plotting the tip position (z axes) generated by the servo system.
State of the art
28
Another approach to control the tip-to-sample distance was developed by McKelvey et
al. [107]. This intermittent contact mode of SECM (IC-SECM) is also based on a sinusoidal
oscillation of the tip normal to the sample surface. The intermittent contact of the tip with the
sample surface leads to a dampening of the tip oscillation. This dampening is used as a feedback
signal for the lock-in closed loop operating circuit to control the relative tip position during
simultaneous mapping of the surface topography and local electrochemical activity. In contrast to
the TPM-based approach, this method uses a non-electrochemical signal for the feedback loop
regulation of the tip-to-sample distance.
An alternative methodology for tip-to-sample distance control utilizes a near-field shearforce
interaction between a SECM tip vibrating at its resonance frequency and the sample surface. A
number of different strategies were reported for the detection of the dampening of the tip vibration
upon approach to the surface. The first article about an optical detection scheme was published by
Ludwig et al. [108] and a few years later this methodology was further developed by Hengsten-
berg et al. [109]. The authors used a piezoelectric tube to agitate the vibration of the SECM tip at
its resonance frequency. To monitor the vibration amplitude a red laser beam was focused onto
the lower end of the glass insulation of the UME leading to the projection of the generated Fresnel
diffraction pattern onto a split photodiode. The difference in the currents measured on both
segments of the photodiode was analyzed with a lock-in amplifier with respect to the agitation
frequency. The magnitude and phase shift of the photodiode signal were used as measures of the
tip vibration. This approach was successfully applied to precisely position the SECM tip for
surface modification [89], positioning of non-amperometric SECM tips [109] and imaging of the
topography and activity of single secretory cells [110].
Katemann and co-workers [111] reported the application of an integrated, non-optical
piezoelectric shearforce-based detection system as an efficient and very sensitive approach to
detect the interaction between the SECM tip and the sample surface. A piezoelectric plate was
glued on the specially designed brass holder with a cylindrical hole well fitted to the outer
diameter of the SECM tip. A set of two small screws was used to fix the holder on the electrode
body. The best sensitivity to shearforces was achieved when two brass holders were mounted on
the microelectrode body in such a way that the piezoelectric plates were located opposite to each
other. The performance of the described methodology was illustrated by simultaneous imaging of
an array consisting of four Pt band electrodes with high lateral resolution for topography and local
electrochemical activity. This non-optical piezoelectric shearforce-based detection system was
successfully applied by Katemann et al. [112] for the high-resolution constant-distance imaging of
State of the art
29
a three-dimensional microstructure with an array of hexagonal holes (LIGA) and by Eckhard et
al. [113] for the high-resolution imaging of corrosion pits with overall dimensions of only a few
micro-meters.
Recently, Nebel and co-workers [114] extended the non-optical shearforce-based constant-
distance mode of SECM to a four dimensional shearforce-based constant-distance mode (4D
SF / CD SECM). This new mode was designed to probe the SECM tip response at several
predefined distances away from the sample surface at each point of a x,y-scanning grid. To
control the tip-to-sample distance the shearforce interaction between the oscillating SECM tip and
the sample surface was used. After the tip reached a point of closest distance (≈ 200 nm away
from the sample surface) a set of stepwise tip retractions was performed. At each new position the
current response and relative tip position were recorded. Thus, a set of 3D SECM images could be
obtained at known distances between the tip and the sample surface. The feasibility of this
imaging mode was demonstrated by high-resolution visualization of a Pt band electrode array.
2.2.2. Operation modes of the SECM
2.2.2.1. Feedback mode (FB)
In the feedback mode of SECM a constant potential (steady-state FB mode) [8] or a series of
potential pulses (chronoamperometric FB mode) [115] are applied to the SECM tip while it moves
across the sample surface in an electrolyte that contains an electrochemically active species. If the
applied potential is anodic enough to invoke diffusion controlled oxidation of R (reduced form of
the species) to O (oxidized form) the steady-state current measured at the tip can be described by
equation 2.11. This current is a function of the electrochemical activity of the sample surface and
the tip-to-sample distance. When the tip approaches an electrochemically active sample surface,
the oxidized form (O) formed at the tip can be re-reduced at the sample. This leads to the
formation of an additional amount of R which can diffuse back to the tip. This causes iT to be
higher than iT,∞ (bulk current). This effect is called positive feedback (Fig 2.5). The ratio of iT to
iT,∞ is a function of the tip-to-sample distance (d) where decreasing of d leads to larger feedback
currents. On the other hand, the feedback mode of SECM can also be used with electrochemically
inactive samples. In this case, approaching the sample surface lowers iT as the surface partially
blocks the hemispherical diffusion field as the tip-to-sample distances become smaller (negative
feedback). Thus, measurement of iT can provide information about the topography of a
homogenous sample surface and its electrical / chemical properties.
State of the art
30
Figure 2.5. Schematic representation of positive and negative feedback modes of SECM (R – the reduced form of the electrochemically active species, O – the oxidized form, iT,∞ - the bulk current at the SECM tip,
iT - the steady-state current at the SECM tip during approaching to the sample surface).
The first detailed theoretical evaluation of the feedback mode was reported by Kwak and
Bard [22]. The steady-state current measured at a SECM tip during operation in the feedback
mode was calculated by solving the steady-state diffusion equations using the finite element
method (FEM) with an exponentially expanding grid for both electrochemically active and
inactive samples. Based on these calculations working curves of iT as a function of the tip-to-
sample distance (d) and the radius of the tip active area (rtip) were constructed. Simulated curves
were compared with experimental data obtained from oxidation of ferrocene in acetonitrile
containing 25 mM TBABF4 at a 10 µm Pt tip above a Pt surface and an oxidized silicon wafer.
The theoretical calculations revealed the independence of the normalized tip current from the
diameter of the glass sheath at the very top of the SECM tip (Dout) for an electrochemically active
sample (positive FB). Thus, when rtip is known, the absolute tip-to-sample distance can be
obtained by fitting experimental data with a derived function. This allows the direct determination
of the sample topography without any additional calibrations. However, in the case of an insulator
(negative FB), the normalized tip current depends on the diameter of the glass sheath that
surrounds the tip (Dout) and decreases with increasing Dout. This effect was explained by the
increased blockage of diffusion pathways to the active tip surface with increasing Dout. Thus,
probing the sample topography and determination of the tip-to-sample distance over an insulating
sample requires knowledge of both rtip and Dout or an empirical calibration curve should be
recorded prior to the actual measurement.
In recent years, several attempts have been made to improve the accuracy and applicable range
for the originally reported approximations for SECM tips with conventional geometries, i.e. flat
disk-shaped UME or spherical tips [116]. Expressions were derived for pure positive
State of the art
31
FB [117], [118] and for pure negative FB [119], [120] while both included the RG ratios of the
used tips (RG = Dout / 2rtip). Figure 2.6 shows theoretical approach curves for electrochemically
active and inactive samples simulated for SECM tips with RG ratios ranging from 1.5 to 100. The
obtained results are very similar to those reported in the original work [22], however, an increased
accuracy of the derived approximations revealed a minor dependence of the normalized current on
Dout for positive FB. Additionally, as the set of data required for simulation of the current-distance
curves does not contain the RG ratio of the used SECM tip, UMEs with any RG values could be
used while the direct determination of the RG ratio of the used tip can be performed based on the
recorded approach curve.
Figure 2.6. Theoretical approach curves for an electrochemically inactive (1-4) and electrochemically active (5-8) sample simulated for SECM tips with different RG values (RG = 1.5 for #1 and #5,
5.1 for #2 and #6, 10.2 for #3 and #7, 100 for #4 and #8) [121].
A big advantage of the feedback mode is the possibility to apply it to investigate almost any
sample surface with high lateral resolution, especially when the sample cannot be polarized at an
externally applied potential as this may cause unwanted reactions or when it contains small
isolated conductive zones. In FB mode an electrochemically active sample does not necessarily
need to be polarized using a potentiostat, as most of the conductive area is located far away from
the tip and the sample is already polarized at a potential that is negative to the formal potential
E0of the R / O couple (this is valid in a solution containing mainly R). In this case, the localized
reduction of O generated by the tip in the gap between the tip and sample can be driven by the
oxidation of R at places at the sample surface located far away from the tip region.
One of the first applications of the FB mode of SECM (reaction-rate imaging) was the study of
the dependence of the steady-state current measured at the SECM tip on the nature of the sample.
State of the art
32
In the case of an electrochemically active sample the presence of catalytic particles with enhanced
activity or adsorbates on the sample surface causes variations in the electron transfer (ET) rates
which lead to changes in the measured current [122]. As the ET rate depends on the electrode
potential, the proper choice of applied potential can “make” an area on a heterogeneous surface of
the sample more active while the surrounding area will be visualized as less active. The feasibility
of the described approach was demonstrated by Wipf and Bard [123] for a composite electrode of
Au regions embedded in a glassy carbon matrix.
The investigation of ET kinetics is one of the most widely used applications of the FB mode of
SECM. Originally, theoretical treatments were developed and experimental studies were
performed for reversible [124] and quasi-reversible [125] heterogeneous ET processes. This
approach was based on application of a constant potential at the UME sufficient for the diffusion-
controlled conversion of a target species, e.g. oxidation of R to O. When the SECM tip is
positioned close enough to the surface of an electrochemically active sample, the positive FB
effect occurs according to the potential at the sample, i.e. the amount of O reduced back to R at
the surface of the sample depends on the applied potential. As the SECM tip can be positioned a
few fractions of a micrometer above the sample surface very high rates of diffusion back to the tip
are promoted. This makes the FB mode very attractive for investigation of fast heterogeneous ET
rates up to 1cm s-1.
A similar approach was used to measure the kinetics of homogeneous first-order (ErCi, [126])
and second-order (ErC2i, [127]) irreversible chemical steps following reversible ET. The concept
of this approach is similar to that used to measure ET rates but the species of interest may undergo
first- or second-order chemical reactions which lead to the formation of a non-electroactive
product on the way from the tip to the sample surface. Thus, during the experiment there is a
competition. O generated at the UME can travel to the electrochemically active sample surface, be
converted to R and R travels back to the tip (positive FB). Alternatively, O does not reach the
sample because of a chemical reaction leading to conversion of O to a product. The latter lowers
the amount of R in the gap region leading to a negative FB. Measurement of the steady-state
current as a function of d and comparison to the unperturbed positive FB provides the rate
constant (k) for the decomposition of O. For such studies the diffusion time of O to the sample (τ,
τ ~ d2/D0, where D0 is the diffusion coefficient of O) and the life time of O (~1/k for a first order
reaction and ~1/kC for a second-order reaction) are two essential parameters. Comparison of the
ratios of these two parameters provides very useful dimensionless parameters such as d2k/D0
(first-order reaction) or d2kC/D0 (second-order reaction). Thus, if this ratio is much larger than 1,
State of the art
33
an electrochemically inactive behavior is observed as O does not reach the sample surface. On the
other hand, if it is much lower than 1, then an electrochemically active behavior is observed as O
is converted back to R on the sample surface. Most useful measurements can be done where d2k/D
or d2kC/D ≈ 1 (the SECM tip should be very close to the sample surface), as in this case very fast
kinetic reactions with rate constants up to 107s-1 can be measured.
Another widely used application of the FB mode of SECM is the investigation of redox
reactions associated with surface-constrained enzymes based on measurements of the enzyme
kinetic or local distribution of active sites. The most commonly used enzymes are
oxidoreductases. These enzymes catalyze the transfer of electrons from an available redox
mediator (Med) to a substrate molecule which is highly specific for any given enzyme. If the
mediator can be reversibly converted from one redox form to the other and if high substrate
concentrations are available to keep the enzyme in fully reduced or oxidized forms, then Med
generated at the SECM tip can provide the reducing or oxidizing equivalents to the enzyme. This
leads to the initiation of the enzyme-catalyzed reaction and, if the SECM tip is positioned near the
surface, the current measured at the tip reflects the rate at which the enzyme reacts with the
mediator. Thus, high reaction rates (increased current, positive FB) indicate high enzyme activity
or high surface coverage, whereas low rates cause negative FB. To avoid possible contributions of
an electrochemically active sample surface to the observed positive FB, the enzyme should be
immobilized on an insulating surface. The feasibility of this approach was originally demonstrated
by Bard and co-workers [122] for glucose oxidase (GOD). Additionally, over the last 20 years
similar approaches were used to investigate many other enzymes such as PQQ-dependent glucose
dehydrogenase (PQQ-GDH) [128], horseradish peroxidase (HRP) [129] and nitrate reducta-
se [130].
More recently the feedback mode of SECM was applied to investigate self-assembled
monolayers on Au surfaces [131 - 133], to monitor charge transfer across polymer coatings on
conductive samples in corrosion research [134 - 136] and to probe the hydrogen evolution
reaction (HER) [137] as well as the hydrogen reduction reaction (HRR) [138].
2.2.2.2. Direct mode
The direct mode of the SECM uses the SECM tip as counter electrode which is positioned just
above the sample surface (d ≈ rtip), which is polarized at a predefined potential and functions as
working electrode. This causes a current at the tip equal to that at the sample but with opposite
polarity. Thus, the distribution of electric field lines is confined to a small volume between the tip
State of the art
34
and the sample electrode. This makes the direct mode a promising tool for the modification of
samples through local deposition or dissolution phenomena. Opposite to the feedback mode,
electrochemical reactions at the SECM tip and the sample electrode do not necessarily need to be
oxidation or reduction reactions of the same redox couple.
Similar to the direct mode of SECM is so-called the restricted diffusion mode which was
originally reported by Hüsser and co-workers [139] as a tool for high-resolution deposition of
metals in polymer films on conductive surfaces and also for localized etching of metal surfaces
with dimensions in the sub-µm range. The authors applied this approach to deposit continuous
patterns of Ag, Au, Pd and Cu on conductive surfaces. A silver electrode coated with a thin
Nafion® film was used as the sample electrode while Ag ions were electrochemically reduced at
the SECM tip. The counter reaction at the sample electrode was the dissolution of Ag. A similar
approach was used by Wuu and co-workers for the high-resolution deposition of polyaniline on
Pt [140].
The application of the direct mode of SECM for microstructuring of the sample surface was
applied by Schuhmann and co-workers [141], [142] to electrochemically induce deposition of
electrically conductive polymers on a conductive surface. To avoid complete depletion of the
monomer in the gap between the tip and sample often occurring during potentiostatic operation a
pulse sequence was developed. Application of this pulse sequence enables resupply of the
monomer to the interelectrode space after an initial oxidation pulse. The height of galvanostatic or
potentiostatic pulses was adjusted according to the oxidation potential of the used monomer
(pyrrole) while the duration of individual pulses was defined by the diffusion properties of the
monomer. The described procedure was successfully applied to connect Au bands deposited on a
glass surface but separated from each other by 100 µm. These experiments confirmed that pulse
profiles allow microstructuring even on insulating samples [143]. Additionally, three-dimensional
structures could be created by retracting the SECM tip during the electropolymerization process
using the shearforce positioning mode [89].
The direct mode of SECM therefore opens up the prospect to fabricate miniaturized biosensors
and biochips. Schuhmann and co-workers [142] applied the direct mode to localize the immo-
bilization of glucose oxidase (GOD) by entrapping the enzyme during the electrochemically
induced polymerization of pyrrole monomers. This approach was further improved by covalent
binding of periodate oxidized GOD to the surface of functionalized polypyrrole electro-
chemically deposited on the surface of a gold sample [144]. Covalent attachment of the enzyme to
the polymer matrix was achieved through the formation of a Schiff base between aldehyde groups
State of the art
35
at the surface of the enzyme and the terminal amino group of the deposited polymer. Successful
immobilization of the enzyme was confirmed using the enzyme-mediated positive feedback mode
of SECM.
Another widely used application of the direct mode is local structuring of self-assembled thiol
monolayers. Wittstock and co-workers [131] reported the localized electrochemically induced
desorption of self-assembled dodecylthiolate monolayers from a monolayer-covered Au surface.
The modified surface has been imaged using the feedback mode of SECM. It was discovered that
electrochemically induced desorption of alkanethiolate monolayers was localized to an area which
is about 2–3 times larger in diameter than the diameter of the microelectrode used for structuring.
To improve the lateral resolution of this technique Wilhelm and Wittstock [132] applied an
alternating voltage (2-10 kHz, ±1 V) between the Au surface covered with an alkanethiolate
monolayer and the SECM tip which was positioned at a distance d ≈ rtip above the sample. The
inhomogeneous field distribution led to localization of the desorption process directly beneath the
UME. The authors explained this effect as the collapse of the sample polarization once the
desorbed area reached the size of the ultramicroelectrode, as the SECM tip can deliver only a
limited current. If another thiol is present in solution, the Au surface cleared during the structuring
step was refilled with the secondary thiol [145]. Thus, this technique can be used to modify
functionalized thiol monolayers that constitute anchor groups for the attachment of biochemical
functional units directly after the structuring step [146]. This makes the direct mode of SECM a
very attractive tool for the fabrication of miniaturized biosensors and biochips.
2.2.2.3. Generation-collection mode (GC)
In contrast to the feedback mode of SECM, the generation-collection (GC) mode is applied in
an electrolyte that initially does not contain any electroactive species that can be converted by an
applied potential at either the SECM tip or the sample surface. The main principle of this mode is
to generate an oxidizable or reducible substance at the sample surface or at the SECM tip which is
subsequently detected at the opposite electrode. Two different GC modes can be distinguished
depending on the arrangement of the experiment (Fig. 2.7). If the compound of interest is
generated at the SECM tip and is converted at the sample surface the mode is called tip-
generation / sample-collection mode (TG-SC). On the other hand, if the sample was used to
generate a substance that can be converted at the UME located close to the sample surface the
mode is called sample-generation / tip-collection mode (SG-TC).
State of the art
36
Figure 2.7. Schematic representation of TG-SC mode (A) and SG-TC mode (B) of the SECM.
The ability of the TG-SC mode of SECM to generate a reactive etchant in a small gap between
the SECM tip and the sample surface has been widely used to localize structuring of the sample
surface by local oxidative dissolution or chemical oxidation. In some cases, regeneration of the
reagent at the sample surface can be used as an electrochemical feedback to control the tip-to-
sample distance by monitoring the tip current. The TG-SC mode was originally applied by Lin et
al. [147] for the localized microstructuring of a n-GaAs surface. Later, electrochemical etching of
Cu [148] and Si [149] samples was reported. Detailed descriptions of other applications of the
TG-SC mode of the SECM to microstructure surfaces can be found in review articles by Krämer
et al. [150] and Radtke et al. [151].
Microstructuring of the sample surface can be also performed through application of a
precursor either entrapped in a thin polymer film or directly deposited on the sample surface and
converting it to the desired structure by the localized generation of an oxidant or a reductant. To
deposit metallic Pd and Au on a sample surface Mandler and Bard [152] loaded polyvinyl-
pyridine films with [PdCl4]2- and [AuCl4]
- while the electrochemical reduction of metal complexes
was performed by local generation of [Ru(NH3)6]2+. Marck and co-workers [153] reported
formation of polythiophene micropatterns by locally induced polymerization of a monomer film
initially deposited on the sample surface. Local change of the pH value in the interelectrode space
was used by Zhou and Wipf [154] for localized deposition of polyaniline films on different
sample surfaces. The ability of the SECM to allow microstructuring and subsequent functional
characterization of the developed structures makes it a very valuable tool for the high-resolution
fabrication of modified interfaces.
One widely used application of the TG-SC mode is the investigation of catalytic activities of
different materials for the oxygen reduction reaction. As the ORR is an irreversible reaction, it
A B
State of the art
37
cannot be investigated using conventional FB modes of SECM. However, in the TG-SC mode
galvanostatic generation of oxygen by water electrolysis at the SECM tip can be applied to inject
oxygen into the gap between the tip and the sample while continuous measurement of the cathodic
current at the sample provides information about the local catalytic activity for ORR. This
approach was reported by Fernandez and Bard [155] for the investigation of the local catalytic
activity of highly dispersed fuel cell electrocatalysts in acidic electrolyte. Additionally, based on
explicit finite difference method (FEM) simulations Fernandez and Bard [156] developed the
theoretical background for localized quantitative kinetic studies using the TG-SC mode of SECM.
Experimental evaluation of the developed methodology used the oxygen reduction reaction in
phosphoric acid. Later, the TG-SC mode was used by Fernandez et al. [157] to investigate the
local ORR activity of metallic catalysts that consisted of binary and ternary combinations of Pd,
Au, Ag and Co deposited on the surface of a glassy carbon plate. Catalyst libraries were
composed based on reported thermodynamic guidelines while all experiments were performed in
acidic electrolyte using a new fast speed imaging mode. Additionally, the same methodology was
used by Fernandez et al. [158] to visualize the local ORR activity of non-platinum electrocatalysts
such as Pd-Ti and Pd-Co-Au in acidic electrolyte and also to investigate oxygen-reducing
enzymes for biofuel cells [159].
Recently, a new version of the TG-SC mode called the surface interrogation mode of SECM
(SI-SECM) was reported by Rodriguez-Lopez and Bard [160]. This mode is based on the
production of an electroactive compound, e.g. an oxidized form, at the SECM tip positioned in
close proximity to the unbiased sample. The subsequent reduction of the electroactive species at
the sample surface back to its reduced form causes a positive feedback as long as another
electroactive substance adsorbed on the sample surface is not completely oxidized. The amount of
adsorbed compound can be calculated by integrating the positive feedback current. This technique
was evaluated using electrochemical decomposition of formic acid at an unbiased Pt surface. The
detection of adsorbed hydrogen revealed that the decomposition mechanism of formic acid on the
Pt surface occurred via dehydrogenation steps. A similar approach was used by Wang et al. [161]
to investigate the poisoning of a Pt-based catalyst for ORR by adsorbed CO. The oxidizing agent
(Br2) was electrogenerated at the SECM tip while the adsorbed CO was detected and quantified by
analysis of the transient positive feedback current caused by Br- regenerated at the sample surface.
Electrochemical processes that are coupled with homogenous chemical reactions are most
commonly investigated through application of feedback mode [126] and the closely related TG-
SC mode of SECM [127]. Martin and Unwin [162] extended the theoretical basis of the sample-
State of the art
38
generation / tip-collection (SG-TC) mode originally reported by Engstrom and co-workers [163]
and demonstrated its applicability in the investigation of electrode processes that involve a
subsequent irreversible first-order chemical reaction (ErCi type). During kinetic measurements a
species O is generated at the macroscopic sample through the diffusion-limited oxidation with
potential-step control of a species R. On its way from the sample surface to the SECM tip a
species O can undergo a first-order chemical reaction in solution that leads to a decreased
collection of the species O at the tip. The interelectrode separation can be determined from the
transient response of the tip and chemical reaction rates can be assessed by subsequent fitting of
the recorded transient response of the UME. Alternatively, the chemical reaction rates can be
determined by simple measurement of the peak time and the corresponding peak current response
in a single transient. Additionally, high-precision measurements of differences in the diffusion
coefficients of the oxidized and reduced forms of the same redox couple were reported by Martin
and Unwin [23] following analysis of the time-dependent tip current response.
Similar to the FB mode, the SG-TC mode of SECM is widely used to investigate the kinetics
and distribution of immobilized enzymes. However, unlike the FB mode, it can be applied to
characterize enzymes immobilized on conductive samples (no influence of the mediator
regeneration on the conductive sample) and the results of SG-TC experiments are less depended
on the tip diameter. However, reliable quantitative evaluation of the experimental data requires
immobilization of the enzyme in a microstructure that can form a steady-state diffusion profile.
Reports on application of the SG-TC mode to investigate a wide range of enzymes, such as
glucose oxidase [164], PQQ-dependent glucose dehydrogenase [128], horseradish peroxi-
dase [165], alkaline phosphatase [166], galactosidase [167] and laccase [168] can be found in
literature.
Sanchez-Sanchez and co-workers [169] developed a methodology to quantify the reaction
intermediates generated in solutions at small samples (≈ 100 µm) using the SG-TC mode of
SECM coupled with linear voltammetry. The results of digital simulation of the collection
efficiency were found to be in good agreement with experimental results obtained using
ferrocenemethanol as a redox mediator. The developed methodology was successfully evaluated
through the quantification of hydrogen peroxide formed during the oxygen reduction reaction on a
Hg electrode in acidic electrolyte. The number of electrons transferred during the reduction of a
single oxygen molecule was determined to be around two which indicates that two electron
pathway processes dominate this reaction. The same methodology was applied by Sanchez-
Sanchez and Bard [170] to quantify the number of electrons transferred during the ORR on eight
State of the art
39
different materials in acidic electrolyte. It was found that only in the case of Pt and Pd-Co alloy
this number is close to 4 while for all other tested materials it has intermediate values between 2
and 4 as a function of the applied potential.
Recently, the application of the SG-TC mode to investigate oxygen evolution electrocatalysts
in acidic media was reported by Minguzzi et al. [171]. A library of pure IrO2 and Sn1−xIrxO2
combinatorial mixtures was used to prepare arrays of electrocatalyst spots. The activity of
individual spots for the oxygen evolution reaction (OER) was determined by localized detection
of the formed oxygen at the SECM tip. To increase the throughput of screening a thin Au layer
was deposited on the external wall of the SECM tip and kept at a constant cathodic potential to
reduce oxygen that diffused from neighboring spots under mass transfer controlled conditions. By
doing so, the tip shield consumes oxygen coming from the neighboring spots in the array while
the Pt disk microelectrode selectively detects the activity of the spot below the tip. It was shown
that the reported simulations and experimental results are in good agreement demonstrating the
effectiveness of the tip shield approach in combination with the SG-TC mode of the SECM for
visualization of local OER activity of the composite materials.
2.2.2.4. Alternating current mode (AC-SECM)
The alternating current mode of SECM is based on the application of an alternating potential to
the SECM tip immersed in an electrolyte solution while the alternating current response is
recorded and analyzed. For this an external function generator or an internal oscillator of the lock-
in amplifier (LIA) acts as the source of the sinusoidal perturbation signal which is transferred to
the potential input channel of the potentiostat. The alternating current measured at the SECM tip
is detected by the LIA at a frequency equal to that of the perturbation signal. The LIA output
signal contains information about the magnitude of the measured alternating current (R) and its
phase shift (θ) relative to the phase of the applied alternating voltage. In contrast to conventional
SECM, there is no need for the presence of a redox mediator in the electrolyte or polarization of
the sample under investigation. Both facts are considered as general advantages of the alternating
current mode of SECM.
Originally, the dependence of the AC-signal on the tip-to-sample distance was used by
Horrocks and co-workers [172] to precisely position an amperometric biosensor above the sample
surface to detect H2O2 diffusion. An alternating potential was applied to the SECM tip while the
solution resistance between the tip and the sample was conductometrically measured in a two-
electrode arrangement. An effect similar to the negative FB was observed during the approach
State of the art
40
towards an insulating surface which led to a decay of the AC-signal. This was explained by the
blockage of the field lines by the sample surface at sufficiently small tip-to-sample distances. The
authors demonstrated variation of the solution resistance as a function of the tip-to-sample dis-
tance and confirmed the applicability of the theoretical approximations originally developed for
negative FB theory to fit normalized conductance approach curves (R∞/R(d), where R∞ and R(d)
are resistance values measured with a tip being in a bulk solution and at the distance d away from
the sample). Later, monitoring of the AC-signal to position a potentiometric sensor above the
sample surface was used by Kashyap et al. [173]. Additionally, Osbourn et al. [174] applied the
same methodology to position a Pt microelectrode.
Investigation of sample topography using the AC mode of SECM was reported for the first
time by Alpuce-Aviles and Wipf [175]. To monitor the tip impedance a high-frequency
alternating voltage was applied at the SECM tip while a piezo-based feedback controller was
implemented to keep the impedance at a constant value. Thus, a constant tip-to-sample distance
could be maintained during an area scan. However, it was found that the AC-signal depends on
the electrochemical nature of the sample which leads to different responses over areas with
different conductivity. Nevertheless, the constant-impedance mode was successfully applied to
image model neurons immobilized on a homogenous insulating sample [176].
The application of the AC mode of SECM to distinguish conducting and non-conducting areas
on sample surfaces was reported for the first time by Katemann and co-workers [24]. The authors
performed area scans over a model sample containing Pt band electrodes of 25 µm width
separated by an insulating surface. As all experiments were performed in solutions with low ionic
strength (1 mM NaCl), the measured current magnitude (R) was strongly associated with the
electrolyte resistance (Rs) while the contribution of the double layer capacitance (Cdl) at
sufficiently high perturbation frequencies, compared to the overall cell impedance, was negligible.
Similar observations to that made by Horrocks et al. [172] were made by approaching exclusively
conductive or insulating areas on the model sample. By approaching an insulating surface the
current flow between the SECM tip and the counter electrode was hindered which lead to a decay
of the R value. On the other hand, by approaching a Pt band electrode the increase of the R value
was registered as proximity to a conductive surface enhances the current flow by reducing the
electrolyte resistance. Both effects are schematically summarized in figure 2.8. Similar to the
feedback mode of SECM these effects were denoted as negative feedback-type response (over
insulating surfaces) and positive feedback-type response (over conducting surfaces). Both types of
feedback behavior were also observed for the phase shift (θ) of the measured alternating current.
State of the art
41
Figure 2.8. Schematic representation of the current flow by approaching the SECM tip to an insulating (red line, negative feedback-type response) and a conductive
(blue line, positive feedback-type response) surfaces [adopted from 184].
The ability of AC-SECM to detect the distribution of electrochemical activity on the sample
surface was applied by the same group to visualize precursor sites of localized corrosion on
lacquered tinplates [177] and to investigate localized corrosion phenomena on the surface of oxide
film passivated NiTi shape memory alloy (Nitinol) [178] with high lateral resolution. Later, AC-
SECM was applied to detect inhomogeneity in the passive oxide layer of NiTi shape memory
alloy with high spatial resolution [179]. It was shown that the imaged microscopic corroding spots
on the surface carry a high reactivity to the onset of pitting corrosion upon anodic polarization of
the sample.
The alternating current mode of SECM was further developed by Baranski and
Diakowski [180] and equivalent circuits for the description of AC-SECM experiments were
formulated. The authors reported a four electrode electrochemical cell configuration for AC
experiments that offer the possibility of sample polarization at a predefined potential. By
monitoring the second harmonic of the alternating current the difference in electrochemical
activity of two different metals could be visualized.
To overcome the influence of the tip-to-sample distance on the measured AC-signal while
maintaining the sensitivity of AC-SECM to differences in local conductivity the measurement
should be performed in a constant distance mode and control of the tip-to-sample distance should
rely on a non-electrochemical signal. This approach was developed by Katemann et al. [111] and
was reported as the non-optical shearforce-dependent constant-distance mode of SECM. The
potential of this approach was demonstrated by the successful visualization of 25 µm Pt band
electrodes deposited on an insulating surface by constant-distance AC-SECM using a 10 µm Pt
UME and later local distribution of the conductivity on the same sample was visualized using a
State of the art
42
350 nm Pt nanoelectrode [181]. Constant-distance AC-SECM was applied by Eckhard et al. [113]
for high-resolution imaging of local pit initiation and instant pit growth on a surface of 316 Ti SS
specimen embedded in PVC.
To further improve the lateral resolution of AC-SECM Eckhard et al. [182] implemented the
alternating current mode of SECM in an AFM set-up. A combined AFM-SECM probe with an
integrated Pt ring microelectrode recessed approximately 1 µm from the apex of a conventional
AFM cantilever tip was used during both contact AFM and AC-SECM experiments. A detailed
description of the AFM-SECM probe fabrication procedure can be found elsewhere [183]. The
performance of the developed AFM-SECM system was evaluated by simultaneously mapping
local distributions of conductivity and topography of a model sample consisting of Au band
electrodes of 500 µm width deposited on a glass surface. As the Pt ring microelectrode does not
touch the sample surface simultaneous data acquisition using the AFM and AC-SECM operation
modes is possible. The control circuit of the AFM scanner enables the constant tip-to-sample
distance operation of the combined AFM-SECM system while the constant distance between the
Pt ring microelectrode and the sample surface is defined by design of the AFM-SECM probe.
Thus, the measured AC-signal is only influenced by local distribution of the conductivity on the
sample surface. High lateral resolution of the AC-SECM technique was tested on a periodically
micro-patterned silicon nitride / Au layer sample with recessed 1 µm disk electrodes [182]. Both
experiments revealed very high lateral resolution in combination with high-contrast electro-
chemical imaging of the distribution of local conductivity.
Since its introduction in 2002, many applications of AC-SECM were reported. However, there
was no quantitative theoretical model and some reported results were contradictory, especially,
different observations during approaches to conductive samples were reported. The alternating
current measured during the AC-SECM experiment is directly proportional to the admittance and
inversely proportional to the overall impedance (Z) of the electrochemical cell. The overall cell
impedance is mainly influenced by the solution resistance (Rs) and the double layer capacitance
(Cdl). To access the local distribution of the electrochemical activity on the sample surface with
high lateral resolution SECM tips with small diameters of active Pt area (Pt Dout) should be
considered. Decrease of the Pt Dout leads to an increase of the Rs while the Cdl will decrease.
However, the impedance of both parameters will increase with decreasing electrode size [184]. On
the other hand, the impedance of the Rs does not change with an increase of the perturbation
frequency while Cdl can be diminished by increasing the frequency. Thus, to maximize the
electrochemical contrast during AC-SECM experiments the contribution of the solution resistance
State of the art
43
to the overall cell impedance should be maximized while the contribution of the double layer
capacitance should be minimized. This can be achieved by using low-conductivity electrolytes
and high perturbation frequencies. A detailed investigation of the electrochemical activity contrast
in AC-SECM on the perturbation frequency was reported by Eckhard et al. [185]. During the
measurements the perturbation frequency was varied from 543 Hz to 49.6 kHz. It was found that
at frequencies below 9.11 kHz a negative feedback response was observed at the AFM-SECM
probe inside the Au microelectrode while at frequencies above this threshold frequency a positive
feedback response was registered. This observation can be explained by suppression of the
capacitive contribution of Cdl to the overall cell impedance at high perturbation frequencies.
Further improvement of the AC-SECM technique focused on enhancement of the electro-
chemical contrast. Usually, AC-SECM measurements were performed at one fixed perturbation
frequency chosen empirically based on previous experiments (many single measurements
performed at varied frequencies). However, this does not necessarily mean that the optimum
electrochemical contrast would be obtained at a given frequency. To obtain the optimum electro-
chemical contrast in a single AC-SECM experiment Eckhard et al. [186] developed a four-
dimensional mode of AC-SECM (4D AC-SECM). This mode is based on sweeping of the
perturbation frequency in a full spectrum at each measurement point and recording a set of AC-
signal responses to provide a four-dimensional data set. Visual evaluation of 2D or 3D represent-
tations of the recorded data offers easy identification of the optimum electrochemical contrast and
the corresponding perturbation frequency. The performance of this approach was evaluated by
successful visualization of corrosion precursor sites, local defects in protective organic coating
and actively corroding pits on a 304 stainless steel sample.
One of the most commonly used approaches for the evaluation of the imaging quality in AC-
SECM is to record current-distance approach curves (operation in the distance domain) over areas
of interest on the sample surface at one fixed perturbation frequency. As an alternative to this
approach, Eckhard et al. [187] proposed the representation of surface approach curves in the
domain of real and imaginary parts of the impedance (operation in the impedance domain). The
selection criterion for conditions that offer the optimal electrochemical contrast is based on an
evaluation of the maximal distance between two of these approach curves. The experimental
conditions that offer the maximum contrast are considered to be optimal. To evaluate the usability
of this developed concept a couple of approach curves towards a Zn plate insulated with a non-
conductive polymer were recorded over intact areas and over scratched surfaces exposing metallic
Zn. For comparison, an area scan in the 4D mode of AC-SECM was performed over the same
State of the art
44
surface area. Experimental conditions associated with optimal electrochemical contrast observed
during 4D imaging were found to be identical to those identified by the new methodology prior to
the actual 4D AC-SECM measurement.
More recently, a new mode of AC-SECM was reported by Gebala et al. [188]. This mode
offers the possibility to simultaneously record the overall impedance of the sample while probing
alternations in the concentration gradient of a redox couple and alternating electric fields caused
by the local electrochemical activity of the sample surface. The new AC mode uses the
application of an alternating current to the sample while the resulting alternating potential is
recorded at a positioned SECM tip. As the lateral resolution of this technique is not intrinsically
limited by the tip size, it is a straightforward step on the way towards local impedance
measurements with high lateral resolution. Additionally, as no redox reactions take place at the
SECM tip, the new AC mode of SECM has no instrumental limitations, such as the threshold of
current detection or the impedance of the tip. The performance of the developed technique was
evaluated by successful visualization of the transition edge between a thin Pt film and the
insulating surface of the underlying sample using a 1 µm Pt SECM tip.
2.2.2.5. Redox-competition mode (RC-SECM)
The oxygen reduction reaction (ORR) is one of the most widely studied reactions in modern
electrochemistry [189]. Originally, imaging of the local catalytic activity for oxygen reduction
was introduced by Fernandez and Bard [155] using the tip-generation / sample-collection (TG-
SC) mode of SECM (see above). The local generation of O2 at the prepositioned Pt SECM tip was
used while the cathodic current was measured at the polarized sample. The main drawback of this
technique was overlapping of the analytical signal with background currents, which was
especially remarkable on larger samples. Additionally, the TG-SC mode of SECM offers rather
low lateral resolution since a decrease of the tip size will lower the analytical signal and make it
smaller than the background current. To overcome these limitations the transient redox-
competition mode of SECM (RC-SECM) was introduced by Eckhard et al. [25] as a tool to
visualize the local catalytic activity of a sample surface towards the ORR.
The redox-competition mode of SECM is a bipotentiostatic experiment where the local
catalytic information at the sample is transduced via a cathodic current measured at the SECM tip.
During the experiment the sample is polarized at a predefined potential which is sufficiently low
to invoke oxygen reduction while the tip competes with the sample for dissolved oxygen within
the gap between the positioned SECM tip and the sample surface (Fig. 2.9). If the tip crosses the
State of the art
45
diffusion zone above a catalytically active area on the sample surface, the measured cathodic
current (iTip) will decrease as the local concentration of O2 within a gap between the tip and the
sample is depleted. In the case of very active samples complete consumption of oxygen in the
vicinity of the sample surface will result in zero current at the tip which leads to poor sensitivity
towards local catalytic activity.
Figure 2.9. Schematic representation of the transient redox-competition mode of the SECM (RC-SECM).
To avoid complete oxygen depletion in the gap between the microelectrode and the sample and to
increase the contrast, a potential pulse profile is applied at the tip (Fig. 2.10 A). The applied pulse
profile consists of three potential pulses. The first potential pulse (P1) is a conditioning potential
which is applied to restore the diffusional equilibrium after movement of the SECM tip during
scanning. The second potential pulse (P2) is an injection pulse which is sufficiently anodic to
electrochemically split water leading to injection of a portion of oxygen into the gap between the
SECM tip and the sample. The potential and the duration of P2 should be carefully adjusted to
avoid local saturation of the electrolyte with oxygen as this may lead to the formation of gas
bubbles. The third potential pulse (P3) is the measurement pulse which is sufficiently cathodic to
invoke oxygen reduction. Figure 2.10 B shows a typical decay curve recorded at the SECM tip
during P3. The steep decay of the cathodic current at the beginning of the P3 pulse represents, in
addition to the decay of the double-layer charging current, very fast depletion of oxygen in the gap
between the SECM tip and the sample caused by diffusion of oxygen into the bulk solution as
well as oxygen reduction at both electrodes. After a short period of time (≈ 50 ms to 75 ms)
almost no current was recorded which indicates complete loss of the generated oxygen. Thus,
high-speed data acquisition during P3 is a prerequisite for good sensitivity during RC-SECM
experiments. A common duration of the P3 pulse is about 400 ms while a predefined sampling
rate of the AD / DA data acquisition board allows one data point to be recorded every 4 ms and
leads to overall 100 measurement points per single oxygen decay curve. Hence, one local chrono-
State of the art
46
amperogram is recorded over each measurement point of the x,y- grid. Software based analysis of
all chronoamperograms results in 100 snapshots of local oxygen consumption at the sample
surface. To determine the moment during P3 that offers the highest contrast, a normalized current
decay curve at the most active area on the sample surface was correlated with a normalized decay
curve of the mean value of the whole image. Based on the representation of absolute difference
between both decay curves as a function of the pulse duration the optimum moment was
identified [25].
Figure 2.10. Potential pulse profile applied at the SECM tip to avoid complete depletion of oxygen in the gap between the tip and the sample (A). Typical oxygen decay curve recorded during the measurement pulse P3 (B).
To evaluate the performance and sensitivity of the developed technique, a model sample based
on electrochemically deposited platinum on the surface of a polished glassy carbon (GC) plate
that resulted in an inhomogeneous spot with subtle variations in catalyst loading was used. RC-
SECM experiments were performed using a 10 µm Pt SECM tips while different polarization
potentials were applied at the GC plate. The highest contrast was observed 24 ms after application
of the P3 potential pulse. Most probably, capacitive charging of the sample masks the oxygen
reduction current at shorter times, whereas depletion / diffusion of the injected oxygen at longer
times lead to the detection of almost zero current. The area with highest Pt loading could be
visualized even at low cathodic polarization [25]. This demonstrates the capability of RC-SECM
to visualize even slight variations in catalyst loading. The redox-competition mode of the SECM
provides qualitative information about the local catalytic activity while quantitative interpretation
of the measured cathodic current remains difficult.
The RC mode of SECM was successfully applied by Karnicka et al. [26] to visualize the spatial
distribution of the biocatalytic activity of bilirubin oxidase / Os-based redoxpolymer assemblies
A
B
State of the art
47
designed for application as ORR catalyst in biofuel cathodes. Okunola et al. [28] used RC-SECM
to investigate the local electrocatalytic activity of different metalloporphyrin spots towards
oxygen reduction. It was found that Mn-based tetratolyl porphyrin exhibited the highest catalytic
activity among similar Fe- and Co-based porphyrins. Recently, Nagaiah et al. [29] applied the
redox-competition mode of SECM to map the local catalytic ORR activity of Ag-Pt particles
homogeneously dispersed over a GC surface. It was found that particles with Ag-Pt compositions
of 40:59 exhibited the highest catalytic activity which was even higher than that of pure Pt. These
findings were in a good agreement with data evaluated by quantitative RDE analysis.
More recently, the RC mode of SECM was adapted to visualize the local catalytic activity of
catalysts in 400 mM HCl solution under specific conditions similar to those present during
industrial application of oxygen depolarized cathodes for chlorine production [30]. High concen-
trations of chloride ions do not allow the application of the originally reported pulse profile as the
P2 pulse (injection of oxygen) leads to oxidation of chloride ions to highly corrosive chlorine and
results in chemical dissolution of the underlying Pt catalyst. To avoid this, the potential pulse
profile was modified and potential pulses were adjusted to appropriate conditions. The local
catalytic activity of Pt-Ag nanoparticles deposited on the GC plate was successfully visualized
using a modified potential pulse profile while pure Pt deposits were unstable in 400 mM HCl
solution.
The oxygen reduction reaction on the surface of a metal sample can follow a 2-electron transfer
pathway which leads to the formation of hydrogen peroxide or a 4-electron transfer pathway
which leads to the formation of water. The branching ratio between these two pathways is of a key
importance as the produced H2O2 can lead to degradation of the catalyst material and limit its
applicability. Eckhard and Schuhmann [190] extended the originally reported potential pulse
profile to a sequence of potential pulses to enable high lateral resolution mapping of local ORR
activity and visualization of local H2O2 formation. The developed sequence of potential pulses
consists of a pulse profile originally reported for the RC mode of SECM which is applied at the
beginning of the sequence while the subsequent detection of hydrogen peroxide is based on a SG-
TC experiment. Directly after the oxygen competition pulse an equilibration and a short oxygen
injection pulses are applied while actual detection of H2O2 is performed by the application of an
anodic pulse which facilitates the oxidation of hydrogen peroxide at the SECM tip. Hence, the
anodic current measured at the SECM tip will be a measure of the amount of H2O2 produced at
the sample surface. To evaluate the ability of developed pulse sequence to visualize local ORR
activity and to probe the selectivity of different catalytic materials one Pt spot and one Au spot
State of the art
48
were electrochemically deposited on a glassy carbon plate. A set of RC-SECM experiments with
the new pulse sequence was performed by increasing the cathodic polarization of the sample. It
was found that at low potentials (-0.2 V vs. Ag / AgCl) only Pt can invoke diffusion controlled
oxygen reduction while Au remains inactive for ORR. At sufficiently cathodic polarization of the
sample (-0.4 V vs. Ag / AgCl) both metals exhibited remarkable catalytic activity with Pt being
more active than Au. However, these two metals showed completely different selectivity.
Subsequent detection of hydrogen peroxide revealed much more H2O2 being formed over the Au
spot while almost no H2O2 was detected over the Pt spot. These findings are in very good
agreement with the well-known property of Au to facilitate the 2-electron transfer reaction while
Pt predominantly catalysis the 4-electron reduction of oxygen. Thus, the extended pulse sequence
of RC-SECM is a powerful tool to sequentially visualize local ORR activity and investigate the
catalyst selectivity.
The electrocatalytic oxygen reduction at metalloporphyrins largely depends on the central
metal ion. To investigate the selectivity of tetratolyl porphyrins (TTP) with Mn, Co and Fe central
metal ions Okunola et al. [28] applied the extended pulse sequence of RC-SECM. A set of RC-
SECM experiments was performed on different porphyrin spots electrochemically deposited on
the surface of an ITO plate by increasing cathodic polarization of the sample. The smallest anodic
current was detected over a Mn-based TTP spot while the Co-based TTP was found to produce
the highest amount of hydrogen peroxide. This observation was in a good agreement with
additional experiments performed using a rotating disk electrode.
More recently, Guadagnini et al. [27] exploited the RC mode of SECM to visualize the local
electrocatalytic activity of a thin Prussian blue (PB) spot electrochemically deposited on a glassy
carbon plate towards the reduction of hydrogen peroxide. The originally reported potential pulse
profile was modified by excluding the oxygen injection pulse while a set of RC-SECM
experiments was performed in a phosphate buffered saline solution (PBS) that contained 1 mM
H2O2. During the experiments the sample was polarized at a potential sufficient to invoke the
reduction of hydrogen peroxide on the PB spot but too low for the same reaction to occur at the
unmodified surface of a GC plate. The depletion of hydrogen peroxide in the gap between the PB
film and the SECM tip positioned just above it leads to the decay of anodic current recorded
during the H2O2 measurement pulse. Additionally, the same detection principle was applied to
successfully visualize the spatial distribution of the electrochemical activity towards H2O2
reduction on an electrochemically deposited PB spot covered with a polymer-entrapped glucose
oxidase film as a model biosensor.
Problem identification
49
3. Problem identification
In recent years, SKPFM has been shown to be a powerful method for the determination of local
potential differences in many alloy systems. Frankel and co-workers [191] reported a good
correlation between corrosion potentials for a series of different metals measured while immersed
in a corrosive electrolyte and the CPD distribution measured upon emersion using the conven-
tional SKPFM technique under ambient conditions. This gave rise to the idea that cathodic and
anodic sites can be identified via SKP or SKPFM while the intensity of the local contrast between
different phases, such as the Al matrix and an intermetallic particle, can provide an indication of
the strength of a local galvanic element during immersion in corrosive electrolyte [192]. Follo-
wing this, a similar approach was used to investigate a number of different magnesium [193],
aluminum [194], and steel alloys [195].
It is well known in electrochemistry that potentials of metals immersed in an electrolyte are
determined by the equilibrium of anodic and cathodic reactions. Therefore, corrosion potentials
will critically depend on environmental factors such as the pH value and the concentration of
chloride ions present in a given electrolyte. Hence, no general correlation can be expected [196],
although in many publications such a correlation appears to exist for potentials measured with
SKP or SKPFM on freshly polished surfaces of different Al based alloys [197 - 199], Mg based
alloys [37], and Ti based alloys [200].
Iron is active in acidic electrolytes while niobium is active in alkaline solutions. Comparison of
the Volta potential difference measured by SKP on Fe and Nb samples upon emersion in
electrolytes with different pH values indicate higher activity of iron at pH 2 while already at pH 4
the niobium starts to be more active. However, electrochemical measurements of the corrosion
potentials performed on samples immersed in the same electrolytes indicated the opposite
behavior (at pH 4 Fe is active while Nb is passive) [196]. Thus, applying the correlation approach
reported by Frankel et al. [191] one would wrongly indicate Nb as being active (local anode) in an
electrolyte with pH 4 with Fe being passive (local cathode).
Obviously, a correlation between the CPD measured by SKPFM on freshly prepared alloy
surfaces and its later corrosion behavior is not necessarily straight forward and additional
techniques must be used to unequivocally elucidate the corrosion processes. In the case of alloys
that contain small intermetallic compounds of µm or even nm dimensions, the investigation of
corrosion processes is even more complicated and often requires the synthesis of pure bulk phases
identified in an alloy of interest. Thus, it becomes evident, that in addition to high-resolution
mapping of the CPD distribution on the sample surface, further information on the electro-
Problem identification
50
chemical or electrocatalytic activity at a similar lateral resolution is required. Additionally, one
should keep in mind that even for relatively large model samples such as Cu dots of several
hundred micrometers deposited on an Al sample the obtained potential contrast delivered by the
SKPFM technique reported to be prone to large errors due to topographical artifacts and stray
capacitances [65]. On the other hand, measurements of the CPD distribution performed using the
conventional SKP technique are less sensitive to topographical features on the sample surface as
also to parasitic stray capacitances.
As shown in chapter 2.2, SECM is a powerful technique for the investigation of localized
electrochemical processes. The AC mode of the SECM has particularly been shown to provide
excellent contrast between regions with different conductivity. This allows very efficient
localization of targeted Cu rich areas in different Al alloy systems. RC-SECM is certainly an ideal
tool to investigate electrochemical activity at localized galvanic elements and visualize local
catalytic activity towards the oxygen reduction reaction. Thus, integration of SKP and SECM into
a combined set-up provides a unique system that being able to deliver complementary information
about both localized corrosion processes and local catalytic activity with a high lateral resolution.
One of the most complicated challenges to develop such an integrated SKP-SECM system is to
mechanically combine the sinusoidal vibration of the conductive Kelvin probe above and
orthogonal to an electrically conductive sample (SKP mode) with ”static” scanning across the
surface of a sample using an insulated disk-shaped ultramicroelectrode (SECM mode). A multi-
functional measurement head that contains a working electrode, which can be operated in
dynamic or static regimes, is supposed to be a very promising solution.
The aim of this work is to develop and evaluate the applicability of an integrated SKP-SECM
system which would fulfill the following requirements:
• Subsequent application of the combined SKP-SECM system using the same electrode in
both SKP and SECM operation modes over exactly the same area of the sample surface.
• Operation of the combined SKP-SECM system under a controlled atmosphere and high relative air humidity.
• Precise control of the tip-to-sample distance combined with ability to follow the sample
topography.
• High-lateral resolution (up to sub-µm range) in both modes of operation.
Results and Discussion
51
4. Results and Discussion
4.1. Preliminary investigations
Prior to development of the integrated SKP-SECM system the investigation of possible
correlations between local surface potential differences, measured by SKP under ambient
conditions, and corrosion reactions at heterogeneous surfaces during immersion in a corrosive
electrolytes was performed using two conventional set-ups. A model sample that mimicked the
interaction of Al and Cu in Al alloys during the heterogeneous corrosion process in an electrolyte
containing chloride ions was chosen to evaluate the complementary nature of the information
received from combined SKP / SKPFM and SECM measurements.
4.1.1. Combined SKP, SKPFM and SECM measurements
High-performance alloys, such as aluminum alloys or high-strength steels are characterized by
their rather complex phase structures. Galvanic coupling between different phases plays a crucial
role in the performance of materials during immersion into electrolytes while corrosive attack is
taking place. For instance, Al-Cu and Al-Cu-Mg alloys suffer from localized corrosion due to
selective dissolution of the Al matrix and intermetallic particles (IMPs) which is initialized and
driven by cathodically active IMPs [201]. To optimize the corrosion protection of these alloys a
better understanding of the role and interaction of the localized galvanic activities is necessary.
However, due to the small size of such IMPs direct study by in-situ electrochemical methods is
challenging and complicated.
To mimic an Al based alloy model samples (Type A test sample) based on 200 nm thin Cu
islands (≈ 500 µm x 200 µm) deposited on freshly evaporated Al samples were prepared. The size
of Cu structures was much larger than the dimensions of intermetallic particles on the surface of
technical alloys. This was intentional to be well within the resolution range of the standard SKP
and SECM techniques later used to access the CPD distribution and for localized electrochemical
experiments. A detailed description of the preparation procedure of the model samples is
presented in chapter 7.3.
Figure 4.1 A shows an SEM image of the Cu islands deposited on the surface of an Al sample.
The red rectangle marks an area investigated using conventional SKPFM to assess the surface
potential contrast between the Al and the Cu region (Fig. 4.1 B). Due to the higher atomic weight,
the Cu regions give a bright contrast in the backscattered imaging mode (BSI) as compared with
the lighter atomic weight Al matrix. A potential contrast between the deposited Cu structures and
Results and Discussion
52
488.7µm 21
2.7
µm
150µm
A
Po
ten
tia
l /
VS
HE
the surface of the Al specimen of around 400 mV was observed by SKPFM. To probe the CPD
distribution over a much larger area than that investigated during the SKPFM measurements, the
model sample was transferred into the conventional SKP system and an area scan was performed
using a 100 µm Cr/Ni tip under ambient conditions (Fig. 4.2). Prior to the actual measurement, the
SKP system was calibrated using a drop of saturated CuSO4 solution placed in a small Cu vial
(E˚Cu2+/
Cu = 0.32 VSHE).
Figure 4.1. SEM micrograph of the prepared model sample (A, brighter structures correspond to Cu islands).
CPD mapping performed over the sample area marked with the red rectangle using SKPFM (B) (measured by C. Senöz [202]).
Figure 4.2. Area scan over the model sample performed with SKP using a 100 µm Cr/Ni tip under ambient conditions. The measured CPD distribution is reported vs. SHE
(measured by C. Senöz [202]).
The sample surface (Al) displays potentials down to values of around -0.6 VSHE while the
nobler Cu islands were detected to have more positive potential of around +0.1 VSHE. As reported
by Rohwerder and co-workers [196], potentials measured in air on the passive surface of metals
are about 0.05 V to 0.1 V more positive than the flatband potential of the corresponding oxide.
B
Results and Discussion
53
This, in turn, depends on the electronic structure of the oxide as well as on details of the growth
conditions and ageing process. For some metals the potentials of the passive layers are quite
reproducible. For example, values of 0 to 0.1 VSHE are characteristic for Fe. However, other
metals show a much wider variation and values as low as -1 VSHE can be measured on freshly
evaporated or polished Al samples. With time, the measured potential can easily increase up to
-0.4 VSHE as a result of a decrease of defects in the electronic structure of the oxide caused by
further oxidation during atmospheric exposure [196]. For the native oxide on Cu, potential values
in the range of 0.2 VSHE to 0.4 VSHE are typical. Hence, the measured values are in good
agreement with those reported in literature considering that the Al sample was freshly prepared.
The contrast in potential between the Al and Cu regions derived by SKP lies in the range of up to
0.6 V, while the contrast obtained by SKPFM is only 2/3 of that CPD value. This is in agreement
with earlier reports that potential contrasts measured by SKPFM are prone to artifacts [65], even if
the characteristic length scales are microscopic and not nanoscopic. In any case, a clear contrast is
delivered by both methods with Cu exhibiting as expected the nobler potential.
In the case of real Al alloys, many different phases are distributed on the sample surface,
giving a more complex potential contrast relative to the Al matrix. In order to assess the
corresponding differences in electrochemical activity during immersion in a corrosive electrolyte,
an in-situ electrochemical technique with a suitably high lateral resolution is needed. The
application of different SECM modes offers a pathway for such an in-situ technique.
The 4D-AC-SECM technique is an alternating current mode of the SECM which was
successfully applied to visualize differences in local electrochemical activities on surfaces [186].
During scanning, a frequency spectrum of the imposed alternating voltage between 273 Hz and
96485 Hz in logarithmic scale was applied over each measurement point on the surface and a
corresponding spectrum of AC responses was recorded. During the experiment the sample was
kept at open circuit potential (OCP). The simplest equivalent circuit for an electrochemical cell in
the absence of any redox mediator is a resistor and a capacitor in series. At low perturbation
frequencies, the tip double layer capacitance dominates the overall cell impedance. The magnitude
of the alternating current increases with increasing frequency because of the inverse proportional
dependency of the capacitance on the excitation frequency. Beyond a certain threshold frequency,
the capacitance of the electrode tip starts to be permeable for the alternating current [186].
Figure 4.3 shows the absolute difference in the magnitude of the AC response which is plotted as
a function of the applied alternating voltage frequency. The highest contrast between the Cu and
the pure Al surface was obtained at 2.31 kHz. The decay in the current magnitude observed above
Results and Discussion
54
2.31 kHz is most likely caused by inevitable stray capacitances and / or because of bandwidth
limitations of the used potentiostat [186].
Figure 4.3. The absolute difference in the magnitude of the AC response plotted as a function of the applied alternating voltage frequency (273 Hz to 96485 Hz) recorded with a 25 µm Pt SECM tip positioned
over the Cu structure on the Al surface (1 mM NaCl, Vpp = 0.2 V).
The frequency of 2.31 kHz was chosen as agitation frequency to perform classical AC-SECM
measurements on the model sample. A detailed description of the 4D-AC-SECM measurement is
presented in chapter 7.1.4.
Figure 4.4 A shows photographic images of the prepared model samples with the marked areas
showing regions over which area scans in the AC-SECM (#1) and RC-SECM (#2) modes were
performed. Figure 4.4 B shows an area scan performed in the AC mode of SECM using a 25 µm
Pt disk microelectrode. The Cu regions, which appear as depressed regions in the 3D plot, are
nicely resolved due to their higher electrochemical activity in comparison with the surrounding Al
surface covered with a native Al oxide layer. This unequivocally demonstrates that the AC-SECM
measurement at the optimum frequency allows localization of the Cu rich structures on the Al
surface.
Directly after imaging the Al-Cu sample by AC-SECM, and without changing solutions or the
electrode arrangement, a scan in the redox-competition mode of SECM was performed over the
sample area as marked in figure 4.4 A (#2). During scanning, the SECM tip was polarized at
-0.6 V vs. Ag/AgCl while the oxygen reduction current was monitored as a function of the x- and
y-position. A 2D RC-SECM image of this scan is presented in figure 4.5.
Results and Discussion
55
B
Figure 4.4. Photographic images of the prepared model samples (A). Dotted rectangles are marking areas over which area scans in the AC-SECM (#1) and RC-SECM (#2) modes of operation were performed
using a 25µm Pt SECM tip. AC-SECM area scan over area #1 (1 mM NaCl, tip-to sample distance = 2.5µm, Vpp= 200 mV, F = 2.31 kHz) (B).
Figure 4.5. 2D-RC-SECM area scan performed over the sample area as marked in figure 4.4 A (#2) using a 25 µm Pt SECM tip (1 mM NaCl, tip-to sample distance = 2.5 µm, Etip = -0.6 V vs. Ag/AgCl).
Smaller oxygen reduction currents were measured over the deposited Cu structure due to their
higher cathodic activity as a result of the higher oxygen reduction rate on Cu. This leads to lower
oxygen concentrations observed over the Cu regions and therefore lower oxygen reduction
currents measured with the SECM tip.
Cu
Al
#1 #2
A
Results and Discussion
56
4.1.2. Conclusion
A model sample mimicking an Al based alloy was prepared by sputtering pure Cu on a freshly
evaporated Al layer leading to formation of Cu structures with dimensions in the micrometer
range. To probe the CPD distribution over the prepared sample, SKP and SKPFM were applied.
Volta potential differences measured over the Cu structures and Al sample were around 0.6 V in
SKP measurement while the CPD was only about 0.37 V in the case of the SKPFM measurement.
Taking into account the exposure of the prepared sample to ambient air for a few days before the
SKP measurements were performed, it is reasonable to conclude that the measured values are in
good agreement with those reported in literature. To localize the deposited Cu structures, 4D-AC
SECM was employed. The optimal frequency providing the best contrast between Cu covered
areas and a clean Al surface was found to be 2.31 kHz. The electrochemical activity of localized
Cu structures was visualized in the RC mode of SECM directly after the 4D-AC-SECM
measurement using the same 25 µm glass insulated Pt SECM electrode without changing the
electrolyte or replacing the sample. It was shown that sequential use of SKP and in-situ 4D-AC
and RC modes of SECM is a very promising approach to assess the distribution of heterogeneities
on the sample surface and to visualize Faradaic processes taking place at heterogeneous areas
during corrosion of the material in an electrolyte.
Results and Discussion
57
4.2. Development of the integrated SKP-SECM system
In this chapter, the design of the integrated SKP-SECM system will be introduced. Based on
the longtime experience of the collaboration partners at the Max-Plank Institute for Iron Research,
it was decided to integrate the vibrating capacitor technique into the conventional SECM set-up to
measure the contact potential difference in the SKP operation mode. The technical realization and
function of each component of the integrated SKP-SECM system are then discussed in detail.
4.2.1. Design of the integrated SKP-SECM system
The design of the SKP-SECM system was focused on performing SKP and SECM
measurements without the necessity of changing the tip or replacing the sample. The selection of a
tip for the combined system was therefore predicated by the requirements of conventional SKP
and SECM. Due to the fact that SECM measurements require an insulated disk electrode whereas
the demands for a SKP tip are less rigorous, a conventional SECM tip, namely a glass-insulated Pt
disk microelectrode, was selected as the tip for both SKP and the SECM measurements.
It was decided to use the vibrating capacitor technique for the design of the integrated SKP-
SECM system. This technique is based on a sinusoidal vibration of the Kelvin probe above the
sample surface and the resulting alternating current is measured. Application of the compensation
voltage with a particular magnitude causes the alternating current to be equal to zero (“null output
condition”). In conventional SKP systems the null output condition is achieved by means of a
lock-in amplifier (LIA) and a compensation unit utilizing a proportional [203] or an integral
feedback [204]. The main disadvantage of these systems is a small signal-to-noise ratio at the
balance point mainly caused by electromagnetic pickup [205]. Thus, it is very important to
minimize the noise level. To overcome this limitation the use of an off-null mode of operation was
proposed by Ritty et al. [206]. A number N of discrete values of backing potential (U0) ranging
from -2.5 V to +2.5 V are imposed on the system and the corresponding values of the Kelvin
current (IKelvin) are recorded. Based on the statistical linear correlation-regression analysis of the
Kelvin current as a function of the applied U0 the CPD can be derived generally with a good linear
correlation coefficient r2 of higher than 0.9999 and errors of less than 10 mV [206]. A modify-
cation of this method was developed by Baikie et al. [207], [208] and involved setting of the
backing potential at +/- 20 mV from the balance point to permit both a high output signal and
accurate determination of in-phase noise components in the neighborhood of the balance point.
This feature has distinct advantages in systems with piezoelectric drives where the driver pickup
Results and Discussion
58
is the prevalent source of noise [209]. Both approaches to measure the CPD were successfully
implemented into the developed SKP-SECM system.
The choice of the driving mechanism for the scanning Kelvin probe plays an important role
especially considering attempts to minimize the noise level and stray capacitance [205]. Different
driving mechanisms based on an electromagnetic solenoid [210], a voice coil [211], a piezo-
electric system [212 - 214], or an electrostatic system [215] were proposed. Presently, only voice
coil or piezo based oscillating systems are used. In the case of voice coil based oscillating
systems, an alternating voltage is applied to the voice coil thus creating an alternating magnetic
field. The interaction of this field with a magnetic field of the permanent magnet generates a
mechanical force causing the coil and thus the attached Kelvin probe to move back and forth.
Very low off-axis displacements of the vibrating electrode can be achieved by means of the probe
suspension system based on two thin circular disks vibrating in a perfect plane-parallel
fashion [208]. The main advantages are the large amplitudes of oscillation (up to 10 mm), much
smaller electromagnetic pickup, and the use of long mechanical transducers for placement of the
driver far away from the sample. In contrast, piezoelectric based oscillating systems are mainly
used in commercial SKPFM devices. Piezo-driven oscillation systems offer a very high-precision,
a possibility to work at high oscillating frequencies and an easy and fast control. However, their
main disadvantage is a strong electromagnetic pickup due to the relatively high driving voltage for
the piezo element in close proximity to the sample. This limits the CPD resolution and requires
additional effort for shielding such as among others the integration of a µ-metal case around the
driver. To enable the operation of vibrating electrodes at high frequencies and with high precision,
it was decided to use the pre-loaded piezo actuator especially designed for dynamic applications.
For very low off-axis displacement of the vibrating electrode, a suspension system was
implemented in the SKP-SECM measurement head. To minimize the electromagnetic pickup
caused by the piezo actuator it was mounted as far away as possible from the input of the
operational amplifier and careful attention was paid to ensure sufficient shielding of all the
electric components of the Kelvin current amplifier.
The selection of the pre-loaded piezo actuator was based on theoretically calculated
performance parameters, which the actuator should provide for the SKP-SECM system operation
under specific conditions. Operational parameters of conventional SKP systems, provided by the
collaboration partners at the Max-Plank Institute for Iron Research, were used as the basis for
these calculations. The main operational parameters are the tip oscillation frequency (up to 2 kHz)
and the tip oscillation amplitude (up to 30 µm). As the pre-loaded piezo actuator was chosen as
Results and Discussion
59
the driving mechanism for the integrated SKP-SECM system, the actuator restoring force
provided by this actuator and its resonant frequency were also taken into account. The restoring
force provided by the piezo actuator can be calculated from the load applied to the actuator, the
working frequency and the range of motion (Eq. 4.1).
% = ∆ ∗ (&' ∗ )) ∗( (4.1)
Equation 4.1. Restoring force provided by the piezo actuator (∆d – range of motion, f - working frequency, m - load applied to the actuator).
The weight of the SKP-SECM tip was estimated to be 3 g and the required restoring force for the
tip oscillation under desired conditions was 14.2 N. Since the actuator has to move the SKP-
SECM tip, this additional load will influence the resonance frequency of the actuator and, in some
cases, can even limit the possible frequency range. The resonance frequency of the piezo actuator
is dependent on the applied load and its stiffness (Eq. 4.2).
= (&' ∗ √(/)) (4.2)
Equation 4.2. Resonance frequency of the piezo actuator under load (f – resonance frequency, m – load applied to the actuator, k – stiffness of the actuator).
From the many pre-loaded piezo actuators available on the market, the PSt 150/5/60 VS10
actuator from Piezomechanik GmbH (Jena, Germany) offered the optimal performance parame-
ters. It is mechanically pre-loaded with a pre-stress force of 150 N, has a maximal stroke of
60 µm, a stiffness of 8 N/µm, an electric capacitance of 2.4 µF and a resonance frequency of
15 kHz. According to equation 4.2 its resonant frequency under a load of 3 g will be around
8.2 kHz which is far above the desired operational frequency of 2 kHz.
The lateral distribution of the CPD on solids i.e. the value of the displacement current is
directly dependent on the surface area of the Kelvin probe, the oscillation frequency and indirectly
on the distance between Kelvin probe and sample. Reduction of the size of the Kelvin probe
should lead to a substantial decrease in the measured Kelvin current and consequently to lower
sensitivity and accuracy [48]. On the other hand, the lateral resolution of the SKP technique is
mainly limited by the dimension of the Kelvin probe [205]. Thus, for high-resolution
measurements a small tip has to be placed very close to the sample surface. Two different
approaches for the regulation of the distance between the Kelvin probe and the sample have been
Results and Discussion
60
reported. One possibility to control the tip-to-sample distance is taking advantage of the
independence of the quotient of two harmonics of the measured displacement current on the value
of the CPD [216]. An alternative possibility proposed by Bonnet et al. [217] is based on the
addition of an additional signal (Um sin ωmt) to the backing voltage where the frequency ω2 is very
different from the oscillation frequency of the Kelvin probe (ωm << ω1, Eq. 4.3).
. =
(4.3)
Equation 4.3. Alternating current demodulated with a second LIA (Um - additional potential for height regulation, ε - dielectric constant of the medium, ωm - frequency of the
additional potential, A - active tip area, d0 - tip-to-sample distance).
The corresponding signal is only dependent on the distance between the probe and the sample
(d0). Thus, keeping this signal constant by varying the position of the Kelvin probe with an
integrated z-positioning system can be used for feedback-loop regulation of the tip-to-sample
distance [217]. By this, the sample topography can be derived during the SKP scanning making
the CPD independent from the sample topography. The latter distance regulation approach was
integrated into the developed SKP-SECM system.
In order to perform sequentially CPD and topography scanning followed by SECM imaging
without moving the sample and using the same tip, the SKP-SECM system was designed in that
way, that it is possible to choose between the SKP and the SECM mode of operation by actuating
two reed relays. A scheme of the system connections for SKP measurements is shown in
figure 4.6 A. A conventional SECM microelectrode, a pre-loaded piezo actuator, a high-frequen-
cy power supply (LE 150/100 EBW), a two channel oscilloscope for signal control (Voltcraft VC
630-2), a lock-in amplifier (EG&G 5210) and a home-build integrator for measurement of the
CPD were used. A second lock-in amplifier (EG&G 5210) was further integrated for the feedback
regulation of the tip-to-sample distance. Connection to the PC was established via a 16 bit analog-
to-digital conversion card (PCI-DAS6014). The Pt-wire inside the SECM tip is connected via a
reed relay to the high-quality ground of the first LIA while the sample is connected via a second
reed relay to the ultra-low noise operational amplifier (AD 549LH). The alternating Kelvin
current is amplified by the operational amplifier, converted to an alternating voltage and used by
the lock-in amplifier and the integrator for measurement of the CPD.
Results and Discussion
61
A B
Figure 4.6. Schematic representation of the system connection for SKP (A) and SECM (B) measurements.
For measurements in the RC-SECM mode subsequent to measurements in the SKP mode, both
reed relays have to be switched, the electrochemical cell has to be filled with an electrolyte, and a
reference as well as an auxiliary electrode has to be connected to the bipotentiostat (Fig. 4.6 B).
Switching of both reed relays results in the electrical connection of the tip electrode to the
bipotentiostat as working electrode 2 (WE 2) and connection of the sample to the bipotentiostat as
working electrode 1 (WE 1).
4.2.2. Mechanical components of the integrated SKP-SECM system
4.2.2.1. Design and development of the measurement head
The heart of the combined SKP-SECM system is the measurement head (Fig. 4.7). It is made
of stainless steel (DIN 1.4571), has 3 mm thick walls for effective shielding of the pre-loaded
piezo actuator and it consists of three parts (#1, #2, #3) for a facilitated exchange of the SKP-
SECM electrode. Inside the housing, the pre-loaded piezo actuator (#10), a cylindrical holder for
the piezo actuator (#9), a translator (#8), a PVC electrode holder (#7), a conventional SECM tip
(#5) and a 0.1 mm thin stainless steel membrane (#6) for stabilization of the electrode oscillation
orthogonal to the sample surface were placed. The pre-loaded piezo actuator is mounted on an
insulating plate (#11) made of PEEK in order to enable its electrical isolation from the grounded
measurement head. Additionally, the pre-loaded piezo actuator is centered inside of the
measurement head by means of the cylindrical holder made of fiber-glass reinforced Teflon®.
Power supply to the piezo actuator is established via a double shielded cable and a two pin
connector (#12). Connection between the pre-loaded piezo actuator and the working electrode is
Results and Discussion
62
established via a cylindrical translator made of a stiff material (PEEK) in order to ensure a stable
on-axes electrode oscillation. The outer diameter of the translator is fitting well to the opening in
the upper part (#1) of the measurement head and is specially made for stabilization of the
electrode oscillation orthogonal to the sample surface. An additional tool for the stabilization of
the tip oscillations is a brass guide (#4) with an opening that fits well to the working electrode
diameter. To avoid unwanted friction between moving parts during operation, a drop of
lubricating oil is used. A connection from the working electrode to the high-quality ground of the
bipotentiostat was established via a reed relay placed in a small aluminum case at the side of the
measurement head. To position the measurement head above the sample it was mounted in a
special holder fixed on a precision linear stage (LTM 60) driven by a 2-phase step motor. The
installed step motor offers 200 single steps, with every single step additionally divided into 200
micro steps which lead to a smallest possible movement increment of 32.5 nm.
Figure 4.7. Scheme of the measurement head: main body (#1, #2, #3), electrode guide (#4), SECM tip (#5), stainless
steel membrane (#6), PVC electrode holder (#7), translator (#8), holder for the pre-loaded piezo actuator (#9), pre-loaded piezo actuator (#10), insulating plate (#11), connector for the HF power supply (#12).
The technical specification provided by the manufacturer for the stepper motor revealed a
maximum allowed load of 1.5 kg. Based on this, and taking into account the recommended load
reserve of 20 %, the maximum allowed weight of all components mounted on three stepper
motors used in the integrated SKP-SECM system was set to 1.2 kg. The geometry and wall
thickness of the SKP-SECM system measurement head were designed in such a way as to allow
the overall mass of the assembled measurement head (incl. piezo actuator and etc.) to be limited to
0.98 kg. As the holder of the measurement head is made of aluminum and with a weight of 0.1 kg,
the overall load on the stepper motor was kept below 1.2 kg.
Results and Discussion
63
A B C D
4.2.2.2. Assembly procedure for the SKP-SECM measurement head
Figure 4.8 shows the assembly procedure for the SKP-SECM measurement head. This starts
with the mounting of the SKP-SECM tip on the translator (#8) using a PVC holder (#7). In the
next step part #2 is screwed onto part #1 and a Cu wire connected to the Pt wire inside the SKP-
SECM tip is guided through the small opening in the side wall of part #2. After that a thin
stainless steel membrane with a hole in the middle, designed to fit very well to the SKP-SECM
tip, is placed on the top part #2 between two pads of PVC. In the last step, part #3 is screwed onto
part #2 in order to squeeze the membrane between the two pads (an additional O-ring is placed on
top of the upper pad) to additionally stabilize the SKP-SECM tip oscillation by the brass guide.
The three parts #1, #2 and #3 have 6 small slots uniformly distributed on the side wall surface
close to the screw thread. These slots are fitted to an especially designed home-made instrument
(spanner) to allow a secure mounting of all three parts (Fig. 4.9). To securely connect two parts
together they first have to be screwed by hand and then further tightened with two spanners
moved in opposite direction to each other.
Figure 4.8. Photographic images of single SKP-SECM measurement head parts (A – part #1, B – SKP-SECM tip mounted on top of the translator, C – part #2 screwed onto part #1, D – part #3 screwed on top of part #2).
Figure 4.9. Photographic images of the custom designed home-made spanner used to securely mount the three parts of the integrated SKP-SECM system measurement head.
Results and Discussion
64
During mounting of part #3 onto part #2, part #3 should be precisely centered over the SKP-
SECM tip to prevent damage to the very thin top part of the tip. Part #3 is then moved carefully
towards part #2. To enable this careful manipulation an auxiliary device was developed
(Fig. 4.10).
The auxiliary device is made of brass, has a cylindrical form with an inner diameter fitted to
the outer diameter of the SKP-SECM measurement head and has two sets of borings uniformly
distributed in a line on the side walls. The auxiliary device can then be screwed together with part
#2 using two short screws. Part #3 can then be fixed with two long screws in the upper part of the
device, which has three notches (two horizontal and one vertical). Movement of two long screws
inside the notches leads to a corresponding movement of part #3. As both parts (part #2 and the
set-up) are perfectly aligned to each other, part #3 is automatically aligned to part #2 and can be
mounted in a secure way.
Figure 4.10. Schematic diagram of custom designed auxiliary device for the secure mounting of the SKP-SECM system measurement head.
4.2.2.3. Design and development of the sample holder unit
In order to minimize the time dependent stray influence coefficients and thus the stray
capacitance [47], the SKP-SECM system was designed in such a way that the tip electrode is at
low impedance (earth potential) and the sample under investigation is connected to an ultra-low
noise operational amplifier (OA). The sample holder is shown in figure 4.11. Because of the
small value of the Kelvin signal, the amplifier has to have high input impedance. For this purpose
the use of a field effect transistor or a resistor was proposed [210]. The use of an input resistor of
150 MΩ was decided. The OA and all the electrical components of the amplifier are located on
one single circuit board (#7) placed in a grounded aluminum case (DIN 3.1645) of 5 mm thick
walls in order to enable effective shielding. The aluminum case consists of two parts. On top of
Results and Discussion
65
the upper part (#4) an insulating plate made of PVC (#3), a grounded shielding plate made of
brass (#2) and a sample holder made of stainless steel (#1, DIN 1.4571) were placed.
Figure 4.11. Scheme of the sample holder: sample holder (#1), shielding plate (#2), insulating plate (#3), upper (#4) and lower part (#5), P-611.3 NanoCube positioning unit (#6), circuit board (#7),
contact pin (#8) and cylindrical guide (#9).
The contact between the OA and the sample holder is established via a thin contact pin (#8)
isolated from the aluminum case with a cylindrical guide made of PVC (#9). For very precise
movement of the sample and very precise control of the distance between tip and sample, a P-
611.3 NanoCube XYZ-nanopositioning system (#6) with sub-nm lateral resolution was placed
inside the lower part (#5). This multi-axis nanopositioning unit is equipped with a stiff, zero-
stiction, zero-friction guiding system which provides motion with an ultra-high resolution and
settling times of only a few milliseconds.
Based on the technical specification the maximum load of the NanoCube should not exceed
1 kg. Taking into account the recommended load reserve of 20 % the maximum allowed weight of
all components mounted on the NanoCube was limited to 0.8 kg. The design, geometry and the
wall thickness of the sample holder were adjusted so that the overall mass of the sample holder
was 0.75 kg.
Figure 4.12 shows a photograph of the assembled SKP-SECM system. To eliminate electrical
noise the integrated SKP-SECM system was placed inside a grounded Faraday cage made of
1 mm thick zinc plated steel sheets.
Results and Discussion
66
Figure 4.12. Photograph of the integrated SKP-SECM system.
4.2.2.4. Design of the vibration damping unit
To effectively insulate the integrated SKP-SECM system from ground vibrations a special
combined jig was developed (Fig. 4.13). This jig consists of three main parts (A, B and C) with
each part adjusted to insulate from vibrations in a predefined frequency range. Part A is a robust
four-foot frame (#1) welded from 50 x50 mm steel profiles with a height adjustment unit (#2)
built in every foot to level the entire construction. Each upper corner of the frame has a square slot
(#3) with 50 mm high borders for a cylindrically shaped insulating unit (#4). This unit is made of
special kind of rubber that enables an effective insulation of ground vibrations at frequencies
below 10 Hz. The main function of the part A is to maintain the entire set-up and compensate
irregular ground vibrations (people walking through the lab etc.).
Part B is directly mounted on top of four rubber units and consists of a 60 mm thick rectangular
concrete plate (#5) with an opening in the center. Inside this opening a container (#6) of 2 mm
thick stainless steel sheet is placed. To increase the overall mass of part B the container is filled
with sand to an overall mass of 60 kg. Each upper corner of the concrete plate has a height
adjustable platform (#6, 50 mm x 50 mm) with a 15 mm thick composite plate (#7) mounted on it.
This composite plate enables an effective insulation of ground vibrations in an intermediate
frequency range (20 Hz to 60 Hz).
On top of the composite plates a rectangular piece of polished granite (#8) with an overall mass
of around 110 kg is placed (part C). The integrated SKP-SECM system is mounted on the base
Results and Discussion
67
plate (#9) directly on the top of the granite piece with four height-adjustable units (feet). The
upper part of each foot has a screw thread for assembly with the base plate. The lower part has the
form of a sharp cone to create a very small contact area between the base plate and the granite
piece. This design enables effective insulation of ground vibrations in the high-frequency range.
Figure 4.13. Schematic diagram of the vibration damping unit used as the basis for the integrated SKP-SECM system (1 – robust four-foot frame, 2 – height adjustment unit, 3 – square
slot, 4 – insulating unit, 5 – concrete plate, 6 – container, 7 – composite plate, 8 – granite plate, 9 – base plate)
For more detailed information about all mechanical parts used in the integrated SKP-SECM
system (technical drawings with exact dimensions and materials used) refer to chapter 9.1.
4.2.3. Electrical components of the integrated SKP-SECM system
All electrical components used in the integrated SKP-SECM system can be divided into
custom build devices (Kelvin current amplifier and integrator) and commercially available instru-
ments (HF power supply, oscilloscope, lock-in amplifier, step motor controller card, NanoCube
controller card and analog-to-digital conversion card).
4.2.3.1. The design of the Kelvin current amplifier
The design of the Kelvin current amplifier (KC amplifier) was developed based on the
longtime experience of the collaboration partners at the Max-Plank Institute for Iron Research
Results and Discussion
68
(Fig. 4.14). The key component of the KC amplifier is a state of the art ultra-low noise operational
amplifier (OA) AD549LH which offers superior noise characteristics such as an input current
noise of 0.11 fA / √Hz for 1 kHz bandwidth and an input voltage noise of 35 nV / √Hz for 1 kHz
bandwidth. The OA is operated as an I/U-converter where the input current signal (Iinput) is
converted to a proportional voltage output signal (Uoutput). The relationship between the input
current and the output voltage is described by the equation 4.4.
** * ∗ * (4.4)
Equation 4.4. Relationship between the input current and the output voltage
(Uoutput – output voltage, Iinput – input current, R – transimpedance).
To convert a very small current to a measurable voltage, a very large R is needed. However, the
transimpedance has a direct influence on the bandwidth of the I/U-converter and has to be
carefully adjusted to enable the desired bandwidth of the I/U-converter. To maintain a large
transimpedance and wide bandwidth the use of a wide bandwidth operational amplifier and a low
junction capacitance resistor are essential. The input stage of the I/U-converter should have a
much lower leak and bias current than the input current signal [218].
Figure 4.14. Electric circuit diagram of the Kelvin current amplifier.
To adjust the desired bandwidth of the I/U-converter two resistors of 150 MΩ each
(Rtotal = 300 MΩ) connected in series and in parallel to two capacitors of 2.7 pF (Ctotal = 1.35 pF)
connected in series are connected in parallel to the KC amplifier between the negative input and
the output of the OA. An additional function of both capacitors is to protect the OA from
Results and Discussion
69
unwanted output signal oscillations at high operational frequencies. The 100 kΩ resistors
connected to the positive and negative inputs of the OA are used to protect the OA against
occasional high input currents which may lead to unit damage. The power supply of the
operational amplifier is established via two high-precision voltage controllers which offer a DC
voltage of 5 V. To additionally stabilize the applied voltage two capacitors of 10 µF are integrated
between the power supply channels (+5 V and -5 V) and the high quality ground. The
combination of 10 nF capacitor and 100 kΩ resistor connected to the positive input of the OA
functions as a low pass filter to eliminate high frequency noise from the “U-Steuer” signal
(compensation voltage + additional AC signal for the height regulation system). The 220 Ω
resistor connected to the output of the OA protects the amplifier output from overload in case of
short-circuit between the OA input and the ground (i.e. contact of the Kelvin probe with the
sample).
4.2.3.2. The design of the integrator
The integrator (Fig. 4.15) consists of individual units such as a low pass filter unit (#1), a
modulator unit (#2), an integration unit (#3) and a control unit for the Kelvin current
amplifier (#4). The integrator also has an analog panel meter (#5) for on-line control of the
measured Volta potential difference or applied compensation voltage, three switches to control the
operation modes (#6, #7, #C) and 8 BNC connectors to connect to other electronic devices in the
integrated SKP-SECM system.
The integration unit consists of a voltage controller (#A), a time constant regulating unit (#B)
and an operation mode switch (#C). The main function of the integrator is to provide a measured
CPD value based on the lock-in amplifier signal. The output channel of the first lock-in amplifier
(LIA 1) is connected to the input channel of the integrator (BNC #6) and the signal provided by
LIA 1 contains information about the magnitude of the Kelvin current. After filtering with a low
pass filter unit (#1) designed with a cutoff frequency of 2.5 Hz the signal is processed by the
integration unit (#3). According to the value and polarity of this signal, a corresponding compen-
sation voltage (U0) is applied until the LIA 1 signal is negated (automatic mode of operation).
To operate the integrator in automatic mode the following switches should be activated: #6 –
“SKP” position, #C – “Auto” position, #7 – “Potential” position. The measured CPD value will be
indicated on the display of the panel meter and recorded via the integrator output channel (BNC
#1) by the analog-to-digital conversion card. The main function of the low pass filter unit (#1) is
to separate the LIA 1 signal, which contains information about the Kelvin current magnitude,
Results and Discussion
70
from the additional alternating signal for the height regulation system that enables proper
operation of the integration unit. The modulator (#2) is used to add to the applied compensation
voltage an additional alternating voltage for the height regulation system provided by the second
LIA (LIA 2) which has a predefined frequency and amplitude.
Figure 4.15. Electric circuit diagram of the integrator: low pass filter unit (#1), modulation
unit (#2), integration unit (#3), control unit of the Kelvin current amplifier (#4), panel meter display (#5), system operation mode switchers (#6, #7 and #C), voltage
controller (#A) and time regulating unit (#B).
The output channel of the integrator (BNC #2, “U-Steuer”) is connected to the input channel of
the Kelvin current amplifier to provide the compensation voltage and an additional alternating
voltage signal to the operational amplifier for height regulation. To protect the OA from being
accidentally damaged by a too high alternating voltage signal a 10:1 ratio attenuator is integrated
into the modulation unit. The main functions of the KC amplifier control unit (#4) are to provide a
fast check of the OA proper function and to indicate any accidental short-circuit between the
Kelvin probe and the ground (contact between the tip and sample). For this switch #6 should be
activated in position “SKP”, switch #C in position “Hand” and switch #7 in position
”Überwachung”. In the case of proper OA function the applied compensation voltage should be
Results and Discussion
71
displayed on the screen of the panel meter. However, in case of an OA malfunction or short-
circuit between the Kelvin probe and the ground, the displayed potential will be around 14 to
15 Vpp and independent from the compensation voltage set with the voltage controller. If the
voltage value displayed on the panel meter screen is larger than the one set with the voltage
controller but smaller than 14-15 Vpp, then this is an indication of a high-impedance short-circuit
between the Kelvin probe and the sample most likely caused by a drop of condensed water.
An additional function of the integrator is as a power supply for the KC amplifier provided by
the built-in 12 V DC power supply unit. However, due to the better signal to noise ratio obtained
from using two high-precision voltage controllers that each offer 5 V, the power supply unit was
disconnected from the KC amplifier and was used only as a power supply for other electric
components of the integrator.
For more detailed information about custom build devices used in the integrated SKP-SECM
system (circuit diagram with specification of electric components) refer to chapter 9.2.
4.2.3.3. The high-frequency power supply
The high-frequency (HF) power supply LE 150/100 EBW was chosen due to its ability to
provide an alternating high-voltage signal (up to 150 Vpp) in a frequency range needed for
optimal operation of the pre-loaded piezo actuator. Figure 4.16 represents the performance of the
HF power supply as a function of the alternating signal frequency under a load of 2.4 µF equal to
the capacitance of the pre-loaded piezo actuator.
The calculation of the needed HF power supply performance was done on operational data
provided by MPIE from conventional SKP systems. The maximal Kelvin probe oscillation
frequency and the maximal Kelvin probe oscillation amplitude were set to 2 kHz and 30 µm
correspondingly.
As the pre-loaded piezo actuator has a maximal stroke of 60 µm with an applied voltage of
150 Vpp and taking into account the recommended middle stroke operation range of 15 µm to
45 µm, then the maximum alternating voltage needed is 112 Vpp (corresponding to a piezo
actuator stroke of 45 µm). Based on the performance curve provided in figure 4.16 the desired
voltage of 112 Vpp can be achieved by the HF power supply in a frequency range up to 2.1 kHz.
The desired DC current offset, superimposed to the external signal (important in the case of
actuator operation in the middle stroke range), can be adjusted manually and the offset value is
displayed on the LCD screen. The amplitude of the resulting alternating voltage applied to the
piezo actuator can be also controlled manually. One remarkable difference between this HF power
Results and Discussion
72
supply and others available on the market is that the external signal is not applied symmetrically
to the DC current offset signal (equal parts of the alternating signal are applied above and below
the DC current offset signal), but the full amplitude of the external signal is applied below the DC
current offset signal. This is very important to take in to the consideration when calculating the
tip-to-sample distance where the Kelvin probe is closest to the sample surface.
Figure 4.16. The performance of the used HF power supply under 2.4 µF load (red cross indicates
the position on the performance curve corresponding to the potential of 112 Vpp and 2.1 kHz).
4.2.3.4. The lock-in amplifier
Since the original magnitude value of the Kelvin current is in a fA range, a sensitive device
which is able to amplify and recover a signal in the presence of overwhelming noise background
is needed. An analog dual phase lock-in amplifier EG&G 5210 was chosen because this
instrument offers superior performance in a frequency range from 0.5 Hz to 120 kHz, has a
voltage sensitivity from 100 nV to 3 V, a current sensitivity from 10 fA to 3 µA and output time
constants from 1 ms to 3 ks. One of the advantages of this lock-in amplifier is the possibility to
manually control all of the instrument settings using the instrument front panel. The integrated
SKP-SECM system utilizes two EG&G 5210 lock-in amplifiers connected according to the
scheme presented in figure 4.17. The output of the Kelvin current amplifier is connected to the
current input of the LIA 1. The KC amplifier amplifies the Kelvin current (alternating current) and
converts it to the corresponding alternating voltage. The signal is then filtered by the band pass
filter of the lock-in amplifier according to the oscillation frequency of the internal frequency
generator, used as a signal source for the HF power supply (i.e. the frequency of Kelvin probe
Results and Discussion
73
oscillation). Since every mechanical and electrical component of the integrated SKP-SECM
system has its own time constant, the alternating signal provided by the KC amplifier has a
different phase in comparison to the alternating signal provided by the LIA 1 frequency generator.
Thus, it is very important to perform the automated phase shift correction procedure before the
actual measurement in the SKP operation mode is performed. The LIA 1 then measures the
magnitude of the phase shift corrected signal and provides this data to the integrator.
Figure 4.17. Scheme of the lock-in amplifiers connection within the integrated SKP-SECM
system (BNC 3, BNC 5 and BNC 6 are inputs of the integrator, AD 3 is the AD input of the ADC card).
The main function of LIA 2 is to generate the additional alternating signal for the height
regulation system and to provide information about the magnitude of the feedback signal. The
frequency generator output of LIA 2 is connected to the input of the modulator unit (BNC #3) and
provides an alternating signal with a frequency of 10 Hz and amplitude of 1.768 Vrms. One should
pay attention to the fact that the values of applied alternating signals displayed on the LCD
displays of both LIAs are presented in Vrms (the root mean square of the values for one time
period of the sine wave). To avoid the disorder in voltage units recalculating all voltage values to
Vpp (peak-to-peak, 1 Vrms = 1.414 Vpp) is recommended. The signal generated by LIA 2 is
processed by the modulator unit (combined with the compensation voltage signal) and provided to
the Kelvin current amplifier. After passing the operational amplifier and going through LIA 1 it
returns to LIA 2. The phase shift corrected magnitude of this signal is proportional to the tip-to-
sample distance and is used by the height regulation system. For more details about the settings
and working parameters of both LIAs refer to chapter 4.3.1.
Results and Discussion
74
4.2.3.5. Peripheral components Step motors controller card
To position the SKP-SECM tip above the sample surface three step motors were utilized in the
integrated SKP-SECM system. For fast and precise control of all motors a PCI-SM32 motor
controller card provided by OWIS (Staufen, Germany) was employed. This is a PCI card with a
build-in microprocessor for continuous and independent control of each step motor, calculation of
the required movement length and processing of signals submitted by optionally connected end
position controllers. The step motor controller is protected against electrostatic charge to a certain
degree.
NanoCube controller card
A P-611.3 NanoCube XYZ-nanopositioning system provided by PI (Karlsruhe, Germany) with
a lateral resolution in a sub-nm range was implemented into the integrated SKP-SECM system to
provide very precise and accurate positioning of the Kelvin probe tip / working electrode above
the sample surface. The operation of the NanoCube is performed by the E-760.3SV NanoCube
controller card also provided by PI which is directly mounted on the main board of the PC. This
controller card is equipped with three low-noise piezo amplifiers for the independent operation of
each piezo element and three position servo-controller circuits based on strain gauge sensors (SGS
sensors) for closed-loop control of the actual position of each piezo element. All functions of the
E-760.3SV NanoCube controller card are accessible via the computer-bus interface. Operation of
the NanoCube controller card can be performed with the software module implemented into the
software written in-house for operation of the integrated SKP-SECM system.
Data acquisition board
To enable very fast and accurate acquisition of the data provided by the integrator, both lock-in
amplifiers (SKP operation mode) and the bipotentiostat (SECM operation mode), a PCI-DAS6014
data acquisition board (DAQ board) provided by Measurement Computing Corporation (MA,
USA) was implemented into the integrated SKP-SECM system. The PCI-DAS6014 is a 16
channel 16-bit DAQ board with 8 digital I/O and two analog outputs which provide up to 200 kHz
data acquisition rate (sampling rate) for analog inputs or up to 10 kHz for digital outputs. It offers
four input ranges from ±50 mV up to ±10 V and can be controlled by the software written in-
house for the operation of the integrated SKP-SECM system.
Results and Discussion
75
4.3. Performance of the integrated SKP-SECM system
In this section the setting procedure and operational aspects of the developed SKP-SECM
system required for reliable measurement of the CPD are described. Additionally, the influence of
a number of factors / parameters on the measured CPD value will be discussed in detail. It will be
shown, how the variation of the tip-to-sample distance can influence the measured CPD value.
Approach curves to the sample surface are recorded before the actual measurement is started
using the online control of the tip-to-sample distance based on the application of an additional
alternating voltage. This can provide reliable information for the measurement of the CPD at a
constant and predefined tip-to-sample distance. To ensure the reliability of the determined CPD
values measured with the developed SKP-SECM system, the calibration of the system was
performed using a set of pure metal reference samples (Ni, Cu and Pt). Data from literature for
similar samples were used for comparison. The nulling technique, which was implemented into
the developed SKP-SECM system, relies on the measurement of the CPD based on the
compensation of the generated Kelvin current by application of a compensation voltage. Thus, the
actual measurement of the CPD is performed at the equilibrium point at which the signal-to-noise
ratio is far away from the optimum one. To estimate the influence of the parasitic signals a set of
measurements using the off-null technique was performed.
4.3.1. Settings and first tests of the developed SKP-SECM system
The developed SKP-SECM system is a rather complex set-up and contains many independent
electrical devices (two lock-in amplifiers, a high-frequency power supply and an integrator). To
enable the proper function of the complete set-up, each of these devices has to be adjusted in
accordance with the requirements of the other components of the overall system. As the lock-in
amplifiers process the main analytical signal provided by the Kelvin current amplifier and delivers
it to the integrator for the CPD measurement, the proper set-up of both lock-in amplifiers is the
priority. A complete set of optimized settings for the operation of the integrated SKP-SECM
system using a 25 µm glass insulated Pt SKP-SECM electrode will be discussed for each lock-in
amplifier (LIA 1 and LIA 2). For better overview of the LIA 1 settings, all activated functions and
settings will be grouped according to the sections on the device front-panel.
Results and Discussion
76
Section “Sensitivity”
• The output of the Kelvin current amplifier is connected to the “A” input of the LIA 1 and a single-ended voltage mode is activated.
• The sensitivity is set to 10 mV. • Option “Float” is activated which leads to a connection of the input connector shell to
the chassis of the device via a 1 kΩ resistor. This option is very useful for reducing noise signals caused by ground loops.
Section “Filters”
• Activation of option “BP” leads to operation of the built-in filter as a band-pass filter with a frequency of the internal oscillator.
• Activation of options “F” and “2F” leads to the operation of the line frequency
rejection filters at 50 Hz and 100 Hz simultaneously. • Activation of option “Track” leads to automatic tuning of the signal channel filter to
the reference frequency of the internal oscillator. • The signal at the “SIG MON” BNC connector corresponds to that provided by the
Kelvin current amplifier which is already filtered by the band-pass filter of the signal channel filter of LIA 1. This is very useful to control the shape and magnitude of the Kelvin current amplifier signal by means of an oscilloscope.
Section “Display 1”
• The parameter “OSC LVL” is set to 0.728 V which leads to an oscillation amplitude of the Kelvin probe of about 12 µm due to the DC offset of the HF power supply of 90.5 mV.
• The parameter “OSC F” is set to 1200 Hz which leads to the generation of a reference
signal with a frequency of 1200 Hz by the internal oscillator. The Kelvin probe oscillation frequency is equal to the frequency of the reference signal.
• The “+90 ” key adds 90 to the previous setting of the reference phase shifter each
time it is pressed. By proper setting of the reference phase shifter the input signal and the reference signal generated by the internal oscillator are brought into phase at the phase sensitive detector. In this case, a maximum positive indication on Display 1 will be reached. Only in this case a reliable value of the CPD will be provided at the output channel of the integrator. However, performing calibration measurements with reference samples of known CPD is highly recommended.
Results and Discussion
77
Section “Reference”
• Activation of option “INT” leads to an adjustment of the reference channel to the output of the internal oscillator.
Section “Output”
• The CH 1 output of the LIA 1 is connected to the input channel “A” of the LIA 2 and to the input channel of the integrator (BNC #6). When the proper settings are made in section “Display 2”, the option “X Y V” should be activated. The signal provided at the CH 1 output represents the magnitude of the input signal measured by LIA 1 and is indicated in mVrms units.
Section “Display 2”
• The time constant is set to 10 ms. This setting provides the optimal ratio between the operational speed of the LIA 1 and stability of the signal at the CH 1 output.
• Activation of the option “12 dB” leads to the roll-off rate of the output filter to be set
to 12 dB / octave which provides a better filtering performance. • When the option “NORM” is activated the dynamic reserve is set to normal operation
mode. The dynamic reserve and output stability are tradeoff parameters, so the option “NORM” gives an output stability of 50 ppm / ˚C.
• The signal at the “OSC OUT” BNC connector is the reference signal generated by the
internal oscillator. This signal is provided to the input channel of the HF power supply.
The optimized settings of the LIA 2 will be discussed in a similar way as in the case of LIA 1,
however, without additional explanation of the meaning of each single option.
Section “Sensitivity”
• The sensitivity is set to 1 V. • Option “Float” is activated.
Section “Filters”
• Options “BP”, “F”, 2F” and “Track” are activated.
Results and Discussion
78
Section “Display 1”
• The parameter “OSC LVL” is set to 1.768 V and the parameter “OSC F” is set to 7.89 Hz. This leads to the generation of an alternating signal with a magnitude of 2.5 Vpp by the internal oscillator. This signal is used for the tip-to-sample distance regulation during the operation of the combined SKP-SECM system in the SKP mode.
Section “Reference”
• Option “INT” is activated.
Section “Output”
• The CH 1 output of the LIA 2 is connected to the AD channel of the analog-to-digital conversion card.
Section “Display 2”
• The time constant is set to 300 ms.
• The options “12 dB” and “NORM” are activated.
• The signal at the “OSC OUT” BNC connector is provided to the input channel of the integrator (BNC #3).
The high-frequency power supply implemented into the combined SKP-SECM system uses the
LIA 1 alternating signal, provided by the internal oscillator, as the input signal. This signal is
amplified by the HF power supply and used as the actuation signal for the pre-loaded piezo
actuator. It was found that the output signal of the HF power supply depends not only on the input
signal, but also on the manually adjusted DC offset (Table 1).
Table 1. The HF power supply DC offset and the corresponding piezo actuator offset. The corresponding HF power supply output signal was measured at a 1 Vpp input signal.
Results and Discussion
79
The HF power supply output signal presented in Table 1 was measured with a 1 Vpp input
signal at different DC offset. As the corresponding output signal of the HF power supply is
represented in Vpp, the recalculation of the LIA 2 signal (which can only be set in Vrms) should be
considered to enable the oscillation of the Kelvin probe tip with desired amplitude.
Reliable measurements of the CPD in the SKP operation mode of the combined SKP-SECM
system could not be performed immediately after switching on the system and setting all
operational parameters. Figure 4.18 represents the measured CPD and the distance dependent
signal both measured over a clean reference sample (polished piece of pure Ni) as a function of
the data acquisition time using the glass insulated 25 µm Pt SKP-SECM tip. Switching the
integrator to the automatic mode leads to a steep drop of the measured CPD value followed by a
slow increase over 75 min until the signal is stabilized and the drift of the measured CPD value is
below 0.5 mV per min.
Figure 4.18. CPD (A) and distance dependent signal (B) measured over a clean reference sample
(polished piece of pure Ni) represented as a function of the data acquisition time using a glass insulated 25 µm Pt SKP-SECM tip. The equilibration of the integrated
SKP-SECM system needs more than 75 min.
The main reason for the observed signal drift could be electrostatic charging of the insulating
glass sheath at the very top of the Pt SKP-SECM electrode. The distance dependent signal
exhibited a similar tendency of signal drift over time and the stabilization of the HR signal was
reached after 75 min. Thus, to enable reliable CPD measurements using the combined SKP-
SECM system with the glass insulated Pt SKP-SECM electrodes, an equilibration of over 75 min
is necessary.
After equilibration of the combined SKP-SECM system, a controlled approach of the 25 µm
glass insulated Pt SKP-SECM tip to the surface of the reference sample (a polished piece of pure
A B
Results and Discussion
80
Ni) was performed. To decrease the approach time, the starting position of the tip is preset as
close as possible to the sample surface. As the approach to the sample surface is usually
performed by the operation of the integrator in the manual mode, only the change of the distance
dependent signal (HR signal) is observed (Fig. 4.19). Decreasing of the tip-to-sample distance
leads to an increase of the HR signal up to contact with the sample surface. As soon as the contact
is established, a very abrupt decay of the HR signal occurs. This indicates the moment where no
further approach to the sample surface should be performed. The relative z-axis position of the tip,
at which contact was registered, is defined as “zero position” and used as the starting point for tip
withdrawal to the desired distance. Thus, based on the recorded approach curve the desired tip-to-
sample distance during subsequent measurements in SKP operation mode can be adjusted
precisely.
Figure 4.19. Approach curve to the surface of the reference sample (a polished piece of pure Ni) performed
with the 25 µm glass insulated Pt SKP-SECM tip in the SKP operation mode.
The contact between the tip and the sample surface for a few seconds leads to an overload of
the operational amplifier and an incorrect function of the combined SKP-SECM system which
causes an erroneous measurement of the CPD. Figure 4.20 represents the measured CPD and the
distance dependent signal (measured HR signal) both recorded as a function of time over the
reference sample following contact of the SKP-SECM tip with the sample surface for a few
seconds. The “saturation” of the operational amplifier leads to erroneously negative values of the
measured CPD. The drift of the CPD after contact with the sample surface was observed over a
long period of time and stabilization of the combined SKP-SECM system required more than
2.5 h until the change of the measured CPD value was <0.5 mV per min. During this time no
reliable measurements in the SKP mode could be performed. The stabilization of the measured
HR signal was observed after 20 min. Thus, to minimize the effect of the OA saturation, the
Results and Discussion
81
B
approach towards the sample surface has to be performed with a low speed and the contact time
between the Kelvin probe tip and the sample surface has to be minimized.
Figure 4.20. CPD (A) and measured HR signal (B) recorded as a function of time over the reference sample (a polished piece of pure Ni) after the contact of the SKP-SECM tip with the sample
surface for a few seconds. After contact a time interval of 165 min is required for the equilibration of the integrated SKP-SECM system.
Modification of the sample holder, described in chapter 4.3.6.1, could minimize the effect of the
OA saturation and led to much shorter stabilization times for the combined SKP-SECM system
after contact between tip and sample. In some cases, no saturation of the AO was observed.
The approach to the surface of the Ni reference sample when performed by operation of the
integrator in the automatic mode (CPD is continuously measured) revealed the dependence of the
measured CPD value on the tip-to-sample distance. To minimize the influence of this effect on the
measured CPD during scanning in the SKP mode, the tip-to-sample distance should be
continuously monitored and kept constant. To control the tip-to-sample distance an additional
signal (Um sin ω2t), where the frequency ω2 is much lower than the oscillation frequency of the
Kelvin probe tip (10 Hz vs. 500 Hz), was added to the backing voltage (U0). It has been shown by
Bonnet et al. [217], that the corresponding alternating signal demodulated by a lock-in amplifier at
the frequency ω2 is only dependent on the distance between the Kelvin probe and the sample. To
investigate the influence of the additional alternating signal (applied HR signal) on the
measurement of the CPD itself, the CPD over the reference sample was monitored continuously
while the magnitude of the applied HR signal was changed from 0 mVpp to 250 mVpp (Table 2).
The CPD measured without addition of the alternating signal is presented as 0 mV and used as
a reference. All measured values of the CPD after application of the additional signal with a
corresponding magnitude are presented relative to the reference value. Application of an
additional alternating signal influenced the measured CPD and leads to an approximately 35 mV
A
Results and Discussion
82
more negative CPD value than without addition of the HR signal. The magnitude of the applied
HR signal had almost no influence on the variation of the CPD value changes, but it influenced
the resulting magnitude of the measured HR signal with an almost linear relationship. Higher
magnitudes of the applied additional signal lead to a higher value of measured HR signal and
allowed control of the tip-to-sample distance with a higher precision.
Table 2. Magnitude of the applied additional alternating signal, the corresponding CPD and the magnitude of the measured HR signal all recorded over the Ni reference sample using a 125 µm Pt SKP-SECM tip
after equilibration of the system. The CPD measured without addition of the alternating signal is presented as 0 mV (reference value), all other CPD values are represented relative to
the reference value. Grey color indicates the optimal condition.
The maximum magnitude of the additional alternating signal which can be applied using the
EG&G 5210 lock-in amplifier implemented into the combined SKP-SECM system is 2 Vrms.
Recalculation of this value into the peak-to-peak voltage and processing the resulting signal by the
10:1 ratio attenuator built-in to the integrator will lead to an alternating signal with a maximum
magnitude of 283 mVpp. Thus, to perform a reliable measurement of the CPD under precise
control of the tip-to-sample distance, a reference sample has to be used and an additional HR
signal with a magnitude of 250 mVpp has to be applied. The representation of the CPD measured
over the sample versus the CPD of the reference sample neglects the influence of the applied HR
signal on the relative value of the measured contact potential difference.
To investigate the influence of the frequency of the applied HR signal on the measured contact
potential difference, the CPD was continuously monitored over the Ni reference sample while the
frequency of the added alternating signal was changed from 5.67 Hz to 9.87 Hz (Table 3). At
frequencies below 5 Hz an oscillation of the measured CPD value with a magnitude above 1 mV
was observed which is not acceptable. It was found that an increase of the frequency of the
applied HR signal leads to a slight change of the measured CPD value. However, this could be
caused by the increase of the HR signal frequency and/or by the ongoing stabilization of the
combined SKP-SECM system.
Results and Discussion
83
Table 3. Magnitude of the applied additional alternating signal, the corresponding CPD value and the magnitude of the measured HR signal recorded over the reference sample. The CPD value measured without
addition of the alternating signal is presented as 0 mV (reference value), all other CPD values are represented relative to the reference value. Grey color indicates the optimal condition.
Additionally, it was found that the frequency of the applied HR signal has an influence on the
maximum value of the measured HR signal just before contact of tip and sample. A set of
approach curves to the Ni reference sample was recorded with a glass insulated 125 µm Pt SKP-
SECM tip while changing the frequency of the applied HR signal (Fig. 4.21). The increase in the
frequency of the applied HR signal from 5.67 Hz to 9.87 Hz leads to a decrease of the maximal
value of the measured HR signal by more than 20 % from 640 mV to 500 mV. Thus, to enable
high stability of the measured CPD and, at the same time, to maintain high sensitivity of the
distance control system, all further measurements were performed by applying an alternating HR
signal with a frequency of 7.89 Hz and a magnitude of 250 mVpp.
Figure 4.21. Approach curves to the surface of the Ni reference sample performed with a 125 µm glass insulated Pt SKP-SECM tip in the SKP mode using different
frequencies of the applied HR signal.
One of the key components of the combined SKP-SECM system is the integrator. As the
integrator can be operated in an automatic mode, which continuously measures the CPD by
Results and Discussion
84
applying a backing voltage (U0) to compensate the generated Kelvin current, it has a built-in unit
for the control and setting of the time constant (TC). The TC of the integrator defines the speed at
which the applied backing voltage is changed. As the integrator is based on a feedback loop
circuit, a too small time constant, i.e. too fast change of the applied U0, will lead to an oscillation
of the measured CPD value as a stable equilibration point (value of U0) will not be reached. On
the other hand, if the TC of the integrator is too big then a long time is required to reach a stable
equilibration point. To investigate the influence of the TC of the integrator on the stability of the
CPD measurement, a set of CPD measurements was performed over the Ni reference sample after
equilibration of the system using a 25 µm glass insulated Pt SKP-SECM electrode while changing
the TC setting of the integrator from 1 to 12 (Table 4). If the TC of the integrator is below 4 no
stable value of the CPD could be measured. In the case of TC larger than 11 no stabilization of the
measured CPD value was reached after 270 s.
The time constant of the integrator plays a more important role in the measurement of the CPD
during a line or an area scan in the SKP mode. In this case the TC of the integrator should be
small enough to enable a fast and reliable measurement of the CPD, but not too high as this leads
to a long measurement time. It was found, that a TC setting of 6 offers an optimal compromise
between the stability of the measured CPD value (< 1 mV) and the measurement time per single
measurement point.
Table 4. TC and the corresponding CPD measured over the Ni reference sample using a 25 µm glass insulated Pt SKP-SECM tip after the equilibration
of the system. Grey color indicates the optimal condition.
Results and Discussion
85
A B
4.3.2. Calibration of the combined SKP-SECM system
To evaluate the performance of the combined SKP-SECM system and to ensure the reliability
of the measured CPD values, calibration of the system was performed with a set of reference
samples and a 25 µm glass insulated Pt SKP-SECM tip. Polished pieces of pure copper, nickel
and platinum (99.999 %) were used as reference samples (Fig. 4.22 A). In the case of the Pt
sample, a 1 mm Pt wire was pressed into a stainless steel body to reduce the overall cost of the
sample (bright horizontal line). The combined SKP-SECM system was operated in the SKP mode
with an optimized set of settings described in chapter 4.3.1. Based on the approach curves to the
reference samples the tip-to-sample distance was adjusted to 5 µm. Before every measurement the
“Auto Phase” function was used separately on each lock-in amplifier to automatically set the
reference phase shifter responsible for the adjusting of the input signal and the reference signal
generated by the internal oscillator of each lock-in amplifier. Proper setting of the reference phase
shifter leads to the maximum magnitude of the Kelvin current indicated on Display 1 of LIA 1 and
to a maximum magnitude of the HR signal indicated on Display 1 of LIA 2.
After equilibration of the system a set of line scans with 25 measurement points per line was
performed over three randomly chosen areas of each reference sample. The average values of the
measured CPD are presented in figure 4.22 B (red columns) relative to the CPD of Ni (reference
value taken as zero).
Figure 4.22. Photographic image of reference samples (polished pieces of pure Ni, Cu and Pt) used for the calibration of the developed SKP-SECM system (A). Comparison of the Volta potential values of
reference samples obtained using the integrated SKP-SECM system (red columns) with the data reported by Schmutz et al [191] (green columns for similar samples
measured using a commercial SKPFM system (B).
Results and Discussion
86
The literature values of the CPD obtained on similar samples using a commercial SKPFM
system were used for a comparison (green columns). The standard deviation of the CPD values
measured over the bulk metal samples was in all cases below 10 mV, which indicates a high
reproducibility for the measurements. The obtained values of the CPD of the bulk metal samples
correlate well with literature data and verify the reliability of the measurements of the CPD using
the developed SKP-SECM system. The deviation in CPD values for bulk Pt sample obtained
during the calibration measurement and the data provided by Schmutz et al. [191] could be
attributed to differences in the sample preparation procedures which lead to variations in the
sample surface conditions.
The observed increase of the measured CPD values (from Ni towards Pt) is in good correlation
with the well-known increase in the nobility of these metals. This proves the correct setting of the
integrated SKP-SECM system and confirms the correspondence of the potential value provided by
the integrator with the CPD of the sample under investigation without the need for additional
changes of the polarity.
4.3.3. Off-null measurement of the Volta potential
The measurement of the Volta potential difference could also be performed using the so called
“off-null” technique developed by B. Ritty [206]. This technique is based on the application of
discrete values of backing potentials (U0) ranging from -2.5 V to +2.5 V and measuring the
corresponding values of the Kelvin current (IKelvin). The statistical linear correlation-regression
analysis of the measured Kelvin current as a function of the applied U0 allows the determination
of the original CPD value as the value which corresponds to the crossing point of the linear
function fitted to the experimental data and the x-axis (i.e. the value of the -U0 which corresponds
to the zero Kelvin current). As reported by the author a good linear correlation coefficient r2 of
higher than 0.9999 and errors of less than 10 mV could, generally, be obtained.
Originally, this technique was developed to increase the lateral resolution by the application of
small area Kelvin probe tips while the classical nulling technique reached limitations caused by
the poor signal-to-noise ratio in the case of tips with reduced dimensions. The main advantage of
this technique is that the determination of the CPD could be performed with a good signal-to-
noise ratio and not at the equilibrium point where this ratio is the worst one. Thus, a comparison
of the contact potential difference measured using the off-null technique (best signal-to-noise
ratio) and the CPD values determined using the nulling technique (worst signal-to-noise ratio) can
Results and Discussion
87
provide information for the estimation of the influence of the parasitic signals on the measurement
of the CPD.
To perform the measurement of the CPD using the off-null technique some components of the
integrated SKP-SECM system were reconnected. The output channel 1 of LIA 1 was connected to
the AD channel of the analog-to-digital conversion card (ADC card) and an additional connection
was established between the input of the Kelvin current amplifier and the DA channel of the ADC
card. Thus, a set of predefined backing potentials could be directly applied to the Kelvin current
amplifier via the ADC card and the corresponding magnitude of the resulting signal could be
recorded simultaneously. To enable a reliable measurement of the CPD the integrator must be
switched into the manual operation mode and a potential of zero volt should by applied using the
built-in voltage controller. As the signal provided by the Kelvin current amplifier is modulated
with the frequency of the HR signal used for the tip-to-sample distance control, applying the HR
signal should be avoided. Otherwise, a rather high fluctuation of the measured magnitude of the
resulting signal leads to an erroneous measurement of the CPD. Before the measurement of the
CPD was performed, the tip-to-sample distance was preset to 5 µm based on an approach curve
recorded using the nulling technique.
The measurement of the Volta potential difference over the Ni reference sample was performed
using a 25 µm glass insulated Pt SKP-SECM tip by applying a set of potential pulses between -
1.9 V and +1.9 V via the ADC card. The number of pulses was varied between 5 and 9 and the
values of applied potentials were equidistantly distributed over the whole range. Figure 4.23
represents an example of the measurement of the Volta potential difference using the off-null
technique.
Using a statistical linear correlation-regression analysis the measured value of the Volta
potential difference was calculated to be -187 mV. For comparison a measurement of the CPD
over exactly the same position of the reference sample at the same tip-to-sample distance was
performed using the nulling technique and the value of the CPD was determined to be -168 mV.
As the applied HR signal influences the measurement of the Volta potential difference, the control
of the tip-to-sample distance was deactivated during this experiment. The observed difference
between the CPD measured with these two different techniques is about 19 mV. A slight
difference between the Volta potential difference measured using the off-null and the nulling
techniques indicates a small influence of a parasitic signal. Despite the slight deviation in the CPD
measured using the two different techniques, the representation of the contact potential difference
measured over the sample versus the CPD of the Ni reference sample neglects the influence of a
Results and Discussion
88
parasitic signal on the relative value of the measured CPD. Additionally, it was found that the
linear correlation coefficient r2 is bigger than 0.998 in all cases, independent of the number of
applied pulses. To reduce the measurement time needed for a single point measurement up to
2.5 s, the optimal number of applied pulses should be equal to five and the length of each potential
pulse should be reduced to 500 ms.
Figure 4.23. Measurement of the Volta potential difference over the Ni reference sample using a 25 µm glass insulated Pt SKP-SECM tip and the off-null technique. The number of applied potential
pulses is 9, length of every potential pulse is 500 ms, r2 is 0.999.
The application of the off-null technique in combination with the developed SKP-SECM
system enables measurement of the CPD at the best signal-to-noise ratio and provides very
reliable results. However, the utilization of the off-null technique has a few disadvantages. First of
all, the statistical linear correlation-regression analysis, despite its high complexity in the case of
an area scan with a few hundred measurement points, requires a completed set of measurements to
be proceeded. Thus, online monitoring of the Volta potential difference is not possible and the
result of the measurement can be analyzed only after the experiment is finished. Secondly, the
required measurement time is at least two times larger than for the nulling technique. Finally, as
the additional alternating signal, needed for the proper function of the tip-to-sample distance
control, has a big influence on the measured CPD value, no online control of the tip position
above the sample surface can be performed. This limits the application of the off-null technique
and makes the investigation of flat samples solely possible using the software based tilt correction
procedure performed before the measurement was started.
Results and Discussion
89
4.3.4. Test measurements with 125 µm / 25 µm Pt SKP-SECM electrodes
To evaluate the performance of the SKP-SECM system the Type B test sample (100 nm Pt
layer partially deposited on a 200 nm W layer) was used. It was fixed on the sample holder with
conductive carbon filled cement (Leit-C). An additional drop of the conductive glue was placed
on the edge of the test sample in order to establish direct electrical contact between the metal layer
and the sample holder. On top of the test sample an O-ring (Din = 7 mm) made of rubber was fixed
with a nail lacquer as a cell for later SECM experiments. During SKP measurements a piece of
pure Ni (99.999 %) fixed on the sample holder with the same conductive glue was used as a
reference sample. The upper side of the reference sample was ground with 1500 grit abrasive
paper and polished with different grades of alumina paste (3 µm, 1 µm, 0.3 µm and 0.05 µm) to
obtain a mirror finish (Fig. 4.24). Before each experiment the reference sample was cleaned using
the same procedure as described for the test sample.
Figure 4.24. Photographic image of the sample holder with Ni reference sample and mounted test sample for measurement in the SKP mode of the integrated SKP-SECM system.
After a controlled approach of the 125 µm Pt SKP-SECM tip to the test sample, a scan in the
SKP mode was performed over the sample area marked in figure 4.25 A. During the scan the
CPD was recorded as a function of the x- and y-positions of the tip. The tip-to-sample distance
was controlled by the addition of an alternating signal to the backing voltage. As the signal
demodulated by the lock-in amplifier is only dependent on the distance between the SKP tip and
the sample, keeping this signal constant by varying the position of the SKP tip with an integrated
z-positioning system allowed the area scan to be performed with a constant tip-to-sample distance.
The absolute tip-to-sample distance was then calculated using an approach curve.
Results and Discussion
90
A
As the scanning Kelvin probe is a very sensitive technique and the results of the measurements
may be affected by many factors such as acoustic noise or electromagnetic fields caused by the
piezo actuator, by electromagnetic and electrostatic fields arising from the stepper motors and
cables, the measured CPD values have to be validated against a standard. Thus, in order to
compare the measured CPD values obtained with the SKP-SECM system a Ni plate was added as
reference sample. Figure 4.25 A represents a line scan over the marked sample area performed
with the SKP-SECM set-up in the SKP operation mode. The CPD of different metals measured
vs. Ni as a reference sample are available in [191] (Fig. 4.25 B). A ∆CPD between W and Pt of
about 350 mV was detected. The measured ∆CPD between W and Pt is around 340 mV and
agrees well with the literature data [191].
Figure 4.25. SKP line scan over the marked area on the test sample (A) and the CPD of different
metals measured vs. Ni as a reference sample from [191] (B).
Immediately after measurements in the SKP mode the reed relays were switched to a position
enabling the connections as shown in figure 4.6 B for SECM measurements. All experiments
were carried out in a four-electrode cell configuration with the test sample as working electrode
one (WE 1), the SECM tip as working electrode two (WE 2), a Pt-wire as counter electrode and
Ag/AgCl/ 3 M KCl as reference electrode. The electrochemical cell comprised an O-ring placed
on the test sample and was filled with 150 µL of 50 mM phosphate buffer (pH 7) (Fig. 4.26).
A control program developed in Visual Basic 6.0 was used for fast data acquisition and control
of all settings. SECM experiments were performed with x- and y-increments of 150 µm and at a
tip-to-sample distance of 5 µm. Control of the tip-to-sample distance was performed based on a
set of recorded approach curves towards the sample surface. In order to avoid a background
current shift caused by an uncompensated tilt between the scanning plane of the SECM tip and the
A
B
Results and Discussion
91
test sample a software based tilt correction procedure was used which is in detail described in
chapter 4.7.2.1. In order to avoid complete oxygen depletion at the tip a multipotential pulse
profile was applied to the tip while the sample was continuously polarized to its predefined
constant potential. The first potential pulse (P1, 0 V vs. Ag/AgCl) is a conditioning potential
applied for 500 ms during which no oxygen reduction takes place at the tip. The main function of
P1 is the restoration of the diffusional equilibrium after movement of the SECM tip during
scanning. The second potential pulse (P2, -0.6 V vs. Ag/AgCl) is the measurement pulse applied
for 400 ms which is sufficiently cathodic to invoke oxygen reduction.
Figure 4.26. Photographic image of the electrochemical cell for measurements in the SECM operation mode of the integrated SKP-SECM system. The electrochemical cell consists of the sample
holder, Ni reference sample, test sample and is filled with electrolyte solution.
Figure 4.27 shows a line scan over the marked sample area from the W coated area to the Pt
coated area performed in the redox-competition mode of the SECM. Two areas with different
catalytic activity towards the ORR are clearly visible. The transition of the current correlate well
with the transition of the CPD values from the SKP measurement.
After the initial functioning of the SKP-SECM system was proven using 125 µm Pt SKP-
SECM electrode an area scan on the same test sample was performed in the SKP mode and
subsequently in the RC-SECM mode using a 25 µm Pt-disk electrode with a displacement
increment of 10 µm (Fig. 4.28). To enhance the electrochemical contrast between W and Pt an
additional oxygen injection pulse (P2, 1.2 V vs. Ag/AgCl) was applied before the measurement
pulse (P3, -0.6 V vs. Ag/AgCl). Single line scans over the marked area crossing the W to Pt
coated edge clearly demonstrated the increased resolution. The resolved width of the W/Pt edge
was estimated as the distance between two crossing points of three tangents of the recorded CPD
Results and Discussion
92
Pt W
values plotted as a function of the x-axes position of the tip during the measurement. In the case of
the SKP measurement the width of the transition between W and Pt layers was around 120 µm
while it was around 20 µm in the RC-SECM mode. Optical microscopy revealed an outer
diameter of about 120 µm for the glass sheath at the very top of the tip. After these experiments
were repeated a couple times with the same results, the observed difference in the resolved width
of the same edge between W and Pt layers was attributed to the influence of the insulating glass
sheath at the very top of the Pt SKP-SECM electrode affecting the resolution of the integrated
SKP-SECM system in the SKP operation mode. In contrast, the resolution in the RC-SECM mode
is dominated by size of the Pt-disk itself.
Figure 4.27. Line scan over the marked area of the sample surface performed in SKP (black triangles) and RC-SECM (blue circles) modes with a 125 µm SKP-SECM tip and displacement increments of 150 µm.
The tip-to-sample distance was 2.5 µm for SKP and 5 µm for SECM measurements.
Figure 4.28. Line scans over the marked area of the sample surface performed in SKP (black triangles) and in RC-SECM (blue circles) operation modes with a 25 µm Pt SKP-SECM electrode and displacement increments
of 10 µm. The tip-to-sample distance was 2.5 µm in case of the SKP measurement and 5 µm in case of the SECM measurement.
W Pt
Results and Discussion
93
To prove this assumption a set of line scans across the test sample were performed using 25 µm
Pt disk electrodes with predefined outer diameters for the glass sheath at the very top of the Pt
SKP-SECM electrodes (110 µm, 180 µm and 260 µm). All measurements were performed at a
constant tip-to-sample distance of 3 µm using a polished piece of pure Ni as a reference sample.
Results of this measurement are presented in Table 5. In all three cases the resolved width of the
edge between W and Pt layers is slightly bigger than the outer diameter of the glass sheath. This
could be explained by the presence of an electrostatic charge on the disk shaped insulating glass
surface at the very top of the Pt SKP-SECM electrode which is influencing the lateral resolution
of the integrated SKP-SECM system in the SKP operation mode.
Table 5. Dout of the glass sheath at the very top of the Pt SKP-SECM electrode and the corresponding resolved width of the edge between W and Pt layers.
4.3.5. Compensation of the electrostatic charging of the Pt SKP-SECM electrodes
As shown in previous section, the lateral resolution of the integrated SKP-SECM system in the
SKP operation mode is influenced by electrostatic charging of the insulating glass surface at the
very top of the Pt SKP-SECM tip. To remove or compensate this effect two strategies were tested.
The first strategy was based on the property of water saturated air to discharge electrostatic
charges on glass surfaces. To prove the efficiency of this approach a new sample holder design
was developed. An additional modification of the sample holder unit enabled an increase in the
relative air humidity to 95.5% by purging water saturated air through the compact closed chamber
placed around the sample. The second strategy to compensate the electrostatic charging was the
application of an external compensation voltage on a thin Ag layer deposited only on the side
walls of the insulation glass sheath of the Pt SKP-SECM tip. The technical realization and the
results of both strategies will be presented and discussed in detail in chapter 4.3.6.
Results and Discussion
94
4.3.6. Application of water saturated air
4.3.6.1. New design of the sample holder
To overcome the influence of electrostatic charging of the glass sheath at the very top of the Pt
SKP-SECM electrode on the lateral resolution of the integrated SKP-SECM system the property
of water saturated air to discharge the electrostatically charged glass surface was tested. As the
relative air humidity around the sample should be kept constant and be above 95 %, a new sample
holder design was developed. The new sample holder is based on a 2 mm thin brass plate with a
1.5 cm2 area and is directly connected to the operational amplifier input via a thin contact pin
equipped with a small spring (Fig. 4.29).
Figure 4.29. Photographic images of the sample holder (A), two inlets for water saturated air (B) and grounded aluminum chamber (C).
The main function of this spring is to protect the rather brittle Pt SKP-SECM tip from being
damaged by accidental contact with the sample surface. Additionally, this sample holder design
enables the electrical connection length between the sample and the input of the operational
C
A B
Results and Discussion
95
amplifier to be minimized, which plays a crucial role in the reduction of the noise level. An
additional grounded Al chamber was placed around the sample holder to create a closed
environment for H2O saturated atmosphere above the sample. This chamber has two inlets for
water saturated air. The air was saturated by purging air through three gas-wash bottles filled with
deionized water. An additional cylindrical brass shield was placed around the sample holder to
avoid vibration of the sample holder caused by purging with the H2O saturated air.
The influence of the H2O saturated air on the humidity level close to the sample surface was
monitored with a miniaturized humidity sensor (Fig. 4.30). After starting the air pump the relative
humidity level inside the chamber close to the sample surface reached a constant value of about
96 % within one hour.
Figure 4.30. Relative humidity inside the chamber close to the sample surface measured with a miniaturized humidity sensor.
4.3.6.2. Influence of H2O saturated air on the SKP-SECM system operation
The presence of water saturated air around the sample surface influenced the main measured
values recorded during operation of the integrated SKP-SECM system in the SKP operation
mode. The first experiments revealed a strong influence of H2O saturated air on the distance
dependent signal (HR signal) observed immediately after the air pump was switched. Purging of
water saturated air leads to a sharp increase of the measured HR signal while switching off the air
pump leads to a steep decrease of the HR signal. As the sample holder is surrounded with a
protective shield, the increase of the HR signal should not be caused by movement of the sample
holder towards the Pt SKP-SECM tip, which could conceivably arise as a consequence of the
interaction between the sample holder and the air flow.
Results and Discussion
96
A B
One possible reason for the described HR signal behavior could be the hydrophobicity of the
carbon filled conductive cement (Leit-C) used to fix the sample on the sample holder. The uptake
of water from H2O saturated air may lead to swelling of the carbon filled cement layer and by this
to a decrease of the tip-to-sample distance. However, the increase in the HR signal could also be
caused by an increase in the relative air humidity around the sample after initiation of pumping.
To test this, the carbon filled adhesive material was exchanged with a hydrophobic silver filled
thermoplastic resin (Acheson 1415). The pump was switched on and off consecutively without
changing the tip-to-sample distance (Fig. 4.31).
Figure 4.31. Influence of water saturated air on the HR signal measured over the reference sample fixed on the
sample holder with carbon filled conductive cement (A) and silver filled thermoplastic resin (B). Red arrows indicate switching of the pump.
Based on the approach curve to the sample surface, the increase of the HR signal from 52 mV
to 100 mV in the case of the carbon filled conductive cement would correspond to a lowering of
the tip-to-sample distance by 10.6 µm (Fig. 4.31 A). However, in the case of the Ag filled
thermoplastic resin the increase of the HR signal from 96 mV to 144 mV would correspond to a
reduction of the tip-to-sample distance by 4.6 µm (Fig. 4.31 B). As the tip-to-sample distance was
kept constant using the positioning system of the integrated SKP-SECM system, the observed
increase of the HR signal in the case of the hydrophobic Ag filled thermoplastic resin could only
be caused by an increased relative air humidity which leads to a capacitance change between the
tip and the sample.
It is clear that both the change of capacitance and swelling of the carbon filled material are
responsible for the observed increase in the HR signal. Thus, to avoid the unwanted change of the
tip-to-sample distance caused by the presence of the H2O saturated air, the hydrophobic Ag filled
thermoplastic resin was used for all further measurements.
Results and Discussion
97
The observed increase in the HR signal caused by purging of water saturated air has a positive
aspect for the tip-to-sample distance regulation. Figure 4.32 shows approach curves to a polished
piece of pure Ni performed both with and without purging of H2O saturated air using the same
25 µm Pt SKP-SECM electrode for each measurement.
Figure 4.32. Approach curves to the reference sample surface recorded both with (black line) and without (red line) purging of water saturated air through the chamber around the sample.
Purging of water saturated air leads to maximum HR signal value of 505 mV, which is reached
shortly before the contact between the tip and the sample. Without purging, the maximum HR
signal value was only 160 mV. Thus, high relative air humidity inside the chamber leads to an
almost three times higher value of the distance dependent signal resulting in a higher sensitivity of
the tip-to-sample distance control system.
A second remarkable effect of purging of water saturated air is a slight increase of the SKP-
SECM system equilibration time needed to stabilize the measured CPD value, from 75 to 95 min.
The HR signal stabilization was reached after 20 min of the equilibration time. This is then
followed by a minor decrease of the signal of less than 5 mV within an hour. Figure 4.33 shows
the measured CPD and the distance-dependent signal recorded as a function of time after the air
pump was switched on.
One possible reason for the increased equilibration time of the integrated SKP-SECM system
could be slight drift of the relative air humidity around the glass insulated Pt SKP-SECM tip. As
discussed in section 4.3.6.1, measurement of the relative air humidity inside the chamber close to
the sample surface indicated that the humidity reached 96 % within an hour. However, it is still
possible, that the level of the relative air humidity was slightly increasing further over time and
hence influencing the quantity of the electrostatic charge present on the glass sheath at the very
Results and Discussion
98
B
top of the tip. As the scanning Kelvin probe is very sensitive to any electrostatic charge, even a
small change in operational conditions could cause a noticeable change of the measured CPD
values.
Figure 4.33. Stabilization of the measured CPD value (A) and the HR signal (B) after switching on the air pump in the SKP operation mode. Equilibration of the system was reached within 95 min.
The third remarkable effect of purging water saturated air is the elimination of the influence of
the tip-to-sample distance on the measured CPD. Figure 4.34 shows the measured CPD values
recorded as a function of the tip-to-sample distance both with (red triangles) and without (black
dots) purging of water saturated air through the chamber. If water saturated air was not purged
through the chamber a strong decrease in the measured CPD value of over 150 mV was observed.
Figure 4.34. Approach curves to the Ni reference sample with (red triangles) and without (black dots) purging of water saturated air recorded using a glass insulated 25 µm Pt SKP-SECM tip.
A
Results and Discussion
99
This effect could be attributed to the presence of the electrostatic charges on the insulating
glass surface at the very top of the Pt SKP-SECM tip. In the presence of water saturated air a very
small change of measured CPD values (below 2 mV per 1 µm distance) was observed. Thus, the
described modification of the integrated SKP-SECM system allowed to overcome the dependency
of the measured Volta potential difference on the tip-to-sample distance at close proximity to the
sample surface. Additionally, after this modification the noise level was decreased by up to
10 µV.
To prove the influence of the water saturated air on the measurements of the Volta potential
difference a set of calibration measurements was performed that used the same bulk metal
samples (Ni, Cu and Pt) under the same conditions with the same settings of the developed SKP-
SECM system as described in chapter 4.3.1 to calibrate the SKP-SECM system under ambient
conditions (Fig. 4.35).
Figure 4.35. Calibration of the developed SKP-SECM using a set of bulk metal samples (Ni, Cu and Pt) performed using a glass insulated 25 µm Pt SKP-SECM tip with (red columns) and without (green columns)
purging of water saturated air through the chamber around the sample. The CPD of the bulk Ni sample is taken as a reference value equal to zero.
The CPD of the bulk Ni sample was taken as a reference value equal to zero. The comparison of
the CPD values measured with (red columns) and without (green columns) purging of H2O
saturated air through the chamber around the sample revealed a very good agreement between the
two sets of measurements. This confirms the absence of any influence of the water saturated air on
the measured CPD values. At the same time, very good reproducibility of the measurements of the
Volta potential differences performed under different conditions proves a very high reliability of
using this set of bulk metal samples for calibration of the developed SKP-SECM system.
Results and Discussion
100
4.3.6.3. Lateral resolution of the modified SKP-SECM system
To demonstrate the influence of purging of water saturated air on the lateral resolution of the
modified SKP-SECM system a set of line scans was performed across the test sample using a
glass insulated 25 µm Pt SKP-SECM electrodes with predefined outer diameters of the glass
sheath at the very top of the tips (110 µm, 180 µm and 260 µm). All measurements were
performed with a constant tip-to-sample distance of 3 µm while a polished piece of pure Ni was
used as a reference sample. The results of these measurements are presented in Table 6.
Table 6. Dout of the glass sheath at the very top of the Pt SKP-SECM electrode and the corresponding resolved width of the edge between W and Pt layers.
The resolved width of the W / Pt edge was found to be equal to the outer diameter of the glass
sheath at the very top of the Pt SKP-SECM tip in all three cases. This can be attributed to the
remaining electrostatic charge present on the glass surface on the top side of the Pt SKP-SECM
electrode. Obviously, purging with water saturated air appears to be insufficient to completely
discharge the electrostatic charge on the glass sheath at the front side of the tip, despite the fact
that the measured CPD values are almost independent of the tip-to-sample distance at small
working distances.
4.3.7. Application of the external compensation voltage
An attempt to compensate the remaining electrostatic charge on the glass surface of the
insulating sheath at the very top of the Pt SKP-SECM tip was made by applying an external
compensation voltage. For this a thin silver layer was chemically deposited on the side walls of
the insulating glass sheath of the Pt SKP-SECM tip following a procedure reported by Turcu et
al. [219] (Scheme 1). After deposition, the Ag layer was connected by means of a 50 µm thin Pt
wire to an external power supply (Fig. 4.36). The front side of the tip was ground on 2000 grit SiC
abrasive paper to expose the disk shaped platinum surface. Additionally, the front side of the tip
was polished with different grades of alumina paste (3 µm, 1 µm, 0.3 µm, and 0.05 µm). Detailed
description of the deposition procedure is discussed in section 7.2.8.
Results and Discussion
101
+,-.,-./0/001 +,).
-0/0/0001 2+, -3,)4
1234567/00001+,
Scheme 1. Schematic representation of the chemical deposition procedure of the silver layer on the surface of the side walls of the insulating glass sheath at the very top of the Pt SKP-SECM tip.
Figure 4.36. Photographic image of the silver layer chemically deposited on the surface of the side walls of the insulating glass sheath at the very top of the Pt SKP-SECM tip (A). Photographic image of the electrical
connection of the silver layer with the external power supply by means of a 50 µm Pt wire (B).
Applying an external potential had an influence on the measured Volta potential difference
where the measured CPD value was found to be a linear function of the applied external potential
in the tested potential range (-1.5 V to +1.5 V). In order to investigate further the influence of the
external potential on the lateral resolution of the integrated SKP-SECM system a set of external
potentials was applied and at each applied potential a line scan was performed over the Type B
test sample (100 nm Pt layer partially deposited on top of 200 nm W layer). The results of these
measurements are presented in Table 7.
Table 7. Applied external compensation voltage and the corresponding resolved width of the edge between W and Pt layers. For these experiments the Pt SKP-SECM tip with Dout of 110 µm was used.
A B
Results and Discussion
102
It was found that, independent of the applied external potential, the resolved width of the edge
between W and Pt layers was equal to the outer diameter of the insulating glass sheath at the very
top of the Pt SKP-SECM tip. Thus, the lateral resolution of the integrated SKP-SECM system in
the SKP mode of operation cannot be improved by applying an external compensation voltage.
4.4. Optimization of the integrated SKP-SECM system
The lateral resolution of the integrated SKP-SECM system in the SKP mode of operation is
limited by the Dout of the insulating glass sheath at the very top of the Pt SKP-SECM tip. To
overcome this limitation the insulating glass sheath should be removed from the very top of the
tip. However, this will lead to the significant decay of the measured Kelvin current as the effective
area of the Kelvin probe will be reduced. To increase the signal-to-noise ratio all mechanical and
electrical components of the integrated SKP-SECM system were therefore revised.
4.4.1. Optimization of mechanical components
During optimization of mechanical components of the integrated SKP-SECM system one
major improvement was caused by a more efficient stabilization of the on-axis tip oscillation
orthogonal to the sample surface. A new elongated brass guide, placed at the very top of the SKP-
SECM measurement head, was developed. This had a larger contact area of around 28 mm2 with
the tip and stabilizes the tip oscillation at the position as close as possible to the very top of the tip
without influencing the measurements.
To improve the stabilization of the oscillating SKP-SECM tip by means of the brass guide
placed at the very top of the measurement head a high-precision drilling bit was used. In
combination with careful sorting of the glass capillaries used for fabrication of the SKP-SECM
tips the best possible fitting of the tips to the opening in the brass guide was achieved with a
precision of up to 10 µm. The viscosity of the lubricating grease was additionally optimized to
enable very smooth motion of the SKP-SECM tip inside the brass guide.
4.4.2. Optimization of electrical components
Noise reduction was the main focus in the optimization of the electrical components. For this,
all power supply cables were additionally shielded (double shielding) and the power supply cable
for the piezo actuator was triple shielded. The position of the operational amplifier (OA) inside
Results and Discussion
103
the sample holder was optimized to enable a connection as short as possible between the sample
and the input of the OA. Additionally, all conductive parts inside the Faraday cage were single
grounded to a common ground.
The power supply of the OA was decreased from 12 V to 5 V and additionally stabilized by a
high-precision voltage stabilizer. All power supply cables inside the pre-amplifier compartment
were additionally grounded. After all optimization procedures the noise level was lowered by up
to 3 µV.
4.5. Development of “glass free” Pt SKP-SECM electrodes
To overcome the influence of the glass sheath at the very top of the SKP-SECM tip on the
lateral resolution of the integrated SKP-SECM system a new tip design became necessary
allowing a significant decrease of the diameter of the glass sheath while simultaneously providing
a disk-shape electrode tip for the anticipated SECM measurements. In this section, the discussion
will focus on the development of “glass free” Pt SKP-SECM electrodes. Technical solutions used
for removal of the glass sheath at the very top of the tip and the reduction of the outer diameter of
the exposed Pt wire will be presented. The final step in the fabrication procedure of the insulated
“glass free” Pt SKP-SECM electrodes is the coating of the Pt wire side walls with a thin layer of
insulating material. An overview of tested materials and the performance of the best one will be
discussed.
4.5.1. The concept of “glass free” Pt SKP-SECM electrodes
The concept of “glass free” SKP-SECM tips is based on etching of glass insulated Pt disk
microelectrodes in concentrated HF (40 %) leading to dissolution of the glass sheath at the very
top of the tip for about 250 µm to 300 µm and partial exposure of the Pt wire (Fig. 4.37). If the
desired outer diameter of the “glass free” Pt SKP-SECM electrode is smaller than the original
diameter of the Pt wire, electrochemical etching and subsequent electrochemical micropolishing
can be applied to thin the Pt wire. As the vibrating capacitor technique requires two planar
surfaces in a parallel geometry, the reduction of the Dout of the Pt wire should be performed
without sharpening of the disk surface at the very top of the tip. The exposed Pt wire should be
subsequently covered with a thin layer of an insulating material that will be stable in the
electrolyte solution to be used later. Parameters such as the hydrophobicity of the insulating layer,
resistance to penetration of ions and/or molecular oxygen dissolved in the electrolyte and
Results and Discussion
104
chemical stability in solutions at a certain pH value have to be taken into account. The coating of
the exposed Pt wire should be performed in such a way that only the side walls of the exposed Pt
wire are insulated and the disk surface at the very top of the Pt SKP-SECM electrode remains
uncoated.
Figure 4.37. Schematic representation for the concept of fabricating “glass free” SKP-SECM tips (1 - etching of the glass sheath at the very top of the tip in conc. HF; 2 - electrochemical etching
of the Pt wire; 3 – electrochemical micropolishing of the Pt wire; 4 – insulation of the exposed Pt wire side walls with a thin insulating layer).
4.5.2. Etching of Pt SKP-SECM electrodes in concentrated HF
Precise control of the immersion depth of the tip into concentrated HF is of key importance to
obtain a reproducible length of the exposed Pt wire. For this, a special holder equipped with a
precise micrometer gauge was developed. A reproducible etching length of the tip of below
300 µm was achieved by controlled approach of the 25 µm Pt disk microelectrode to the HF
solution / air interface to the very first contact with the HF solution surface (Fig. 4.38).
Figure 4.38. Photographic image of a 25 µm Pt disk microelectrode after etching in concentrated HF for 10 min.
Results and Discussion
105
Immediately after the contact HF solution covered the very top part of the tip due to capillary
forces leading to a complete etching of the glass sheath around the Pt wire within 10 min.
Additional rotation of the tip leads to a decrease of the etching length to about 150 µm. However,
this may also lead to the formation of a conical gap between the glass sheath and the Pt wire.
To achieve reproducible etching of the glass sheath at the very top of Pt disk microelectrodes
glass insulated electrodes with long tips and Dout of the glass sheath of below 100 µm should be
used. In the case of microelectrodes with thicker glass sheath it is rather difficult to achieve a
reproducible etching of the tip and the length of the exposed Pt wire is difficult to control. Starting
from Dout of the glass sheath > 200 µm non-uniform etching of the tip can take place which leads
to the formation of a conical gap between the glass sheath and the Pt wire (Fig. 4.39, marked with
a red circle). Such tips can be still used; however, greater attention should be paid to ensure proper
isolation of the exposed Pt wire side walls during the next preparation step. Once a tip is damaged
(i.e. the exposed Pt wire is no longer straight or broken) it can be polished again up to the exposed
Pt disk surface and subsequently etched in concentrated HF. In general, such a “recovery
procedure” could be performed only once as polishing of the tip for the third time (2nd recovery)
will lead, in most cases, to a tip with a very thick glass sheath.
Figure 4.39. Photographic images of etched in concentrated HF Pt disk microelectrodes made using 25 µm Pt wire
(A – original Dout of the glass sheath at the very top of the tip before etching in concentrated HF ≈ 120 µm; B– original Dout ≈ 220 ). Part of the tip with a conical gap between the Pt wire and the
glass sheath is marked with a red circle.
In the case of SKP measurements the lateral resolution of the integrated SKP-SECM system is
mainly limited by the size of the Kelvin probe. Thus, to improve the lateral resolution of the
integrated SKP-SECM system a Pt wire with a diameter of a few µm should be used for the
fabrication of the Pt SKP-SECM tip or the outer diameter of the 25 µm Pt wire should be reduced
after removing the insulating glass sheath from the very top of the Pt SKP-SECM tip. The
fabrication of the glass insulated Pt disk UMEs using 10 µm Pt wire is challenging. Thus, the
B A
Results and Discussion
106
reduction of Dout of the Pt wire exposed after the etching of the tip in concentrated HF presents a
more efficient method to fabricate Pt SKP-SECM tips with the desired Dout of the Pt wire smaller
than 10 µm.
4.5.3. Reduction of the outer diameter of Pt SKP-SECM electrodes
The fabrication of sharp metallic tips is essential for techniques such as field ion micro-
scopy [220] and scanning tunneling microscopy [221]. As the desired tip geometry plays a crucial
role in high-resolution measurements, many different techniques have been proposed for the
fabrication of sharp tips with different form, length and aspect ratio [222]. These fabrication
techniques can be divided in three main groups: electrochemical etching [223], chemical
etching [224] and mechanical shaping [225]. Techniques such as ion milling [226], cathode
sputtering [227], physical vapor deposition [228], electro-beam deposition [229] and electron
beam-induced deposition [230] have also been reported for sharp tip fabrication. However, these
techniques need rather expensive equipment and are now seldom used. A 25 µm Pt wire used for
Pt UME fabrication has very low mechanical stability and mechanical sharpening of the tip cannot
be used. Thus, only chemical or electrochemical etching procedures can be applied. Each of these
techniques has its advantages and disadvantages. However, only by using electrochemical etching
one can precisely control the length of the tip by controlling the insertion depth of the tip in the
electrolyte.
Electrochemical etching and electrochemical polishing are generally well defined phenomeno-
logically and are distinctly different from each other [222]. Electrochemical etching occurs in the
lower current-density regime of the typical current-density vs. applied DC voltage relationship.
Generally, electrochemical etching leads to a fast and selective removal of material at different
local areas characterized by different parameters, such as crystallographic orientation, composi-
tion or microstructure, and results in a particularly rough topography [227]. Electrochemical
polishing takes place in the middle current-density range and tends to remove material more
rapidly from rough places leading to smoother surfaces [231]. Thus, combination of both methods
seems to be the best choice for fast and controllable fabrication of smooth sharp tips.
Since the electrochemical processes are complex, many parameters such as active electrolyte
composition and concentration, solution viscosity and solution temperature, presence of organic
solvents, type and magnitude of the applied voltage (AC, DC, pulse sequence) and material and
geometry of the counter electrode should be considered. However, there is no universally
applicable choice of parameters suitable for every material.
Results and Discussion
107
In this work only Pt wires with Dout of 25 µm were used as the starting material for fabrication
of SKP-SECM tips. To optimize the tip fabrication procedure a set of electrodes was made with
the Pt wire protruding out of the glass body for approximately 10 to 15 mm. For a detailed
description of the fabrication procedure of such electrodes refer to chapter 7.2.5.
Many different electrolytes were proposed to fabricate sharp PtIr STM tips such as: chloride
based solutions (NaCl and CaCl2) [232], NaBr [233], cyanide based solutions [234] and even
molten salt etchants [235]. Since cyanide based solutions are highly toxic and molten salts can be
hazardous to handle, only aqueous electrolytes containing NaCl or CaCl2 were tested for the
fabrication of sharp Pt tips. The first set of experiments was performed using saturated aqueous
solutions of NaCl or CaCl2 without addition of any organic solvents. A Pt mesh with a geometric
area of about 1 cm2 was placed at the bottom of the electrochemical cell to act as the counter
electrode while the Pt wire electrode was acting as the anode. An Agilent 33120A high-frequency
generator (Agilent Technologies, CA, USA) was used as source of alternating voltage and varying
etching times and applied potentials were tested. The shape of the etched tips was assessed
afterward using optical microscope. In general, lower voltages caused the formation of longer tips.
However, independent of the electrolyte composition and etching time all tips were inhomoge-
neously conically etched and therefore not useful for further SKP-SECM tips fabrication. The
inhomogeneous etching could be caused by the evolution of gas bubbles at the specimen surface
leading to a localization of the etching process by partial blocking of the wire surface combined
with a disturbed transport of fresh active solution [90].
Good results were achieved using a procedure previously described by Nam et al. [232] for
etching PtIr tips in CaCl2 aqueous solution that contains a small amount of acetone. The addition
of acetone enables control over bubble evolution and leads to a fast transport of gas bubbles from
the specimen surface into the electrolyte. Tips of the desired geometry and length were obtained
using an optimized procedure described by Libioulle et al. [236]. Following dissolution of 7 g
CaCl2 in 20 ml of water and subsequent addition of 20 ml of acetone the solution was stirred for
20 s to allow saturation of the aqueous phase with the organic solvent. Acetone was used to
improve the reproducibility of the etching procedure, as the bubbles stream in aqueous solutions
disturbs the material dissolution leading to fluctuations of the tip shape [237]. For electro-
chemical etching the aqueous phase was used as electrolyte. The immersion depth of the tip was
precisely controlled by using a holder equipped with a micrometer gauge. The tip was dipped into
the electrolyte for 150 µm while a sinusoidal alternating voltage with a frequency of 40 Hz and
amplitude of 4.3 Vpp was applied until the anodic current dropped to 120 µA. This procedure led
Results and Discussion
108
to cylindrical etching of the Pt wire yielding a long tip with a flat disk surface with a diameter of
about 15 µm at the very top of the tip and smooth parallel side walls (Fig. 4.40).
Figure 4.40. Photographic images of electrochemically etched Pt wire (A, etching was performed in 7 g. CaCl2 / 20 mL H2O / 20 mL acetone mixture) and the same Pt wire after subsequent electro-
chemical micropolishing (B, micropolishing was done in 0.5 M H2SO4 for 90 s).
Etching of the Pt wire for longer time leads to sharpening of the tip and should be avoided. To
further decrease the Dout of the etched tip the procedure for electrochemical micropolishing of Pt
tips as described by Libioulle et al. [238] was optimized. The tip was dipped into 0.5 M H2SO4 to
an immersion depth of approximately 150 µm while short anodic pulses with a frequency of
4 kHz and an amplitude of 3.5 Vpp were applied for 90 s. This procedure leads to a slow and very
homogeneous etching of the Pt wire yielding an elongated tip with a flat disk surface on the very
top, parallel side walls and a Dout of 8 µm.
4.5.4. Insulation of Pt SKP-SECM electrodes
Insulation of the exposed Pt wire side walls was a very challenging issue. The coating of the Pt
wire side walls with a thin layer of insulating material should be stable in the electrolyte of
interest. Additionally, only the disk surface at the very top of the tip should remain uncoated.
Simultaneously, the insulation should be very thin to limit the overall outer diameter of the tip, as
a larger outer diameter lowers the lateral resolution of the integrated SKP-SECM system.
Due to similar requirements for the insulation of the tips for electrochemical scanning
tunneling microscopy (ECSTM) the literature in this field was carefully reviewed. Since the
pioneering work by Sonnenfeld et al. [239] many new methods for tip insulation were described
and many potentially insulating materials were tested. These methods can be divided into three
main groups: “dip coating” methods based on dipping of the tip into the liquid insulating
A B
Results and Discussion
109
compound [240], “melt coating” methods based on the coating of the tip with melted glass [241]
or polymer [242] and methods based on electrochemical deposition of an electrophoretic paint on
the exposed tip surface [94].
Compounds such as nail varnish [243], paraffin [91] and epoxy resin [244] were tested as
insulation materials used in “dip coating” based methods for ECSTM tips fabrication. However,
all tested materials produced bulky and uneven films sometimes leading to electrolyte
contamination. Apiezon wax was frequently used as insulating material by transferring the drop of
hot wax onto the tip [245] or by raising the tip through a melted drop of the wax [246]. Some of
the described materials were also tested during this work as potential candidates for insulating the
exposed Pt wire.
More precise and localized isolation of the exposed Pt wire can be obtained by electroche-
mically induced deposition of an electrophoretic paint (EDP). EDPs were introduced in industry
as cost efficient coatings for the protection of car bodies from corrosion processes [247]. Two
different types of electrophoretic paints are available on the market: cathodic EDP based on
positively charged polymers with amino groups containing side chains and negatively charged
anodic EDP with carboxylic group containing side chains [248]. If the amino groups are
protonated (cathodic EDP) or the carboxylic groups are deprotonated (anodic EDP), than charged
polymer micelles are formed. Electrochemically induced pH-modulation of the solution in the
diffusion zone in front of the electrode surface caused by the oxidation or reduction of water leads
to pronounced changes in the polymer solubility and, finally, to precipitation on the electrode
surface. Subsequent curing of the deposited polymer layer leads to hardening of the film and the
formation of a dense insulating layer.
Anodic EDPs have been used to fabricate nanoelectrodes with active areas up to a few
nanometers. Widely used anodic EDPs are Glassophor® ZQ 84-3211 [101], its modern version
Glassophor® ZQ 84-3225 [249] developed by BASF (Ludwigshafen, Germany) and Clearclad
HSR [250] manufactured by LVH coating (Coleshill, UK). Since Glassophor® ZQ 84-3211 is no
longer available on the market, only Glassophor® ZQ 84-3225 and Clearclad HSR were tested for
insulating the exposed Pt wire. In both cases good insulation of the exposed Pt wire could be
achieved after optimization of the originally reported procedures. Cyclic voltammetry in 5 mM
[Ru(NH3)6]3+ / 100 mM KCl solution, which is commonly used to investigate fabricated
nanoelectrodes, revealed Faradaic currents in the range of tenths of pA for fully insulated Pt wire
and around 25 nA after cutting the insulated Pt wire with a scalpel. However, chronoampero-
metric experiments performed in 100 mM Na2SO4 / 5 mM NaCl solution with fully insulated Pt
Results and Discussion
110
wires polarized at -0.6 V vs. Ag /AgCl (3 M KCl) reference electrode revealed much larger
cathodic currents than expected for a 25 µm Pt disk UME. The measured currents increased over
time indicating swelling of the insulating film and an enhancement of oxygen diffusion toward the
Pt wire surface. This effect was observed for both anodic EDPs and further optimization of the
deposition / curing parameters could not solve this issue.
A set of fifteen commercially available materials were tested for their ability to be deposited on
the Pt wire surface to form a thin insulating layer and for their resistivity in 100 mM
Na2SO4 / 5 mM NaCl (Table 8). The stability of the insulation layer was investigated in a three-
electrode cell configuration with the insulated Pt wire electrode as working electrode, a Pt-wire as
counter electrode and Ag/AgCl/ 3 M KCl as reference electrode. A constant potential of -0.6 V or
a sequence of potential pulses (EPulse = -0.6 V, tPulse = 0.5 s), were applied to enable O2 reduction
at the exposed Pt surface. The O2 reduction current measured at the tip was recorded as a function
of either the time or the potential pulse duration using fast data acquisition of up to 100 kHz. Each
material was tested with a set of 3 electrodes dipped into the solution for approximately 500 µm.
Before insulation the exposed Pt wire was cleaned by subsequent dipping of the tip in to acetone,
isopropanol and water for 60 s each. Coating of the Pt wire side walls was performed by means of
a specifically developed set-up enabling the fixed tip to be moved towards a U-shaped holder
containing a small droplet of the insulating material. The set-up was placed under an optical
microscope to allow control of the coating procedure. After curing of the insulation layer the
completely insulated Pt wire was cut with a razor blade in order to expose the disk shaped Pt
surface at the very top of the wire.
The applicability of the material as insulating coating for the exposed Pt wire was evaluated
based on following criteria:
• Insulation layer quality (information about the possibility of the deposition of thin insulating film without droplets or air bubbles).
• Efficiency of the insulation (information is based on the CV obtained in 5 mM [Ru(NH3)6]
3+ / 100 mM KCl solution directly after cutting of the fully insulated Pt wire).
• Stability of the insulation (information is based on the comparison of the measured cathodic currents at the beginning and the end of the chronoamperometric and the pulse measurements).
Results and Discussion
111
Material Insulation layer
quality
Efficiency of the
insulation
Stability of the
insulation Applicability
Epoxy resin good good bad bad
Acrylic resin bad - - -
Polyurethane resin good good bad bad
Polyester resin good bad bad bad
Polystyrene / Toluene good good good good
PMMA based
photoresist good good bad bad
Polyacryl based
photoresist good bad bad bad
Sylgard 184 good good bad bad
Silicone A good good bad bad
Silicone B bad - - -
Varnish G 371 bad - - -
SX AR-PC 5000/40 good good bad bad
Apiezon Wax in
Toluene good good good good
Nail varnish A good good good good
Nail varnish B good good very good very good
Table 8. Comparison of the insulation performance of the tested materials.
A very thin layer of nail varnish B (Maybelline Jade) offered the best insulating properties
among all tested materials. Despite the rather viscous consistency of the nail varnish a thin
insulating layer could be deposited on the Pt wire surface without forming droplets or air bubbles.
Once the best insulating material was selected and the coating procedure was optimized the “glass
free” Pt SKP-SECM tip with Dout of the etched Pt wire of about 5 µm was insulated. Since the
length of the exposed Pt wire was around 200 µm and the very top of the Pt wire had the desired
Dout, cutting of the wire after insulation could not be performed. Thus, to insulate only the side
walls of the exposed Pt wire it was inserted in to the droplet of the nail varnish in such a way that
Results and Discussion
112
only the Pt wire side walls were coated and the flat disk surface at the very top of the Pt wire
remained uncoated (Fig. 4.41).
Figure 4.41. Photographic image of a “glass free” Pt SKP-SECM tip insulated with nail varnish.
Figure 4.42 represents a cyclic voltammogram of the 5 µm “glass free” SKP-SECM tip
insulated with the nail varnish in 5 mM [Ru(NH3)6]3+ / 100 mM KCl at a scan rate of 50 mV/s.
Figure 4.42. Cyclic voltammogram of a nail-varnish insulated “glass free” tip recorded in 5 mM [Ru(NH3)6]Cl3 / 5 mM NaCl at a scan rate of 50 mV / s.
The measured current of about -8.5 nA corresponds to a diameter of the exposed electrode surface
of about 8.5 µm assuming a disc microelectrode. However, the diameter estimated with an optical
microscope was around 5 µm. The difference is most probably caused by an additionally exposed
electrode surface at the side walls of the Pt wire at the very top of the tip.
A long term stability measurement was performed in 100 mM Na2SO4 / 5 mM NaCl with the
tip constantly polarized at -0.6 V vs. Ag/AgCl/ 3 M KCl as reference electrode (Fig. 4.43). A
slight decay of the measured cathodic current was observed over the first 30 min resulting in a
Results and Discussion
113
current plateau. We assume that the plateauing of the current was caused by a local pH change
resulting from the O2 reduction at the active tip surface. After 50 min a minor increase of the
cathodic current was observed indicating enhanced ORR. This could be caused by an increased
diffusion of O2 through the insulating layer and its subsequent reduction at the Pt side walls or an
initial degradation of the nail varnish at the very top of the tip caused by the local pH change.
Figure 4.43. Cathodic current measured at the insulated 5 µm “glass free” SKP-SECM tip as function of the experiment time. Experiments were performed in O2 saturated solution containing
100 mM Na2SO4 / 5 mM NaCl, Etip = -0.6V vs. Ag / AgCl (3 M KCl).
All insulated “glass free” Pt SKP-SECM tips used during combined SKP and SECM
experiments were coated with a thin layer of this nail varnish.
4.6. Optimization of working parameters
The optimization of the working parameters of the integrated SKP-SECM system was
performed using the “glass free” SKP-SECM tip in accordance with the Kelvin equation
(Eq. 4.5).
58729:; = (∆6+ 7)89+∆<
< (4.5)
Equation 4.5. Kelvin equation where IKelvin represents the Kelvin current, ∆Ψ is the CPD, U0 is the variable backing potential, ε is the dielectric constant of the medium, ω is the oscillation
frequency of the tip, A is the active tip area, ∆d is the tip oscillation amplitude, and d0 is the tip-to-sample distance.
Results and Discussion
114
Evidently, to maintain a constant Kelvin current and, at the same time, decrease the active area of
the working electrode one can increase the oscillation frequency of the Kelvin probe, increase the
amplitude of the tip oscillation, or decrease the tip-to-sample distance.
Figure 4.44 shows the magnitude of the measured Kelvin signal as a function of the tip-
oscillation frequency (red triangles) and as a function of the tip-oscillation amplitude (black dots).
Figure 4.44. Magnitude of the measured Kelvin signal as functions of the tip-oscillation frequency (red triangles) and the tip-oscillation amplitude (black dots). Both experiments were
performed using a 25 µm “glass free” Pt SKP-SECM tip.
The magnitude of the Kelvin signal is a linear function of the oscillation frequency only at
small frequencies up to 450 Hz. At frequencies above 1000 Hz a saturation of the signal was
observed. To obtain the maximum Kelvin current we decided to operate the system at 1200 Hz.
The influence of the tip-oscillation amplitude on the magnitude of the measured Kelvin current
showed a linear behavior at small amplitudes between 1.2 µm and 4.5 µm as also in the range
between 6 µm to 10 µm. At tip-oscillation frequencies above 10 µm a small increase of the
measured Kelvin signal followed by signal saturation was observed. To obtain the maximum
Kelvin signal we decided to operate the integrated SKP-SECM system with 12 µm tip-oscillation
amplitude.
4.7. Investigation of the optimized SKP-SECM system performance
This section describes the performance of the optimized SKP-SECM system and its operation
with the “glass free” Pt SKP-SECM electrodes. The consequences of removing the insulating
Results and Discussion
115
glass sheath from the very top of the Pt SKP-SECM tip is discussed based on approach curves to a
Type C test sample and line scans performed with the “glass free” Pt SKP-SECM electrodes with
different outer diameters ranging from 8 µm to 50 µm.
4.7.1. Operation of the integrated SKP-SECM system using “glass free” electrodes
Figure 4.45 represents the stabilization of the CPD value recorded over the Ni reference
sample using a “glass free” 25 µm Pt SKP-SECM tip as a function of time while purging water
saturated air through the chamber around the sample.
Figure 4.45. Stabilization of the measured CPD value while purging of water saturated air through the chamber around the sample. Equilibration of the combined
SKP-SECM system was reached within 20 min.
Elimination of the glass sheath from the very top of the tip led to a decrease in the time needed
for equilibration of the combined SKP-SECM system from 95 min (glass insulated 25 µm Pt
SKP-SECM tip) to 20 min (“glass free” 25 µm Pt SKP-SECM tip). Thus, operation of the
integrated SKP-SECM system with the “glass free” Pt SKP-SECM tips enables the reduction of
the equilibration time of the integrated SKP-SECM system by more than four times in comparison
with the glass insulated Pt SKP-SECM electrodes.
To prove the influence of water saturated air on the measured CPD value and the measured HR
signal, the pumping of H2O saturated air was switched on after equilibration under ambient
conditions. Figure 4.46 represents the measured CPD and HR signal values recorded as a function
of time.
Results and Discussion
116
A B
Figure 4.46. Influence of water saturated air pumped through the chamber on the measured CPD value (A) and on the distance dependent signal (B). Red arrow indicates the switching on of the pump.
At the beginning of the measurement the relative air humidity inside the chamber was equal to
the ambient humidity and after 9 min the pumping of water saturated air through the chamber was
initiated. This led to a rapid decrease of the measured CPD value and stabilization of the signal
was reached after about 7 min. The presence of water saturated air around the sample caused a
decrease of the measured CPD value of approximately 16 mV in comparison with SKP
measurements under ambient conditions. At the same time, the change of the relative air humidity
inside the chamber had almost no influence on the distance dependent signal (HR signal) and only
a slight increase of less than 3 mV was observed (Fig. 4.46 B). The increase of the relative air
humidity inside the measurement chamber does not lead to an increase of the measured HR signal
or to higher vertical resolution. The application of “glass free” Pt SKP-SECM electrodes enables
rapid switching between the operation of the developed system under ambient conditions and at
high air humidity inside the measurement chamber while purging water saturated air. This
dramatically shortens the equilibration time of the system after altering the experimental
conditions.
The performance of the optimized SKP-SECM system was investigated using the “glass free”
Pt SKP-SECM tips. As the insulating glass sheath was removed from the very top of the tip, the
exposed Pt wire of the “glass free” Pt SKP-SECM tips is highly sensitive to mechanical loads and
could easily be deformed by intensive contact with the sample surface. To enable precise control
of the tip-to-sample distance an approach curve to the sample surface is essential for “glass free”
tips. Thus, it is very important to precisely control the moment of contact between the tip and the
sample surface while recording approach curves. It was found, that the measured CPD value is
very sensitive to the contact of the tip with the sample surface as it causes a shortcut of the
measurement circuit. This leads to an overload of the OA and to a steep drop of the measured
Results and Discussion
117
CPD value. Figure 4.47 represents an approach curve to the surface of the Ni reference sample
recorded using a 50 µm “glass free” Pt SKP-SECM tip with a z-increment of 10 nm while purging
water saturated air through the chamber around the sample. Simultaneously, the measured value
of the CPD was recorded. The approach to the sample surface was stopped at the moment a
significant drop of the measured CPD value was observed.
Figure 4.47. Approach curve to the Ni reference sample performed using a 50 µm “glass free” Pt SKP-SECM tip with a z-increment of 10 nm (black dots) while purging water saturated air through the chamber.
During the approach the measured CPD value was recorded (red dots).
The moment of the contact between the tip and the surface of the reference sample could not be
detected based on the approach curve alone. However, a significant drop in the measured CPD
value of about 500 mV was observed over the last 20 nm of the approach distance which indicates
a shortcut of the circuit. Investigation of the Pt wire at the very top of the “glass free” Pt tip under
optical microscope revealed the absence of any mechanical damage of the tip. Thus, a rapid
change of the measured CPD value during the approach to the sample surface is a very reliable
and precise parameter for the control of the contact between the Kelvin probe and the sample.
A set of approach curves toward the Ni reference sample were recorded using Pt SKP-SECM
tips of varying diameter ranging from 8 µm to 50 µm. To obtain a better comparison of the data
with electrodes of different diameter the measured values were normalized with respect to the
maximum value of the HR signal with the signal from the tip just before coming into contact with
the sample surface taken as the reference value equal to one (Fig. 4.48). In the case of the 25 µm
and 50 µm Pt SKP-SECM tips, the normalized approach curves look similar with a slow increase
in the HR signal during the tip approach to the sample surface. However, the 8 µm tip exhibits a
very steep increase of the HR signal over 500 nm above the sample surface, this leads to a very
high-resolution of the SKP-SECM system over the z-axis. As the information about contact with
Results and Discussion
118
the sample surface in the case of the 8 µm tip can only be obtained at very close proximity to the
sample surface, one should be extremely careful during the approach. Therefore, z-increments of
5 nm are highly recommended.
Figure 4.48. Normalized approach curves to the sample surface obtained using “glass free” SKP-SECM tips with different diameters of the Pt wire (1 - Dout = 50 µm; 2 – Dout = 25 µm; 3 – Dout = 8 µm). The tip
oscillation amplitude was 12 µm in case 1 and 2, and 5 µm in case 3.
Additionally, it was found that proper operation of the integrated SKP-SECM system using 8 µm
Pt SKP-SECM tip is possible at small tip oscillation amplitudes between 4 µm and 8 µm. In the
case of a larger tip oscillation amplitude no reproducible approach curves could be obtained and
contact of the tip with the sample was registered at different values of the HR signal. This makes
the estimation of the tip-to-sample distance based on the approach curves recorded at large tip
oscillation amplitudes not reliable.
Analysis of the recorded approach curves revealed the continuous change in the measured CPD
values during the approach to the sample surface independent of the outer Pt wire diameter of the
“glass free” Pt SKP-SECM electrodes used. Table 9 represents the calculated change of the
measured CPD value per µm distance for “glass free” tips of varying outer diameters. Experi-
ments were performed both with and without purging water saturated air through the chamber
around the sample.
Surprisingly, the observed drift of the CPD was more pronounced in the case of purging water
saturated air. The decrease of Dout of the exposed Pt wire at the very top of the “glass free” Pt
SKP-SECM tip leads to stronger increase of the CPD drift up to 100 mV per one micrometer
distance in the case of a 8 µm tip. Thus, to avoid erroneous measurement of the Volta potential
difference caused by the influence of the tip-to-sample distance on the measured CPD values, the
Results and Discussion
119
operation of the integrated SKP-SECM system should be performed at a constant tip-to-sample
distance.
Table 9. Dout of the “glass free” Pt SKP-SECM tips and the corresponding calculated change of the measured CPD value per µm distance with and without purging of
water saturated air through the chamber.
4.7.2. Constant distance operation of the integrated SKP-SECM system
Reducing the outer diameter of the used “glass free” Pt SKP-SECM tips causes the measured
CPD value to be strongly dependent of the tip-to-sample distance. In order to simultaneously
enable high-resolution measurements and to eliminate the distance dependence of the measured
CPD values a very precise control of the tip-to-sample distance in combination with the constant
distance operation of the combined SKP-SECM system should be ensured. The adjustment of the
desired tip-to-sample distance based on the measured HR signal in combination with a recorded
approach curve to the sample surface was already discussed in detail. The application of the
software based tilt correction procedure or the software based feedback controller when both
implemented into the integrated SKP-SECM system could provide such precision. The technical
realization, advantages and disadvantages of each of these solutions are discussed in detail in this
section.
4.7.2.1. Software based tilt correction procedure
One of the possibilities to perform the measurement of the Volta potential difference over an
area on the sample surface at a constant tip-to-sample distance is based on the application of the
software based tilt correction procedure. This procedure should be performed before the start of
the actual measurement and is based on the recording of three approach curves to the sample
surface performed over three edges of the rectangular scanning area (points P0, P1 and P2). The z-
axes position at which the contact between the tip and the sample was observed should be saved in
a specially programmed software module. Based on this data the software calculates the plane that
describes the tilt of the sample surface (Fig. 4.49). The desired tip-to-sample distance at which the
Results and Discussion
120
tip should be retracted from the sample surface during the measurement can be freely selected.
During scanning the tip will follow the calculated plane positioned above the sample surface at the
defined tip-to-sample distance thus avoiding the risk of impact against the tilted sample surface.
Figure 4.49. Schematic diagram of the software based tilt correction procedure. The calculated plane describing the tilt of the sample surface is marked in green, the scanning plane positioned
above the sample surface at the predefined distance is marked in yellow.
The procedure for operation of the combined SKP-SECM system at a constant tip-to-sample
distance is rapid for setting the required parameters (only three approach curves to the sample
surface) and the positioning of the tip above the sample surface at a predefined distance during the
scan is therefore very quick. However, it has a few disadvantages. The most remarkable one is
that it can only be used with sufficiently planar samples, as topographical features on the sample
surface are not considered during setting of the required parameters of the scanning plane.
Additionally, in the case of the application of the “glass free” Pt SKP-SECM tip performing three
approach curves to the sample surface until contact between both is achieved possibly results in a
risk of the tip being damaged by the contact with the sample.
4.7.2.2. Software based feedback controller
An alternative solution for performing the measurements of the Volta potential difference over
an area of the sample surface at a constant tip-to-sample distance is based on the application of
software based feedback controller. Three different types of software based feedback controllers
were implemented into the integrated SKP-SECM system (a normalized feedback controller, an
absolute feedback controller and a relative feedback controller). One of the most commonly used
types of feedback controller is the relative feedback controller. This is a software based module
which controls the position of the Kelvin probe above the sample surface using the measured HR
signal. Before an actual measurement is started, the value of the HR signal corresponding to the
Results and Discussion
121
desired tip-to-sample distance and the hook movement (the retraction distance of the tip before
movement to the next measurement point) should be manually defined. Usually, this requires an
approach curve to the sample surface to be performed or an already known value of the measured
HR signal can be used. During scanning the controller enables the approach of the tip to the
sample surface until the defined value of the measured HR signal (desired distance between the tip
and the sample) is reached (Fig. 4.50, step 1).
Figure 4.50. Schematic representation of the operation principles of the software based feedback controller. The red plane indicates the defined scanning position of the tip above the sample surface. Numbers and
arrows are indicating the sequence and the direction of the tip movement (1 – approaching to the sample surface until the defined value of the measured HR signal is reached, 2 – hook
movement of the tip to the defined distance above the sample surface, 3 – movement of the tip to the next measurement point).
Only after that the actual measurement of the Volta potential difference is performed. Before
progressing to the next measurement point the tip is retracted from the sample surface by a
defined distance (hook movement, step 2) to ensure the secure motion of the tip between two
measurement points (step 3). The required length of the hook movement can be easily defined
from the approach curve and the change of the measured HR signal during movement of the tip
above two measurement points without changing the tip-to-sample distance.
To ensure the proper operation of the feedback controller a few operational parameters should
be adjusted in advance:
• Safety limit – the maximal value of the measured HR signal which can be reached while approaching to the sample surface. To avoid collision of the tip with the sample this value should be at least 25 % smaller than the maximal measured HR signal reached during an approach curve just prior to contact of the tip with the sample. As soon as the safety limit is reached, the tip will be immediately retracted to a safe position above the sample surface, the measurement will be stopped and the tip will be moved to the start point of the entire scan.
• Proportional constant (Kp) – the proportional value between the tip-to-sample distance and the corresponding change of the measured HR signal (µm / mV) for a defined part of the approach curve. It shows the distance which the tip should be moved to change the value of
Results and Discussion
122
the measured HR signal by 1 mV. The value of Kp can be calculated from the approach curve for the desired tip-to-sample distance at which the operation of the integrated SKP-SECM system is planned. To ensure smooth and precise positioning of the tip it is recommended to set the Kp approximately 5 times lower than the calculated value. Good results were achieved with the Kp equal to 0.005 while using the NanoCube for positioning of the tip.
• Feedback search increment – the length of the single movement of the tip after which the
value of the measured HR signal is compared with the desired value. If the desired value of the HR signal is not reached, the next single movement of the tip is performed. It is recommended to set this value equal to Kp or slightly larger. If the feedback search increment is smaller than Kp, no proper positioning of the tip is achieved. Good results were achieved with the feedback search increment equal to 0.005 while using the NanoCube for positioning of the tip.
• Loop delay – the waiting time of the feedback controller between two single movements of
the tip needs to provide the lock-in amplifier 2 (LIA 2) enough time to properly process and submit the value of the measured HR signal to the controlling unit of the integrated SKP-SECM system. Good results were achieved with the loop delay equal to 50 ms with a time constant of the LIA 2 equal to 300 ms.
To ensure proper operation of the feedback controller the scanning direction of the tip has to be
adjusted to the opposite direction of the sample tilt, i.e. with each following movement of the tip
the tip-to-sample distance should become smaller. If the scanning direction of the tip is adjusted to
the direction of the sample tilt, then no hook movement is required. However, in this case after the
finishing the first line scan the tip is moved to the beginning of the second line and, dependent of
the sample tilt, collision with the sample surface is possible. This issue requires further
optimization of the software module.
The most important advantage of the software based feedback controller is the possibility to be
applied to investigations of samples with different topographical features at a constant tip-to-
sample distance. As only preliminary information of the sample topography is required to set the
operational parameters, the software based feedback controller is a more flexible and reliable tool
for constant distance measurements than the software based tilt correction procedure. However,
adjustment of the tip position above the sample surface at a defined distance over each
measurement point requires more time which sometimes leads to much longer overall measure-
ment times.
Results and Discussion
123
4.7.3. Lateral resolution of the optimized SKP-SECM system
The lateral resolution of the optimized SKP-SECM system was investigated with line scans
over a Type C test sample and the resolved width of the W/Pt edge was calculated as described
earlier (Fig. 4.51). These results were normalized with respect to the maximum CPD value,
measured over the Pt part of the sample (taken as 1) and the minimum CPD value, measured over
the W part of the sample (taken as 0). All measurements were performed with a tip oscillation
frequency of 1200 Hz while the tip oscillation amplitude was varied.
Figure 4.51. Normalized line scans performed over the test sample using “glass free” Pt SKP-SECM tips with different diameters of the Pt wire (1 - Dout = 50 µm; 2 – Dout = 25 µm;
3 – Dout = 8 µm). The tip-oscillation amplitude was 12 µm in case 1 and 2,
and 5 µm in case 3. The tip-to-sample distance was 1.5 µm in case 1,
0.8 µm in case 2, and 0.1 µm in case 3.
To allow a better overview the values of the calculated resolved width of the W/Pt edge are
summarized in Table 10. The resolved width of the W / Pt edge is equal to the Dout of the used
“glass free” Pt SKP-SECM tip at a tip-to-sample distance of 1.5 µm in the case of a 50 µm
diameter tip, at a tip-to-sample distance of 0.8 µm in the case of a 25 µm diameter tip and at a tip-
to-sample distance of 0.1 µm in the case of 8 µm diameter tip.
Table 10. Dout of the Pt wire used for fabrication of the “glass free” SKP-SECM tips, the corresponding calculated resolved width of the W/Pt edge and the used tip-to-sample distance.
Results and Discussion
124
Excellent agreement between the resolved width of the W / Pt edge and the corresponding Dout
of the used “glass free” Pt SKP-SECM tip proves a very high efficiency of technical solutions
implemented into the integrated SKP-SECM system for stabilization of the Kelvin probe
oscillation orthogonal to the sample surface.
4.7.4. Vertical resolution of the optimized SKP-SECM system
To investigate the accuracy of positioning of the “glass free” SKP-SECM electrodes during the
operation of the integrated SKP-SECM system in the SKP mode a line scan of 2500 µm length
with 100 measurement points was performed over a 200 nm thick layer of W sputtered on a
surface of a thermally oxidized silicon wafer (Type B test sample) using a 25 µm “glass free” Pt
SKP-SECM tip while purging with water saturated air through the chamber around the sample
(Fig. 4.52).
Figure 4.52. Line scan performed over a piece of commercially available silicon wafer using a 25 µm “glass free” Pt SKP-SECM tip while purging water saturated air through the chamber. The tip-to-sample distance was controlled by the software based feedback controller with its set
value based on the recorded approach curve to 300 nm, the loop delay parameter of the feedback controller was 100 ms. Correction of the sample tilt was performed
manually by extracting a slope of about 25 µm from the recorded data.
To define the desired tip-to-sample distance with high precision, an approach curve to the
sample surface was performed before the actual measurement after equilibration. Based on the
recorded approach curve the distance between the tip and sample during the measurement was
adjusted to 300 nm using the software based feedback controller. To ensure a high precision of the
positioning of the tip above the sample surface, the loop delay parameter of the feedback
controller was set to 100 ms while keeping the time constant of the second lock-in amplifier at
Results and Discussion
125
300 ms. As the scanning plane of the tip is not perfectly parallel to the sample surface due to a
possible tilt, the software based feedback controller was programmed in such a way that before
movement of the tip to the next measurement point it was retracted from the sample surface by
500 nm to avoid occasional crashing of the tip into the sample surface. All other operational
parameters were set to optimal values as discussed in chapter 4.3.1.
As the sample surface was not exactly parallel to the scanning plane of the Pt SKP-SECM tip,
the recorded data was corrected according to the tilt of the sample and a slope of about 25 µm
over the 2500 µm measurement range was used. The average standard deviation of the z-position
of the tip over 10 lines is 22 nm while the difference between the maximum and minimum values
is 102 nm (indicated by two blue dashed lines). Thus, the accuracy of the tip positioning in the
SKP mode using a 25 µm “glass free” Pt SKP-SECM tip is below ±51 nm.
To evaluate the vertical resolution of the developed system a line scan of 45 µm length was
performed in the SKP mode after equilibration over the edge of a 100 nm thin structured Pt layer
deposited on top of a W layer (Type C test sample, Fig. 4.53) using a 8 µm “glass free” Pt SKP-
SECM tip with a movement increment of 1 µm while purging water saturated air (Fig. 4.54).
Figure 4.53. Photographic image of the test sample (Type C, structured Pt layer of 100 nm thickness deposited on top of a 200 nm W layer using PVD). The red rectangle marks the area on the
sample surface where the line scan over the Pt edge was performed.
To define the desired tip-to-sample distance with high precision, an approach curve to the
sample surface was performed before the actual measurement after equilibration. Based on the
recorded approach curve the distance between the tip and sample during the measurement was
adjusted to 300 nm using the software based feedback controller. To ensure a high precision of the
positioning of the tip above the sample surface, the loop delay parameter of the feedback
controller was set to 100 ms while keeping the time constant of the second lock-in amplifier at
Results and Discussion
126
300 ms. As the scanning plane of the tip is not perfectly parallel to the sample surface due to a
possible tilt, the software based feedback controller was programmed in such a way that before
movement of the tip to the next measurement point it was retracted from the sample surface by
500 nm to avoid occasional crashing of the tip into the sample surface. All other operational
parameters were set to optimal values as discussed in chapter 4.3.1.
The thickness of the Pt structure was calculated from the data obtained by the quartz crystal
microbalance (QCM) placed inside the PVD chamber close to the sample. The tip-to-sample
distance during the measurement was adjusted to 50 nm using the software based feedback
controller with an approach curve recorded before the actual measurement was started. The loop
delay parameter of the feedback controller was set to 500 ms while the time constant setting of the
second lock-in amplifier was increased to 1000 ms to enable very precise positioning of the tip.
All other operational parameters of the integrated SKP-SECM system were set to the optimal
values discussed in section 4.3.1.
Figure 4.54. Line scans performed over the edge of the test sample (Type C, 100 nm Pt layer deposited on top of a W layer) using a 8 µm “glass free” Pt SKP-SECM tip in the SKP mode while purging water saturated
air (A) and using the commercial AFM system (B). The tip-to-sample distance was controlled by the software based feedback controller and was set based on the recorded approach
curve to 50 nm, the loop delay parameter of the feedback controller was500 ms. The correction of the sample tilt was done manually by extracting
a slope of about 0.5 µm from the recorded data.
For comparison, the same sample was investigated using a commercial AFM system,
NanoWizard 3 (JPK Instruments, Berlin, Germany). For better comparison of both measurements
the average position of the tip and the cantilever recorded over the part of the sample covered with
W layer was set as zero position. Based on this the height of the Pt edge was calculated as the
distance between the average positions of the tip and the cantilever recoded over the parts of the
sample covered with different metals.
A B
Results and Discussion
127
The line scan over the edge of the 100 nm thin Pt structure in the SKP mode revealed an
average thickness of the Pt layer of about 83 nm while the distance between the maximum and
minimum position of the tip during the measurement was 105 nm. In the case of the AFM
measurement the average thickness of the Pt layer was 87 nm while the distance between the
maximum and minimum position of the cantilever was about 91 nm. Comparison of the data
obtained using both techniques clearly demonstrates very high vertical resolution of the integrated
SKP-SECM system which is, to some extent, comparable with that of the AFM system in the case
of the application of “glass free” Pt SKP-SECM electrodes enabling very precise positioning of
the Kelvin probe tip above the sample surface.
Results and Discussion
128
4.8. Application of the SKP-SECM system
In this chapter, two applications of the developed SKP-SECM system are presented. The first
example focuses on linking contact potential difference (work function) with catalytic activity
towards oxygen reduction reaction (ORR) for novel high-performance catalyst which is
subsequently probed using the integrated SKP-SECM system in the SKP and RC-SECM modes of
operation. The second example focuses on localization of a single Al2CuMg (S-phase) crystal in a
solid solution Al matrix performed in the SKP operation mode utilizing the “glass free” SKP-
SECM tips. Subsequent visualization of the in situ consumption of molecular O2 dissolved in the
electrolyte at the surface of this single crystal has been performed using the redox-competition
mode of the SECM.
4.8.1. Investigation of high-performance catalyst for oxygen reduction reaction
4.8.1.1. State of the art in ORR electrocatalysis
The rapidly growing population of the earth, intensively increasing rates of industrialization
and constantly decreasing amount of available fossil fuels as well as numerous ecological and
environmental factors attract significant attention to the development for alternative energy
sources [251]. Proton exchange membrane fuel cells (PEMFCs) are promising clean power
sources for both automotive and stationary applications. However, state of the art Pt catalysts
which are currently used at the cathode side of fuel cells, where the oxygen reduction reaction
takes place, are expensive. This is the main barrier on the way to a wider commercialization of
PEMFCs [252]. Application of a state of the art Pt-based catalyst requires approximately 0.5 mg
Pt per 1 cm2 of electrode area leading to an overall loading of 50 g Pt for a 100 kWh
vehicle [253]. However, to become economically attractive, the amount of the noble metal in a
PEMFC should be comparable to that of those platinum group metals currently used in catalytic
converters for internal combustion engines [254]. In the case of the PEMFC anode, where
oxidation of molecular hydrogen takes place, the loading could be decreased to 0.05 mg Pt per
cm2 without a significant loss of performance [255]. On the other hand, to reduce the loading of
Pt-based catalyst at the cathode side to enable wider commercialization of vehicles powered by
PEMFCs, the activity of ORR catalysts should be 4 to 10 times higher than that of state of the art
Pt-based catalysts [256]. Analysis of the polarization curves of state of the art Pt-based PEMFCs
reported by Gasteiger et al. [257] reveals a significant drop of the working potential from the
theoretical thermodynamic limit of 1.17 V to Ucell = 0.57 V corresponding to j = 1.5 A cm-1. If the
Results and Discussion
129
overpotential due to transport limitations and hydrogen oxidation is eliminated, 2/3 of this drop is
caused by the overpotential originating from the oxygen reduction reaction. Thus, it is clear that
the ORR is the main obstacle to wider commercialization of PEMFCs. Therefore, new catalysts
with an improved ORR activity are highly desired.
To develop new catalysts with improved ORR activity, it is very important to understand the
oxygen reduction reaction in detail. This is of great importance from a fundamental point of view
in order to develop and verify the theory that describes the performance of already known
catalysts. The next step in the development of such a theory would be the prediction of the
composition and properties of novel materials with enhanced ORR activity.
The reduction of a single O2 molecule requires the transfer of four electrons and four protons
leading to the formation of two water molecules (Eq. 4.6):
+ +
→ (4.6)
Equation 4.6. The overall reaction describing electrochemical reduction of a single O2 molecule.
Experimental investigation of the ORR mechanism in-situ is challenging and often controversial
due to difficulties in detecting all reaction intermediates [258]. Therefore, a simplified theoretical
analysis of the electrochemical oxygen reduction reaction uses a pathway with four single stages,
with each step describing a transfer of one proton and one electron that involves the formation of
several intermediates such as *O, *OOH, *OH (where *O, *OOH, *OH are oxygen, hydroxyl and
superhydroxyl groups adsorbed at the active sites) (Scheme 2) [254].
DFT calculations performed by Stephens and co-workers [254] for Pt3Y(111) with 25 % Y
located in the subsurface layer and ¼ ML *O pre-adsorbed on the surface reveals that ∆G1 and
∆G4 (formation of *OOH and reduction of *OH) are greater than zero. This leads to an almost
zero current at applied potentials above 1 V vs. RHE. Increase of the driving force (i.e. decrease
of the applied potential) will lead to gradual decrease of ∆G1 and ∆G4 until both terms become
negative at 0.81 V. From this model, the reduction of *OH is recognized as the “potential
determining step” while ∆G4 is the final step to ensure all ∆Gx become negative. Thus, based on
this theoretical model, an ideal ORR catalyst should bind *OH and *O intermediates slightly
weaker than Pt3Y(111). However, as Norskov and co-workers have shown [259], it is not possible
to vary the adsorption energies of the ORR intermediates (∆G*O, ∆G*OH, ∆G*OOH) independently
of each other by changing the electronic structure of the catalyst surface. This means that the
Results and Discussion
130
A B
electrode surface which binds *O stronger than Pt3Y(111) will also bind *OH or *OOH stronger
as each intermediate adsorbate binds to the active site via an oxygen atom.
Scheme 2. Schematic representation of the four steps for the reduction of one oxygen molecule (associative or peroxo mechanism) with several intermediates, such as *O, *OH and *OOH (oxygen, hydroxyl and super hydroxyl groups adsorbed at the active sites). ∆Gx represents
the free energy of a specified single stage where * denotes an active site on the catalyst surface. Indices (1) and (2) indicate water molecules that are formed as
consequence of *OOH and *OH reduction, respectively [259].
As shown by Norskov and co-workers [259], there is roughly a linear correlation between the
binding energies of *O and *OH (Fig. 4.55 A).
Figure 4.55. Trends in ORR activity of pure metals plotted as a function of both the O and the OH
binding energies (A). Volcano plot describing the ORR activity of pure metals as a function of the oxygen binding energy. Both graphs are copied from the original
work by Norskov and co-workers [original data from 259].
Thus, knowledge of the binding energy of one intermediate of the ORR reaction can enable an
accurate estimation of the binding energy of the other intermediates. Based on these data, a model
Results and Discussion
131
(volcano plot) that describes the ORR catalyst activity as a function of the theoretically calculated
oxygen binding energy (∆G*O) was developed [259] (Fig. 4.55 B). In general, volcano plots are
widely used in electrocatalysis [260 - 263] as they illustrate the Sabatier principle which
determines the optimal catalyst as a catalyst that provides a compromise between binding and
activation of the reactants. At the same time, the catalyst should not bind too strong as this can
lead to poisoning of its surface by the reaction products [264].
A widely used representation of this model utilizes the *OH binding energy as the main
catalytic descriptor. To compare the performance of ORR catalysts, the experimental activity is
measured under predefined conditions (U = 0.9 V vs. RHE) using a rotating disc electrode (RDE)
in oxygen saturated 0.1 M HClO4 solution. The experimental values are then plotted as a function
of the calculated difference ∆G*OH - ∆GPt,*OH, where ∆G*OH is the theoretically calculated
hydroxyl binding energy on the surface of the investigated catalyst and ∆GPt,*OH corresponds to a
Pt (111) single crystal surface as a reference catalytic system. Figure 4.56 shows the volcano plot
for six Pt-based catalysts with the highest activity ever reported for the ORR [265]. The best ORR
activity reported up to now was found to be on the vacuum-annealed Pt3Ni(111) single crystal
surface: this is approximately 10 times higher than that of the Pt(111) single crystal surface [266].
Figure 4.56. Volcano plot for six catalysts exhibiting the highest ORR activity ever reported (experimental enhance-
ment of ORR activity as a function of difference in hydroxyl binding energies on the catalyst surface and on the Pt(111) single-crystal surface, adopted from [265]): Pt3Ni(111) – vacuum annealed Pt3Ni(111)
single-crystal [266],Cu-Pt(111) NSA – Cu based near-surface alloy formed using the Pt(111) single-crystal [267], Pt3Y(pc) – polycrystalline sputter-cleaned Pt3Y [260], dealloyed PtCu nanoparticles [268], Pt5La(pc) – sputter-cleaned polycrystalline Pt5La single phase [254],
Pt3Ni(pc) – vacuum-annealed polycrystalline Pt3Ni alloy [269].
As can be seen from simple theoretical estimations (Fig. 4.56), the ORR activity of an optimal
catalyst (the surface of such a catalyst would bind *OH 0.1 eV weaker than a Pt(111) single
crystal surface) should be up to 40 times higher than that of Pt(111). However, a review of the six
Results and Discussion
132
catalysts which exhibit the highest ever reported ORR activity shows that the maximum
theoretical activity appears to be unachievable, i.e. the simple theoretical model should be
adjusted for catalysts located close to the optimum. For example, despite Cu-Pt(111) NSA [267] is
binding *OH almost 0.1 eV (0.111 eV) weaker than the Pt(111) single crystal surface, it has a
19 % lower ORR activity than the Pt3Ni(111) single crystal surface [266] which binds *OH
0.136 eV weaker than a Pt(111) single crystal surface. On the other hand, the surface of dealloyed
PtCu nanoparticles [268] binds *OH-intermediates 0.08 eV weaker than the Pt(111) single crystal
surface, but has 23 % lower ORR activity than a sputter-cleaned polycrystalline Pt3Y
surface [260] which binds *OH 0.06 eV weaker than the Pt(111) single crystal surface or 0.02 eV
stronger than the surface of dealloyed PtCu nanoparticles. According to the volcano plot, the
sputter-cleaned polycrystalline Pt3Y surface should have a lower ORR activity than dealloyed
PtCu nanoparticles.
Nevertheless, despite the volcano plot description [259] for the ORR activity of catalysts
focuses on only one descriptor, such as the *OH binding energy while kinetic parameters are not
taken into account, it remains one of the most reliable models for both the description of known
trends in ORR activity and for the design of new materials with improved ORR activity.
The successful design of novel catalysts requires an excellent link between theoretical conside-
rations and models (prediction), comprehensive characterization of the surfaces of resultant
materials (confirmation) and evaluation of the ORR activity of these materials (evaluation).
These three factors form a triumvirate for the development of novel catalysts with improved ORR
activity (Fig. 4.57).
Figure 4.57. Schematic representation of the triumvirate combining theoretical predictions with experimental confirmation of developed material properties and evaluation of the ORR activity as a methodo-
logical approach for designing novel materials with enhanced ORR activity.
Results and Discussion
133
A very good example of a successful application of this triumvirate is shown in figure 4.58 for
the development of a new catalyst with improved ORR activity in the work published by Stephens
et al. [267]. In this publication, the authors described the development and evaluation of the
performance of Cu-Pt(111) near-surface alloy (NSA) that exhibit ORR activity up to 8 times
higher than that of a Pt(111) single crystal surface. Prediction about possible improvements in the
ORR activity of the Pt(111) single crystal surfaces was based on theoretical calculations originally
performed by Norskov et al. [259]. The authors performed DFT calculations and evaluated the
influence of the coverage of solute Cu atoms located in the subsurface layer (2nd
Pt layer) on the
*OH binding energy for different NSA surfaces.
Figure 4.58. Schematic representation of successful application of the triumvirate for development of novel catalyst with enhanced ORR activity (experimental data are taken from [267].
It was found that approximately ½ ML Cu present in the Pt(111) single crystal sub-surface
layer weakened the *OH binding energy of the surface by 0.1 eV which would lead to enhanced
ORR activity. As shown by Rossmeisl and co-workers [270], on a completely homogeneous Cu-
Pt(111) single crystal surface, the shift in the electrode potential required to reach 1/6 ML
coverage of *OH (∆U1/6 ML OH) corresponds to a shift in the *OH binding energy (∆∆EOH) relative
to Pt(111) single crystal surface. Based on this methodology, experimental characterization of Cu-
Pt(111) NSA surfaces was performed using cyclic voltammetry measurements in oxygen free
0.1 M HClO4 solution. These measurements confirmed the DFT-predictions for Cu coverage up to
2/3 ML while it was not possible to experimentally reach higher Cu coverage than that. To
evaluate the performance of the NSAs, the standard procedure based on the application of the
Results and Discussion
134
RDE technique in oxygen saturated 0.1 M HClO4 solution was used.
Despite the superior ORR activity of the Cu-Pt(111) near-surface alloy catalysts reported by
Stephens et al. [267], further work is needed to implement this catalyst in commercially viable
PEMFCs. One of the biggest limitations is the cost of bulk Pt(111) single crystal sample used as
the starting material for the preparation of the Cu-Pt(111) NSA (1200 € per single crystal of
5 mm x 3mm, 1.3 g.). Even taking into account the 8 times increase in ORR activity of Cu-
Pt(111) NSA vs. Pt(111) single crystal surface, the overall price of such a catalyst per gram of Pt
or per cm2 of active sample surface is far too high for successful commercialization in PEMFCs.
However, application of Cu-Pt(111) NSA catalyst in PEMFCs could become economically viable
if it would be possible to produce such a catalyst not only on the surface of bulk Pt(111) single
crystals, but also on the surface of supported Pt-nanoparticles or thin films as well. As a first step
towards up-scaling and implementation of NSA catalysts in PEMFCs, it would be reasonable e.g.
to implement this catalyst at the surface of Pt(111)-like thin films formed on top of thin Pt films
that were originally polycrystalline Pt. The latter should be deposited on a conductive supporting
material which is stable under the operational conditions of PEM fuel cells. In this case, the
superior ORR activity of Cu-Pt(111) NSA catalyst in combination with the low overall loading of
Pt could provide the next step to successful commercialization of PEMFCs.
Application of cyclic voltammetry in oxygen free 0.1 M HClO4 solution is an elegant method
to experimentally probe surface properties, especially the *OH binding energy shift relative to the
Pt(111) surface based on the shift of the potential required to reach 1/6 ML *OH coverage.
However, such a reversible and predictable formation of the compact monolayer of hydroxyl ions
is possible only on a few well-oriented surfaces of Pt and Pt alloys. In the case of polycrystalline
ORR catalysts such as Pt3Y or Pt5La, which are only slightly less active than the best reported Pt-
based catalysts, adsorption of *OH is difficult to separate from the adsorption of *O which leads
to complications in the interpretation of cyclic voltammograms. Additionally, cyclic voltammo-
grams of well-oriented Pt surfaces have been studied for more than 30 years in many different
electrolytes. As a result, almost every single voltammetric feature can be correlated with a
particular physicochemical process with a high degree of reliability [271]. However, novel
polycrystalline Pt-based alloys such as Pt3Y or Pt5La have been known for just a few years. Thus,
the correlation of observed voltammetric peaks with the corresponding processes is a very
challenging issue that requires additional fundamental studies. This limits the application of cyclic
voltammetry for an experimental estimation of the *OH binding energy shift that leads to
complications linking experimental observations of the sample activity with theoretical models.
Results and Discussion
135
Thus, the use of the *OH binding energy as a universal catalytic descriptor is somehow limited to
well-known materials (mostly single crystal surfaces of Pt and Pt alloys): experimental estimation
of *OH binding energy requires reliable identification of voltammetric features that correspond to
*OH adsorption. Moreover, this methodology meets obvious difficulties for other materials such
as bulk Cu or Ag in acidic media. The latter materials do not show any voltammetric features
attributed to *OH- and *O adsorption in acidic media. This issue raises the need for a more
convenient, and ideally universal catalytic descriptor which can be experimentally determined
under the conditions of further catalyst application. Ideally, this descriptor needs to be measured
using a reliable technique accessible for many research groups worldwide.
The work function (Φ) is one of the first descriptors originally proposed by Bockris [272] and
later further evaluated by Trasatti [273] as a universal catalytic descriptor for the explanation
(depiction) of the electrocatalytic activity of different metals. In his widely cited work [273],
Trasatti proposed the use of the work function as an ideal parameter which could conceptually be
the origin of a wide range of activities or, in principle, could explain any property of metals. This
is due to the fact that the work function describes the binding energy of electrons near the Fermi
levels to the material interior. Generally, these electrons are exchanged during the occurrence of
contact potential differences, electrochemical reactions or the formation of chemisorption bonds.
One famous example of the application of the work function as a catalytic descriptor is the
hydrogen evolution reaction (HER). The representation of the exchange current density as a
function of Φ can indeed be used to describe trends in the catalytic activity of pure metals towards
electrochemical hydrogen evolution reaction in acidic solution (Fig. 4.59 A) [273]. Based on their
work function values, all presented transition and sp metals can be clearly divided into two groups
with mercury being outside of both groups. Although a significant distinction between transition
and sp metals is obvious, it does not suggest anything in terms of reaction mechanism which is
likely not the same for all metals. However, considering only the dependence of the hydrogen
evolution reaction exchange current on the heat of adsorption of hydrogen at the metal surface, a
division of metals into separate groups can be made based on a reaction mechanism leading to a
volcano plot with Pt almost at the peak of the volcano (Fig. 4.59 B). Thus, consideration of only
the free energy of electrons (nature of the metal) without any necessary reference to the detailed
nature of the reaction is sufficient to separate metals into well-defined groups while the
mechanism of the hydrogen evolution reaction is only the consequence of how the heat of
adsorption of hydrogen depends on the work function of the used metal [273]. This makes the
work function a good candidate for a universal catalytic descriptor of metals with respect to the
Results and Discussion
136
electrochemical hydrogen evolution reaction.
Figure 4.59. The dependence of the hydrogen evolution exchange current on the work function for different metal surfaces (A). When immersed into an acidic electrolyte, the metals from the group A are positively charged
while metals from the group B are negatively charged. The volcano plot for hydrogen evolution reaction represents hydrogen evolution exchange current as a function of the heat of adsorption of hydrogen
on the metal surface (B). Original data was taken from [273].
The work function of materials is a property which can be calculated by ab initio methods
[274], [275] and the work function of many materials have already been measured and tabulated
[276], [277]. However, it should be noted, that work function values of model surfaces reported in
literature were predominantly measured under UHV conditions. As a result, it is difficult to relate
these values to material activities measured in less ideal environments or under real operational
conditions.
4.8.1.2. Concept of research
The general concept of the work performed for this thesis is based on the application of the
developed SKP-SECM system to directly probe surface properties (experimental confirmation) by
measuring the CPD in the SKP operation mode. Subsequent experimental evaluation of the
surface ORR activity is performed in the redox-competition mode of the SECM over the same
area of the probed surface to elucidate the connection between surface properties and material
activities for oxygen reduction reaction (Fig. 4.60). Linking the measured CPD with the
calculated work function / binding energies for intermediates of electrocatalytic reaction using the
integrated SKP-SECM technique feeds back into theoretical modeling to provide an insight for the
design of novel catalytic materials with improved ORR activity.
A
B
A B
Results and Discussion
137
Figure 4.60. Schematic representation of the concept of application of the developed SKP-SECM system
for in-situ probing of surface properties and experimental investigation of ORR activity over the same area of the probed surface to link experimental parts of the triumvirate
with theoretical predictions.
Results and discussion presented in the following sections will focus on three main topics:
• Development of a procedure for the quick and simple fabrication of Pt(111)-like thin films on the surface of a thin polycrystalline Pt film deposited on the oxidized surface of a commercial Si wafer. The primary focus of this fabrication procedure is to prepare the sample with a high coverage of well-ordered Pt(111) terraces. This procedure is neces-sary to provide well-defined and reproducible surfaces for SKP-SECM measurements (Chapter 4.8.1.3).
• Preparation of a highly active ORR catalyst using the Pt(111)-like thin films as a starting
material (Chapter 4.8.1.4).
• Experimental probing of the surface properties by investigation of the Volta potential difference and evaluation of the catalyst activity towards the oxygen reduction reaction over the same area using the developed SKP-SECM system (Chapter 4.8.1.5).
4.8.1.3. Preparation of Pt(111)-like thin films
A simple and quick method for the preparation of large Pt(111)-like thin films (~1cm2) on Si/Ti
substrates for electrochemical and/or electrocatalytic experiments is presented. This method
involves physical vapor deposition followed by flame annealing, electrochemical cleaning and a
short heat treatment under a controlled atmosphere to reduce the number of point defects at the
sample surface, especially on the Pt(111) terraces. Careful selection of the substrate, surface
preparation and cooling atmosphere allows production of Pt(111)-like thin films which show
voltammetry features typical of well-ordered bulk Pt(111) single crystal electrodes in 0.1 M
Results and Discussion
138
HClO4. The prepared Pt(111)-like thin films are characterized using scanning electron microscopy
(SEM), atomic force microscopy (AFM) and cyclic voltammetry (CV).
4.8.1.3.1. State of the art in Pt(111)-like surface preparation
Platinum based electrodes are often used as model objects in electrochemistry, particularly for
research in the field of electrocatalysis [278]. The main reason for this is the unique ability of Pt to
catalyze many reactions which are essential for a future hydrogen economy especially oxygen
reduction [279], hydrogen oxidation [256], or alcohol oxidation [198]. Modern model studies in
electrocatalysis involve well defined single crystalline surfaces to relate surface structure
properties with material catalytic activity [266]. Among all currently known Pt single crystal
surfaces, Pt(111) is most widely used as the starting point for the development of new catalysts
with improved ORR and HER activity. Modification of Pt(111) single crystal surfaces using
controllable surface [280] or sub-surface alloying [267], as well as deposition of well-defined
atomic overlayers [281] have been shown to be very promising and efficient approaches towards
further optimization of catalytic activity.
The preparation of high-quality Pt(111) single crystal surfaces was originally developed by
Clavilier and co-workers [282] using small bead-type single crystals. This rather simple technique
for the reproducible preparation of smooth single crystalline surfaces which can be performed in
any laboratory around the world caused a significant leap in progress in understanding of
electrocatalytic properties of solids and allowed significant advances to be made in physical
electrochemistry [271].
Current state of the art techniques for preparation of high-quality model Pt(111) single crystal
surfaces require a sequence of delicate procedures that involve melting of raw Pt material,
crystallization of the melt which leads to the formation of a bulk crystal followed by X-ray
diffraction (XRD) based adjustment of the crystal to desired crystallographic orientation and
subsequent polishing to a surface roughness down to 30 nm. To check the quality of the prepared
single crystal surfaces, the specific adsorption of *H, *OH and *O on the Pt(111) single crystal
surface in clean aqueous electrolytes such as HClO4, NaOH, or H2SO4 which lead to well-known
“finger-print” voltammetric profiles can be used. As shown by Climent and Feliu [271], well-
ordered Pt(111) single crystal surfaces are easier to obtain on smaller crystals than on larger
surfaces. The shape of Pt(111) “finger-print” voltammetric profiles are extremely sensitive to the
presence of surface defects and any impurities, as these factors affect the dynamics of the ordered
two-dimensional adsorbate phase formation, particularly in HClO4 solutions [283].
Results and Discussion
139
Preparation of Pt(111)-like thin films on different substrates using physical vapor deposition
followed by long high-temperature annealing and/or flame annealing procedures have already
been reported [284 - 286]. However, the voltammetric profiles of the prepared films reveal
relatively defective surfaces that exhibit almost indistinguishable voltammetric features attributed
to the two-dimensional disorder / order phase transitions in either *OH or SO42- adsorbate layers.
This voltammetric features are known to be very sharp and characteristic in the case of high-
quality Pt(111) single crystal surfaces [286 - 288].
In summary, to transfer the knowledge and techniques developed using small model Pt(111)
single crystal electrodes (typically 0.2 mm to 1 cm in diameter) to surfaces comparable with those
normally used in industrial applications, it is highly desirable to develop a new procedure for the
preparation of Pt(111)-like thin films with large well-defined smooth surfaces made of Pt(111)
facets having a maximum number of catalytic centers. Additionally, this procedure is necessary to
provide well-defined and reproducible surfaces for SKP-SECM measurements used in this work.
4.8.1.3.2. Preparation of high-quality Pt(111)-like thin films
The main procedures for the Pt(111)-like film preparation are schematically shown in
figure 4.61. Clean commercial Si(100) substrate (4 inch Si wafer, orientation ±0.5°) covered with
a thick SiO2 layer (thermal oxidation, 1500 nm ± 5%) as a barrier layer was used as a starting
material. The oxidized surface of the substrate was coated with a 10 nm thin Ti layer (99,995%) to
improve Pt film adhesion. A 200 nm thick Pt film (99.995%) was then deposited on top of Ti
layer. A detailed description of all relevant deposition parameters is provided in section 7.3.4.
After sputtering, the Pt covered side of the Si wafer was protected with a thin layer of a PMMA-
based photoresist AR-P 617 to avoid contamination of the surface by dust particles or cooling
agents during the subsequent cutting of the wafer into small pieces (ca. 1.5 cm x 0.8 cm) using an
automated high-precision saw. The first step in the preparation of high-quality Pt(111)-like thin
films is flame annealing of the samples in air for approximately 10-15 seconds using a portable
torch supplied with butane / oxygen mixture. The flame of the torch was oriented perpendicular to
the sample surface and 10 periodic heating and cooling sequences were applied.
The second step in the preparation procedure is focused on the electrochemical cleaning of the
flame annealed samples in an electrochemical cell to remove any surface contaminations from the
laboratory atmosphere or other sources. Before electrochemical cleaning of the samples, the
glassware of the electrochemical cell was cleaned with a fresh “piranha” solution (3:1 mixture of
Suprapur® 96 % H2SO4 and TraceSelect Ultra® 30°% H2O2, 12 h).
Results and Discussion
140
Step 1
Step 2
Step 3
Figure 4.61. Schematic representation of main procedures for Pt(111)-like film preparation [289].
Flame annealed samples were cycled between
-0.67 V and 0.8 V vs. a mercury-mercury sulfate
(MMS) reference electrode in 0.1 M HClO4 solution
prepared using Siemens© UltraClear water
(0.055°µS/cm, TOC content < 1 ppb) and Suprapur®
HClO4 until a stable reproducible electrode response
was observed. After the electrochemical cleaning all
samples were carefully rinsed with Siemens© UltraClear
water and dried in either Ar/H2 (6.0, 5°% H2) or Ar/CO
(5.0, 0.1°% CO) gas streams. Afterwards, the cleaned
samples were transferred into a furnace equipped with clean quartz glassware for thermal
treatment. Before heating of the samples, the quartz tube was cleaned in the same way as the
electrochemical cell using “piranha” solution. During the third step, the cleaned samples were
heated in a tube furnace to 400°C for 3-5 min in a flow of either Ar/H2 or Ar/CO and cooled down
under one of these controlled atmospheres (Fig. 4.62). Swagelok® stainless steel tubes were used
to supply Ar, Ar/H2 and Ar/CO gas mixtures into the cell and the furnace. Plastic tubing was
Figure 4.62. Photographic image of the Pt(111)-like thin film electrode.
Results and Discussion
141
avoided to minimize outgassing of possible organic contaminations.
4.8.1.3.3. Characterization of high-quality Pt(111)-like thin films
The prepared samples were characterized using cyclic voltammetry in a three-electrode cell
configuration with the sample acting as the working electrode, a 500 µm Pt-wire as counter
electrode and the MMS reference electrode (kept in a separate compartment and separated from
the working solution with a ceramic insert). Before electrochemical experiments, all glassware
was cleaned with “piranha” solution similar to cleaning procedure for the electrochemical cell
described earlier. To remove SO42--ions, multiple heating, ultrasonic treatments and rinsing with
Siemens© UltraClear water were performed. For benchmark experiments, a commercial Pt(111)
single crystal with a diameter of 5 mm (oriented to < 0.1°, polished down to 30 nm roughness)
obtained from Mateck (Jülich, Germany) was used. The experimental protocol for the preparation
and electrochemical cleaning procedure of this crystal are discussed in chapter 7.3.5 and were
similar to those reported by Bondarenko and co-workers [258]. All experiments were performed
in 0.1°M HClO4 solution prepared using Suprapur® HClO4 and Siemens© UltraClear water.
The perchloric acid is known to be a model electrolyte for many electrocatalytic studies.
Figure 4.63 shows a typical cyclic voltammogram of Pt(111) single crystal in 0.1°M HClO4 used
as benchmark.
Figure 4.63. Typical cyclic voltammogram of a commercial Pt(111) single crystal oriented to < 0.1° (dE/dt = 50°mV/s, 0.1°M HClO4). Areas of the cyclic voltammogram marked with red squares
are attributed to hydrogen adsorption / desorption (A), adsorption / desorption of *OH (B) and of *O (C). Green arrows indicate “butterfly” peaks correlated with the quality
of the Pt(111) single crystal surface.
Three distinct regions are distinguishable in the voltammogram shown in figure 4.63. These
A
B
C
Results and Discussion
142
B A
are associated with hydrogen adsorption / desorption (A, between -0.67°V and -0.3 V vs. MMS),
adsorption / desorption of *OH (B, between -0.2°V and 0.1°V vs. MMS) and adsorption of O* (C,
between 0.2 V and 0.4 V vs. MMS). The height and sharpness of the “butterfly” peaks at around
0.08 V vs. MMS are often correlated with the quality of the Pt(111) single crystal surface and the
cleanliness of the system.
Figure 4.64 shows cyclic voltammograms of two typical Pt(111)-like thin films to demonstrate
the quality and the reproducibility of the preparation procedure. The characteristic “butterfly”
peaks exhibited in the prepared Pt(111)-like thin films are less pronounced (smaller / less sharp)
than those for the benchmark Pt(111) single crystal. However, they are clearly distinguishable and
are associated with the order / disorder phase transition at Pt(111) terraces. As “butterfly” peaks
are very sensitive to long-range order, their presence clearly indicates long range ordering in the
prepared films. Additionally, noticeable differences between the cyclic voltammograms are also
observable in the hydrogen adsorption / desorption regions between -0.67°V and -0.3°V vs. MMS
and between 0.1 V and 0.3 V vs. MMS. These differences can be explained by the presence of
defects at Pt(111) terraces as large-area thin films are likely to have more surface defects than
well-oriented small bulk Pt(111) single crystals [271], [283], [290].
Figure 4.64. Typical cyclic voltammograms of prepared Pt(111)-like thin films (dE/dt = 50°mV/s, 0.1°M HClO4). Voltammetric peaks observed at -0.6 V (blue square) and -0.45 V (orange square) are characteristic for
Pt(110) and Pt(100) -like defects. Red squares indicate the area of CV attributed to the initial adsorption of *O while green arrows indicate “butterfly” peaks correlated with the
quality of the Pt(111) single crystal surface.
Additionally, small voltammetric peaks at -0.6 V (blue square) and -0.45 V (orange square) both
vs. MMS reference electrode are characteristic of Pt(110) and Pt(100)-like defects and indicate the
presence of areas on the surface of the prepared Pt(111)-like thin film with orientations of surface
Pt atoms that differ from Pt(111). Nevertheless, these are the first published cyclic
Results and Discussion
143
A B
voltammograms of Pt(111)-like thin films that demonstrate characteristic sharp “butterfly” peaks
where the shape approaches that of well-oriented bulk Pt(111) single crystals [289].
Longer annealing times were found to give no improvement to the CV response of the Pt(111)-
like thin films, however, pre-treatment of the sample in an Ar/CO atmosphere for 5 min appeared
to yield films with better voltammetric features characteristic of Pt(111) single crystal surfaces.
Annealing in Ar/CO atmosphere at 400°C for longer times did not give any additional impro-
vement in the electrochemical response of the samples.
To gain insights into the morphology of the prepared Pt(111)-like thin films, scanning electron
microscopy (SEM) was used. Figure 4.65 shows SEM micrographs of a polycrystalline Pt thin
film after sputtering of pure Pt on the surface of an oxidized Si wafer (Fig. 4.65 A) and of
Pt(111)-like thin film after heat treatment under controlled atmosphere (Fig. 4.65 B). The freshly
deposited Pt surfaces were found to be composed of small Pt clusters with dimensions of tenth of
nanometers. However, no evidence of any significant surface ordering was observed. Flame
annealing of the samples led to much better ordering of Pt atoms on the thin film surface and the
formation of large smooth terraces with a few small defects at the surface. However, cyclic
voltammetry of these surfaces indicate a polycrystalline Pt surface that requires further treatment
to obtain a stable reproducible response, as discussed above. Following annealing in a tube
furnace for 3-5 min at 400°C in a flow of Ar/H2 mixture and subsequent cooling under controlled
atmosphere resulted in the characteristic Pt(111) CV response being observed.
Figure 4.65. SEM micrographs of a polycrystalline Pt thin film after sputtering of pure Pt on the surface of an oxidized Si wafer (A) and of a Pt(111)-like thin film after heat treatment
under controlled atmosphere (B).
500 nm
Results and Discussion
144
AFM was used to investigate the topography of prepared Pt(111)-like thin films. Figure 4.66
shows a typical AFM area scan performed over one of many randomly chosen positions on the
Pt(111)-like thin film surface.
Figure 4.66. AFM area scan image over one of many randomly chosen positions on the surface of a prepared Pt(111)-like thin film (A). White lines labeled with (a), (b) and (c) are marking positions on the sample surface which corresponding line profiles are presented as profile (a), (b) and (c). The x- and z-values
were converted into atomic radii of single Pt atoms (Pt at. diam.) [289].
For a better visualization of the length of the monoatomic terraces and the smoothness of the
sample surface, recorded x- and z-values were converted into atomic diameters of single Pt atoms.
Three line profiles were extracted from the area scan data set corresponding to three neighboring
plateaus. Line profiles of two plateaus (Fig. 4.66 a and Fig. 4.66 c) reveal long range order with
large atomically flat terraces with a width of up to 100 Pt atoms separated by single monoatomic
steps. Imaging of the defect site also displayed the presence of steep polyatomic steps with a
height of up to five Pt atoms (Fig. 4.66 b).
A detailed description of the AFM equipment used and other relevant operational parameters
are listed in chapter 7.1.4.
To summarize, a quick and simple method for the formation of large Pt(111)-like thin films has
been developed. Flame annealing of the sample enables reorganization of Pt atoms and leads to
the formation of large Pt(111) terraces, which are known to be more stable at the surface of the Pt
Results and Discussion
145
thin films. Further electrochemical cleaning removes unwanted surface contaminations and
additionally eliminates undercoordinated Pt atoms. Final annealing at 400°C under controllable
Ar/H2 or Ar/CO atmospheres reduces the number of point defects at the surface, particularly on
Pt(111) terraces. Each processing stage was found to be critical to produce Pt thin films which
show voltammetric features typical for small well-ordered Pt(111) single crystal electrodes in
0.1M HClO4.
Scale-up of this technique to form larger electrodes should be relatively simple, making this
procedure promising for the transfer of knowledge and approaches developed for small model
Pt(111) single crystal electrodes to surfaces comparable to those normally used in industrial
catalytic applications.
4.8.1.4. Preparation of PtCu NSA on Pt(111)-like thin films
4.8.1.4.1. State of the art in ORR catalysts
As mentioned above, one of the main barriers to successful commercialization of PEM fuel
cells equipped with state of the art Pt-based catalysts is the high loading of noble metal required to
reach the desired power performance of the system which leads to high overall costs. Ideally, one
would use catalysts based on more abundant elements resulting in much lower overall costs for
the catalyst. However, under real operational conditions for PEMFCs (U0 = 0.9 V, pH ≈ 0) only
Pt, Au and Ir are thermodynamically stable in bulk metal form [291] while even Pt starts to
corrode at ~ 1 V [292]. There are a few examples of non-metal catalysts such as metalorganic
complexes [293], oxides [294] and graphene-based materials [295], [296] which demonstrate an
ORR activity comparable, or even better to that of pure Pt. However, these catalysts suffer from
poor stability, especially in acidic solutions.
Improvement of the catalytic activity of pure Pt by alloying with less noble solute metals such
as Co, Ni, Fe, Cu and Y is one of the most actively developing areas in modern electrocata-
lysis [257], [267], [297]. Oxygen reduction catalysts based on Pt-alloys have a surface layer that is
almost always entirely Pt with a core composed of Pt and solute metal atoms (Pt alloy). The
surface Pt layer provides kinetic stability to atoms of the solute metal against dissolution as they
are thermodynamically unstable at the surface at operational conditions of PEMFCs [291]. There
are two main approaches to form a Pt overlayer around the Pt alloy core: either leaching of solute
metal from the surface of the Pt alloy in acidic electrolyte [297] or vacuum annealing of the Pt
alloy in an inert or reducing atmosphere [298]. DFT calculations support the notion that the Pt
Results and Discussion
146
atoms in the Pt overlayer on these catalysts exhibit mildly weaker binding to *O and *OH than
pure Pt [260], [268].
Application of solute metal leaching was reported for the first time by Toda and co-
workers [297] in 1999 and this leached structure was indicated by Stamenkovic and co-
workers [298] to be a “Pt-skeleton” surface. Typically, the Pt overlayer has a negligible amount of
solute metal atoms and has a thickness of about 1-2 nm. Since the first report, many leached “Pt-
skeleton” structures with different solute metals have been reported [268], [299], [300]. An
alternative route to form a thin Pt overlayer is based on vacuum annealing of Pt alloys which leads
to exchange of solute metal atoms present in the surface layer with Pt atoms. This leads to
migration of the solute metal atom into the second layer and subsequent formation of a monolayer
thick so-called “Pt-skin” composed of Pt atoms [298]. The main driving force for the exchange of
solute metal atoms originally present in the surface layer with Pt atoms is the lower surface energy
of Pt atoms relative to solute metal atoms.
Adjustment of *O, *OH or *OOH binding energies on Pt surface by alloying of Pt with less
noble metals can be explained by corresponding modification of the electronic structure of the Pt
overlayer for both Pt-skeleton and Pt-skin surfaces [301]. Such modification can be described by
ligand [302] and strain effects [303]. The ligand effect occurs in the case where the electronic
structure of surface atoms (which act as active sites for a given electrochemical reaction) is
modified by neighboring atoms of dissimilar elements. This effect is more pronounced in ORR
catalysts with a Pt overlayer when the solute metal atoms are in the second atomic layer and is
almost negligible when the solute atoms are embedded in the fourth atomic layer [304]. The strain
effect is present when the catalyst is strained parallel to its surface, i.e. the lattice parameters at the
surface are changed by alloying with solute metal and tend toward that of the bulk. Thus, a change
in the bulk composition (creating a core with a smaller lattice parameter than by Pt) results in a
corresponding change in the interatomic Pt-Pt distance at the surface of the catalyst leading to
weakening of the binding of ORR intermediates on the catalyst surface (compressive lattice strain)
while tensile strain will have the opposite effect [303].
Ligand and strain effects are rather difficult to separate, as almost all modern ORR catalysts
(“Pt-skin” and “Pt-skeleton” catalysts) have a thin Pt layer on their surface and are enriched with a
solute metal core. Thus, their surfaces should always be strained as their bulk lattice parameters
are dissimilar to that of Pt. On the other hand, most active forms of Pt3Co- and Pt3Ni-based
catalysts have vacuum annealed Pt-skin structures with an enrichment of the solute metal in the
Results and Discussion
147
second atomic layer [297], [305]. This suggests that the ligand effect is also important for the very
high ORR activity of these catalysts.
Nonetheless, there are two known examples of ORR catalysts, where it was possible to
improve the catalytic activity of material using only one of the above-mentioned effects. Stephens
and co-workers [267] investigated the effect of subsurface alloying on the ORR activity using a
Pt(111) single crystal to prepare a Cu-Pt(111) near-surface alloy. The alloy was formed by
underpotential deposition (UPD) of sub-monolayer amounts of Cu and subsequent annealing of
the crystal under a controlled atmosphere. Using angle resolved X-ray spectroscopy (AR-XPS)
and rotating disk electrode (RDE) techniques the authors found that an improvement of the ORR
activity of up to 8 times compared to Pt(111) single crystal surface was observed when Cu atoms
were stabilized in the subsurface layer. Since the bulk of the crystal was composed of only Pt
atoms, there was no bulk lattice strain suggesting that the improved ORR activity was primarily
caused by the ligand effect.
One of the most comprehensive studies into the influence of the strain effect on the ORR
activity was reported by Strasser and co-workers [268]. The authors investigated dealloyed PtCu
nanoparticles prepared using acidic leaching of pre-annealed PtCu precursors. Ex-situ analysis
suggested that the nanoparticles had a thick Pt skin and a Cu-rich core. As the lattice parameter of
Cu is smaller than that of Pt, this caused a lateral strain to the Pt surface atoms and lead to a
weakening of the binding of *O, *OH and *OOH to Pt surface atoms hence resulting in 4 to 6
times enhancement in ORR activity over that of a Pt(111) single crystal surface.
To select an ORR catalyst with higher ORR activity than a Pt(111) single crystal surface, the
literature was reviewed with the focus on ORR catalysts of different types with reported ORR
activity greater than that of Pt(111) single crystal surfaces (Fig. 4.67). For a better comparison of
data, the catalytic activity of all presented materials was normalized with respect to the activity of
a Pt(111) single crystal surface and polycrystalline alloys were marked with “*” while catalysts
with specifically oriented surfaces like (111) were indicated with “**”.
Detailed analysis of data revealed the Pt3Ni (111) single crystal reported by Stamenkovic et
al. [266] and Cu-Pt (111) near-surface alloy (NSA) reported by Stephens and co-workers [267] as
the two most active ORR catalysts ever reported. Both catalysts have specifically oriented
surfaces suggesting Pt(111) single crystal surfaces as the basis while alloying with late transition
metals as a powerful tool for fine tuning of ORR activity.
Results and Discussion
148
Figure 4.67. Comparison of ORR activity of well-oriented and polycrystalline Pt-based catalysts exhibiting an ORR activity higher than that for the Pt(111) single crystal surface (Pt3Ti, Pt3Ni, Pt3Fe and
Pt3Co – vacuum-annealed polycrystalline alloys of Pt with late transition metals [269]; Pt3Sc and Pt3Y – sputter-cleaned polycrystalline alloys of Pt with Sc and Y [260]; Pt5La – sputter-cleaned
polycrystalline alloy of Pt with La [254]; dealloyed PtCu nanoparticles [268]; Cu-Pt(111) NSA – Cu based near-surface alloy formed using the Pt(111) single crystal [267];
Pt3Ni(111) – vacuum annealed Pt3Ni(111) single crystal [266].
The catalyst of choice for further investigations using the integrated SKP-SECM system should
not only have a very high overall ORR activity. The preparation procedure of this catalyst should
allow application of the developed Pt(111)-like thin films as a starting material. Since the
Pt3Ni (111) is an alloy of pure Pt and Ni, formation of this catalyst using Pt(111)-like thin films
would involve rather complex preparation procedures and sophisticated equipment such as a UHV
chamber equipped with a sputtering system etc. On the other hand, preparation of Cu-Pt (111)
near-surface alloy was originally performed using the Pt(111) single crystal as a starting material.
This suggests that it should be possible to prepare Cu-Pt(111) NSA using the developed Pt(111)-
like thin films. Thus, further discussion will focus on results of the preparation and
characterization of Cu-Pt(111) near-surface alloy thin films prepared using Pt(111)-like thin films
as a starting material.
Results and Discussion
149
Step 1 Step 2
4.8.1.4.2. Preparation of Cu-Pt(111) NSA on Pt(111)-like thin films
The procedure for the preparation of Cu-Pt(111) near-surface alloy thin films is schematically
presented in figure 4.68. The Pt(111)-like thin films formed by means of a simple preparation
procedure as described in detail in chapter 4.8.1.3.2 were used as the starting material.
Figure 4.68. Schematic representation of the preparation procedure for formation of Cu-Pt(111) near-surface alloy thin films using Pt(111)-like thin film electrodes as a starting material.
Prior to the actual preparation of the NSA thin film all Pt(111)-like thin film electrodes were
cycled in 0.1M HClO4 between 0.05 V and 1.12 V vs. RHE to electrochemically clean the surface
and to verify the presence of voltammetric “butterfly” peaks characteristic for Pt(111) facets.
The first step in the preparation of the Cu-Pt(111) NSA thin film is based on the
electrochemical deposition of a monolayer of Cu onto the Pt(111)-like thin film electrode surface
by means of underpotential deposition while applying a potential of 0.33 V vs. RHE for 3 min in
an electrolyte containing 1 mM Cu2+ and 0.1 M HClO4. During UPD, deposition of the metal was
limited to monolayer quantities by ensuring the electrode potential is too anodic for bulk metal
deposition.
UPD of Cu from acidic perchlorate electrolytes is not complicated by specific adsorption of Cl-
containing anions which are present in the solution. This has been shown in a separate
investigation [306] using electrochemical quartz crystal microbalance combined with
electrochemical impedance spectroscopy (EQCM-EIS) [307], [308].
Results and Discussion
150
To ensure the deposition of a full monolayer of Cu, the potential was held at the cathodic limit
(0.33 V vs. RHE) for 3 min and the charge under the subsequent anodic stripping peak allowed
confirmation of complete monolayer formation. To remove the intact Cu monolayer present on
the sample surface, the Pt(111)-like thin film electrode was rinsed with Siemens© UltraClear water
while keeping the working electrode under potential control (0.33 V vs. RHE) to avoid oxidation
of the deposited monolayer of Cu. To allow the potential of the working electrode to be controlled
while the sample is in or not in the electrolyte a dummy electrode (Pt wire) present in the
electrochemical cell was connected to the working electrode. After rinsing, the sample was dried
in an Ar/H2 (6.0, 5 % H2) atmosphere for 30 min. Once dry, the electrode was disconnected from
the potentiostat and placed in a tube furnace equipped with clean quartz glassware to allow
heating and cooling of the thin film samples under a controlled atmosphere. Prior to heating of the
sample, all the glassware inside the furnace was cleaned using the same procedure as discussed
above for cleaning of the electrochemical cell. The second step in the preparation of Cu-Pt(111)
NSA thin films is to anneal the sample under a reducing atmosphere (Ar/H2, 6.0, 5 % H2) at
400°C. The temperature within the furnace was maintained at 400°C; and once an Ar/H2(5 %)
atmosphere was established, the sample was introduced to the heated section of the furnace for
2 min. After 2 min, the sample was removed from the furnace and allowed to cool for 5 min under
controlled atmosphere. This reductive annealing forces Cu atoms to move preferentially into the
second Pt-layer (sublayer) as shown schematically in figure 4.68. As shown by Stephens and co-
workers [267], annealing of the sample leads to stabilization of approximately 2/3 ML of Cu in
the second layer while the rest of the Cu atoms migrate into the bulk Pt-phase. After annealing the
sample was reintroduced to the electrochemical cell for electrochemical characterization.
4.8.1.4.3. Characterization of the Cu-Pt(111) near-surface alloy
Electrochemical characterization of the prepared Cu-Pt (111) near-surface alloy thin films was
performed using CV between 0.05 V and 1.12 V vs. RHE in saturated with nitrogen or oxygen
0.1 M HClO4 solutions under controlled atmosphere. CVs recorded in N2-saturated 0.1 M HClO4
solution provided evidence for the successful formation of Cu-Pt(111) NSA thin films based on a
comparison of potentials required for stripping of 1 ML Cu deposited via Cu UPD on Pt(111)-like
thin film and on the prepared Cu-Pt(111) NSA thin film in O2-free 0.1 M HClO4 (Fig. 4.69). In
the case of the Pt(111)-like thin film a sharp anodic peak corresponding to irreversible dissolution
of Cu into solution as Cu2+ was observed at 0.79 V vs. RHE. This data is in a good agreement
with the stripping potential of 1 ML Cu reported by Markovic and co-workers [309] for well-
Results and Discussion
151
ordered Pt(111) single crystal surface. In the case of Cu-Pt(111) NSA thin films, the Cu
dissolution peak shifted to a more cathodic potential and centered around 0.73 V vs. RHE. The
lower potential required for stripping of 1 ML Cu indicates a weaker interaction of Cu atoms with
the Pt surface caused by higher nobility of the surface as a result of the NSA formation.
Figure 4.69. Stripping of Cuad (1 ML) from a Pt(111)-like thin film (black line) and a Cu-Pt(111) NSA thin film (red line) in O2-free 0.1 M HClO4, T = 295 K, dE/dt = 50 mV/s.
Figure 4.70 shows cyclic voltammograms of the prepared Cu-Pt(111) near-surface alloy and a
Pt(111)-like thin film recorded in N2-saturated 0.1 M HClO4 solution under controlled
atmosphere.
Figure 4.70. Cyclic voltammograms of a Pt(111) benchmark single crystal (black line) and the formed Cu-Pt(111) NSA (red line) recorded in O2-free 0.1 M HClO4 (T = 295 K, dE/dt = 50 mV/s).
Results and Discussion
152
The overall shape of the CV recorded using the prepared Cu-Pt(11) NSA thin film is in good
agreement with data reported by Bondarenko and co-workers [200]. Detailed analysis of the CV
revealed no voltammetric features which would indicate that Cu is stripped from the surface of the
NSA suggesting complete diffusion of the deposited ML of Cu into the bulk of the Pt(111)-like
thin film and formation of a compact Pt overlayer (Pt-skin) on the sample surface. Additionally,
no sharp “butterfly” peaks characteristic of Pt(111) single crystal surface as recorded on the
Pt(111)-like thin film electrode before fabrication of the NSA were observed. Instead, broad peaks
were observed which indicate the changing strength of the interaction of the Cu-Pt(111) NSA thin
film surface with *OH and *H. Shifts of the *H adsorption peak (in cathodic direction) and of the
OH* adsorption peak (in anodic direction) confirm that the presence of Cu in the subsurface layer
destabilizes these adsorbates on the surface of the formed Cu-Pt(111) NSA thin films similar to
features observed on Pt3Ni(111) single crystal surfaces [266].
The usual method to test the ORR activity of Pt-based catalysts is to use a rotating disc
electrode (RDE). However, the use of commercial RDE equipment with large thin-film samples
was prohibitively difficult. Therefore, to elucidate the activity of prepared Cu-Pt(111) near-
surface alloy thin films cyclic voltammetry and square-wave voltammetry (SWV) were utilized in
0.1 M HClO4 solutions oxygenated for at least 10 min prior to the electrochemical experiments.
The sample was cycled between 0.05 V and 1.0 V vs. RHE and all electrochemical experiments
were performed under oxygen atmosphere.
Figure 4.71 shows anodic scans of cyclic voltammograms recorded using a Pt(111)-like thin
film (black line) and a Cu-Pt(111) near-surface alloy thin film (red line).
Figure 4.71. Anodic scans of cyclic voltammograms of a Pt(111)-like thin film (black line) and a Cu-Pt(111) NSA (red line) performed in O2-saturated 0.1 M HClO4 (T = 295 K, dE/dt = 50 mV/s). For better
comparison both data have been normalized by the maximal value of the limiting current density [310].
Results and Discussion
153
Due to small variations in the limiting current density both data have been normalized by the
maximal value of the limiting current density [310]. The observed half-wave potential for the
Pt(111)-like thin film is in excellent agreement with that observed on well-ordered bulk Pt(111)
single crystals (~ 0.86 V vs. RHE) as reported by different research groups [254], [266]. This is
the highest ORR activity reported for Pt-based thin films up to date and measured under similar
conditions. The observed half-wave potential for the prepared Cu-Pt(111) NSA is approximately
0.89 V vs. RHE, which is slightly lower than that reported by Stephens and co-workers [267] for
the Cu-Pt(111) NSA prepared on well-ordered bulk Pt(111) single crystals (~0.9 V vs. RHE,
measured at 60°C using RDE technique). On the other hand, the observed shift of the half-wave
potential for the prepared Cu-Pt(111) NSA is around 30 mV in anodic direction indicating
remarkably improved ORR activity of the formed Cu-Pt(111) NSA thin film in comparison to the
initial Pt(111)-like thin film surface. The prepared Cu-Pt(111) NSA thin film and its ORR activity
were stable for at least 3 h while cycling the potential between 0.05 V and 1.0 V vs. RHE on
oxygenated 0.1 M HClO4 solution under oxygen atmosphere.
Figure 4.72 shows square-wave voltammograms (SWVs) recorded using the developed
Pt(111)-like thin film (black line) and prepared Cu-Pt(111) NSA thin film (red line). For square-
wave experiments the potential limits were 0.05 V and 1.0 V both vs. RHE while the pulse height
and width were 12.5 mV and 500 ms respectively, the step height was 2.5 mV. The peak
separation observed in SWVs for the oxygen reduction reaction shows the same anodic shift of
about 30 mV. This additionally confirms that the prepared Cu-Pt(111) NSA thin film is more
active than the initial Pt(111)-like thin film surface.
Figure 4.72. Square-wave voltammograms of a Pt(111)-like thin film (black line) and a Cu-Pt(111) NSA
thin film (red line) performed in O2-saturated 0.1 M HClO4 (T = 295 K, pulse height and width were 12.5 mV and 500 ms respectively, step height was 2.5 mV).
Results and Discussion
154
The developed Cu-Pt(111) NSA thin film catalyst shows an ORR activity which is ~5.5 times
higher than that of state of the art Pt nanostructured thin films [311]. This corresponds to a
turnover frequency of ~17e-/s per active site on the Cu-Pt(111) NSA thin film surface at 295˚K at
0.9 V vs. RHE in O2 saturated 0.1 M HClO4 solution under oxygen atmosphere. Further optimi-
zation of the Cu content in the second layer could potentially further enhance the ORR activity of
the formed Cu-Pt(111) NSA thin films.
To conclude, a simple and quick method for the formation of large Cu-Pt(111) NSA thin films
using Pt(111)-like thin films as a starting material was developed and evaluated. In these NSA
thin films, the solute metal Cu is preferentially located in the second platinum layer and protected
by a Pt skin from being dissolved at operation conditions of PEM fuel cells where it is
thermodynamically unstable (0.9 V vs. RHE). The Cu-Pt(111) NSA film acts as a highly active
and fairly stable electrocatalyst for the reduction of oxygen with activity and stability which
approach those for bulk single crystalline Pt-alloy surfaces and ~5 times more active than state-of-
the art Pt thin films reported up to date.
4.8.1.5. Investigation of Cu-Pt(111) NSA using the integrated SKP-SECM system
In this section, the link between the contact potential difference (work function) and the
location of Cu atoms relative to Pt atoms on the surface of Pt(111)-like thin films and the resulting
ORR activity is evaluated using the integrated SKP-SECM system. To evaluate the sensitivity of
the SKP-SECM system, different quantities of Cu ranging from 1 ML to approximately 11 ML
were deposited on half of the surface of Pt(111)-like thin film electrodes. The lateral distribution
of the Volta potential difference over the sample surface was investigated in the SKP mode. To
assess the influence of Cu atoms located in the second Pt layer on the Volta potential difference of
the sample Cu-Pt(111) NSA thin films were formed on Pt(111)-like thin film electrodes. Area
scans over the sample were performed in the SKP mode while the catalytic activity of the Cu-
Pt(111) NSA thin films was visualized in the redox-competition mode of SECM performed in
oxygen saturated 0.1 M HClO4 solution using the 25 µm Pt SKP-SECM tips.
4.8.1.5.1. Influence of Cu atoms on the contact potential difference
To evaluate the influence of monolayers of Cu on the surface of the Pt(111)-like thin films on
the resulting CPD across the sample surface, two Pt(111)-like thin film electrodes were partially
(½) covered with 1 ML and approximately 11 ML of Cu. Deposition of 1 ML Cu on the electrode
Results and Discussion
155
surfaces was performed using Cu UPD with parameters similar to those used to prepare the Cu-
Pt(111) NSA, while 11 ML were deposited by applying a slightly more cathodic potential as for
Cu UPD of 1 ML. The amount of deposited Cu was calculated based on the anodic stripping
charge. Lateral distribution of the CPD across the sample surface was investigated using the
25 µm glass insulated Pt SKP-SECM tip while a polished piece of pure Ni was used as a reference
(Fig. 4.73). During the measurement, water saturated air was continuously purged through the
chamber around the sample and the software based feedback controller was activated to enable
automatic adjustment of the tip-to-sample distance to 5 µm. To avoid vibration of the sample by
its interaction with the air streams a brass made shield was placed around the sample holder.
Figure 4.73. Photographic image of the Pt(111)-like thin film electrode partially (½, right side) covered with 1 ML Cu by means of Cu UPD. A polished piece of pure Ni was used as reference sample.
Figure 4.74 shows area scans performed over the Pt(111)-like thin film electrode surface
partially (½, right side) covered with 1 ML Cu (A) and approximately 11 ML Cu (B). It is clear
that even a small difference in the amount of Cu deposited on the surface of the Pt(111)-like thin
film leads to remarkable differences in the measured CPD. Deposition of 1 ML Cu on the surface
of the Pt(111)-like thin film electrode led to a shift of the measured CPD to negative direction of
around 100 mV indicating “lower nobility” of the surface covered with 1 ML Cu. In the case of
11 ML of Cu most probably deposited in the form of Cu clusters on the sample surface, the shift
of the measured CPD in the negative direction was around 280 mV indicating a much more
pronounced reduction of the surface nobility relative to the clean Pt(111)-like thin film surface.
Both experiments confirmed the very high sensitivity of the developed SKP-SECM system
towards extremely small amounts of material deposited on the surface of the sample. Moreover, it
is not only possible to distinguish between a clean Pt(111)-like thin film surface and one covered
Results and Discussion
156
with 1 ML Cu, but also to probe differences in amounts of Cu deposited on the sample surface
(1 ML vs. 11 ML).
Figure 4.74. Area scans performed over a Pt(111)-like thin film electrodes partially (½, right side) covered with 1 ML Cu (A) and 11 ML Cu (B) using a 25 µm Pt SKP-SECM tip in the SKP mode. A polished piece of pure
Ni was used as reference while water saturated air was continuously pumped through the chamber around the sample. The tip-to-sample distance was controlled by a software based feedback controller and
adjusted to 5 µm. Blue spheres indicate Pt atoms while orange spheres symbolize Cu atoms.
To evaluate the influence of Cu atoms located in the second Pt layer on the resulting Volta
potential difference across the sample surface, 1 ML Cu was deposited on the surface of the
Pt(111)-like thin film electrode using Cu UPD. Subsequent annealing of the sample at 400˚C
under controlled atmosphere (Ar/H2, 5 %) caused migration of Cu atoms into the second Pt layer
and led to formation of a Cu-Pt(111) NSA. Figure 4.75 shows an area scan performed over a
Pt(111)-like thin film electrode with one half of the sample surface (right side) modified with the
Cu-Pt(111) NSA. The presence of Cu atoms in the second Pt layer caused changes in the
properties of the Pt(111)-like thin film surface which led to a shift of the measured CPD in
positive direction by around 150 mV. This indicates an “increased nobility” of the sample surface.
A B
Results and Discussion
157
Figure 4.75. Area scan performed over a Pt(111)-like thin film electrode partially (½, right side) modified to
a Cu-Pt(111) NSA thin film using a 25 µm Pt SKP-SECM tip in the SKP mode. Water saturated air was continuously pumped through the chamber around the sample while the tip-to-sample distance was
controlled by a software based feedback controller and adjusted to 5 µm. Blue spheres are indicating Pt atoms while light blue spheres correspond to Pt surface atoms (Pt skin)
with Cu atoms (orange spheres) being located in the second Pt layer.
For a better overview, the results were summarized in one diagram (Fig. 4.76). Additionally, an
area scan was performed over a polished piece of pure Cu under the same conditions to measure
the CPD of a polycrystalline bulk Cu sample. All CPD values recorded during an area scan were
averaged to provide one single value for the diagram. It is obvious that an area on the surface of
the Pt(111)-like thin film electrode covered with a monolayer Cu is around 100 mV “less noble”
than the unmodified electrode surface, while further deposition of Cu clusters leads to a more
pronounced “decrease in nobility” by around 180 mV. In the case of the bulk polycrystalline Cu
sample, the measured CPD value was around 540 mV lower than that of the clean Pt(111)-like
thin film. It should be noted that the CPD observed for polycrystalline Cu differs from that
observed under UHV conditions (≈1 eV smaller than a Pt(111) single crystal surface) [312],
although this is to be expected as the surfaces used in this work will inevitably have some
adsorbates on the surface. This fact highlights the requirement to extend studies of surface
characterization to real active surfaces as opposed to relying on surface characterization of “ideal”
surfaces.
Results and Discussion
158
Cu-Pt(111) NSA
1ML Cu (Cu overlayer)
11ML Cu (Cu clusters)
Polycrystalline Cu
Pt(111)-like thin film
∆Φ ≈ 150mV
∆Φ ≈ 100mV
∆Φ ≈ 180mV
∆Φ ≈ 260mV
Figure 4.76. Comparison of CPD values measured over different samples using the 25 µm Pt SKP-SECM tip in the SKP mode. All measurements were performed using the Ni reference sample. The
measured CPD values are presented relative to the CPD of the Pt(111)-like thin film.
4.8.1.5.2. Visualization of the ORR activity of Cu-Pt(111) NSA thin films
To investigate the catalytic activity of the Cu-Pt(111) NSA formed on the Pt(111)-like thin
films, an O-ring was fixed on the sample surface using nail varnish. This creates a small
electrochemical cell for SECM experiments directly after the measurement of the CPD in the SKP
mode. However, it was found that the presence of nail varnish causes contamination of the sample
surface and the measured Volta potential differences across the unmodified Pt(111)-like thin film
surface and the Cu-Pt(111) NSA become indistinguishable. On the other hand, if only the O-Ring
was placed on the sample surface, then a good contrast between unmodified and modified areas
on the sample surface was observed. However, the weak contact between the O-Ring and the
sample surface led to leakage of the electrolyte after a short period of time. To overcome this
limitation a short piece (L = 5 mm) of glass tube (Din = 5 mm, Dout = 8 mm) was used as an
electrochemical cell (Fig. 4.77). To establish a good contact between the glass tube and the
sample surface, a thin layer of a hydrophobic polymer Sylgard 184 was deposited on the lower
edge of the glass tube.
Results and Discussion
159
Figure 4.77. Photographic image of the Pt(111)-like thin film electrode partially (½, right side) modified to a Cu-Pt(111) NSA thin film. A piece of glass tube placed directly on the sample surface was used as
electrochemical cell during SECM experiments.
Visualization of the catalytic ORR activity of the prepared sample was performed in the redox-
competition mode of the SECM (Fig. 4.78).
Figure 4.78. Schematic representation of the redox-competition mode of the SECM on the Pt(111)-like thin film electrode partially (½, right side) modified to a Cu-Pt(111) NSA thin film.
All experiments were carried out in a four-electrode cell configuration with a 25 µm glass
insulated Pt SKP-SECM tip as working electrode one (WE 1), the thin film electrode as working
electrode two (WE 2), a Pt-wire as counter electrode and a Ag/AgCl/ 3 M KCl as reference
electrode. To avoid the presence of chloride ions in the HClO4 solution, a double junction
reference electrode was used with the outer chamber filled with 0.1 M HClO4. The
electrochemical cell was filled with 200 µl of oxygen saturated 0.1 M HClO4 solution and a
sequence of potential pulses was applied at the tip (P1 = 0.65 V, t1 = 1 s; P2 = 1.5 V, t2 = 0.2 s;
Results and Discussion
160
P3 = 0.2 V, t3 = 0.5 s) while the sample was polarized at a constant potential of 0.175 V vs.
Ag/AgCl (0.025 V < OCPsample). The first potential pulse (P1, 0.65 V) is a conditioning potential
applied for 500 ms to restore the diffusional equilibrium after movement of the SECM tip during
scanning. The second potential pulse (P2, 1.5 V) is the injection pulse which leads to oxidation of
water and injection of oxygen into the gap between tip and sample. The third potential pulse (P3,
0.2 V) is the measurement pulse applied for 500 ms which is sufficiently cathodic to invoke
oxygen reduction. As the sample is continuously polarized at a potential sufficient for the oxygen
reduction reaction to proceed, application of the measurement pulse at the tip (P3, 0.2 V) also
leads to oxygen reduction at the tip hence resulting in a competition between the tip and the
sample for dissolved O2. With increased local activity of the sample, the remaining amount of
oxygen which is accessible for the tip decays leading to smaller cathodic currents measured at the
tip over areas with higher ORR activity.
Figure 4.79 shows line scans over the Pt(111)-like thin film electrode one half of which (right
side) was modified to Cu-Pt(111) NSA performed in the SKP (solid dots) and the RC-SECM
(circles) modes. To avoid a background current shift caused by an uncompensated tilt between the
scanning plane of the SECM tip and the electrode surface a software based tilt correction
procedure was used while the tip-to-sample distance was adjusted to 2 µm.
Figure 4.79. Line scans over the Pt(111)-like thin film electrode one half of which (right side) was modified to a Cu-Pt(111) NSA thin film performed in SKP (solid dots) and RC-SECM (circles) modes.
The best contrast between the unmodified surface of the Pt(111)-like thin film and the surface
being modified to a Cu-Pt(111) NSA was achieved at 65 ms after applying the measurement pulse
at the SECM tip. The higher cathodic currents observed over the Pt(111)-like thin film clearly
Results and Discussion
161
indicate higher catalytic ORR activity of the Cu-Pt(111) NSA thin film compared to that on the
unmodified sample surface. Thus, using the integrated SKP-SECM system, a good correlation
between the work function and the catalytic ORR activity of the surface was demonstrated.
4.8.1.5.3. Conclusion
Experiments performed on Pt(111)-like thin film electrodes which were partially covered with
extremely small amounts of Cu (1 ML and 11 ML Cu) demonstrated the superior sensitivity of the
integrated SKP-SECM system. The developed system is not only able to distinguish between
unmodified Pt(111)-like thin film surface and areas on the sample surface covered with 1 ML Cu,
but also to probe extremely small differences in amounts of deposited Cu.
Direct probing of surface properties in combination with evaluation of surface ORR activity
over the same area of the investigated surface using the integrated SKP-SECM system was
successfully applied to elucidate a connection between the surface property (work function) and
the catalytic ORR activity of the Pt(111)-like thin film electrodes and that of the Cu-Pt(111) NSA
thin films. Visualization of the catalytic ORR activity performed in the RC-SECM mode revealed
higher ORR activity on the surface of Cu-Pt(111) NSA. Insights acquired from the combined
SKP-SECM technique should feed back into theoretical modeling to provide the basis for the
design of novel catalytic materials with enhanced ORR activity.
Results and Discussion
162
4.8.2. Combined high-resolution SKP-SECM investigations of local corrosion
4.8.2.1. State of the art in research on Al alloys
Precipitate-hardened aluminum alloys of the 2XXX series and particularly the AA2024-T3
alloy are widely used in the aircraft industry due to their very high strength-to-weight ratio. The
main alloying elements are copper, magnesium, manganese and iron, leading to nominal
compositions of commercially available alloys of 3.8-4.9 % Cu, 1.2-1.8 % Mg, 0.3-0.9 % Mn,
0.5 % Fe, 0.5 % Si, 0.25 % Zn, 0.1 % Cr, 0.05 % Ti by weight balanced with Al. Despite their
superior mechanical properties these alloys are extremely susceptible to localized corrosion in
chloride containing media. Numerous corrosion processes were observed for these alloys such as
pitting corrosion [197], intergranular corrosion [313], exfoliation [314] and stress corrosion
cracking [315].
The presence of alloying elements is the primary reason for developing a heterogeneous
microstructure during solidification of the alloy leading to formation of coarse insoluble
intermetallic particles (IMPs). Two dominant types of IMPs can be found in AA2024-T3 alloys.
The first type consists of particles which contain only Al, Cu and Mg atoms called S-phase
(Al2CuMg). These particles usually have a spherical shape and, as reported by Buchheit et al. [3],
cover 2.8 % of the total alloy surface. The second type are Al-Cu-Mn-Fe particles which are
assigned as Al6(Cu, Mn, Fe). These particles usually have a blocky and angular shape and cover
2.7 % of the total alloy surface [316].
The corrosion of AA2024-T3 alloys during immersion in chloride containing electrolyte
initiates with the preferential release of Al and Mg from the S-phase IMPs. This leads to the
formation of a porous Cu reach particle remnant and redistribution of Cu around the pit.
Enrichment with Cu leads to an increase in the cathodic activity of the remnants towards oxygen
reduction. Increased O2 reduction further promotes local alkalization, which causes the formation
of a deep trench in the aluminum matrix encircling the particle to such an extent that, in some
cases, S-phase particles are detached [317].
Over the last 15 years combinations of AFM and SKPFM have been extensively used for
topographical analysis and corresponding Volta potential difference mapping in air to study the
localized corrosion of AA2024 alloys. The pioneers of this approach were Schmutz and
Frankel [191], who first demonstrated the existence of a linear relationship between Volta
potential differences, measured in air by SKPFM, and the corrosion potential of various metals
immersed in aqueous electrolyte solution. Schmutz and Frankel used these observations as the
basis for the investigation of local galvanic interactions on metallic surfaces. Metals with a more
Results and Discussion
163
positive contact potential difference relative to a reference are defined to be nobler. A freshly
polished AA2024-T3 specimen surface exhibited CPD values which are 0.2 V more positive for
S-phase IMPs and around 0.3 V nobler for iron containing particles than the aluminum matrix. In
the case of exposure to an electrolyte containing chloride ions, the more noble areas on the surface
will behave as local cathodes, inducing O2 reduction on their surface, while the less noble areas
act as local anodes displaying preferential dissolution. However, in some cases no potential
difference was observed between particles and the surrounding matrix on corroded samples [201].
This effect can be explained based on the observations by Guillaumin et al. [318] concerning the
oxide thickness gradient over S-phase IMPs and the influence of the surface oxide thickness and
chemical oxide composition on the measured CPD values. Hence, great care should be taken
when interpreting Volta potential difference mapping performed with SKPFM on the alloy surface
following exposure to electrolyte solutions. Despite SKPFM shows high lateral resolution the
failure to trace the topography, which tends to happen during scanning over pits on corroded
specimens, can lead to erroneously high CPD values [319].
One of the main strategies used for investigating the corrosion behavior of AA2024 alloys
consists of three steps. Firstly, the location of the IMPs under investigation on the freshly polished
specimen surface is derived either by SEM and/or by Volta potential difference mapping using
SKPFM. Secondly, the sample is exposed to the electrolyte of interest to invoke corrosion.
Thirdly, the corrosion products are quantified and the redistribution of alloying elements in the
vicinity of pits is performed using electron probe microanalysis (EPMA) in combination with
energy-dispersive x-ray analysis (EDX) [3], [320] or secondary ion mass spectrometry
(SIMS) [321]. Despite the high lateral resolution of these ex situ techniques, the obtained
complementary data do not provide information about the in situ corrosion processes which occur
during exposure of the specimen to the electrolyte. A number of approaches have been used to
obtain such in situ information. Schneider et al. [322] have used confocal laser scanning
microscopy (CLSM) for the in situ investigation of the morphology of the attack at and around
IMPs simultaneously providing information on the dissolution rate of different IMPs.
Alternatively, Büchler et al. [323] visualized the localized corrosion behavior of AA2024 alloys
with fluorescence and near-field scanning optical microscopy (NSOM) in the presence of a pH
sensitive fluorescent dye. The fluorescent patterns that were formed around single IMPs during
the corrosion attack visualized the solubility transition of Al3+ ions in the electrolyte at high lateral
resolution.
Results and Discussion
164
Non-optical in situ methods which have been used to study localized corrosion include
electrochemical methods such as potentiodynamic polarization (PDP), which provides informa-
tion on the corrosion rate and the polarization resistance (Rp), and electrochemical impedance
spectroscopy (EIS) [324], which provides a description of the transfer function as the response of
a system to an AC perturbation [325]. Since the O2 reduction reaction at the surface of IMPs is the
dominant reaction during corrosion of AA2024-T3 alloys the investigation of this reaction has
played a very important role in clarifying the corrosion mechanism. Jakab et al. [326] performed
rotating disk electrode (RDE) experiments using disks machined out of an AA2024-T3 alloy. The
ORR on the heterogeneous AA2024-T3 specimen exhibits charge-transfer, mixed and mass-
transfer-controlled regimes. Dufek et al. [327], on the other hand, applied cyclic voltammetry and
linear sweep voltammetry (LSV) in combination with a wall-jet flow cell to investigate the
influence of O2 reduction on oxide growth and the dissolution processes at the surface of an
AA2024-T3 specimen. The current response of the sample was caused by the competition
between interfacial etching of the surface oxide, a consequence of local alkalization promoted by
the ORR, and anodic oxide growth on the alloy surface.
These electrochemical methods have, however, low lateral resolution and their application for
the in-depth investigation of the corrosion processes on a single IMP is not possible. To increase
the lateral resolution of the potentiodynamic polarization technique Suter et al. [328] used a
microelectrochemical cell limiting the exposed area between the electrolyte and the alloy sample
to a few hundreds of square micrometers. The authors reported the onset of pitting at S-phase
inclusions occurring within 0.2 V of the pitting potential recorded for large scale measurements,
while the onset of pitting at areas without inclusions occurred at potentials higher than 0.5 V vs.
SCE. The highest reported resolution was obtained using a capillary with 20 µm diameter. Even
with this resolution approximately 20 % of the response from the exposed specimen surface
originated from AlCuFeMn inclusions, with the remaining 80% caused by the surrounding solid
solution matrix. While this resolution is sufficiently high for investigation of the onset of pitting at
different IMPs in an AA2024-T3 alloy, it is still not sufficient for in situ investigation of the
electrochemical processes which occur only on the surface of single IMP. Llevbare et al. [329]
applied the potentiodynamic polarization technique to investigate the ORR kinetics on an
AA2024-T3 specimen. They also synthesized Al-Cu, Al-Cu-Mg and Al-Cu-Mn-Fe intermetallic
phases and the desired intermetallic phases were selectively addressed using different
metallographic masking techniques.
Results and Discussion
165
Since being introduced by Bard et al. [8], very few papers concerning the application of SECM
for the investigation of the corrosion processes of AA2024-T3 were published until now.
Seegmiller et al. [330] applied SECM to study the heterogeneous cathodic activity at AA2024-T3
using a 10 µm Pt microelectrode as SECM tip. When the UME was polarized at +0.65 V vs. SCE,
the protonated form of (dimethylaminomethyl) ferrocene (DMAFc+) can be oxidized to
DMAFc2+. The AA2024-T3 sample was polarized at -0.75 V vs. SCE, where DMAFc2+ is reduced
back to DMAFc+ only at cathodically active regions on the specimen surface. Comparison of the
SECM images with data obtained by SEM and EDX showed good correlation between regions of
high cathodic activity and the location of the IMPs. However, the reported lateral resolution was
insufficient to resolve a single IMP and a good contrast was achieved only over rather big groups
of particles located in proximity to each other.
In summary, the in situ visualization of the consumption of molecular O2 on the surface of a
single S-phase IMP, with a sufficiently high lateral resolution, still remains a very challenging
issue and has not been reported previously. Thus, the results and discussion presented in the
following part of this work focuses on the application of the developed SKP-SECM system for
high-resolution visualization of the in situ oxygen reduction reaction on the surface of a single S-
phase crystal, as well as on the surface of the surrounding solid solution Al matrix using insulated
“glass free” Pt SKP-SECM tips. As the size of the S-phase particles in commercially available
aluminum alloys ranges from tenths of nanometers to the sub-µm range, to access electrochemical
information from a single S-phase particle using the integrated SKP-SECM system an aluminum
bulk alloy with single S-phase crystals ranging from 50 µm to 100 µm randomly distributed in the
solid solution Al matrix had to be synthesized prior to the combined SKP-SECM measurements.
4.8.2.2. Synthesis of bulk Al2CuMg single crystals
Synthesis of pure, bulk Al2CuMg intermetallic phase is a very complicated issue and very few
reports on the successful synthesis are available. Many attempts to produce large S-phase crystals
by solidification of a liquid melt containing single components with an atomic ratio
Al:Cu:Mg = 2:1:1 failed due to the formation of heterogeneous ingots containing intimate
mixtures of several different phases. Buchheit and co-workers [331] reported the synthesis of
large lenticular S-phase crystals with dimensions ranging from 100 µm to 200 µm in width and
several millimeters in length using isothermal treatment of an Al-24.5 Cu-10.1 Mg melt at 510˚C
for 65 h.
Results and Discussion
166
In a collaboration with Dr. Ceylan Senöz (MPIE, Düsseldorf) single S-phase crystals were
synthesized by melting the corresponding proportions of pure metal powders, namely 46.4 g Al
(99.99 %), 38 g Cu (99.99 %) and 18 g Mg (99.99 %) by heating a homogenized mixture of
powders at a rate of 8˚C min-1 to 950˚C and maintaining the temperature at 950˚C for 15 min to
ensure complete melting. The melt was then mechanically agitated and cooled down to room
temperature. The homogenization of the melt was done using isothermal treatment by heating to
510˚C for 65 h, leading to segregation into the solid S-phase crystals and an Al-rich liquid [332].
Long isothermal treatment was used to obtain the desired size of single S-phase crystals. Finally,
the melt was water-quenched to freeze the microstructure and the chemical composition of the
desired Al2CuMg intermetallic phase. The ingot was then cut into smaller pieces and samples
were mechanically ground with successively finer SiC abrasive paper lubricated with kerosene.
Polishing to a smooth surface was performed with non-aqueous concentrated diamond based
slurries 3 µm, 1 µm and 0.25 µm and DP lubricant. The polished samples were degreased by
ultrasonic cleaning in pure ethanol and dried in an argon stream.
The synthesis of bulk Al2CuMg crystals described above is very similar to that originally
described by Buchheit and co-workers [331]. However, two notable modifications were made to
the originally reported synthesis procedure. Firstly, a different atomic ratio of the single metals
(Al : Cu : Mg = 1 : 0.819 : 0.388 vs 1 : 0.941 : 0.412 used in [331]) for the formation of the melt
before the solidification process was used. Secondly, following the complete melting of the metal
mixture the melt was cooled down to room temperature and then heated again to 510˚C for 65 h in
order to form smaller S-phase crystals with dimensions of around 50 µm to 100 µm, well suited
for further combined SKP-SECM experiments.
4.8.2.3. Characterization of bulk Al2CuMg single crystals
The size and chemical composition of the synthesized Al2CuMg intermetallic phase was
investigated with EPMA analysis (Fig. 4.80). Single Al2CuMg crystals mostly displayed an oval
shape with dimensions of around 60 µm to 100 µm in width and 100 µm to 150 µm in length.
Single S-phase crystals were predominantly separated by the pure solid solution Al matrix.
However, occasionally areas enriched in Mg were also observed (Fig. 4.80 A and Fig. 4.80 D,
area marked with a dotted circle).
Results and Discussion
167
D
Figure 4.80. SEM image of the synthesized Al bulk alloy (A, Al2CuMg single crystals appear as bright areas while the solid solution Al matrix appears as dark areas). EPMA analysis of synthesized
Al2CuMg phases with B, C and D being 2D-EDX mapping of Al, Cu and Mg. Red (A) and yellow (D) dotted circles indicate an area enriched in Mg.
For a better overview of the chemical composition of the single Al2CuMg crystals in
comparison with the available literature data, results of the EPMA analysis performed on 25
randomly chosen particles are shown in Table 11. The average chemical composition of the
synthesized Al2CuMg single crystals is in a very good agreement with the data reported by
Lacroix and co-workers [333] which represents the average chemical composition of 20 single S-
phase intermetallics found in a statistical study of a commercially available AA2024-T351 alloy.
Table 11. Average chemical composition of 25 randomly chosen Al2CuMg single crystals compared with data from literature.
A
C
B
Results and Discussion
168
The crystallographic structure of the synthesized Al2CuMg intermetallic phase was determined
by using standard X-ray diffraction analysis on small portions of the pulverized bulk ingot
(Fig. 4.81). The peak positions for the S-phase single crystals and solid solution Al matrix
obtained by XRD analysis were analyzed and are in a good agreement with the reference pattern
from the database provided by Bruker. It is believed that the few additional reflections from non-
identified phases are caused by small deviations in the composition from the equilibrium
composition [334] or the presence of “planar, stacking fault” type defects [335].
Figure 4.81. XRD spectra of the synthesized Al2CuMg intermetallic phases (triangles are indicating peaks identified to belong to the S-phase). The non-identified peaks are marked with U.
To cross check the phase composition of the synthesized Al2CuMg single crystals
backscattered electron Kikuchi patterns (BEKPs) were recorded (Fig. 4.82).
Figure 4.82. Backscattered electron Kikuchi patterns recorded using the synthesized Al2CuMg intermetallic phases
(A) and simulated by the software pattern of the orthorhombic Al2CuMg phase (B).
A B
Results and Discussion
169
The experimental patterns of 25 randomly chosen single S-phases were analyzed and indexed
according to the positions of the zone axes using the algorithm provided by the software package
provided by Carl Zeiss. Finally, simulated patterns were compared with the experimental patterns.
The experimental patterns were consistent with the pattern for orthorhombic Al2CuMg phase
available in the data base provided by Carl Zeiss Company.
4.8.2.4. SKP-SECM measurements on synthesized Al2CuMg single crystals
The first step in the combined investigation of the synthesized Al bulk alloy was performing
SKP measurements on the surface of the freshly polished sample. Despite very careful preparation
of the sample using high-quality SiC abrasive paper with successively finer grains and diamond
based water-free polishing slurries with particle sizes as small as 250 nm in combination with a
range of polishing clothes, the large variation in the hardness of the solid solution matrix and
Al2CuMg single crystals caused the S-phase crystals to slightly protrude out of the sample surface.
The preliminary SKP measurements, performed using an insulated “glass free” 5 µm Pt SKP-
SECM tip while purging water saturated air through the chamber around the bulk Al alloy,
revealed topographical variations of up to 300 nm across the sample and intensive corrosion of
Al2CuMg single crystals (Fig. 4.83). Figure 4.83 A shows a photographic image of the polished
sample surface before the SKP measurement. This image was made using a polarization filter
which provided Al2CuMg single crystals with different crystallographic orientations slightly
different coloring (from bright-grey to dark-grey). Slight protruding of the S-phases out of the
solid solution Al matrix can be recognized by barely visible shadow lines at the edges of
Al2CuMg crystals as the light source was not completely orthogonal to the sample surface.
Figure 4.83 B shows a photographic image of the sample surface after the SKP measurement.
This image was also made with a polarization filter which leads to a coloring of the corrosion
tranches in black while leaving the S-phase crystals and the solid solution Al matrix bright. To
avoid corrosion attack on Al2CuMg crystals the air supply to the chamber around the sample was
modified (Fig. 4.84). Additional valves and tubing were connected in such a way that it becomes
possible to choose between operation of the integrated SKP-SECM system at ambient air
conditions (with / without pumping of the air through the chamber) or with pumping of water
saturated air, or under a desired gas atmosphere (dry or H2O saturated).
Results and Discussion
170
100 µm 100 µm
Figure 4.83. Photographic images of the bulk Al alloy before (A) and after (B) the SKP measurement (H2O saturated air was continuously pumped through the chamber around the sample). Polarization filters were used to amplify
the coloring of single Al2CuMg crystals (A) as also of corrosion tranches all over the sample (B).
Figure 4.84. Schematic diagram of tubing and valves connection enabling operation of the integrated SKP-SECM system under ambient air conditions (with or without pumping of the air) and
with water saturated air being pumped through the chamber, or under desired gas atmosphere (dry or water saturated gas).
An uneven sample surface can lead to an unwanted collision of the tip with the sample which
leads to saturation of the operational amplifier and measurements with erroneously high CPD
values. To avoid this, the software based feedback controller responsible for the tip-to-sample
distance regulation was programmed in such a way that the scanning tip was retracted 1 µm from
the sample surface before moving to the next measurement point. Based on the differences in the
measured CPD values of the solid solution Al matrix and the Al2CuMg single crystals (∆ CPD)
three single S-phase crystals were localized (Fig. 4.85). It was found that, as a result of the higher
Cu content, these Al2CuMg phases were 0.35 V nobler than the solid solution Al matrix.
A B
Results and Discussion
171
Figure 4.85. Line scan over the synthesized Al2CuMg bulk sample in the SKP mode performed using a 5 µm insulated “glass free” SKP-SECM tip (tip-to-sample distance = 250 nm, tip-oscillation frequency = 1200Hz,
tip-oscillation amplitude = 5 µm). Darker areas correspond to the solid solution matrix, brighter areas correspond to the Al2CuMg single crystals. Arrows indicate sites of the sample where the
approach curves were recorded in the feedback mode of SECM.
In the second step, SECM measurements were performed in the feedback mode of SECM over
exactly the same sample area as the measurements performed in the SKP mode. Figure 4.86
represents approach curves to the sample surface recorded in 2.5 mM [Fe(CN)6]4-/ 100 mM
Na2SO4 / 5 mM NaCl over the positions marked with arrows in figure 4.85 distingui-shing
between the solid solution Al matrix and a single S-phase crystal. The tip was polarized at
+0.35 V vs. Ag/AgCl allowing the diffusion-controlled oxidation of Fe2+ to Fe3+ while the sample
was kept at OCP (≈ -0.8 V vs. Ag/AgCl). At a certain tip-to-sample distance the Fe3+ generated at
the tip can reach the sample surface and be reduced to Fe2+ at sample areas active to cathodic
reactions (i.e. single Al2CuMg crystals). The local increase in the Fe2+ concentration will lead to a
positive feedback current as the tip approaches the surface of a single S-phase crystal. Negative
feedback is observed during the approach to the solid solution Al matrix surface. All approach
curves were recorded to the tip contact with the sample surface leading to short-circuiting of the
cell which was registered as an overload of the potentiostat. The information obtained by the
controlled approach to the sample surface was then additionally used for the software based tilt
correction procedure to keep the tip-to-sample distance constant during scanning.
Results and Discussion
172
Figure 4.86. Current-distance curves obtained using a 5 µm insulated “glass free” SKP-SECM tip (1 – over the solid solution Al matrix, negative feedback; 2 – over Al2CuMg single crystal, positive feedback). The tip was
polarized at +0.35 Vvs. Ag/AgCl (diffusion controlled oxidation of [Fe(CN)6]4-); 2.5 mM [Fe(CN)6]
4- / 100 mM Na2SO4 / 5 mM NaCl; sample was kept at OCP; iT, - current at the tip; iT,∞ = 2.5 nA;
d – tip-to-sample distance; reff = 2.55 nA).
Directly after determination of the relative sample position a line scan in the feedback mode of
the SECM with a length of 325 µm was performed over the sample surface at a tip-to-sample
distance of 1 µm. Schematic representation of this experiment is shown in figure 4.87.
Figure 4.87. Schematical representation of the experiment performed over the
synthesized Al bulk alloy in the feedback mode of SECM.
During scanning the tip was polarized at a potential allowing the diffusion-controlled oxidation of
Fe2+ to Fe3+ (Etip = +0.35 V vs. Ag/AgCl, Esample = OCP) while the anodic current measured at the
tip was monitored as a function of the x-position of the tip. At a certain tip-to-sample distance, the
Fe3+ generated at the tip can reach the sample surface and be reduced to Fe2+ at regions of the
Results and Discussion
173
sample that are active to cathodic reactions. The local increase in the Fe2+ concentration will lead
to an increase of measured anodic current. On the other hand, if the surface of the sample
underneath the tip does not exhibit a cathodic activity, the decrease of the measured current will
be observed as the sample surface starts to block the diffusion of Fe2+ to the active surface of the
SECM tip.
The same three S-phase crystals as identified during the SKP measurement could be localized
at identical positions with an excellent contrast and high lateral resolution due to higher anodic
currents over single Al2CuMg crystals (Fig. 4.88).
Figure 4.88. Line scans over the synthesized Al bulk alloy in the feedback mode of SECM (solid line) and SKP (dashed line) mode performed using an insulated 5 µm “glass free” Pt SKP-SECM tip
(tip-to-sample distance during SECM measurement = 1 µm; Etip = +0.35 V, Esample = OCP2.5 mM [Fe(CN)6]
4- / 100 mM Na2SO4 / 5 mM NaCl).
Following the line scan in the feedback mode of the SECM the solution inside of the
electrochemical cell was exchanged with an O2 saturated solution containing 100 mM Na2SO4/
5 mM NaCl and a line scan was recorded in the redox-competition mode of the SECM over
exactly the same area on the sample surface as during the previous measurements. The tip was
polarized at -0.6 V vs. Ag/AgCl leading to the diffusion controlled reduction of oxygen at the
active tip surface, while the sample was polarized at 0.05 V below its OCP (Esample = -0.85 V vs.
Ag/AgCl). During scanning the O2 reduction current was recorded at the tip as a function of the x-
position of the tip. If the tip will be positioned in close proximity to the single Al2CuMg crystal
which is active for the reduction of oxygen, the cathodic current at the tip will decrease as
compared with the ORR current measured over the solid solution Al matrix. This difference in the
measured ORR currents will result from the enhanced consumption of dissolved oxygen on the
Results and Discussion
174
surface of a single S-phase crystal leading to a smaller amount of remaining oxygen available for
the tip (Fig. 4.89).
Figure 4.89. Schematic representation of the experiment performed over the synthesized Al bulk alloy in the RC mode of SECM.
The same three S-phase single crystals identified during the previous SKP and feedback-mode
SECM measurements could be localized at identical positions with a very good contrast and high
lateral resolution due to the smaller cathodic currents observed over single Al2CuMg crystals
caused by their higher ORR activity relative to the solid solution Al matrix (Fig. 4.90).
Figure 4.90. Line scans over the synthesized Al bulk alloy in the RC-SECM (solid line) and the SKP (dashed line)
modes performed using an insulated 5 µm “glass free” Pt SKP-SECM tip (tip-to-sample distance during SECM measurement = 1 µm, Etip = -0.6 V vs. Ag/AgCl, Esample = -0.85 V vs. Ag/AgCl (0.05 V below OCP),
O2 saturated 100 mM Na2SO4 / 5 mM NaCl solution).
Results and Discussion
175
A
D
B
C
Investigation of the Al bulk sample directly after experiments in the feedback and redox-
competition modes of SECM in electrolytes containing 5 mM NaCl, but where one was saturated
with air and the other with oxygen, revealed different behavior of the synthesized alloy.
Figure 4.91 A shows a photographic image of the sample surface directly after the SECM
measurement in feedback mode (Esample = OCP). A few small corrosion pits were observed on the
surface of single S-phase crystals while the solid solution Al matrix remained intact.
Figure 4.91. Photographic images of the surface of Al bulk sample after SECM experiments performed in air
saturated electrolyte containing 2.5 mM [Fe(CN)6]4- / 100 mM Na2SO4 / 5 mM NaCl in feedback
mode (A) and in oxygen saturated solution consists of 100 mM Na2SO4 / 5 mM NaCl in redox-competition (B-D) mode.
On the other hand, figures 4.91 B, C and D show photographic images of the same sample
directly after two successive SECM measurements in feedback and RC-SECM modes. To
promote the oxygen reduction reaction during RC-SECM experiments the sample was polarized at
-0.85 V vs. Ag/AgCl (0.05 V below OCP). Much larger corrosion pits were observed in
combination with deposition of corrosion products around the pits on the surface of single S-phase
crystals indicating the preferential dissolution of Al2CuMg single crystals following breakthrough
Results and Discussion
176
of the protective oxide layer. This leads to redistribution of Cu originally present in the Cu-rich S-
phase crystals all around the corrosion pits (yellow to orange colored aureole). Enrichment with
Cu causes an enhancement in the cathodic activity of the sample surface covered with Cu clusters
towards oxygen reduction. Increased O2 reduction promotes local alkalization which further
increases the dissolution rate of S-phase crystals. This effect can be clearly seen in figure 4.91 C
where the increased pH level inside the crack leads to active dissolution of the inner edges of a
single Al2CuMg crystal. Further alkalization will lead to the formation of a deep trench on the
aluminum matrix encircling the particle. As can be seen in figure 4.91 D, in some cases,
dissolution of the solid solution Al matrix leads to detachment of S-phase particles. These
observations of the corrosion behavior for synthesized single S-phase crystals, randomly
distributed in solid solution Al matrix, are in a good agreement with data previously reported by
Buchheit and co-workers [331].
4.8.2.5. Conclusion
To access electrochemical information from a single S-phase crystal an Al bulk alloy was
synthesized by solidification of a liquid melt of pure Al, Cu and Mg. A long term isothermal
treatment was applied to form single Al2CuMg crystals with the desired dimensions and chemical
composition. The obtained Al2CuMg single crystals were randomly distributed in the solid
solution Al matrix and mostly displayed an oval shape with dimensions of around 50 to 100 µm
width and 100 to 150 µm length. Their average chemical composition was in very good agreement
with literature data for S-phase IMPs in commercially available AA2024-T351 aluminum alloys.
Using insulated “glass free” Pt SKP-SECM tips, single S-phase crystals were successfully
localized in the SKP mode of the SKP-SECM system. Higher cathodic activity on the surface of
the single Al2CuMg crystal was indirectly visualized in the feedback mode of the SECM and the
in situ consumption of dissolved O2 at the surface of a single S-phase crystal was visualized in the
redox-competition mode of SECM. Smaller cathodic currents recorded over a single S-phase
crystal indicate a more intensive consumption of O2 at the surface of S-phase crystals than on the
surrounding solid solution Al matrix. This confirms the action of Al2CuMg phases as local
cathodes during active corrosion process.
Conclusions
177
5. Conclusions
Prior to the development of the integrated SKP-SECM system preliminary experiments were
performed on a model sample to evaluate the applicability of a sequential employment of SKP
based techniques with in-situ 4D-AC and RC modes of the SECM (Chapter 4.1). A model
sample that mimics an Al based alloy was prepared by sputtering pure Cu through a mask onto a
freshly evaporated Al layer which led to the formation of micrometer sized Cu islands. Mapping
of the CPD distribution over the prepared sample was performed using conventional SKP and
SKPFM. The ∆CPD measured over the Cu structures and Al sample was around 0.6 V being in a
good agreement with data reported in literature. To localize the deposited Cu structures the 4D-
AC mode of SECM was employed. The optimal frequency that provided the best contrast between
Cu covered areas and Al surface was found to be 2.31 kHz. Two Cu regions were resolved with a
high contrast based on their higher conductivity in comparison with the surrounding Al surface
covered with a native Al oxide layer. Results of the AC-SECM experiments clearly confirmed the
high sensitivity of the AC mode toward distribution of local conductivity making this mode of
SECM an ideal tool for localization Cu rich structures on the surface of an Al alloy. The catalytic
ORR activity of the localized Cu structures was visualized in the RC mode of SECM directly after
the 4D-AC-SECM measurement using the same 25 µm glass insulated Pt SECM electrode
without changing solutions or replacing the sample. RC-SECM experiments revealed higher ORR
activity on the surface of Cu islands than on the surrounding Al surface. The obtained results
clearly indicate the complementary nature of the information about the distribution of
heterogeneities on the sample surface and local faradaic processes that take place on these
heterogeneous areas during active corrosion in an electrolyte received from combined
SKP / SKPFM and SECM measurements.
The design and development of the integrated SKP-SECM system focused on the ability of the
new system to perform sequential measurements in SKP and SECM modes over exactly the same
area on a sample surface without the necessity of changing the tip or replacing the sample
(Chapter 4.2.1). Based on the longtime experience of collaboration partners at the Max-Plank
Institute for Iron Research the vibrating capacitor technique was integrated into the SECM set-up.
To minimize the influence of time dependent stray coefficients, and thus stray capacitances, the
integrated SKP-SECM system was designed in such a way that the Kelvin probe was at low
impedance (earth potential) while the sample under investigation was connected to the input of an
ultra-low noise operational amplifier with high impedance. As the SECM experiments require an
Conclusions
178
insulated disk electrode whereas the demands for a SKP tip are less rigorous, a glass insulated Pt-
disk microelectrode was selected as tip for both SKP and SECM measurements.
To measure the contact potential difference both the nulling mode and the off-null mode of
operation were successfully implemented into the developed SKP-SECM system (Chapter 4.2.1).
To enable the operation of the vibrating electrode at high frequencies and to maintain high
positioning precision, a pre-loaded piezo actuator especially designed for dynamic applications
was used. The choice of the actuator was based on theoretical calculations of the main operational
parameters. Very low off-axis displacement of the vibrating electrode was achieved using a high-
efficiency suspension system implemented in the measurement head. A choice between operation
of the developed SKP-SECM system in SKP or in SECM modes can be made by actuating two
reed relays. Additionally, to perform experiments in the SECM mode the electrochemical cell
must be filled with an electrolyte and a reference as well as an auxiliary electrode needs to be
connected to the bipotentiostat.
The technical realization and function of each mechanical (Chapter 4.2.2) and electrical
(Chapter 4.2.3) component of the integrated SKP-SECM system were discussed in detail. Great
attention was paid to the design and development of a multifunctional measurement head that
enables operation of the Kelvin probe / working electrode in dynamic or static regimes
(Chapter 4.2.2.1). The vibration amplitude of the Kelvin probe can be adjusted by using a high-
frequency power supply while controlling the vibration frequency with the lock-in amplifier. The
position of the measurement head above the sample surface is controlled by three step motors
while positioning of the tip and the tip-to-sample distance adjustment can be performed using a
NanoCube positioning unit which offer a lateral resolution in the sub-nm range. Detailed
description of the assembling procedure for the measurement head was given (Chapter 4.2.2.2).
To maximize the signal-to-noise ratio the operational amplifier was mounted in a well-shielded
sample holder unit as far away as possible from the piezo actuator (Chapter 4.2.2.3). To
efficiently insulate the integrated SKP-SECM system from ground vibrations a special combined
jig consisting of three main units with each unit selectively adjusted to insulate vibrations in a
predefined frequency range was developed (Chapter 4.2.2.4). Deliberate selection of materials
used for fabrication of any mechanical component of the system ensured the maximized
performance of the integrated SKP-SECM system. Additionally, great attention was paid to keep
the weight of individual components at least 20 % below the maximum load specified by the
manufacturer especially in the case of sensitive components such as step motors and the
NanoCube positioning unit.
Conclusions
179
To maximize the performance of the integrated SKP-SECM a combination of custom build
electrical components such as a Kelvin current amplifier and an integrator as well as a few
commercially available instruments such as a high-frequency power supply, a lock-in amplifier, a
step motors controller, a NanoCube controller and an analog-to-digital conversion card were used.
The design of the Kelvin current amplifier (KC amplifier) was based on a state of the art ultra-low
noise operational amplifier AD549LH which offers superior performance and noise characte-
ristics. The electric circuit diagram of the KC amplifier and function of each circuit component
were discussed in detail (Chapter 4.2.3.1). The integrator was designed in such a way that manual
or automatic modes of operation can be selected while proper function of the operational amplifier
or an accidental short-circuit between the Kelvin probe and the sample can be monitored online.
The electric circuit diagram of the integrator and function of each unit were discussed in detail
(Chapter 4.2.3.2). The choice of specific high-frequency power supply was based on theoretical
calculations of required performance which relied on the technical specification of the piezo
actuator (Chapter 4.2.3.3). To amplify and recover information about the induced Kelvin current
in the presence of overwhelming background noise the integrated SKP-SECM system utilizes two
analog dual phase lock-in amplifiers. The scheme for the lock-in amplifiers connection within the
integrated SKP-SECM system and also the processing procedure of the Kelvin current are
discussed in detail (Chapter 4.2.3.4). Additionally, performance and some selected technical
specifications of peripheral electrical components implemented into the developed SKP-SECM
system such as a step motors controller, a NanoCube controller and an analog-to-digital
conversion card were also discussed (Chapter 4.2.3.5).
To enable proper functioning of the integrated SKP-SECM system each built-in
component / device was adjusted in accordance to the requirements of the other components. The
setting procedure and the operation aspects of the developed SKP-SECM system were discussed
in detail (Chapter 4.3.1). To ensure the reliability of the CPD values measured with the
developed SKP-SECM system, system calibration was performed using a set of pure bulk metal
reference samples (Ni, Cu and Pt). The obtained values for the CPD were in good agreement with
the data presented in the literature for similar samples (Chapter 4.3.2). To estimate the influence
of the parasitic signals potentially present in the developed SKP-SECM system a set of
measurements using the nulling technique (unfavorable signal-to-noise ratio) and the off-null
technique (best signal-to-noise ratio) were performed. The observed difference between CPD
values measured over a Ni reference sample was about 19 mV. This indicated a rather small
influence of the parasitic signal present in the integrated SKP-SECM system on measured CPD
Conclusions
180
values (Chapter 4.3.3). However, representation of the CPD measured over the sample under
investigation vs. the CPD of the Ni reference sample completely neglects the presence of any
disturbing signals.
The lateral resolution of the developed SKP-SECM system was evaluated using 25 µm and
125 µm glass insulated Pt SKP-SECM tips in both SKP and SECM modes. With SECM the
resolved width of the W / Pt edge (reference sample for evaluating the lateral resolution) was
found to be equal to the diameter of the active area of the electrode while in the SKP mode the
same edge was resolved to be equal to the outer diameter of insulating glass sheath at the very end
of the tip. This effect was attributed to electrostatic charging of the glass surface (Chapter 4.3.4).
To overcome this effect two different approaches were tested. The application of water saturated
air was evaluated after development of a new sample holder. It was found that purging H2O
saturated air through the chamber around the sample leads to almost three times increase of the
distance dependent signal resulting in a higher sensitivity of the tip-to-sample distance control
system. Additionally, elimination of the influence of the tip-to-sample distance on the measured
CPD value was observed. The modification of the sample holder unit additionally decreased the
noise level by up to 10 µV (Chapter 4.3.6). The second strategy to compensate the electrostatic
charging of the glass surface was based on application of an external compensation voltage
applied to a thin Ag layer deposited only on the side walls of the insulation glass sheath of the Pt
SKP-SECM tip (Chapter 4.3.7). Unfortunately, neither the first nor second approach could
completely discharge the glass surface at the very top of the tip.
To improve the lateral resolution of the integrated SKP-SECM system a new concept of “glass
free” Pt SKP-SECM tips was developed (Chapter 4.5.1) while the noise signal was decreased by
up to 3 µV after revision and optimization of all mechanical (Chapter 4.4.1) and electrical
(Chapter 4.4.2) components of the developed SKP-SECM system. The concept of the “glass
free” Pt SKP-SECM tips is based on etching the very top part of the glass insulated Pt-disk
microelectrodes in concentrated HF (40 %) for about 250 µm to 300 µm which leads to partial
exposure of the Pt wire (Chapter 4.5.2). If the desired outer diameter of the “glass free” Pt SKP-
SECM electrode was smaller than the original diameter of the Pt wire, the Pt wire was thinned
using electrochemical etching and subsequent electrochemical micropolishing. Both procedures
were optimized to reproducibly etch the Pt wire to give “glass free” Pt SKP-SECM tips with
parallel side walls and a planar surface at the very top of the tip (Chapter 4.5.3). Subsequent
coating of the exposed Pt wire was performed in such a way that only the side walls of the
exposed Pt wire were insulated while the Pt-disk surface at the very top of the tip remained
Conclusions
181
uncoated. A set of fifteen commercially available materials were tested for their ability to be
deposited on the Pt wire surface to form a thin insulating layer and for their resistivity in 100 mM
Na2SO4 / 5 mM NaCl solution. Surprisingly, a thin layer of nail varnish showed the best insulating
properties amongst all tested potential candidates (Chapter 4.5.4). Using the novel “glass free” Pt
SKP-SECM tips the working parameters of the integrated SKP-SECM system were optimized
(Chapter 4.6). The maximum Kelvin current was achieved by operating the system at 1200 Hz
with 12 µm amplitude of the tip oscillation.
The performance of the optimized SKP-SECM system was evaluated using the “glass free” Pt
SKP-SECM tips with a range of outer diameters from 8 µm to 50 µm (Chapter 4.7). The
elimination of the glass sheath from the very top of the tip led to a decrease of the equilibration
time of the system in the SKP mode of more than four times to 20 min. Additionally, an increase
of the relative air humidity inside the measurement chamber did not lead to an increase of the
measured HR signal which makes application of the “glass free” Pt SKP-SECM electrodes
especially attractive when fast switching between experiments under ambient conditions and at
the high relative air humidity is needed (Chapter 4.7.1). On the other hand, the CPD values
measured over the reference Ni sample were found to be dependent on the tip-to-sample distance.
To simultaneously enable high-resolution measurements and to eliminate the described
dependency a software based tilt correction procedure (Chapter 4.7.2.1) and a software based
feedback controller module (Chapter 4.7.2.2) were implemented into the integrated SKP-SECM
system.
Evaluation of the lateral resolution of the integrated SKP-SECM system was performed with
the “glass free” Pt SKP-SECM electrodes which showed excellent agreement between the
resolved width of the W / Pt edge and the outer diameter of the Pt wire at the very top of the tip
(Chapter 4.7.3). The accuracy of positioning the “glass free” Pt SKP-SECM electrode during the
operation of the integrated SKP-SECM system in the SKP mode was estimated by analyzing 1000
data points recorded during an area scan over a flat sample (Chapter 4.7.4). The average standard
deviation of the z-position of the tip was 22 nm while the maximal difference between two values
was about 102 nm. The vertical resolution of the integrated SKP-SECM system during operation
in the SKP mode was investigated with line scans over the edge of 100 nm thick structured Pt
layer deposited on top of the underlying W layer using a 8 µm “glass free” Pt SKP-SECM tip. For
comparison, an area scan over the same edge was performed using a conventional AFM set-up in
the tapping mode. The obtained results (∆Z = 83 nm vs. ∆Z = 87 nm) clearly demonstrate very
high vertical resolution of the integrated SKP-SECM system (Chapter 4.7.4).
Conclusions
182
Two examples were chosen to demonstrate the performance and applicability of the integrated
SKP-SECM system developed and optimized during this work. The first example focused on
linking the contact potential difference (work function) with catalytic activity towards the oxygen
reduction reaction of a novel high-performance catalyst subsequently probed using the developed
system in the SKP and in the redox-competition mode of the SECM (Chapter 4.8.1). Prior to the
actual SKP and SECM experiments the preparation procedure of a novel Pt(111)-like thin film
electrodes as samples for the following synthesis of the high-performance ORR catalyst was
developed (Chapter 4.8.1.3.2). The prepared Pt(111)-like thin films were investigated using
SEM, AFM and cyclic voltammetry. Investigation of the topography of the Pt(111)-like thin films
revealed long range order with large atomically flat terraces of up to 100 Pt atoms separated by
single monoatomic steps. Analysis of the electrochemical response of the new Pt(111)-like thin
film electrodes indicated characteristic “butterfly” peaks with height and sharpness comparable to
that of the state of the art well-ordered Pt(111) single crystals (Chapter 4.8.1.3.3).
A recently reported Cu-Pt near-surface alloy (NSA) was prepared on the Pt(111)-like thin film
electrodes by electrochemical underpotential deposition of a monolayer of Cu followed by
annealing of the sample under a reducing atmosphere of Ar / H2. Thermal treatment of the
prepared sample was applied to force migration of the deposited Cu atoms into the second Pt-layer
(Chapter 4.8.1.4.2). Cyclic voltammetry on the prepared Cu-Pt(111) NSA revealed ORR activity
about 5.5 times higher than that of the state of the art Pt nanostructured thin films
(Chapter 4.8.1.4.3).
The Pt(111)-like thin film electrodes were applied to evaluate the sensitivity of the developed
SKP-SECM system towards different amounts of Cu ranging from 1 ML to about 11 ML
deposited on the electrode surface. It was shown that the SKP-SECM system is able to probe and
identify extremely small differences in the amounts of Cu present on the surface of the novel
Pt(111)-like thin films (Chapter 4.8.1.5.1). To assess the influence of Cu atoms located in the
second Pt layer on the local CPD distribution, area scans over the Cu-Pt(111) NSA thin films were
performed in the SKP mode while local catalytic activity of the sample surface was visualized in
an oxygen saturated 0.1 M HClO4 solution using RC-SECM. It was shown that the presence of Cu
atoms in the second Pt layer caused remarkable changes in the properties of the Pt(111)-like thin
film surface and led to a shift of the measured CPD in positive direction by around 150 mV
(increased nobility). The RC-SECM experiments confirmed the enhanced ORR activity of the
Pt(111)-like thin film surface modified with Cu-Pt(111) NSA (Chapter 4.8.1.5.1). Thus, the
connection between surface property (work function) and local catalytic ORR activity of Pt(111)-
Conclusions
183
like thin films and also of the Cu-Pt(111) NSA thin films was successfully elucidated using the
integrated SKP-SECM system.
The second example focused on localization of a single Al2CuMg crystal in a solid solution Al
matrix performed in the SKP mode using the insulated “glass free” SKP-SECM tips and
subsequent localized visualization of in situ consumption of O2 at the surface of this single crystal
using RC-SECM (Chapter 4.8.2). Single S-phase crystals were synthesized by melting pure metal
powders while a long isothermal treatment was applied to segregate the melt into the solid S-
phase and an Al-rich liquid. A final water quenching was used to freeze the desired microstructure
and chemical composition of the Al2CuMg single crystals (Chapter 4.8.2.2). The size and
chemical composition of the synthesized S-phase crystals were investigated with EPMA analysis.
It was found that the synthesized Al2CuMg crystals mostly displayed an oval shape with
dimensions around 60 µm to 100 µm in width and 100 µm to 150 µm in length while their
chemical composition was in good agreement with data from literature (Chapter 4.8.2.3). Three
Al2CuMg single crystals were successfully localized in the SKP mode. Subsequent experiments
on the localized S-phase crystals in the feedback mode of SECM revealed higher cathodic activity
on their surface while the in situ consumption of dissolved O2 on the surface of single S-phase
crystals was visualized using RC-SECM (Chapter 4.8.2.4). Smaller cathodic currents were
recorded over a single Al2CuMg crystal which indicated more intensive consumption of O2 on the
surface of S-phase crystals than on the surrounding solid solution Al matrix. This confirms the
action of Al2CuMg phases as local cathodes during active corrosion of Al alloys containing such
Cu-rich phases.
Outlook
184
6. Outlook
The integrated SKP-SECM system developed during this work was shown to be a powerful
tool for the investigation of fundamental properties of sample surfaces such as the contact
potential difference (work function) and electrochemical activity with high lateral resolutions for
both. In the near future, the ability to localize single Al2CuMg crystals is planned to be applied for
the time resolved monitoring of Cu release from single S-phase crystal during active corrosion of
synthesized Al bulk alloy to gain further insights into the kinetics of the corrosion process.
Successful application of the integrated SKP-SECM system to localize areas with enhanced
nobility followed by high-resolution in situ visualization of local consumption of O2 can be used
to investigate new catalytic materials concerning enhanced ORR activity. Additionally, after
minor modification of some electrical components, the developed SKP-SECM system can also be
applied for spatial resolution investigations of delamination processes which take place on coated
samples.
Experimental part
185
7. Experimental part
7.1. Materials
7.1.1. Analytical reagents Acetone J. T. Backer, Netherlands
Ammonium hydroxide (NH4OH) J. T. Backer, Netherlands
Calcium chloride (CaCl2) J. T. Backer, Netherlands
Copper (II) oxide (Cu2O) Sigma Aldrich, Germany
Ethanol J. T. Backer, Netherlands
Glucose Merck Group, Germany
Hexaammineruthenium (III) chloride ([Ru(NH3)6]Cl3) ABCR, Germany
Hydrochloric acid (HCl, 38%) VWR International, Germany
Hydrofluoric acid (HF, 40%) VWR International, Germany
Hydrogen peroxide (TraceSelect Ultra® H2O2, 30°%) Sigma Aldrich, Germany
Iso-propanol J. T. Backer, Netherlands
Perchloric acid (Suprapur® HClO4, 1 M) Merck Group, Germany
Potassium chloride (KCl) J. T. Backer, Netherlands
Potassium hexacyanoferrate (II)(K4[Fe(CN)6]Cl4) Riedel-de-Haen, Germany
Silver nitrate (AgNO3) J. T. Backer, Netherlands
Sodium chloride (NaCl) J. T. Backer, Netherlands
Sodium sulfate (Na2SO4) J. T. Backer, Netherlands
Sulphuric acid (Suprapur®H2SO4, 96 %) Sigma Aldrich, Germany
Water (Siemens® UltraClear) Siemens AG, Germany
7.1.2. Solutions
All solutions for electrochemical experiments described in this work were prepared using
Siemens© UltraClear water (0.055 µS/cm, TOC content < 1 ppb).
The quality of the microelectrodes was controlled by cyclic voltammetry in a 5 mM
[Ru(NH3)6]3+ / 100 mM KCl solution. The solution was stored prepared in a light impermeable
container and stored in a fridge.
A “piranha” solution consisting of a mixture of Suprapur® H2SO4 (96%) and TraceSelectUltra®
H2O2 (30%) was used to clean all glassware for 12 h before each electrochemical experiment.
Experimental part
186
Solutions of 0.1 M HClO4 were prepared using Suprapur® HClO4 (Merck) while copper (II)
oxide (99.9999%) was used to prepare solutions of 1 mM CuClO4 + 0.1M HClO4. Prior to
electrochemical experiments all solutions were degassed thoroughly with Ar (5.0) while the
electrochemical cell was kept under an Ar atmosphere at all times.
7.1.3. Consumables AFM cantilevers (ACTA) Applied Nanostructures, USA
Ag / AgCl reference electrode Schott AG, Germany
Ag filled thermoplastic resin (Acheson 1415) Plano GmbH, Germany
Apiezon wax M&I Materials Ltd., USA
Ar/CO (5.0, 0.1°% CO) Air Liquid GmbH, Germany
Ar/H2 (6.0, 5°% H2) Air Liquid GmbH, Germany
Borosilicate glass capillaries Hilgenberg, Germany
Carbon filled cement (Leit-C) Plano GmbH, Germany
Clearclad HSR LVH coating LTD, UK
Cu, Ni, Pt, Ag Goodfellow, Germany
Diamond based polishing suspension Struers GmbH, Germany (3 µm, 1 µm and 0.25 µm) DP lubricant Struers GmbH, Germany
Glass tube (Din = 5 mm, Dout = 8 mm) Schott AG, Germany
Glassophor® ZQ 84-3225 BASF SE, Germany
Lacomit varnish G 371 Agar Scientific, UK
Metal powders of Al, Cu, Mg (99.99%) Goodfellow, Germany
MMS reference electrode Schott AG, Germany
Nail varnish (Maybelline Jade) L'ORÉAL Deutschland GmbH, Germany
Photoresist AR-P 617 Allresist GmbH, Germany
Polishing alumina paste LECO, Germany (3 µm, 1 µm and 0.3 µm) Polishing cloth LECO, Germany
Pt target for magnetron sputtering FHR, Germany
Ti target for magnetron sputtering FHR, Germany
Pt(111) single crystal Mateck GmbH, Germany
Scalpel Aesculap AG, Germany
Experimental part
187
SiC abrasive paper (600 to 2000 grid) Starcke GmbH & Co KG, Germany
Polishing paper (1 µm, 3 µm) 3M, USA
Swagelok® stainless steel tubes Swagelok Company, USA
Sylgard 184 Dow Corning GmbH, Germany
Ti target for magnetron sputtering FHR, Germany
Gas torch Conrad Electronics, Germany
Two-component epoxy glue (UHU plus) UHU GmbH & Co KG, Germany
Two-component epoxy glue (UHU endfest) UHU GmbH & Co KG, Germany
Vibration damping composite plate Bilz vibration Technology AG, Germany
7.1.4. Instrumentation AFM (Digital Instruments Dimension 3100) Veeco Instruments, USA
AFM (NanoWizard 3) JPK Instruments, Germany
Air pump (Tetrateck APS 400) Tetra GmbH, Germany
Analog two channel lock-in amplifier (EG&G 5210) AMETEK, USA
Analog-to-digital conversion card (PCI-DAS6014) Measurement Computing Corp., USA
Bipotentiostat (Jaissle PG 100) Jaissle Elektronic GmbH, Germany
Confocal magnetron sputter (ATC-2200-V) AJA International, USA
DC power supply (2 x 0 -30 V, 2.5 A) Conrad Electronics, Germany
Digital dual-phase lock-in amplifier (7280) AMETEK, USA
eSEM Dual Beam™ Quanta 3D FEG FEI, Japan
Fiber optic light source (EK-1) Euromex, Netherlands
Heating gun Conrad Electronics, Germany
High-frequency generator (33120A) Agilent Technologies, USA
High-frequency power supply (LE 150/100 EBW) Piezomechanik GmbH, Germany
Home-build integrator MPIE Düsseldorf, Germany
Home-build Kelvin current amplifier MPIE Düsseldorf, Germany
JPK SPMControl Station III JPK Instruments, Germany
Experimental part
188
Linear stage (LTM 60) OWIS GmbH, Germany
NanoCube controller card (the E-760.3SV) PI GmbH & Co. KG, Germany
Operational amplifier (AD 549LH) Analog Devices, Germany
Potentiostat (µ-Autolab) Eco Chemie B. V., Netherlands
Potentiostat (Autolab PGSTAT 12) Eco Chemie B. V., Netherlands
Potentiostat (EG&G 273A) Princeton Applied Research, USA
Pre-loaded piezo actuator (PSt 150/5/60 VS10) Piezomechanik GmbH, Germany
Scanning Kelvin probe K&M SoftControl, Germany
SEM (Zeiss Leo Gemini 1550 VP) Carl Zeiss AG, Germany
SKPFM controller (Nanoscope IV) Veeco Instruments, USA
Step motors controller card (PCI-SM32) OWIS GmbH, Germany
Two channel oscilloscope (Voltcraft VC 630-2) Conrad Electronic SE, Germany
Ultrasonic bath Schalltech, Germany
Vibration dampening table (Vario Control Micro 40) Halcyonics, Germany
X-ray detector (EDAX Genesis XM2i) EDAX, USA
X-ray diffraction spectrometer (D8) Bruker AXS, USA
XYZ-nanopositioning system (P-611.3 NanoCube) PI GmbH & Co. KG, Germany
SECM measurements
For SECM experiments a home-build SECM set-up, described previously [24] was used either
in the alternating current mode (AC-SECM) or redox competition mode (RC-SECM) of the
SECM. The key component of the system for operation in AC-SECM mode is a 7280 dual-phase
digital lock-in amplifier (Ametek, USA), which generates the sinusoidal excitation voltage. The
oscillator output of the lock-in amplifier was fed into an EG&G 273A potentiostat (Princeton
Applied Research, USA) via its external voltage input which transfers the modulated potential to
the SECM tip. The current flowing between the SECM tip and the counter electrode was
measured by the potentiostat and fed back into the lock-in amplifier. The amplifier provides the
current magnitude (R) and the phase shift (θ) of the measured alternating current with respect to
the excitation signal. For the RC-SECM measurements a PG100 bipotentiostat (Jaissle Elektronik,
Experimental part
189
Germany) was used to operate the electrochemical cell. During scanning the SECM tip was
polarized at -0.6 V vs. Ag/AgCl while the oxygen reduction current was monitored as a function
of the X and Y tip position. All measurements were carried out in a one-compartment
electrochemical cell, in a three-electrode configuration with the SECM tip as working electrode
(WE), a Pt-wire as counter electrode (CE) and a chloridized silver wire as a Ag/AgCl pseudo-
reference electrode (RE). The SECM tips were glass-insulated Pt-disk microelectrodes fabricated
from 25 µm diameter Pt-wire as described in the experimental section. A PC in combination with
in-house developed software programmed in Microsoft Visual Basic 6.0 was used to control the
system parameters and for fast data acquisition.
SEM and EDX measurements
Scanning electron microscopy measurements were performed using an eSEM Dual Beam™
Quanta 3D FEG set-up (FEI, Tokyo, Japan). The system was equipped with an energy dispersive
X-ray detector (EDAX Genesis XM2i; EDAX, NJ, USA) which was used to obtain element maps
of the surfaces to identify the chemical composition of the S-phase single crystals. A beam
voltage of 15 kV and 60 µA beam current were used for the acquisition of the SEM images.
X-ray diffraction (XRD) measurements
Bulk intermetallic samples were pulverized and characterized by standard X-ray powder
diffraction with a D8 X-ray diffraction spectrometer (Bruker AXS, WI, USA). The obtained
spectra were analyzed using the DIFFRACplus software and the reference data from the library
provided by Bruker Company were used to identify the observed peak positions.
Electron backscatter diffraction (EBSD) measurements
For the phase identification of the synthesized intermetallic samples the backscattered electron
Kikuchi patterns (BEKPs) were obtained using a Zeiss Leo Gemini 1550 VP (Carl Zeiss AG,
Germany) equipped with an EBSD detector. The recorded patterns were analyzed and indexed
with the provided software. Final verification was achieved by comparing the simulated pattern of
Al2CuMg single phase with the recorded experimental pattern.
Atomic Force Microscopy (AFM) measurements
AFM images were acquired using a NanoWizard 3 AFM controlled by a JPK SPMControl
Station III (JPK Instruments, Germany) mounted on a vibration dampening table Vario Control
Experimental part
190
Micro 40 (Halcyonics, Germany). Cantilevers were made of doped single crystal silicone (type
ACTA, Applied Nanostructures, USA), with a tip-radius <10 nm, a spring constant of ~40 N/m
and an Al coating on the reflecting side. All images were acquired ex situ using tapping mode
under ambient laboratory conditions with a scanning speed of 100 nm/s.
SKPFM measurements
All SKPFM measurements were performed in air with a Digital Instruments Dimension 3100
AFM (Veeco Instruments, USA) equipped with a Nanoscope IV controller using two pass lift-
mode. During the first scan the collection of topographical data was performed in the tapping
mode while during the second scan the tip was lifted to a predefined distance so that electrostatic
interactions between the tip and the sample were recorded and displayed in terms of contact
potential difference between the tip and the sample surface. Co/Cr coated tips with a resonance
frequency of 60-75 kHz were used.
Scanning Kelvin probe (SKP) measurements
Scanning Kelvin probe measurements were conducted using a SKP system obtained by K&M
SoftControl, Germany) using a 100 µm Cr / Ni wire as Kelvin probe. Prior to measurements the
SKP system was calibrated using a drop of saturated CuSO4 solution placed in a small Cu vial
(E˚(Cu2+/Cu) = 0.32 VSHE).
7.1.5. Software
Citavi 3 Swiss Academic Software, Switzerland
Database of EBSD spectra Carl Zeiss AG, Germany
Database of XRD spectra (DIFFRACplus ) Bruker AXS, USA
GPES 4.9 Eco Chemie B. V., Netherlands
Microsoft Office 2010 Microsoft, USA
Nova 1.7 Eco Chemie B. V., Netherlands
Origin 8.5 OriginLab Corp., USA
Visual Basic 6.0 Microsoft, USA
Windows XP Microsoft, USA
Experimental part
191
7.2. Fabrication of electrodes
7.2.1. Glass insulated Pt-disk microelectrodes
Glass insulated Pt-disk microelectrodes were fabricated following the procedure previously
described by Kranz et al. [143]. Borosilicate glass capillaries (L = 100 mm, Dout = 1.5 mm,
Din = 0.75 mm) were conically pulled by melting the glass with a heating coil while
simultaneously pulling on the unmounted side of the capillary. Short pieces (≈ 10 mm) of Pt wire
with diameters ranging from 25 µm to 500 µm were placed into the pulled part of the capillary
and subsequently sealed into the glass using a heating coil. During the sealing procedure an air
pump was used to lower the pressure inside the capillary to guarantee a tight sealing of the Pt wire
in borosilicate glass. The sealed end of capillary was then carefully polished with sandpaper until
a disk-shaped surface of Pt was exposed. The unsealed end of the Pt wire was electrically
connected to 0.5 mm thick Cu wire using silver paint. To additionally stabilize the Cu wire inside
the capillary a drop of two-component epoxy glue was placed at the unsealed end of the capillary.
Before each experiment the working electrode was cleaned by polishing with different grades of
alumina paste (3 µm, 1 µm, 0.3 µm and 0.05 µm) and subsequently ultrasonicated in 2-propanol
and deionized water for 5 min. After cleaning, all electrodes were dried in an argon stream.
7.2.2. Glass insulated Pt SKP-SECM electrodes
Glass-insulated Pt-disk microelectrodes fabricated as previously described with 25 µm, 50 µm
or 125 µm diameter Pt-wires were the basis for the glass insulated Pt SKP-SECM tips. Before
fabrication all glass capillaries (L = 100 mm, Dout = 1.5 mm) were sorted to ensure optimal fitting
to the brass guide on top of the SKP-SECM measurement head. Using a special homemade set-up,
the SECM tip was fixed with two-component epoxy glue in a cylindrical holder (L = 2 cm, Dout =
6 mm) made of polyvinylchloride (PVC). This holder has a screw thread and can be mounted on
the PEEK translator. Before each experiment the working electrode was cleaned by polishing it
with different grades of alumina paste (3 µm, 1 µm, 0.3 µm and 0.05 µm) and subsequently
ultrasonicated in 2-propanol and deionized water for 5 min. After cleaning, all electrodes were
dried in argon stream.
7.2.3. “Glass free” Pt SKP-SECM electrodes
Glass-insulated Pt-disk microelectrodes fabricated as previously described using 25 µm
diameter Pt wire were used as the basis for the development of the proposed “glass free” Pt SKP-
Experimental part
192
SECM tips. Before fabrication all glass capillaries (L = 100 mm, Dout = 1.5 mm) were carefully
selected to ensure optimal fitting to the brass guide on top of the SKP-SECM measurement head.
The concept of the “glass free” SKP-SECM tips is based on etching of the glass sheath at the very
top of glass insulated SECM tip for about 250 µm using concentrated HF (40 %). Before etching
in HF the microelectrode was polished with different grades of alumina paste (3 µm, 1 µm,
0.3 µm and 0.05 µm) to ensure mirror quality of exposed Pt disk surface. After polishing all
microelectrodes were ultrasonicated in acetone, 2-propanol and deionized water for 5 min.
Additionally, after cleaning all electrodes were dried in an argon stream to remove water from the
exposed Pt surface. Directly after etching all microelectrodes were cleaned again with the same
procedure, however, they were not dried in an argon stream as this lead to bending of the exposed
Pt wire.
7.2.4. Insulated “glass free” Pt SKP-SECM electrodes
Insulated “glass free” Pt SKP-SECM electrodes were fabricated using “glass free” Pt SKP-
SECM electrodes by isolation of the exposed Pt wire side walls with a thin layer of nail varnish
(Maybelline Jade). During the insulation procedure the greatest care was taken to avoid the
blocking of the disk area at the very top of the Pt wire by the nail varnish. To ensure this, the
insulation procedure was performed under an optical microscope.
7.2.5. Pt wire electrodes
Optimization of electrochemical etching and electrochemical micropolishing procedures was
performed using Pt wire electrodes based on 25 µm Pt wire protruding from the borosilicate glass
capillary for approximately 15 mm. The contact between the Pt wire and 0.5 mm thick Cu wire
inside the capillary was established with Ag filled epoxy glue. To avoid leakage of the etching
solution inside the capillary the lower part of it was sealed with two-component epoxy glue. Prior
to each optimization step the very top of the Pt wire was cut with a razor blade to simulate the
same geometry of exposed Pt wire as by electrochemical etching and electrochemical
micropolishing of the “glass free” Pt SKP-SECM electrodes.
7.2.6. Miniaturized reference electrodes
The miniaturized reference electrode was fabricated using a borosilicate Pasteur pipette, a
silver wire of approximately 4-5 cm in length and 0.5 mm in diameter and also a ceramic frit. The
Experimental part
193
porous ceramic frit was placed inside the thinnest part of the Pasteur pipette and heated using a
gas torch to tightly seal the frit into the pipette glass. Great attention was paid to ensure tight
sealing of the frit as this is very important to avoid leakage of electrolyte. Subsequently, the sealed
end of the pipette was carefully polished on emery paper until a disk-shaped surface of the
ceramic frit was exposed. The Ag wire was coiled to obtain a spiral and connected to a thick Cu
wire (4 cm in length, 1 mm in diameter) by soldering. The connection area was insulated using
Lacomit varnish G 371 while the Cu wire was fixed inside the plastic tube which fitted very well
to the widest part of the Pasteur pipette. The pipette was filled with 3 M KCl solution and
connected to a vacuum pump to suck electrolyte through the dry frit to remove air. Chloridization
of the Ag wire was performed in an electrolytic setup by immersing the Ag wire (anode) with a Pt
wire counter electrode (cathode) in a mixture of 3 M KCl and 0.1 M HCl. A thin layer of AgCl
was deposited on the Ag wire by applying a potential of 5 V for about 1 min followed by
application of a potential of 10 V for 10 min. Afterwards, the chloridized Ag wire was thoroughly
rinsed with deionized water and placed into the pipette filled with 3 M KCl. After fabrication, the
miniaturized reference electrode was characterized by measuring its potential vs. a commercial
Ag / AgCl / 3 M KCl reference electrode using a multimeter. The reference electrodes were used
if the potential difference was no bigger than 10 mV. The potential of the miniaturized reference
electrode was checked on a regular basis and the electrolyte inside of it was exchanged before
each experiment.
7.2.7. Miniaturized Pt counter electrodes
A miniaturized Pt counter electrode was fabricated by coiling a Pt wire of approximately 3-
4 cm in length and 0.25 mm in diameter to a spiral while 0.5 mm thick Cu wire was connected by
soldering. The connection area was insulated using Lacomit varnish G 371.
7.2.8. Deposition of a Ag layer on the glass surface of the Pt SKP-SECM tip
A compact layer of Ag was chemically formed on the surface of an insulating glass sheath of
the Pt SKP-SECM tip using Tollen´s reaction. To enhance the adhesion of metal layer, the
roughness of the glass surface was increased by gently rubbing the side walls of the tip with
emery paper (grade 600). Prior to formation of silver layer, the Pt SKP-SECM tip was wiped with
a soft tissue to remove the grinding dust and subsequently rinsed with acetone. Tollen´s reagent
was prepared by mixing 2 ml of 5 %AgNO3 with 10 %NaOH. The NaOH was added gradually
until no more dark colored precipitate (Ag2O) was formed. Stepwise addition of small portions
Experimental part
194
28 % NH4OH led to dissolution of Ag2O and formation of colorless [Ag(NH3)2OH]. Formation of
the metallic Ag layer was initiated by addition of a few drops of 10 % glucose solution as a
reducing agent. To enhance the deposition rate of Ag on the glass surface of the Pt SKP-SECM
tip, the body of the electrode was pre-heated with a heat gun and then 1/3 of the tip (lower part)
was quickly immersed into the freshly prepared Tollen´s reagent for a few seconds. Immediately
after emersion, the Pt SKP-SECM tip was heated with a heat gun and after cooling down to room
temperature, rinsed with distilled water. To form a uniform and compact layer of Ag, the plating
procedure was repeated 3 to 5 times.
7.3. Fabrication and preparation of samples
7.3.1. Single Cu islands deposited on an Al layer (Type A test sample)
Freshly evaporated 500 nm thin Al layers deposited on clean glass plates were used as the
substrate for the deposition of 200 nm thin Cu islands. Prior to deposition of Al a 10 nm thin Cr
layer was deposited onto the glass surface to improve adhesion of the Al layer. The deposition of
Cu structures was performed using physical vapor deposition (PVD) through masks with oval
shaped 500 µm x 200 µm holes, which resulted in a homogeneous distribution of Cu islands on
the Al sample surface. Both metals were deposited using high-purity metal pieces (99.99%) as
sources for evaporation.
7.3.2. Pt layer partially deposited on a W layer (Type B test sample)
Thermally oxidized (1500 nm SiO2) four inch silicon wafer was covered with a 10 nm thin Cr
adhesion layer and subsequently covered with a 200 nm layer of W by magnetron sputtering. Half
of the wafer surface was covered by a mask and a 100 nm Pt film was deposited. After metal
deposition the metal film was protected by a thin layer of a PMMA-based photoresist and was cut
in 1.2 cm x 1.2 cm pieces in such a way, that one half of the sample was covered by Pt and the
other by W. Before each experiment the test sample was cleaned by subsequent ultrasonication in
acetone, 2-propanol and deionized water for 5 min. After cleaning the active area of the test
sample was dried with argon.
7.3.3. Pt grid structure deposited on a W layer (Type C test sample)
Thermally oxidized (1500 nm SiO2) four inch silicon wafer was covered with a 10 nm thin Cr
adhesion layer and subsequently covered with a 200 nm layer of W by magnetron sputtering. The
Experimental part
195
deposited metal layer was covered by a thin PMMA-based positive photoresist using a conven-
tional spin coating technique and a laser writer was used to write a test structure with lines of
varying widths. The photoresist areas which were exposed to the laser beam were then removed
by washing in acetone (lift-off process). A 100 nm thin Pt layer was then deposited on the wafer
surface. Any Pt deposited on the remaining photoresist surface was removed with dissolution of
the underlying photoresist during a lift-off process. Before each experiment the test sample was
cleaned by subsequent ultrasonication in acetone, 2-propanol and deionized water for 5 min. After
cleaning the test sample was dried in an argon stream.
7.3.4. Polycrystalline Pt thin film sample
High-quality polycrystalline Pt thin film samples were fabricated on clean commercial Si(100)
substrates (orientation ±0.5°) covered with a thick SiO2 layer (1500 nm ± 5%) as a barrier layer.
The oxidized surface of the substrate was coated with a 10 nm thin Ti layer (99,995 %) to improve
Pt film adhesion and a 200 nm thick Pt film was then deposited (99.99 %). Deposition parameters
were as follows: room temperature (24°C), base vacuum 1·10-7 Torr, pressure 0.133 Pa, with 40
sccm Ar (99.9999 %), substrate rotation 30 rpm, deposition rate for Ti 0.017 nm/s, deposition rate
for Pt 0.15 nm/s, 100 W DC. ATC-2200-V confocal magnetron sputtering machine was used
(AJA International, USA) for the film formation. After sputtering the Pt side of the wafer was
protected with a PMMA-based photoresist and the wafer was cut into small 1.5 cm x 0.8 cm
pieces.
7.3.5. Bulk Pt(111) single crystal
Before each experiment bulk Pt(111) single crystal was annealed at 850 ± 20˚C in a Ar / H2
(5 % of H2) atmosphere for 2 to 6 min, followed by annealing in a Ar / CO (0.1% of CO)
atmosphere for an additional 2 min at 850 ± 20˚C. To minimize contamination during the course
of CV measurements, the crystal was always manipulated without exposure to air. Therefore, it
was always kept in a controlled atmosphere, minimizing random contamination from the
laboratory environment. A hanging meniscus configuration was used for electrochemical chara-
cterization of the Pt(111) surface.
Experimental part
196
7.3.6. Pt(111)-like thin film sample
For the fabrication of Pt(111)-like thin films a high-quality polycrystalline Pt thin film sample
was flame annealed in air for approximately 10 s to 15 s using a butane / oxygen flame providing
10 periodic heating and cooling sequences. During annealing the flame was orientated perpen-
dicular to the surface of the polycrystalline Pt thin film sample. After flame annealing, the samples
were transferred into an electrochemical cell for electrochemical cleaning to remove any surface
contaminations from the laboratory atmosphere or other sources. All samples were cycled
between -0.67 V and 0.8 V vs. MMS reference electrode until a stable electrode response was
observed. After that the samples were carefully rinsed with Siemens© UltraClear water and dried
in Ar / H2 (6.0, 5 % of H2) atmosphere. Afterwards, the samples were transferred into the furnace
for thermal treatment. Heating of samples was performed in a clean quartz tube at 400°C for 3 to
5 min while subsequent cooling was done under controlled atmosphere of Ar / H2 (6.0, 5 % of H2).
The sample was then extracted from the furnace and reintroduced to the electrochemical cell.
7.3.7. Cu-Pt near-surface alloy on Pt(111)-thin film sample
Prior to fabrication of Cu-Pt(111) thin film NSA all Pt(111)-like thin film electrodes were
electrochemically cleaned by cycling between 0.05 V and 1.12 V vs. RHE in 0.1M HClO4 until
reproducible characteristic peaks for Pt(111) facets were observed. The potential was then held at
1.0 V vs. RHE while a 1 mM CuClO4 in 0.1 M HClO4 solution was introduced into the electro-
chemical cell. The potential was then cycled between 0.33 V and 1.0 V vs. RHE for sequential
underpotential deposition and stripping of Cu monolayers. To ensure the deposition of a full
monolayer of Cu the potential was held at the cathodic limit for 3 min, the charge under the
subsequent anodic stripping peak allowed confirmation of complete monolayer formation. After
3 min the working electrode (under potential control at all times) was rinsed with Siemens©
UltraClear water. To avoid oxidation of the Cu monolayer during the following drying step, a
dummy Pt electrode connected to the working electrode was present in the cell. This allowed the
potential of the working electrode to be held under control when the Pt thin film sample was not
in the electrolyte solution. Drying of the sample was performed in an Ar/H2 atmosphere for
30 min while the potential was held at 0.33 V vs. RHE. Once dry the electrode was disconnected
and placed in a tube furnace for the annealing step. The furnace equipped with standard clean
quartz glassware was used to allow heating and cooling of the thin film samples under a
controlled atmosphere. The temperature within the furnace was maintained at 400°C while the
sample was introduced into the heating zone after an Ar / H2 (5% of H2) atmosphere had been
Experimental part
197
established. The optimal heating time was found to be about 2 min. After the annealing step the
sample was removed from the furnace and allowed to cool for 5 min under the controlled
atmosphere (Ar / H2, 5% of H2). The sample was then extracted from the furnace and reintroduced
to the electrochemical cell. The formed NSA thin film was characterized by cycling between
0.05 V and 1.12 V vs. RHE in 0.1 M HClO4 solution.
7.3.8. Al2CuMg single crystals in solid solution Al matrix
Bulk S-phase single crystals were synthesized by melting the corresponding proportions of the
pure elements, namely 46.4 g. 99.99 % Al, 38 g. 99.99 % Cu and 18 g. 99.99 % Mg by heating at
a rate of 8˚C min-1 to 950˚C and maintaining the temperature at 950˚C for 15 min to ensure
complete melting. The melt was then mechanically agitated and cooled down to room tempe-
rature. The homogenization of the melt was done using isothermal treatment by heating to 510˚C
for 65 h, leading to segregation into the solid S-phase and an Al-rich liquid. Long isothermal
treatment was used to obtain the desired size of single S-phase crystals. Finally, the melt was
water quenched to freeze the desired microstructure and the chemical composition of synthesized
Al2CuMg single crystals. The ingot was then cut into smaller pieces and samples were
mechanically ground with successively finer SiC abrasive paper lubricated with kerosene.
Polishing to a smooth surface was performed with non-aqueous concentrated diamond based
slurries (3 µm, 1 µm and 0.25 µm) and DP lubricant. The polished samples were degreased by
ultrasonic cleaning in pure ethanol and dried in an argon stream.
References
198
8. References [1] J. Bohannon, Science 2005, 309, 376. [2] Y. Baek, G. S. Frankel, Journal of the Electrochemical Society 2003, 150, B1. [3] R. G. Buchheit, R. P. Grant, P. F. Hlava, B. Mckenzie, G. L. Zender, Journal of the
Electrochemical Society 1997, 144, 2621. [4] K. R. Trethewey, D. A. Sargeant, D. J. Marsh, A. A. Tamimi, Corrosion Science 1993, 35,
127. [5] H. S. Isaacs, Journal of the Electrochemical Society 1988, 135, 2180. [6] H. S. Isaacs, M. W. Kendig, Corrosion 1980, 36, 269. [7] W. A. Zisman, Review of Scientific Instruments 1932, 3, 367. [8] A. J. Bard, F. R. Fan, J. Kwak, O. Lev, Analytical Chemistry 1989, 61, 132. [9] M. Schmidt, M. Nohlen, G. Bermes, M. Bohmer, K. Wandelt, Review of Scientific
Instruments 1997, 68, 3866. [10] A. Ilie, A. Hart, A. J. Flewitt, J. Robertson, W. I. Milne, Journal of Applied Physics 2000,
88, 6002. [11] G. Koley, M. G. Spencer, Journal of Applied Physics 2001, 90, 337. [12] H. Heil, J. Steiger, S. Karg, M. Gastel, H. Ortner, H. von Seggern, M. Stossel, Journal of
Applied Physics 2001, 89, 420. [13] M. Pfeiffer, K. Leo, N. Karl, Journal of Applied Physics 1996, 80, 6880. [14] D. M. Taylor, Advances in Colloid and Interface Science 2000, 87, 183. [15] C. G. Vayenas, S. Bebelis, S. Ladas, Nature 1990, 343, 625. [16] M. Stratmann, H. Streckel, Corrosion Science 1990, 30, 681. [17] G. Williams, H. N. McMurray, Journal of the Electrochemical Society 2001, 148, B377. [18] G. Williams, H. N. McMurray, D. Hayman, P. C. Morgan, PhysChemComm 2001, 6, 1. [19] S. Yee, R. A. Oriani, M. Stratmann, Journal of the Electrochemical Society 1991, 138, 55. [20] W. Schmidt, M. Stratmann, Corrosion Science 1998, 40, 1441. [21] B. Lagel, I. D. Baikie, U. Petermann, Surface Science 1999, 435, 622.
References
199
[22] J. Kwak, A. J. Bard, Analytical Chemistry 1989, 61, 1221. [23] R. D. Martin, P. R. Unwin, Analytical Chemistry 1998, 70, 276. [24] B. B. Katemann, A. Schulte, E. J. Calvo, M. Koudelka‐Hep, W. Schuhmann, Electro-
chemistry Communications 2002, 4, 134. [25] K. Eckhard, X. Chen, F. Turcu, W. Schuhmann, Physical Chemistry Chemical Physics
2006, 8, 5359. [26] K. Karnicka, K. Eckhard, D. A. Guschin, L. Stoica, P. J. Kulesza, W. Schuhmann,
Electrochemistry Communications 2007, 9, 1998. [27] L. Guadagnini, A. Maljusch, X. Chen, S. Neugebauer, D. Tonelli, W. Schuhmann,
Electrochimica Acta 2009, 54, 3753. [28] A. O. Okunola, T. C. Nagaiah, X. Chen, K. Eckhard, W. Schuhmann, M. Bron,
Electrochimica Acta 2009, 54, 4971. [29] T. C. Nagaiah, A. Maljusch, X. Chen, M. Bron, W. Schuhmann, ChemPhysChem 2009,
10, 2711. [30] A. Maljusch, T. C. Nagaiah, S. Schwamborn, M. Bron, W. Schuhmann, Analytical
Chemistry 2010, 82, 1890. [31] G. S. Frankel, Journal of the Electrochemical Society 1998, 145, 2186. [32] R. Hausbrand, M. Stratmann, M. Rohwerder, Journal of the Electrochemical Society 2008,
155, C369. [33] IUPAC. Compendium of Chemical Terminology, 2nd ed. (the "Gold Book"). Compiled by
A. D. McNaught, A. Wilkinson. Blackwell Scientific Publications, Oxford (1997) [34] C. Herring M. H. Nichols, Review of Modern Physics 1949, 21, 185. [35] E. B. Hensley, Journal of Applied Physics 1961, 32, 301. [36] R. H. Fowler, Physical Reviews 1957, 107, 1553. [37] E. W. Müller and T. T. Song, Field Ion Microscopy, Elsevier, New York (1969). [38] W. Thompson, Philosophical Magazine 1898, 46, 82. [39] W. A. Zisman, Review of Scientific Instruments 1932, 3, 367. [40] J. Bonnet, J. M. Palau, L. Soonckindt, L. Lassabatere, Journal of Physics E: Scientific
Instruments 1977, 10, 212. [41] T. Delchar, A. Eberhagen F. C. Tompkins, Journal of Scientific Instruments 1963, 40, 105.
References
200
[42] C. W. Oatley, Proceedings of the Royal Society 1936, A155, 218. [43] M. Green, Solid State Surface Science, Marcel Dekker, New York, Vol.1 (1969). [44] H. Shelton, Physical Review 1957, 107, 1553. [45] P. A. Anderson, Physical Review 1935, 17, 958. [46] I. Langmuir, K.H. Kingdon, Physical Review 1929, 34, 129. [47] N. A. Surplice, R. J. D´Arcy, Journal of Physics E: Scientific Instruments 1970, 3, 477. [48] B. Ritty, F. Wachtel, R. Manquenouille, F. Ott, J. B. Donnet, Journal of Physics E:
Scientific Instruments 1982, 15, 310. [49] J. O. Bockris, S. D. Argade, Journal of Chemical Physics 1968, 49, 5133. [50] J. O. Bockris, Energy Conversion 1970, 10, 41. [51] J. O. Bockris, M. A. Habib, Journal of Electroanalytical Chemistry 1975, 65, 473. [52] E. Gileady, G. Stoner, Journal of Electroanalytical Chemistry 1972, 36, 492. [53] S. Trasatti, Journal of Electroanalytical Chemistry 1974, 52, 313. [54] S. Trasatti, Journal of Electroanalytical Chemistry 1975, 66, 155. [55] S. Trasatti, Electrochimica Acta 1991, 36, 1659. [56] A. Leng, H. Streckel, M. Stratmann, Corrosion Science 1999, 41, 547. [57] A. Hadjadj, P. R. Cabarrocas, B. Equer, Review of Scientific Instruments 1995, 66, 5272. [58] K. B. Johnson, W. N. Hansen, Review of Scientific Instruments 1995, 66, 2967. [59] H. Baumgärtner, H. D. Liess, Review of Scientific Instruments 1988, 59, 802. [60] H. N. McMurray, G. Williams, Journal of Applied Physics 2002, 91, 1673. [61] M. Nonnenmacher, M. P. Oboyle, H. K. Wickramasinghe, Applied Physics Letters 1991,
58, 2921. [62] M. Nonnenmacher, J. Greschner, O. Wolter, R. Kassing, Journal of Vacuum Science &
Technology B 1991, 9, 1358. [63] H. O. Jacobs, P. Leuchtmann, O. J. Homan, A. Stemmer, Journal of Applied Physics 1998,
84, 1168. [64] H. O. Jacobs, H. F. Knapp, A. Stemmer, Review of Scientific Instruments 1999, 70, 1756.
References
201
[65] D. B. Blucher, J. E. Svensson, L. G. Johansson, M. Rohwerder, M. Stratmann, Journal of
the Electrochemical Society 2004, 151, B621. [66] M. Rohwerder, E. Hornung, M. Stratmann, Electrochimica Acta 2003, 48, 1235. [67] A. K. Henning, T. Hochwitz, J. Slinkman J. Never, S. Hoffmann, P. Kaszube, C. Daghlian,
Journal of Applied Physics 1995, 77, 1888. [68] T. Hochwitz, A. K. Henning, C. Levey, C. Daghlian, J. Slinkman, Journal of Vacuum
Science & Technology B 1996, 14, 457. [69] C. Sommerhalter, T. Glatzel, T. W. Matthes, A. Jager-Waldau, M. C. Lux-Steiner, Applied
Surface Science 2000, 157, 263. [70] A. Chavez-Porson, O. Vatel, M. Tanimoto, H. Ando, H. Iwamura, H. Kanbe, Applied
Physics Letters 1995, 67, 3069. [71] M. Fujihira, H. Kawate, Thin Solid Films 1994, 242, 163. [72] G. Williams, H. N. McMurray, D. A. Worsley, Journal of Forensic Sciences 2001, 46,
1085. [73] W. Fürbeth, M. Stratmann, Corrosion Science 2001, 43, 207. [74] K. Hayashi, N. Saito, H. Sugimura, O. Takai, N. Nakagiri, Ultramicroscopy 2002, 91, 151. [75] W. Kutner, Electrochimica Acta 1992, 37, 1109. [76] G. Grundmeier, K. M. Jüttner, M. Stratmann in Corrosion and Environmental Degrada-
tion, M. Schütze (ed.), Vol.1, Wiley-VCH, Weinheim (2000). [77] M. Stratmann, H. Streckel, Corrosion Science 1990, 30, 697. [78] M. Stratmann, H. Streckel, K. T. Kim, S. Crockett, Corrosion Science 1990, 30, 715. [79] A. Leng, M. Stratmann, Corrosion Science 1993, 34, 1657. [80] H. Leidheiser, in Corrosion mechanisms, F. Mansfeld (ed.), Marcel Dekker, New York
(1987). [81] R. Feser, M. Stratmann, Materials and Corrosion 1991, 42, 187. [82] M. Stratmann, R. Feser, A. Leng, Electrochimica Acta 1994, 39, 1207. [83] H. Leidheiser, W. Wang, L. Igetoft, Progress in Organic Coatings 1983, 11, 19. [84] A. Leng, H. Streckel, M. Stratmann, Corrosion Science 1999, 41, 579. [85] A. Leng, H. Streckel, K. Hofmann, M. Stratmann, Corrosion Science 1999, 41, 599.
References
202
[86] R. C. Engstrom, T. Meaney, R. Tople, R. M. Wightman, Analytical Chemistry 1987, 59, 2005.
[87] J. Newman, Journal of the Electrochemical Society 1966, 113, 501. [88] Handbook of the Electrochemistry, Ed. C. S. Zoski, Elsevier, New York (2006) [89] C. Kranz, H. E. Gaub, W. Schuhmann, Advanced Materials 1996, 8, 634. [90] M. Fotino, Applied Physics Letters 1992, 60, 2935. [91] B. L. Zhang, E. K. Wang, Electrochimica Acta 1994, 39, 103. [92] R. Kazinczi, E. Szocs, E. Kalman, P. Nagy, Applied Physics A - Materials Science &
Processing 1998, 66, S535. [93] F. R. Fan, J. Kwak, A. J. Bard, Journal of the American Chemical Society 1996, 118, 9669. [94] C. E. Bach, R. J. Nichols, W. Beckmann, H. Meyer, A. Schulte, J. O. Besenhard, P. D.
Jannakoudakis, Journal of the Electrochemical Society 1993, 140, 1281. [95] H. L. Bonazza, J. L. Fernandez, Journal of Electroanalytical Chemistry 2010, 650, 75. [96] Y. H. Shao, M. V. Mirkin, G. Fish, S. Kokotov, D. Palanker, A. Lewis, Analytical
Chemistry 1997, 69, 1627. [97] X. J. Zhang, B. Ogorevc, Analytical Chemistry 1998, 70, 1646. [98] B. B. Katemann, T. Schuhmann, Electroanalysis 2002, 14, 22. [99] L. P. Bauermann, W. Schuhmann, A. Schulte, Physical Chemistry Chemical Physics 2004,
6, 4003. [100] A. Schulte, R. H. Chow, Analytical Chemistry 1998, 70, 985. [101] A. Schulte, R. H. Chow, Analytical Chemistry 1996, 68, 3054. [102] E. M. Hussien, W. Schuhmann, A. Schulte, Analytical Chemistry 2010, 82, 5900. [103] A. Schulte, M. Nebel, W. Schuhmann, in Annual Review of Analytical Chemistry, Vol. 3,
E. S. Yeung, R. N. Zare (ed.), 2010. [104] D. O. Wipf, A. J. Bard, Analytical Chemistry 1992, 64, 1362. [105] D. O. Wipf, A. J. Bard, D. E. Tallman, Analytical Chemistry 1993, 65, 1373. [106] C. Lee, D. O. Wipf, A. J. Bard, K. Bartels, A. C. Bovik, Analytical Chemistry 1991, 63,
2442. [107] K. McKelvey, M. A. Edwards, P. R. Unwin, Analytical Chemistry 2010, 82, 6334.
References
203
[108] M. Ludwig, C. Kranz, W. Schuhmann, H. E. Gaub, Review of Scientific Instruments 1995, 66, 2857.
[109] A. Hengstenberg, C. Kranz, W. Schuhmann, Chemistry - A European Journal 2000, 6,
1547. [110] A. Hengstenberg, A. Blochl, I. D. Dietzel, W. Schuhmann, Angewandte Chemie - Int. Ed.
2001, 40, 905. [111] B. B. Katemann, A. Schulte, W. Schuhmann, Chemistry - A European Journal 2003, 9,
2025. [112] B. B. Katemann, A. Schulte, W. Schuhmann, Electroanalysis 2004, 16, 60. [113] K. Eckhard, M. Etienne, A. Schulte, W. Schuhmann, Electrochemistry Communications
2007, 9, 1793. [114] M. Nebel, K. Eckhard, T. Erichsen, A. Schulte, W. Schuhmann, Analytical Chemistry
2010, 82, 7842. [115] A. J. Bard, G. Denuault, R. A. Friesner, B. C. Dornblaser, L. S. Tuckerman, Analytical
Chemistry 1991, 63, 1282. [116] S. Daniele, I. Ciani, D. Battistel, Analytical Chemistry 2008, 80, 253. [117] J. L. Amphlett, G. Denuault, Journal of Physical Chemistry B 1998, 102, 9946. [118] C. Lefrou, Journal of Electroanalytical Chemistry 2006, 592, 103. [119] Y. H. Shao, M. V. Mirkin, Journal of Physical Chemistry B 1998, 102, 9915. [120] R. Cornut, C. Lefrou, Journal of Electroanalytical Chemistry 2007, 608, 59. [121] G. Wittstock, M. Burchardt, S. E. Pust, Y. Shen, C. Zhao, Angewandte Chemie - Int. Ed.
2007, 46, 1584. [122] A. J. Bard, F. R. Fan, D. T. Pierce, P. R. Unwin, D. O. Wipf, F. M. Zhou, Science 1991,
254, 68. [123] D. O. Wipf, A. J. Bard, Journal of the Electrochemical Society 1991, 138, L4. [124] D. O. Wipf, A. J. Bard, Journal of the Electrochemical Society 1991, 138, 469. [125] A. J. Bard, M. V. Mirkin, P. R. Unwin, D. O. Wipf, Journal of Physical Chemistry 1992,
96, 1861. [126] P. R. Unwin, A. J. Bard, Journal of Physical Chemistry 1991, 95, 7814. [127] F. Zhou, P. R. Unwin, A. J. Bard, Journal of Physical Chemistry 1992, 96, 4917.
References
204
[128] C. A. Zhao, G. Wittstock, Analytical Chemistry 2004, 76, 3145. [129] H. Shiku, T. Matsue, I. Uchida, Analytical Chemistry 1996, 68, 1276. [130] G. Wittstock, T. Wilhelm, S. Bahrs, P. Steinrucke, Electroanalysis 2001, 13, 669. [131] G. Wittstock, R. Hesse, W. Schuhmann, Electroanalysis 1997, 9, 746. [132] T. Wilhelm, G. Wittstock, Microchimica Acta 2000, 133, 1. [133] H. Yamada, M. Ogata, T. Koike, Langmuir 2006, 22, 7923. [134] M. Arca, M. V. Mirkin, A. J. Bard, Journal of Physical Chemistry 1995, 99, 5040. [135] M. Tsionsky, A. J. Bard, D. Dini, F. Decker, Chemistry of Materials 1998, 10, 2120. [136] M. Quinto, S. A. Jenekhe, A. J. Bard, Chemistry of Materials 2001, 13, 2824. [137] F. Li, I. Ciani, P. Bertoncello, P. R. Unwin, J. Zhao, C. R. Bradbury, D. J. Fermin, Journal
of Physical Chemistry C 2008, 112, 9686. [138] J. Zhou, Y. Zu, A. J. Bard, Journal of Electroanalytical Chemistry 2000, 491, 22. [139] O. E. Husser, D. H. Craston, A. J. Bard, Journal of the Electrochemical Society 1989, 136,
3222. [140] Y. M. Wuu, F. R. Fan, A. J. Bard, Journal of the Electrochemical Society 1989, 136, 885. [141] C. Kranz, M. Ludwig, H. E. Gaub, W. Schuhmann, Advanced Materials 1995, 7, 38. [142] W. Schuhmann, C. Kranz, H. Wohlschlager, J. Strohmeier, Biosensors & Bioelectronics
1997, 12, 1157. [143] C. Kranz, M. Ludwig, H. E. Gaub, W. Schuhmann, Advanced Materials 1995, 7, 568. [144] C. Kranz, G. Wittstock, H. Wohlschlager, W. Schuhmann, Electrochimica Acta 1997, 42,
3105. [145] T. Wilhelm, G. Wittstock, Electrochimica Acta 2001, 47, 275. [146] T. Wilhelm, G. Wittstock, Angewandte Chemie - Int. Ed. 2003, 42, 2247. [147] C. W. Lin, F. R. Fan, A. J. Bard, Journal of the Electrochemical Society 1987, 134, 1038. [148] D. Mandler, A. J. Bard, Journal of the Electrochemical Society 1989, 136, 3143. [149] Y. B. Zu, L. Xie, B. W. Mao, Z. W. Tian, Electrochimica Acta 1998, 43, 1683. [150] S. Kramer, R. R. Fuierer, C. B. Gorman, Chemical Reviews 2003, 103, 4367.
References
205
[151] V. Radtke, J. Heinze, International journal of Research in Physical Chemistry & Chemical
Physics 2004, 218, 103. [152] D. Mandler, A. J. Bard, Journal of the Electrochemical Society 1990, 137, 1079. [153] C. Marck, K. Borgwarth, J. Heinze, Advanced Materials 2001, 13, 47. [154] J. F. Zhou, D. O. Wipf, Journal of the Electrochemical Society 1997, 144, 1202. [155] J. L. Fernandez, A. J. Bard, Analytical Chemistry 2003, 75, 2967. [156] J. L. Fernandez, A. J. Bard, Analytical Chemistry 2004, 76, 2281. [157] J. L. Fernandez, D. A. Walsh, A. J. Bard, Journal of the American Chemical Society 2005,
127, 357. [158] J. L. Fernandez, V. Raghuveer, A. Manthiram, A. J. Bard, Journal of the American
Chemical Society 2005, 127, 13100. [159] J. L. Fernandez, N. Mano, A. Heller, A. J. Bard, Angewandte Chemie - Int. Ed. 2004, 43,
6355. [160] J. Rodriguez-Lopez, A. J. Bard, Journal of the American Chemical Society 2010, 132,
5121. [161] Q. Wang, J. Rodriguez-Lopez, A. J. Bard, Journal of the American Chemical Society 2009,
131, 17046. [162] R. D. Martin, P. R. Unwin, Journal of the Chemical Society, Faraday Transactions 1998,
94, 753. [163] R. C. Engstrom, M. Weber, D. J. Wunder, R. Burgess, S. Winquist, Analytical Chemistry
1986, 58, 844. [164] J. Wang, L. H. Wu, R. L. Li, Journal of Electroanalytical Chemistry 1989, 272, 285. [165] H. Shiku, Y. Hara, T. Matsue, I. Uchida, T. Yamauchi, Journal of Electroanalytical
Chemistry 1997, 438, 187. [166] C. A. Wijayawardhana, G. Wittstock, H. B. Halsall, W. R. Heineman, Electroanalysis
2000, 12, 640. [167] C. Zhao, G. Wittstock, Angewandte Chemie - Int. Ed. 2004, 43, 4170. [168] W. Nogala, M. Burchardt, M. Opallo, J. Rogalski, G. Wittstock, Bioelectrochemistry 2008,
72, 174. [169] C. M. Sanchez-Sanchez, J. Rodiriguez-Lopez, A. J. Bard, Analytical Chemistry 2008, 80,
3254.
References
206
[170] C. M. Sanchez-Sanchez, A. J. Bard, Analytical Chemistry 2009, 81, 8094. [171] A. Minguzzi, M. A. Alpuche-Aviles, J. R. Lopez, S. Rondinini, A. J. Bard, Analytical
Chemistry 2008, 80, 4055. [172] B. R. Horrocks, D. Schmidtke, A. Heller, A. J. Bard, Analytical Chemistry 1993, 65, 3605. [173] R. Kashyap, K. Gratzl, Analytical Chemistry 1999, 71, 2814. [174] D. M. Osbourn, R. H. Sanger, P. J. Smith, Analytical Chemistry 2005, 77, 6999. [175] M. A. Alpuche-Aviles, D. O. Wipf, Analytical Chemistry 2001, 73, 4873. [176] R. T. Kurulugama, D. O. Wipf, S. A. Takacs, S. Pongmayteegul, P. A. Garris, J. E. Baur,
Analytical Chemistry 2005, 77, 1111. [177] B. B. Katemann, C. G. Inchauspe, P. A. Castro, A. Schulte, E. J. Calvo, W. Schuhmann,
Electrochimica Acta 2003, 48, 1115. [178] A. Schulte, S. Belger, M. Etienne, W. Schuhmann, Materials Science and Engineering: A
2004, 378, 523. [179] S. Belger, A. Schulte, C. Hessing, M. Pohl, W. Schuhmann, Materialwissenschaft und
Werkstofftechnik 2004, 35, 276. [180] A. S. Baranski, P. M. Diakowski, Journal of Solid State Electrochemistry 2004, 8, 683. [181] M. Etienne, A. Schulte, W. Schuhmann, Electrochemistry Communications 2004, 6, 288. [182] K. Eckhard, H. Shin, B. Mizaikoff, W. Schuhmann, C. Kranz, Electrochemistry Communi-
cations 2007, 9, 1311. [183] A. Lugstein, E. Bertagnolli, C. Kranz, B. Mizaikoff, Surface and Interface Analysis 2002,
33, 146. [184] K. Eckhard, W. Schuhmann, Analyst 2008, 133, 1486. [185] K. Eckhard, C. Kranz, H. Shin, B. Mizaikoff, W. Schuhmann, Analytical Chemistry 2007,
79, 5435. [186] K. Eckhard, T. Erichsen, M. Stratmann, W. Schuhmann, Chemistry - A European Journal
2008, 14, 3968. [187] K. Eckhard, W. Schuhmann, M. Maciejewska, Electrochimica Acta 2009, 54, 2125. [188] M. Gebala, W. Schuhmann, F. La Mantia, Electrochemistry Communications 2011, 13,
689. [189] A. Bauer, K. Lee, C. Song, Y. Xie, J. Zhang, R. Hui, Journal of Power Sources 2010, 195,
3105.
References
207
[190] K. Eckhard, W. Schuhmann, Electrochimica Acta 2007, 53, 1164. [191] P. Schmutz, G. S. Frankel, Journal of the Electrochemical Society 1998, 145, 2285. [192] J. H. de Wit, Electrochimica Acta 2004, 49, 2841. [193] F. Andreatta, I. Apachitei, A. A. Kodentsov, J. Dzwonczyk, J. Duszczyk, Electrochimica
Acta 2006, 51, 3551. [194] P. Campestrini, E. P. van Westing, H. W. van Rooijen, J. H. de Wit, Corrosion Science
2000, 42, 1853. [195] N. Sathirachinda, R. Gubner, J. Pan, U. Kivisakk, Electrochemical and Solid State Letters
2008, 11, C41. [196] M. Rohwerder, F. Turcu, Electrochimica Acta 2007, 53, 290. [197] P. Leblanc, G. S. Frankel, Journal of the Electrochemical Society 2002, 149, B239. [198] S. Srinivasan, Fuel cells, Springer, New York (2006). [199] F. Andreatta, H. Terryn, J. H. de Wit, Electrochimica Acta 2004, 49, 2851. [200] A. S. Bondarenko, I. E. L. Stephens, L. Bech, I. Chorkendorff, Electrochimica Acta, 2012,
DOI: j.electacta.2012.02.095. [201] P. Schmutz, G. S. Frankel, Journal of the Electrochemical Society 1998, 145, 2295.
[202] C. Senöz, A. Maljusch, M. Rohwerder, W. Schuhmann, Electroanalysis 2012, 24, 239. [203] J. C. Mitchinson R. D. Pringle, W. E. Farvis, Journal of Physics E ‐ Scientific Instruments
1971, 4, 525. [204] B. H. Blott, T. J. Lee, Journal of Physics E ‐ Scientific Instruments 1969, 2, 785.
[205] H. A. Engelhardt, P. Feulner, H. Pfnur, D. Menzel, Journal of Physics E ‐ Scientific
Instruments 1977, 10, 1133. [206] B. Ritty, F. Wachtel, F. Ott, R. Manquenouille, J. B. Donnet, Review of Scientific
Instruments 1980, 51, 1421. [207] I. D. Baikie, PhD thesis, University of Twente, 1988. [208] I. D. Baikie, K. O. van der Werf, H. Oerbekke, J. Broeze, A. van Silfhout, Review of
Scientific Instruments 1989, 60, 930. [209] I. D. Baikie, E. Venderbosch, J. A. Meyer, P. J. Estrup, Review of Scientific Instruments
1991, 62, 725. [210] J. H. Parker, R. W. Warren, Review of Scientific Instruments 1962, 33, 948.
References
208
[211] T. Fort, R. L. Wells, Surface Science 1968, 12, 46. [212] K. Besocke, S. Berger, Review of Scientific Instruments 1976, 47, 840. [213] S. Saito, T. Soumura, T. Maeda, Journal of Vacuum Science & Technology A: Vacuum
surfaces and Films 1984, 2, 1389. [214] K. Germanova, C. Hardalov, V. Strashilov, B. Georgiev, Journal of Physics E: Scientific
Instruments 1987, 20, 273. [215] Y. L. Yousef, A. Mishriki, S. Aziz, H. Mikhail, Journal of Scientific Instruments 1965, 42,
873. [216] H. Baumgartner, Measurement Science & Technology 1992, 3, 237. [217] J. Bonnet, L. Soonckindt, L. Lassabatere, Vacuum 1984, 34, 693. [218] J. Ren, PhD thesis, Universität der Bundeswehr München, 1995. [219] F. Turcu, A. Schulte, W. Schuhmann, Analytical and Bioanalytical Chemistry 2004, 380,
736. [220] E. W. Müller, K. Bahadur, Physical Review 1956, 102, 624. [221] G. Binning, H. Rohrer, C. Gerber, E. Weibel, Physical Review Letters 1982, 49, 57. [222] A. J. Melmed, Journal of Vacuum Science & Technology B 1991, 9, 601. [223] E. W. Müller, T. T. Song, Field Ion Microscopy and Applications, Elsevier, New York
(1969). [224] J. L. Vosson, W. Kern, Thin Film Processes, Academic, New York (1978). [225] T. Hibi, K. Ishikawa, Journal of Electron Microscopy 1960, 9, 81. [226] M. G. Burke, D. D. Sieloff, S. S. Brenner, Journal de Physique 1986, 47, 459. [227] E. W. Müller, Journal of Physics 1937, 106, 132. [228] A. J. Melmed, Journalof Chemical Physics 1963, 38, 1444. [229] K. D. Rendulic, E. W. Muller, Journal of Applied Physics 1967, 38, 550. [230] Y. Akama, E. Nishimura, A. Sakai, H. Murakami, Journal of Vacuum Science &
Technology A-Vacuum Surfaces and Films 1990, 8, 429. [231] M. K. Miller, G. D. W. Smith, Atom Probe Microanalysis: Principles and Applications to
Materials Problems, Materials Research Society, Pittsburg (1989).
References
209
[232] A. J. Nam, A. Teren, T. A. Lusby, A. J. Melmed, Journal of Vacuum Science &
Technology B 1995, 13, 1556. [233] R. M. Penner, M. J. Heben, N. S. Lewis, Analytical Chemistry 1989, 61, 1630. [234] H. Lemke, T. Goddenhenrich, H. P. Bochem, U. Hartmann, C. Heiden, Review of
Scientific Instruments 1990, 61, 2538. [235] S.P. Kounaves, Platinum Metals Review 1990, 34, 131. [236] L. Libioulle, Y. Houbion, J. M. Gilles, Review of Scientific Instruments 1995, 66, 97. [237] M. Fotino, Applied Physics Letters 1992, 60, 2935. [238] L. Libioulle, Y. Houbion, J. M. Gilles, Journal of Vacuum Science & Technology B 1995,
13, 1325. [239] R. Sonnenfeld, P. K. Hansma, Science 1986, 232, 211. [240] M. P. Green, K. J. Hanson, D. A. Scherson, X. Xing, M. Richter, P. N. Ross, R. Carr, I.
Lindau, Journal of Physical Chemistry 1989, 93, 2181. [241] K. Itaya, S. Sugawara, Chemistry Letters 1987, 1927. [242] M. J. Heben, M. M. Dovek, N. S. Lewis, R. M. Penner, C. F. Quate, Journal of
Microscopy 1988, 152, 651. [243] C. M. Vitus, S. C. Chang, B. C. Schardt, M. J. Weaver, Journal of Physical Chemistry
1991, 95, 7559. [244] A. A. Gewirth, D. H. Craston, A. J. Bard, Journal of Electroanalytical Chemistry 1989,
261, 477. [245] J. Wiechers, T. Twomey, D. M. Kolb, R. J. Behm, Journal of Electroanalytical Chemistry
1988, 248, 451. [246] L. Nagahara, T. Thundat, S. M. Lindsay, Review of Scientific Instruments 1989, 60, 3128. [247] F. Beck, Electrochimica Acta 1988, 33, 839. [248] C. Kurzawa, A. Hengstenberg, W. Schuhmann, Analytical Chemistry 2002, 74, 355. [249] C. E. Bach, R. J. Nichols, H. Meyer, J. O. Besenhard, Surface & Coatings Technology
1994, 67, 139. [250] S. L. Chen, A. Kucernak, Electrochemistry Communications 2002, 4, 80. [251] U. Eberle, R. von Helmolt, Energy & Environmental Science 2010, 3, 689.
References
210
[252] F. T. Wagner, B. Lakshmanan, M. F. Mathias, Journal of Physical Chemistry Letters 2010, 1, 2204.
[253] H. A. Gasteiger, D. R. Baker, R. N. Carter in Hydrogen Fuel Cells: Fundamentals and
Applications, D. Stolten (ed.),Wiley-CPH, Weinheim (2010). [254] I. E. L.Stephens,; A.S. Bondarenko, U. Grønbjerg, J. Rossmeisl, I. Chorkendorff, Energy
Environmental Science 2012, 5, 6744. [255] K. C. Neyerlin, W. Gu, J. Jorne, H. A. Gasteiger, Journal of the Electrochemical Society
2007, 154, B631. [256] H. A. Gasteiger, N. M. Markovic, Science 2009, 324, 48. [257] H. A. Gasteiger, S. S. Kocha, B. Sompalli, F. T. Wagner, Applied Catalysis B -
Environmental 2005, 56, 9. [258] A. S. Bondarenko, I. E. L. Stephens, H. A. Hansen, F. J. Perez-Alonso, V. Tripkovic, T. P.
Johansson, J. Rossmeisl, J. K. Norskov, I. Chorkendorff, Langmuir 2011, 27, 2058. [259] J. K. Norskov, J. Rossmeisl, A. Logadottir, L. Lindqvist, J. R. Kitchin, T. Bligaard, H.
Jonsson, Journal of Physical Chemistry B 2004, 108, 17886. [260] J. Greeley, I. E. L. Stephens, A. S. Bondarenko, T. P. Johansson, H. A. Hansen, T. F.
Jaramillo, J. Rossmeisl, I. Chorkendorff, J. K. Norskov, Nature Chemistry 2009, 1, 552. [261] J. K. Norskov, T. Bligaard, A. Logadottir, Kitchin, JR, J. G. Chen, S. Pandelov, Journal of
the Electrochemical Society 2005, 152, J23. [262] A. Logadottir, T. H. Rod, J. K. Norskov, B. Hammer, S. Dahl, C. J. Jacobsen, Journal of
Catalysis 2001, 197, 229. [263] T. Bligaard, J. K. Norskov, S. Dahl, J. Matthiesen, C. H. Christensen, J. Sehested, Journal
of Catalysis 2004, 224, 206. [264] P. Sabatier, Chemische Berichte 1911, 44, 1984. [265] J. Henry, A. Maljusch, M. Huang, W. Schuhmann, A. S. Bondarenko, ACS Catalalysis
2012, 2, 1457. [266] V. R. Stamenkovic, B. Fowler, B. S. Mun, G. Wang, P. N. Ross, C. A. Lucas, N. M.
Markovic, Science 2007, 315, 493. [267] I. E. L. Stephens, A. S. Bondarenko, F. J. Perez-Alonso, F. Calle-Vallejo, L. Bech, T. P.
Johansson, A. K. Jepsen, R. Frydendal, B. P. Knudsen, J. Rossmeisl, I. Chorkendorff, Journal of the American Chemical Society 2011, 133, 5485.
[268] P. Strasser, S. Koh, T. Anniyev, J. Greeley, K. More, C. Yu, Z. Liu, S. Kaya, D. Nordlund,
H. Ogasawara, M. F. Toney, A. Nilsson, Nature Chemistry 2010, 2, 454.
References
211
[269] V. Stamenkovic, B. S. Mun, K. J. Mayrhofer, P. N. Ross, N. M. Markovic, J. Rossmeisl, J. Greeley, J. K. Norskov, Angewandte Chemie - Int. Ed. 2006, 45, 2897.
[270] J. Rossmeisl, G. S. Karlberg, T. Jaramillo, J. K. Norskov, Faraday Discussions 2008, 140,
337. [271] V. Climent, J. M. Feliu, Journal of Solid State Electrochemistry 2011, 15, 1297. [272] J. O. Bockris, Transactions of Faraday Society 1947, 43, 417. [273] S. Trasatti, Journal of Electroanalytical Chemistry 1972, 39, 163. [274] L. Reinaudi, M. Delpopolo, E. Leiva, Surface Science 1997, 372, L309. [275] M. V. Mamonova, V. V. Prudnikov, Fizika Metallov i Metallovedenie 1998, 86, 33. [276] B.E. Nieuwenh, R. Bouwman, W. M. Sachtler, Thin Solid Films 1974, 21, 51. [277] B.E. Nieuwenh, O.G. Vanaarde, W. M. Sachtler, Chemical Physics 1974, 5, 418. [278] D. M. Kolb, Angewandte Chemie – Int. Ed. 2001, 40, 1162. [279] J. K. Norskov, T. Bligaard, J. Rossmeisl, C. H. Christensen, Nature Chemistry 2009, 1, 37. [280] J. Greeley, T. F. Jaramillo, J. Bonde, I. B. Chorkendorff, J. K. Norskov, Nature Materials
2006, 5, 909. [281] L. A. Kibler, M. A. El-Aziz, R. Hoyer, D. M. KOLB, Angewandte Chemie – Int. Ed. 2005,
44, 2080. [282] J. Clavilier, D. Armand, S. G. Sun, M. Petit, Journal of Electroanalytical Chemistry 1986,
205, 267. [283] A. Berna, V. Climent, J. M. Feliu, Electrochemistry Communications 2007, 9, 2789. [284] G. Beck, H. Fischer, E. Mutoro, V. Srot, K. Petrikowski, E. Tchernychova, M. Wuttig, M.
Ruehle, B. Luerssen, J. Janek, Solid State Ionics 2007, 178, 327. [285] M. Trassin, N. Viart, C. Ulhaq-Bouillet, G. Versini, S. Barre, C. Leuvrey, G. Pourroy,
Journal of Applied Physics 2009, 105. [286] B. Braunschweig, A. Mitin, W. Daum, Surface Science 2011, 605, 1082. [287] K. Uosaki, S. Ye, H. Naohara, Y. Oda, T. Haba, T. Kondo, Journal of Physical Chemistry
B 1997, 101, 7566. [288] M. Fayette, Y. Liu, D. Bertrand, J. Nutariya, N. Vasiljevic, N. Dimitrov, Langmuir 2011,
27, 5650.
References
212
[289] A. Maljusch, J. B. Henry, W. Schuhmann, A. S. Bondarenko, Electrochemistry
Communications 2012, 16, 88. [290] J. Clavilier, A. Rodes, K. Elachi, M. A. Zamakhchari, Journal de Chimie Physique 1991,
88, 1291. [291] M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, National Associa-
tion of Corrosion Engineers, Houston (1974). [292] Y. Shao-Horn, W. C. Sheng, S. Chen, P. J. Ferreira, E. F. Holby, D. Morgan, Topics in
Catalysis 2007, 46, 285. [293] R. Bashyam, P. Zelenay, Nature 2006, 443, 63. [294] Y. Gorlin, T. F. Jaramillo, Journal of the American Chemical Society 2010, 132, 13612. [295] M. Lefevre, E. Proietti, F. Jaouen, J. P. Dodelet, Science 2009, 324, 71. [296] G. Wu, K. L. More, C. M. Johnston, P. Zelenay, Science 2011, 332, 443. [297] T. Toda, H. Igarashi, H. Uchida, M. Watanabe, Journal of the Electrochemical Society
1999, 146, 3750. [298] V. R. Stamenkovic, B. S. Mun, K. J. J. Mayrhofer, P. N. Ross, N. M. Markovic, Journal of
the American Chemical Society 2006, 128, 8813. [299] I. E. L. Stephens, A. S. Bondarenko, L. Bech, I. Chorkendorff, ChemCatChem 2012,
DOI: 10.1002/cctc.201100343. [300] C. Wang, M. Chi, G. Wang, D. van der Vliet, D. Li, K. More, H.-H. Wang, J. A. Schlueter,
N. M. Markovic, V. R. Stamenkovic, Advanced Functional Materials 2011, 21, 147. [301] T. Bligaard, J. K. Norskov, Electrochimica Acta 2007, 52, 5512. [302] Kitchin, JR, J. K. Norskov, M. A. Barteau, J. G. Chen, Journal of Chemical Physics 2004,
120, 10240. [303] M. Mavrikakis, B. Hammer, J. K. Norskov, Physical Review Letters 1998, 81, 2819. [304] A. Schlapka, M. Lischka, A. Gross, U. Kasberger, P. Jakob, Physical Review Letters 2003,
91, 016802. [305] V. R. Stamenkovic, B. S. Mun, M. Arenz, K. J. J. Mayrhofer, C. A. Lucas, G. Wang, P. N.
Ross, N. M. Markovic, Nature Materials 2007, 6, 241. [306] B. B. Berkes, A. Maljusch, W. Schuhmann, A. S. Bondarenko, Journal of Physical
Chemistry C 2011, 115, 9122. [307] M. Huang, J. B. Henry, B. B. Berkes, A. Maljusch, W. Schuhmann, A. S. Bondarenko,
Analyst 2012, 137, 631.
References
213
[308] M. Huang, J.B. Henry, B.B. Berkes, A. Maljusch, W. Schuhmann, A.S. Bondarenko. Simultaneous acquisition of impedance and gravimetric data for the characterization of the electrode / electrolyte interfaces, in Lecture Notes on Impedance Spectroscopy.
Measurement, Modeling and Applications, O. Kanoun (ed.), Vol. 3, CRC Press (2012), accepted.
[309] N. Marcovic, P. N. Ross, Langmuir 1993, 9, 580. [310] A. Kuzume, E. Herrero, J. M. Feliu, Journal of Electroanalytical Chemistry 2007, 599,
333. [311] D. van der Vliet, C. Wang, M. Debe, R. Atanasoski, N. M. Markovic, V. R. Stamenkovic,
Electrochimica Acta 2011, 56, 8695. [312] M. T. Paffett, C. T. Campbell, T. N. Taylor, S. Srinivasan, Surface Science 1985, 154, 284. [313] X. Zhao, G. S. Frankel, B. Zoofan, S. I. Rokhlin, Corrosion 2003, 59, 1012. [314] A. B. Bayoumi, Engineering Fracture Mechanics 1996, 54, 879. [315] X. Liu, G. S. Frankel, B. Zoofan, S. I. Rokhlin, Corrosion Science 2007, 49, 139. [316] C. Blanc, B. Lavelle, G. Mankowski, Corrosion Science 1997, 39, 495. [317] R. G. Buchheit, M. A. Martinez, L. P. Montes, Journal of the Electrochemical Society
2000, 147, 119. [318] V. Guillaumin, P. Schmutz, G. S. Frankel, Journal of the Electrochemical Society 2001,
148, B163. [319] T. H. Muster, A. E. Hughes, Journal of the Electrochemical Society 2006, 153, B474. [320] C. Blanc, S. Gastaud, G. Mankowski, Journal of the Electrochemical Society 2003, 150,
B396. [321] L. Lacroix, L. Ressier, C. Blanc, G. Mankowski, Journal of the Electrochemical Society
2008, 155, C131. [322] O. Schneider, G. O. Ilevbare, Scully, JR, R. G. Kelly, Journal of the Electrochemical
Society 2004, 151, B465. [323] M. Buchler, J. Kerimo, F. Guillaume, W. H. Smyrl, Journal of the Electrochemical Society
2000, 147, 3691. [324] J. H. W. de Wit, Electrochimica Acta 2001, 46, 3641. [325] G.S. Frankel, Journal of ASTM International 2008, 5, 1. [326] M. A. Jakab, D. A. Little, J. R. Scully, Journal of the Electrochemical Society 2005, 152,
B311.
References
214
[327] E. J. Dufek, J. C. Seegmiller, R. C. Bazito, D. A. Buttry, Journal of the Electrochemical
Society 2007, 154, C458. [328] T. Suter, R. C. Alkire, Journal of the Electrochemical Society 2001, 148, B36. [329] G. O. Llevbare, J. R. Scully, Corrosion 2001, 57, 134. [330] J. C. Seegmiller, D. A. Buttry, Journal of the Electrochemical Society 2003, 150, B413. [331] R. G. Buchheit, L. P. Montes, M. A. Martinez, J. Michael, P. F. Hlava, Journal of the
Electrochemical Society 1999, 146, 4424. [332] C. R. Brooks, Heat treatment, Structure and Properties of Nonferrous alloys, ASM,
Metals Park, OH (1982). [333] L. Lacroix, L. Ressier, C. Blanc, G. Mankowski, Journal of the Electrochemical Society
2008, 155, C8. [334] K. S. Vecchio, D. B. Williams, Metallurgical Transactions A - Physical Metallurgy and
Material Science 1988, 19, 2885. [335] W. A. Cassada, G. J. Shiflet, E. A. Starke, Scripta Metallurgica 1986, 20, 751.
Appendix
215
#1
#3
#2
#6
#4
#8
#7
#9
#5
#10
9. Appendix
9.1. Technical drawings
SKP-SECM measurement head
Appendix
216
SKP-SECM measurement head, part #1
Appendix
217
SKP-SECM measurement head, part #2
Appendix
218
SKP-SECM measurement head, part #3
Appendix
219
SKP-SECM measurement head, part #4
Appendix
220
SKP-SECM measurement head, part #5
Appendix
221
SKP-SECM measurement head, part #6, #8, #9
Appendix
222
SKP-SECM measurement head, part #7
Appendix
223
SKP-SECM measurement head, part #10
Appendix
224
SKP-SECM measurement head (unit for reed relay)
Appendix
225
Adaptor plate for SKP-SECM measurement head
Appendix
226
Holder unit for SKP-SECM measurement head
Appendix
227
#1
#2
#3
Sample holder unit
Appendix
228
Sample holder unit, part #1
Appendix
229
Sample holder unit, part #2
Appendix
230
Sample holder unit, part #3
Appendix
231
Sample holder unit, part #4
Appendix
232
9.2. Electronic plans
Electronic circuit diagram of the integrator
Appendix
233
Connection diagram for individual components of the integrator
Appendix
234
Layout for custom fabrication of the circuit board for the integrator
Appendix
235
Connection diagram for individual components of the Kelvin current amplifier
Appendix
236
Layout for custom fabrication of the circuit board for the KC amplifier (1)
Layout for custom fabrication of the circuit board for the KC amplifier (2)
Appendix
237
9.3. List of abbreviations AC Alternating current
AC-SECM Alternating current mode of SECM
ADC Analog-to-digital conversion card
AFM Atomic Force Microscopy
AR-XPS Angle resolved X-ray spectroscopy
BSI Backscattered imaging mode
CE Counter electrode
CF Carbon fiber
CLSM Confocal laser scanning microscopy
CPD Contact potential difference
CV Cyclic voltammetry
DAQ Data acquisition board
DC Direct current
DFT Density functional theory
ECSTM Electrochemical Scanning Tunneling Microscopy
EDP Electrophoretic deposition paint
EDX Energy-dispersive X-ray analysis
EIS Electrochemical Impedance Spectroscopy
EPMA Electron probe microanalysis
EQCM Electrochemical quartz crystal microbalance
ET Electron transfer
FB Feedback mode of SECM
FEM Finite element method
GC Glassy carbon
GDH Glucose dehydrogenase
GOD Glucose oxidase
HER Hydrogen evolution reaction
HF High frequency
HOPG Highly ordered pyrolytic graphite
HRR Hydrogen reduction reaction
IC-SECM Intermittent contact mode of SECM
IDA Interdigitated array
Appendix
238
IMP Intermetallic particle
ITO Indium tin oxide
LCD Liquid crystal display
LEIS Localized Electrochemical Impedance Spectroscopy
LIA Lock-in amplifier
LSV Linear sweep voltammetry
ML Mono layer
MMS Mercury-mercury sulfate reference electrode
MPIE Max-Plank Institute for Iron Research
NSA Near-surface alloy
NSOM Near-field scanning optical microscopy
OA Operational amplifier
OCP open circuit potential
ORR Oxygen reduction reaction
PB Prussian Blue
PBS Phosphate buffered saline solution
PDP Potentiodynamic polarization
PEEK Polyether ether ketone
PEMFC Proton exchange membrane fuel cell
PVC Polyvinylchloride
PVD Physical vapor deposition
RC-SECM Redox-competition mode of SECM
RDE Rotating disk electrode
RE Reference electrode
RHE Reference hydrogen electrode
SCE Standard calomel electrode
SECM Scanning Electrochemical Microscopy
SEM Scanning Electron Microscopy
SGS Strain gauge sensor
SG-TC Sample-generation / tip-collection mode of SECM
SHE Standard hydrogen electrode
SIMS Secondary ion mass spectrometry
SI-SECM Surface interrogation mode of SECM
SKP Scanning Kelvin probe
Appendix
239
SKPFM Scanning Kelvin Probe Force Microscopy
SRET Scanning Reference Electrode Technique
SS Stainless steel (inox steel)
STM Scanning Tunneling Microscopy
SVET Scanning Vibrating Electrode Technique
SWV Square-wave voltammetry
TC Time constant
TG-SC Tip-generation / sample-collection mode of SECM
TPM Tip-position modulation
TTP Tetratolyl porphyrin
UHV Ultra high vacuum
UME Ultramicroelectrode
UPD Underpotential deposition
WE Working electrode
XRD X-ray diffraction
9.4. List of symbols ∆d Tip oscillation amplitude (SKP mode)
a Diameter of the active area of the tip (SECM mode)
A Active tip area (SKP mode)
C Capacitance
C Concentration of the active species converting at the SECM tip
Cdl Double layer capacitance
d Tip-to-sample distance (SECM mode)
d0 Main tip-to-sample distance (SKP mode)
D0 Diffusion coefficient of the active species or mediator
Dout Outer diameter
EAbs Absolute electrode potential
Ecorr Corrosion potential
Esample Sample potential (SECM mode)
Etip Tip potential (SECM mode)
F Faraday constant
f Frequency
Appendix
240
IKelvin Kelvin current (SKP mode)
iT Steady-state current at the SECM tip close to the sample surface
iT,∞ Steady-state current at the SECM tip in bulk solution
k Stiffness of the piezo actuator
k Reaction rate constant
Kp Proportional constant
m Load applied to the piezo actuator
n Number of transferred electrons
rtip Radius of the active area of the tip (SECM mode)
R Transimpedance
R Current magnitude (SECM mode)
Rp Polarization resistance
Rs Electrolyte resistance
S Sensitivity (SKP mode)
U0 Variable backing potential (SKP mode)
UCPD Potential gradient (SKP mode)
Um Additional alternating voltage (SKP mode)
Vpp Voltage (peek-to-peak)
Vrms Voltage (root mean square)
Z Impedance
ε Dielectric constant of the medium
θ Phase shift (SECM mode)
µe Chemical potential
τ Diffusion time
Φ Work function
φ Galvani potential
χ Dipole or surface potential
Ψ Volta potential or the outer potential
ω Angular vibration frequency of the tip (SKP mode)
ω0 Frequency of the additional alternating voltage (SKP mode)
∆φ Galvani potential difference
∆Ψ Volta potential difference or contact potential difference
Appendix
241
9.5. List of publications
Patents 1. R. Weber, J. Kintrup, W. Schuhmann, M. Bron, A. Maljusch, C.N. Tharamni, Electrode
material, electrode, and method for hydrogen chloride electrolysis. PCT Int. Appl. (2010), 25pp. CODEN: PIXXD2 WO 2010020365 A1 20100225 AN 2010:241121.
2. R. Weber, J. Kintrup, W. Schuhmann, M. Bron, A. Maljusch, C.N. Tharamni, Electrode
material, electrode, and method for hydrogen chloride electrolysis. Ger. Offen. (2010), 12pp. CODEN: GWXXBX DE 102008039072 A1 20100225 AN 2010:234350.
Published 1. J. B. Henry, A. Maljusch, M. Huang, W. Schuhmann, A. S. Bondarenko, Thin-Film Cu-
Pt(111) Near-Surface Alloys: Active Electrocatalysts for the Oxygen Reduction Reac-
tion, ACS Catalysis 2012, 2, 1457. 2. M. Huang, A. Maljusch, J. B. Henry, W. Schuhmann, A. S. Bondarenko, Probing electrode
/ electrolyte interface during intercalation of Cu into Te, Electrochemistry Communi-
cations 2012, 20, 92. 3. A. Maljusch, J. B. Henry, W. Schuhmann, A. S. Bondarenko, A quick method for the
preparation of Pt(111)-like thin films, Electrochemistry Communications 2012, 16, 88. 4. M. Huang, J. B. Henry, B. B. Berkes, A. Maljusch, W. Schuhmann, A. S. Bondarenko,
Towards a detailed in situ characterization of non-stationary electrocatalytic systems, Analyst 2012, 137 (3), 631.
5. C. Senöz, A. Maljusch, M. Rohwerder, W. Schuhmann, SECM and SKPFM Studies of the
Local Corrosion Mechanism of Al Alloys - A Pathway to an Integrated SKP-SECM System, Electroanalysis 2012, 24, 239.
6. A. Maljusch, B. Schönberger, A. Lindner, M. Stratmann, M. Rohwerder, W. Schuhmann, An
integrated SKP-SECM system: development and first applications, Analytical Chemistry 2011, 83 (15), 6114.
7. B. B. Berkes, A. Maljusch, W. Schuhmann, A. S. Bondarenko, Simultaneous acquisition of
impedance and gravimetric data in a cyclic potential scan for the characterization of non-stationary electrode/electrolyte interfaces, The Journal of Physical Chemistry B 2011, 115 (18), 9122.
8. A. Maljusch, T. C. Nagaiah, S. Schwamborn, M. Bron, W. Schuhmann, Pt-Ag Catalysts as
Cathode Material for Oxygen-Depolarized Electrodes in Hydrochloric Acid Electrolysis, Analytical Chemistry 2010, 82(5), 1890.
Appendix
242
9. T. C. Nagaiah, A. Maljusch, X. Chen, M. Bron, W. Schuhmann, Visualization of the Local
Catalytic Activity of Electrodeposited Pt-Ag Catalysts for Oxygen Reduction by means of SECM, ChemPhysChem 2009, 10(15), 2711.
10. L. Guadagnini, A. Maljusch, X. Chen, S. Neugebauer, D. Tonelli, W. Schuhmann, Visuali-
zation of electrocatalytic activity of microstructured metal hexacyanoferrates by means of redox competition mode of scanning electrochemical microscopy (RC-SECM), Electrochimica Acta 2009, 54(14), 3753.
Accepted 1. A. Maljusch, C. Senöz, M. Rohwerder, W. Schuhmann, Combined high resolution SKP-
SECM investigations for the visualization of local corrosion processes, Electrochimica
Acta 2012.
2. M. Huang, J. B. Henry, B. B. Berkes, A. Maljusch, W. Schuhmann, A. S. Bondarenko, Simultaneous acquisition of impedance and gravimetric data for the characterization of the electrode / electrolyte interfaces, in: Lecture Notes on Impedance Spectroscopy. Measure-
ment, Modeling and Applications, Volume 3 / O. Kanoun (ed.), CRC Press (2012).
In preparation 1. A. Maljusch, J. B. Henry, W. Schuhmann, A. S. Bondarenko, SKP-SECM: linking surface
properties and electrocatalytic activity, JACS, 2012. 2. J. B. Henry, A. Maljusch, W. Schuhmann, A. S. Bondarenko, Preparation of thin-film Cu-
Pt(111) near-surface alloys: towards up-scaling model single crystal surfaces, Electrochimica Acta, 2012.
3. M. Huang, A. Maljusch, S. Chakraborty, R. Schmidt, W. Schuhmann, A. S. Bondarenko,
Intercalation of Cu into Te films: combined theoretical and electrochemical study, Electrochimica Acta, 2012.
Appendix
243
9.6. Oral and poster presentations
Oral presentations 1. A. Maljusch, C. Senöz, M. Rohwerder, W. Schuhmann, “Combined high resolution SKP-
SECM investigations for the visualization of local corrosion processes”, Gordon Research
Seminar on Aqueous Corrosion 2012, New London (MA), USA (2012).
2. A. Maljusch, C. Senöz, M. Rohwerder, W. Schuhmann, “An integrated SKP-SECM system: design, development, optimization and first applications”, 62
nd Annual Meeting of the ISE,
Niigata, Japan (2011). 3. A. Maljusch, C. Senöz, M. Rohwerder, W. Schuhmann, “Combination of Scanning Electro-
chemical Microscopy and Scanning Kelvin Probe”, Workshop on SECM, Bochum, Germany (2011).
4. A. Maljusch, C. Senöz, M. Rohwerder, W. Schuhmann, “An integrated SKP-SECM system:
design, development, optimization and first applications”, IMPRS SurMat Seminar 2011, Meschede, Germany (2011).
5. A. Maljusch, C. Senöz, M. Rohwerder, W. Schuhmann, “An integrated SKP-SECM system:
design, development, optimization and first applications”, 6th
SECM workshop, Frejus, France (2010).
6. A. Maljusch, C. Senöz, M. Rohwerder, W. Schuhmann, “SKP-SECM: System Development,
Optimization and First Applications”, 61st Annual Meeting of the ISE, Nice, France (2010).
7. A. Maljusch, C. Senöz, M. Rohwerder, W. Schuhmann, “SKP-SECM: System Design, Deve-
lopment, Optimization and First Applications”, Electrochemistry 2010, Bochum, Germany (2010).
8. A. Maljusch, C. Senöz, M. Rohwerder, W. Schuhmann, “SKP-SECM: System Development
and First Applications”, Second Regional Symposium on Electrochemistry, Belgrade, Serbia (2010).
9. A. Maljusch, C. Senöz, M. Rohwerder, W. Schuhmann, “SKP-SECM: System Development
and First Applications”, IMPRS SurMat Seminar 2010, Meschede, Germany (2010). 10. A. Maljusch, C. Senöz, M. Rohwerder, W. Schuhmann, “SKP-SECM: System
Development and First Applications”, 216th meeting of the ECS, Vienna, Austria (2009). 11. A. Maljusch, C. Senöz, M. Rohwerder, W. Schuhmann, “SKP-SECM: System Development
and First Applications”, IMPRS SurMat Seminar 2009, Kleve, Germany (2009). 12. A. Maljusch, C. Senöz, M. Rohwerder, W. Schuhmann, “SKP-SECM: System Development
and First Applications”, International evaluation of the IMPRS SurMat, Düsseldorf, Germany (2009).
13. A. Maljusch, K. Eckhard, S. Schwamborn, M. Bron, W. Schuhmann, “Visualization of local
catalyst activity towards oxygen reduction in hydrochloric acid solution with RC-SECM”, 59
th Annual Meeting of the ISE, Seville, Spain (2008).
Appendix
244
Posters 1. A. Maljusch, C. Senöz, M. Rohwerder, W. Schuhmann, “Combined high resolution SKP-
SECM investigations for the visualization of local corrosion processes”, Gordon Research
Conference on Aqueous Corrosion 2012, New London (MA), USA (2012).
2. L. Guadagnini, A. Maljusch, D. Tonelli, W. Schuhmann, “Visualization of local electro-catalytic activity of metal hexacyanoferrates by means of Scanning Electrochemical Microscopy”, 5th SECM workshop, New York, USA (2009).
3. A. Maljusch, K. Eckhard, M. Bron, W. Schuhmann, “Visualization of local catalyst activity towards oxygen reduction in hydrochloric acid solution with RC-SECM”, FUNDP 2007, Namur, Belgium (2007.)
4. A. Maljusch, K. Eckhard, M. Bron, W. Schuhmann, “Visualization of local catalyst activity towards oxygen reduction in hydrochloric acid solution with RC-SECM” SMCBS 2007, Wlodovice, Poland (2007).
Appendix
245
9.7. Curriculum Vitae
Artjom Maljusch
Date and place of birth 11.08.1983, Belarus Nationality Belorussian E-Mail [email protected]
Doctoral Studies (PhD) in Electroanalytical Chemistry (10/2007 – 08/2012) AG Elektroanalytik & Sensorik, Faculty of Chemistry and Biochemistry, Ruhr-Universität Bochum, Bochum, Germany Supervisor: Prof. Dr. Wolfgang Schuhmann
Master of Science (M.Sc.) in Analytical Chemistry (10/2005 – 09/2007) AG Elektroanalytik & Sensorik, Faculty of Chemistry and Biochemistry, Ruhr-Universität Bochum, Bochum, Germany Supervisor: Prof. Dr. Wolfgang Schuhmann Title of thesis: „Visualization of local activity of oxygen reduction catalyst in hydrochloric solution“ Overall mark: 1.3 (very good), thesis mark: 1.0 (very good)
Bachelor of Science (B.Sc.) in Analytical Chemistry (09/2001 – 09/2005) Faculty of Chemistry, Belarusian State University, Minsk, Belarus Supervisor: Prof. Dr. Vladimir A. Vinarskij Title of thesis: „Amperometric biosensors based on imidazole-osmium complexes“ Overall mark: 9.4 (excellent, top 5%), thesis mark: 10 (excellent)
Scientific courses
T 1 (Basic concepts in materials science), mark: 1.3 (very good)
T 2 (Physical chemistry of surfaces and interfaces), mark: 1.0 (excellent, top 10%)
T 3 (Multiscale modeling), mark: 1.3 (very good)
T 4 (Deposition and properties of thin films and SAMs), mark: 1.0 (excellent, top 15%)
Fellowships Research fellowship by International Max-Plank Research School SurMat (06/2008 – 05/2011) DAAD fellowship for continuing education (10/2006 – 09/2007) Research fellowship by Belarusian State University (04/2005 - 08/2005)