Surface functionalised and
nanostructured Transparent
Conductive Oxides - towards a
platform for (bio)electrocatalysis
vorgelegt von
Master of Research (M.Res) in Molecular Modelling and
Materials Science
Tomos Gwilym Ab Alun Harris
ORCID: 0000-0003-2989-4127
von der Fakultät II - Mathematik und Naturwissenschaften
der Technischen Universität Berlin
zur Erlangung des akademischen Grades
Doktor der Naturwissenschaften
- Dr. rer. nat. -
genehmigte Dissertation
Promotionsausschuss:
Vorsitzender: Prof. Dr. Roel van der Krol
Gutachterin: Prof. Dr. Anna Fischer
Gutachter: Prof. Dr. Peter Hildebrandt
Gutachterin: Prof. Dr. Ulla Wollenberger
Tag der wissenschaftlichen Aussprache: 23. November 2018
Berlin 2019
ii
iii
ABSTRACT
Storage of (renewable) energy in the form of chemical bonds in electrofuels, their later
conversion back to electricity, or the generation of renewable chemical feedstocks from CO2, N2
etc. all require the use of stable and efficient catalysts. In particular, there is urgent need for the
development of efficient, precious-metal free catalysts. Enzymes and inorganic (e.g. biomimetic)
molecular catalysts can provide the necessary high turnover frequencies and selectivities, as
well as low overpotentials. The immobilisation, or heterogenisation, of enzymatic and molecular
catalysts on electrode surfaces is imperative, thereby increasing the number of addressable
active sites and facilitating electron transfer, especially when using hydrophobic catalysts in
aqueous conditions. Immobilisation can be carried out on conductive metal oxides, thus allowing
the combination of catalytic properties with certain electrode properties e.g. high surface areas
due to nanostructuration, transparency in the case of transparent conductive oxides (TCOs), or
photoadsorption in the case of narrow band semiconductors, which can drive reactions with light.
Current strategies for immobilising electrocatalysts on metal oxide surfaces use anchoring
groups such as phosphonic acids, carboxylic acids, silanes and their derivatives. While each
approach has its own merit e.g. high stability or good charge transfer properties, one usually
comes at the expense of the other. Development of new approaches is hence required to build
chemically and electrochemically robust interfaces which allow efficient and fast charge transfer
rates. In this context, the development of in situ spectroscopic techniques will provide real-time
monitoring of interface formation, catalyst binding and system evolution under reaction
conditions, thus allowing for a rational design of immobilisation strategies, far beyond a simple
trial-and-error approach.
In this work, electrografting of diazonium salts on nanostructured gold, indium tin oxide (ITO) and
mesoporous antimony-doped tin oxide (me-ATO) thin film electrodes was successfully used to
immobilise (bio)electrocatalysts. In situ surface enhanced infrared absorption (SEIRA)
spectroelectrochemistry demonstrates the remarkably broad electrochemical stability window of
3.0 V for the electrografted interfaces on gold in acetonitrile (ranging from -2.2 to +0.8 V (vs
Fc/Fc+)). Using the same in situ technique, a radical scavenger is shown to moderate interface
formation on gold, reducing heterogeneous charge transfer resistance enough to enhance direct
electron transfer between an immobilised redox active protein, namely an oxygen-tolerant [NiFe]
hydrogenase, and the electrode. This moderation approach is further employed successfully on
ITO. Furthermore, diazonium salts are used to immobilise a precious-metal free Mn-bipyridyl CO2
reduction molecular catalyst on gold, and a Cu-bipyridyl oxygen evolution reaction (OER)
molecular catalyst on ITO.
Precious-metal free oxygen reduction reaction (ORR) and OER ‘Hangman’ molecular catalysts
are also immobilised on me-ATO using electrografting. In situ attenuated total reflection-IR
spectroscopy, as well as IR spectroelectrochemistry, demonstrates the wide hydrolytic stability
window (in a pH range of ca. 2.5 - 12), and the broad electrochemical stability window of at least
3.0 V in aqueous media (ranging from -0.73 to 2.23 V (vs RHE)), and 2.9 V in organic media
(ranging from -1.3 V to 1.6 V (vs Fc/Fc+)) of the electrografted interfaces. In situ resonance
Raman spectroelectrochemistry was further used to determine the co-ordination of the
immobilised species and to demonstrate their excellent electrochemical accessibility. These
techniques were complemented by electrochemical, UV-vis and XPS measurements, as well as
applied theoretical calculations.
The work culminated in functioning precious-metal free electrocatalytic devices and the
demonstration that in situ spectroscopy and spectroelectrochemistry are powerful tools for
researchers working on oxide-based electrochemical devices. Furthermore, it was shown that
electrografting of diazonium salts results in highly stable interfaces suitable for catalyst
immobilisation on both gold and TCO materials.
iv
ZUSAMMENFASSUNG
Das Speichern von (erneuerbarer) Energie in Form von chemischen Bindungen in
Elektrotreibstoffen, sowie deren spätere Umwandlung zurück in Elektrizität oder in die
Erzeugung von erneuerbaren, chemischen Rohstoffen aus CO2, N2, usw., erfordern die
Anwendung von stabilen und effizienten Elektrokatalysatoren. Insbesondere besteht alleinschon
aus Kostengründen ein dringender Bedarf für die Entwicklung von effizienten und auch
edelmetallfreien Katalysatoren. Enzyme und anorganische (z. B. biomimetische) molekulare
Katalysatoren können die erforderlichen hohen Umsatzfrequenzen und Selektivitäten wie auch
niedrige Überpotentiale vorweisen. Daher ist eine Immobilisierung oder ‚Heterogenisierung’ von
enzymatischen und molekularen Katalysatoren auf Elektrodenoberflächen zwingend erforderlich,
wodurch die Anzahl der adressierbaren aktiven Zentren erhöht wird, und der elektronische
Kontakt erleichtert wird, insbesondere wenn hydrophobe Katalysatoren unter wässrigen
Bedingungen verwendet werden sollen. Die Immobilisierung kann auf leitfähigen Metalloxiden
durchgeführt werden, wodurch die katalytischen Eigenschaften mit bestimmten
Elektrodeneigenschaften kombiniert werden können, z.B. hohe Oberflächen durch
Nanostrukturierung, Transparenz im Fall von transparenten leitfähigen Oxiden (TCO), sowie
Photoabsorption bei schmalbandigen Halbleitern, die entsprechenden Reaktionen dann auch mit
Licht steuern können.
Aktuelle Strategien zur Immobilisierung von Elektrokatalysatoren auf Metalloxidoberflächen
verwenden Ankergruppen wie Phosphonsäuren, Carbonsäuren, Silane und ihre Derivate.
Während jeder dieser Immobiliserungsansätze spezifische Vorteile mit sich bringt, z.B. in Bezug
auf hohe Stabilität oder gute Ladungsübertragungseigenschaften, geht die Optimierung einer
dieser Eigenschaften meist zu Lasten einer anderen. So ist zum Aufbau chemisch und
elektrochemisch robuster Grenzflächen, die effiziente und schnelle Ladungsübertragungsraten
ermöglichen sollen, auch die Entwicklung neuer Immobilisierungstechniken erforderlich. In
diesem Zusammenhang ermöglichen in situ spektroskopische Techniken durch eine
Echtzeitüberwachung der Grenzflächenbildung, der Katalysatoranbindung und der
Systementwicklung unter Reaktionsbedingungen ein rationales Design von
Immobilisierungsstrategien, welche weit über einen einfachen empirischer Ansatz hinaus gehen.
In dieser Arbeit wurde das sogenante „Elektrografting“, ein elektrisch gesteuertes „Aufbringen“
von Diazoniumsalzen an nanostrukturierten Gold-, Indiumzinnoxid- (ITO) und mesoporösen
Antimon-dotierten Zinnoxid (me-ATO) -Dünnschichtfilm-Elektroden erfolgreich zur
Immobilisierung von BioElektrokatalysatoren verwendet. Durch in situ oberflächenverstärkte
Infrarot- Absorptions-Spektroskopie (SEIRA) konnte ein bemerkenswert breites
elektrochemisches Stabilitätsfenster von 3.0 V für solche „elektrogegrafte“ Grenzflächen an Gold
in Acetonitril (von -2.2 bis +0.8 V vs Fc/Fc+) gezeigt werden. Unter Verwendung derselben in
situ-Technik wurde gezeigt, dass ein Radikalfänger die Grenzflächenbildung auf Gold
einschränkt, wodurch der heterogene Ladungsübertragungswiderstand ausreichend verringert
wird, um den direkten Elektronentransfer zwischen einem immobilisierten redox-aktiven Protein
und der Elektrode zu verbessern, nämlich einer sauerstofftoleranten [NiFe]-Hydrogenase. Dieser
Ansatz wurde erfolgreich für die Oberflächenfunktionalisierung an ITO angewandt. Darüber
hinaus werden Diazoniumsalze zur Immobilisierung von einem edelmetallfreien, molekularen
Mn-Bipyridyl-CO2-Reduktionskatalysator und einem Cu-Bipyridyl-Sauerstoffproduktionsreaktion
(OER) - Katalysator auf Gold und ITO verwendet.
Edelmetallfreie, molekulare Sauerstoffreduktionsreaktion (ORR) und OER "Hangman"-
Katalysatoren werden ebenfalls auf me-ATO mittels Elektrografting immobilisiert. In situ-IR-
Spektroskopie im Einsatzmodus der abgeschwächten Totalreflexion sowie IR-
Spektroelektrochemie zeigen ein breites hydrolytisches Stabilitätsfenster dieser Grenzflächen (in
einem Bereich von ca. 2,5 bis 12 pH-Einheiten) auf, wie auch ein breites elektrochemisches
Stabilitätsfenster von mindestens 3,0 V in wässrigen Medien (von -0,73 bis 2,23 V vs RHE) und
v
2,9 V in organischen Medien (von -1,3 bis 1.6 V vs Fc/Fc+) der Grenzflächen. Weiterhin wurde in
situ-Resonanz-Raman-Spektroelektrochemie verwendet um den Koordinationszustand der
immobilisierten Spezies zu bestimmen und ihre ausgezeichnete, elektrochemische
Zugänglichkeit zu demonstrieren. Diese Techniken wurden durch elektrochemische, UV-Vis- und
XPS-Messungen sowie durch theoretische Berechnungen ergänzt.
Die Arbeit gipfelte so in funktionierenden, edelmetallfreien elektrokatalytischen Geräten und dem
Nachweis, dass In situ-Spektroskopie und Spektroelektrochemie leistungsfähige Techniken für
Forscher sind, die an oxid-basierten elektrochemischen Aufbauten arbeiten. Darüber hinaus
wurde gezeigt, dass das „Elektrografting“, das sogenannte elektrisch-gesteuerte „Aufbringen“
von Diazoniumsalzen zu sehr stabilen Grenzflächen führt, die sich sowohl für Gold- als auch für
TCO-Materialien zur Immobilisierung von Katalysatoren eignen.
vi
PUBLICATIONS
Parts of this work are published in the following articles
T. G. A. A. Harris, R. Götz, P. Wrzolek, V. Davis, P. Hildebrandt, M. Schwalbe, I. Weidinger, I
Zebger, A. Fischer.
Title: “Robust electrografted interfaces on metal oxides for electrocatalysis – an in situ
spectroelectrochemical study”. J. Mater. Chem. A, 2018, 6 (31), 15200–15212.
T. G. A. A. Harris, N. Heidary, J. Kozuch, S. Freilingsdorf, O. Lenz, P. Hildebrandt, I. Zebger, A.
Fischer.
Title: “In-situ spectroelectrochemical studies into formation and stability of robust diazonium-
derived interfaces on gold electrodes for the immobilization of an oxygen-tolerant hydrogenase”.
ACS Appl. Mater. Interfaces, 2018, 10 (27), 23380-23391.
T. G. A. A. Harris, S. Rauwerdink, A. Tahraoui, S. Freilingsdorf, O. Lenz, P. Hildebrandt, I.
Zebger, A. Fischer. Title: “Electrografted interfaces on metal oxides for enzyme immobilisation
and bioelectrocatalysis”. Manuscript in preparation
Other publications
H. Gatemala, T. G. A. A. Harris, L. Pardo-Perez, C. Querebillo. S. Frielingsdorf, O. Lenz, A.
Fischer, I. Weidinger, S. Ekgasit, N. Heidary, H. K. Ly, I. Zebger. “A Graphene Oxide Modified
Hybrid Surface as an Alternative Platform for SEIRA Spectro-electrochemical Studies”.
Manuscript in preparation
vii
SELECTED TALKS
“In situ spectroelectrochemical approches to study molecular interfaces on conductive oxides –
Immobilisation of electrocatalysts using diazonium chemistry“, 4th International Symposium on
Chemistry for Energy Conversion & Storage, Berlin, January 28 – 31 2018.
“In situ spectroelectrochemical approches to study molecular interfaces on conductive oxides –
Immobilisation of electrocatalysts using diazonium chemistry“, 2017 Fall Meeting – Materials
Research Society (MRS), Boston, USA, November 26 – December 1 2017.
“Spectroelectrochemical Insights into the Formation of Robust Interfaces for the Immobilization of
Biological and Molecular Electrocatalysts“, Electrochemistry 2016 - Gesellschaft Deutscher
Chemiker (GDCh), Goslar, September 26 – 28 2016.
“Spectroelectrochemical Insights into the Formation of Robust Interfaces for the Immobilization of
Biological and Molecular Electrocatalysts“, 229th Spring Meeting - Electrochemical Society
(ECS), San Diego, USA, May 29 – June 3 2016.
SELECTED POSTERS
“Diazonium salts for attaching molecular catalysts to metal oxide surfaces – an in situ
spectroelectrochemical study”, GDCh-Wissenschaftsforum Chemie 2017, Berlin, September 10 –
14, 2017.
“Highly stable interfaces for the immobilisation of molecular catalysts on metal oxide materials –
an in situ spectroelectrochemical study”, 5th International Conference on Multifunctional, Hybrid
and Nanomaterials, Lisbon, Portugal, March 6 – 10, 2017
viii
ABBREVIATIONS
4-BABD – 4-benzamidobenzenediazonium
tetrafluoroborate
4-NBD – 4-nitrobenzenediazonium
tetrafluoroborate
Ag/AgCl – silver/silver chloride electrode
ATO – antimony-doped tin oxide
ATR – attenuated total reflection
BE – binding energy
bpy-diazo – 4-([2,2’-Bipyridine]-4-
carboxamido)benzenediazonium
tetrafluoroborate
CME – chemically modified electrode
CNT – carbon nanotubes
CPE – constant potential electrolysis
CV – cyclic voltammetry
DFT – density functional theory
DI – deionised
DPPH – 2,2-diphenyl-1-picrylhydrazyl
EFC – enzymatic fuel cell
EISA – evaporation induced self-assembly
ESI-MS – electrospray ionisation mass
spectroscopy
ET – electron transfer
Fc/Fc+ – ferrocene/ferrocenium
FTO – fluorine-doped tin oxide
GC – glassy carbon
HER – hydrogen evolution reaction
HOR – hydrogen oxidation reaction
HSM – high surface area nanomaterial
Im-diazo – 4-(1H-imidazol-1-
yl)benzenediazonium tetrafluoroborate
IR – infrared
ITO – indium tin oxide
MBH – membrane-bound hydrogenase
me-ATO – mesoporous antimony-doped tin
oxide
MWCNT – multi-walled carbon nanotubes
OER – oxygen evolution reaction
ORR – oxygen reduction reaction
PB – phosphate buffer
PFV – protein film voltammetry
pl-ITO – planar indium tin oxide
RCT – charge transfer resistance
RDE – rotating disc electrode
Re – Ralstonia eutropha
RHE – reversible hydrogen electrode
rR – resonance Raman
SAM – self-assembled monolayer
sc-ITO – spin-coated indium tin oxide
SEIRA - surface enhanced infrared
absorption
SEM – scanning electron microscopy
TBAF – tert-butylammonium
hexafluorophosphate
TBAP – tert-butylammonium perchlorate
TCO – transparent conductive oxide
TOF – turnover frequency
UV-Vis – ultraviolet visible
XPS – X-ray photoelectron spectroscopy
ΔEP – peak separation
TABLE OF CONTENTS
1 Introduction and motivation …………………………………………………………1
Theory
2.1 Transparent conductive oxides (TCOs) .................................................................. 4
2.1.1 Conductivity in semiconductors ........................................................................... 4
2.1.2 The band structure of semiconductors and the Fermi energy ............................. 5
2.2 Electrode processes ................................................................................................. 8
2.2.1 Fermi energy and electrochemical potential ........................................................ 8
2.2.2 The electrode-electrolyte interface .................................................................... 10
2.2.3 The relation between current and potential ....................................................... 14
2.3 Surface modification of electrode materials ........................................................ 17
2.3.1 Chemically modified electrodes ......................................................................... 17
2.3.2 Langmuir-Blodgett assembly ............................................................................. 18
2.3.3 Self-assembly .................................................................................................... 19
2.3.4 Modification of TCO’s ........................................................................................ 20
2.3.5 Electrochemical modification ............................................................................. 23
2.3.6 Modification using diazonium salts .................................................................... 24
2.3.7 Electrochemical grafting using diazonium salts ................................................. 26
2.3.8 Electrochemical response of adsorbed monolayers .......................................... 27
2.3.9 Immobilisation of electrocatalysts on electrode surfaces ................................... 29
2.4 Spectroscopic methods ......................................................................................... 33
2.4.1 Spectroelectrochemistry .................................................................................... 33
2.4.2 UV-Vis spectroscopy ......................................................................................... 33
2.4.3 Vibrational spectroscopy .................................................................................... 35
2.4.4 Infrared spectroscopy ........................................................................................ 36
2.4.5 Surface Enhanced Infrared Absorption (SEIRA) spectroscopy ......................... 38
2.4.6 Raman spectroscopy ......................................................................................... 39
2.4.7 X-ray photoelectron spectroscopy ..................................................................... 41
2.5 Hydrogenases ......................................................................................................... 42
2.6 Molecular catalysts ................................................................................................. 48
2.6.1 Oxygen reduction reaction at metal macrocycles .............................................. 48
2.6.2 Water oxidation molecular catalysts .................................................................. 50
3 Materials, instruments and methods……………………………………………..51
3.1 Electrode synthesis ................................................................................................ 51
3.1.1 Gold SEIRA film preparation .............................................................................. 51
3.1.2 Indium tin oxide electrode preparation ............................................................... 51
3.1.3 Antimony-doped tin oxide electrode preparation ............................................... 52
3.2 Precursor synthesis ............................................................................................... 53
3.2.1 Diazonium salts ................................................................................................. 53
3.3 Protein purification and catalyst preparation ...................................................... 54
3.3.1 Membrane-Bound Hydrogenase (MBH) from Ralstonia eutropha ..................... 54
3.3.2 Hangman complexes ......................................................................................... 55
3.4 Electrochemistry ..................................................................................................... 55
3.5 Electrode modification ........................................................................................... 56
3.5.1 Functionalisation of gold .................................................................................... 56
3.5.2 Functionalisation of ITO ..................................................................................... 57
3.5.3 Functionalisation of ATO ................................................................................... 58
3.5.4 Enzyme immobilisation ...................................................................................... 58
3.6 Spectroscopic characterisation ............................................................................ 59
3.6.1 ATR-IR and SEIRA ............................................................................................ 59
3.6.2 Density functional theory ................................................................................... 60
3.6.3 X-ray photoelectron spectroscopy ..................................................................... 60
3.6.4 Resonance Raman ............................................................................................ 61
3.6.5 UV-Vis ................................................................................................................ 61
4 Spectroelectrochemcial investigation into the electrochemical grafting of
diazonium salts on electrodes for the immobilisation of biological and
molecular catalysts……………………………………………………………………….62
4.1 Introduction ............................................................................................................. 62
4.2 Electrochemical reduction of diazonium salts on gold ...................................... 66
4.3 Electrochemical stability of electrografted diazonium interfaces on gold ....... 68
4.4 Effect of radical scavenger on interface formation ............................................. 71
4.5 Spontaneous adsorption of diazonium salts on gold ......................................... 75
4.6 Immobilisation of hydrogenase on electrografted diazonium interfaces ......... 77
4.7 Immobilisation of molecular catalysts on gold using electrografting of
diazonium salts ................................................................................................................. 80
4.8 Conclusions ............................................................................................................ 87
5 Electrochemical grafting of diazonium salts on transparent conductive
oxide (TCO) electrodes for the immobilisation of enzymes………………………88
5.1 Introduction ............................................................................................................. 88
5.2 Effect of radical scavenger on interface formation and immobilisation of
hydrogenase ...................................................................................................................... 89
5.3 Conclusions .......................................................................................................... 106
6 Electrochemical grafting of diazonium salts on transparent conductive
oxide (TCO) electrodes for the immobilisation of molecular catalysts………..108
6.1 Introduction ........................................................................................................... 108
6.2 Immobilisation of a molecular oxygen evolution reaction catalyst on indium tin
oxide (ITO) using the electrografting of a diazonium salt ........................................... 112
6.3 Chemical and electrochemical stability of electrografted diazonium interfaces
on mesoporous antimony-doped tin oxide (me-ATO) ................................................. 118
6.4 Fe-Hangman complexes immobilised on planar and porous TCOs using the
electrografting of a diazonium salt for oxygen reduction reaction ............................ 128
6.5 Co Hangman complex immobilised on me-ATO using the electrografting of a
diazonium salt for the oxygen evolution reaction ....................................................... 149
6.6 Conclusions .......................................................................................................... 161
7 Appendices…………………………………………………………………………..164
7.1 Appendix to chapter 4.1 ....................................................................................... 164
7.2 Appendix to chapter 4.2 ....................................................................................... 167
7.3 Appendix to chapter 4.3 ....................................................................................... 168
8 Bibliography 171
1
2
3
1
4 Introduction and motivation
Chapter 1
Introduction and motivaton
As societies begin to decarbonise their energy systems there is an increasing focus on methods
for storing and concentrating renewable energy, or lack thereof. While the cost of solar and wind
energy has plummeted to such an extent that they have become more competitive than natural
gas in some areas1, they are still intermittent in supply (often resulting in oversupply) and located
far from demand. One solution is to store electrical or solar energy in the form of chemical bonds
in electrofuels for later conversion into power or heat in fuel cells or through combustion.
Electrofuels are generally compatible with current industrial infrastructure and supply chains and
can therefore be decoupled from the electricity grid e.g. in the form of liquid fuels or for use as
chemical feedstocks in industry. Even if currently more costly than storing energy in batteries,
electrofuels are a complementary approach, especially in situations where high energy densities
are required (e.g. in aviation fuels), or in the conversion of CO2 captured from unavoidable
sources (e.g. cement kilns or steel plants).2
Scheme 1
An overview of the different pathways available for the generation of electrofuels and chemical feedstocks from
renewable energy via electrolysers or photoelectrochemical cells (PECs), and their potential deployment in the
economy, within existing industrial infrastructure and supply chains.
Scheme 1 gives an overview of how energy can be stored in the form of chemical energy using
electrolysers or photoelectrochemical cells (PECs) and how these chemicals can be used for
2
different purposes. Typical reactions include the cleavage of water via the so called “oxygen
evolution reaction” (OER), yielding molecular oxygen O2, protons and electrons. The protons can
be further reduced via the corresponding “hydrogen evolution reaction” (HER), yielding molecular
hydrogen H2, a clean and efficient energy carrier, or can alternatively participate in the reduction
of CO2, yielding various energy carriers such as methane, methanol, or chemical feedstocks such
as formic acid, carbon monoxide or ethylene.3 Such reactions, as well as the respective “oxygen
reduction reaction” (ORR) and “hydrogen oxidation reaction” (HOR) that convert the chemical
energy back into useable electrical energy (i.e. in fuel cells), require the use and development of
efficient electrocatalysts, which allow them to proceed at high rates (turnover frequencies, TOF’s)
with low diving forces (overpotentials), thus making them more economically viable. Many of the
most efficient catalysts available today for the aforementioned reactions are based on
prohibitively expensive precious metals, giving an impetus for the development of efficient,
precious-metal free catalysts.
Enzymes, as possible blueprints from nature, and inorganic (e.g. biomimetic) molecular catalysts
can exhibit the high TOF’s and selectivities, as well as low overpotentials. Furthermore, these
properties can be easily fine-tuned chemically by modification of the 1st or the 2nd coordination
sphere. Many spectroscopic methods can easily be deployed with such catalysts, providing in-
depth analysis of reaction mechanisms, which is otherwise much more difficult with
heterogeneous catalysts, and thus facilitating rational catalyst design. The immobilisation, or
heterogenisation of enzymatic and molecular catalysts on electrode surfaces is imperative,
thereby increasing the number of addressable active sites and facilitating charge transfer
between the catalyst and the electrode. This is especially important for reactions in water using
hydrophobic catalysts. Immobilisation can be carried out on conductive metal oxides or other
electrode substrates, thus allowing the combination of catalytic properties with certain electrode
properties e.g. high surface areas due to nanostructuration, and transparency in the case of
transparent conductive oxides (TCOs), or photoabsorption in the case of narrow band
semiconductors, which can drive reactions with light. In particular, TCOs such as antimony-doped
tin oxide (ATO) are low-cost and highly stable4, and therefore of interest in prototype devices.
Current strategies for attaching electrocatalysts to oxide surfaces employ anchoring groups such
as phosphonic acids, carboxylic acids, silanes and their derivatives.5,6 While each approach has
its own merit in terms of stability or charge transfer properties, one usually comes at the expense
of the other, or may require complex synthetic chemical approaches or additional stabilisation
steps. Development of new approaches is hence required to build chemically and
electrochemically robust interfaces that allow efficient and fast charge transfer rates.6,7 In this
context, the development of in situ spectroscopic techniques providing real-time monitoring of
interface formation, stability and evolution under reaction conditions will allow for a rational design
of immobilisation strategies, far beyond simple trial and error approaches.
3
Of particular interest are in situ surface sensitive infrared (IR) spectroscopic and
spectroelectrochemical methods, which can provide the aforementioned real-time monitoring of
the interface, as well as monitoring of catalyst immobilisation and mechanistic insights into the
catalytic processes occurring at the interface. Such detailed information cannot be obtained
through conventional characterisation approaches used with oxide electrodes, such as UV-vis
spectroelectrochemistry or ex situ x-ray photoelectron spectroscopy (XPS).
In this work, diazonium electrografting was successfully used to immobilise a number of precious-
metal free catalysts on gold and indium tin oxide (ITO) and ATO conductive oxide electrodes,
including catalysts active for the OER and ORR, as well as a hydrogenase enzyme active for the
HOR. I -situ surface sensitive IR methods were used to characterise the electrografted interfaces
and demonstrate their high electrochemical and chemical stabilities. In situ resonance Raman
was further used on mesoporous ATO to characterise the immobilised species and demonstrate
their excellent electrochemical accessibility. Both of these techniques are complemented by
electrochemical, UV-vis and XPS measurements, as well as theoretical calculations. This
culminated in functioning precious-metal free electrocatalytic devices and the demonstration that
in situ IR spectroscopy is a powerful tool for researchers working on oxide-based electrochemical
devices, including electrolyzers, sensors, or dye-sensitized solar cells and photocatalytic cells i.e.
anywhere inorganic-molecular or inorganic-biological interfaces play an important role in a
device’s performance.
4
5 Theory
Chapter 2
Theory
2.1 Transparent conductive oxides (TCOs)
2.1.1 Conductivity in semiconductors
When a potential difference is applied across a material, an electric field exerts a force on the
charge carriers causing them to accelerate and gain momentum. As the charge carriers move
through the crystal lattice they will encounter phonons, lattice impurities and defects, all of which
will alter their momentum. Assuming momentum is completely lost after each collision, the
average carrier velocity, or drift velocity 𝑣𝑣", can be written as:
𝒗𝒗𝐝𝐝=𝒒𝒒𝒒𝒒𝛕𝛕
𝒎𝒎𝟏𝟏
where 𝐸𝐸 is the applied electric field, 𝑞𝑞 the carrier charge, 𝑚𝑚 the charge carrier mass, and τ the
mean free time, or scattering time, between collisions, which is in turn related to the mean free
path 𝑙𝑙. This can also be written in terms of carrier mobility 𝜇𝜇2 such that:
𝜇𝜇2=𝑞𝑞τ
𝑚𝑚2
Assuming all carriers move with their drift velocity, the current density 𝐽𝐽5 (in Am-2) of the carriers
moving in the x direction can be written as the product of the density of electrons 𝑛𝑛 and the drift
velocity:
𝐽𝐽5=𝑛𝑛𝑞𝑞𝑣𝑣"3
Combining equations 1 and 3 gives the relationship between current density and applied
electric field:
𝐽𝐽5=𝑛𝑛𝑞𝑞9τ
𝑚𝑚𝒒𝒒𝒙𝒙4
which gives a linear relationship, which is in fact Ohms law 𝐽𝐽 = 𝜎𝜎𝒒𝒒. Conductivity 𝜎𝜎 can thus be
written as8,9:
𝜎𝜎 = 𝑞𝑞τ
𝑚𝑚𝑛𝑛𝒒𝒒𝒙𝒙= 𝜇𝜇2𝑛𝑛𝒒𝒒𝒙𝒙5
5
2.1.2 The band structure of semiconductors and the Fermi energy
Free atoms have well defined energy levels. When n identical atoms come together and interact,
the Pauli principle dictates that they form n degenerate energy levels. On bringing a very large
number of identical atoms together in a solid, the energy difference between the degenerate
energy levels becomes so small that they can essentially be considered as continuous energy
bands with forbidden energy gaps between them. In general, when considering a solid, only two
bands need to be considered: the highest occupied band, known as the valence band (VB), and
the next highest band, which is the lowest unoccupied band, and known as the conduction band
(CB). Between the valence band maximum (VBM) and the conduction band minimum (CBM)
exists a region with no electronic states that is known as the band gap. The Fermi level describes
the electrochemical potential of electrons in a solid, or the energy required to add or remove an
electron from that solid. The probability of an electronic state with energy 𝐸𝐸 being occupied is
given by the Fermi-Dirac distribution:
F 𝐸𝐸 = 1
1 + exp 𝐸𝐸 − 𝐸𝐸D
𝑘𝑘F𝑇𝑇
6
where 𝑘𝑘F is the Boltzmann constant, 𝑇𝑇 is the temperature and 𝐸𝐸D is the Fermi energy level. At
𝑇𝑇 = 0 K, F𝐸𝐸 is a step function equal to 1, meaning that all the electronic states up to 𝐸𝐸D are
occupied and, therefore, 𝐸𝐸D describes the maximum energy level occupied by an electron.9
Scheme 2
Energy band diagram for a metal, a semiconductor and an insulator, indicating the position of the Fermi level 𝐸𝐸D
relative to the band edges (the valence band maximum 𝐸𝐸IFJ and the conduction band minimum 𝐸𝐸KFJ and the
band gap 𝐸𝐸L of the material.
In order for a material to conduct electricity, electrons must be able to gain momentum on
application of an electric field and move to higher unoccupied energy states. In the case of
metals, 𝐸𝐸D lies within an energy band meaning unoccupied states exist above 𝐸𝐸D, therefore
allowing electrical conductivity. In an insulator, 𝐸𝐸D lies within the band gap so that at 0 K all of the
6
valence band states are occupied and all of the conduction band states are unoccupied, meaning
the material cannot conduct electricity. At finite temperatures, electrons can be thermally excited
into states above the Fermi level. Now 𝐸𝐸D gives the energy level at which there is a 50%
probability of being occupied. The Fermi-Dirac function (equation 6) varies exponentially away
from 𝐸𝐸D, therefore the degree of conductivity depends on the band gap 𝐸𝐸L of the material.8,9
Semiconductors are said to have a bandgap values of 0 < 𝐸𝐸L < 4 eV and insulators 𝐸𝐸L ≥ 4 eV.10
TCOs exhibit high electrical conductivity while maintaining transparency in the visible range, and
therefore must have an optical band gap greater than 3.2 eV. Energy band diagrams for a metal,
a semiconductor and an insulator are illustrated in Scheme 2.
Charge neutrality requires the number of electrons n occupying states above 𝐸𝐸D to be equal to
the number of unoccupied states, or holes, beneath 𝐸𝐸D. If the density of states at the CBM and
the VBM in a material are identical, as illustrated in the simplified model in Scheme 2, then 𝐸𝐸D will
sit exactly in the middle of the bandgap i.e. 𝐸𝐸D=MNOPQMROP
9. The charge carrier density in the
conduction band can be obtained by approximating the Fermi-Dirac integral using a Boltzmann
distribution, where the density is an exponential function of the distance between 𝐸𝐸D and 𝐸𝐸KFJ:
𝑛𝑛 ≈ 𝑁𝑁KFJexp −𝐸𝐸KFJ − 𝐸𝐸D
𝑘𝑘F𝑇𝑇6
Similarly, the number of holes occupying the valence band is given by:
𝑝𝑝 ≈ 𝑁𝑁IFJexp −𝐸𝐸D− 𝐸𝐸IFJ
𝑘𝑘F𝑇𝑇7
where 𝑁𝑁KFJ and 𝑁𝑁IFJ are the effective density of states at the conduction band minimum and
valence band maximum, respectively. This approximation breaks down when 𝐸𝐸D becomes too
close to either the conduction band minimum and valence band maximum, as is often the case
for doped semiconductors.
7
Scheme 3
The band structure, density of states N(E) and Fermi Dirac distribution F(E) for an intrinsic semiconductor, an n-
type doped extrinsic semiconductor, and a p-type doped extrinsic semiconductor. The additional donor and
acceptor states introduced through doping are indicated by Ed and Ea, respectively.
There are two cases in which 𝐸𝐸D can deviate from the middle of the band gap: either when the
density of states at the band edges are not equal to one another, or when electronic states exist
somewhere inside the formally forbidden band gap. While the former case only leads to a small
deviation from 𝐸𝐸D=MNOPQMROP
9, the latter case can push 𝐸𝐸D closer to either band edge. These
states can be introduced through doping, i.e. through introducing electron donor or acceptor
impurities into the materials lattice (n-type or p-type doping), to give extrinsic semiconductors. In
the case where there is no significant doping of the semiconductor, it is called an intrinsic
semiconductor. The additional donor or acceptor states introduced by dopant species do not sit in
the valence or conduction bands. If the distance between e.g. the donor state and the conduction
band minimum is in the order of 𝑘𝑘F𝑇𝑇, then electrons can be thermally excited from the donor state
and into the delocalised conduction band, and therefore participate in charge transport. If the
donor densities become large enough, the electronic states introduced inside the band gap can
interact forming their own electronic band, where electrons can be ionised to or from.
Metal oxide semiconductors differ to conventional semiconductors, such as Si, in that the
electronegativity of the oxygen results in a total or partial transfer of the valence electrons from
the metal ion to the oxygen giving rise to their ionic bonding character. Conductivity in oxides
8
therefore arises from cationic or anionic dopants, or intrinsic point defects such as oxygen
vacancies. When doping an oxide with aliovalent dopant atoms (i.e. atoms having a different
valency), there must either be an ionic or an electronic compensation mechanism that preserves
charge and lattice stoichiometry.11,12 n-type conductivity can be induced in undoped indium oxide
(In2O3) through the introduction of oxygen vacancies, as described using the Kröger-Vink notation
shown below (for further information on these notations see 13):
OX
Y⇌ VX
•• +12O9 L + 2𝑒𝑒^8
The oxygen anions leave doubly ionised vacancy sites and two free electron charge carriers upon
their removal from the crystal structure. Alternatively, doping with aliovalent cations can also
induce compensation mechanisms in semiconductors. n-type conductivity can be induced in this
way in indium oxide through the introduction of tin oxide (SnO2), yielding indium tin oxide (ITO)14:
2SnO9→2Sncd
•+3OX
Y+12O9 L + 2𝑒𝑒^9
n-type conductivity can also be induced in undoped tin oxide SnO2 through the introduction of
oxygen vacancies, like in In2O3, or through the introduction of antimony, yielding antimony-doped
tin oxide (ATO):
2Sb9Og→2Sbhd
•+4OX
Y+12O9 L + 2𝑒𝑒^10
2.2 Electrode processes
2.2.1 Fermi energy and electrochemical potential
The electrochemical potential of a charged species i measures the partial molar Gibbs free
energy of that species in a phase with an inner potential 𝜑𝜑:
𝜇𝜇l=𝜕𝜕𝐺𝐺
𝜕𝜕𝑛𝑛lo,q,r,s
11
The electrochemical potential can also be expressed as the components associated with the
chemical species and its charge:
𝜇𝜇l= 𝜇𝜇l+ 𝑧𝑧l𝐹𝐹𝜑𝜑 12
where 𝜇𝜇l is the chemical potential, and the second term represents the electrical work done in
transferring one mole of charge 𝑧𝑧l𝐹𝐹, where 𝑧𝑧l is the charge of the species i, and 𝜑𝜑 is the local
electrostatic potential, and 𝐹𝐹 is the Faraday constant.15 The chemical and electrochemical
potential of an uncharged species is the same. For a metal at 𝑇𝑇 = 0 K, the highest occupied state
9
is the Fermi level 𝐸𝐸D, and therefore 𝜇𝜇c =𝐸𝐸D, since any electron added must occupy the Fermi
level. At finite temperatures, the values of 𝜇𝜇l and 𝐸𝐸D will differ by an amount in the order of (𝑘𝑘𝑇𝑇)2,
which is negligible in most cases. The work function 𝜙𝜙 is defined as the minimum work required
to remove an electron from inside the metal to a state just outside the surface of the metal,
therefore 𝐸𝐸D = − 𝜙𝜙. If the reference point is taken as the vacuum level, then 𝐸𝐸D = −𝜙𝜙 − 𝑒𝑒w𝜓𝜓, with
the second term representing the extra work required to take the electron from the vacuum level
to the surface of the metal.16
As for all chemical equilibria, the chemical potential 𝜇𝜇 for two phases in contact with each other
must be equal i.e. 𝜇𝜇y= 𝜇𝜇F. When two metals are brought into contact with each other, electrons
will flow from one to another to achieve equilibrium, resulting in excess charge at the interface on
the surface of the metals. The difference in potential is known as the contact potential.
The electrochemical potential of electrons in a redox electrolyte is given by the Nernst equation
𝐸𝐸z{"|Y = 𝐸𝐸}z{"|Y +𝑅𝑅𝑇𝑇
𝑛𝑛𝐹𝐹ln [O]
[R] 13
where 𝑛𝑛 is the number of electrons transferred in the half-cell reaction, 𝐸𝐸} is the standard
potential, and [O] and [R] are the concentrations of the oxidised and reduced redox
species.𝐸𝐸z{"|Y = 𝜇𝜇z{"|Y and can be considered as the Fermi level of the redox couple, provided
the same energetic reference point is used. The energy of an electron in a solid-state material is
measured with respect to the vacuum level, whereas the electrochemical potential of a redox
species is usually measured against the standard hydrogen electrode (SHE), which lies at -4.5
eV with respect to the vacuum level. The Fermi level of a redox species can therefore be given by
𝐸𝐸D,z{"|Y = −4.5eV − q𝐸𝐸}
z{"|Y 14
where q is the elementary charge of an electron.
Scheme 4
The equilibration of potential for a platinum electrode in contact with the iron ferricyanide/ferrocyanide redox
couple in solution.
10
When a metal is in contact with a redox electrolyte, electrons will also flow from one phase to
another until an equilibrium in electrochemical potential is reached. Take the example of the iron
ferricyanide/ferrocyanide redox couple that is in contact with a platinum electrode, as illustrated in
Scheme 4. Electrons will be transferred from the highest occupied molecular orbital (HOMO) of
the ferrocyanide species to the metal electrode. This electrostatic charging will raise the Fermi
level of the electrode to reflect the electrochemical potential of the electrolyte redox species.
Scheme 5
The application of (a) a negative overpotential (bias) to the platinum electrode, which allows electron transfer from
the metal to the electrolyte LUMO, or (b) a positive overpotential (bias), which allows electron transfer from the
electrolyte HOMO to the metal.
At equilibrium, there is no net flow of electrons between the electrode and the electrolyte redox
species. If a negative potential is applied to the metal electrode, the Fermi level will increase
above the electrochemical potential of the lowest unoccupied molecular orbital (LUMO) of the
ferricyanide species and electrons will be transferred from the platinum to the LUMO, resulting in
the flow of reduction current, as illustrated in Scheme 5a. Similarly, if a positive potential is
applied, the Fermi level will move below the HOMO of the ferrocyanide and electrons will be
transferred from the HOMO to the platinum, resulting in the flow of oxidation current, as illustrated
in Scheme 5b.
2.2.2 The electrode-electrolyte interface
Scheme 6
The metal-solution interface represented as a capacitor with a (a) negative and a (b) positive charge 𝑞𝑞m on the
metal surface.
11
When a metal is brought in contact with a solution, an excess charge 𝑞𝑞Ü will develop at the metal
surface, and an equivalent but opposite charge 𝑞𝑞á (i.e. 𝑞𝑞Ü = -𝑞𝑞á) will develop at the electrolyte
interface due to the equilibration of potential between the two phases, similar to when two metals
are brought in contact. The charge on the metal represents an excess or depletion of electrons
and will reside in a very shallow region at the surface (< 0.1 Å), while the charge in solution is
comprised of an excess of anions or cations within the vicinity of the electrode surface. The
charged species and the corresponding dipoles present at the metal-electrolyte interface are
called the electrical double layer, and can be regarded somewhat as a capacitor. The charges 𝑞𝑞Ü
and 𝑞𝑞á are often given as charge densities 𝜎𝜎 in µC/cm2, and the related double-layer capacitance
𝐶𝐶" at a given potential is given in µF/cm2. Scheme 7 shows a schematic representation of the
double layer model at the metal-electrolyte interface.
Scheme 7
Schematic representation of the double layer model at the metal-electrolyte interface showing the inner Helmholtz
plane (IHP), which contains solvent molecules and specifically adsorbed ionic and molecular species, and the
outer Helmholtz plane (OHP), which contains solvated ions. The charge density of these layers is indicated by 𝜎𝜎.
Reproduced with permission from 17, copyright John Wiley and Sons 2001.
The solution side of the double layer is itself comprised of different layers, including the inner
Helmholtz layer closest to the interface, which contains solvent molecules or ionic species that
are specifically adsorbed, and is defined by the inner Helmholtz plane (IHP) at a distance x1 from
the interface. The total charge density associated with the species located within this layer is
𝜎𝜎l (µC/cm2). Solvated ions can approach the metal surface up to a distance of x2 from the
interface, which is known as the outer Helmholtz plane (OHP), and these non-specifically
adsorbed solvated ions make up the diffuse layer that extends to the bulk of the solution where
there are no longer any perturbations in the ionic structure. The total charge density associated
12
with the species located within the diffuse layer is 𝜎𝜎", and the total excess charge density of the
solution components of the double layer 𝜎𝜎á is given by:
𝜎𝜎á= 𝜎𝜎l+ 𝜎𝜎"= −𝜎𝜎Ü(15)
The thickness of the diffuse layer depends on the total ionic concentration in the solution.
Concentrations of greater than 10 mM result in a diffuse layer of less than ~100 Å.
Scheme 8
The potential (𝜙𝜙) distribution at the metal-electrolyte interface in the absence of specifically adsorbed ions.
Reproduced with permission from 17, copyright John Wiley and Sons 2001.
Scheme 8 displays the potential distribution at the metal-electrolyte interface. As electroactive
species can only approach the OHP, the potential drop across the double layer (known as the
barrier height ∆𝜙𝜙) can reduce the potential that this species will experience, and can, thus,
influence the rate of redox processes that happen between non-specifically adsorbed
electroactive species and the electrode. Furthermore, when a potential is applied to the metal, the
change in the Fermi level of the metal will change the potential distribution and thus induce a
reconstruction of the double layer, maintaining a neutral charge. The charging current can often
be significant, with currents greater than the faradaic current for redox processes taking place.17
13
Scheme 9
The (a) potential and (b) charge distribution at the semiconductor-electrolyte interface. Reproduced with
permission from 18, copyright John Wiley and Sons 2015.
The charger carrier density in semiconductors is much less than in metals, meaning that when a
semiconductor is brought in contact with a solution, the excess charge induced by the equilibrium
of potential between the two phases will be distributed over a greater distance than in a metal.
For a moderately doped semiconductor, this so-called space charge region can extend 10-1000
nm inside the semiconductor region. The electrochemical potential of electrons near this region is
different than those in the semiconductor bulk, and this phenomenon is similar to the diffuse
double layer formed on the solution side of an electrode in solution. An example of the potential
and charge distribution in the space charge region and diffuse layer is shown in Scheme 9.
As most of the potential drop occurs in the space charge region, rather than at the semiconductor
solution interface (i.e. in the Helmholtz layer), the position of the band at the interface does not
change. Consider when the electrochemical potential of a solution is within the bandgap and lies
below the Fermi level of the semiconductor resulting in electrons being transferred to the solution
phase; the positive charge at the semiconductor surface causes the band energies to become
more negative with increasing distance inside the bulk, where it then remains flat without the
influence of any charges or fields. This phenomenon is known as band bending and is illustrated
for both a n- and p-type semiconductor in Scheme 10. As a result, in the given example, any
excess electron in the space charge region would move towards the bulk of the semiconductor in
the direction that is consistent with the existing electric field, while an excess hole in the bulk
would move towards to the interface. The potential at which no excess charge exists in the
semiconductor is the potential of zero charge. As there are no fields and no space-charge region
under such conditions, the bands themselves are not bent, and this potential is known as the flat-
band potential.17
14
Scheme 10
Band bending taking place at the semiconductor-electrolyte interface for a semiconductor with a Fermi energy EF
in contact with a redox species with an electrochemical potential of 𝜇𝜇{ laying i n-between t he s emiconductor
bandgaps. The situation before and after contact is shown for an n-type semiconductor in (a) and (c),
respectively, and for a p-type semiconductor in (b) and (d). The arrows represent the direction of the electron
flow. Reproduced with permission from 19, copyright Elsevier 2007.
2.2.3 The relation between current and potential
For a non-spontaneous cell-reaction to take place, an overpotential
h
must be applied, which is
the magnitude of potential applied over the equilibrium potential 𝐸𝐸wi.e.
𝜂𝜂 = 𝐸𝐸 − 𝐸𝐸w16
Take the example of the oxygen reduction reaction (ORR) at platinum or carbon; the
overpotential
h
for ORR obtained at carbon is significantly greater than at platinum, despite
having a similar 𝐸𝐸D, the identical oxygen LUMO and, thus, the identical thermodynamic energy
requirement. The extra potential is required to overcome the greater kinetic barrier for ORR at
carbon compared to that at platinum, as illustrated in Scheme 11.
15
Scheme 11
Current-potential curves for the oxygen reduction reaction (ORR) at (a) platinum and (b) carbon electrodes, as
well as band diagrams, illustrating the difference in activation energy required to overcome the kinetic barrier for
ORR at each electrode.
The overall current 𝑖𝑖 for a reaction O + e- D R is the difference between the cathodic and anodic
currents, 𝑖𝑖z{" and 𝑖𝑖|Y, respectively:
𝑖𝑖 = 𝑖𝑖z{" − 𝑖𝑖|Y 17
Each current is proportional to its corresponding heterogeneous rate constant 𝑘𝑘z{" and 𝑘𝑘|Y,
respectively, such that:
𝑖𝑖z{" = 𝐹𝐹𝐴𝐴𝑘𝑘z{" O18
and
𝑖𝑖|Y =𝐹𝐹𝐴𝐴𝑘𝑘|Y R19
where 𝐴𝐴 is the surface area of the electrode, O and [R] are the concentrations of the oxidised or
reduced species at the surface of the electrode, and 𝐹𝐹 is the Faraday constant. The rate
constants 𝑘𝑘z{" and 𝑘𝑘|Y can be written as a function of the standard heterogeneous rate constant
𝑘𝑘|:
𝑘𝑘z{" = 𝑘𝑘|e
Qèê MQMë
ío 20
and
𝑘𝑘|Y = 𝑘𝑘|e
ìQè ê MQMë
ío 21
16
where 𝛼𝛼 is a dimensionless parameter that is a measure of the symmetry of the energetic barrier
for electron transfer. The overall current is thus given by:
𝑖𝑖 = 𝐹𝐹𝐴𝐴𝑘𝑘|O e
Qèê MQMë
ío − R e
ìQè ê MQMë
ío 22
At equilibrium, there is no net flow of current due to the cathodic and anodic currents being equal
and opposite in value such that:
𝑖𝑖w= 𝑖𝑖z{" = 𝑖𝑖|Y 23
𝑖𝑖w is known as the exchange current and can be written in terms of the bulk concentrations of the
oxidised and reduced species:
𝑖𝑖w=𝐹𝐹𝐴𝐴𝑘𝑘|Oïñóò
ìQè Rïñóò
è24
From the equations 22 and 24, the current-potential equation can be described in terms of
overpotential 𝜂𝜂 rather than equilibrium potential:
𝑖𝑖 = 𝑖𝑖w
O
Oïñóò
𝑒𝑒
Qèêô
ío −R
Rïñóò
𝑒𝑒
ìQè êô
ío 25
where 𝜂𝜂 = 𝐸𝐸 − 𝐸𝐸w. The first term in the brackets describes the cathodic component of the current
at a given potential, while the second term similarly describes the anodic component. If there are
no mass-transfer limitations (i.e. the solution is stirred well, or the currents are low), the
concentration of the oxidised and reduced species at the surface and in the bulk will be similar,
such that O /[O]ïñóò and R /[R]ïñóò will be approximately 1 and equation 25 will reduce to the
Butler-Volmer equation:
𝑖𝑖 = 𝑖𝑖w𝑒𝑒
Qèêô
ío − 𝑒𝑒
ìQè êô
ío 26
For large values of 𝜂𝜂, either the cathodic or the anodic component in equation 26 will become
significantly greater than the other back reaction, such that:
𝑖𝑖|Y = 𝑖𝑖w𝑒𝑒
Qèêô
ío 27
𝑖𝑖z{" = −𝑖𝑖w𝑒𝑒
ìQè êô
ío 28
or
17
𝜂𝜂 = 𝑅𝑅𝑇𝑇
𝛼𝛼𝐹𝐹 ln 𝑖𝑖w−𝑅𝑅𝑇𝑇
𝛼𝛼𝐹𝐹 ln 𝑖𝑖|Y 29
𝜂𝜂 = − 𝑅𝑅𝑇𝑇
1 − 𝛼𝛼 𝐹𝐹 ln 𝑖𝑖w−𝑅𝑅𝑇𝑇
1 − 𝛼𝛼 𝐹𝐹 ln 𝑖𝑖z{" 30
Scheme 12
Tafel plots for the anodic and cathodic directions of the reaction O + e- D R.
A plot of log(𝑖𝑖) against 𝜂𝜂, also known as a Tafel plot, will thus allow the exchange current 𝑖𝑖w to be
determined from the y-intercept, as shown in Scheme 12. At low overpotentials (< 50 mV), the
plots deviate from a linear behaviour as the back reaction ceases to be negligible. At large
overpotentials the slope can also deviate due to mass-transfer limitations. The Tafel slope
provides information on the response of the system to the applied potential and is highly
dependent on the mechanism of the electrochemical process taking place. For multistep
electrochemical processes that involve the transfer of several electrons or protons, the Tafel
slope can provide an insight into the rate determining step and the mechanistic aspects of an
electrochemical process.20
2.3 Surface modification of electrode materials
2.3.1 Chemically modified electrodes
A chemically modified electrode (CME) is a conducting or semiconducting material that has been
modified with an adlayer of a monomolecular, multi-molecular, ionic or polymeric film that alters
the electrochemical, optical or other properties of the electrode interface.21 The substrates are
derived from conventional electrode materials, while the adlayers may be very diverse in their
nature and their properties. Modification may impart a number of improvements or properties on
the electrode, such as increased sensitivity or selectivity, chemical and electrochemical stability,
broadened useable potential windows and antifouling properties. Surface characterisation
techniques, including electrochemical, spectroscopic or spectroelectrochemical techniques, have
become essential tools not only to verify the function of CME’s but also provide a basis upon
which improvements or refinements of the modification approach can be made in order to
enhance performance. A schematic representation of a CME is shown in Scheme 13, with an
18
anchoring group that binds the adlayer to the electrode surface, functional groups that impart
particular properties or functions to the CME, and a spacer (such as an alkyl chain) that typically
exists between the anchoring group and the functional groups.
Scheme 13
Schematic representation of a chemically modified electrode.
The surface of electrodes can be modified in different ways, such as through:
1) Physisorption – long range or weak van der Waals attraction between the adsorbate and
the substrate.
2) Chemisorption – the strong adsorption of molecules on the surface through the
spontaneous formation of a chemical bond.
3) Covalent attachment – functional groups on the surface may be modified through the
formation of a new, covalent chemical bond.
4) Polymer layers – attachment via physisorption, chemisorption or physical anchoring e.g.
inside pores at the surface of the electrode, or due to the low solubility of the polymer in
the contact/electrolyte solution.
Modification of oxides can be done using a number of chemistries including silanes,
phosphonates, carboxylates, catechols, acetylacetonates, hydroaxamates, alkenes, alkynes,
amines, thiols, Grignard agents and more.22–28 The following sections will focus on the most well-
established methods for modifying electrode materials.
2.3.2 Langmuir-Blodgett assembly
The Langmuir-Blodgett (L-B) assembly technique was originally developed using amphiphilic
molecules containing a hydrophobic tail and a hydrophilic head, such as surfactants and fatty
acids; however, other systems such as nanoparticles or biological species may also be used.
These species become spontaneously orientated at a liquid/gas interface e.g. when the
19
hydrophilic head group of a sparingly water-soluble species becomes partially solubilised and the
hydrophobic tail extends into the gaseous phase in order to reduce the free energy of the system.
Intermolecular interactions may exist between the orientated molecules. Homogenous films or
films of mixed composition can be prepared. Amphiphiles commonly used to form L-B films
include acids (CnH2n+1COOH), alcohols (CnH2n+1OH), esters (CnH2n+1COOR), amides
(CnH2n+1CONH2), amines (CnH2n+1NH2) and nitriles (CnH2n+1CN).21
Figure 1
An example of a pressure-area isotherm for a monolayer of 1,2-dipalmitoyl-sn-glycero-3-phosphocholine on water
at 25 °C, with the different phases assigned as gaseous, liquid-expanded (LE), the co-existence of LE and liquid-
condensed (LC), LC, and collapsed phases.29 Reproduced from 29 with permission from the PCCP Owner
Societies.
A solution containing the molecules of interest is added to a solvent, such as water, and the
lateral pressure at the gas/liquid interface is increased by using a moveable barrier in a Langmuir
trough. As the molecular film is compressed, it goes through a series of phase transitions, as
exemplified in Figure 1. Once a highly ordered, compressed film is formed, it can be transferred
to the surface of an electrode by withdrawing the electrode substrate transversely out of
(receding deposition) or advancing the substrate into (extending deposition) the liquid layer, while
maintaining a constant pressure by moving the barrier. Receding deposition results in the
hydrophobic tail depositing on the substrate surface, while extending deposition results in the
hydrophilic head groups depositing on the substrate surface. In addition to simple monolayers,
multi-layers and complex architectures can also be achieved using L-B assembly.21 Given that
the film is held in place through van der Waals forces, they suffer from poor thermal and
mechanical resistance and may easily be removed from the surface using certain solvents.
2.3.3 Self-assembly
Modification of a substrate via the spontaneous adsorption of molecules is termed as self-
assembly. The assembly mechanism may take place from the liquid phase or from the vapour
phase. These molecules are typically bound to the surface of the substrate via chemisorption. In
addition to chemisorption, intermolecular interactions within the film also provide a driving force
20
for the film formation and play an important role in the interface structure. When monolayers are
formed as the result of self-assembly, they are known as self-assembled monolayers (SAMs).
The molecules chemisorb onto the substrate surface via a head group and are separated from a
tail group by a spacer, as illustrated in Scheme 14. The spacer is typically an alkyl chain;
however, it can also be a different kind of molecular skeleton. Lateral interactions between the
spacer chain and between tail groups orientate the molecules perpendicular (usually with a
certain tilt angle) to the surface and can potentially facilitate an increase in coverage and
thickness up to packing density limits. Functional groups on the tail end of the SAM may impart
certain functionalities or properties to the electrode.30
Scheme 14
Schematic diagram of a single-crystalline SAM of an alkanethiol adsorbed on a (111) Au surface. Reprinted with
permission from 30, copyright 2005 American Chemical Society.
Commonly used adsorbates in SAMs are thiolates on metals, such as Au or Ag30; however, a
number of different species are used on different substrates, including silane SAMs on oxides31,
carboxylic acid SAMs on oxides or metals such as Ti or Ni30,31, or terminal-alkynes on Si22 or
Au.32
2.3.4 Modification of TCO’s
Chemical modification of oxides have been widely used to control their surface and interface
properties and has found use in applications ranging from organic electronic devices, photovoltaic
cells and biosensors, to devices with supported catalysts.33 In particular, TCO electrodes such as
ITO, TiO2, SnO2, ZnO etc. have been modified in order to enhance heterogeneous electron
transfer processes with solution species that are thermodynamically favoured but inhibited due to
kinetics e.g. hydrophobic solution species.34 Confinement of redox species at the surface of these
materials is meant to allow fast and reversible redox processes and mediation of charge transfer
with solution species. Satisfactory electron transfer rates can be achieved by virtue of the redox
species being in close proximity to electronically active sites on the metal oxide. Different oxides
may differ significantly in terms of their properties, such as their conductivity, chemical or
21
mechanical stability etc.; however, most oxides have a similar propensity for surface modification
as a result of surface hydroxyl groups (-OH), which can act as anchoring points for densely
packed monolayers. A broad range of different chemistries may be used to modify oxides.
Covalent modification can be carried out using silanes or alkenes/alkynes, or alternatively the
surface can be modified through chemisorption of species such as carboxylic acids, phosphonic
acids, thiols or amines.31,34
Scheme 15
Schematic representation of (a) an ideal metal oxide surface and (b) an oxide surface that has been partially
hydrolysed resulting in hydroxyl groups, in addition to possible physisorbed metal hydroxide species.
Scheme 15a is a representation of an ideal oxide surface. Hydrolysis of bridging oxygen species
at the surface can result in the formation of hydroxyl species, while extensive hydrolysis can lead
to the formation of fully hydrolysed, metal hydroxide species that may in turn physisorb on the
surface, as shown in Scheme 15b. These hydroxide species, as well as variations in electron-rich
defect or dopant sites, lead to heterogeneity in the electron transfer rates or chemical
compatibility of oxide surfaces.34 As such, electrical ‘hot spots’ and ‘dead spots’ may exist in
varying degrees on oxide surfaces, such as on ITO.35 Indium oxide has a favourable equilibrium
constant for hydrolysis (K = 4.3 x 101), while the fully hydrolysed indium hydroxide species
In(OH)3 has a very low solubility product (Ksp = 1.3 x 10-37). On the other hand, the hydrolysis of
tin oxide is significantly less favourable (K = 4.3 x 10-10), while the fully hydrolysed Sn(OH)2 or
Sn(OH)4 species have a higher solubility than In(OH)3.34 Such variations in the surface state of
different oxides that have been synthesised or pre-treated using different methods should be
taken into consideration when using chemical modification methods.
2.33..44..11
SSiillaanneess
Silanes attach to oxide surfaces through metal-oxygen-silicon bonds that form on reaction with
surface hydroxyl groups, with the removal of one or more leaving group. Suitable leaving groups
on the silane include chloro or alkoxy groups (e.g. methoxy or ethoxy groups), and each silane
modifier can form multiple bonds with the surface, as shown in Scheme 16. The functional group
22
(R) may be diverse in nature and surface modification may be achieved in solution or in the
vapour phase. Different mechanisms can take place depending on the reaction conditions and
particularly the humidity or water content of the reaction media. Condensation of neighbouring
species during or after the modification process may lead to cross-linking, which can yield
hydrolytically stable, polymeric monolayers on the oxide surface; however, this may also result in
disordered multilayers due to the tendency of the silanes to extensively hydrolyse. Such lack of
control can be problematic when attempting to attaching redox species to oxide surfaces.31,34
While silanes suffer from reproducibility issues and limited hydrolytic stabilities, caged
trialkoxysilanes called silatranes have been shown to result in interfaces on that are more stable
than carboxylate or phosphonate derived interfaces on oxides n TiO2.25,36
Scheme 16
Schematic representation of the different possible binding motifs of silanes on an oxide surface, including the
cross-linking of neighbouring species. The silane molecule depicted on the left contains methoxy leaving groups,
however different leaving groups are possible.
2.33..44..22
CCaarrbbooxxyylliiccaacciiddss
Carboxylic acids chemisorb on oxides thanks to their amphoteric nature. Different
possible binding motifs are shown in Scheme 17; however, this scheme is not exhaustive.
Carboxylic acids may bind through ester bonds with the metal, through the coordination of the
carboxylate oxygen to two metal centres via a bridging motif or chelation to one metal
centre, or simply through hydrogen bonding between the acid and the surface hydroxyl
groups. Carboxylic acid SAMs can be deposited on oxides using Langmuir-Blodgett techniques
or using solution or gas-
phase processes. Annealing often results in improved interfaces. Carboxylic acids have
been widely used in a number of applications, including as anchors for sensitising
semiconducting
oxides with dyes, thereby giving rise to dye-sensitised solar cells (DSSCs).34 While
the modification process with carboxylic acids is straightforward and self-limiting, the
resulting interfaces are not very stable, particularly at non-acidic pH’s. In fact, carboxylic acid
SAMs exhibit stabilities that are around an order of magnitude less than for phosphonate or
silane based
interfaces.37,38
23
Scheme 17
Schematic representation of different possible binding motifs of carboxylic acids on an oxide surface, including (a)
hydrogen bonding, (b) ester bonds to a metal centre, and (c) bridging chelation of two metal centres by the
carboxylate ion.
2.33..44..33
PPhhoosspphhoonniiccaacciiddss//pphhoosspphhoonnaatteess
Scheme 18
Schematic representation of the binding mechanism of phosphonic acids to a Lewis acidic metal oxide, with
mono-, bi- and tri-dentate binding modes due to differing degrees of heterocondensation with neighbouring
hydroxyl groups.
Phosphonic acids and their ester derivatives are known to chemisorb strongly onto oxide
surfaces. Both dissolved phosphonic acids and phosphonate esters can self-assemble on oxide
surfaces and the outcome can be highly dependent on the type of solvent used, as well as its
dielectric constant, the pH, and temperature. Weak van der Waals or hydrogen bonding
interactions can often be superseded by inducing the formation of stronger covalent bonds
through annealing. On Lewis acidic oxides, the phosphoryl oxygen (P=O) coordinates to a Lewis
acidic site on the surface, which leads to the P atom becoming more electrophilic and inducing
heterocondensation with neighbouring hydroxyl groups, as illustrated in Scheme 18. On oxides
lacking Lewis acidity, initial hydrogen bonding interactions with the surface induce
heterocondensation reactions, and this can also be enhanced by annealing. The P-O-M bonds
formed are strong and a large number of different binding motifs are possible. Phosphonic acids
and esters tend to form monolayers and, due to the stability of the P-O-M bond, are more
resistant to hydrolysis than silane-derived interfaces; however, this only applies in pH’s below
around 5 to 6, thus limiting their use at basic pH’s.25,38,39
2.3.5 Electrochemical modification
Electrochemical modification (or electrografting) allows the surfaces of conductive materials to be
modified with organic layers by transferring electrons between the electrode and the modifying
reagent, or by using a reducing or oxidising agent. A large number of chemistries can be used on
24
a broad number of materials, ranging from carbon to metals, silicon and oxides.40 Examples of
oxidative electrochemical modifications include the use of amines on glassy carbon (GC), Au or
Pt,41 or carboxylates and alcohols on carbon materials.40 Reductive electrochemical modifications
include the use of vinylic species, which has been carried out on a number of metal, carbon,
silicon and semiconductor substrates.42 However, the most well-known and studied reductive
electrochemical reduction chemistry is that of diazonium salts.
2.3.6 Modification using diazonium salts
Aryl diazonium salts are a group of compounds Ar-N2+X- sharing a diazonium N2+ functional
group, where X- is an organic or inorganic counter anion, as shown in Scheme 19. Their
existence has been known since their first use in the synthesis of organic dyes in 1858.
Diazonium salts react with activated aromatic molecules via an electrophilic aromatic substitution
reaction to form azo compounds, a reaction known as azo coupling, which is still hugely important
in the synthesis of dyes. The diazonium group also activates the aryl ring to nucleophilic aromatic
substitution and can be displaced by a broad number of species, including halogens, CN, NO2,
OH, SH etc.
Scheme 19
The structure of a diazonium salt.
In 1992 Pinson and co-workers described the reaction mechanism of the modification of carbon
electrodes by diazonium salts,43 which led to a number of publications being published on the
modification of surfaces with diazonium salts in the proceeding decades. They have found use in
a wide variety of applications, including chemo- and biosensors, anti-corrosion coatings, and
molecular electronics, to name but a few, and their use has been widely reviewed.40,44,45 Their
synthesis from the corresponding aromatic amines (anilines) is straightforward and can be carried
out via diazotisation in acidic aqueous media in the presence of sodium nitrite (NaNO2), or in
aprotic media in the presence of tert-butyl nitrite ((CH₃)₃CONO) or nitrosonium tetrafluoroborate
NOBF4. Aqueous solutions of diazonium salts are generally unstable above 5 oC and thus must
be used in situ in the solution it was prepared in. The use of a stabilising anion, such as the
tetrafluoroborate anion (BF4-), can lead to enhanced thermal stability and allows the salts to be
isolated and used at room temperatures. The functional group (R) does not necessarily have to
be in the para-position of the aryl ring, and the ring itself may have more than one substituent.
A broad number of materials have been modified using diazonium salts, including electrodes
such as carbon (including GC, highly oriented pyrolytic graphite46, pyrolysed photoresist films47,
25
mesoporous carbon48), metals (including noble metals, such as Au49, Pt50 and Ag51, as well as
base metals, such as Cu52, Fe41, Ni41, and Zn41), semiconductors (such as Si, gallium/indium
arsenide53, or boron-doped diamond54) and oxides (including ITO55, tin oxide56, fluorine-doped tin
oxide57 and TiO258), as well as other substrates such as organic materials (including
polyethylene59, polypropylene59, Teflon60), nitrides59 and carbides59. Diazonium salts have also
been used to modify particles in solution, including metallic or oxide nanoparticles, carbon
nanotubes (CNTs) and nanodiamonds.61 The reaction of the diazonium salt with a substrate or
particle can be a spontaneous chemical or electrochemical (i.e. at open circuit potentials)
reaction, or induced e.g. by external application of electrochemical potential, UV light,
microwaves, ultrasound or mechanical grafting. Most studies and applications are carried out
using carbon or metallic materials, and these are mainly modified by applying an electrochemical
potential or by spontaneous processes.
Scheme 20
An overview of the different mechanisms by which surfaces can be modified by diazonium salts.
An overview of the different mechanism by which a surface can be modified by diazonium salts is
shown in Scheme 20. The surface can either be modified by the adsorption of the diazonium
cation, or by the adsorption of a diazonium decomposition (dediazoniation) product.
Dediazoniation can be heterolytic, whereby an aryl cation is formed, or homolytic, whereby an
aryl radical is formed. Heterolytic decomposition can be thermally induced, while homolytic
decomposition can be induced by electrochemical reduction at an electrode or by a reducing
agent, or alternatively by photolysis. The dediazoniation products can be highly reactive and form
polymeric, branched structures as a result of addition at the meta-position of the aryl ring of
already-grafted species. Furthermore, the diazonium cation can also add at the meta-position
resulting in azo-linkages within the layer. The nature of the resulting film is highly dependent on
the conditions used, as well as the diazonium salt itself and the nature of the substrate. Films can
range from a monolayer thickness up to thicknesses of 100’s of nm’s.
26
2.3.7 Electrochemical grafting using diazonium salts
Electrochemical modification using diazonium salts is usually carried out in acetonitrile; however,
it can also be carried out in other aprotic solvents, aqueous acidic solutions or ionic liquids. As
previously mentioned, the diazonium salt undergoes homolytic decomposition upon reduction and
an aryl radical is formed (illustrated again in Scheme 21 for clarity). When a potential is applied at
an electrode, the reaction becomes self-limiting as the surface of the electrode is passivated and
the charge transfer resistance of the interface inhibits the reduction of further diazonium species.
This can be seen in cyclic voltammograms or chronoamperometry as a rapid decrease in
cathodic current density, as demonstrated in Figure 2.
Scheme 21
The electrochemical reduction of a diazonium salt at an electrode.
Figure 2
The first and second cyclic voltammograms of 1 mM 4-benzamidobenzenediazonium tetrafluoroborate in
acetonitrile (0.1 M TBAF) using a gold working electrode. Scan rate = 50 mV/s.
The reduction potential of a diazonium salt can be correlated to the Hammett constant of its
substituent.62 Comparing the reduction potentials (Table 1) of the diazonium salts used in this
study (Scheme 22) using the same planar ITO working electrode, it is possible to see that the
reduction potentials span a wide potential range from -0.92 to -0.25 V (vs Fc/Fc+), and that the
order the potentials do indeed correlate to the expected electronic inductive effects of each
respective diazonium salt.
27
Scheme 22
The diazonium salts used in this study in order of increasing reduction potential, with the most electron
withdrawing substituents on the right.
Table 1
The electrochemical reduction potentials of the diazonium salts used in this study in acetonitrile, determined using
cyclic voltammetry with a planar ITO working electrode.
Diazonium salt
Reduction potential (V vs Fc/Fc+)
4-nitrobenzenediazonium
tetrafluoroborate (4-NBD)
-0.25
4-(1H-imidazol-1-
yl)benzenediazonium
tetrafluoroborate (Im-diazo)
-0.56
4-([2,2’-Bipyridine]-4-
carboxamido)benzenediazonium
tetrafluoroborate (bpy-diazo)
-0.64
4-benzamidobenzenediazonium
tetrafluoroborate (4-BABD)
-0.92
2.3.8 Electrochemical response of adsorbed monolayers
The electrochemical response of redox species for the reaction O + ne- D R can be heavily
influenced by the adsorption of O or R on the surface of the electrode. Assuming negligible
current from dissolved O or R species, and that the coverages of the redox species ΓX and Γú are
independent of potential, i.e. there is no electrochemically-induced desorption/deactivation or
adsorption/activation, and that all of the reduced adsorbed-O become adsorbed-R, i.e.
−𝜕𝜕ΓX𝑡𝑡
𝜕𝜕𝑡𝑡 =𝜕𝜕Γú𝑡𝑡
𝜕𝜕𝑡𝑡 =𝑖𝑖
𝑛𝑛𝐹𝐹𝐴𝐴 31
then the current-potential equation can be given by:
𝑖𝑖 = 𝑛𝑛9𝐹𝐹9
𝑅𝑅𝑇𝑇
𝜐𝜐𝐴𝐴Γ∗𝑏𝑏X
𝑏𝑏úexp 𝑛𝑛𝐹𝐹
𝑅𝑅𝑇𝑇𝐸𝐸 − 𝐸𝐸|
1 + 𝑏𝑏X
𝑏𝑏úexp 𝑛𝑛𝐹𝐹
𝑅𝑅𝑇𝑇𝐸𝐸 − 𝐸𝐸|
932
28
where 𝜐𝜐 is the scan rate, 𝑏𝑏X and 𝑏𝑏ú are parameters related to the Gibbs free energies of
adsorption of O and R, and Γ∗ is the total coverage of species (i.e. ΓX+ Γú at a given time t). A
current-potential curve for such a system is shown in Scheme 23.
Scheme 23
A cyclic voltammetry curve for reduction and oxidation of the adsorbed species O showing the peak current and
full width at half maximum (FWHM). Reproduced with permission from 17, copyright John Wiley and Sons 2001.
The peak current 𝑖𝑖° is given by:
𝑖𝑖 = 𝑛𝑛9𝐹𝐹9
4𝑅𝑅𝑇𝑇 𝜐𝜐𝐴𝐴Γ∗32
and the peak potential 𝐸𝐸°is given by:
𝐸𝐸°= 𝐸𝐸|−𝑅𝑅𝑇𝑇
𝑛𝑛𝐹𝐹ln 𝑏𝑏X
𝑏𝑏ú
33
The peak current is proportional to the scan rate 𝜐𝜐, whereas for diffusing species it is proportional
to the square root of the scan rate 𝜐𝜐ì/9. Once corrected for any residual current, such as charging
current, the area beneath the reduction wave represents the total charge 𝑄𝑄 required to fully
reduce the adsorbed layer and can thus be used to calculate the total coverage:
29
Γ∗=𝑄𝑄
𝑛𝑛𝐹𝐹𝐴𝐴 34
For species whose adsorption is described by the Langmuir isotherm and follows ideal Nernstian
behaviour, the cathodic peak is equal to the anodic peak (𝐸𝐸°£=𝐸𝐸°§) and the full width at half
maximum (FWHM) for a cathodic or anodic wave of is given by:
∆𝐸𝐸°D•¶J = 3.53 𝑅𝑅𝑇𝑇
𝑛𝑛𝐹𝐹=90.6
𝑛𝑛mV at25℃ 35
The degree to which the peak position 𝐸𝐸° w ill d eviate f rom 𝐸𝐸| wi ll de pend on th e re lative
adsorption strength of O and R. If 𝑏𝑏X = 𝑏𝑏ú then 𝐸𝐸° = 𝐸𝐸|. If 𝑏𝑏X > 𝑏𝑏ú then 𝐸𝐸° will sh ift to more
negative potentials, while if 𝑏𝑏ú > 𝑏𝑏X then 𝐸𝐸° wil l shi ft to mor e pos itive pot entials. Lat eral
interactions between the redox species in the layer can also have an effect on the curve shape.
Attractive interactions will lead to a reduction in ∆𝐸𝐸°D•¶J, while repulsive interactions will lead to
an increase in ∆𝐸𝐸°D•¶J. Experimentally, modified electrodes deviate from such ideal behaviour
due to a number of complex factors, such as limited mass and charge transport, structural
changes upon reduction or oxidation, and inhomogeneity of the electrode and the adsorbed
species.17
2.3.9 Immobilisation of electrocatalysts on electrode surfaces
The immobilisation, or heterogenisation, of enzymatic and molecular catalysts on electrode
surfaces is imperative, thereby increasing the number of addressable active sites, separating
half-reactions and facilitating electronic contact, especially when using hydrophobic catalysts in
aqueous conditions. A number of different approaches can be used to immobilise electrocatalysts
on electrode surfaces taking advantage of many of the aforementioned chemistries. For the most
part, molecular (homogenous) electrocatalysts used in energy conversion and storage reactions
contain one or more metal centres coordinated by organic ligands5,63–66 (although some examples
of metal-free molecular systems do exist67,68).
The immobilisation of electrocatalysts on surfaces can be broadly arranged into three categories:
i) the adsorption or covalent attachment of an electrocatalyst on to a surface via an
anchoring group or reactive group on one or more of its ligands,
ii) the formation of covalent bonds between the electrocatalyst and reactive groups on
the surface, or
iii) the in situ assembly of an electrocatalyst using ligands adsorbed or covalently
attached to the surface.
30
Scheme 24
Schematic representation of three different examples where electrocatalysts have been immobilised on different
electrode surfaces via an anchoring group or a reactive group on one or more of their ligands: (a) a manganese
carbonyl bipyridyl CO2 reduction catalyst immobilised on a carbon nanotube (CNT) electrode via a pyrene
anchoring group69, (b) a cobalt cobaloxime catalyst hydrogen evolution reaction (HER) catalyst immobilised on a
mesoporous ITO electrode via phsophonic acid anchoring groups70, and (c) an iridium bipyridyl oxygen evolution
reaction (OER) catalyst immobilised on a glassy carbon (GC) electrode via the electrochemical grating of a
diazonium salt71.
Several examples of electrocatalysts immobilised on different electrode surfaces via an anchoring
group or reactive group on one or more of their ligands are displayed in Scheme 24. Scheme 24a
shows an example by Reisner and co-workers of a manganese carbonyl bipyridyl CO2 reduction
catalyst immobilised on a carbon nanotube (CNT) electrode using (non-covalent) p-p interactions
via a pyrene anchoring group attached to the bipyridine ligand.69 Meanwhile, Scheme 24b shows
another example by Reisner and co-workers of a cobalt cobaloxime hydrogen evolution reaction
(HER) catalyst immobilised on a mesoporous ITO electrode via chemisorption of phosphonic acid
anchoring groups70. Finally, Scheme 24c shows an example by Hinds, Lin and co-workers of an
iridium bipyridyl oxygen evolution reaction (OER) catalyst immobilised on a glassy carbon (GC)
electrode via the electrochemical grating of a diazonium group located on the bipyridine ligand.71
Scheme 25
Schematic representation of two different examples, where electrocatalysts have been immobilised via the
formation of covalent bonds between the electrocatalysts and reactive groups on the electrode surfaces: (a) a
nickel bisdiphosphine–based mimic of the active site of hydrogenase that catalyses the hydrogen oxidation
reaction (HOR) and hydrogen evolution reaction (HER) immobilised on a multi-walled carbon nanotube (MWCNT)
electrode via amide coupling to surface amino-groups72, and (b) a ruthenium oxygen evolution reaction (OER)
catalyst immobilised on a glassy carbon (GC) electrode via azide-alkyne cycloaddition (or click reaction).73
31
Scheme 25 displays examples of electrocatalysts immobilised on different electrode surfaces via
the formation of covalent bonds between the electrocatalysts and reactive groups on the
electrode surfaces. Scheme 25a shows an example by Fontecave and co-workers of a nickel
bisdiphosphine–based mimic of the active site of hydrogenase that catalyses the hydrogen
oxidation reaction (HOR) and hydrogen evolution reaction (HER) immobilised on a multi-walled
carbon nanotube (MWCNT) electrode via amide coupling between a carboxylic acid group on one
of the catalysts ligands and amino groups grafted to the surface of the MWCNT.72 Meanwhile,
Scheme 25b shows an example by Sun and co-workers of a [RuII(pdc)(pic)3] (pdc = 2,6-pyridine-
dicarboxylate, pic = 4-picoline) OER catalyst immobilised on a glassy carbon (GC) electrode via
copper-catalysed azide-alkyne cycloaddition (also known as a ‘click’ reaction) between an alkyne
group located on a ligand of the catalyst and surface grafted azide groups.73
Scheme 26
Schematic representation of two different examples of the in situ assembly of electrocatalysts on electrode
surfaces using ligands adsorbed or covalently attached to the surface: (a) a copper phenolato oxygen reduction
reaction (ORR) catalyst formed by incubating ligands that were immobilised on a multi-walled carbon nanotube
(MWCNT) electrode via a pyrene anchoring group with CuCl274, and (b) a bifunctional cobalt terpyridyl hydrogen
evolution reaction (HER) and CO2 reduction catalyst formed by incubating terpyridine ligands that were
immobilised on a glassy carbon (GC) electrode via the electrochemical grafting of a diazonium salt with CoCl275.
Finally, Scheme 26 displays examples of electrocatalysts immobilised on electrode surfaces via
in situ assembly using ligands adsorbed or covalently attached to the surface. Scheme 26a
shows an example by Thomas, Le Goff and co-workers of a copper phenolato oxygen reduction
reaction (ORR) catalyst formed by incubating ligands that were immobilised on a MWCNT
electrode via a pyrene anchoring group with CuCl2.74 In a similar manner, Scheme 26b shows an
example by Fontecave and co-workers of a cobalt terpyridyl HER and CO2 reduction catalyst
formed by incubating terpyridine ligands that were immobilised on a GC electrode via the
electrochemical grafting of a diazonium salt with CoCl2.75
Some of these aforementioned approaches can also be used to immobilise macromolecular
catalysts, such as enzymes, onto electrode surfaces, e.g. the grafting of diazonium-modified
horseradish peroxidase onto gold via electrochemical grafting76, or the commonly used approach
GCGC
i) Electrochemical
grafting
ii) CoCl
2
(metallation)
M = Co
i) Pyrene adsorption
ii) CuCl
2
(metallation)
(a) (b)
32
of forming of covalent bonds between amino-acids on the surface of proteins and reactive groups
on the electrode surface. A number of coupling agents may be used for the latter approach,
including glutaraldehyde77, N-Hydroxy-succinimide (NHS)78 and N-ethyl-N'-(3-
dimethylaminopropyl) carbodiimide (EDC) hydrochloride79 agents. Otherwise, immobilisation also
can be achieved using non-covalent adsorption driven by hydrophobic forces, electrostatic
interactions, hydrogen bonding, van der Waals forces or affinity binding. Physical entrapment or
encapsulation may be used as well. In depth descriptions of all of these approaches can be found
elsewhere.3,80,81 Enzyme immobilisation using electrostatic binding interactions will briefly be
discussed further as this approach is applied in this work.
Ionisable functional groups on amino acids on the surface of proteins may be charged under
certain pH’s, depending on their pK values.82 The coincident presence of electrostatic groups on
the surface of an electrode can lead to electrostatic interactions between the electrode and
individual amino acids on the enzymes’ surface, or the enzyme as whole (i.e. if there is an overall
polar charge distribution). This attraction and subsequent contact between the enzyme and the
surface can lower the free energy of the system, and, in the case of an overall polar charge
distribution across the enzyme, it can lead to an immobilisation of the enzyme with a preferential
orientation, as illustrated in Scheme 27.81,83 Interactions may be controlled by changing the pH or
ionic strength of the solution, or the surface of the electrode. Certain functional groups may be
introduced on the surface of the electrode to induce preferential orientation of enzymes, as
demonstrated for a hydrogenase immobilised on SAM-modified gold electrodes by Heidary et
al.84,85
Scheme 27
A schematic representation of an enzyme immobilised on an amino-functionalised electrode surface with a
preferential orientation due to the electrostatic interaction between the overall positively charged surface (due to
partial protonation of the amino functional groups) and the enzyme (which exhibits an overall polar charge
distribution).
33
2.4 Spectroscopic methods
2.4.1 Spectroelectrochemistry
Spectroelectrochemistry (SEC) encompasses a whole range of different spectroscopic
techniques that can be used to simultaneously obtain spectroscopic as well as electrochemical
data for an electrochemical reaction inside an adequate cell. In situ techniques involve the study
of a system while the electrode is immersed in the electrolyte of an electrochemical cell, and
allows systems to thus be spectroscopically studied under potential control. This includes: UV-vis
spectroscopy, infrared (IR) spectroscopy, Raman spectroscopy, X-ray adsorption spectroscopy
(XAS), electron paramagnetic spectroscopy (EPR), luminescence spectroscopy and ellipsometry.
Ex situ techniques involve the removal of the electrode from the electrolyte and allow
spectroscopic studies to be carried out in air or under vacuum. This includes: X-ray photoelectron
spectroscopy (XPS), low energy electron diffraction (LEED) and others. Removal of the electrode
from the electrolyte or exposure to vacuum may change the nature of the electrode surface
significantly.
The electrochemical cell setup is most often specific to each technique, and as such there is no
standard setup. For example, in situ UV-Vis SEC requires the use of an optically transparent
electrode (OTE). Light is transmitted through the electrode surface and the electrolyte, and the
absorbance changes are measured using the detector of the spectrometer. In addition to
transmission, measurements can also be done using (internal) reflection spectroscopy. The OTE
in such systems may be a TCO, such as ITO or TiO2, or indeed any other conductor, provided it
is sufficiently thin. Metal thin films such as Au or Pt are used, as well as fine wire meshes or
minigrids that contain regions of alternating transparency in the cross-section of the light
beam.86,87 The diffusion layer for redox species in such systems can become larger than the
openings in the minigrids. In general, species being probed by in situ SEC methods may be
dissolved in the bulk of the solution or the diffusion layer, or they may be specifically adsorbed at
the surface of an electrode. Certain techniques may be more sensitive to one of these, such as
the use of UV-Vis or IR reflection SEC, which can be more sensitive to species closer to the
surface of the working electrode. A number of different electrochemical techniques can in
principle be employed in parallel, including voltammetry (cyclic voltammetry (CV), linear sweep
voltammetry (LSV), chronoamperometry etc.), chronocoulometry, chronopotentiometry and step
techniques.
2.4.2 UV-Vis spectroscopy
Absorption of UV and visible light involves the excitation of electrons in atoms and molecules
from ground states to excited states. In chromophores, incident radiation can excite electrons
from bonding (s or p) orbitals or non-bonding orbitals, to anti-bonding (s* or p*) orbitals. In the
34
case of transition metal complexes, incident light can excite electrons from the ligand molecular
orbitals to the metal d-orbitals, if the ligand orbitals are all occupied. Likewise, incident light can
excite an electron from metal d-orbital to a low-lying empty ligand molecular orbital. Such ligand-
to-metal or metal-to-ligand charge-transfer electronic transitions result in intense adsorption
bands. Adsorption may also take place due to excitation of electrons from metal d-orbitals to d-
orbitals of higher energy.
In UV-Vis spectroscopy, a sample is irradiated with light of continuously varying wavelengths and
the adsorption of light at a given wavelength 𝜆𝜆 is given by the absorbance 𝐴𝐴(𝜆𝜆):
𝐴𝐴𝜆𝜆 = log 𝐼𝐼w
𝐼𝐼36
where 𝐼𝐼w is the intensity of the UV-Vis light source and 𝐼𝐼 is the intensity of the detected light after
passing through the sample. For dilute solutions, the adsorption follows the Beer-Lambert Law:
𝐴𝐴𝜆𝜆 = log 𝐼𝐼w
𝐼𝐼= 𝜀𝜀 𝜆𝜆 𝑙𝑙𝑙𝑙 37
where 𝜀𝜀 is the molar absorption coefficient, 𝑙𝑙 is the concentration of the absorbing species and 𝑙𝑙
is the (optical) path length of the absorbing solution.
Scheme 28
Potential energy diagrams showing vertical transitions between electronic states for a (a) 0-0 transition and a (c)
0-2 transition to a vibrationally excited state. The corresponding absorption spectra for the (b) 0-0 transition and
(d) 0-2 transition with the dotted lines representing the absorption lines observed for a vapour, and the broadened
solid spectra representing absorption spectra for a species in solution. Adapted with permission 88, copyright
Wiley 2001.
35
The motion of electrons is much more rapid than those of nuclei. The excitation of an electron to
an empty orbital takes about 10-15 s, whereas the characteristic time for a molecular vibration is
around 10-10 – 10-12 s. Therefore, the Franck-Condon principle states that electronic transitions
are most likely to occur with no change in the position of the nuclei of the molecule and its
environment. The resulting state is called a Franck-Condon state and the transition involved is
called a vertical transition, as illustrated in Scheme 28. At room temperature, most of the
molecules are in the vibrational ground state and so in addition to the purely electronic 0-0
transition shown, other vibrational transitions can take place whose intensities depend on the
relative position and shape of the potential energy curves.88
Usually, broad absorption bands are observed in UV-Vis spectroscopy and this can be attributed
to homogenous and inhomogeneous broadening. Homogenous broadening results from the
existence of a continuous set of vibrational sublevels in each electronic state, whereas
inhomogeneous broadening results from interactions between molecules in solution and the
change in the solvation shell of a chromophore.
2.4.3 Vibrational spectroscopy
Vibrational spectroscopy is an invaluable, energy sensitive probe of molecular structures. It
depends on periodic changes in the dipole moments (in infrared (IR) spectroscopy) or
polarisabilities (in Raman spectroscopy) caused by molecular vibrations, and changes in
frequencies during adsorption (in IR spectroscopy) or scattering (in Raman spectroscopy) of
electromagnetic radiation. Each atom in a molecule has 3 degrees of freedom, resulting in 3N
degrees of freedom for N atoms. Since 3 degrees of freedom are translational and 3 are
rotational, 3N-6 are fundamental vibrations for non-linear molecules (linear molecules have 3N-5
vibrations since rotation about the bond axis is not possible). These vibrations are normal modes,
meaning all atoms vibrate with the same frequency and different amplitudes, while the centre of
mass does not change. An expression can be derived from Hooke’s law that gives the
wavenumber of a vibration:
𝜐𝜐=1
2𝜋𝜋
𝑘𝑘
𝜇𝜇(38)
where 𝑘𝑘 is the force constant of the spring and 𝜇𝜇 is the reduced mass ≤≥≤¥
≤≥µ≤¥
of the diatomic
molecules. Higher frequency vibrations are indicative of a stronger bond or a lower mass of the
molecule. The potential energy of a vibration can be determined using the harmonic oscillator
approximation. Solving the Schrödinger equation for the harmonic oscillator results in:
36
𝐸𝐸∂lï = 𝑛𝑛 + 1
2ℎ𝑣𝑣 (39)
where ℎ is Planck’s constant and 𝑛𝑛 is the vibrational quantum number (𝑛𝑛 = 0, 1, 2 …). Molecular
vibrations are therefore quantised with equal distances between each energy level. Although the
harmonic oscillator is a good approximation to describe molecular vibrations, they are essentially
anharmonic.
2.4.4 Infrared spectroscopy
IR spectroscopy is widely used to characterise organic, inorganic and biological species, for
which the mid-IR is useful (wavelengths from 2.5 to 25 μM, which correspond to wavenumbers of
4000 to 400 cm-1). For a molecule to absorb IR radiation and undergo a transition to a vibrational
excited state, the electric field of the incident radiation must be oscillating at the same frequency
as a vibration in the molecule (i.e. resonant absorption). Another condition is that the molecular
dipole moment must change during the vibrational oscillation. The transition dipole moment
couples the total molecular wave function of the initial and final vibrational states of the molecule,
Ψ∫ and Ψª, via the molecular dipole moment operator, 𝜇𝜇:
𝜇𝜇∫ª = Ψ∫𝜇𝜇 Ψª(40)
The transition dipole moment must therefore be non-zero (𝜇𝜇∫ª ≠ 0) in order for a transition to
occur. For a polyatomic molecule, the dipole moment is a vector quantity with three Cartesian
components, and is the sum of the charges on each 𝛼𝛼th atom 𝑒𝑒è and its distance from the centre
of mass of the molecule 𝑞𝑞è:
𝜇𝜇 = 𝑒𝑒è𝑞𝑞è
è
(41)
For a transition to occur, the change in the dipole moment for a vibration must have a component
in the same direction as the oscillation of the incident electric field. The probability 𝐴𝐴 of a
transition occurring is proportional to the square of 𝜇𝜇∫ª, and for randomly orientated molecules to
absorb unpolarised light, the probability is thus proportional to the sum of the transition
probabilities along the three components:
𝐴𝐴 ∝ 𝜇𝜇5ª∫
9+ 𝜇𝜇æª∫
9+ 𝜇𝜇øª∫
9(42)
Molecular vibrations of polar molecules generally correspond to the strongest infrared absorption
bands. Molecular vibrations that don’t alter the dipole moment, and which are thus IR silent,
37
include homonuclear diatomic molecules, or the symmetric O=C=O stretch of CO2. As can be
inferred from equation 38, the position of a vibration is related to the mass of the atoms involved,
with lighter atoms vibrating at higher frequencies, as well as the strength of the bond (e.g. an sp,
sp2 or sp3 bond). So-called ’group frequencies’, which are characteristic of certain functional
groups (e.g. C=O, -CH3 etc.), may be used to assist in the interpretation of IR spectra.
Scheme 29
Schematic representation of attenuated total reflection (ATR) of incident IR radiation in an optical prism at an
interface to an optically less dense medium. The incoming, unpolarised IR beam is defined by both perpendicular
(s) and parallel (p) components relative to the plane of incidence.
Studying electrode-electrolyte interfaces using IR SEC can be complicated by the absorption of
IR radiation by the bulk solution (especially when water is used, which adsorbs strongly
throughout most of the mid-IR region). This can be overcome by using an attenuated total
reflection (ATR) setup for IR studies. Total internal reflection of radiation takes place at an
interface to an optically less dense medium when the angle of incidence is higher than a critical
angle 𝜃𝜃£, as illustrated in Scheme 29. The electric field of the reflected radiation penetrates the
less dense medium without carrying any energy, and thus its intensity decays exponentially in the
direction normal to the plane of incidence:
𝐸𝐸 = 𝐸𝐸wexp −𝑧𝑧
𝑑𝑑°
38
where 𝑑𝑑° is the penetration depth at which the amplitude of the field has decayed to ca. 37%
(~expQì) of its initial value. The penetration depth of this evanescent wave is a function of the
angle of incidence 𝜃𝜃, the wavelength of the radiation 𝜆𝜆 and the ratio of the refractive indices of the
optically more dense and less dense media, 𝑛𝑛ì𝑎𝑎𝑛𝑛𝑑𝑑𝑛𝑛9, respectively:
𝑑𝑑°=𝜆𝜆
2𝜋𝜋 sin9𝜃𝜃 − 𝑛𝑛ì
𝑛𝑛w
939
38
An internal reflection element (IRE) or an ATR prism is coated with an electrode layer that is in
contact with the electrolyte in a spectroelectrochemical cell. Typical materials used as IREs
include silicon (𝑛𝑛hl = 3.42)89, germanium (𝑛𝑛ƒ{ = 4.0)89 or zinc selenide (𝑛𝑛≈dh{ = 2.43)90. For a Si
prism, an angle of incidence of 60 o (as used in this work) results in a penetration depth of
between 2.6 and 0.7 μm for the spectral region of 1000 to 4000 cm-1. To carry out SEC, the
electrode material must be conductive and transparent for the IR beam. These conditions were
fulfilled in this study either by using a very thin film (see below) of Au, or films of the transparent
conductive oxides ITO (𝑛𝑛c∆X ~ 1.5 in near-IR)91 or ATO (𝑛𝑛y∆X ~ 2 – 4.5 in mid-IR)92. The ATO
used in this study is highly porous, meaning that the refractive index will be reduced by the
presence of solvent within the pores.
2.4.5 Surface Enhanced Infrared Absorption (SEIRA) spectroscopy
The optical properties of molecules can be dramatically altered when adsorbed on rough metal
surfaces, islands or particles. This occurs in surface enhanced Raman scattering (SERS), which
was first observed in 197393, where enhancement in the Raman scattering of molecules of a
million can be achieved compared to the free molecules. In 1980, it was discovered that a similar
effect happens in the mid-IR spectral region, with an enhancement in IR adsorption of around 10-
1000 times.94 It was shown in studies correlating metal thickness and surface structure to the
enhancement that this effect exists over a short-range which is within ca. 8 nm of the metal
surface.95
Scheme 30
Schematic representation of the electromagnetic mechanisms of SEIRA on metal particles or islands. Incident IR
radiation induces a dipole p in the particle, which in turn generates a local electric field around the particle that
leads to enhanced adsorption in adsorbed molecules. Molecular vibrations can furthermore induce additional
dipoles 𝛿𝛿p that perturb the optical properties of the metal. Adapted with permission from94.
The observed surface enhancement is presumably provided by a sum of individual effects94:
39
1) Plasmon resonance (electromagnetic effect) – the metal particles are polarised by the
electric field of incident IR radiation due to the excitation of collective electron resonances (or
localised plasmon modes). This induces a dipole in the metal, which in turn generates a local
electromagnetic field around the particle that is stronger than that of the incident radiation.
The induced electromagnetic field 𝐸𝐸ó|£§ó is polarised along the surface normal to it at every
point on the particle surface. The magnitude of the field depends on the induced dipole 𝑝𝑝 and
decays sharply with the distance 𝑑𝑑 from the surface94:
𝐸𝐸ó|£§ó 9=4𝑝𝑝9
𝑑𝑑»40
The enhancement factor 𝐹𝐹 at a distance 𝑑𝑑 from the surface for a particle with a radius 𝑎𝑎w can
be estimated by94:
𝐹𝐹h…cúy 𝑑𝑑 = 𝐹𝐹h…cúy 0𝑎𝑎w
𝑎𝑎w+ 𝑑𝑑 41
The enhanced electromagnetic field leads to an enhanced absorption by the molecules. The
polarisation of the field along the surface normal means that only absorption due to
vibrational modes perpendicular to the surface are enhanced, rather than modes parallel to
the surface. This is known as the surface selection rule, which can be used to gain insight
into the orientation of molecules at the surface of plasmonic structures.
2) Perturbation of the optical properties of the metal (electromagnetic effect) – the
dipoles of adsorbed molecules on the surface of a metal can induce dipoles that perturb the
optical properties of the metal itself. This perturbation is enhanced around the vibrational
frequencies of the molecule and in effect amplifies the corresponding absorption due to
molecular vibrations.
3) Chemical mechanism – Chemisorption of molecules on the surface of the metal can lead
to interactions that change the polarisability of the molecules due to donor-acceptor
interactions, leading to an increase in the absorption coefficient.
2.4.6 Raman spectroscopy
Raman spectroscopy is another, complementary form of vibrational spectroscopy that provides
chemical and structural information like IR spectroscopy; however, rather than arising from a
change in the dipole moment of a molecule, Raman spectroscopy arises from the change in
polarisability of a molecule. A sample is irradiated with monochromatic (laser) light and most of
this light passes through the system or is elastically scattered due to Rayleigh scattering with no
40
change in energy; however, a small amount of light is inelastically scattered due to the Raman
effect.
Scheme 31
Schematic representation of the Raman scattering effect and resonance Raman scattering effect.
The Raman effect is illustrated in Scheme 31. The incident photons of light excite electrons in the
molecule, which is followed by the immediate re-emission of the photons as scattered light.
Rayleigh scattering results in electrons in the ground state becoming excited and returning to the
original ground state with no loss of energy (𝜈𝜈w= 𝜈𝜈w). Raman scattering results in the excited
electrons falling to a final energy state other than that original state. Stokes Raman scattering
involves electrons excited from the ground state falling to a vibrational level with emission of a
photon with less energy (𝜈𝜈w= 𝜈𝜈w− 𝜈𝜈∂lï), whereas anti-Stokes Raman scattering involves
electrons excited from a vibrational level falling to the ground state with emission of a photon with
greater energy (𝜈𝜈w= 𝜈𝜈w+ 𝜈𝜈∂lï). In general, the intensity of scattering due to Stokes Raman
scattering is greater than that due to anti-Stokes scattering; however, the ratio changes as the
temperature of the sample rises and molecules become more vibrationally excited.
As only a very small proportion of photons undergo Raman scattering, techniques are employed
to enhance scattering, including surface enhanced Raman spectroscopy (SERS) and resonance
Raman (RR) spectroscopy. SERS functions in a similar way to SEIRA(S), with surface plasmon
resonance at the surface of nanostructured metals inducing large electromagnetic fields that
enhance scattering, or due to donor-acceptor interactions with adsorbed molecules. Resonance
Raman scattering occurs when the wavelength of the incident light is similar to that of an
electronic transition of the molecule, as illustrated in Scheme 31. Vibrational modes associated
with excited electronic states are greatly enhanced, such as in chromophores like porphyrins
where charge-transfer electronic transitions enhance metal-ligand stretching modes, in addition to
vibrational modes associated with the ligand. The probability of Raman scattering can be
enhanced by a factor up to 106 using RR.
41
2.4.7 X-ray photoelectron spectroscopy
In X-ray photoelectron spectroscopy (XPS), a sample is irradiated with a high energy
monochromatic X-ray source of a known wavelength (usually Al Ka rays of 1486.7eV) in vacuum,
which results in the ejection of electrons from core levels of atoms in the sample into the vacuum.
The probability of these photoelectrons escaping without being inelastically scattered and losing
energy decays exponentially with depth, meaning that XPS is a surface sensitive technique (< 10
nm). Because the energy of the incident irradiation is known (ℎ𝜐𝜐), t he k inetic e nergy o f t he
emitted photoelectron 𝐸𝐸ò can be measured and the binding energy (BE) of the electron 𝐸𝐸F can be
determined according to:
𝐸𝐸F= ℎ𝜐𝜐 − 𝐸𝐸ò+ 𝜙𝜙 42
where 𝜙𝜙 is the work function of the spectrometer and sample.
An XPS spectrum plots the number of ejected photoelectrons against their binding energy. Each
element produces a set of well-defined peaks with characteristic binding energies corresponding
to different atomic orbital levels e.g. 1s, 2s, 2p, 3s etc. The number of ejected photoelectrons
corresponds directly to the amount of element in the sample and, therefore, atomic percentage
value can be determined. The binding energies of electrons can provide information on the
chemical state of elements in a sample. Changes in the oxidation state or chemical environment
of an atom can lead to characteristic chemical shifts that can be compared to known materials.
For example, it is possible to spectroscopically distinguish the different chemical environments of
nitrogen in an amine, an imine or a protonated amine. In general, higher oxidation states lead to a
greater coulombic interaction between the emitted electron and the ion core and, therefore, leads
to higher binding energies.
The ejection of electrons from p, d and f subshells leads to spectral peaks that are split into two
components. The unpaired spin of an electron in a final state can be up or down, and when it
occupies an orbital with non-zero orbital angular momentum there will be coupling between the
unpaired spin and the orbital angular momentum, or spin-orbit coupling. The generated states
𝑗𝑗µ= 𝑙𝑙 + ì
9 and 𝑗𝑗Q= 𝑙𝑙 + ì
9 are not degenerate and will therefore be observed in the spectrum. The
lowest energy component is the one with the maximum 𝑗𝑗 and the intensity ratio between the two
components is given by the ratio of the multiplicities 9ÃÕµì
9ÃŒµì =9œµ9
9œ , which in turn determines the
relative probability of transition to these states upon photoionisation.
42
2.5 Hydrogenases
The interconversion of chemical and electrical energy in fuel cells is seen as an alternative
method of deriving power over the conventional combustion of fossil fuels. Use of hydrogen as an
energy carrier results in relatively high efficiencies and the emission of water rather than CO2.
Conventional fuel cells use Pt to catalyse the hydrogen oxidation reaction (HOR) at the cathode
and the oxygen reduction reaction (ORR) at the anode. Hydrogenases are enzymes that can
catalyse the interconversion of molecular hydrogen into protons and electrons with similar
activities per active site to those of Pt catalysts96:
H2 ⇄ 2H+ + 2e- Eo = 0 V (vs NHE)
Scheme 32
A schematic representation of a hydrogen-powered enzymatic cell with a hydrogenase to catalyse the hydrogen
oxidation reaction (HOR) at the anode, and a bilirubin oxidase to catalyse the oxygen reduction reaction (ORR) at
the cathode. In this case, an (optional) polymer electrolyte membrane (PEM) separates the two compartments.
Reproduced with permission from 97, copyright Royal Society of Chemistry 2017.
As such, hydrogenases have been proposed as alternative catalysts for the HOR. Because Pt
catalyses both HOR and ORR at the cathode and anode, a proton exchange membrane (PEM) is
required to separate both compartments in order to avoid the formation of mixed potentials due to
cross-over, and to allow the selective transport of protons to the cathode. Due to the high
specificity/selectivity of enzymes for a given substrate, an enzyme-based fuel cell (e.g. with
bilirubin oxidase at the cathode) can use mixed fuel-oxidant feeds (in contrast to e.g. platinum). In
addition, the absence of a PEM can save costs and offers potential for miniturisation.98 A
schematic representation of such an enzymatic fuel cell (in this case with a PEM) is shown in
Scheme 32. In addition to being used in fuel-cells, hydrogenases can also be used to generate
hydrogen (hydrogen evolution reaction – HER) either through fermentation using organic
substrates99, or using solar energy or electricity.100–102
43
There are three major types of hydrogenases: [Fe]-, [FeFe]- and [NiFe]-hydrogenases. Due to the
different binding of hydride at the active site, different catalytic activities for the HOR and HER are
observed for different hydrogenases.103 The [FeFe]-hydrogenase from Desulfovibrio
desulfuricans is active for the HOR and HER under 1 bar of H2, whereas [NiFe]-hydrogenases
show little or zero activity for the HER under the same conditions due to inhibition by H2. Currents
for the HER can be observed for [NiFe]-hydrogenases under inert atmospheres, e.g. a small
current can be observed for the [NiFe]-hydrogenase from Ralstonia eutropha. This preference for
catalysis in one direction is known as the catalytic bias of the enzyme.96 It was shown for the O2-
tolerant [NiFe] hydrogenase from Escherichia Coli that the catalytic bias of [NiFe]-hydrogenases
is pH sensitive, with the bias towards the HER maximised at low pH’s.104
Scheme 33
The overall structure of the heterodimeric membrane-bound hydrogenase (MBH) from Ralstonia eutropha with the
cellular localisation of the enzyme shown in the upper left corner. The large subunit harbouring the [NiFe] catalytic
active site (the molecular structure of which is drawn on the right) is coloured in blue, while the smaller subunit
harbouring the three Fe-S cluster electron relay chain is coloured in green. The active site, enlarged for better
visibility on the right, is connected through the electron relay to the respiratory chain via a b-type cytochrome.
Adapted with permission from 105, copyright Springer Nature 2011.
The focus in this thesis will be on [NiFe]-hydrogenase. [NiFe]-hydrogenases contain an active site
with an Fe atom with one CO and two CN- ligands, and a Ni atom that is bridged to the Fe atom
via two cysteine thiolate donors, as shown for the membrane-bound hydrogenase (MBH) from
Ralstonia eutropha in Scheme 33. The Ni is further ligated by two more cysteine ligands that
attach the entire active site to the protein backbone of the enzyme. The active site is
electronically wired via an electron transfer relay chain of Fe-S clusters to a redox pool. In a
bacterial membrane, this would be a cytochrome that is linked to the respiratory chain via
electron-shuttling quinones inside the membrane.106 In an enzymatic fuel cell or other electronic
device, this would ideally be an electrode, as illustrated later in Scheme 35.
44
Scheme 34
The catalytic states for a [NiFe]-hydrogenases showing the deactivation and subsequent reactivation mechanism
of the oxidised states. The catalytic cycle is indicated by the black arrows in the hydrogen oxidation reaction
(HOR) direction. Deactivation is indicated by either the red or green arrows. In the case that electrons are readily
available at the active site, the Ni-B state is formed with a hydroxide ligand, which can rapidly be reactivated upon
a one electron, one proton reduction back to the Ni-SIa state. In the case that not enough electrons are available
(i.e. in O2 sensitive hydrogenases), the Ni-A state is formed, which requires low potentials and extended
reactivation times to form the Ni-SIa state. Anaerobic inactivation (e.g. potential-induced) is indicated by the dotted
lines. Reproduced with permission from 103, copyright American Chemical Society 2014.
In [NiFe]-hydrogenases, O2 reacts with the active site giving a mixture of inactive states, the so-
called Ni-A or Ni-B states, depending on the nature of the oxygen ligand that bridges the Ni and
Fe atoms (for the proposed key states existing in the catalytic cycle of [NiFe]-hydrogenases see
Scheme 33). Certain [NiFe]-hydrogenases exhibit O2 tolerance, which means that they form the
Ni-B state (with a hydroxide ligand), which can be rapidly reactivated back to the Ni-SIa state
upon a one electron, one proton reduction, removing the O2 as water and giving rise to the O2
tolerance. O2 sensitive hydrogenases form the Ni-A state (which is believed to contain a peroxo
ligand) in a poorly reversible process that requires low potentials and extended reactivation times
to form the Ni-SIa state.107 The Ni-B state only forms upon the availability and immediate delivery
of electrons and protons at the active site. In the case that not enough electrons are available (i.e.
in O2 sensitive hydrogenases), the Ni-A state is formed.108,109 This ‘fast’ delivery of electrons in
the case of the O2 tolerant hydrogenases is the result of the unique [4Fe-3S] proximal cluster
(shown in Scheme 33), which differs significantly in its electronic structure from the [4Fe-4S]
cubane proximal clusters found in O2 sensitive hydrogenases function. This cluster can
undergo two one-electron redox transitions, instead of a single one-electron redox transition, and
45
is thus capable of providing two electrons for complete oxygen reduction, with the other two
electrons coming from the active site and the medial [3Fe4S]-cluster).105,109,110
The Knallgas bacterium Ralstonia eutropha (Re) H16 contains at least three such O2 tolerant
[NiFe]-hydrogenases105, with the MBH specifically being used in this study.
Scheme 35
Schematic representation of an enzyme with an Fe-S electron relay chain immobilised on an electrode in a direct
electron configuration. Possible limitations are illustrated by a series of resistors, where WE represents interfacial
electron transport, Wcat the intramolecular electron transport, and Wtrans the mass transfer of substrates and
products. Adapted with permission from 107, copyright Springer 2014.
Enzymatic electrode devices (e.g. enzymatic fuel-cells, sensors) usually employ an electrode that
is coated with a layer of enzymes. The enzymes may be orientated on the electrode in a direct
electron transfer configuration or in an indirect configuration (see Scheme 36). A direct
configuration means that there is a direct exchange of electrons between the electrode and the
enzyme without the need of any additional electroactive mediators to ‘shuttle’ the electrons
between the electrode and the enzymes’ active site, which is the case for an indirect
configuration. Direct electron transfer via electron tunnelling is only possible if either the active
site of the enzyme or the electron transfer relay chain (if there is one) is positioned close enough
to the electrode surface. The interfacial electron transfer 𝑘𝑘…∆ rate decreases exponentially with
distance 𝑑𝑑 from the electrode surface as follows:
𝑘𝑘…∆ 𝑑𝑑 = 𝑘𝑘wexp −𝛽𝛽𝑑𝑑 43
where 𝑘𝑘w is the maximum value at closest contact, and 𝛽𝛽 is a factor that depends on the height of
the energy barrier and the nature of the medium between the redox site and the electrode.96,111
46
Enzymes are immobilised on different types of electrode materials using different approaches.
They can adsorb onto metal oxide and carbon materials often without any prior modification of
the surface, whereas metal electrodes like gold or silver often require a SAM to render the
surface biocompatible and prevent the unfolding of the protein backbone, in particular during the
application of electric fields. Enzymes can also be immobilised on an electrode in a bio-
membrane like structure, such as a surfactant film or a lipid bilayer.112 The binding interaction
between the enzyme and the electrode may be covalent or non-covalent (for more details see
section 5.3.9). For example, at moderate pH’s, an amino-functionalised surface will have a
positive charge and will therefore interact with negatively charged patches on the surface of a
protein e.g. the negatively charged area around the distal Fe-S cluster of the Desulfofovibrio
gigas [NiFe]-hydrogenase.96 The Re MBH lacks such a distinctive negatively charged surface
patch.85 The type of linker molecules or interface used to modify the electrode will have an effect
on 𝑘𝑘…∆, with conjugated molecules having lower values of 𝛽𝛽 than non-conjugated molecules, and,
in general, larger values of 𝑘𝑘…∆.
Protein film voltammetry (PFV) is often used to study the electrochemical and catalytic behaviour
of enzymes on electrodes. The maximum electrical current (when mass transport is not limiting,
e.g. when using a rotating disc electrode) corresponds to the catalytic activity of the enzyme, and,
if the coverage of enzyme Γ is known, then the catalytic current 𝑖𝑖£§“ can be used to determine the
turnover frequency 𝑘𝑘£§“:
𝑘𝑘£§“ =𝑖𝑖£§“
𝑛𝑛𝐹𝐹𝐴𝐴Γ44
where 𝐴𝐴 is the electrode surface area and 𝑛𝑛 is the number of electrons transferred during the
reaction.96
Figure 3
Protein film voltammogram of a [FeFe]-hydrogenase from the Desulfovibrio desulfuricans covalently attached to a
pyrolytic graphite edge (PGE) electrode at 10 °C under 1 bar H2 in pH 7.0 with a 2500 rpm electrode rotation rate.
The reversible cleavage of H2 via the hydrogen oxidation reaction (HOR) and hydrogen evolution reaction (HER)
is indicated by positive and negative currents, respectively. The switch potential for reductive reactivation of the
hydrogenase is indicated by Eswitch. Reproduced with permission from 103, copyright American Chemical Society
2014.
47
Figure 3 shows a protein film voltammogram of a hydrogenase displaying characteristic
reversible cleavage of H2 via the HOR and the HER. Hydrogenases typically operate at the
thermodynamic potential for both of these reactions, as indicated by the straight trace through the
zero-current line. At high anodic overpotentials, the hydrogenase undergoes anaerobic oxidative
inactivation via the reversible formation of the Ni-B state, and thus the current drops accordingly.
On the reverse sweep, the hydrogenase is reductively reactivated close to the switch potential,
Eswitch, with the removal of the bridging hydroxide ligand.96
Scheme 36
Schematic representation of electrocatalysis proceeding through direct electron transfer between an electrode
and an immobilised MBH from Ralstonia eutropha (left) or through mediated electron transfer via a mediator (in its
oxidised MedOX or reduced MedRED state) in solution.
As previously mentioned, enzymes that are immobilised on an electrode in an indirect electron
transfer configuration may require the use of electroactive mediators, such as methylene blue or
methyl viologen, to ‘shuttle’ electrons between the active site and the electrode. When there is a
potential difference between the electrode and the enzyme, the mediator is cycled between its
oxidised and reduce state, as illustrated in Scheme 36. The mediator must be capable of fast and
reversible electron transfer with little overpotential, meaning the choice of mediator and its redox
potential is critical.113 These mediators may be present in solution, attached to the protein, or
present in a polymeric matrix such as a hydrogel that also contains the enzyme.114 Such matrices
can also afford additional stability against inactivation or damage to enzymes, such as oxygen
protection for an O2 sensitive hydrogenase.115 For diffusion-controlled mediators, interfacial
electron transfer rates of the mediator at the electrode-electrolyte interface, as well as electron at
the enzyme, will be determined by factors such as hydrophobic or electrostatic interaction, or the
nature of an interface (e.g. the charge transfer resistance experienced by the mediator).
Methylene blue is used as a mediator with the MBH from the Ralstonia eutropha hydrogenase
because it overlays with the potential range in which the HOR efficiency is at its greatest. The
electrocatalytic currents due to direct electron transfer (DET) can be compared to the currents
due to mediated electron transfer (MET) in the presence of a mediator to get an idea about
whether the enzymes in a system are in a direct or indirect electron transfer configuration.
48
2.6 Molecular catalysts
2.6.1 Oxygen reduction reaction at metal macrocycles
Catalysts are used on the cathode of a fuel cell to catalyse the reduction of oxygen via the ORR.
The electrochemical reduction of oxygen is a multi-electron reaction that can proceed via two
pathways: a two-electron reduction to give peroxide, or a four-electron reduction to give water.
Two-electron pathway:
Acid (pH 0)
O
2
+ 4H+ + 2e- ⟶ 2H
2
O
2
Eo = 0.67 vs. NHE
Base (pH 14)
O
2
+ H
2
O + 2e- ⟶ HO
2
- +
OH-
Eo = -0.065 vs. NHE
Four-electron pathway:
Acid (pH 0)
O
2
+ 4H+ + 4e- ⟶ 2H
2
O
Eo = 1.229 vs. NHE
Base (pH 14)
O
2
+ 2H
2
O + 4e- ⟶ 4OH-
Eo = 0.401 vs. NHE
In order to provide the maximum free energy in a fuel cell, it is necessary to reduce O2 via the
four-electron pathway. This requires the O-O bond to be broken and the nature of the catalyst
being used will have a strong influence on the pathway followed. Most electrode materials
catalyse the two-electron pathway. The O2 molecule has a bond order of two, with two unpaired
electrons in a doubly degenerate antibonding p* orbital and four bonding electrons, which
explains its high stability (118 kcal/mol). When O2 is reduced on a catalyst surface such as Pt,
electrons are added to the antibonding p* orbitals e.g. through the back bonding of transition
metal d-orbitals, weakening and lengthening the O-O bond and increasing the likelihood of
oxygen reduction taking place. A number of possible interactions between O2 and the surface of
a transition metal are possible, including end-on or side-on interaction with a single metal atom,
or a bridging interaction between two metal atoms (see Scheme 37). It has been proposed that
oxygen adsorbs on Pt via a bridging interaction, which favours oxygen reduction via a four-
electron pathway.116
49
Scheme 37
Possible side-on and head-on modes of interaction of O2 with a transition metal centre, indicating the respective
molecular orbitals involved. Adapted with permission from 117, copyright Elsevier 1992.
The high cost of Pt and other noble metals used for ORR has led to efforts to develop non-noble
metal catalysts based on N4 metal macrocycles, inspired by biological catalyst like cytochrome c
and haemoglobin. Furthermore, N4 metal macrocyclic complexes exhibit a high tolerance to
methanol compared to platinum catalysts, making them interesting for methanol fuel-cell
applications. A number of N4 metal macrocyclic complexes have been shown to be active for
ORR. Fe and Co porphyrins and phthalocyanines catalyse the reduction of oxygen either via two-
or four-electron pathways, depending on the pH of the electrolyte. Certain complexes only
catalyse the reaction via a two-electron pathway, as is the case for Co, Ni and Cu
phthalocyanines.118
A one-electron reduction pathway of O2 to give a superoxide occurs through an outer-sphere
process, whereas a two- or four-electron pathway takes place through an inner-sphere process.
Similar to Pt, O2 interacts with an N4 metal macrocycle by binding to the central metal d-orbitals
either via an end-on or side-on interaction, weakening and lengthening the O-O bond. The
strength of the interaction will depend on the electronic density of the d-orbitals. The metal should
be in a M2+ oxidation state, and an adduct is formed as follows116:
M2+ + O2 ⇄ M3+-O2- or M2+-O2
This adduct should be short-lived so that the interaction does not block the active site and is
reduced as follows:
50
M3+-O2- + e- ⇄ or M2+ + intermediates
The final step is for the ORR in alkaline conditions and could also involve M2+-O2, while in acidic
conditions protons are involved in the process.118
For Co complexes, oxygen reduction occurs at potentials more negative than the Co2+/Co3+
couple, whereas for Fe complexes, reduction usually takes place close to the Fe2+/Fe3+ couple. At
very negative overpotentials, H2O2 becomes the main product of the reaction.116
5.6.2 Water oxidation molecular catalysts
Water oxidation is the first step in photosynthesis, where water is used as a source of electrons
and protons:
2H2O ⟶ O2 + 4H+ + 4e- Eo = 1.229 vs. RHE
The reaction is thermodynamically and kinetically very demanding with many intermediate steps
resulting in slow kinetics. Use of a catalyst that stabilises these intermediates can lower the
kinetic barrier and lead to reduced overpotentials. Technologies based on the production of
sustainable fuels from renewable energy or sunlight rely on having an abundant source of
protons and electrons. Heterogeneous catalysts that catalyse the water oxidation reaction include
transition metal oxides such as iridium, ruthenium, nickel or cobalt oxides.119 Ir and Ru catalysts
tend to operate in acidic conditions, while Ni and Co catalysts tend to operate in basic conditions
(but not always). The first homogenous water oxidation catalyst developed was Meyer’s “Blue
Dimer” in 1982, which is based on Ru.120 Many other catalysts have been developed since then
that are highly active for water oxidation, in particular many Ir-based catalysts.121 Given the
scarcity and high cost of Ru and Ir, efforts have been made to develop non-noble metal catalysts
based on Fe, Mn (like nature’s oxygen-evolving photosystem II complex), Co and Cu. These
efforts have recently been reviewed by Crabtree, Brudvig and co-workers.119
51
6 Materials, instruments and methods
Chapter 3
Materials, instruments and methods
3.1 Electrode synthesis
3.1.1 Gold SEIRA film preparation
An ATR Si prism coated was coated with a nanostructured gold film via electroless deposition.122
The surface of the Si prism was first polished with alumina powder (Microgrit WCA-9, grain size:
ca. 6 μm) and then thoroughly rinsed with water. The polished surface was then etched with a
400 g/L NH4F solution for 2 min to remove the silicon oxides on the surface to leave a bare silicon
surface. The gold film was deposited by heating the prism up to 65 ˚C in a water bath and
dropping a gold plating solution (aqueous 1:1:1 volume ratio of 2% (w/w) HF, 0.03 M
NaAuCl4·H2O and a reduction solution comprised of 0.3 M Na2SO3, 0.1 M Na2S2O3·5H2O, and 0.1
M NH4Cl) on the surface for 1 min. The film is formed as a result of the reduction of AuIII species
in the plating solution results and the concomitant oxidation of the Si surface122:
Si0(s) + 6F−(aq) → SiF2−(aq) + 4e−
AuCl−4 (aq) + 3e− → Au0(s) + 4Cl−(aq)
The gold films were cleaned electrochemically by repeated cycling between 0 and 1.4 V (vs
Ag/AgCl 3M KCl) at 100 mV/s in 0.1 M HClO4 under Ar bubbling until reproducible CV traces
were obtained. The gold surface area was calculated using the Au(III) oxide reduction peak by
using the following formula A = Q1/Q0, where Q1 is the measured charge and Q0 is the charge
required to reduce 1 cm2 of Au(III) oxide (Q0 = 400 µCcm-²).123
3.1.2 Indium tin oxide electrode preparation
The ITO electrodes used in this work are ITO-coated glass electrodes (8-12 Ω/sq) purchased
from Sigma Aldrich (hereby denoted pl-ITO). ITO films are generally deposited on substrates
using sputtering, whereby particles are ejected from a solid target material in a vacuum chamber
as a result of bombardment using energetic particles, such as gas ions, and deposited on all
surfaces in the chamber, including any substrate present. As sputtering thus requires specialised
equipment, planar ITO films were deposited on a Si prism for ATR-IR measurements using a sol-
gel method from the literature.124 A 10% Sn-doped In2O3 (ITO) indium-tin-acetylacetonate
solution, where Sn% = Sn/(Sn+In), was prepared by dissolving 20 mmol indium (III) nitrate
hydrate (In(NO3)3·xH2O, 99.99%, Sigma Aldrich) and 2.2 mmol tin (IV) chloride hydrate
52
(SnCl4·xH2O, Sigma Aldrich) in acetylacetone (Sigma Aldrich) and stirring overnight at 60 ˚C,
resulting in a transparent brown solution. In a modification to the original literature method, and to
improve the adhesion of this solution to the Si surface, an aliquot of warm indium-tin-
acetylacetonate was diluted with an aliquot of toluene, before spin-coating on a freshly-polished
Si-prism. The coated prism was then immediately placed on a hotplate set at 300 ˚C for 5
minutes and then calcined in air at 400 ˚C for 10 minutes, resulting in a blue-tinted film. A second
calcination was carried out under a reducing atmosphere (10% H2, 90% N2) at 300 ˚C for 90 min
with a ramp-rate of 20 ˚C/min. The resulting films (hereby denoted sc-ITO) were characterised
using scanning electron microscopy (SEM).
3.1.3 Antimony-doped tin oxide electrode preparation
Crystalline 8% Sb-doped SnO2 (ATO) nanoparticles, where Sb% = Sb/(Sb+Sn), were
synthesised via a non-hydrothermal synthesis route by dissolving 6.25 mmol of tin tetrachloride
(SnCl4 Sigma Aldrich) and 0.55 mmol on antimony acetate (Sigma Aldrich) in 5 mL toluene in a
glovebox and adding this solution dropwise into 15 mL benzyl alcohol under continuous stirring.
The glass container with the resulting clear solution was removed from the glovebox and
transferred into a Teflon-lined autoclave, sealed and heated at 150 oC for 3 hrs. The resulting
brown particles were removed and sequentially cleaned using 15 min of ultrasonication twice in
toluene and three times in acetone. Centrifugation was used for 20 min at 5,000 rpm to separate
the supernatant from the particles.
Colloidally stable solutions of ATO nanoparticles were formed by sonicating and stirring 160 mg
ATO nanoparticles and 80 mg Pluronic F127 (Sigma Aldrich) in 1 mL THF with a few microliters
of conc. HCl added to aid the dispersion. Mesoporous ATO (me-ATO) films were electrogred on
planar ITO-coated glass substrates (pl-ITO, 8-12 Ω/sq, Sigma-Aldrich) by evaporation induced
self-assembly (EISA) whereby the substrates were immersed into the solutions using a dip-coater
with a 200 mm/min withdrawal rate in a 50% relative humidity at 20 oC. The resulting films were
aged for 12 hrs in air at 100 oC and calcined at 450 oC for 30 min with a temperature ramp rate of
0.6 oC/min. The resulting transparent films were characterised using scanning electron
microscopy (SEM). pl-ITO were cleaned sequentially by sonication for 5 min in water, ethanol and
acetone before coating with me-ATO or uncoated before electrochemical measurements.
me-ATO films for attenuated total reflectance Infrared spectroscopy (ATR-IR) and rotating disk
electrode (RDE) experiments were electrogred on an un-doped Si prism and glassy carbon tips
(5 mm diameter, Pine Research) by spin coating the same colloidal solution of ATO nanoparticles
as that used on pl-ITO, followed by the same ageing and calcination steps.
The me-ATO films on ITO used in section 9 were synthesised by Victoria Davis (from the group
of Prof. Dr. Anna Fischer, Albert-Ludwigs-Universität Freiburg).
53
3.2 Precursor synthesis
3.2.1 Diazonium salts
4-benzamidobenzenediazonium tetrafluoroborate (4-BABD) and 4-(1H-imidazol-1-
yl)benzenediazonium tetrafluoroborate salt (Im-diazo) were synthesised by diazotising the
corresponding amines 1 4’-aminobenzanilide (Sigma-Aldrich) and 2 4-(1H-Imidazol-1-yl)aniline
(Sigma-Aldrich) using nitrosonium tetrafluoroborate (NOBF4). The method used was adapted
from literature methods for the synthesis of other diazonium salts.52,125 Briefly, the amine was
dissolved in a minimum amount of acetonitrile (MeCN) in a round-bottomed flask at -41 °C and
kept under an inert Ar atmosphere on a Schlenk line. 1.1 equivalents of NOBF4 dissolved in
acetonitrile were added under stirring, and the mixture stirred for 30 min at -41 °C. After warming
to room temperature, diethyl ether was added to precipitate the diazonium salt. The filtrate was
filtered using a Büchner funnel, washed with diethyl ether and allowed to dry overnight under
vacuum. The brightly coloured products (pink and orange-brown for 4-BABD and Im-diazo,
respectively) were stored at 4 °C in the dark.
Scheme 38
The synthesis of (a) 4-benzamidobenzenediazonium tetrafluoroborate (4-BABD) and (b) 4-(1H-imidazol-1-
yl)benzenediazonium tetrafluoroborate (Im-diazo) from the corresponding amines 4’-aminobenzanilide (1) and 4-
(1H-Imidazol-1-yl)aniline (2).
Figure 4
Electrospray mass spectrum of 4-BABD.
54
Figure 5
Electrospray mass spectrum of Im-diazo.
Solutions of 4-BABD and Im-diazo were prepared in acetonitrile and characterised using
electrospray ionisation mass spectroscopy (ESI-MS) (see Figure 4 and Figure 5). Peaks due to
the intact diazonium cations are present in both spectra at m/z 224.08 and 171.07, respectively.
Peaks for the aryl cations that form as the result of dediazoniation taking place during the
ionisation process are also visible for 4-BABD and Im-diazo at m/z 196.08 and 143.06,
respectively, and such cationic species have previously been observed in the literature.126 In the
case of Im-diazo, there is a greater contribution from product fragments that form due to the
reaction between the aryl cation and solvent molecules. This reflects the higher reactivity of the
aryl cation due to less stabilisation via inductive effects from the para-substituent compared to 4-
BABD. The difference in the inductive effect of each substituent can clearly be seen in the
electrochemical reduction potentials measured for each salt in Table 1 (page 27).
4-([2,2’-Bipyridine]-4-carboxamido)benzenediazonium tetrafluoroborate (bpy-diazo) was
synthesised by Dr. Anja Sokolowski (formerly from the group of J.Prof. Dr. Carl Christoph
Tzschucke, Freie Universität Berlin) from N-(4-Aminophenyl)-[2,2’-bipyridine]-4-carboxamide.
3.3 Protein purification and catalyst preparation
3.3.1 Membrane-Bound Hydrogenase (MBH) from Ralstonia eutropha
The heterodimeric MBH was purified using a Strep-tag II peptide as an affinity tag, which was
attached to the protein by genetically engineering the small subunit of the MBH (HoxK). The
growth conditions and purification are described in detail by Goris et al.109 The samples were
purified by Dr. Stefan Frielingsdorf (from the group of Dr. Oliver Lenz, Technische Universität
55
Berlin). The purified MBH heterodimer samples were buffered in 50 mM KPO4 buffer at pH 5.5
with 150 mM NaCl and subsequently stored in liquid N2.
3.3.2 Hangman complexes
The different Hangman porphyrin complexes used in this work (FePOMe and CoPOMe) were
synthesised by Dr. Pierre Wrozlek (from the group of Dr. Matthias Schwalbe, Humboldt -
Universität zu Berlin) using methods published in the literature.127–129
Scheme 39
The molecular structures of the Hangman porphyrin complexes used this work.
3.4 Electrochemistry
Electrochemical measurements were performed using a three-electrode set-up that employed a
working electrode (a gold SEIRA film, me-ATO film, pl-ITO film or sc-ITO film), a Pt wire as
counter electrode and an adequate reference electrode. Control of the working electrode was
achieved using either a Metrohm µAutolabIII potentiostat, or a CHI potentiostat. For
measurements in aqueous solvents an Ag/AgCl 3M KCl (Dri-Ref, WPI) reference electrode was
used, while for measurements in acetonitrile, either an Ag/AgNO3 0.1 M AgNO3 (homemade)
reference electrode or a Ag/AgCl 3M KCl (Leak-Free, Warner Instruments) reference electrode
was used. Potentials in acetonitrile were measured relative to the Fc/Fc+ couple of a 0.1 or 1 mM
ferrocene solution in acetonitrile (0.1 M TBAF or TBAP). For electrochemical impendence spectra
using a 1 mM Fc, a μAutolabIII (Metrohm) was used and controlled with FRA2 software and
spectra were recorded in a frequency range of 0.00001 to 1000 kHz with an amplitude of 25 mV
(rms).
For ATR-IR and SEIRA spectroelectrochemical measurements, a home-made electrochemical
cell was used. For more details, see 6.6.1
Electrochemical measurements using pl-ITO (i.e. on a glass substrate) or me-ATO films on pl-
ITO as a working electrode were performed in a home-made PTFE cell. A Viton® O-ring was
56
used to seal the spectroelectrochemical cell on top of the electrodes, forming a working electrode
with a geometric area of ca. 0.4 cm2. The geometric surface area was used to calculate current
densities. For UV-Vis spectroelectrochemical measurements, a home-made electrochemical cell
was used. For more details see 6.6.5.
For rotating disk electrode (RDE) measurements (using me-ATO films coated on a glassy carbon
RDE tips), an RDE setup from PINE Research Instrumentation was used. The geometric surface
area was used to calculate current densities.
3.5 Electrode modification
The electrochemical reduction potentials of all the diazonium salts used in this study in
acetonitrile, determined using cyclic voltammetry with a pl-ITO working electrode, were shown
previously in in Table 1 (page 27).
3.5.1 Functionalisation of gold
4-NTP and 4-ATP SAMs were deposited overnight on gold from 1mM ethanolic (HPLC grade)
solutions and thoroughly rinsed with copious amounts of ethanol.
Electrochemical grafting and desorption measurements were performed in acetonitrile with 0.1 M
tert-butylammonium hexafluorophosphate (TBAF) electrolyte. All electrochemical measurements,
unless stated otherwise, were performed under inert argon atmosphere.
Benzanilide interfaces were electrochemically grafted on gold from 1 mM 4-BABD solutions in
acetonitrile by cycling two-times between +0.12 and -0.88 V (vs vs Fc/Fc+) at a scan rate of 50
mV/s.
Nitrophenyl interfaces were electrochemically grafted from 1 mM 4-NBD solutions in acetonitrile
with and without the addition of different molar equivalents of the radical scavenger 2,2-diphenyl-
1-picrylhydrazyl (DPPH) using chronoamperometry by applying a potential of -0.445 V (vs Fc/Fc+)
for 60 s, followed by rinsing the electrode in copious amounts of acetonitrile, followed by ethanol.
The nitrophenyl interface was spontaneously deposited from a 20 mM 4-NBD solution in
acetonitrile without any application of potential for 60 min, followed by copious rinsing in
acetonitrile and ethanol. Nitrophenyl interfaces were electrochemically reduced to aminophenyl
interfaces by cycling the electrode between 0 and -1.2 V (vs Ag/AgCl 3M KCl) at a scan rate of 50
mV/s in a 0.1 M NaClO4 1:9 ethanol:water solution. Final SEIRA spectra of the reduced interfaces
were recorded in dry acetonitrile. Details regarding the MBH immobilisation and
bioelectrocatalysis are given in 6.5.4.
57
Bipyridine interfaces were electrochemically grafted from 1 mM bpy-diazo solutions in 0.1 M
TBAF by cycling five times between 0.351 and -0.948 (vs Fc/Fc+) at a scan rate of 50 mV/s.
Manganese carbonyl bipyrydyl interfaces were formed by incubating the electrografted bipyridine
interface with a 1 mM Mn(CO)5Br solution in acetonitrile for 60 min, followed by copious rinsing in
acetonitrile. Manganese carbonyl bipyrydyl interfaces were also electrochemically grafted from
mixed 1 mM bpy-diazo and 1 mM Mn(CO)5Br solutions, which were allowed to react for 60 min
prior to electrochemical grafting. Electrocatalysis was carried out in acetonitrile (0.1 M TBAP)
after bubbling the solution with CO2.
3.5.2 Functionalisation of ITO
Electrochemical grafting and desorption measurements were performed in acetonitrile with 0.1 M
tert-butylammonium perchlorate (TBAP) electrolyte. All electrochemical measurements, unless
stated otherwise, were performed under inert argon atmosphere. Prior to modification, the pl-ITO
electrodes were sonicated for 5 min sequentially in water, ethanol and acetone, before drying
under a stream of nitrogen. Nitrophenyl interfaces were electrochemically grafted from 1 mM 4-
NBD solutions in acetonitrile (0.1 M TBAP) with and without the addition of different molar
equivalents of DPPH using chronoamperometry by cycling between 0.14 and -0.76 V (vs Fc/Fc+)
at a scan rate of 50 mVs-1, followed by rinsing the electrode in copious amounts of acetonitrile,
followed by ethanol. In the same way as was carried out on gold, the nitrophenyl interfaces were
electrochemically reduced to aminophenyl interfaces by cycling the electrode between 0 and -
1.2 V (vs Ag/AgCl 3M KCl) at a scan rate of 50 mV/s in a 0.1 M NaClO4 1:9 ethanol:water
solution. Details regarding the enzyme immobilisation and bioelectrocatalysis are given in 6.5.4.
Bipyridine-functionalised interfaces were electrochemically grafted from 1 mM bpy-diazo solutions
in acetonitrile (0.1 M TBAP) by cycling between -0.16 and -1.06 V (vs Fc/Fc+) at a scan rate of 50
mVs-1, followed by rinsing the electrode in copious amounts of acetonitrile, followed by ethanol.
After modification, the bipyridine-functionalised interface was incubated with 2 mM CuSO4 (in DI
water) for 10 min followed by copious rinsing with water. Electrocatalysis was performed in 0.1 M
NaOH in air.
Imidazole-functionalised interfaces were electrochemically grafted from 1 mM Im-diazo solutions
in acetonitrile (0.1 M TBAP) by using chronoamperometry and applying a potential of -0.56 V (vs
Fc/Fc+) for 120 s, followed by rinsing the electrode in copious amounts of acetonitrile and
dichloromethane. Alternatively, interfaces were also electrografted from 1 mM Im-diazo solutions
in acetonitrile (0.1 M TBAP) by applying a linear sweep voltammogram between 0.04 V and -0.86
V (vs Fc/Fc+) at a scan rate of 50 mV/s. FePOMe was immobilised on pl-ITO using two different
strategies: (a) a two-step ‘post-coordination’ process whereby imidazole-functionalised interfaces
were first electrografted from 1 mM Im-diazo using the aforementioned methods, followed by
incubation with 1.2 mM FePOMe; and (b) a one-step process whereby an interface is
58
electrochemically grafted on pl-ITO directly from 1 mM Im-diazo and 1.2 mM FePOMe (after
mixing them for 5 min i.e. pre-coordinating the iron centre to the axial ligand). In each case a
potential of -0.56 V (vs Fc/Fc+) is applied for 120 s, followed by copious rinsing with acetonitrile
and dichloromethane. Electrocatalysis was carried out in a pH 7 0.1 M phosphate buffer (PB).
3.5.3 Functionalisation of ATO
Electrochemical grafting and desorption measurements were performed in acetonitrile with 0.1 M
TBAP electrolyte. All electrochemical measurements, unless stated otherwise, were performed
under an inert argon atmosphere.
Imidazole-functionalised interfaces were electrochemically grafted from 1 mM Im-diazo solutions
in acetonitrile (0.1 M TBAP) by using chronoamperometry and applying a potential of -0.42 V (vs
Fc/Fc+) for 120 s, followed by rinsing the electrode in copious amounts of acetonitrile and
dichloromethane. FePOMe was immobilised on me-ATO using two different strategies: (a) a two-
step ‘post-coordination’ process whereby imidazole-functionalised interfaces were first
electrografted from 1 mM Im-diazo using the aforementioned method, followed by incubation with
1.2 mM FePOMe; and (b) a one-step process whereby an interface is electrochemically grafted
on me-ATO directly from 1 mM Im-diazo and 1.2 mM FePOMe (after mixing them for 5 min i.e.
pre-coordinating the iron centre to the axial ligand). In each case a potential of -0.42 V (vs Fc/Fc+)
is applied for 120 s, followed by copious rinsing with acetonitrile and dichloromethane.
Electrocatalysis was carried out either in pH 7 0.1 M PB, or in 0.1 M KOH under O2 bubbling.
CoPOH was immobilised on me-ATO using the two-step ‘post-coordination’ process whereby
imidazole-functionalised interfaces were first electrografted from a 1 mM Im-diazo solution,
followed by incubation with 1.2 mM CoPOH. Electrocatalysis was carried out in 0.1 M NaOH in
air.
3.5.4 Enzyme immobilisation
MBH was immobilised by adsorption onto the electrode surfaces for protein film voltammetry
(PFV) and spectroelectrochemical studies.
MBH was immobilised on aminophenyl-functionalised SEIRA gold electrodes (functionalised
using the method introduced in 6.1.1) by incubating the electrodes in a 1 µM solution of MBH in
10 mM phosphate buffer (PB) pH 7 at 5 °C for 15 min. Electrodes were rinsed with 10 mM PB
pH 7 and bio-electrocatalysis was performed at 5 mV/s in 10 mM PB pH 5.5 at 25 °C, after
saturation of the buffer with O2-free Ar or O2-free H2 gas.
MBH was immobilised on bare pl-ITO and sc-ITO, as well as aminophenyl-functionalised pl-ITO
and sc-ITO electrodes (functionalised using the method introduced in 6.1.1) by incubating the
59
electrodes in a 1 µM solution of MBH in 10 mM phosphate buffer (PB) pH 7 at 5 °C for 15 min.
Electrodes were rinsed with 10 mM PB pH 7 and bio-electrocatalysis was performed at 10 mV/s
in 10 mM PB pH 5.5 at 25 °C, after saturation of the buffer with O2-free Ar or O2-free H2 gas.
Blank measurements of the modified electrodes without MBH were recorded in the same
conditions.
3.6 Spectroscopic characterisation
3.6.1 ATR-IR and SEIRA
All ATR-IR and SEIRA measurements were performed with a Kretschmann ATR-type
configuration implementing a Si prism, as described elsewhere.130 The Si ATR prism is a
trapezoidal Si crystal (L × W × H = 25 mm × 20 mm × 10 mm) and was coated with either a
SEIRA gold, sc-ITO or me-ATO film using the methods outlined in 6.1. Either a Bruker IFS66v/s
or a Tensor 27 spectrometer equipped with a liquid N2 cooled photoconductive Mercury Cadmium
Telluride (MCT) detector was used. A globar was used as an IR radiation source in both
spectrometers and the prism was irradiated at an angle of incidence of 60˚, with a resulting cross-
sectional measuring area of 7 × 3 mm. A temperature-controlled electrochemical cell was used,
maintaining a temperature of 25 oC, unless stated otherwise. Spectra were collected between
4000 and 1000 cm-1 with a spectral resolution of 4 cm-1 and each spectrum made up of 400
scans, which takes 3 min to accumulate, were averaged per IR spectrum with spectra allowed to
stabilise before a final spectrum was recorded. For spectroelectrochemical measurements, a
home-made spectroelectrochemical cell was used with a three-electrode configuration. Control of
the working electrode was achieved using a Metrohm µAutolabIII potentiostat. A PTFE-coated O-
ring was used to seal the spectroelectrochemical cell on top of the coated prism, forming a
working electrode with a geometric area of ca. 0.79 cm2. The surface area of the SEIRA gold
films were determined electrochemically (vide infra). Spectra were evaluated using OPUS 7.0
software (Bruker).
The electrochemical stability window of the electrografted interfaces on gold and me-ATO were
determined by applying increasingly anodic or cathodic potential steps for a set amount of time
and then recording IR adsorption/SEIRA spectra after each application. These spectra were
recorded under a constant potential to avoid charge-induced differences in the orientation of the
organic molecules at the interface due to fluctuating open circuit potentials, or, in the case of me-
ATO, changes in the transparency of the film under different potentials. Desorption
measurements on modified gold SEIRA electrodes were carried out in acetonitrile (0.1 M TBAF)
using potential application steps of 15 s. Desorption measurements on me-ATO were carried out
in both pH 7 PB buffer (0.1 M) and acetonitrile (0.1 M TBAP) using potential application steps of
120 s. In all cases, separate samples were prepared and measured for desorption measurements
in each cathodic and anodic direction and in different media.
60
3.6.2 Density functional theory
To ease the assignment of the measured SEIRA spectra, theoretical IR spectra of benzanilide-
Au, nitrobenzene-Au, nitrothiophenol-Au, aminobenzene-Au, aminothiophenol-Au, 4-mercapto-N-
phenylquinone-Au and nitrosobenzene-Au were calculated using density functional theory (DFT)
in vacuum. These DFT calculations were performed by Dr. Jacek Kozuch (formerly from the
group of Prof. Peter Hildebrandt). The thiol or benzene hydrogen was substituted by an Au atom
to account for structural changes upon binding to the Au surface. Geometry optimisation and
vibrational analysis were performed on the BP86 level of theory using Gaussian
09.{FormattingCitation} For C, H, N, and O atoms the 6-31g* basis set was chosen. For the
heavier S the TZVP basis set, and for Au LanL2DZ (using a pseudo core potential) were
employed.{FormattingCitation} Geometry optimisations were performed using the keywords
“opt=tight” and “int=ultrafine” before calculating the vibrational frequencies. Experimental
vibrational frequencies were assigned by analyzing the potential energy distribution of the normal
modes obtained from the DFT calculations.
To assign the measured ATR-IR spectra on Im-interfaces on me-ATO, DFT calculations
were performed in vacuum. These DFT calculations were performed by Robert Götz (from the
group of Prof. Dr. Inez Weidinger, Technische Universität Dresden/Technische Universität Berlin.
A phenyl-imidazole species bound to an Sn(OH)3 cluster via a Sn-O-C bond was used to account
for structural changes upon binding to the ATO surface. Geometry optimisation and vibrational
analysis were performed using the BP86 level of theory through Gaussian 09.131–133 For C, H,
N, and O atoms the 6-31g* basis set was employed, while for Sn the LanL2DZ (using a
pseudo core potential) was employed.134–136 Geometry optimisations were performed using the
keywords “opt=tight” and “int=ultrafine” before calculating the vibrational frequencies.
3.6.3 X-ray photoelectron spectroscopy
X-ray photoelectron spectroscopy (XPS) was carried out using a Thermo Scientific K-Alpha
instrument with monochromatic Al-Kα radiation at 1486.6 eV on an area of 400 µm2. All spectra
were charge corrected relative to the C-C component of the C 1s fitted spectra (285 eV). Survey
scans were recorded with a resolution of 3 eV, while high resolution scans of the individual
elements were recorded with a resolution of 5 eV. All spectra were charge corrected relative to
the C-C component of the C 1s fitted spectra (285 eV). Spectra were fitted using the CasaXPS
software (2.3.16) with mixed 30%/70% Gaussian/Lorentzian profiles and a Shirley background. N
1s spectra were fitted without any constraints. O 1s spectrum of the unmodified me-ATO were
fitted using a Sb 3d5/2 spin-orbit component with a constrained area and peak position (9.39 eV
separation from the Sb 3d3/2 component). The modified me-ATO O 1s spectrum was fitted using
the same Sb 3d3/2/5/2 spin-orbit separation and the ratio between the bulk lattice oxygen
61
component and the Sb 3d5/2 was constrained to the same ratio determined for the unmodified
me-ATO.
3.6.4 Resonance Raman
Resonance Raman measurements of the FePOMe species were performed by Robert Götz,
while measurements of the CoPOH species were performed by Dr. Patrycja Kielb (formerly from
the group of Prof. Dr. Inez Weidinger). Measurements were carried out using a confocal Raman
spectrometer (LabRam HR-800, Jobin Yvon) equipped with a liquid N2 cooled CCD Symphony
detector. The samples were excited using the 413 nm line of the Kr ion continuous wave laser
and the laser beam was focussed on the sample using a 20´ Olympus objective. The scattered
light was collected in a 180˚ back-scattering geometry. The laser power was adjusted for each
experiment, with a final laser power of between 1-2 MW. Spectra were calibrated against the
Raman shift of Hg, which is positioned at 435.834 nm. Spectra were collected with a spectral
resolution of 1-2 cm-1. Spectroelectrochemical measurements were performed using the home-
made PTFE cell mentioned previously with a three-electrode configuration. The cell consisted of
a Ag/AgCl 3M KCl (Leak-Free, Warner Instruments) reference electrode, a platinum wire counter
electrode a me-ATO film on ITO-coated glass as a working electrode. Control of the working
electrode was achieved using a Metrohm potentiostat. The cell was rotated by means of a
rotating table to avoid photoinduced processes, such as photodegradation or photoreduction.
Measurements were performed at room temperature. Spectra were evaluated using the home-
made qpipsi software, with component fit analysis used to assign porphyrin vibrational modes.
Peaks were fitted with Lorentzian curves.
3.6.5 UV-Vis
UV-vis spectroelectrochemical measurements were performed in a UV-Vis spectrometer (Agilent)
using a home-made spectroelectrochemical cell with a three-electrode configuration. The cell
consisted of a 3,500 uL optical glass cuvette, a Ag/AgCl 3M KCl (Leak-Free, Warner Instruments)
reference electrode, a platinum wire counter electrode and a me-ATO film on ITO-coated glass
as a working electrode. PTFE film was used to obtain a measureable surface area. Control of the
working electrode was achieved using a Metrohm µAutolabIII potentiostat.
62
7 Spectroelectrochemcial investigation into the electrochemical grafting of diazonium salts on electrodes for the
immobilisation of biological and molecular catalysts
Chapter 4
Spectroelectrochemcial investigation into the
electrochemical grafting of diazonium salts on
electrodes for the immobilisation of biological and
molecular catalysts
4.1 Introduction
In electrocatalysis, the interface between redox catalyst and electrode surface plays a critical role
in terms of catalytic activity, stability and hence applicability in (bio-) technological applications.
Despite its higher cost, gold still finds many applications as an electrode material due to its
inherent inertness, low toxicity and applicability in a large number of spectroscopic and analytical
techniques, including surface plasmon resonance spectroscopy, surface enhanced infrared and
(resonance) Raman spectroscopies.30 In order to produce functional electrocatalytic devices
based on molecular or enzymatic catalysts and take advantage of their particular activities or
study the structure or the mechanism by which these species function (e.g. by using
spectroscopic techniques) it is necessary to attach them in a stable and direct electron transfer
(DET) favourable configuration onto an electrode surface. In this context, thiols and other
organosulfur compound self-assembled monolayers (SAMs) and are widely utilised in order to
functionalise gold surfaces. However, SAMs suffer from poor chemical and electrochemical
stabilities and, therefore, have limited application in electrochemical devices.137 To overcome this
limitation, much effort has been invested into finding alternative methods for modifying gold
electrode materials. One very promising method is based on diazonium chemistry, which is
applicable not only on gold but also on a broad range of other materials, as mentioned previously
in (5.3.7, page 26).45
Electrografted diazonium interfaces take advantage of a strong covalent carbon-electrode bond
formed between the diazonium bearing organic molecule and the substrate interface; a bond that
has been shown to be very resistant to oxidation. Consequently, such interfaces were shown to
exhibit excellent chemical, electrochemical and physical stabilities on a wide range of materials,
including gold and carbon. However, a direct proof by any spectroscopic technique was not
provided so far and the true electrochemical stability window of such diazonium-derived
interfaces on gold remains unknown.62,138–140
The potential to immobilise enzymes on electrografted diazonium interfaces via electrostatic
adsorption or covalent coupling has been demonstrated on carbon materials as well as on
63
gold,141–147 and many examples exist where diazonium chemistry has been used to attach
molecular complexes and catalysts to carbon materials.71,75,148–152 One main disadvantage of
diazonium chemistry is, however, the tendency to form thick multilayers due to branching
polymerisation reactions, especially on gold. This often results in interfaces with insufficient
conductivity and, hence, slow or suppressed heterogeneous electron transfer kinetics between
the immobilised redox species and the electrode. Small changes in the deposition method or
conditions can lead to significantly different outcomes in terms of interface structure, electron
transfer properties and hence applicability, as was demonstrated by Brozik and co-workers for
pyrroloquinoline quinone immobilised on gold.153 De Lacey and co-workers obtained catalytic
current for a laccase immobilised on gold by electrochemically depositing sub-monolayers using
diazonium salts and filling in the areas in-between with thiol SAMs. Such mixed interfaces will,
however, only be as stable as the thiol component. Alternatively, enzymes or molecules may
themselves be functionalised with diazonium groups and directly electrografted onto an electrode
surface, as demonstrated for diazonium-modified horseradish peroxidase grafted onto gold.76 For
large molecules like enzymes, this approach leads to a distribution of orientations on the
electrode surface due to the lack of site-specific modification, and furthermore, the harsh
modification conditions may not be suitable for many enzymes.76 Correct orientation of the redox
active enzyme at the electrode surface is crucial for ensuring high direct electron transfer rates
between the electrode and the redox centre of the protein, thus eliminating the need for redox
mediators. Orientation control can in general be achieved through the correct selection of surface
functionality and adsorption conditions (such as pH etc.), as shown for the electrostatically
adsorbed bilirubin oxidase on SAM-coated gold electrodes.154
64
Scheme 40
Overview of the different reactions and modification approaches used in this chapter: (a) shows the formation of
diazonium- and thiol-derived interfaces on gold and their subsequent electrochemically-induced desorption, (b)
shows the formation nitro-functionalised electrografted diazonium interfaces on gold with or without the addition of
a radical scavenger to moderate interface growth, (c) shows the formation of amino-functionalised interfaces from
the nitro-functionalised electrografted diazonium interfaces via electrochemical reduction for the immobilisation of
membrane-bound hydrogenase (MBH), and (d) shows the immobilisation of a molecular catalyst on gold using
similar diazonium chemistry.
The first part of this chapter will focus on the electrochemical stability of electrografted diazonium
interfaces on gold electrodes and will compare them to thiol SAMs (Scheme 40a). To do so, a
spectroelectrochemical method using a combination of electrochemical polarisation and in situ
SEIRA spectroscopy was used. This technique is featured throughout this chapter and allows one
65
to follow the formation of interfaces and observe structural changes, e.g. under different chemical
or electrochemical conditions, and is ideally suited to determine the electrochemical stability of
the electrografted interfaces. Amine-terminated interfaces were used in this chapter as it has
been shown that amine-terminated aliphatic SAMs facilitate the immobilisation of the oxygen-
tolerant [NiFe] membrane bound hydrogenase (MBH), suitable for biotechnological fuel cell
applications, in a direct-electron transfer configuration (Scheme 1c).85,155,156 4-aminothiophenol
(4-ATP) is structurally more similar to diazonium salts than aliphatic amine thiols due to the
presence of the phenyl ring and is therefore used for better comparison. Amino-terminated
diazonium-derived interfaces are usually deposited by first functionalising the surface with a 4-
nitrobenzene diazonium salt, such as 4-nitrobenzene tetrafluoroborate (4-NBD), and then
electrochemically reducing the deposited nitro-groups to amino-groups (Scheme 1c). Isolating or
synthesising diazonium salts from p-phenylenediamine is difficult due to the presence of two
phenyl-amine groups, both of which may be diazotised. As such, a benzanilide diazonium salt 4-
benzamidobenzenediazonium tetrafluoroborate (4-BABD) was synthesised and used instead for
the corresponding electrochemical stability study (Schemes 1a, 2a).
In the second part of this chapter, the effect of adding a radical scavenger during electrochemical
grafting of gold with 4-NBD is studied (Scheme 40b). It was shown by Breton and co-workers that
the addition of a radical scavenger such as 2,2-diphenyl-1-picrylhydrazyl (DPPH) during the
reduction of 4-NBD on glassy carbon (GC) inhibits the polymerisation mechanism by which
multilayers are formed, and thereby promotes monolayer formation.157 This method was deployed
on gold in this thesis in an attempt to obtain thinner interfaces that facilitate the direct electron
transfer between an immobilised MBH enzyme and the electrode surface (Scheme 1c). As
previously mentioned, this enzyme can be used in the reversible cleavage of dihydrogen with
nearly no overpotential, i.e. close to thermodynamic potential, and is, therefore, of interest in
applications involving the hydrogen oxidation reaction (HOR) as well as the hydrogen evolution
reaction (HER). In situ SEIRA measurements were carried out in combination with protein film
voltammetry (PFV), allowing structure-function studies to be carried out in order to determine the
best modification conditions for achieving the highest catalytic currents from the immobilised
enzyme.
In the last part of this chapter, diazonium salts are assessed for their suitability for immobilising
molecular complexes and catalysts on gold electrodes (Scheme 40d). In situ SEIRA
spectroelectrochemistry can be used to yield important structural insights into the mechanisms by
which surface immobilised molecular and bio-catalysts function. A number of electrode materials
used in the field of electrocatalysis are not IR transparent, such as carbonaceous materials
(although studies have been carried out on carbon black and carbon nanotubes in ATR
mode158,159), or lack the surface enhancement properties of nanostructured gold. Thiol-based
66
SAMs are not suitable for immobilising catalysts used in a number of important reactions, such as
the CO2 reduction reaction (COR), the HER and the oxygen evolution reaction (OER), due to
their electrochemical instability in the potential ranges that these reactions take place in. Cathodic
electrochemical desorption typically takes place at 0 ± 0.25 V (vs RHE) in 0.1 M KOH.160–162 A
bipyridine-containing ligand bpy-diazo was synthesised for this purpose as bipyridine ligands are
found in many molecular complexes that are relevant for electrocatalysis. In particular, carbonyl
bipyridyl complexes of metals such as Re163 and Ru164 have been shown to be excellent
molecular catalysts for the reduction of CO2. Recently, it was shown that manganese carbonyl
bipyridyl complexes are also excellent catalysts for the selective electrochemical reduction of CO2
to CO under mild conditions,165 which generated significant interest given that Mn is an earth-
abundant element, with numerous new derivatives and mechanistic studies being published.166–
171 Infrared spectroelectrochemical methods are particularly useful in this regard given the IR-
activity of the carbonyl vibrational modes and those of intermediate species formed during the
catalytic cycle. Kubiak and co-workers have carried out a number of studies using IR
spectroelectrochemical methods to study such homogenous catalytic species in solution.166,169,171
More recently, manganese carbonyl bipyridyl complexes were immobilised in Nafion/multi-walled
carbon nanotube (MWCNT) membranes by Cowan and co-workers167,172, as well as on
mesoporous TiO2173 and carbon nanotubes (CNTs)159 electrodes by Reisner and co-workers. In
the latter two examples, the bipyridine ligands were modified with phosphonic acid or pyrene
moietiesm, respectively, to allow immobilisation via chemisorption or p-p stacking. In this thesis,
using the previously mentioned bpy-diazo linker, Mn(bpy)(CO)3Br was immobilised on gold
starting from Mn(CO)5Br via different methods, studied using SEIRA spectroscopy and assessed
for catalytic activity.
4.2 Electrochemical reduction of diazonium salts on gold
A freshly prepared, bare gold SEIRA electrode coated on a Si-prism was incubated with 4-BABD
within the spectroelectrochemical cell in acetonitrile (0.1 M TBAF), and, upon incubation, SEIRA
spectroscopy showed that 4-BABD species are adsorbed onto the electrode surface from solution
(Figure 6). This is indicated by the appearance of the corresponding v(N≡N) stretching vibration
band at 2262 cm-1 and bands corresponding to the benzanilide moiety, as summarised in
Table 1. Band assignments were made with the assistance of DFT calculations (see Appendix
1a) and refer in the allocation given here to the dominant vibrational modes.
67
Scheme 41
Formation of diazonium and SAM-derived interfaces on gold. These are formed by (a) electrochemical reduction
of 4-BABD and (b) spontaneous adsorption of 4-ATP.
Figure 6
In situ SEIRA spectra of a gold thin film electrode incubated with 1 mM 4-BABD in acetonitrile
(0.1 M TBAF) before (black trace) and after application (green trace) of two CVs at a scan rate of
50 mV/s between +0.12 and -0.88 V (vs Fc/Fc+), starting in the cathodic direction (see inset for
CVs). The reference spectrum was recorded for unmodified gold in acetonitrile (0.1 M TBAF).
The negative bands corresponding to the displaced acetonitrile solvent molecules at the interface
are asterisked.
As seen in Figure 6 (inset), electrochemical reduction resulted in the passivation of the electrode.
This is indicated by a sharp decrease in reductive current upon subsequent cycling and by the
growth of the benzanilide band intensities, along with the simultaneous disappearance of the
v(N≡N) band and two negative bands (asterisked) that surround it, corresponding to acetonitrile
solvent displaced from the interface. After electrochemical reduction, non-covalently bound
species were removed by cycling the gold electrode several times to mild anodic potentials in
fresh acetonitrile (0.1 M TBAF). For comparison, thiol-derived SAMs were deposited via the
spontaneous adsorption of 4-ATP. The successful formation of these interfaces is seen in the
SEIRA spectra (Figure 7c and d) and is indicated by the appearance of characteristic absorption
bands corresponding to the aminothiophenol moiety, as summarised in Table 2.
68
Table 2
Assignment of the major absorption bands to the characteristic dominating vibrational modes made with the
assistance of DFT calculations for Au covering interfaces deposited by electrochemical reduction of 1 mM 4-
BABD or by spontaneous adsorption of 1 mM 4-ATP on gold electrodes.
Sample
Band position (cm-1)
Band assignment
Dominant vibration
4-BABD
1493
d
(C-H)+v(C=C)
ar
C-H bending coupled
with a C=C aromatic
stretching
1525
d
(N-H)
N-H bending vibration
1598
v(C=C)
ar
+
d
(N-H)
C=C aromatic stretching
coupled with a N-H
bending
1672
v(C=O)
C=O stretching
2262
v(N≡N)
N≡N stretching
4-ATP
1488
d
(C-H)+v(C=C)
ar
C-H bending coupled
with a C=C aromatic
stretching
1593
v(C=C)ar
C=C aromatic stretching
1627
d
(NH2)
NH
2
bending
3370
v
s
(NH
2
)
NH
2
symmetric
stretching
4.3 Electrochemical stability of electrografted diazonium interfaces on gold
In order to assess the anodic and cathodic stability of the electrografted diazonium interfaces and
the SAMs, SEIRA spectra were recorded under a constant potential of -0.58 V (vs Fc/Fc+) after
polarising the respective modified electrodes at different cathodic and anodic potentials (in 100
mV steps) for 15 s (Figure 7a – d) to avoid charge-induced differences in the orientation of the
organic molecules at the interface due to fluctuating open circuit potentials. Plots of the SEIRA
band intensities against the applied potentials allow one to follow potential-induced changes in
the interface structure and composition and thereby assess their electrochemical stability range
(see Appendix 2a - d).
69
Figure 7
In situ SEIRA spectra recorded at -0.58 V (vs Fc/Fc+) (0 V (vs Ag/AgCl 3M KCl)) in acetonitrile (0.1 M TBAF)
following the interfacial desorption induced by electrode polarisation at different potentials: behaviour of an
electrografted 4-BABD diazonium interface under (a) cathodic and (b) anodic polarisation and behaviour of a 4-
ATP SAM-based interface under (c) cathodic and (d) anodic polarisation (asterisked bands assigned to the
dimerised 4-mercapto-N-phenylquinone monoimine (NPQM) species). (e) Normalised v(C=C)ar absorption band
intensities at 1598 and 1593 cm-1 plotted against polarisation potential for spectra (a) (b), and (c) (d), respectively.
In order to better compare the difference in the electrochemical stabilities of the Au¾C and
Au¾S bound interfaces, the normalised SEIRA intensities of an aromatic C=C stretching
vibration, common to both molecular systems, are plotted against the applied potential (Figure
7e). As one can see, the electrografted diazonium interface is stable over a much broader
potential window in acetonitrile than the SAM-derived interface. The electrografted diazonium
interface is stable up to +0.8 V (vs Fc/Fc+), above which gold oxidation occurs, and down to -2.2
V (vs Fc/Fc+), after which a sharp decrease in band intensity is observed due to reductive
70
desorption. The small decrease in band intensity observed between -0.6 and 0.8 V (vs Fc/Fc+) in
Figure 3 may be due to potential-induced desorption of remaining physisorbed species that resist
post-modification rinsing and electrochemical treatments; such species could arise from
homogeneous coupling between reactive species in solution. It has been suggested by others
that this accounts for such a loss of material observed during sonication or refluxing.139 In
contrast, the SAM is only stable up to around 0 V (vs Fc/Fc+), at which point oxidative desorption
begins to occur. In acetonitrile, there is no spectroscopic evidence of oxidised species such as
sulfonites (S=O vibration), as is the case for oxidative desorption in aqueous solvents.162 Two
new absorption bands (asterisked) at 1690 cm-1 and 1514 cm-1 begin to appear at around 0.4 V
(vs Fc/Fc+) and may be attributed to a dimerised species, 4′-mercapto-N-phenylquinone
monoimine (NPQM), that is known to be formed as a result of head-to-tail coupling between
adjacent oxidised aminothiophenol species.174–176
Figure 8
Calculated spectrum of 4-mercapto-N-phenylquinone (NPQM) monoamine bound to a gold atom using DFT and
assignment of the main absorptions to the dominating, characteristic vibrational modes.
The corresponding calculated spectrum of this species bound to gold (Appendix 3) supports the
hypothesis that a dimerised species NPQM is formed, with an absorption band at 1656 cm-1
characteristic for the v(C=O) stretching vibration of the quinone, and another band at 1526 cm-1
related preferentially to a v(C=C) aromatic stretching vibration coupled with a C=N stretching
vibration of the imine bond. In the cathodic range, significant desorption of the thiolate begins to
occur at around -1.2 V (vs Fc/Fc+) and the entire desorption process takes place over a wider
potential window than that observed for the reductive desorption of the electrografted diazonium
interface. This can be explained by the reversible nature of the one-electron thiolate
desorption/re-adsorption process. The exact potential at which thiolate species are reduced
depends not only on the substrate-adsorbate interaction but also on the adsorbate-adsorbate
intermolecular interactions.161
N
Au
O
71
4.4 Effect of radical scavenger on interface formation
Figure 9
(a) SEIRA spectra recorded in acetonitrile of interfaces that were deposited on gold electrodes by electrochemical
reduction of 1 mM 4-NBD with increasing molar equivalents of added DPPH (coloured traces), as well as a SAM
deposited by spontaneous adsorption of 1 mM 4-NTP (black trace), (b) cyclic voltammograms of 1 mM
K3[Fe(CN)6] recorded at 50 mV/s in 0.1 M PB (pH 7) of the same interfaces and (c) plots showing the change in
the vs(NO2) band intensity, as well as the change in the peak separation ΔEp for the [Fe(CN)6]4-/3- couple with
increasing concentration of DPPH.
Electrochemical reduction of 1 mM 4-NBD results in the passivation of the gold surface with a
dense polymeric interface, as has previously been reported.139 SEIRA spectra were recorded
after the reduction with and without the addition of different molar equivalents of the radical
scavenger DPPH (see Figure 3(a)). The spectrum of gold modified with 4-NTP is included for
comparison. Band assignments were made with the assistance of DFT calculations (see
Appendix 1b and c) and are in line with previous literature reports.139
72
Table 3
Assignments of the most prominent absorption bands to the dominant characteristic vibrational modes were made
with the assistance of DFT calculations for interfaces deposited on gold by electrochemical reduction of 1 mM 4-
NBD or spontaneous adsorption of 1 mM 4-NTP.
Sample
Band position (cm-1)
Band assignment
Dominant vibration
4-NBD
1351
v
s
(NO
2
)
Symmetric NO
2
stretching
1526
v
as
(NO
2
)+v(C=C)
ar
Antisymmetric NO
2
stretching coupled with
a C=C aromatic
stretching
1598
v(C=C)
ar
+v
as
(NO
2
)
C=C
aromatic
stretching coupled with
an antisymmetric NO2
stretching
4-NTP
1346
v
s
(NO
2
)
Symmetric NO
2
stretching
1518
v
as
(NO
2
)+v(C=C)
ar
Antisymmetric NO
2
stretching coupled with
a C=C aromatic
stretching
1572
v(C=C)
ar
C=C
aromatic
stretching
1592
v(C=C)
ar
+v
as
(NO
2
)
C=C
aromatic
stretching coupled with
an antisymmetric NO2
stretching
Having a closer look, one can see a difference in the ratio between the vs(NO2) and the
vas(NO2)+v(C=C)ar adsorption band intensities – I[vs(NO2)]/I[vas(NO2)+v(C=C)ar] – for each
interface. This ratio is related to the way molecules arrange at the corresponding interface.
Indeed, due to the SEIRA surface selection rules, only modes which preferentially exhibit dipole
moment changes perpendicular to the surface are enhanced, indicating that, with a large value of
I[vs(NO2)]/I[vas(NO2)+v(C=C)ar], 4-NTP forms a monolayer where the molecules are arranged
uniformly. Conversely, with a value of I[vs(NO2)]/I[vas(NO2)+v(C=C)ar] of almost 1, diazonium salts
form disordered, polymeric interfaces, where the formation of a preferential orientation is not
possible.139
73
Scheme 42
Proposed mechanism by which radical scavengers inhibit polymerisation during the electrochemical grafting of
diazonium salts on electrode surfaces. For monolayer coverage, the rate of phenyl radical adsorption on the
electrode surface (ks) must be greater than the rate of radical coupling to the radical scavenger (kr), which in turn
must be greater than the rate of radical coupling to already grafted phenyl moieties (kp). The coupling product of
DPPH and the nitrophenyl radical, 2-[4-(4-nitro)phenyl]-2-phenyl-1-picrylhydrazine, was detected by Breton and
co-workers in experiments carried out using GC electrodes.157
It was shown by Breton and co-workers that the addition of a radical scavenger such as DPPH
during the reduction of 4-NBD on GC inhibits the polymerisation mechanism by which multilayers
are formed, and thereby promotes monolayer formation.157 This effect would, if occurring in the
case of the gold electrodes presented here, result in an increase in
I[vs(NO2)]/I[vas(NO2)+v(C=C)ar]. However, as seen in Figure 9a, when gold is used as substrate,
I[vs(NO2)]/I[va(NO2)+v(C=C)ar] remains roughly the same at all concentrations of DPPH, while the
band intensities decrease proportionally until at 1 equivalent of DPPH almost no material is
electrografted on the gold surface anymore. CVs of a ferricyanide electrochemical probe (see
Figure 9b) measured for the electrografted diazonium interfaces deposited in the presence of
various DPPH concentrations confirm this finding: the peak-to-peak separation (ΔEp) of the
oxidation/reduction features of the [Fe(CN)6]4-/3- (Fe2+/3+) couple decreases with increasing DPPH
concentration, indicating a substantial decrease in the charge transfer resistance RCT of the
interface, until at 1 equivalent DPPH the electrochemical response is almost the same as for an
unmodified electrode. A CV of a gold electrode modified with a 4-NTP SAM is included for
comparison.
Scheme 43
The proposed structures for interfaces electrochemically grafted from 4-NBD, with and with the addition of a
radical scavenger.
74
The behaviour observed for gold is in sharp contrast to the one observed for GC where surface
coverage and RCT reach a steady-state minimum corresponding to approximate monolayer
coverage. For such a monolayer coverage, the rate of phenyl radical adsorption on the electrode
surface (ks) must be greater than the rate of radical coupling to the radical scavenger (kr), which
in turn must be greater than the rate of radical coupling to already grafted phenyl moieties (kp),157
as illustrated in Scheme 42. In the case of gold, the obtained results indicate that the rate of
coupling with the radical scavenger is greater than with the gold surface, so that both surface
adsorption of radicals and polymerisation is suppressed, leading to less material being deposited
in a less dense, albeit still polymeric, interface, as illustrated in Scheme 43. This difference in
behaviour can be related to the lower adsorption energy of the phenyl moiety on gold (predicted
to be 24 kcal/mol) compared to greater adsorption energies on carbon materials (predicted to be
as great as 63 kcal/mol for graphene), meaning that radical adsorption is energetically favoured
on GC.177,178
Scheme 44
Chemical structure of the radical scavenger galvinoxyl.
Figure 10
SEIRA spectra recorded in acetonitrile of interfaces deposited by electrochemical reduction of 1 mM 4-NBD under
the same conditions with increasing molar equivalents of the radical scavenger galvinoxyl added.
Even though a monolayer coverage could not be achieved on gold, this method of using radical
scavenger leads to reproducible, less dense interfaces with decreased interfacial charge transfer
resistance. Trials using the less reactive radical scavenger galvinoxyl (for structure see Scheme
44)showed a similar behaviour, and comparable results as with DPPH could be achieved using
75
higher radical scavenger concentrations, as shown in Figure 10. Thus, the addition of radical
scavenger can reduce the RCT of the deposited interfaces sufficiently to allow the use of the
functionalised interfaces for the immobilisation and utilisation of electroactive species, while still
taking advantage of the interfaces superior stability.
4.5 Spontaneous adsorption of diazonium salts on gold
Figure 11
(a) SEIRA spectra recorded in acetonitrile of interfaces deposited by the spontaneous adsorption of 4-NBD from a
20 mM solution in acetonitrile, before (green trace) and after (blue trace) rinsing in water. (b) Spectrum calculated
using DFT of a nitrophenyl moiety bound to a gold atom via a terminal azo N=N bond.
It has been widely reported that diazonium salts adsorb spontaneously onto a range of materials
including GC,56,179–181 various metals52,56 and gold.52,182,183 The exact mechanism by which these
interfaces are formed on gold has not been fully elucidated yet, but is known to occur via different
routes in various media and conditions, with possible routes involving the adsorption of cationic,
radical or diazonium species at the surface of the gold, or at already grafted/adsorbed moieties.
Figure 11a displays the SEIRA spectrum of gold after incubation with a 20 mM 4-NBD solution in
acetonitrile for 60 min (at this point no further increase in adsorption intensity can be observed)
and thorough rinsing with fresh acetonitrile. Overall band intensities are low compared to the
electrochemically grafted interfaces (<5 mOD), while the v(C=C)arom, v(NO2)a and v(NO2)s bands
appear at or near the same wavenumbers. An increase in the value of
I[vs(NO2)]/I[vas(NO2)+v(C=C)ar] to 1.7 is observed, suggesting an increased ratio of monomeric
surface-bound species compared to the electrochemically grafted interfaces. A broad band of
medium intensity appears at 1396 cm-1, which is absent from the electrochemically grafted
interfaces. This value is similar to bands observed at 1377 and 1392 cm-1 by Calvo and co-
workers in IR absorption spectra of 4-carboxybenzene diazonium and 4-carboxy-2,3,5,6-
tetrafluorobenzene diazonium tetrafluoroborate salts on gold, which are claimed to be due to the
v(N=N) stretching vibration of azo-linkages, with indications given to the existence of branched
azo-bonds (i.e. to already-grafted moieties) and terminal Au-N=N- azo-bonds.184 The bands
observed here and by Calvo and co-workers appear at much lower wavenumbers than one would
76
normally expect for azo-bonds, and indeed, a calculated spectrum for Au-N=N-PhNO2 (Figure
11b) shows a band for the v(N=N) stretching vibration at 1559 cm-1 (which overlaps with the
vas(NO2) vibration at 1557 cm-1). This phenomena was explained by Calvo and co-workers as
resulting from conjugation through the azo bond.184
Scheme 45
Proposed structure of interface deposited from the spontaneous adsorption of 4-NBD on gold from an acetonitrile
solution, before and after rinsing with water.
The binding of the nitrophenyl moieties to the gold via terminal Au-N=N- azo-bonds would explain
the increase in I[vs(NO2)]/I[vas(NO2)+v(C=C)ar] compared to the electrochemically grafted
interfaces. Upon rinsing the electrode with water there was a marked decrease in intensity for all
bands and a decrease in I[vs(NO2)]/I[vas(NO2)+v(C=C)ar] from 1.7 to 1.2, which was also
accompanied by the almost-complete disappearance of the v(N=N) band at 1396 cm-1 (Figure
11a), suggesting the removal or hydrolysis of monomeric, terminal bound azo-bound species
from the gold (as illustrated in Scheme 45). Any remaining species must be bound to the surface
via Au−C bonds and, therefore, be deposited from cationic- or radical-phenyl species, which can
form as a result of the spontaneous homo- or heterolytic decomposition of diazonium species.
Palacin and co-workers show evidence from XPS of some species bound to gold via an azo-bond
deposited from solutions of diazonium salts in both water and acetonitrile.52 They propose an
interaction between a vacant 2p orbital on the terminal nitrogen of the diazonium species and the
full 5d orbitals of gold. Such a binding motif would correspond well with the observation of a band
at 1396 cm-1, with the release of electrons from the gold 5d orbitals into the vacant nitrogen 2p
orbital weakening the N=N bond and leading to a decrease in the wavenumber of the v(N=N)
band. Due to the instability of these interfaces in aqueous environments, they were deemed
unsuitable for further utilisation for the immobilisation on MBH.
77
4.6 Immobilisation of hydrogenase on electrografted diazonium interfaces
As previously mentioned, electrografted diazonium interfaces have been used to immobilise
enzymes on a range of electrode materials.141–146 It was previously shown by Heidary et al. that it
is possible to immobilise the strep-tag MBH onto amino-terminated SAMs on gold, resulting in a
compact enzyme monolayer. The enzyme does not denature upon immobilisation and allows
interfacial, preferentially direct electron transfer processes, and it keeps its catalytic functionality.
Hence, second, ternary and quaternary structures are preserved.84,85
Scheme 46
Electrochemical reduction of a nitrophenyl interface electrografted on gold and the subsequent immobilisation of
MBH under orientation control.
Nitrophenyl moieties can be reduced chemically or electrochemically in protic conditions to
aminophenyl moieties (see Scheme 46), as has been widely reported.185,186,147 Reduction of the
nitrophenyl moieties is necessary to allow the adsorption and immobilisation of the MBH onto the
surface via electrostatic and hydrogen bonding interactions between the surface of the enzyme
and the amino-terminated surface. The electrodes that were electrografted with 4-NBD and
DPPH were subsequently electrochemically reduced by cycling the electrode between 0 and -1.2
V (vs Ag/AgCl 3M) in a 0.1 M NaClO4 1:9 ethanol:water solution. SEIRA spectra were recorded
after reduction and are shown in Figure 12a. As for the previous spectra, band assignments were
made with the assistance of DFT calculations (see Appendix 1d). A pronounced decrease in the
band intensity of the vs(NO2) and vas(NO2) absorption bands is observed upon electrochemical
reduction, as previously reported by Vaz-Domínguez et al.147 The v(C=C)ar band at 1600 cm-1
remains with similar intensity, while a broad shoulder appears around 1626 cm-1 corresponding to
the
d
(NH2) band. It must be noted that the vs(NO2) adsorption at 1350 cm-1 does not vanish
completely after reduction, indicating the presence of residual NO2 within the interfaces, in line
with previous studies. Nevertheless, the contribution is negligible. In line with this conclusion, it is
also possible to see a shoulder close to the
d
(C-H)+v(C=C)ar band at 1520 cm-1, which may
correspond to the the v(N=O) stretching vibration of incompletely reduced nitrosophenyl species
(a spectrum for this species was calculated and is shown in Appendix 3). The SEIRA spectra of
the electrochemically reduced interface is similar to that of a SAM deposited from the
78
spontaneous adsorption of 4-ATP (Appendix 1e). All band intensities decrease proportionally with
the amount of DPPH added during the nitrophenyl interface formation.
Figure 12
(a) SEIRA spectra in acetonitrile of the amino functionalised interfaces deposited by the electrochemical grafting
of 1 mM 4-NBD with increasing molar equivalents of DPPH added that were subsequently electrochemically
reduced in 0.1 M NaClO4, 1:9 ethanol:water solution, (b) SEIRA spectra in buffer (pH 7, 10 mM phosphate buffer
(PB)) of MBH adsorbed onto the same amino functionalised interfaces after rinsing with buffer (inset: shows a
characteristic active site band of the enzyme, namely the v(CºO) stretching vibration of the Fe-coordinated
carbon monoxide ligand, which can be detected at 1948 cm-1 for the so called “Nir-B” state of the oxidised
enzyme) and (c) plots showing the change in the v(C=C)ar band intensities, as well as the amide I and II band
intensities, for all spectra with increasing concentration of DPPH.85
In order to evaluate the suitability of the electrografted interfaces for enzyme immobilisation and
bio-electrocatalysis, we studied the electrostatically-controlled adsorption of the MBH onto the
amino functionalised surfaces from a bulk solution at pH 7. Previous studies demonstrated that
SEIRA spectroelectrochemistry is a powerful tool for the investigation and knowledge-based
optimisation of protein immobilisation on electrode surfaces for optimally exploiting enzymatic
processes in bioelectronic devices.84,85 Additionally, it is an invaluable tool for elucidating
underlying molecular catalytic mechanisms by directly probing active site states. In case of
hydrogenases this is possible due to the absorption properties of Fe-bound CN- and CO
ligands.187 The appearance of two broad absorption bands at 1658 and 1550 cm-1 in the SEIRA
spectra (see Figure 12b) corresponding to the amide I and amide II bands of the protein
backbone proves the successful binding of the MBH to the electrografted diazonium interface.
The appearance of the band at 1948 cm-1, related to the v(CºO) stretching vibration of the CO
ligand coordinated to the [Ni-Fe] active site and characteristic for the highest oxidised ‘Nir-B’
79
state,188 indicates that also the enzyme’s active integrity is preserved upon surface
immobilisation.130,189 These results are in good agreement with the previous SEIRA study by
Heidary et al. reporting the electrostatic immobilisation of MBH onto amino-1-hexanethiol gold
modified electrodes under the same conditions. These results clearly underline the “binding”
similarity between diazonium-derived and thiol-derived amino-terminated interfaces.85 The pKa of
an amino-1-hexanethiol SAM was determined at close to 6 and is assumed to be 8% protonated
at pH 7.190 pKa values for amino functionalised surfaces derived from the aliphatic-amine
containing 4-aminobenzyldiazonium and 4-(2-aminoethyl)benzenediazonium have been
previously reported to be around 10.0 and 10.5 respectively191, while a pKa of 6.9 was determined
for 4-ATP SAMs on gold.192 We assume that the pKa values of our aminophenyl surfaces are at
least greater than 6 and therefore a significant proportion of the surface will be protonated leading
to an electrostatic binding. A decrease in MBH adsorption is observed with decreasing
aminophenyl density (i.e. when increasing amounts of scavenger are added), as indicated by a
decrease in amide band intensity (plotted in Figure 12c). It is tentatively proposed that this is a
result of decreasing electrostatic attraction, due to a decrease in protonated aminophenyl
moieties at the interface.
Figure 13
Protein film voltammograms recorded in 10 mM PB buffer at pH 5.5 saturated with H2 with a scan rate of 5 mV/s
on MBH adsorbed onto amino-functionalised interfaces (a) after and (b) before addition of the redox mediator
methylene blue. (c) Plots showing the change in the maximum electrocatalytic current for H2 oxidation reaction
(HOR) due to direct electron transfer and the corresponding MBH amide I and II band intensities with increasing
concentration of DPPH added during interface formation.
80
Cyclic voltammograms of the samples incubated with MBH were conducted in a H2 saturated,
aqueous PB buffer at pH 5.5 and are shown in Figure 13a. Activity has previously been shown to
be highest at pH 5.5.193 The sigmoidal shape of the CV traces indicates the electrocatalytic
oxidation of H2 into protons (H+) and electrons as a result of direct electron transfer (DET)
between the catalytic redox centre of the MBH and the electrode. The interface remains stable in
the potential range probed. Similar overpotentials to those reported by Heidary et al. for amino-
terminated SAM derived surfaces are observed.85 In Heidary’s study, a peak current of 1.8
µA/cm2 was obtained,85 whereas in the present study, peak currents between 1.3 and 5.4 µA/cm2
were obtained. These differences must result from the difference in surface protonation (e.g.
resulting from different pKa values), interface thickness (around 1 nm for a SAM, whereas up to
2.5 nm can be obtained for the polymeric interface139), and lastly differences in charge transfer
resistance. The lower currents of around 1.3/1.4 µA/cm2 observed for the thick interfaces
obtained with 0 mM or 0.25 molar equivalent of DPPH reflect the higher charge-transfer
resistance of these interfaces, despite them having relatively more adsorbed MBH catalyst (as
revealed by SEIRA). Typically tunnelling can take place across distances of 2-3 nm.194,195 An
increase in current to 5.6 µA/cm2 with 0.375 equivalent DPPH is observed, despite a decrease in
adsorbed MBH (again, as revealed by SEIRA). This either indicates a decrease in the charge
transfer resistance of the interface, or a decrease in the distance between the distal cluster of the
electron transfer conduit of the adsorbed MBH and the electrode surface, or a combination of
both. The current then decreases to 3 µA/cm2 with 0.5 equivalent DPPH, despite the fall in the
charge transfer resistance of the interface, which corresponds to a decrease in MBH adsorption
(again, as revealed by SEIRA). The change in current and MBH adsorption with DPPH
concentration is plotted in Figure 13c. The results shown here suggest that there is a trade-off
between coverage of the electrode surface (and thus the adsorption of MBH) and the charge-
transfer resistance off the interface (and thus the electrochemical accessibility of the adsorbed
species).
Addition of the redox mediator methylene blue results in an increase in current as a result of
mediated electron transfer (MET), as shown for the interfaces deposited with 0 and 0.375
equivalents of DPPH in Figure 13b. The proportional increase observed is much greater than that
observed on by Heidary on the amino-terminated SAM-on gold, and might be attributed to the
polymeric nature of the interface, which results in a substantial number of enzyme molecules
immobilised in greater distance from the electrode surface.85
4.7 Immobilisation of molecular catalysts on gold using electrografting of diazonium salts
A freshly prepared, bare nanostructured gold SEIRA electrode coated on a Si-prism was
incubated with bpy-diazo in acetonitrile within the spectroelectrochemical cell, and, upon
incubation, SEIRA spectroscopy was used to observe the adsorption of the 4-bpy species onto
81
the electrodes surface (Figure 14b, pink trace). This is indicated by the appearance of
characteristic absorption bands corresponding to the 4-([2,2’-Bipyridine]-4-carboxamido)benzene
moiety, as summarised in Table 1. The band assignment was made with aid of DFT calculations
(for more details, see Appendix 4). As can be seen in Figure 14a, electrochemical reduction
resulted in the passivation of the electrode. This is indicated by a sharp decrease in reductive
current upon subsequent cycling and by the growth of the related 4-([2,2’-Bipyridine]-4-
carboxamido)benzene band intensities displayed in Figure 14b. The growth in SEIRA
spectroscopic band intensities decreases with each consecutive CV due to passivation of the
electrode. The electrode was subsequently rinsed in acetonitrile and non-covalently bound
species were removed by cycling the gold electrode several times to mild anodic potentials in
fresh TBAF, resulting in the spectrum shown in Figure 14c.
Scheme 47
The formation of manganese carbonyl bipyrydyl interfaces via the electrochemical grafting of bpy-diazo via a ‘pre-
coordination’ or ‘post-coordination’ method.
82
Figure 14
(a) CVs showing the electrochemical reduction of a 1 mM solution of bpy-diazo in acetonitrile (0.1 M TBAF) using
the gold SEIRA electrode as a working electrode. Scan rate = 50 mV/s. (b) The corresponding SEIRA spectra of
the surface species on gold in a 1mM bpy-diazo soluton in acetonitrile before electrochemical grafting (pink
trace), and after each consecutive CV/reduction (lilac to teal traces). The inset plots the absorbance of the
d
(N-
H)+v(C=C)ar band at 1531 cm-1 after each CV. (c) SEIRA spectrum of the electrode after modification and rinsing
in acetonitrile.
Table 4
Major band assignments made with the assistance of DFT calculations for interfaces deposited by
electrochemical grafting of 1 mM bpy-diazo on gold electrodes.
Sample
Band position (cm-1)
Band assignment
Dominant vibrational modes
bpy-diazo
1325
v(C=C)
ar
C-H bending coupled with a C=C
aromatic stretching
1410
d
(C-H)
C-H bending vibration of the amide
1459
d
(C-H)
bpy
C-H bending vibration of the first
bipyridine ring
1480
d
(C-H)
bpy
C-H bending vibration of the
second bipyridine ring
1531
d
(N-H)+v(C=C)
ar
N-H bending vibration of the amide
coupled with the C=C aromatic
stretching of the phenyl ring
1554
v(C=C)
bpy
+v(C=N)
bpy
C=C and C=N stretching of the first
bipyridine ring
1597
v(C=C)
ar
+
d
(N-H)
C=C aromatic stretching of the
phenyl ring coupled with the N-H
bending vibration of the amide
1681
v(C=O)
bpy
+
d
(N-H)
C=O stretching of the carbonyl
coupled with the N-
H bending
vibration of the amide
83
To determine the electrochemical window in which the electrografted bpy-interfaces on gold is
stable, modified electrodes were cycled to cathodic and anodic potentials, as shown in Figure
15a, and SEIRA spectra were recorded after each cycle. Difference spectra were calculated (with
respect to the initially deposited interface, Figure 15b) and exhibited little change for electrodes
cycled to 0.76 V (vs Fc/Fc+), above which gold oxidation occurs, while band intensities for the 4-
([2,2’-Bipyridine]-4-carboxamido)benzene moiety begin to decrease (as indicated by the negative
bands in the difference spectra) when the electrode was cycled to potentials of -2.14 V (vs
Fc/Fc+) and lower. The thereby determined electrochemical stability potential window
corresponds well to the potential window of -2.2 to +0.8 V (vs Fc/Fc+) as determined for interfaces
electrografted on gold from 4-BABD in section 7.2.
Figure 15
(a) CVs acetonitrile (0.1 M TBAF) under Ar bubbling showing the electrochemical stability window of interfaces
electrografted from bpy-diazo on gold. Scan rate = 50 mV/s. (b) SEIRA difference spectra of the interfaces on
gold calculated (with respect to the initially deposited interface) after consecutive CVs to increasingly cathodic
potentials showing the removal of bpy species from the surface of the electrode.
As previously mentioned, manganese carbonyl bipyridyl complexes are excellent catalysts for the
selective electrochemical reduction of CO2 to CO under mild conditions.165–171 A gold SEIRA
electrode was electrochemically grafted with bpy-diazo by applying 2 CVs and then incubated
with a 1 mM Mn(CO)5Br solution in acetonitrile for 2 hours inside the spectroelectrochemical cell
and subsequently rinsed repeatedly with further acetonitrile (hereby named the ‘post-
coordination’ method, Scheme 47a). Two prominent bands appear at 2043 and 1953 cm-1, with a
shoulder band at around 1927 cm-1, in the SEIRA spectrum of the electrode after incubation,
which are in the spectral region expected for carbonyl v(C≡O) stretching vibrations. Furthermore,
changes are visible in the lower wavenumber region containing bands belonging to vibrational
modes of the 4-([2,2’-Bipyridine]-4-carboxamido)benzene moiety, thereby suggesting that a
coordination of Mn species to bipyridine groups on the surface of the electrode has occurred.
Literature values for the v(C≡O) bands in Mn(CO)5Br are given at 2138, 2052 and 2007 cm-1 196
and are not observed here. Literature values for the v(C≡O) bands of fac-[MnBr(2,2’-
bipyridine)(CO)3 in acetonitrile with 0.1 M TBAF are at 2028, 1933 and 1923 cm-1,165 while values
of 2030, 1946 and 1930 cm-1 were obtained for fac-[MnBr(4,4’-bis(phosphonic acid)-2,2’-
84
bipyridine)(CO)3 immobilised on mesoporous TiO2.173 The higher wavenumbers reported here
may be explained by the electron-withdrawing nature of the bipyridine amide-substituent, which
reduces the electron density at the Mn metal centre and thus reduces backbonding of Mn d-
orbital electrons into the antibonding π* orbitals of the CO ligands, leading to stronger C≡O bond
and higher energy stretching bands. This would also explain the higher wavenumbers compared
to fac-[MnBr(2,2’-bipyridine)(CO)3 for fac-[MnBr(4,4’-bis(phosphonic acid)-2,2’-bipyridine)(CO)3,
which also contains electron-withdrawing phosphonic acid substituents.
Figure 16
SEIRA spectra (a) of the electrografted bipyridine interface on gold in acetonitrile (0.1 M TBAF) before (teal trace)
and after incubation (pink tail) with a 1 mM solution of Mn(CO)5Br in acetonitrile for 2 hrs (‘post-coordination’
method), including a difference spectrum (dotted pink trace), and (c) of the Mn-bipyridine interfaces electrografted
on gold via both the ‘post-coordination’ and ‘pre-coordination’ methods.
Mn-bpy interfaces were also electrografted on gold SEIRA electrodes directly from pre-mixed 1
mM bpy-diazo and 1 mM Mn(CO)5Br solutions (hereby named the ‘pre-coordination’ method,
Scheme 47b). A gold electrode was electrochemically grafted with the Mn(CO)5Br bpy-diazo
solution by applying 2 CVs and rinsed repeatedly in acetonitrile. Carbonyl v(C≡O) bands appear
in the SEIRA spectrum of the electrode after surface modification, albeit at slightly shifted
wavenumbers to those obtained using the ‘post-coordination’ method, in addition to bands
corresponding to the 4-([2,2’-Bipyridine]-4-carboxamido)benzene moiety, indicating that the ‘pre-
coordination’ was successful. v(C≡O) bands appear at 2050 and 1964 cm-1, with a less prominent
shoulder band compared to that observed for the ‘post-coordination’ method. Overall the intensity
ratio between the bands corresponding to the 4-([2,2’-Bipyridine]-4-carboxamido)benzene moiety
and the v(C≡O) bands is less, which suggests that ‘pre-coordination results in less coordinated
Mn species.
85
Figure 17
CVs in acetonitrile (0.1 M TBAF) of (a) unmodified gold, and gold electrograftd with a Mn-bipyridine interface
using the ‘pre-coordination’ method, under Ar bubbling, (b) unmodified gold, and gold electrografted with a Mn-
bipyridine interface using the ‘pre-coordination’ method, under CO2 bubbling, and CVs of (c) unmodified gold and
gold electrografted with only a bipyridine interface, under CO2 bubbling. Scan rate = 50 mV/s.
CVs in acetonitrile (0.1 M TBAF) under Ar bubbling of a Mn-bipyridine interface electrografted on
gold via the ‘pre-coordination’ method (Figure 17a, teal trace) show a reversible peak at around -
1.3 V (vs Fc/Fc+). It is difficult to determine if this peak is related to the reduction of MnI to Mn0
species, as a CV of the electrode before electrografting also reveals an irreversible feature
around the same potential (Figure 17a, light teal trace). However, the reversible peak does
become irreversible in CO2, as reported for Mn carbonyl bipyrydyl species (Figure 17b), and is
absent in CVs of a gold electrode electrografted with only a bipyridine interface (Figure 17c). It
should be noted that it was not possible to obtain gold SEIRA electrodes that are completely
featureless in CVs in acetonitrile, regardless of how many cleaning cycles were applied after
electroless deposition on the gold. In either case, there was a small shift in onset potential under
CO2 bubbling for the Mn-bipyridine species, compared to the unmodified gold, which becomes
active for CO2 reduction around -1.8 V (vs Fc/Fc+). This shift may also be a result of the solvent
adsorbing water from the atmosphere during the course of the experiment. Passivation of the
gold surface does suppress reduction of CO2 at the gold surface to a large extent, as can be
86
seen in the CVs of gold under CO2 bubbling before and after electrografting with a bipyridine
interface i.e. without Mn (Figure 17c).
Scheme 48
Proposed catalytic mechanism for the reduction for CO2 using a Mn carbonyl bipyrydyl catalyst with a bulky
bipirydyl ligand (in this case mesby (6,6′-dimesityl-2,2′-bipyridine)), which eliminates dimerisation at the Mn
centre. Reprinted with permission from 169, copyright 2014 American Chemical Society.
It has been shown that upon initial one-electron reduction at around -1.6 V (vs Fc/Fc+),
manganese carbonyl bipyrydyl species form dimeric species, which can be further reduced to
form doubly-reduced monomeric species.165 Reisner and co-workers suggest that the low
overpotential obtained for electrocatalytic CO2 reduction using fac-[MnBr(4,4’-bis(phosphonic
acid)-2,2’-bipyridine)(CO)3 immobilised on TiO2 is due to the temporary desorption of the
phosphonic-acid bound species from the surface, which allows the formation of dimeric species
that, in the presence of water, react directly with CO2 to form CO.173 In the case presented here, it
is not expected that the Mn-bpy bound to the gold will show any lability due to the strength of the
rigid Au-C bond at the interface (the electrografted diazonium interfaces on gold do not show the
same flexibility as, say, thiol SAMs do). As such, further reduction of the monomeric manganese
carbonyl bipyrydyl species to the catalytically active, doubly-reduced monomeric species would
need to take place at lower potentials, which would in turn result in a larger overpotential. A study
by Kubiak and co-workers for a manganese carbonyl bipyrydyl species in solution with bulky
substituents on the bipyridine that eliminate dimerisation at the Mn centre shows that the doubly-
reduced Mn species forms at around -2 V (vs Fc/Fc+)169, which is close to the stability limit of the
interface. Furthermore, the formation of thick, branched, polymeric interfaces may have resulted
in most of the Mn centres being redox inactive, which would explain the absence of defined redox
peaks for the Mn0/+1 transition. For these reasons, further spectroelectrochemical measurements
were not carried out.
87
4.8 Conclusions
In situ SEIRA spectroscopy was used to provide an insight into the formation and structure of
interfaces electrografted on gold electrodes via electrochemical reduction of diazonium salts,
while in situ spectroelectrochemical investigations allowed the determination of the
electrochemical stability window of such interfaces, which was determined to be 3.0 V wide in
acetonitrile. This is the first time that spectroscopic methods have been used in this way to
determine the electrochemical stability window of electrografted diazonium interfaces and this
results in larger sets of information than can be obtained using conventional electrochemical
methods (e.g. regarding structure). The superior stability of the electrografted diazonium
interfaces over thiol-derived SAMs demonstrated herein emphasises their potential for use in
electrochemical devices.
A radical scavenger was deployed during electrochemical grafting of the 4-nitrobenzene
diazonium salt, resulting in the moderation of interfacial growth and a concomitant decrease in
the charge transfer resistance of the interface. After subsequent conversion into biocompatible
amino-functionalised interfaces, this led to an increase in the direct electron transfer between the
electrode and an immobilised oxygen-tolerant hydrogenase that could be correlated to the
concentration of scavenger added during interface deposition and the structural/functional
properties of the interface, highlighting the potential for in situ spectroelectrochemical methods to
be used in the design and optimisation of interfaces for electrochemical devices.
In addition to using electrografted diazonium interfaces for immobilising enzymes on the surface
of gold electrodes, a manganese carbonyl bipyrydyl was also immobilised as a model molecular
catalyst for the electrochemical CO2 reduction reaction. Here it could be demonstrated, using in
situ SEIRA spectroelectrochemical methods, that it is possible to use a bipyridine-containing
diazonium salt to immobilise and subsequently characterise a manganese carbonyl bipyrydyl
interface on the surface of gold, either via a one-step ‘pre-coordination’ method, or a two-step
‘post-coordination’ method. However, due to the limited electrochemical stability window in the
region of interest for the catalytic reduction of CO2 it was not possible to study this system further
under turnover conditions.
Despite not being able to use the electrografted diazonium interfaces to study the manganese
carbonyl bipyridine system under turnover conditions, the broad potential stability window of
these interfaces permits many other catalytic systems or reactions to be studied using
electrochemical or spectroelectrochemical methods under highly anodic and cathodic potentials
that are not possible using thiol SAMs. This includes the study of enzymes, such as the study of
hydrogenases under hydrogen evolution conditions.
88
8 Electrochemical grafting of diazonium salts on transparent conductive oxide (TCO) electrodes for the
immobilisation of enzymes
Chapter 5
Electrochemical grafting of diazonium salts on
transparent conductive oxide (TCO) electrodes for
the immobilisation of enzymes
5.1 Introduction
Electrocatalytic devices based on enzymes, such as enzymatic fuel cells (EFCs) or biosensors,
use enzymes to catalyse the oxidation or reduction reactions taking place and do this without
noble metals that are otherwise often used as catalysts e.g. in proton exchange membrane fuel
cells (PEMFCs). The concept of a hydrogenase-based EFC was already introduced in section 1.5
(page 42). Such devices are attractive not only due to potential cost savings that can be achieved
e.g. through avoiding the use of noble metals, but also because they can be used in implantable
or portable electronic devices (i.e. in physiological conditions), and, in the case of biosensors,
they may show particularly high selectivity towards specific analytes due to specific binding
affinities of enzymes to specific substrates.197,198 Glucose-powered EFCs that run off glucose and
oxygen in the extra-cellular body fluid have been proposed as alternative power sources to
batteries for medical implants. A prominent example is that of a glucose-oxidase/laccase-based
EFC that successfully operated implanted in a rat with a power output of 38.7 µW.199 In order to
make such devices scalable and cost-efficient to produce, low cost electrode materials must be
used. A number of high surface area nanomaterials (HSMs), such as carbon materials like
carbon nanotubes (CNTs), have been used and developed as bioelectrodes in recent years, in
particular due to the numerous advantages that they offer in terms of protection against enzyme
denaturation, enhanced loadings and increased current outputs due to enhanced surface areas,
as well as improved electrical contact between the enzymes and the electrode material.200 In this
context, TCO materials, such as ITO or ATO, are highly suited for use as bioelectrode materials
due to the ease with which nanostructured TCOs can be fabricated, their broad potential range
(especially under highly anodic potentials), and their biocompatibility and ease with which
enzymes adsorb on native oxide surfaces. The inherent transparency of these materials also
allows their use in photoelectrochemical devices and in spectroelectrochemical investigations e.g.
of enzyme behaviour.201–209 The use of low cost TCOs, such as ATO or FTO, offer potentially
large cost-savings compared to conventional electrode materials like Au, and could be used in
low-cost, disposable devices.
89
While certain enzymes may adsorb on native oxide surfaces, the lack of control over the
orientation of the enzyme can lead to poor electronic coupling between the enzyme and the
electrode, as recently demonstrated for certain heme proteins on mesoporous ITO by Reisner
and co-workers.208 Orientation control for enzymes immobilised on mesoporous ITO electrodes
was demonstrated by Schuhmann, Shleev and co-workers by using silanes to functionalise the
ITO surface with amine or epoxide functional groups prior to immobilisation of a cellobiose
dehydrogenase and a bilirubin oxidase.209 These functionalities further allowed the covalent
attachment of the enzymes to the electrode. Fattakhova-Rohlfing and co-workers also covalently
attached cytochrome c to mesoporous ATO, which had been functionalised with amine functional
groups using a silanisation procedure, and demonstrated greater stability to leaching compared
to cytochrome c immobilised on the native, bare ATO surface.206
Despite the widespread use of diazonium salts to modify carbon materials, only few examples
exist where metal oxide materials have been modified, including titanium dioxide, ITO and
fluorine-doped tin oxide (FTO).55,57,210–214 Conventional approaches to functionalising oxides,
such as using phosphonate, carboxylates and silanes, have their drawbacks, including varying
stabilities as mentioned in section 5.3.4 (page 20), or poor electronic coupling, as will be
discussed further in section 9.6 (page 161). Using the specific example of the silanes used to
functionalise the ITO and ATO in the work of Schuhmann, Shleev and co-workers, as well as
Fattakhova-Rohlfing and co-workers, drawbacks include a propensity to polymerise, as well as a
weak electronic coupling between the conductive oxide and the immobilised redox species.
In an attempt to develop a new approach to immobilising enzymes on TCOs electrodes, planar
ITO electrodes were electrochemically modified using a similar approach to that previously
developed for gold electrodes (section 7.4, page 71). In that work, a radical scavenger was added
in an attempt to inhibit the polymerisation mechanism whereby multilayers are formed, and
thereby promote monolayer formation. The ITO electrodes are similarly electrochemically grafted
with 4-NBD to introduce nitrophenyl moieties that are electrochemically reduced to aminophenyl
moieties in order to allow the immobilisation of the MBH enzyme under orientation control and to
allow protein film voltammetry (PFV) to be carried out. A greater understanding of how such
electrografted diazonium interfaces function on simple, planar surfaces will allow their eventual
use on more complex TCO HSMs for use in functional devices.
5.2 Effect of radical scavenger on interface formation and immobilisation of hydrogenase
pl-ITO electrodes were electrochemically grafted with 1mM 4-NBD using cyclic voltammetry
with the addition of 0, 1 or 2 molar equivalents of the radical scavenger DPPH. As can be
seen in Figure 18, addition of DPPH changes the reduction peak shape and position
compared to the trace obtained with only 4-NBD. With 2 molar equivalents of DPPH
added (lilac trace), the reversible 1 electron reduction and re-oxidation of the DPPH itself can
be observed overlapping
90
with the irreversible reduction wave of the 4-NBD. As a reference, a CV showing the redox
behaviour of 1 mM DPPH with no 4-NBD is also shown in Figure 18 (black trace, short dots).
Passivation of the pl-ITO electrodes upon repeated cycling can be seen Figure 19.
Figure 18
CVs of a 1 mM solution of 4-NBD in acetonitrile (0.1 M TBAP) with increasing molar equivalents of DPPH added
(coloured traces), and a CV of only 1 mM DPPH (black trace, short dots) using pl-ITO as a working electrode.
Scan rate = 50 mV/s.
Figure 19
CVs of 1 mM solutions of 4-NBD in acetonitrile (0.1 M TBAP) using pl-ITO as a working electrode. Scan rate = 50
mV/s.
Passivation of the electrodes grafted with 4-NBD (and DPPH) is further confirmed by comparative
CV measurements of the unmodified and modified pl-ITO electrodes (modified with 0 or 1 molar
equivalent DPPH) in the presence of the ferrocene/ferrocenium (Fe/Fc+) redox probe (Figure
20a). In addition to an increase in peak separation ΔEP, a significant decrease of 59% in the
ferrocene oxidation current is observed for pl-ITO electrografted with no added DPPH, while a
91
smaller decrease of 35% is observed for pl-ITO modified with 1 molar equivalent of DPPH added.
First of all, this result confirms that electrochemical grafting of 4-NBD results in the passivation of
the ITO surface, and second of all it indicates that the addition of the radical scavenger DPPH
leads to the deposition of a thinner or less dense interface.
Figure 20
CVs of 1 mM ferrocene in acetonitrile (0.1 M TBAP) using unmodified pl-ITO (black dotted trace), or pl-ITO
electrografted with (pink trace) or without (dark green trace) the addition of 1 molar equivalent DPPH, as a
working electrode (pl-ITO/NP = nitrophenyl). Scan rate = 50 mV/s. (b) Nyquist plots comparing the
electrochemical impedance of the unmodified and modified electrodes in 1 mM ferrocene and the corresponding
Bode plots (c-e) used to calculate the solution resistance RS, the charge transfer resistance RCT as well as the
double layer capacitance Cdl.
92
Stevenson and co-workers demonstrated that planar ITO electrodes could be modified via the
electrochemical grafting of the 4-NBD aryl diazonium salt in acetonitrile, with films of thicknesses
in the order of 1-6 nm deposited on the surface. It was suggested that these films were strongly
physisorbed to the ITO surface, since X-ray photoelectron spectroscopy (XPS) did not reveal any
significant changes in the In 3d or Sn 3d spectra, or any apparent evidence of In-O-C or Sn-O-C
bonds in the O 1s spectra. More recently, Stevenson and co-workers used iodonium
tetrafluoroborate salts to electrochemically modify ITO.215 Iodonium salts form aryl radical species
upon electrochemical reduction that are analogous to those formed by the electrochemical
reduction of diazonium salts and, once formed, the adsorption or interaction of these radical
species with surfaces or other species in solution or at the interface can be considered the same.
XPS indicate that Sn-C or In-C bonds are not formed, while there are suggestions that In-O-C or
Sn-O-C species may form. The differentiation of these latter species is hindered due to strong
overlapping contributions from the nitro groups in the O 1s spectra.215 Notwithstanding, a
mechanism was proposed by Stevenson and co-workers for the abstraction of hydrogen from the
ITO surface and the subsequent formation of covalent bonds between the ITO and the aryl
species, as shown in Scheme 49.
Scheme 49
Proposed hydrogen abstraction mechanism proposed by Stevenson and co-workers showing how aryl radicals
abstract hydrogen from surface hydroxyl groups, leaving radical oxygen species, which may couple with further
aryl radicals to form covalent M-O-C bonds. Reprinted with permission from 215, copyright 2015 American
Chemical Society.
Electrochemical impedance spectroscopy (EIS) was used to measure the electron transfer
properties of the electrografted interfaces using a similar approach to that used by Stevenson and
co-workers for ITO modified with an iodonium salt.215 The impedance response of the unmodified
and modified pl-ITO electrodes (modified with 0 or 1 molar equivalent DPPH) is shown in the
Nyquist plot in Figure 20b. The plots are characterised by a semi-circular trace in the complex
impedance plane at high frequencies, corresponding to the sum of the double layer capacitance
Cdl and the charge transfer resistance of the electrode RCT, and a Warburg line at low frequencies
due to impedance caused by diffusion of ions. These features indicate that a Randle equivalent
circuit may be used to model the properties of these systems.216 The impedance data are
represented in Bode plots in Figure 20c – d, showing how Cdl, RCT and the solution resistance RS
for each electrode was calculated. These values are summarised in Table 5.
93
Table 5
The values of RS, RCT and Cdl calculated from the impedance data for unmodified and modified pl-ITO electrodes.
Sample
R
S
(Ω)
R
CT
(Ω)
C
dl
(µF/cm2)
pl-ITO/NP (0 mM DPPH)
196
4337
2.69
pl-ITO/NP (1 mM DPPH)
229
1871
8.20
pl-ITO
186
359
185.00
Electrochemical grafting of ITO with 4-NBD without any addition of DPPH results in an order-of-
magnitude increase in RCT, as well as a very large decrease in Cdl, which suggests the deposited
interface interrupts the adsorption of electrolyte ions onto the native oxide surface. Addition of 1
molar equivalent of DPPH during electrochemical grafting significantly reduces RCT in agreement
with the deposition of a thinner or less dense interface. Similar reductions in RCT were observed
by Breton and co-workers when electrochemically grafted GC electrodes with 4-NBD in the
presence of molar equivalents of DPPH.157
Figure 21
Electrochemical reduction of the nitrophenyl (NP) moieties to aminophenyl moieties in a 0.1 M NaClO4, 1:9
ethanol:water solution, and (b) the corresponding surface coverage of electroresponsive nitrophenyl moieties
Γ
NO2 derived using the integrated charge of the nitrophenyl reduction peaks (as shown for 0 mM DPPH in (a)).
Scan rate = 50 mV/s.
As shown already for gold in section 7.6 (page number 77), the electrografted nitrophenyl
moieties can be reduced chemically or electrochemically in protic conditions via an irreversible
multi-electron reduction to aminophenyl (-NH2) and hydroxyaminophenyl (-NHOH) moieties, as
has been widely reported147,185,186,217, and this results in amino functionalised interfaces that allow
94
the electrostatically-controlled adsorption of the MBH. Figure 21a shows the electrochemical
reduction of these nitrophenyl moieties using cyclic voltammetry, where the first reduction wave
corresponds to mixed four and six electron reductions of nitrophenyl moieties to
hydroxyaminophenyl and aminophenyl moiteties, respectively. Equations (1) and (2) show both
reduction processes involved. Instead of undergoing further reduction to aminophenyl moieties,
hydroxyaminophenyl moieties may undergo a two-electron oxidation to nitrosophenyl moieties
with a certain reversibility, as shown in equation (3).
ITO-Ph-NO2 + 4H+ + 4e- ® ITO-Ph-NHOH + H2O (1)
ITO-Ph-NHOH + 2H+ + 2e- ® ITO-Ph-NH2 + H2O (2)
ITO-Ph-NHOH ⇌ ITO-Ph-NO + 2H+ + 2e- (3)
Integrating the charge of the first reduction wave and the subsequent smaller oxidation wave
gives the charge for a total six-electron redox reaction, which may be used to calculate the
surface coverage
Γ
NO2 of electrochemically accessible nitrophenyl moieties.217,218 The values of
Γ
NO2 calculated using this method are plotted in Figure 21b as a function of molar equivalents of
DPPH used during the electrochemical grafting. A significant reduction in coverage can be
observed from around 1.7 ´ 10-9 mol/cm2 for pl-ITO electrografted without DPPH to 1.1 ´ 10-9
mol/cm2 for pl-ITO electrografted with one molar equivalent of DPPH added. For pl-ITO
electrografted with two molar equivalents of DPPH added, a coverage of around 0.7 ´ 10-9
mol/cm2 is obtained, which corresponds to a near monolayer coverage and is the same value
(0.65 ´ 10-9 mol/cm2) as that obtained by Breton and co-workers on GC with added DPPH. An
ideal close packed monolayer of nitrophenyl moieties is expected to give a surface coverage of
1.2 ´ 10-9 mol/cm2; however, given the rapid nature of the adsorption of radical species on
surfaces and the strong bonds that they form, an ordered self-assembly of molecules is not
expected to take place and therefore a value of less than 1.2 ´ 10-9 mol/cm2 will always be
achieved for non-branched interfaces. The differences in the reduction peak potentials and
shapes are further indications of differences in the interface structures.219 Further measurements
would need to be carried out with greater molar equivalents of DPPH added to observe if a
plateau in coverage is reached at higher concentrations. This plateau occurs at around 1-2 molar
equivalents of DPPH in the case of GC. It would seem as though the behaviour observed on ITO
is similar to that observed on GC, where the rate of phenyl radical adsorption on the electrode
surface (ks) is greater than the rate of radical coupling to the radical scavenger (kr), which in turn
is greater than the rate of radical coupling to already grafted phenyl moieties (kp) (as illustrated
previously in Scheme 42, page 73).
50
0
N-.. -50 0.7 x 10·• mol/cm'
-100
-= 3.1 x 10·• moVcm'
-150
·200
-pl-lTO/NP (OmM DPPH),0.1 M TBAF
-pl-lTO/NP (0 mM DPPH), 0.1 M TBAF
-250
-2.5 -2.0 -1.5 -1.0 -0.5
E (V vs FclFc")
Figure 22
CVs in acetonitrile (0.1 M TBAF) under Ar of the nitrophenyl (NP) interfaces on pl-lTO electrodes
electrochemically grafted with 1 mM 4-NBD with (pink trace) or without (dark green trace) the addition of DPPH.
The surface coverages of electroresponsive n�rophenyl moieties r NCO indicated here were calculated using
equation (4) and the integrated charge of the (semi)-reversible N02 reduction peak. Scan rate= 50 mV/s.
In aprotic solvents, such as acetonitrile, nitrophenyl moieties on electrodes can be reversibly
reduced by one electron to the radical anion:
(4)
which may also be used to calculate I'N02.217 CVs of the unmodified and modified pl-lTO
electrodes (electrografted with O or 1 molar equivalent DPPH) in acetonitrile are shown in Figure
22. The semi-reversibility observed here may be attributed to residual water in the acetonitrile,
which may protonate Ph-N02
'"'.220 The integrated cathodic peaks give coverages of 3.1 x 10·9
moVcm2 for pl-lTO electrografted without DPPH and 0.7 x 1 O.Q mol/cm2 for pl-lTO modified with
one molar equivalent of DPPH added, with peak separations !iE
p of 88 mV and 215 mV,
respectively. While the coverage obtained using this method for pl-lTO electrografted with one
molar equivalent of DPPH added is the same as that obtained from the reduction of the
nitrophenyl moieties in water, the higher coverage of 3.1 x 10·9 mol/cm2 for pl-lTO modified
without DPPH is closer to the 3.4 x 10·9 mol/cm2 originally reported by Stevenson and co-workers
for pl-lTO electrodes electrochemically grafted with 4-NBD, and suggests that the measurements
in water underestimate the coverage. Underestimated coverages determined using voltammetry
in protic solvents have previously been reported for thick films on carbon electrodes and were
attributed to the presence non-electroactive, or non-electrochemically accessible (electronically
contacted) moieties.219·221 Certain parts of the interface may be particularly hydrophobic in thick
films and therefore more difficult to irreversibly reduce in protic media. Differences in coverages
determined for thick films in protic and aprotic solvents are therefore highly likely, given the
differences in the ion or solvent molecules that diffuse in the interface and the subsequent
95
96
changes in the intermolecular interactions between the redox species and the dielectric constant
of the interface etc. In any case, these results indicate that addition of a radical scavenger during
the electrochemical grafting of ITO with 4-NBD results in a decrease in coverage
Γ
NO2, with near
monolayer values approached using a higher concentration of DPPH.
Figure 23
(a) Protein film voltammograms recorded in 10 mM PB buffer at pH 5.5 saturated with H2 with a scan rate of 10
mV/s of strep-MBH adsorbed onto aminophenyl (AP)-functionalised pl-ITO interfaces, and plots of electrocatalytic
HOR current due to direct electron transfer as a function of (b) concentration of DPPH added during interface
formation and (c) the coverage
Γ
NO2 of the pl-ITO electrodes.
In order to evaluate the suitability of the electrografted interfaces for enzyme immobilisation and
bio-electrocatalysis, strep-tagged MBH was electrostatically absorbed onto unmodified pl-ITO
and amino functionalised pl-ITO electrodes from a bulk solution at pH 7 in a similar way to that
used in section 7.6 on functionalised gold. PFVs of the electrodes incubated with MBH were
recorded in a H2 saturated, aqueous PB buffer at pH 5.5 and are shown in Figure 23. The
sigmoidal shapes of the CV traces indicate the electrocatalytic oxidation of H2 into protons (H+)
and electrons as a result of direct electron transfer (DET) between the catalytic redox centre of
the strep-MBH and the electrode in a similar way to that observed for strep-MBH on the modified
gold electrodes in section 7.6, as well as by Heidary, Utesch et al on SAM-modified gold
97
electrodes.85 The behaviour is also similar to that observed by Heidary for strep-MBH on planar
tin-rich ITO.84 There is a slight anodic shift in overpotential for HOR of around 50 mV compared to
those obtained on gold. A plot of the peak current as a function of DPPH amount added during
electrochemical grafting is shown in Figure 23b, and a plot of peak current as a function of the
interface coverage (using the higher coverage calculated using CVs in acetonitrile for the thickest
interface electrografted without the addition of DPPH) is shown in Figure 23c. The peak currents
increase with increasing concentrations of DPPH added during electrografting, from 0.90 µA/cm2
for 0 equivalents DPPH, to 2.49 µA/cm2 for 1 equivalent and finally to 7.09 µA/cm2 for 2
equivalents DPPH; the highest overall current recorded was 10.94 µA/cm2 for the unmodified
ITO. A peak current of around 2.3 µA/cm2 was obtained by Heidary for strep-MBH on planar tin-
rich ITO.84
On gold, the peak current was shown to increase as the interface density decreased, presumably
due to the decrease in RCT; however, at the lowest density interface there was a subsequent drop
in current due to a decrease in MBH adsorption. Proteins adsorb well onto native oxide surfaces,
as has been widely reported201–209, therefore it is not surprising that the highest current density is
observed on the unmodified ITO, as the MBH will also experience the lowest RCT due to the
absence of any grafted interface. While further data points should be collected to get a better fit of
the data, there appears to be a sigmoidal shape to the plot of peak current as a function of
interface coverage, which would be consistent with a simple model where the peak currents tend
to zero as the coverage and thickness of the interface becomes high, and tend to a maximum as
the coverage tends to zero. In reality, the interface itself will have an effect on the orientation of
the MBH and, therefore, will affect the current density to a certain degree.
Addition of the redox mediator methylene blue (MB) results in an increase in current as a result of
mediated electron transfer (MET) for all samples, as shown in Figure 24. The currents are
compared to the DET currents in Table 6, and the enhancement in current as a result of adding
the mediator is indicated for each electrode. The values obtained by Heidary for strep-MBH
immobilised on tin-rich ITO and SAM-coated Au are also included for comparison.84 For the
results obtained in this work, there is an overall decrease in the current enhancement as the
interface coverage decreases, as would be expected as RCT experience by the MBH falls. For
thicker interfaces, a large increase in current is anticipated as the mediator allows electrons to be
shuttled to and from the non-electrically contacted MBH adsorbed on top of the interface.
98
Figure 24
Protein film voltammograms recorded in 10 mM PB buffer at pH 5.5 saturated with H2 with a scan rate of 10 mV/s
on MBH adsorbed onto aminophenyl (AP)-functionalised interfaces, and unmodified ITO before (solid traces) and
after (dashed traces) addition of the redox mediator methylene blue.
Table 6
Electrocatalytic HOR peak current due to direct electron transfer (DET) and mediated electron transfer (MET) and
the resulting enhancement in current for strep-MBH immobilised on each sample. Values obtained for by Heidary
for strep-MBH immobilised on tin-rich ITO and SAM-coated Au are also included for comparison.
Sample
DET (
µµ
A/cm2)
MET (
µµ
A/cm2)
Enhancement
0 mM DPPH
0.90
14.17
x 15.7
1 mM DPPH
2.49
12.28
x 4.9
2 mM DPPH
7.09
36.58
x 5.2
Bare ITO
10.94
29.31
x 2.7
Bare tin-rich ITO84
2.3
9.7
x 4.2
NH
2
-SAM coated Au84
2
4
x 2
NH
2
/OH-SAM (1:9) coated Au84
3.3
9.8
x 3
Care should be exercised when interpreting the enhancement in currents given the changing
nature of the electrografted interface on the pl-ITO and the corresponding change in RCT
experienced by the mediator. For the denser interfaces, electron transfer between the electrode
and the mediator may be inhibited and this may in turn influence the kinetics of the HOR (see
Figure 20 and Table 5 for the influence of the interface structure on the electron transfer with the
Fc/Fc+ redox couple). This can also be inferred from the appearance of a reversible redox peak
for the methylene blue on the unmodified ITO that is absent for the modified electrodes.
99
The increase in catalytic current observed for the bare ITO suggests that either a significant
amount of the MBH molecules are orientated in a non-direct electron transfer configuration i.e.
with the distal cluster pointing away from the electrode surface, or that multi-layers of MBH
adsorb onto the electrode surface, or a combination of both. The formation of multi-layers would
explain why the currents observed for MET are substantially larger than those obtained by
Heidary for monolayers (confirmed by atomic force microscopy) of strep-MBH immobilised on
planar tin-rich ITO or NH2/OH-SAM (1:9 ratio) coated Au (29.3 µA/cm2 vs 9.7 µA/cm2 and 9.8
µA/cm2, respectively).84
Figure 25
SEM images of the (a) commercial planar ITO-coated glass (pl-ITO), and (b) spin-coated ITO deposited on a Si
prism.
To follow the adsorption of the MBH on modified and unmodified ITO using IR spectroscopy, ITO
was spin coated onto a Si-prism (hereby denoted sc-ITO) following a modified literature method
from Xu, Cheng and co-workers that produces ITO films with a low surface roughness,
comparable to sputtered ITO.124 SEM images of the pl-ITO electrodes used in this study
compared to SEM images of the sc-ITO on a Si-prism are shown in Figure 25. The images of the
sc-ITO resemble those reported in by Xu, Chening and co-workers, with an apparent lower-
surface roughness than the pl-ITO. The sc-ITO electrodes were studied using in situ ATR IR in a
Kretschmann-type configuration.
100
Figure 26
ATR IR absorption spectrum of spin-coated ITO (sc-ITO) (a) in acetonitrile after electrochemical grafting with 1
mM 4-NBD with 1 equivalent DPPH added to give nitrophenyl (NP)-functionalised ITO, and (b) in water after
electrochemical reduction of the nitrophenyl (NP) moieties to aminophenyl (AP) moieties.
A sample was successfully electrochemically grafted with 1 mM 4-NBD in the presence of 1 mM
DPPH, as can be seen by the appearance of absorption bands characteristic of a nitrophenyl
interface in the IR adsorption spectrum in Figure 26. Due to the lack of enhancement like in
SEIRA spectroscopy, the adsorption intensities obtained are very low, and no polarised light was
used. Thus, no information can be obtained about any preferential orientation of the molecules on
the ITO surface. Compared to the nitrophenyl interfaces measured on Au, there is a shift of 6-8
cm-1 to lower wavenumbers for each of the bands marked in Figure 26. The band positions on Au
and ITO are compared in Table 7.
Table 7
A comparison between the IR adsorption band positions of nitro-phenyl moieties electrografted on Au and ITO.
Sample
Band position (cm-1)
Au
ITO
Band assignment
4-NBD
1351
1344
v
s
(NO
2
)
1526
1518
v
as
(NO
2
)+v(C=C)
ar
1598
1592
v(C=C)
ar
+v
s
(NO
2
)
The nitrophenyl moieties were electrochemically reduced in 0.1 M NaClO4, 1:9 ethanol:water and
a new IR adsorption spectrum was recorded (Figure 26b). This spectrum was recorded in water,
unlike the spectrum of the aminophenyl interface on gold in Figure 12a in section 7.6 (page 78),
which was recorded in dry acetonitrile, resulting in a broad band around 1640 cm-1 due to water
that obscures the broadened
d
(NH2) band for the partially protonated aminophenyl moieties.
While there are significant changes in the baseline after reduction of the interface, making the
characterisation of the spectrum difficult, it is possible to identify bands characteristic of the
aminophenyl interface, including the v(C=C)ar band at 1594 cm-1 and the
d
(C-H)+v(C=C)ar band at
1502 cm-1. The IR adsorption spectra of aminophenyl interfaces on gold and ITO recorded in
101
water are compared in Figure 27a, while the corresponding difference spectra calculated after
reduction of the aminophenyl interface are compared in Figure 27b. The IR adsorption spectra
appear similar, while the behaviour during reduction also appears similar, as confirmed by looking
at the second derivatives in Figure 27b.
Figure 27
(a) ATR IR absorption spectra in water comparing the aminophenyl (AP) interface electrografted on spin-coated
ITO (sc-ITO) and gold. (b) Difference spectra recorded in 0.1 M NaClO4, 1:9 ethanol:water after electrochemical
reduction of the nitrophenyl (NP) moieties to the aminophenyl (AP) moieties. The second derivatives of the
difference spectra are also shown.
Upon incubating the unmodified and amino-functionalised sc-ITO electrodes with strep-MBH at
pH 7, two broad absorption bands appear at 1650 and 1544 cm-1 (Figure 28) corresponding to
the amide I and amide II bands of the protein backbone, indicating that the MBH binds to the
unmodified and amino-functionalised surfaces. Compared to the spectra recorded on gold, these
bands are shifted 4-8 cm-1 to lower wavenumbers. Spectra were recorded at 3-minute time
intervals over the course of the incubation process and the amide I and II band intensities were
plotted as a function of time and fitted using a biexponential function:
A = A0 + A1 exp(-t/τ1) + A2 exp(-t/τ2)
where A0, A1 and A2 are the final adsorption intensities for the individual adsorption processes,
and τ1 and τ2 are the time constants for the respective adsorption processes. What is clear from
the plots and the kinetic data analysis is that the first adsorption process occurs at a faster rate
on both electrodes compared to the second, slower adsorption process.
102
Figure 28
ATR IR absorption spectra in buffer (pH 7, 10 mM PB) of MBH being adsorbed (a) onto bare sc-ITO and (c) onto
amino functionalised sc-ITO (modified with 1 molar equivalent DPPH) (inset: shows a characteristic active site
band of the enzyme, namely the v(CºO) stretching vibration of the Fe-coordinated carbon monoxide ligand, which
can be detected at 1948 cm-1 with a shoulder at 1934 cm-1 for the so called “Nir-B” and the “Nir-S” states of the
oxidised enzyme). Kinetic analysis of the adsorption of the MBH onto the (b) bare sc-ITO and (d) the amino
functionalised sc-ITO (using 1 mM DPPH).
Table 8
Kinetic analysis of the MBH amide I and II band intensities recorded during adsorption on the unmodified and
amino functionalised sc-ITO electrodes. A biexponential fitting (A = A0 + A1 exp(-t/τ1) + A2 exp(-t/τ2)) was applied
over the first 21 minutes of MBH absorption in both cases.
Sample /
band
A
0
(mOD)
A
1
(mOD)
τ
1
(min)
A
2
(mOD)
τ
2
(min)
Bare ITO
1650 cm-1
1545 cm-1
3.97 ± 0.18
3.16 ± 0.33
1.19 ± 0.06
0.99 ± 0.08
0.70 ± 0.11
0.92 ± 0.16
2.78 ± 0.13
2.17 ± 0.25
13.98 ± 1.75
15.97 ± 4.42
1 mM DPPH
1650 cm-1
1544 cm-1
2.37 ± 0.04
1.91 ± 0.01
1.39 ± 0.10
1.03 ± 0.04
0.82 ± 0.12
0.61 ± 0.08
0.98 ± 0.07
0.88 ± 0.03
6.21 ± 1.07
5.29 ± 0.34
A0, A1 and A2 are the final adsorption intensities for the individual adsorption processes, and τ1 and τ2 are the
time constants for the respective adsorption processes.
103
The first process occurs at a slightly faster rate on the amino-functionalised sc-ITO compared to
the unmodified sc-ITO, while the second, slower adsorption process occurs at a substantially
slower rate on the amino-functionalised sc-ITO. These results suggest that while strep-MBH
monolayer formation is fast on both substrates, substantial multilayer formation is preferentially
observed in the case of the unmodified sc-ITO, which would explain the substantial mediated
currents obtained for unmodified pl-ITO in Figure 24. The reasons for the preferential multilayer
growth of strep-MBH on unmodified ITO remains unclear; however, the difference in
immobilisation observed for the unmodified and amino-functionalised ITO shows that surface
modification clearly impacts on the immobilisation behaviour of the MBH. One possible
explanation may be that the orientation at which the first monolayer of MBH adsorbs onto the
unmodified ITO further facilitates adsorption of more MBH molecules on top of them. The surface
charge of the unmodified ITO is likely different to the amino functionalised ITO, and thus likely to
lead to a different distribution of orientations of MBH molecules. The unmodified ITO will mainly
have hydroxyl groups present and exposed metal sites, which may interact with amino-acids on
the protein back-bone. For strep-MBH adsorbed on tin-rich ITO by Heidary, a nearly mono-
exponential increase in band intensity was observed, where a large proportion of the MBH
molecules are adsorbed in an unfavourable orientation, given the 5-fold increase in HOR current
upon addition of MB mediator.84 A similar behaviour was observed for MBH adsorbed on a
hydrophobic alkane-thiol SAM84 and may point to the hydrophobic nature of the tin-rich ITO
surface.
The appearance of a band at 1948 cm-1, related to the v(C=O) stretching vibration of the CO
ligand coordinated to the [Ni-Fe] active site and characteristic for the highest oxidised ‘Nir-B’ state
for both the unmodified sc-ITO and the amino functionalised sc-ITO indicates that the enzyme’s
integrity is preserved upon surface immobilisation in both cases.130,189 A small shoulder band at
1934 cm-1 in the case of the unmodified sc-ITO may be due to the ready silent ‘Nir-S’ state.188,222
No shoulder band was observed by Heidary for strep-MBH immobilised in tin-rich ITO.84
104
Figure 29
First 5 cycles of the protein film voltammetry recorded in 10 mM PB buffer at pH 5.5 saturated with H2 with a scan
rate of 10 mV/s for strep-MBH adsorbed onto unmodified and amino-functionalised pl-ITO.
Further differences in the behaviour of the MBH on modified and unmodified ITO can be
observed in repeated PFV cycles of MBH in H2 on the unmodified and amino functionalised pl-
ITO electrodes. The shape of the traces on unmodified pl-ITO slowly change upon repeated
cycling, with the electrocatalytic current increasing with each subsequent cycle. This behaviour
points to a gradual change in the orientation of the MBH molecules in the first monolayer and is in
stark contrast to that observed on the amino functionalised pl-ITO electrodes, where the trace
shape and electrocatalytic currents remain relatively stable. The change in orientation on the
unmodified ITO may be promoted by a change in the MBH-surface interaction energy landscape
induced upon changing the pH of the buffer from pH 7, at which the MBH was adsorbed onto the
surface, to pH 5.5, at which the electrochemistry is performed. The stabilities in current exhibited
here upon repeated cycling are in contrast to the behaviour of strep-MBH on tin-rich ITO, which is
reported to decrease by 50% after 5 minutes of potential application. A lower decrease of 12%
was reported for the selectively bound his-tagged MBH on tin-rich ITO.84
105
Figure 30
(a) Constant potential electrolysis (CPE) recorded at -0.15 V (vs Ag/AgCl 3M) in 10 mM PB buffer at pH 5.5
saturated with H2 for MBH adsorbed on unmodified pl-ITO (black trace) and amino functionalised pl-ITO (using 1
molar equivalent DPPH, violet trace), and (b) protein film voltammetry recorded before (dotted traces) and after
CPE, scan rate = 10 mV/s.
After repeated cycling, constant potential electrolysis (CPE) was performed for 1000 s at -0.15 V
(vs Ag/AgCl 3M) using the MBH immobilised on the unmodified pl-ITO electrode and the pl-ITO
electrografted with the addition of 2 molar equivalents DPPH. A comparison of the PFV traces
recorded before and after CPE (Figure 30) show that while the trace shape doesn’t change in the
case of the amino functionalised ITO (other than for an 8% decrease in maximum current that
reflects the decrease in activity over the course of the CPE), the trace shape for the unmodified
pl-ITO changes substantially with a slight increase in maximum current. Furthermore, above a
potential of -0.15 (vs Ag/AgCl 3M) there is a substantial decrease in current due to the reversible,
electrochemically induced anaerobic inactivation of the hydrogenase to the Ni-B state. This
suggests that the overpotential for deactivation decreases, which may result from a reorientation
on of MBH on unmodified pl-ITO. This would indicate a different interaction between the MBH
and the unmodified and amino-modified interfaced, i.e. a stronger interaction for the latter and
higher energetic barriers for reorientation. In any case, these results indicate that diazonium salts
with various functionalities may be used to immobilise enzymes via electrostatic interactions on
TCO materials in an electrochemically accessible manner, and potentially allows for
immobilisation via covalent bonds (e.g. through amide coupling).
106
Scheme 50
Proposed structures of interfaces electrochemically grafted onto ITO surfaces via the electrochemical reduction of
diazonium salts.
As previously mentioned, close packing of the interface is not possible during the electrochemical
grafting of the ITO, and, as more radical scavenger is added during modification, the coverage
eventually plateaus at a sub-monolayer coverage. This will result in a ‘mixed’ interface, where a
certain area of the ITO substrate is left exposed (e.g. with hydroxyl groups), which could, for
example, affect the surface charge of the electrode. ITO has been reported to have a point of
zero charge (PZC) of around 7223; while the PZC of the amino-functionalised surface still needs to
be measured.
5.3 Conclusions
Interfaces were electrografted on ITO electrodes via the electrochemical reduction of diazonium
salts, in this case 4-NBD. The radical scavenger DPPH was added during electrografting
resulting in the moderation of the interface formation with near monolayer coverages achieved,
and in so doing displayed similar behaviour to carbon electrode materials.157 After subsequent
conversion to amino-functionalised interfaces, an oxygen-tolerant hydrogenase was immobilised
on the electrodes and the behaviour of the enzyme in electrocatalytic tests was related to the
coverage and properties of the interfaces. Substantial differences were observed between the
hydrogenase immobilised on unmodified ITO and amino functionalised ITO, with indications of
stronger electrostatic interactions between the MBH and the amino functionalised ITO. Charge
transfer resistance could be sufficiently lowered with the radical scavenger to allow high
electrocatalytic currents to be obtained. Diazonium salts may therefore be used to immobilise
107
enzymes on TCO materials instead of other anchoring groups, such as silanes or phosphonates.
Diazonium chemistry may offer potential advantages over these aforementioned anchoring
groups in terms of electrochemical and hydrolytic stability, and charge transfer properties, and
these aspects will be explored further in the next chapter.
108
9 Electrochemical grafting of diazonium salts on transparent conductive oxide (TCO) electrodes for the
immobilisation of molecular catalysts
Chapter 6
Electrochemical grafting of diazonium salts on
transparent conductive oxide (TCO) electrodes for
the immobilisation of molecular catalysts
6.1 Introduction
As mentioned in section 5.3.9 (page 29), the immobilisation, or heterogenisation, of enzymatic
and molecular catalysts on electrode surfaces is imperative in functional devices. Immobilisation
increases the number of addressable active sites, separates half-reactions and facilitates
electronic contact, especially when using hydrophobic catalysts in aqueous conditions.224,225
Conductive metal oxide supports offer many advantages over other electrode materials like gold
(including significantly lower costs and broader possibilities for functionalisation31) or carbon-
based materials (including improved corrosion resistance, particularly under oxidative
conditions).226 Semi-conducting oxides have several properties that make them highly suitable for
energy storage and conversions devices, including transparency or band gaps that allow the
adsorption of visible light for solar-driven reactions, as well as diverse possibilities for
nanostructuraton of the material e.g. using templating methods. Nanostructuration can induce
desirable properties in the material, such as high surface areas (giving HSMs) in order to achieve
a higher density of catalytic species (per cm2 of electrode), and hence higher outputs from e.g.
electrocatalytic devices like electrolysers or fuel-cells.
In order to immobilise catalysts on oxides, a number of anchoring groups are typically used,
including phosphonic acids, carboxylic acids, silanes and their derivatives, amongst
others.5,6,25,28,36,70,227–235 Each approach has its own merit, which may be related to stability or
charge transfer properties; however, one property (stability or good charge transfer properties)
often comes at the expense of another.27,231,236 New approaches must be developed which, in no
particular order, are: (1) sufficiently robust to withstand a range of chemical and electrochemical
conditions, (2) synthetically straightforward to realise, and (3) facilitate fast and efficient charge
transfer.
Aryl diazonium salts have been used to functionalise a broad range of electrode materials,
including carbon materials, metals and silicon because of the strong covalent bond formed
between the phenyl ring of the salt and the electrode surface.45,237,238 However, the formation of
thick, insulating interfaces often prevents their use in the field of electrocatalysis, where a fast
heterogeneous charge transfer between the immobilised redox species and the electrode is
109
required e.g. in electrolysers or sensor devices.237 So far, diazonium salts have been widely used
to modify carbon materials, such as glassy carbon (GC) or carbon nanotubes (CNTs), and they
have recently been used to successfully immobilise molecular catalysts on such materials. This
includes the immobilisation of an iridium-based oxygen evolution reaction (OER) catalyst on
GC71, the immobilisation of a cobalt terpyridine-based catalyst on GC with activity towards proton
and CO2 reduction75, the immobilisation of an iron porphyrin on CNTs with activity for CO2
reduction152, and the immobilisation of iron porphyrin151 and iron phthalocyanine150 species on
CNTs with activity for the oxygen reduction reaction (ORR). Highlighting some of the potential
drawbacks of using carbon materials, the GC electrodes used to immobilise the iridium-based
catalyst and perform water oxidation were rapidly oxidised under the anodic potentials applied,
leading to loss of the catalyst from the electrode surface.71
Figure 31
CVs of commercially available (a) FTO and (b) ITO substrates in pH 1 (0.1 M H2SO4), pH 7.2 - 7.8 (0.1 NaAc) and
pH 13 (0.1 M NaOH) electrolytes indicating their electrochemical activity and electrochemically inert potential
range (adapted from Figure 3 and 4, Benck et al. “Substrate Selection for fundamental studies of electrocatalysts
and photoelectrodes: inert potential windows in acidic, neutral, and basic electrolyte.” PLoS One 9, e107942
(2014)). Scan rate = 25 mV/s. The theoretical potentials for the HER/HOR and OER/ORR reactions are
highlighted, and lie within the inert potential range for FTO and ITO in all electrolytes/pH’s tested.226
110
Figure 32
SEM images of the (a) cross-sectional view and (b) top view of the me-ATO films deposited on a Si wafer
showing the nanostructuration and open, mesoporous structure of the me-ATO films, as well as the film
thickness.
For the following work, mesoporous antimony-doped tin oxide (me-ATO) thin film electrodes were
used as a model nanostructured TCO system due to their high surface areas, interconnected
porosity and ease of synthesis.205,239–241 ATO has gained attention as a catalyst support material
due to its low cost compared to semi-precious indium-based ITO (indium tin oxide) and due to its
high electrical conductivity and stability in the electrochemical range required for a diverse range
of important energy conversion and storage reactions, including ORR and water oxidation.4,242
The electrochemically inert potential window of ITO and fluorine-doped tin oxide (FTO) was
determined by Jaramillo and co-workers and found to encompass the thermodynamic potentials
for the HER/HOR and OER/ORR reactions, as shown in Figure 31.226 The same electrochemical
experiments carried out on me-ATO for this study show that it is stable in a similarly broad
stability window (-0.77 to 1.83 V (vs RHE) in pH 7 0.1 M PB, see Appendix 8). As a result, ATO
has found use as a corrosion-resistant electrode materials in dye sensitised solar cells243,244,
optoelectronic devices241 and for electrocatalysis.4,242 The high transparency of me-ATO should
also allow the use of in-situ UV-Vis, Raman and Infrared spectroelectrochemical techniques to
characterise interfaces at the surface of the ATO during and after electrografting, as well as
under catalytic conditions, and as such these techniques were used in this work.
In order to synthesise the me-ATO electrodes used in this work, ATO nanoparticles synthesised
using a non-aqueous hydrothermal synthesis route were deposited on planar ITO coated glass
slides from a colloidal solution containing the amphiphilic block co-polymer F127, as described in
6.1.3. After calcination, the resulting films were crack-free and had an open, mesoporous
structure and a film thickness of around 160 nm, as shown in Figure 32. The small crystal size of
the ATO (3-4 nm) reduces scattering and makes these films highly suitable for in-situ
spectroscopic studies.
111
Scheme 51
Overview of the different catalytic systems synthesised in this chapter and the diazonium salts used to form them
(bpy-diazo and Im-diazo). For more information see text.
In this chapter, the bpy-diazo ligand used to functionalise gold in section 7.7 is also used to
functionalise ITO to determine if diazonium salts can be used to graft/immobilise molecular
catalysts on these oxide surfaces in a series of electrochemical experiments. While the
theoretical potentials for CO2 reduction lie within the stability window of ITO and FTO, CO2
reduction catalysts usually require very large overpotentials that push the real potential outside
the stability window. As ATO (and ITO and FTO) is not stable in the electrochemical conditions
required for the reduction of CO2, i.e. with the manganese bipyridine catalyst, an attempt is made
to assemble a copper-bipyridine complex for water oxidation purposes instead (as shown in
Scheme 51).
In order to immobilise certain metalloporphyrins, which are a highly important class of molecules
used in a broad range of applications (including DSSCs and catalytic devices116,245) on the TCO
electrodes via axial coordination, an imidazole-containing diazonium salt Im-diazo (4-(1H-
imidazol-1-yl)benzenediazonium tetrafluoroborate) was synthesised from 4-(1H-Imidazol-1-
yl)aniline. Axial metal-ligand coordination has been widely employed to immobilise a range of
molecular catalysts or other molecular complexes70,233,246,247, such as photosensitisers231,236,248,
on solid supports. The resulting reactivities of complexes immobilised in this way can further be
tuned by modulating their electronic properties using ligands with different electron-donor
abilities.249 In fact, it has been shown that axial coordination of imidazole species to iron
tetraphenylporphyrin can anodically shift the Fe2+/3+ redox couple potential, enhance its activity
for ORR and improve its selectivity towards water (as opposed to H2O2).250 First, an iron
112
Hangman porphyrin is immobilised on ITO and ATO, as this species is known to be active for the
ORR.251 A number of electrochemical and spectroelectrochemical measurements (resonance
Raman and UV-Vis) are carried out to obtain insights into the interface formation and its
structure, thus facilitating rational interface design. Secondly, a cobalt Hangman porphyrin is
immobilised on me-ATO, as this species is known be active for the HER,252 and similar cobalt
tetrapyrrole species have also been shown to be active for the OER.253–259 As the name
suggests, hangman porphyrins (depicted in scheme 9) contain a 'hanging’ group positioned at a
fixed distance from the metal centre via a xanthene bridge (i.e. in the second coordination sphere
of the complex), which may introduce certain non-covalent interactions that can influence electron
transfer during reduction and alter the behaviour or catalytic activity of the complex.128,129
As well as using ex situ XPS measurements to characterise the interfaces electrografted on me-
ATO from Im-diazo, in situ ATR-IR spectroelectrochemistry is used to determine their
electrochemical and hydrolytic stabilities. This information is important for determining the types
of catalysts and reactions that can be catalysed using electrografted diazonium interfaces on
oxide materials.
6.2 Immobilisation of a molecular oxygen evolution reaction catalyst on indium tin oxide
(ITO) using the electrografting of a diazonium salt
It was shown recently by Mayer and co-workers that copper–bipyridine–hydroxo complexes can
be formed in situ from simple copper salts and bipyridine at high pH’s, and that they can be used
as homogenous catalysts in the OER with an overpotential of around 750 mV and a turnover
frequency of around 100 s-1.260 An improvement in reducing overpotential to 640 mV was
demonstrated by Lin and co-workers using a 6,6′-dihydroxy-2,2′-bipyridine ligand.261 While the
overpotential for this catalyst is large compared to those typically obtained for ruthenium- or
iridium-based molecular catalysts119, copper is several orders of magnitude lower in price. These
results have spurred further work in the direction of copper-based OER catalysts262–265, with an
overpotential as low as 520 mV being obtained by Crabtree, Brudvig and co-workers for a
Cu(pyalk)2 (pyalk = 2-pyridyl-2-propanoate) OER catalyst.266 A number of publications have
recently demonstrated the successful immobilisation of noble metal-based molecular OER
catalysts on nanostructured TCO materials. Sheehan, Brudvig and co-workers recently
demonstrated the use of iridium-based [Ir(pyalc)(H2O)2(μ-O)]22+ (pylac = 2-(2’pyridyl)-2-
propanolate) OER catalyst immobilised on mesoporous ITO via Ir-O-MOx bonds (where MOx =
metal oxide).235 Batista, Crabtree, Brudvig and co-workers immobilised a Ir-based
pentamethylcyclopentadienyl OER pre-catalysts on mesoporous ITO using a silatrane anchoring
group.232 Meyer and co-workers immobilised a Ru-based [RuII(Mebimpy)(4,4′-
(PO3H2)2bpy)(OH2)]2+ (mebimpy = 2,6-bis(1-methylbenzimidazol-2-yl)pyridine) OER catalyst on
mesoporous ATO using a phosphonic-acid based anchoring group.267 Llobet, Jooss, Meyer and
113
co-workers immobilised a Ru-based 3,5-bis(bipyridyl)pyrazolate (bbp)-based diruthenium OER
precatalyst on mesoporous ITO using a carboxylate anchoring group.268
Figure 33
CVs of a 1 mM solution of bpy-diazo in acetonitrile (0.1 M TBAP) using pl-ITO as a working electrode. Scan rate =
50 mV/s.
Cyclic voltammetry was used to electrochemically graft bpy-diazo on the surface of the pl-ITO, as
shown in Figure 33. The electrode was passivated after 2 CVs, showing similar behaviour to the
gold electrodes in 7.7. The reduction potential of bpy-diazo on pl-ITO is -0.64 V (vs Fc/Fc+)
compared to a potential of -0.53 V (vs Fc/Fc+) on Au. Small variations in diazonium salt reduction
potentials on different substrates are commonly reported.186 The bipyridine interfaces previously
electrografted on gold were already characterised using SEIRA spectroscopy, and, as such, it
can be assumed that similar bipyridine-functionalised interfaces are electrografted in this instance
on pl-ITO. In order to assess the activity of pl-ITO towards the OER before and after
electrochemical grafting, CVs were applied in 0.1 M NaOH, as shown in Figure 34. Bare,
unmodified pl-ITO shows current for OER with an onset potential of around 1.9 V (vs RHE), which
is the same as that reported by Jaramillo and co-workers for ITO.226 After incubating the
unmodified ITO electrode for 10 min in 2 mM CuSO4 (in DI water) and rinsing, the same CV
shows slightly increased current for the OER and a marginally lowered onset potential, as can be
seen in Figure 34. Meyer and co-workers have previously shown CuSO4 and other Cu2+ salt
solutions to be active for the OER at high overpotentials in basic solutions.269 For bare pl-ITO that
has been electrografted with bpy-diazo, the same CVs show decreased currents for the OER,
which indicates that the surface of the ITO is passivated by a bipyridine-functionalised interface.
CVs of the electrografted ITO electrode after incubation with 2 mM CuSO4 (in DI water) for 10
min and rinsing result in greatly increased currents for OER (0.63 mA/cm2 vs 0.05 mA/cm2 at 2 V
(vs RHE)) with the onset potential shifting to 1.73 V (vs RHE), suggesting that addition of Cu and
complexation with the bipyridine-functionalised interface is necessary in order to form a species
that is active for OER. The overpotential required to achieve OER is 0.5 V.
114
Figure 34
CVs of pl-ITO electrodes in 0.1 M KOH (in air): bare (orange trace), after incubating for 10 min in 2 mM CuSO4
(purple trace), after electrografting with bpy-diazo (blue trace), and after electrografting and incubating for 10 min
with 2 mM CuSO4 (turquoise trace). Scan rate = 50 mV/s.
Scheme 52
(a) The copper-bipyridine species present at different pHs in the work of Mayer and co-workers (adapted with
permission from 260, copyright 2012 Macmillan Publishers Ltd). (b) The proposed structures of the immobilised
copper-bipyridine species present in this work.
As previously mentioned, Mayer and co-workers achieved water oxidation using a homogenous
copper–bipyridine–hydroxo catalyst. Salts of the dimeric bis-μ-hydroxide cation [(bpy)Cu(μ-
OH)]22+ are isolated from alkaline solutions of 1:1 Cu2+:bipyridine, where the anion is an acetate,
triflate or sulphate anion, or alternatively generated in situ from the copper salt and bipyridine. It
was shown using electron paramagnetic resonance (EPR) spectroscopy that below pH 8 the
copper species exists as a monomeric aquo-complex [(bpy)Cu(H2O)2]2+. As the pH increases, the
dimeric species are formed [(bpy)Cu(μ-OH)]22+, while at highly alkaline pHs the dominant species
is the monomeric (bpy)Cu(OH)2, which is the catalytically active species.260 The structures of
these species are shown in Scheme 52a. It is proposed that during incubation with the 2 mM
CuSO4 solution (in DI water) the monomeric copper-bipyridine aquo-complexes form at the
electrografted interface and in 0.1 M NaOH they form the monomeric bis-hydroxide structure (as
115
shown in Scheme 52b). The overpotential for OER calculated using the half-peak potential for the
homogenous system by Mayer and co-workers at pH 12.5 is 0.75 V, while the overpotential
calculated using the onset potential is around 0.64 V.260 Comparing the overpotentials of this
homogenous system to the heterogeneous system demonstrated in this work is complicated due
to the absence of a well-defined redox peak in the latter system from which a half-peak potential
can be calculated. These peaks are a result of diffusion limitations of the catalytically active
species from the solution to the electrode surface in the homogenous system, which does not
occur in heterogeneous systems (although mass transfer of reactants and products and lead to
diffusion limitations, this is tends not to be an issue in the case of the OER). Nevertheless, the
difference in onset potential between the homogenous system of Mayer and co-workers and the
immobilised system presented here is determined at around 0.14 V. This decrease in
overpotential may partly be explained by the electron-withdrawing amide substituent on the
bipyridine ligand used in this work, which may stabilise the electron-deficient Cu(III) or Cu(IV)
centre required for catalysis. Mechanistic studies for the copper-6,6′-dihydroxy-2,2′-bipyridine
hydroxide OER can be found in a work by Lin an co-workers.262 The use of electron-donating or
electron-withdrawing groups on a catalysts ligand can have a substantial effect on its activity, as
has been demonstrated for a Re(bpy)(CO)3Cl CO2 reduction reaction catalyst, where the use of
either an electron withdrawing or donating amide substituent on the bipyridine ligand was used.270
Immobilisation of a catalyst on the ITO surface should not affect its catalytic activity per se. In
work by Grätzel and co-workers it was shown for another Re(bpy)(CO)3Cl CO2 reduction catalyst
(with phosphonic acid anchoring groups on the bipyridyl ligand) that there was no shift in the
onset potential for CO2 reduction when immobilised on TiO2 compared to the unbound species
used in solution with a GC working electrode.271 Intermolecular interactions between immobilised
catalytic species may have an effect of their catalytic activity however, as will be demonstrated
later for the Fe Hangman porphyrin ORR catalyst.
A semi-irreversible current at cathodic potentials for the Cu+/Cu2+ couple was observed by Mayer
and co-workers at ca. 0.65 V (vs RHE) in nitrogen. There is no current observable in Figure 34 for
the Cu+/Cu2+ couple, rather there is non-reversible current below 0.7 V vs RHE, which
corresponds to electrocatalytic current for the ORR. Other examples exist in the literature where
copper complexes have been used for the ORR.272–274
116
Figure 35
(a) Controlled potential electrolysis (CPE) at 1 V vs 3M Ag/AgCl (1.95 V vs RHE) of the Cu-bpy species
immobilised on pl-ITO (ITO-bpy-Cu) in 0.1 M NaOH (in air), and (b) CVs of the electrode before CPE (turquoise
trace), after CPE (pink trace) and after subsequent re-incubation for 10 min with 2 mM CuSO4 (teal trace). Scan
rate = 50 mV/s.
Controlled potential electrolysis (CPE) at 1.95 V (vs RHE) was performed in 0.1 M NaOH after
the 2 CVs that were performed initially (Figure 35a). There is a rapid and almost total decrease in
current over the 5 min, indicating that the catalytic species is not stable. Applying another CV
after CPE shows a large decrease in current and a significant anodic shift in the onset potential
(pink trace, Figure 35b). Re-incubating the electrode in 2 mM CuSO4 solution (in DI water) and
rinsing results in the restoration of the current for OER in 0.1 M NaOH (teal trace, Figure 35b),
thereby indicating that the bipyridine ligand is stable on the ITO surface and that the loss of
catalytic activity is presumably due to the loss of copper species e.g. by decomplexation from the
bipyridine terminated interface.
Figure 36
CVs of pl-ITO in 0.1 M NaOH (a) after initial electrografting and incubation for 10 min in 2 mM CuSO4 (turquoise
trace), and (b) after CPE and re-incubation with 2 mM CuSO4 (dark green and pink trace). Scan rate = 50 mV/s.
117
As can be seen in Figure 36, the activities remain more stable after re-incubation with CuSO4,
which could also be an indication that some of the previous lost activity was due to the desorption
of loosely bound or physisorbed bipyridine-ligands from the surface, as has been previously
reported for other electrografted diazonium interfaces on electrodes.215,275 A CV of the electrode
after 24 hrs under non-turnover conditions in 0.1 M NaOH (pink trace, Figure 36b) indicates that
there was no loss of activity during this time (i.e. no substantial hydrolysis of the interface). This is
particularly significant, given that commonly used carboxylate and phosphonate anchoring groups
show very poor stability in basic pH’s.25,39
Both Mayer and co-workers, and Lin and co-workers reported good stabilities for their
homogenous copper bipyridine catalysts, as well as high turnover numbers of 100 and 400 s-1,
respectively.260,262 A 35% decomposition of the former catalyst was partly attributed to the
deposition of copper metal on the Pt counter electrode (due to the absence of a membrane) and
there was no indication of copper oxide deposits on the working electrode or any heterogeneous
catalysis taking place. Given that the homogenous copper bipyridine hydroxide catalysts are
readily formed in situ in solution at basic pH values, it is reasonable to expect that any catalytic
species that decomposes as result of the decomplexation of copper from the bipyridine may
regenerate, providing that the bipyridine ligand is not damaged and the copper remains in
solution as an ion. Loss of copper from the surface of the modified electrodes in this work would
result in a miniscule concentration of Cu in solution (compared to the mM equimolar
concentrations of bipyridine and copper salts used in the homogenous systems), meaning that
the equilibrium for the formation of copper bipyridine species will not be favourable. Another
possibility for the loss of activity may be due to the absence of a supporting electrolyte (other than
NaOH). Coordinating anions may play an important role in stabilising transition states or charged
species, such as aquo complexes.
Figure 37
SEM images of a pl-ITO electrode (a) before and (b) after electrochemical grafting with bpy-diazo, incubation with
CuSO4, and electrolysis in 0.1 M NaOH.
118
Scanning electron microscopy (SEM) images were made of the bare, unmodified pl-ITO and of
the modified pl-ITO electrode after the CPE and other electrochemical experiments. As can be
seen in Figure 37, electrochemical grafting results in the formation of an insulating film on the ITO
that is still present, even after 24 hrs in 0.1 M NaOH. There is no indication of the formation of
oxides, as expected, although it is unlikely such trace amounts of oxide could be observed at
such resolutions. Ideally, XPS would need to be conducted on the modified electrodes before and
after electrolysis to determine the loss of Cu or detect the formation of non-catalytically active
species. Given the apparent stability of the electrografted interface, the addition of copper salt
(e.g. in the form of mM concentrations of CuSO4) to the NaOH electrolyte may result in stable
currents for OER, as this would favour the regeneration of copper bipyridine hydroxide species at
the interface (assuming that this is the reason for the loss of activity observed).
6.3 Chemical and electrochemical stability of electrografted diazonium interfaces on
mesoporous antimony-doped tin oxide (me-ATO)
As mentioned previously in 9.1, the stability of a catalytic interface deposited on an electrode is of
utmost importance when it comes to the fabrication of practical catalytic devices with long-term
stabilities. In order to determine the stability of electrografted diazonium interfaces on TCOs using
ATO as a model electrode material, me-ATO films were electrochemically grafted with Im-diazo
and the electrochemical and chemical stability of the electrografted interface was investigated
using in situ IR spectroscopy in ATR mode.
Figure 38
SEM images of me-ATO films deposited on (a) (b) Si-prisms via spin-coating and (c) (d) ITO-coated glass slides
via dip-coating.
119
The me-ATO films were deposited on a Si-prism via spin-coating, as outlined in 6.1.3. SEM
images (Figure 38 a, b) show that the obtained films are crack-free and exhibit an open, porous
structure similar to films deposited on ITO slides via dip-coating (Figure 38 c, d). The only
difference is the somewhat greater extension of the ATO nanocrystals in the case of spin-coating.
Given the similarity in structure between the films (and that the surfaces are essentially the same)
it can be assumed that the corresponding desorption behaviour of the interfaces on the Si-prism
will be representative of interfaces on samples used in electrochemical experiments.
IR spectra were recorded in ATR mode in acetonitrile (0.1 M TBAP or TBAF) before and after
addition of 1 mM Im-diazo and subsequent electrochemical grafting, and the calculated difference
spectra are shown in Figure 39. Adsorption of the Im-diazo on the ATO surface from solution can
be deduced from the first spectrum, due to the appearance of a band at 2253 cm-1 corresponding
to the diazonium N≡N+ related stretching vibration v(N≡N), as well as bands at lower
wavenumbers corresponding to modes of the respective phenyl and imidazole rings. The v(N≡N)
absorption band is partially obscured by two negative bands corresponding to the d(CCN) and
v(CºN) bands of acetonitrile molecules (asterisked) that are displaced from the surface of the
ATO by the diazonium species. The adsorption of positively-charged species such as diazonium
cations on ATO may be enhanced due to ATO’s inherent negative charge, as previously
observed in a range of pH’s.276 Relative to the rapid, initial adsorption observed, further growth in
band intensity, e.g. due to the spontaneous heterolytic decomposition of Im-diazonium species, is
negligible.
Figure 39
ATR-IR absorbance spectrum of me-ATO in a 1 mM solution of Im-diazo (0.1 M TBAF, Acetonitrile) before
electrochemical grafting is shown in dark green. The corresponding spectrum in fresh 0.1 M TBAP (without Im-
diazo) after electrochemical reduction is shown in black. On the right is the optimised structure of Im-Sn(OH)4 (1),
atom labels: carbon (dark grey), hydrogen (light grey), nitrogen (blue), oxygen (red), tin (green). The
corresponding IR spectrum of 1 calculated using DFT (dotted line) is shown in black (dashed). The inset displays
the overlapping bands at 1247 cm-1 (x 4) overlaid with calculated contributions from v(C-N)+d(C-H) and v(C-
O)+d(C-H) modes. Negative bands resulting from the displacement of acetonitrile molecules from the surface are
asterisked.
120
Electrochemical reduction of Im-diazo was performed by cyclic voltammetry, resulting in a sharp
increase in the intensity of the vibrational bands and a disappearance of the v(N≡N) absorption
band at 2254 cm-1. DFT calculations were carried out to assign the experimentally observed
vibrational modes. As previously mentioned, it is anticipated that phenyl radicals generated by the
electrochemical reduction of diazonium salts will result in the formation of M-O-C bonds (in this
case Sn-O-C) at the interface.215 In order to model such surface-bound species, DFT calculations
were run for the compound 1 (see Figure 39), a phenyl-imidazole species bound to an Sn(OH)3
cluster via a Sn-O-C bond. As can be seen in the IR spectra, there is a reasonably good
agreement between the experimental and calculated spectra despite using a small (less
expensive) monoatomic Sn hydroxide model to simulate the ATO surface instead of a bigger
cluster, such as the larger 9 Ti atom TiO2 cluster recently used by Brudvig and co-workers to
model a molecular species bound to TiO2.232 Notable bands assigned here in the spectrum
include those at 1608 cm-1, corresponding preferentially to the aromatic C=C stretching vibration
v(C=C)ph of the phenyl ring, and those at 1515 cm-1, corresponding predominantly to the coupled
d(C-H), v(C=C)ph and v(C-N) modes of the imidazole ring.
The absorption band intensities appear constant with respect to repeated rinsing in multiple
solvents (acetonitrile, dichloromethane, ethanol), strongly indicating that the electrografted
diazonium interface is covalently attached to the ATO surface. The inset in Figure 39 shows an
absorption centred at 1247 cm-1, which is composed of two overlapping bands at 1263 cm-1 and
1252 cm-1 corresponding to the v(C-N)+d(C-H) and v(C-O)+d(C-H) modes calculated for 1. This
indicates the presence of Sn-O-C bonds at the interface, which supports the assumption that the
phenyl radicals bind to surface hydroxyl groups.
Figure 40
The deconvoluted C 1s and Sn 3d XP spectra of me-ATO before and after electrochemical grafting with Im-diazo.
In order to gain further information about the nature of the interface formed between the linker
molecule and the ATO, X-ray photoelectron spectroscopy (XPS) measurements were conducted
on me-ATO before and after electrochemical grafting with Im-diazo. The samples were freshly
121
prepared and calcined prior to electrografting and duly stored in flushed, sealed glass containers
to reduce the adsorption of adventitious species from the atmosphere. For details on the fitting
procedure used to de-convolute the components of the different spectra, see 6.6.3 (page 41).
Figure 40 shows the C 1s and Sn 3d spectra of me-ATO before and after electrochemical
modification with Im-diazo. XPS is a surface sensitive technique with a limited penetration depth
of a few nm’s and successful modification of the me-ATO can be inferred from the suppression of
the Sn 3d5/2 and Sn 3d3/2 peak intensities after modification, and the increase in the C 1s peak
intensity. Deconvolution of the C1s signal reveals the presence of peaks that correspond to
aliphatic C-C, as well as C-O/C-N, and a π-π* satellite peak at with a binding energy (BE) of
around 292 eV indicative of aromatic carbon species.
Figure 41
The deconvoluted N 1s and O 1s XPS spectra of (a-b) me-ATO before and after electrochemical grafting with Im-
diazo.
Scheme 53
Scheme illustrating the different chemical environments for the nitrogen atoms present in the protonated and un-
protonated linker molecule. The colours of the nitrogen atoms are colour co-ordinated with the corresponding
environments in the fitted XP spectra.
Table 9
Binding energies (in eV) of the individual deconvoluted components of the N 1s and O 1s spectra recorded for
modified and unmodified me-ATO.
Binding energy (eV)
N 1s
O 1s
me-ATO
530.8
532.0
me-ATO/Im
399.2
530.9
400.1
531.9
401.1
402.1
122
The N 1s and O 1s spectra are shown in Figure 41. The N 1s spectra of me-ATO after
electrochemical modification shows peaks at 399.2 and 401.1 eV that can be assigned to the
amine and imine nitrogen environments, further indicating the successful grafting of Im-moieties
onto the ATO surface (the different chemical environments are illustrated in Scheme 53). The
peak with a BE of 402.1 eV is assigned to the iminium environment that results from protonation
of imine nitrogen and represents a shift in BE of 2.8 eV. Summing together the atomic
percentages imine and iminium nitrogen results in an almost 1:1 ratio with the amine nitrogen and
suggests that 14% of the interface molecules are protonated. A peak with a BE of 400.1 eV is
assigned to azo-bridges formed as a result of addition of non-reduced diazonium ions to already-
grafted species.221
XPS studies on ITO modified iodonium salts rule out the formation of In-C or Sn-C bonds at the
interface and suggest instead the formation of M-O-C bonds; however, the results are not
conclusive due to overlapping signals from the nitro groups in the O 1s spectra.215 Given the
absence of oxygen in the Im-diazo used to modify me-ATO, the O 1s spectrum was fitted in an
attempt to observe the Sn-O-C bond. The spectra were fitted taking into account the overlap with
the Sb 3d5/2 peak and the result was a peak at 530.8 eV attributed to the bulk lattice oxygen
component and a peak at 532 eV for surface hydroxide oxygen. While the ratio between these
peaks does decrease after modification of me-ATO, it was not possible to distinguish any Sn-O-C
bonds. Sn-O-C bonds have been previously reported to have a binding energy of 532.2 eV, which
is too close to that of oxygen in surface hydroxide species to allow for a clear differentiation.277
Figure 42
IR spectra calculated using DFT of 1-phenyl-1H-imidazole 2 and its protonated imidazolium analogue 3. Right:
optimised structures of 2 and 3, atom labels: carbon (dark grey), hydrogen (light grey), nitrogen (blue).
IR spectra calculated using DFT of 1-phenyl-1H-imidazole 2 and its protonated imidazolium
analogue 3 indicate that distinct bands exist for each species, as shown in Figure 42. The
noticeable differences between the spectra of the protonated and unprotonated species are due
to the strong coupling of the vibrational modes throughout the molecules. Vibrational modes that
123
are mostly confined to the phenyl ring, such as the v(C=C)ar band at 1604 cm-1 and C-H bending
mode d(C-H) at 1407 cm-1, remain almost unchanged. New bands appear for the protonated
species e.g. at 3540 cm-1- for the N-H stretching mode v(N-H), or at 1531 cm-1 for the imidazole-
ring C-H bending mode d(C-H).
Figure 43
(a) ATR-IR absorbance spectra of an Im-interface electrografted on me-ATO in increasingly basic KOH solutions
and (b) difference spectra showing the change in absorption DA of the interfaces in the increasingly basic
solutions calculated with respect to the initial spectrum of the interfaces in DI water. The broad negative band
around 1600 cm-1 is due to the desorption of water molecules from the surface. pH DI water ~ 6.
Figure 44
(a) ATR-IR absorbance spectra of an Im-interface electrografted on me-ATO in increasingly acidic HClO4
solutions and (b) difference spectra showing the change in absorbance DA of the interfaces in the increasingly
acidic solutions calculated with respect to the initial spectrum of the interface in DI water. (c) Plot of intensity DA
as a function of pH. Bands marked with a star or a diamond correspond to vibrational modes of the protonated Im
species, Im(N-H), or un-protonated species, Im(N), respectively. pH DI water ~ 6.
124
pH dependent spectroscopic stability measurements were conducted in basic and acidic
solutions of KOH and HClO4 (Figure 43 and Figure 44, respectively) to determine the hydrolytic
stability window of the electrografted interfaces, which is of paramount importance in catalytic
applications. In both cases, difference spectra at each pH were calculated with respect to the
initial spectrum recorded for the same interfaces in DI water (pH ~ 6) (Figure 43b and Figure
44b), therefore indicating the change in absorbance ΔA at each pH. The differences in the
spectra of 2 and 3 calculated by DFT due to protonation of the imidazole are also observed
experimentally.
Upon increasing the pH of the solution, an overall decrease in the band intensities corresponding
to both the protonated and deprotonated imidazole species can be observed (Figure 43b). The
spectra stabilise before slow desorption takes place at very high pH’s (>12). Plotting the changes
in ΔA over pH is somewhat complicated due to shifts in the baseline as the pH changes. Upon
decreasing the pH of the solution, the protonation of the imidazole is clearly visible, in particular
due to the appearance of the band at 3562 cm-1, corresponding to N-H stretching mode v(N-H) of
the protonated imidazole ring. Due to the dynamic changes taking place upon changes in pH, ΔA
was plotted versus pH (Figure 43 c) allowing any desorption of linker molecules or instability in
the interface to be followed. There is an increase inΔA for the protonated Im-species (marked by
a star) upon decreasing the pH. This decrease is accompanied by a concomitant decrease in ΔA
for the de-protonated species (marked by a diamond), until a point of inflection is reached around
pH 2, beyond which there is an overall decrease in ΔA for all species, indicating the removal of
the grafted species from the ATO surface. On closer inspection of the plots of ΔA as a function
of pH, one can see they exhibit a sigmoidal shape (see Appendix 7), which is characteristic for a
pH titration. At very low pH’s this sigmoidal shape is disrupted due to desorption of the interface.
Figure 45
Plots of d(C-H) IR band intensities versus time for Im-interface electrografted on me-ATO in (a) 0.1 M HClO4 and
(b) 0.1 M KOH. The wavenumber of the bands used are indicated in the legend. Plots of the natural logarithm (ln)
of the absorbance versus time are shown in the insets.
125
In order to gain an insight into the desorption of the electrografted interface under acidic and
basic conditions and any underlying hydrolytic processes taking place, IR adsorption spectra of
freshly prepared interfaces were measured between various time intervals in 0.1 M HClO4 and in
0.1 M KOH, respectively. The absorbance is proportional to the concentration of molecules in the
interface. The absorbance for particular bands were plotted as a function of time in Figure 45a
and b, with the bands used to plot the change are indicated in the legend. Under basic conditions
(0.1 M KOH) a plot of the natural logarithm (ln) of the absorbance as a function of time is linear
(see inset Figure 45b), and therefore exhibits (pseudo-) first-order kinetics with an apparent
hydrolysis rate of k = 3.77 ´ 10-5 ± 0.34 s-1, which corresponds to a half-life of 5.5. hours. Under
acidic conditions (0.1 M HClO4) a higher-order reaction takes place (Figure 45a)
Figure 46
Schematic representation of possible mechanisms for acid-catalysed (A) and base-catalysed (B) hydrolysis of
electrografted diazonium interfaces on oxide materials, such as ATO.
Assuming hydrolysis of Sn-O-C bonds at the interface of the ATO, a classical SN2-type hydrolysis
is not expected to take place, neither under acidic nor basic conditions, because of the
impossibility of symmetry inversion at the interface of the bulk SnO2 and the linker molecule. A
more plausible mechanism for acid-catalysed hydrolysis involves flank-side attack of the Sn atom
by a water molecule without inversion of symmetry, in which the oxygen atom is protonated in a
first step, as illustrated by mechanism A in Figure 46 and similar to mechanisms proposed for the
hydrolysis of silicates278, where Si has a similar Lewis acidity to Sn. A possible mechanism for
base-catalysed hydrolysis may involve a similar attack of the Sn atom by a hydroxide anion,
resulting in the formation of a 5-coordinate intermediate and a subsequent displacement and
removal of the phenoxide anion, again, as also proposed for silicates.278
126
It is expected that hydrolysis rates will differ depending on the nature of the metal to which the
molecule is grafted (e.g. Sn/Sb/In/Ti etc.), as well as the type of crystallographic facet to which
the anchoring group is bound. In the case where other diazonium salts are used, inductive effects
depending on the nature of the phenyl-substituents (in this case imidazole-1-yl) are also expected
to affect the reactivity of the Sn-O-C bond towards hydrolysis. It is known that the reactivity of
diazonium salts to certain reactions differ depending on the inductive effect of substituents.181
Figure 47
ATR-IR absorbance spectra recorded at 163 mV (vs Ag/AgCl 3M) of Im-interfaces electrografted on me-ATO in
pH 7 (0.1 M PB) after 2-minute-long applications of increasingly (a) cathodic or (b) anodic potential steps. Inset:
plots of band intensities at 1247 and 1306 cm-1 versus applied potential.
Spectroelectrochemical titrations were also carried out in a similar fashion to measure and
determine the electrochemical stability window of the electrografted interfaces on ATO at pH 7
(i.e. in 0.1 M PB electrolyte). Different samples were used for the investigation of the cathodic
and anodic potential regions. Spectra were recorded at an applied potential of 163 mV (vs
Ag/AgCl 3M), which is close to the initial OCP of the functionalised ATO electrode, after having
applied an increasingly positive or negative potential step for 2 minutes each (Figure 47 a and b).
The band intensities of two d(C-H) modes at 1247 and 1306 cm-1 are plotted versus the applied
potential in the insets of Figure 47. These plots indicate only a small decrease in intensity over
the measured potential ranges. This decrease may be attributed to the loss of physisorbed
species, as has been previously reported for other electrografted diazonium interfaces 215,275, or
simply to small changes in the baseline. The electrochemical cleavage of the Au-C interfacial
bond in electrografted diazonium interfaces on gold was observed as a rapid decrease in
absorbance intensities in the IR spectra, as shown in Figure 7e in 7.3, and is not observed in this
case. The electrografted diazonium interfaces on ATO studied here therefore exhibit high
stabilities in a broad potential window (at least -0.73 to 2.23 V vs RHE) at pH 7 (0.1 M PB) that
encompasses the intrinsic electrochemical stability window of ATO (-0.77 to 1.83 V vs RHE)
measured at pH 7 (0.1 M PB) (see Appendix 8). This also encompasses the stability window of
ITO and FTO electrode materials (-0.62 to 1.96 V vs RHE and -0.51 to 1.73 V vs RHE,
127
respectively, as determined by Jaramillo and co-workers) in neutral electrolyte (0.1 M sodium
acetate).226 The stability of these electrografted diazonium interfaces in such a wide potential
window broadens the possibilities for conducting electrocatalysis using covalently immobilised
species on TCO materials.
Figure 48
ATR-IR absorbance spectra of Im-interfaces electrografted on me-ATO in acetonitrile (0.1 M TBAP) after 2-
minute-long applications of increasingly (a) cathodic or (b) anodic potential steps. Insets: plots of the
corresponding band intensities at 1608 cm-1 versus applied potential.
Further measurements were conducted in organic media (acetonitrile, 0.1 M TBAP), as shown in
Figure 48 a and b. The interface is stable over the entire anodic potential range probed from at
least 1.6 V until -1.3 V (vs Fc/Fc+), after which desorption is observed.
To summarise, in situ IR spectroscopic and spectroelectrochemical methods were used to
determine the hydrolytic and electrochemical stability windows of electrografted diazonium
interfaces on ATO. Previous attempts to probe the stability of interfaces on oxides applied mainly
UV-Vis spectroscopy or electrochemical techniques.25,36,231,279,280 Using instead in situ IR
spectroscopy allows one to probe the hydrolytic and electrochemical stability of an anchoring
group without having to rely on chromophoric or electroactive components (which may
themselves be unstable for other reasons) and allows the intrinsic stability of specific
linker/anchoring groups to be determined. The high surface area of the me-ATO films leads to
very high IR absorption band intensities that usually can’t be obtained without some
surface/plasmonic enhancement, as is the case for the Au SEIRA electrodes. For comparison,
the IR absorbance spectra obtained for planar ITO films in chapter 5 just about allow for
characterisation of interfaces formed at their surface.
128
6.4 Fe-Hangman complexes immobilised on planar and porous TCOs using the
electrografting of a diazonium salt for oxygen reduction reaction
Scheme 54
Scheme depicting the different immobilisation strategies employed for grafting/immobilising FePOMe on an
electrode: (A) a two-step ‘post-coordination’ process whereby the electrode is first electrochemically grafted with a
linker molecule Im-diazo, followed by incubation with FePOMe, and (B) a one-step ‘pre-coordination’ process
where the electrode is electrochemically grafted with Im-diazo and FePOMe together. The chemical structure of
the iron Hangman porphyrins is depicted by 1.
As previously mentioned, hangman porphyrin complexes contain a heme-group with a rigid
xanthene scaffold at one of the meso-positions that contains a “hanging” group, which may
introduce certain non-covalent interactions in the second coordination sphere that can alter the
behaviour or activity of the complex.129,251 Iron hangman porphyrins containing a carboxylic acid
or methyl ester group (hereby denoted FePOH and FePOMe, respectively) were synthesised per
published procedures in the group of Dr. M. Schwalbe.127,128 The same complexes were
previously immobilised on SAM-coated silver and gold electrodes by Ly et al251, whereby it was
suggested that the acid group takes part in a proton coupled electron transfer reaction that may
be beneficial in ORR. For these immobilisation studies on TCOs, the ester-containing FePOMe
was used in order to exclude the possibility of the porphyrin binding to oxide surfaces via the
carboxylic acid hanging group.
129
Figure 49
CVs of 1 mM solutions of Im-diazo in acetonitrile (0.1 M TBAP) using me-ATO as a working electrode, with
(purple trace) and without (green trace) addition of 1.2 mM FePOMe. Scan rate = 50 mV/s.
Two distinct immobilisation strategies were developed and employed for immobilising FePOMe
on me-ATO: (A) a two-step ‘post-coordination’ process whereby me-ATO is first electrochemically
grafted with 1 mM Im-diazo in acetonitrile (in a fashion similar to that done in 9.3), thereby
providing a ‘ligand’ terminated interface for axial coordination through the imidazole, followed by
incubation with 1.2 mM FePOMe (Scheme 54A); and (B) a one-step ‘pre-coordination’ process
whereby me-ATO is electrochemically grafted with 1 mM Im-diazo and 1.2 mM FePOMe
together, after mixing them for 5 min i.e. pre-coordinating the iron centre to the axial ligand
(Scheme 54B). In each case, a potential of -0.42 V (vs Fc/Fc+) is applied to the me-ATO working
electrode for 120 s, corresponding to the reduction potential of Im-diazo. CVs of 1 mM Im-diazo in
acetonitrile (0.1 M TBAP), as well as 1 mM Im-diazo and 1.2 mM FePOMe in acetonitrile (0.1 M
TBAP), using me-ATO as a working electrode are shown in Figure 49. An irreversible Im-diazo
reduction peak is clear in the trace for a 1 mM Im-diazo solution, while the trace for a 1 mM Im-
diazo and 1.2 mM FePOMe solution shows the Im-diazo reduction peak overlapping with the
reversible redox peak for the Fe2+/3+ couple of the FePOMe.
130
Figure 50
CVs of a 1.2 mM solution of 1.2 FePOMe in acetonitrile (0.1 M TBAP) using me-ATO as a working electrode (pink
trace). The electrode was rinsed with acetonitrile and a second blank CV was recorded (dashed black trace).
Scan rate = 50 mV/s.
A CV of 1 mM FePOMe without any Im-diazo is shown in Figure 50. A semi-reversible redox
peak for the Fe2+/3+ couple that is around 0.3 V more negative than that measured in the
presence of Im-diazo. It was shown for iron protoporphyrin IX immobilised on either pristine
MWCNTs (via pi-stacking) or on imidazole-functionalised polypyrrole coated MWCNTs (via axial
coordination to the pyrrole) that the Fe2+/3+ redox potential shifts by 0.1 V to more positive
potentials upon coordination with the imidazole terminated surface.281 The 0.3 V shift to more
positive potentials for the FePOMe in the presence of Im-diazo can thus be attributed to the
coordination and/or immobilisation of the FePOMe by Im-species, as well as the otherwise weak
or repulsive interaction between the hydrophobic uncoordinated FePOMe and the hydrophilic
oxide surface. After rinsing the electrode with acetonitrile and applying another CV, it is possible
to see that no FePOMe adsorbed onto the surface of the unmodified ATO.
Figure 51
CVs of 1 mM ferrocene in acetonitrile (0.1 M TBAP) using unmodified me-ATO and me-ATO electrografted with
Im-diazo using -0.42 V (vs Fc/Fc+) for 120 s. Scan rate = 50 mV/s.
131
Copious rinsing in acetonitrile and dichloromethane is used to remove non-covalently bound
species. Passivation of a me-ATO electrode electrochemically grafted with Im-diazo (i.e. before
incubating with FePOMe) is indicated by comparative CV measurements of an unmodified me-
ATO electrode and an electrografted me-ATO electrode in the presence of the
ferrocene/ferrocenium (Fe/Fc+) redox probe (Figure 51). In addition to an increase in peak
separation ΔEP, a significant decrease of 52% in the ferrocene oxidation current is observed for
the electrografted me-ATO electrode, confirming the passivation of the ATO surface.
Figure 52
(a) Photograph of me-ATO coated pl-ITO electrodes after immobilisation of FePOMe using the ‘pre-coordination’
(top) or ‘post-coordination method (bottom) and (b) CVs in acetonitrile (0.1 M TBAP) of FePOMe immobilised on
me-ATO electrografted using the ‘post-coordination’ method (labelled A, green trace) and the ‘pre-coordination’
method (labelled B, green trace). Scan rate = 50 mV/s.
Both the ‘post-coordination’ and ‘pre-coordination’ methods for immobilising FePOMe on me-ATO
result in a dark brown colouration of the films that is resistant to rinsing and soaking in acetonitrile
and dichloromethane, with a darker colour achieved using the ‘pre-coordination’ approach (see
photograph in Figure 52a and UV-Vis spectra in Figure 60). The darker colour and higher
adsorption in the UV-Vis spectra indicate a higher loading of FePOMe using the ‘pre-coordination’
approach.
CVs in acetonitrile of me-ATO modified via ‘post-coordination’ (method A) and ‘pre-coordination’
(method B) are shown in Figure 52b. The CVs stabilise over several cycles and show reversible
redox peaks close to 0.56 V vs. Fc/Fc+ for the Fe(2+/3+) redox couple of the immobilised FePOMe.
Peak currents (IP) increase linearly with applied scan rate for species immobilised via both the
‘post-coordination’ and ‘pre-coordination’ methods (Figure 53 a and b, respectively), as expected
for electron transfer to and from a surface-bound redox species.
132
Figure 53
Plot of peak currents Ip vs scan rates v for FePOMe immobilised on me-ATO via (a) the ‘post-coordination
method’ and (b) ‘pre-coordination’ method.
Surface coverages of electrochemically accessible FePOMe,
Γ
FePOMe, were calculated from
background subtracted CVs and found to be 1.5 and 3.0 × 10-9 mol.cm-2 for ‘post-coordination’
method and ‘pre-coordination’ method, respectively. The electrochemically accessible coverage
of FePOMe will henceforth be denoted by
Γ
CV to distinguish it from the coverage measured
using UV-Vis spectroscopy,
Γ
UV. As previously stated, electron transfer (ET) through insulating
organic interfaces proceeds via a tunnelling mechanism, in which the ET rate kET decays
exponentially with the distance between the redox centre and the electrode surface.282 It is well
known that electrochemical reduction of diazonium salts often leads to thick and often insulating
polymeric interfaces films on a broad range of materials45,55, as was observed on gold electrodes
in chapter 4, and ITO electrodes in chapter 5. Phosphonates, however, bind selectively to surface
oxide sites, meaning that interface formation is self-limiting and that multilayers of chemisorbed
phosphonate species do not readily form. The values of
Γ
CV obtained here for Im-diazo are in line
with those reported in the literature for molecular catalysts immobilised on other porous oxide
electrode materials using phosphonates, when taking into consideration differences in film
thicknesses (catalyst loading increases linearly with the film thickness). A comparison between
the values obtained here and those obtained in the literature (using electrochemical or
spectroscopic means) are summarised in Table 10.
133
Table 10
A comparison between surface coverages
Γ
CV of molecular catalysts and redox species on different porous oxide
electrodes of different film thicknesses from the literature.
Coverages (
Γ
UV) of FePOMe were also calculated for the same me-ATO electrodes using UV-
Vis spectroscopy.
Γ
UV of 1.2 and 3.2 × 10-9 molcm-2 were determined for the ‘post-coordination’
and ‘pre-coordination’ methods, respectively. These values are in line with those calculated using
CVs (
Γ
CV) and suggest that all of the immobilised species are electrochemically accessible using
the ‘post-coordination’ method, while around 75% of the species immobilised using the ‘pre-
coordination’ method are electroactive.
Full-widths at half maximum (FWHMs) of the oxidation and reduction waves, Ea fwhm and Ep fwhm,
accounting for 199 and 197 mV for ‘post-coordination’ and 210 and 211 mV for ‘pre-coordination’,
respectively, are found to be greater than the theoretical 90.6 mV17 expected for a 1-electron
transfer process. For increased coverages, broadening of the FWHMs of the oxidation and
reduction wave are indicative of so-called surface charge effects, which are electrostatic
Redox species
Electrode
Film
thickness
(
µ
µ
M)
Coverage
(× 10-9
molcm-2)
Volume coverage
(× 10-5 molcm-3)
Ref.
[Ru(bpy)
2
(4,40-PO
3
H
2
-
bpy)](PF6)2
-Phosphonate
Porous ITO
0.55
2.5
15.7
5.5
25
160
(UV-Vis)
~ 10
283
[CoIIIBr
2
{(DO)(DOH)pn}]
-Phosphonate
Porous ITO
13
~ 150
(UV-Vis)
~12
70
Various Co catalysts
-Phosphonate
Porous ITO
13
22 – 28
(CV)
~ 2
247
[Ru(Mebimpy)(4,4′-((HO)
2
-
OPCH2)2bpy)(OH2)]2
+ - -
Phosphonate
Porous TiO
2
~ 10
53
(UV-Vis)
~ 5.3
234
NiII bis(diphosphine) complex
-Phosphonate
Porous TiO
2
4
146
(UV-Vis)
~ 37
227
fac-[MnBr(4,4’-bis(PO
3
H
2
)-
2,2’-bipyridine)(CO)3]
-Phosphonate
Porous TiO
2
6
34
(UV-Vis)
~ 6
173
[RuII(Mebimpy)(4,4′-
(PO3H2)2bpy) (OH2)]2+
-Phosphonate
Mesoporous ATO
2
248
(UV-Vis)
1.58
(CV)
~ 12
~ 8
267
Ferrocene carboxylic acid
(via amide-coupling to 3-
aminopropyltrietoxysilane
-Silane
Mesoporous ATO
0.21
0.9 – 0.6
(CV)
~ 3
240
FePOMe (‘post-coordination’)
-Diazonium
Mesoporous ATO
0.16
1.5
(UV-Vis)
1.5
(CV)
~ 9.4
~ 9.4
This
work
FePOMe (‘pre-coordination’)
-Diazonium
Mesoporous ATO
0.16
3.9
(UV-Vis)
3.0
(CV)
~ 24
~ 19
This
work
134
intermolecular interactions between charged species.284 Such behaviour has previously been
reported for immobilised porphyrins on silicon.285,286 As can be seen in Figure 52b, the slightly
broadened FWHMs observed in case of the ‘pre-coordination’ method are in line with the higher
electroactive porphyrin loading
Γ
CV and may as such indeed be related to enhanced crowding. In
addition, broadening may also reflect a certain heterogeneity of the imidazole coordination sites.
Given the propensity for diazonium salts to form thick interfaces during electrochemical grafting of
electrode surface, it is reasonable to expect that ‘pre-coordination’ with FePOMe would lead to
high values of
Γ
CV due to the electrografting of pre-coordinated species directly at the surface.
This would lead to a more even distribution of FePOMe species through the interface, as well as
reduce the distance between the FePOMe redox centres and the electrode surface compared to
the ‘post-coordination’ method, in which FePOMe coordinate on top of the already electrografted
Im-interface in a ‘monolayer-like’ fashion. If the Im-interface is too thick in the latter case, then the
charge transfer resistance might become too great to and could lead to low values of
Γ
CV. The
similarity between the values of
Γ
CV obtained on the me-ATO using both methods, which are
also in line with values of
Γ
CV obtained using phosphonates in the literature, suggest that thin
interfaces are electrografted on me-ATO from Im-diazo (even in the absence of FePOMe). This
compares to the thick, insulating interfaces previously electrografted on planar gold and ITO, and
may result from diffusion limitation of Im-diazo inside the porous structure of the high-surface
area me-ATO during the electrochemical modification step, thereby limiting the interface
thickness formed.
Figure 54
CVs in acetonitrile (0.1 M TBAP) of FePOMe immobilised on (a) pl-ITO using the ‘post-coordination’ method at
constant potential (labelled A, grey trace) or with a linear sweep (labelled Sweep, pink trace). (b) CVs comparing
the current densities obtained for FePOMe immobilised on pl-ITO and me-ATO (blue and green). Scan rate = 50
mV/s.
Applying the ‘post-coordination’ method on planar ITO (pl-ITO) results in a
Γ
CV of 0.5 × 10-10 mol
cm-2, which is 12% of the value expected for a perfectly packed monolayer of hangman porphyrin
species on a planar surface (4.2 × 10-10 molcm-2 – as calculated from dimensions129 determined
135
using single crystal X-ray crystallography) and indeed indicates the electrografting of a thick,
insulating interface (grey trace, Figure 54a). In the case of ITO, a potential of -0.56 V (vs Fc/Fc+)
was applied due to the more negative reduction potential of Im-diazo on ITO (Appendix 9), which
is approximately 0.14 V more negative than that observed on ATO. Applying instead a linear
sweep from 0.04 V to -0.86 V (vs Fc/Fc+) at 50 mV/s in the same concentration of Im-diazo in
acetonitrile results in a reproducible
Γ
CV of 1.2 × 10-10 molcm-2 (pink trace, Figure 54a),
equivalent to 29% of a monolayer of FePOMe. This value is in line with coverages given in
literature for molecular catalysts immobilised on planar ITO and FTO using phosphonate
anchoring groups.234,267,283 Values of
Γ
CV on ITO may be especially limited by electrochemical
‘dead spots’.35,287 As is the case for me-ATO, IP increase linearly with scan rate (Figure 57 b), as
expected for a surface-bound species.
Figure 55
CVs of 0.1 mM ferrocene in acetonitrile (0.1 M TBAP) using unmodified pl-ITO and pl-ITO modified with Im-diazo
after applying a constant potential of -0.56 (vs Fc/Fc+) for 120 s (violet trace) or linear sweep voltammetry from
0.04 V to -0.86 V (vs Fc/Fc+) (pink trace). Scan rate = 50 mV/s.
CVs of the ferrocene/ferrocenium (Fe/Fc+) redox probe before and after electrografting of pl-ITO
with Im-diazo using either electrochemical reduction at constant potential (-0.56 V vs Fc/Fc+) or
the application of a linear sweep from 0.04 V to -0.86 V (vs Fc/Fc+) show a decrease in the
ferrocene oxidation current of 61% and 29%, respectively. The lower decrease in current for the
sweep method indicates the electrografting of a thinner, less insulating interface. This, along with
the results obtained on mesoporous ATO, strongly suggests that a slower interface formation is
necessary to obtain high kET for immobilised species, whether it be a result of diffusion limitation
in a porous material or applied potential.
136
Figure 56
Plots of peak separation ΔEp vs the log of the scan rate log(v) for the Fe (2+/3+) redox couple of FePOMe
immobilised on me-ATO modified via (a) ‘post-coordination’ and (b) ‘pre-coordination’.
CVs of the FePOMe species immobilised on me-ATO were conducted at different scan rates
(Appendix 10). Peak separations ΔEP for the Fe(2+/3+) couple recorded at 10 mV/s for me-ATO
modified using ‘post-coordination’ and ‘pre-coordination’ are small (26 mV and 32 mV,
respectively), while at higher scan rates they deviate from a linear relationship (Figure 56a and
b), indicating a kinetic limitation. Theoretically ΔEP should be zero for redox processes at low
scan rates; however, ΔEP increases as the scan rate increases and the time taken to scan
through the peak becomes comparable or faster than the electron transfer rate.284 The apparent
electron transfer rate constants kc and ka for the cathodic and anodic peaks, as well as the
electron transfer coefficients αc and αa, were calculated using mathematical treatments devised
by Laviron.288 ‘Post-coordination’ results in kc = 4.4 ± 0.7 s-1, αc = 0.62, ka = 2.9 ± 0.1 s-1 and αa =
0.28, while ‘pre-coordination’ results in kc = 4.9 ± 0.1 s-1, αc = 0.81, ka = 2.8 ± 0.1 s-1 and αa =
0.14. The unsymmetrical energetic barrier for electron transfer, as indicated by α, can be
explained by structural changes to the porphyrin, e.g. the axial ligation, upon changes in the Fe-
centre oxidation state. The electron transfer rates calculated here are in line with those obtained
for other redox systems that are immobilised on nanostructured metal oxide electrodes (see
comparisons to literature in Table 11), including Fe-porphyrin containing Fe-mimochrome (4 s-1)
or microperoxidase-11 (10 s-1) immobilised on nanostructured ITO electrodes289,290,
microperoxidase-11 (1.5 s-1) immobilised on mesoporous ATO.291 Attempts to compare the ET
behaviour of FePOMe immobilised on me-ATO via the carboxylic acid anchoring group of the
analogous, carboxylic acid-containing linker molecule 4-(1H-Imidazol-1-yl)benzoic acid failed,
presumably due to the instability of the anchoring group. A similar attempt in the literature to
immobilise a ruthenium dye on me-ATO using a carboxylic acid failed for the same reason.292
137
Table 11
A comparison of the redox potentials E0 and electron transfer rates (kET) for the Fe(2+/3+) redox couple of the
system presented here and different molecular/enzymatic redox species from literature immobilised on different
porous oxide electrodes of different film thicknesses. Potentials are given versus RHE and were calculated using
the Nernst equation.
Figure 57
(a) CVs in acetonitrile (0.1 M TBAP) of FePOMe immobilised on pl-ITO using the linear sweep voltammetry
method and (b) a plot of peak separation Ep vs the log of the scan rate log(v) for the Fe (2+/3+) redox couple. Scan
rate = 50 mV/s.
Peak separation ΔEP for FePOMe immobilised on pl-ITO remains approximately the same at 45
mV from a scan rate of 10 mV/s up to 500 mV/s (Fig. 7 SI), above which faster measurements
could not be carried out, which meant that a kinetically limited kET could not be calculated. The
difference between the redox behaviour of FePOMe on the more conductive pl-ITO and the me-
ATO suggests that the ET rates calculated on me-ATO are limited by the resistivity of the porous
ATO support240,241 (aside from any possible difference in the intrinsic bulk conductivity of ATO
over ITO, the small size of the ATO nanocrystals result in more grain boundaries when
Redox species
Electrode
Film
thickness
(
µµ
M)
E˚(Fe2+/3+)
(V vs RHE)
k
ET
(s-1)
Ref.
Cytochrome C
Mesoporous
ITO
0.17
0.425
1.2
202
Cytochrome C
Mesoporous
ITO
2
0.670
12
204
Fe mimochrome
Mesoporous
ITO
1
0.274
4
289
Microperoxidase-11
Mesoporous
ITO
0.2
0.263
10
293
Cytochrome C
Mesoporous
SnO2
4
0.673
1
201
Fe tetra(2,6-
dihydroxyphenyl) porphyrin
Carbon
nanotubes
-
0.330
-
152
Microperoxidase-11
Mesoporous
ATO
0.42
0.274
1.5
291
FePOMe (A)
Mesoporous
ATO
0.16
0.429
4.4
This work
FePOMe (B)
Mesoporous
ATO
0.16
0.496
4.9
This work
138
assembled in thin films when compared to planar ITO films that are deposited using sputtering
techniques) rather than the charge transfer resistance of the interface. This limitation may also
explain why the values of kc and ka are the same for me-ATO modified using both the ‘pre-
coordination’ and ‘post-coordination’ methods, within the given values of error. For comparison
with another material, electron transfer rates of 0.35 s-1 were obtained for and Fe protoporphyrin
XI immobilised on MWCNTs281, while a rate of 2.9 s-1 was obtained for hemin immobilised on
MWCNTs in another study294. This compares to a transfer rate of 4.9 × 103 s-1 obtained for hemin
immobilised on a glassy carbon electrode295 and highlights the limited conductivity in 3D
structured electrodes compared to planar electrodes.
Comparing the
Γ
CV obtained for 160 nm thick me-ATO films electrografted with the ‘pre-
coordination’ and ‘post-coordination’ methods with planar ITO electrografted with linear sweep
voltammetry, we see a 25 and 12.5-fold increase, respectively. Comparing the same values with
those obtained for planar ITO electrografted with a constant potential, we a see a 30 and 60-fold
increase.
Figure 58
CVs in acetonitrile (0.1 M TBAP) of FePOMe immobilised on (a) me-ATO electrografted using the ‘pre-
coordination’ method B at -0.42 (vs Fc/Fc+) (orange trace) and -0.76 (vs Fc/Fc+) (green trace) and (b) pl-ITO
electrografted using the ‘pre-coordination’ method at -0.56 V (vs Fc/Fc+) (labelled B, purple trace) or
electrografted using linear sweep voltammetry (labelled Sweep, pink trace). Scan rate = 50 mV/s.
Electrografting of pl-ITO via the ‘pre-coordination’ method (i.e. in the presence of FePOMe)
results in a brown surface film that is easily removed/dissolved upon rinsing in dichloromethane,
and results in negligible redox current for the Fe(2+/3+) couple in aceteonitrile (0.1 M TBAP, Figure
58b), indicating the deposition of weakly bound, physisorbed species. A similar behaviour is
observed on me-ATO when a potential more negative (-0.42 V vs (vs Fc/Fc+) than the FePOMe
Fe(2+/3+) couple is applied in the ‘pre-coordination’ method. Given that the applied reduction
potential of Im-diazo on pl-ITO (-0.56 V vs vs Fc/Fc+) is more negative than the potential of the
FePOMe Fe(2+/3+) couple (close to -0.56 V vs vs Fc/Fc+), these results suggest that the
139
application of potentials at or below the redox potential of the species being immobilised results in
electron hopping through the Fe centres of already-grafted FePOMe species. This causes
diazonium species to be reduced away from the electrode surface, which may in turn reduce
other diazonium species or form oligomeric species that aren’t chemisorbed at the surface (as
illustrated in Scheme 55). Electron-hopping has previously been observed for diazonium-bearing
osmium and ruthenium species on a range of electrode materials and resulted in thick, polymeric
films.45,275 The electron transport pathway may take three different routes: through the conjugated
bonds of the macrocycle, via the metal atom, or a mixture of both of these pathways.284 Picot et
al. deposited films on glassy carbon (GC), pyrolised photoresist films (PPF) and ITO electrodes
via the electrochemical reduction of a diazonium meso-substituted free-base porphyrin in
aqueous solvent (0.1 M HCl) and observed the formation of thick, purple films (measured at 20
nm on PPF) that were insoluble in aqueous solvents but easily removed in organic solvents. The
solute was identified as non-chemisorbed H2TPP monomers and it is suggested that these form
as a result of the electrochemically generated radicals undergoing further reduction to carbanions
and abstract a hydrogen from water.296 While the presence of water in the acetonitrile used in the
experiments presented here cannot be excluded and is indeed quite likely, it is not plausible that
this mechanism takes place here as it was shown that an insulating interface was also deposited
on pl-ITO under a potential of -0.56 V (vs vs Fc/Fc+) in the absence of FePOMe (i.e. the radicals
formed are able to adsorb on the oxide surface and form covalent bonds). Identification of the
soluble material should be possible using mass spectrometry techniques and may provide
evidence for the type of mechanism taking place.
Scheme 55
Schematic representation of a possible mechanism for the formation of thick, weakly bound films on pl-ITO and
me-ATO electrodes when applying potentials more negative than the FePOMe Fe(2+/3+) redox couple. Reduced
FePOMe species (i.e. in oxidation state +2) at the electrode interface reduce Im-diazo or FePOMe species further
away from the interface via an electron ‘hopping’ mechanism, which eventually leads to radicals forming in
solution that react with other species in solution, rather than the oxide surface, finally leading to the formation of
some kinds of oligomeric species.
140
The high loadings of FePOMe that could be achieved on me-ATO result in very high resonance
Raman (RR) scattering intensities, thus allowing to ascertain the spin, coordination and redox
states of the porphyrin. For the RR investigation, FePOMe was immobilised on a me-ATO
electrode using the ‘post-functionalisation’ method. In situ RR measurements were performed
under different applied potential steps (Figure 59 a) and a component fit analysis was carried out
to decompose each potential-dependent spectrum into the different contributing ligation, spin and
redox components (Figure 59).251k
Figure 59
(a) Potential-dependent resonance Raman (RR) spectra recorded in pH 7 (0.1 M PB) of FePOMe immobilised on
me-ATO using the ‘post-coordination’ method and (b) the relative contributions from the different FePOMe
spectra components at different potentials.
The component fit analysis was conducted using the same methodology as that used by Ly et al.
for the iron Hangman porphyrins immobilised on imidazole-terminated SAMs on silver
electrodes.251 At OCP (135 mV vs 3M Ag/AgCl) in pH 7 (0.1 M PB), intense marker bands appear
around 1367 cm-1 (v4) and 1561 cm-1 (v2), and a lower intensity band appears around 1495 cm-1
(v3). The main spectral contribution observed comes from the 5-coordinated high-spin (HS) state
of the FePOMe with a small contribution from the 6-coordinated low-spin (LS) state, as evidenced
by a high-frequency shoulder in the v3 band at 1502 cm-1. This indicates that almost all of the
immobilised FePOMe species are axially coordinated to the surface of the functionalised me-ATO
via the electrografted imidazole linker molecule with a 6th free coordination site, which is available
for the binding of oxygen. 6-coordinated species were also overserved by Ly et al.251 and their
molecular origin is at present unknown. Plotting the different contributions versus applied
potential allows us to follow the change in oxidation state of the Fe centre and hence evaluate the
electrochemical accessibility of the immobilised species. As can be observed in Figure 59b, upon
application of cathodic potentials there is an almost complete reduction in intensity of the 5-
coordinated HS oxidised state, indicating that most of the immobilised species are indeed
electrochemically accessible (i.e. electrochemically contacted), despite passivation of the me-
ATO surface during electrografting. Application of spectroelectrochemical methods, such as that
outlined here, provides information on the evolution of immobilised species under different
potentials or reaction conditions, and potentially allows the identification of intermediate species
in catalytic reactions. It further provides precise information on the electrochemical accessibility of
141
the immobilised species (at least in the case of porphyrin-containing species), which may be used
to effectively evaluate different immobilisation strategies.
Figure 60
Spectroelectrochemical UV-vis measurements of FePOMe immobilised on me-ATO using (a) ‘post-coordination’
and (b) ‘pre-coordination’ methods in pH 7 (0.1 M PBS).
UV-Vis spectroelectrochemistry may be used to measure redox changes of certain species, such
as molecular catalysts, immobilised on TCO material surfaces; however, UV-Vis spectroscopy
cannot be used to quantitatively measure the contributions from different redox states of
immobilised porphyrin-type species in the same way as rR spectroscopy due to the broad overlap
of the Soret bands (Figure 60 a, b).
Figure 61
CVs of unmodified me-ATO at pH 7 (0.1 M PB) under Ar and O2 bubbling, as well as me-ATO with FePOMe
immobilised using the ‘post-coordination’ (labelled A, blue trace) and ‘pre-coordination’ (labelled B, green trace)
methods.
To evaluate the electrochemical performance of the modified thin film electrodes for ORR, CV’s
of unmodified and FePOMe-modified electrodes were recorded at pH 7 (0.1 M PB) in nitrogen
and oxygen saturated electrolyte (Figure 61). While unmodified me-ATO shows some cathodic
current in O2 with an onset potential of around 0.4 V (vs RHE), the me-ATO electrodes
142
immobilised with FePOMe display diffusion limited reductive waves with onset potentials of 0.570
and 0.645 V (vs RHE) and maximum currents of 0.9 and 1 mA/cm2 for the ‘post-coordination’ and
‘pre-coordination’ methods, indicating that the immobilised FePOMe-species is catalytically active
for ORR.
Figure 62
CVs of pl-ITO and me-ATO electrodes in pH 7 (0.1 M PB) electrolyte under O2 bubbling with increasing loadings
of immobilised FePOMe.
Table 12
A comparison of the onset potentials for the ORR for FePOMe immobilised on pl-ITO and me-ATO using different
immobilisation methods and the corresponding coverages of FePOMe as determined using CV.
Coverage FePOMe
Γ
Γ
CV
(× 10-9 molcm-2)
Onset potential
(mV) vs. RHE
pl-ITO – ‘post-coordination’
method, constant potential
0.05
-
pl-ITO – ‘post-coordination’
method, linear sweep
0.12
500
me-ATO – ‘post-coordination’
method
1.5
570
me-ATO – ‘pre-coordination’
method
3.0
645
To determine the influence of the catalyst electrode loading on the catalytic performance, CVs
were recorded for the pl-ITO electrodes modified with FePOMe using the ‘post-coordination’ and
‘sweep’ methods. CVs for both the me-ATO and pl-ITO electrodes are shown in Figure 62. The
onset potentials for ORR and the corresponding coverages of FePOMe (
Γ
CV) for each electrode
are shown in Table 12. FePOMe-modified electrodes show improved onset potentials for ORR
with increasing coverages of electroactive FePOMe (i.e. catalyst loading) in the order me-ATO
‘pre-coordination’ > me-ATO ‘post-coordination’ > pl-ITO ‘post-coordination’, with onset potentials
of around 645 mV, 570 mV and 500 mV (vs RHE), respectively. It has previously been shown
that overpotentials for ORR reduce with the amount of metal complex present on the surface,
143
with ORR currents shown to be directly proportional to the amount of complex, indicating that the
reaction is first order with respect to the surface concentration of the catalyst.116,297,298 This is true
for cases where multi-layers are not formed and all of the metal active centres are accessible and
available for catalysis, as is the case shown here. pl-ITO modified using the ‘post-coordination’
method shows negligible ORR activity, presumably due to the very low loading of FePOMe.
Figure 63
CVs of FePOMe immobilised on me-ATO using (a) ‘post-coordination’ method (labelled A, blue traces) and (b)
‘pre-coordination’ method (labelled B, green traces) in aqueous media pH 7 (0.1 M PB) under Ar and O2. Scan
rate = 50 mV/s.
The redox peak potential of a process E1/2 can be shifted to more positive potentials when
immobilised on a surface.284 In the surface activity theory of Brown and Anson, the anodic and
cathodic peak potentials are considered as equal and related to the surface coverage of redox
species (
Γ
) by:
𝐸𝐸°{§ò = 𝐸𝐸w+𝑅𝑅𝑇𝑇
𝑛𝑛𝐹𝐹𝑟𝑟ú− 𝑟𝑟X𝛤𝛤 (45)
where rR and rO are interaction parameters that represent the intermolecular interactions between
reduced and oxidised species, provided that the mixed interactions are the same, i.e. the
oxidised-reduced and reduced-oxidised interactions.299,300 As the anodic and cathodic peak
potentials are not equal, the average formal potential E01/2 is used instead of E0, and the theory
predicts that the peak potential will shift depending on the relative magnitude of rR and rO as the
coverage changes. Abruña and co-workers showed that for a SAM of [Os(bpy)2(dipy)]1+ on
platinum electrodes (with bpy = 2,2’-bipyridine and dipy = 4,4’-trimethylenedipyridine), E1/2
increases as a function of surface coverage in a range of solvents. Coupled with a broader than
ideal (i.e. greater than 90.6 mV) ∆EFWHM, which implies negative interaction parameters, this
indicates, as expected, that the repulsion between the +2 oxidised species is greater than
between the reduced +1 species,300 and indicates that the thermodynamic driving force for the
144
redox transition increases with coverage. Lindsey, Bocian and co-workers showed for porphyrins
bound to metal and semiconductor surfaces via meso-substituents that ∆EFWHM and E1/2 also both
increase as a function of surface coverage.301–303 As shown previously in acetonitrile (with 0.1 M
TBAP electrolyte), ∆EFWHM increases with the coverage of FePOMe on me-ATO, while E1/2
slightly decreases by 3 mV from 36 mV for a coverage of 1.5 × 10-9 molcm-2 (for the ‘post-
coordination’ method) to 33 mV for a coverage ΓCV of 3.0 × 10-9 molcm-2 (for the ‘pre-
coordination’ method). However, as can be seen in Figure 63, in aqueous media (with 0.1 M PB
electrolyte), there is an increase in ∆EFWHM and an increase in E1/2 of almost 70 mV with an
increased coverage of FePOMe, as expected, and this explains why there is a shift in the onset
potential (of 75 mV) for ORR with an increased loading of FePOMe. The insignificant decrease in
E1/2 of 3 mV observed in acetonitrile with an increased loading of FePOMe may be a result of
more complex structural differences, e.g. the presence of multi-layers.
The values of ∆EFWHM for CVs performed in aqueous media are much broader than those
measured in acetonitrile at the same scan rates (see Appendix 10 and Appendix 11). This
phenomenon may be explained by an increase in the strength of the ion pairing between the
FePOMe and the TBAP perchlorate anion, or the intercalation of the less polar acetonitrile
solvent molecules between the FePOMe species, both of which would minimise intermolecular
repulsions between the FePOMe redox species and each other and lead to lower values of
∆EFWHM. Collman, Chidsey, Hal Van Ryswyk and co-workers showed for a ruthenium porphyrin
with hydrophobic toluyl meso-substituents, axially coordinated to imidazole-terminated SAM
coated gold electrodes, that an increase in the strength of the ion pairing between the redox
species and the electrolyte anion can indeed minimise intermolecular repulsion, thereby reducing
∆EFWHM and leading to a decrease in E1/2. In water in particular, hydrophilic electrolytes led to
waves with broadened values of ∆EFWHM that are difficult to differentiate from the baseline.304
Figure 64
CVs of me-ATO in pH 7 (0.1 M PB) and pH 12.8 (0.1 M KOH) under O2 or Ar bubbling with FePOMe immobilised
using the ‘pre-coordination’ method.
145
The highest onset potential for the me-ATO modified electrodes was obtained in 0.1 M KOH (pH
12.8) using the ‘pre-coordination’ method, yielding an onset potential of 810 mV (vs RHE), as
shown in Figure 64. This compares well to previously reported onset potentials of 922 mV (vs
RHE) measured in O2 saturated 0.1 M KOH electrolytes for axial-pyridine coordinated iron
phthalocyanine reported by Liu and co-workers150, or 913 mV for axial-imidazole coordinated iron
porphyrin immobilised on carbon nanotubes by Cho and co-workers.151 In both cases diazonium
salts were used to chemically functionalise (i.e. not electrochemically, as is the case here) the
carbon nanotubes during the catalyst ink preparation, which is possible for carbonaceous
materials. Onset potentials in 0.1 M HClO4 could not be measured for FePOMe immobilised on
me-ATO due to the instability of the interface to hydrolysis at such an acidic pH. Aside from the
activity of the porphyrin itself, a number of other factors may affect the activity of such species
immobilised on nanostructured oxides, such as me-ATO, or CNTs. These include: different
substrate conductivities and electronic coupling between the catalyst and the substrate, which
may affect the electron transfer rate; completely different substrate geometries or different
interface structures, potentially influencing intermolecular interactions between the immobilised
catalytic species; and finally, different loadings of catalyst. The amount of catalyst present in the
herein presented me-ATO system is up to 3.5 μg/cm2 of FePOMe, which is an about two orders
of magnitude less than the 318 μg/cm2 loading found in the work of Cho and co-workers and the
1 mg/cm2 loading found in the work of Liu and co-workers. Other studies have reported onset
potentials of 820 mV305 and 880 mV306 for iron phthalocyanine supported on carbon electrodes
with loadings of 800 and 200 μg/cm2 of catalyst, respectively. Improved onset potentials and
activities could potentially be achieved using thicker films of me-ATO, which can easily be
accomplished using different coating conditions, or other nanostructured TCO films.
Figure 65
(a) CVs of me-ATO in pH 7 (0.1 M PB) electrolyte under O2 bubbling with FePOMe immobilised using ‘post-
coordination’ method (labelled A, blue trace) and ‘pre-coordination’ method (labelled B, green trace). (b)
Chronoamperometry at -0.32 V (vs RHE) for 10 min under O2 bubbling of me-ATO immobilised with FePOMe
using ‘post-coordination’. Inset shows second chronoamperometry applied for a second time to the same
electrode for a further 30 min.
146
Figure 65a shows the first 4 CVs of the me-ATO electrodes immobilised with FePOMe using the
‘pre-coordination’ and ‘post-coordination’ methods. The onset potential for the me-ATO modified
using ‘post-coordination’ method shows significantly better stability upon repeated cycling
compared to the ‘pre-coordination’ method. While the reason for this behaviour currently remains
unclear, it may be related to the previously discussed increased ‘crowding’ of FePOMe species in
the interface deposited using the ‘pre-coordination’ method, which may lead to a re-structuring or
displacement of electro-accessible and catalytically active species over repeated cycling. The
stability of the catalytic species and interface were further investigated using chronoamperometry
recorded at a potential of -323 mV (vs RHE) for the me-ATO electrodes modified using the ‘post-
coordination’ method, with measurements recorded before and after incubation with the
catalytically active FePOMe species (Figure 65b). The current density decreases over time,
which may be a result of: (a) desorption of the interface linker molecules from the oxide surface,
(b)disruption of the linker-catalyst axial coordination molecules (e.g. due to protonation of the
imidazole250 or a change in the orbital overlap resulting from the formation of intermediate Fe-oxo
species during catalysis), or (c) deactivation of the catalytically active species (e.g. due to the
formation of superoxides during catalysis). Imidazole protonation may be enhanced on
hydrophilic substrates such as me-ATO when compared to hydrophobic substrates such as
CNTs. Another reason for the loss of activity may be due to the formation of H2O2 due to ORR
taking place on the ATO (see Figure 61). It has been shown that ITO predominantly forms H2O2
via a 2-electron process during ORR at highly cathodic potentials.307 Further experiments would
need to be carried out to determine the exact reason for this apparent deactivation and loss of
catalytic activity.
147
Figure 66
(a) CVs in acetonitrile (0.1 M TBAP) of FePOMe immobilised on a me-ATO-coated ITO electrode (blue trace) and
a me-ATO-coated GC RDE electrode (pink trace) using the ‘post-coordination’ method (labelled method A), scan
rate = 50 mV/s. (b) LSVs in pH 7 (0.1 M PB) electrolyte under O2 bubbling, and a CV under Ar bubbling, of the
FePOMe immobilised on me-ATO coated on the RDE electrode at different rotation rates, as well as a CV the
unmodified me-ATO under O2 bubbling. Scan rates = 10 mV/s. (c) Tafel plot of the applied potential against the
logarithm of the current for the oxygen reduction reaction (ORR) using FePOMe immobilised on me-ATO coated
on the RDE electrode.
To study the ORR electrocatalytic behaviour and kinetics, a rotating disk electrode (RDE) was
used. me-ATO was coated on the GC RDE electrode using spin-coating, and subsequently
FePOMe was immobilised on its surface using the ‘post-coordination’ method. It was possible to
see by eye that the films were not even in thickness across the whole surface of the GC electrode
and this was likely a result of the GC electrode surface not being completely flat, or due to the
spin-coating procedure. Film thicknesses could not be determined for this electrode. Figure 66a
shows a CV of FePOMe immobilised on the me-ATO coated GC electrode using the ‘post-
coordination’ method compared to a CV of FePOMe immobilised on the me-ATO films on ITO
using the same method. A higher ΓCV of 2.1 × 10-9 molcm-2 was calculated for this electrode using
the geometrical surface area of the GC. Figure 66b shows the linear sweep voltammograms
(LSVs) of the electrode in pH 7 (0.1 M PB) under O2 at different rotation rates. At a rotation rate
of 400 rpm, an onset potential of 730 mV for ORR was obtained, which is 160 mV greater than
that obtained for me-ATO coated ITO electrodes. This difference is likely due to differences in the
me-ATO structure or film thickness, as discussed previously. Slower scan rates of 10 mV/s were
148
used for the LSVs, so the increase in onset potential after each LSV is more pronounced. Due to
this instability, it was not possible to carry out a Koutecký-Levich analysis to determine the mass
transport parameters or the standard rate constant. Currents for ORR should plateau at higher
overpotentials due to diffusion limitation to the catalyst; however, as can be seen in Figure 66b,
the currents do not plateau. This can be attributed to ORR taking place on the ATO substrate
itself at higher overpotentials. A CV of unmodified me-ATO in pH 7 under O2 in Figure 66 (dotted
black trace) shows this behaviour more clearly. Tafel plots give a linear slope with no significant
curvature at low overpotentials, which is similar to those given for other iron porphyrin or
phthalocyanine species immobilised on carbon electrodes via coordination to imidazole or
pyridine interfaces.150,151,250 A slope of 122 mV per decade is obtained, which compares to
around 27 mV per decade for iron phthalocyanine immobilised on CNTs via axial coordination to
pyridine, 67 mV for an iron porphyrin immobilised on CNTs via axial coordination to imidazole and
100 mV for an iron porphyrin immobilised on Vulcan (a carbon material) via axial coordination to
imidazole.150,151,250 Axial coordination to imidazole was shown known to enhance activity for ORR
and improve its selectivity towards water (as opposed to H2O2) in the latter study.250
Figure 67
CVs in acetonitrile (0.1 M TBAP) of FePOMe immobilised on me-ATO using ‘post-coordination’ before and after
cathodic polarisation in pH 7 (0.1 M PB) in air for ca. 2 hrs in pH 7, and then again after re-incubating with fresh
FePOMe.
Despite the exact reason for the deactivation of the catalytic interface remaining unclear thus far,
CVs of me-ATO electrodes modified using the ‘post-coordination’ method before and after
cathodic polarisation for ca. 2 hours show that the lost current density for the Fe(2+/3+) couple
could be restored by re-incubating the electrode in 1 mM FePOMe in acetonitrile (0.1 M TBAP).
This result therefore clearly emphasises the robustness of the grafted Im-diazo interface itself, as
well as the strong bond between the electrografted linker and the metal oxide surface; a result
which is in line with the ATR-IR spectroelectrochemical stability measurements performed
previously (vide supra).
149
6.5 Co Hangman complex immobilised on me-ATO using the electrografting of a diazonium
salt for the oxygen evolution reaction
Cobalt tetrapyrrole species have previously been shown to act as molecular OER catalysts,
whether in solution253–255, immobilised on planar FTO via drop-coating256,257, on bismuth vanadate
via carboxylate anchoring groups located on the porphyrin meso-substituent308, or on CNTs258. In
particular, Cao and co-workers showed for films of porphyrin species immobilised on FTO via
drop-coating that introducing axial ligands at the porphyrin cobalt centre can have a pronounced
effect on the onset potential for OER that is dependent on the electron donating ability of the
ligand.257 As previously mentioned, the rigid xanthene scaffold at one of the meso-positions of the
Hangman porphyrin complexes contains a “hanging” functional group that may introduce certain
non-covalent interactions that can alter the behaviour or activity of the complex. Nocera and co-
workers immobilised cobalt hangman corroles with carboxylic acid hanging groups on planar FTO
electrodes in Nafion and suggest that pre-organisation of water molecules by the hanging group
is helpful for the OER.309 The cobalt Hangman analogue of the FePOMe used in section 9.4
containing a carboxylic acid hanging group instead of a methyl ester is shown in Scheme 56 and
hereby denoted CoPOH. CoPOH has been shown to be an active catalyst for the HER310–312,
however no studies have been published regarding the use of this species for the OER.
Scheme 56
Scheme depicting the strategy employed for immobilising CoPOH on a me-ATO electrode. A two-step ‘post-
coordination’ process is followed whereby the electrode is first electrochemically grafted with a linker molecule Im-
diazo, proceeded by incubation with CoPOH. The chemical structure of the CoPOH is depicted by 1.
An attempt was made to immobilise CoPOH on me-ATO in the same way as the FePOMe
Hangman complex and use the modified electrode for the HER. CoPOH was immobilised on a
me-ATO electrode using the ‘post-functionalisation’ method, as illustrated in Scheme 1. This
method was chosen over the ‘pre-functionalisation’ method as it resulted in a more stable system
in terms of onset potential for the ORR with immobilised FePOMe (see Figure 65, page 145).
After incubating the modified me-ATO in a 1 mM solution of CoPOH, the initially colourless
electrode becomes brown/yellow. The colour is also resistant to rinsing and soaking in different
solvents. In order to check that the CoPOH is bound to the me-ATO via the Im-linker molecules, 1
mM CoPOH in acetonitrile was also drop-cast onto an unmodified me-ATO electrode and allowed
to dry for several hours and then rinsed in dichloromethane. There was no visible colouration of
150
the me-ATO, therefore the sample was measured using resonance Raman using 413 nm
excitation. As can be seen in Figure 68, the resonance Raman intensity is much greater for the
me-ATO electrode that has been electrografted with the Im-linker molecules and incubated with
CoPOH than for the drop-cast sample, indicating that the Im functionality is required to immobilise
the CoPOH successfully in significant amounts. This suggests that the CoPOH is immobilised via
axial ligation of Im to the porphyrin Co-centre.
Figure 68
Resonance Raman spectra recorded in air of CoPOH immobilised on Im-functionalised me-ATO using the ‘post-
coordination method (blue), and CoPOH drop cast onto me-ATO (from a 1 mM solution of CoPOH in acetonitrile)
and subsequently rinsed with acetonitrile (green). The inset shows both spectra normalised to the 𝑣𝑣÷ vibrational
band.
The inset of shows the resonance Raman spectra with intensities normalised on the 𝑣𝑣 vibrational
band. Excluding a broad peak around 1100 cm-1 due to the substrate, the peak positions and the
peak intensity ratios are roughly the same in both spectra, indicating that both species are in the
same oxidation state and have the same coordination number. Intense marker bands appear at
1370 cm-1 (𝑣𝑣÷) and 1569 cm-1 (𝑣𝑣9), and a lower intensity band appears around 1499 cm-1 (𝑣𝑣◊).
The 𝑣𝑣÷ vibrational mode, which is mostly composed of pyrrolic C-N stretching coordinates, is
highly sensitive to the metal centre oxidation state, and the band observed at 1370 cm-1 indicates
an oxidation state of either Co2+ or Co3+.313 A E0 (Co2+/3+) of 0.085 V (vs RHE) was measured by
Kielb314 for cobalt Hangman species with a methyl ester hanging group (CoPOMe), and therefore
at the measured OCP of around 0.93 V (vs RHE), the CoPOH species are expected to be in the
Co3+ oxidation state.
The frequencies of the 𝑣𝑣◊ and 𝑣𝑣9 vibrational modes, which are composed of pyrrolic C-C
stretching coordinates, are sensitive to the coordination and spin state of the metal centre, and
can be correlated to the core sise, which is the distance from the centre of the porphyrin core to
the pyrrole nitrogens (𝑑𝑑K“Qÿ), according to:
𝑣𝑣 = 𝐾𝐾 𝐴𝐴 − 𝑑𝑑K“Qÿ (46)
151
where 𝐴𝐴 and 𝐾𝐾 are mode-specific empirical constants.315 A change in the metal d-orbital
population, e.g. as a result of the substitution of a strong ligand for a weaker ligand, can induce a
change in the spin state of the metal, resulting in a change in the ion radius and a change in the
Ct-N distance, and therefore a shift in frequency. The positions of the bands observed here
indicate that the main spectral contribution is from species that are 6 coordinated low-spin
(LS)316, indicating axial coordination by two solvent molecules. This suggests that, in the case of
the drop-cast sample, the non-specifically adsorbed CoPOH species are coordinated by two
acetonitrile solvent molecules, while in the case of the ‘post-coordination’ sample, the specifically
adsorbed CoPOH species must be coordinated by a surface-grafted Im-linker molecule and an
acetonitrile solvent molecule, as illustrated in Scheme 57. Imidazole and acetonitrile ligands have
a similar ligand-field splitting parameter and are thus very likely to result in the same spin state. A
coverage (
Γ
UV) of CoPOH on the me-ATO of 1.2 × 10-9 molcm-2 was estimated with UV-Vis
spectroscopy using the adsorption coefficient (
ε
) of 1.69 × 104 M-1cm-1 measured for FePOMe
(see UV-vis spectra in Figure 76).
Scheme 57
Schematic illustration of CoPOH molecules (a) non-specifically adsorbed on me-ATO and coordinated by two
acetonitrile solvent molecules, and (b) specifically adsorbed on me-ATO via coordination to a surface-grafted Im-
linker molecule and an acetonitrile solvent molecule.
An attempt was made to conduct HER using the immobilised species in 0.1 M NaClO4 (pH 5).
CVs of the CoPOH immobilised on the me-ATO are shown in Figure 69. No substantial current
for HER can be detected within the potential range probed. For HER to take place, cobalt must
first be reduced to the oxidation state CoI.252 A redox potential E0 (Co+/2+) of -0.325 V (vs RHE)
was measured by Kielb for CoPOMe.314
152
Figure 69
CVs in 0.1 M NaClO4 (pH 5) under Ar bubbling of me-ATO with (green trace) or without (dotted black trace)
CoPOH immobilised using the ‘post coordination’ method. Scan rate = 50 mV/s.
As previously mentioned, cobalt tetrapyrrole compounds have been shown to be good catalysts
for OER, particularly under basic conditions. Resonance Raman spectra of the CoPOH species
immobilised on the me-ATO via the post-coordination method were measured in 0.1 M NaOH,
and, as shown in Figure 70, no substantial changes in the band positions or peak intensity ratios
can be observed in comparison to a spectrum recorded in air. It should be noted that the
spectroscopic pH titrations performed in section 9.3 indicated the desorption of the (metastable)
electrografted Im-linker molecules in 0.1 M NaOH takes place with an apparent hydrolysis rate of
k = 3.77 ´ 10-5 ± 0.34 s-1 and thus a half-life of 5.5 hours, which is significantly longer than the
timeframe of the experiments performed here.
Figure 70
Resonance Raman spectra normalised to the u4 vibrational band of CoPOH immobilised on Im-functionalised me-
ATO using the ‘post-coordination’ method in air and in 0.1 M NaOH, before and after applying a CV between 0.97
and 2.17 V (vs RHE).
153
Figure 71
CVs in 0.1 M NaOH of CoPOH immobilised on Im-functionalised me-ATO using the ‘post-coordination’ method.
Several CVs between 0.97 and 2.17 V (vs RHE) were applied to the sample (see Figure 71). In
the first cycle, an irreversible anodic peak centred around 1.7 V (vs RHE) can be detected that is
absent in the subsequent cycles. In these subsequent cycles, anodic current is observed for OER
with an onset potential of ca. 1.7 V (vs RHE), with the current density increasing with the cycle
number. This represents an overpotential of 0.47 V. The onset potential for OER is compared to
those obtained for other cobalt tetrapyrrole compounds reported in literature in Table 13.
Table 13
A comparison of reported onset potentials for the oxygen evolution reaction (OER) catalysed by cobalt
tetrapyrrole compounds. Potentials are given versus RHE and were calculated using the Nernst equation.
a best onset potential. b shown to be highly dependent on pKa of the buffer used. c calculated using UV-vis with
the adsorption co-efficient (
ε
) of FePOMe in dichloromethane (1.69 × 104 M-1cm-1).
Catalyst
System
Onset potential for
OER (V vs RHE)
Ref.
Co Hangman β-octafluoro
corrole (with carboxylic acid
hanging group)
Heterogenous
3 × 10-8 mol/cm2 in nafion, coated on
planar FTO
1.7 (pH 14)
1.75 (pH 10)
1.85 (pH 7)
309
Soluble Co porphyrins
(various)
Homogenous
1 mM aqueous solution
1.8 (pH 7)a,b
253
Perfluorinated Co
phthalocyanine
Heterogeneous
Up to 1.5 × 10-9 mol/cm2, drop-coated
on planar FTO
1.7 (pH 7)
256
Co 5,10,15-
tris(pentafluorophenyl)-
corrole (with various axial
ligands)
Heterogeneous
2.0 × 10-9 mol/cm2, drop-
coated on
planar FTO
1.75a (pH 7)
257
Co Hangman porphyrin (with
carboxylic acid hanging
group)
Heterogeneous
1.2 × 10-9 mol/cm2 c, immobilised on
via axial coordination to Im-
functionalised me-ATO
1.7 (pH 12.6)
This
work
154
After the first CV, there is a slight shift in the v4 band in Figure 70 from 1370 to 1371 cm-1 and the
v2 band from 1569 to 1570 cm-1, suggesting no overall change in the oxidation state or
coordination number of the immobilised CoPOH species (or the possible oxidation of some Co2+
species to Co3+). The origin of the anodic peak observed in the first CV will be discussed later.
Resonance Raman spectra of the sample were recorded under constant potential electrolysis
(CPE) at different potentials in 0.1 M NaOH, as shown in Figure 72. The corresponding
chronoamperometric data is shown in Figure 73. At a potential of 1.76 V (vs RHE), the current
density for OER is stable and increases slightly as a function of time, while at 2.16 V (vs RHE)
the current density is at its highest but decreases as a function of time. Raman spectra are only
shown for two potentials. On moving from a potential of 1.16 V (vs RHE), i.e. where no OER
occurs, to a potential of 2.16 V (vs RHE), i.e. where OER takes place, no change can be
observed in the Raman spectra. This either suggests that the catalytically active species has a
very short lifetime and can’t be detected spectroscopically, or that the incident light is not
resonant with an electronic transition in the catalytically active species (e.g. an oxide). It has been
proposed in the literature that for Co tetrapyrrole species, Co is active for the OER in oxidation
state CoIV or CoV.253,255,257
Figure 72
Resonance Raman spectra normalised to the u4 vibrational band of CoPOH immobilised on Im-functionalised me-
ATO using the ‘post-coordination method in 0.1 M NaOH under constant potential electrolysis at +1.16 and + 2.16
V (vs RHE).
155
Figure 73
Constant potential electrolysis (CPE) at different potentials in 0.1 M NaOH for CoPOH immobilised on Im-
functionalised me-ATO using the ‘post-coordination’ method.
For a species with a coordination number of 6, the loss of one or two of the axial ligands is a
prerequisite for forming a species with a free coordination site capable of accepting a water
molecule. Upon completing the OER catalytic cycle, the CoPOH species must either be 5-
coordinated or once again form a 6-coordinated species with a water or hydroxide ligand
(assuming the loss of the acetonitrile ligand). The resonance Raman spectra indicate that the
resting species are 6-coordinated low spin. It is thus proposed that, for catalytically active
species, these resting species are coordinated by an Im-linker molecule and a water or hydroxide
ligand (as illustrated in Scheme 58). It is assumed that substitution of an acetonitrile ligand for a
weaker ligand such as a water or a hydroxide ligand will not lower the energy difference between
the eg- and t2g- orbitals so much such that a high spin configuration becomes energetically
favourable.
Scheme 58
Proposed structure of the resting species of catalytically active CoPOH molecules immobilised on me-ATO and
coordinated by an Im-linker molecule and a water or hydroxide ligand.
A catalytic OER cycle mediated by axially-coordinated Co-corroles, as proposed by Cao and co-
workers, is shown in Scheme 59. The nature of the axial ligands are shown to have a substantial
156
effect on the corroles activity for the OER, with electron-donating axial ligands leading to higher
activity, which is likely the result of a strong trans effect weakening the CoV-oxo bond and
reducing the energy barrier for O-O bond formation via nucleophilic attack by water.257
Scheme 59
Catalytic oxygen evolution reaction (OER) cycle mediated by axially-coordinated Co corroles, as proposed by
Cao and co-workers. Adapted from 257 with permission from the PCCP Owner Societies.
To determine whether or not the CoPOH species were immobilised in an electrochemically
accessible manner, resonance Raman measurements were also conducted under the application
of potentials lower than E0 (Co+/2+) (-0.325 V (vs RHE)314). Measurements were conducted in a
sodium borate buffer at pH 10, the same conditions used by Kielb to measure CoPOH
physisorbed on nanostructured Ag electrodes (immobilised by dip-coating) under reducing
conditions using surface enhanced resonance Raman (SERR) spectroscopy.252 As can be seen
in Figure 74, at -0.7 V (vs RHE), there is a downward shift in the positions of the v2 and v4 bands
by 2 and 4 cm-1, respectively. Such a downward shift in the position of the oxidation-state
sensitive v4 band would be expected upon reduction of the initial oxidation state.315,316
Furthermore, there is a significant increase in the v2 to v4 band ratio. The SERR spectra of Kielb
also show such an increase in the v2 to v4 band ratio, as well as the appearance of a shoulder at
1361 cm-1 next to the dominant peak at 1369 cm-1 in the v4 band.252 The downward shift in the
broad v4 band observed in Figure 74 clearly indicates a greater contribution from a lower
wavenumber component, and suggests that a greater proportion of the immobilised CoPOH
species are reduced than in the work of Kielb. This observation is to be expected, given that it is
highly likely that thick multilayers of CoPOH species are deposited on the Ag electrode surface by
157
dip-coating, whereas (close to) monolayer coverage is expected in the case of the Im-
functionalised me-ATO. While the oxidation state of the Co Hangman species cannot be as easily
resolved as for the Fe Hangman species, with a 11 cm-1 difference in the v4 band positions of the
Fe2+ and Fe3+ high spin states, it is clear that the immobilised CoPOH species are
electrochemically accessible.
Figure 74
Resonance Raman spectra normalised to the u4 vibrational band of CoPOH immobilised on Im-functionalised me-
ATO using the ‘post-coordination’ method in 0.1 M sodium borate buffer (pH 10) under the application of a
constant potential of +0.8 and -0.7 V (vs RHE).
As shown in Figure 75, CVs were recorded of the CoPOH modified electrodes after CPE and
reveal the appearance of a reversible redox peak centred at ca. E1/2 = 1.31 V (vs RHE) (∆Ep =
161 mV). In a recent work by Sun and co-workers on the OER electrocatalytic activity of three
different Co porphyrins spin-coated onto FTO substrates, an increase in OER current with an
onset potential of ca. 1.75 V (vs RHE) is observed upon repeated cycling in a borate buffer (pH
9.2), along with the concurrent growth of a reversible redox peak at ca. E1/2 = 1.35 V (vs RHE),
irrespective of the porphyrin used.317 The OER activity and the redox peaks remain after
removing the immobilised porphyrin species by solvent rinsing. In fact, there is even a slight
enhancement in the OER activity. The redox peaks were therefore assigned to the Co2+/3+ couple
of cobalt oxide CoOx that forms due to the decomposition of the cobalt porphyrins. Evidence for
the formation of this highly active ultratrace oxide is made possible due to the use of low energy
synchrotron-based XPS, which has a high surface sensitivity. While the CoPOH species
immobilised on the me-ATO are not soluble in organic solvents due to axial coordination by
surface-bound Im-linker molecules, and thus cannot be removed in the same way to check for
activity due to oxide formation, any decomposition of bound CoPOH can be monitored using in
situ spectroelectrochemistry.
158
Figure 75
CVs in 0.1 M NaOH of the CoPOH immobilised on Im-functionalised me-ATO using the ‘post-coordination’
method before and after the application of constant potential electrolysis (CPE) steps in 0.1 M NaOH (see Figure
73) showing the appearance of a reversible redox peak at ca. E1/2 = 1.31 V (vs RHE) (∆Ep = 161 mV).
Quantitative measurements of the CoPOH species could not be made using resonance Raman
spectroelectrochemistry due to fluctuating Raman intensities as a result of the movement of the
rotating optical stage. As such, UV-vis spectroelectrochemistry was used instead. UV-Vis spectra
were recorded in 0.1 M NaOH before and after the application of an increasing number of CVs
between 0.95 and 2.15 V (vs RHE), as shown in Figure 76a and b. The UV-vis spectra show a
Soret band with an absorbance maximum at 441 nm, which is similar to a maximum of 435 nm
obtained for an aqueous solution of Co3+TMpyP(4), where TMpyP = tetrakis(N-methyl-4-
pyridiniumyl).313 The irreversible oxidation peak observed in the first CV can be decoupled from
any substantial decomposition processes taking place at the immobilised CoPOH by virtue of the
relatively insignificant change induced in the UV-Vis spectrum. The corresponding total charge
passed is equivalent to 0.8 × 10-9 molcm-2, which compares to an estimated total coverage
Γ
UV
of CoPOH on the me-ATO of 1.2 × 10-9 molcm-2.
159
Figure 76
(a) 11 CVs in 0.1 M NaOH of the CoPOH immobilised on Im-functionalised me-ATO using the ‘post-coordination’
method before and after application of a 30 min CPE step at 1.98 V (vs RHE) (shown in (b)), as well as a blank
CV of unmodified me-ATO. UV-Vis absorption spectra recorded before and after each CV are shown in (c), with
difference spectra calculated using the first recorded spectrum shown in the inset. The maximum OER current
density obtained (i.e. at 2.15 V vs RHE) during each CV is plotted in (d), alongside the corresponding change in
the UV-Vis absorption intensity at 442 nm.
Both the maximum current densities for OER and the corresponding change in the UV-Vis
absorption intensities at 442 nm are plotted as a function of the applied cycle number in Figure
76d. The steady increase in current density after the first CV can be correlated to a concurrent
decrease in Soret band intensity, which strongly suggests that the electrochemically induced
decomposition of the CoPOH leads to the formation of a new inorganic catalytic species, such as
CoOx species, as was observed by Sun and co-workers. Such species would exhibit
comparatively weak UV-Vis d-d transition absorption bands and, furthermore, trace amounts of
oxide would not be visible in Raman spectra without any resonance enhancement. The
appearance of a single dominant Q-band at 560 nm in the UV-vis spectra in Figure 76c (and a
weaker second band that is visible in the difference spectra that were calculated using the first
spectrum recorded, shown in the inset) indicates the presence of a metalloporphyrin, rather than
a free-base porphyrin that would display four Q-bands. There are no indications of the CoPOH
transforming into the free-base porphyrin upon repeated cycling, and thus, it is likely that the
porphyrin ring is oxidised upon removal of the cobalt.
160
Figure 77
Tafel plots of the applied potential against the logarithm of the current for the oxygen evolution reaction (OER)
using CoPOH immobilised on Im-functionalised me-ATO after (a) 1 CV, and after (b) 10 CVs.
It is unclear whether the initial OER catalytic current can be attributed to the immobilised CoPOH
or to any trace amounts of CoOx or other species formed during the first cycle. Tafel slope
analysis was carried out after the first CV (and the disappearance of the first irreversible anodic
peak), as well as after 10 CVs to determine whether or not there is a change in the OER kinetics
that would be indicative of different contributions from different catalytic species. A Tafel slope of
439 mV per decade is obtained after the first CV, while after 10 CVs the Tafel slope decreases to
158 mV per decade at low overpotentials and 398 mV per decade at high overpotentials. This
change clearly indicates a change in the rate determining step, and hence the OER mechanism
after repeated cycling, and therefore it is proposed that most of the initial catalytic OER current is
due to the immobilised CoPOH. Cobalt oxides typically exhibit Tafel slopes of between 40-60 mV
per decade318–320, while single-site or molecular OER catalysts exhibit higher Tafel slopes. A
Tafel slope of 118 mV per decade was obtained in pH 7 for single site Co OER catalysts
immobilised directly on mesoporous ITO from a Co[N(SiMe3)2]2 precursor321, 183 mV per decade
in pH 1 for a 3,5-bis(bipyridyl)pyrazolate-based diruthenium OER catalyst immobilised on
mesoporous ITO via a carboxylate anchor268, ca. 120 mV in pH 7 for a perfluorinated Co
phthalocyanine drop-coated on FTO256, and 120 mV in pH 7 for a Co Hangman β-octafluoro
corrole (with carboxylic acid hanging group) in Nafion coated on FTO309. For [Ir(pyalc)(H2O)2(μ-
O)]22+ (pylac = 2-(2’pyridyl)-2-propanolate) immobilised on mesoporous ITO via Ir-O-MOx bonds,
the Tafel slope in pH 2.6 was shown to increase with the ITO film thickness, increasing from 66
mV per decade with a film thickness of 3 μm, to 118 mV per decade for a film thickness of 11
μm.235 This increase is attributed to protons generated by the OER having to diffuse through the
thicker film (at pH 2.6, the pH is less sensitive to proton production from the OER). While this may
not pose as much of a problem at a pH of 12.6 and with a lower film thickness, the current
densities obtained here are up to 5 times higher than those obtained with the Ir catalyst.
Excluding such mass transport effects, the Tafel slope should be independent of the current
magnitude. The large values obtained here for the Tafel slopes may also be the result of a higher
161
resistivity in the me-ATO electrode, as previously discussed in section 9.4. Again, the
mesoporous ITO films in the aforementioned studies were composed of ITO nanocrystals of
around 50 nm in size, and these films are expected to exhibit lower resistivities compared to the
me-ATO films, which are composed of ATO nanocrystals 3-4 nm in size.
Figure 78
A Pourbaix diagram mapping out the possible equilibrium phases of cobalt species in an aqueous electrochemical
system. Reproduced with permission from 322, copyright 2008 Elsevier.
A Pourbaix diagram mapping out the possible equilibrium phases of cobalt species in an aqueous
electrochemical system is shown in Figure 78. In basic pH’s and at high potentials close to or
greater than the thermodynamic potential for OER, cobalt is stable as an oxide or hydroxide,
whereas at lower pH’s or potentials it is stable as a free ion. It is proposed that the abstraction of
Co ions from the porphyrin ring and the formation of Co oxides or hydroxides under anodic
potentials is favoured in basic pH’s, such as those used in this work and in the work of Sun and
co-workers317. The decomposition of cobalt porphyrins leads to the formation of a highly active
OER catalyst. Use of lower pH’s is likely required in order to maintain the molecular character of
cobalt porphyrin and other tetrapyrrole species under OER conditions. Fluorinated Co corroles
and phthalocyanines both show higher OER activity than their non-fluorinated counterparts in the
works of Rodionov and co-workers, and Nocera and co-workers. The lower overpotentials that
can be obtained with these species can further stabilise their molecular character.256,309
6.6 Conclusions
me-ATO and pl-ITO were used as model TCO materials and electrochemically grafted with
diazonium salts in order to immobilise a range of molecular electrocatalysts on their surfaces. In
situ potential-modulated IR measurements in the ATR mode were carried out and demonstrated
that grafted diazonium interfaces on me-ATO are highly stable in a broad electrochemical stability
window, encompassing the stability window of the ITO and FTO. Furthermore, these interfaces
exhibit hydrolytic stability down to a pH of ca. 2.5 and good stability in highly basic pH’s. This is
162
the first time such in situ IR methods have been applied with nanostructured TCO electrodes and
that the stability of interfaces on TCO surfaces has been determined in such a way. In situ IR
spectroscopy yields unprecedented information on interface formation, stability and structure that
is simply not possible using conventional techniques, such as UV-vis spectroscopy or purely
electrochemical methods.
A number of molecular catalysts were immobilised on pl-ITO and me-ATO, including a copper
bipyridine catalyst that is active for the oxygen evolution reaction (OER), and iron and cobalt
Hangman porphyrin catalysts, which are active for the oxygen reduction reaction (ORR) and the
OER, respectively. As the electrografting of interfaces from diazonium salts can often result in
thick, insulating layers on a range of electrode materials, strategies were developed to
functionalise the oxide surfaces in a way that the immobilised molecular species are
electrochemically accessible and able to take part in electrocatalysis. This is the first time that
diazonium salts have been used to successfully immobilise molecular complexes on
nanostructured TCO materials.
The copper-bipyridine catalyst was able to catalyse the OER with an onset potential of 1.73 V (vs
RHE), which represents a lower overpotential than those reported for homogenous copper-
bipyridine systems, most likely due to the electron-withdrawing substituent used to bind the
complex to the ITO surface. While the catalytic currents for this system were unstable,
presumably due to the loss of the copper active site, the diazonium-derived interface itself was
stable under non-turnover conditions for 24 hours in 0.1 M KOH, which is highly significant given
that most known anchoring groups would rapidly desorb under such basic pH’s, including
phosphonic acid derivatives, which are currently the most widely adopted anchoring groups used
in electrocatalytic cells or photo-electrocatalytic cells. Using the iron Hangman species FePOMe,
an onset potential of up to 0.81 V (vs RHE) could be obtained for the ORR for a relatively small
loading of catalyst. The overpotential was shown to decrease with catalyst loading. The cobalt
Hangman species CoPOH was shown to be active for the OER with an onset potential of 1.7 V
(vs RHE). An increase in OER current density as a function of time/applied potential was shown
to be related to the decomposition of CoPOH, which presumably forms a small amount of highly
active cobalt oxide OER catalyst. It is proposed that this reaction is favoured in basic pH’s.
The electronic behaviour and high stabilities observed for the electrografted diazonium interfaces
are likely the result of the binding motif formed with the oxide surfaces, which differs from other
common anchoring groups, such as phosphonates and silanes or purely organic anchoring
groups. Experimental and theoretical studies on TiO2 have shown that organic carboxylate and
hydroxamate anchoring groups exhibit faster interfacial electron transfer because of a better
electronic coupling at the oxide interface compared to phosphonate anchoring groups.27,231,236
163
The poor coupling observed for phosphonate (and silane) anchoring groups may arise from the
tetrahedral geometry of the phosphorous (or silicon) atom and the loss of conjugation through the
oxide.323 However, at the expense of better electronic behaviour, phosphonate and silane-derived
anchoring groups exhibit greater stabilities, with phosphonate anchoring groups showing
stabilities that are orders of magnitude greater than carboxylate anchoring groups.37,38 That being
said, phosphonates show poor stability in pH’s above 5 or 6, which limits their applicability, while
silanes result in bonds which are stronger and have been shown to be stable up to pH 11.25,36,38,39
It is tentatively proposed that the strong M–O–C bond formed upon the electrochemical grafting
of diazonium salts on oxide surfaces not only results in good electronic coupling with immobilised
molecular redox species, but also provides high chemical and electrochemical stabilities. Such
stabilities not attainable with other organic anchoring groups (or indeed phosphonates).
To conclude, the work here illustrates the potential for diazonium salts to be used as novel
anchoring groups on metal oxides materials, e.g. for electrocatalytic applications, as well as
potentially for other applications, e.g. in dye sensitised solar cells and optoelectronic devices.
164
10 Appendices
Appendices
10.1 Appendix to chapter 4.1
NH
Au
O
Au
NO2
S
NO
2
Au
165
Appendix 1
SEIRA spectra in acetonitrile and the corresponding spectra calculated using DFT of (a) interface deposited by
electrochemical grafting of 1 mM 4-BABD, (b) interface deposited by electrochemical grafting of 1 mM 4-NBD, (c)
SAM deposited by spontaneous grafting of 1 mM 4-NTP, (d) amino functionalised interfaces formed by the
electrochemical grating of (b), (e) SAM deposited by spontaneous adsorption of 1 mM 4-ATP.
Au
NH
2
S
NH2
Au
166
Appendix 2
Absorption band intensity against polarisation potential during the desorption of electrografted 4-BABD diazonium
interfaces in acetonitrile (0.1 M TBAP) under (a) cathodic and (b) anodic polarisation and desorption of 4-ATP
SAM-derived interface under (c) cathodic and (d) anodic polarisation.
Appendix 3
Spectra calculated using DFT of nitrobenzene-Au (Au-NP) and the partially reduced nitrosobenzene species (4-
NOP).
NO2
Au
N
Au
O
167
Appendix 4
SEIRA spectrum in acetonitrile and the corresponding spectrum calculated using DFT of an interface deposited
by electrochemical grafting of 1 mM bpy-diazo.
10.2 Appendix to chapter 4.2
Appendix 5
CVs of 1 mM solutions of 4-NBD in acetonitrile (0.1 M TBAP) using pl-ITO as a working electrode. Scan rate = 50
mV/s.
N
NH
O
Au
N
168
Appendix 6
Electrochemical reduction of the nitrophenyl moieties of the interface electrografted on pl-ITO from a 1 mM
solution 4-NBD with no added DPPH to aminophenyl/hydroxyaminophenyl moieties in a 0.1 M NaClO4, 1:9
ethanol:water solution.
10.3 Appendix to chapter 4.3
Appendix 7
Plot of change in absorption intensity DA with decreasing pH for the C-H bending mode d(C-H) at 1545 cm-1.
169
Appendix 8
CVs conducted in 0.1 M PB pH 7 using me-ATO as a working electrode in order to determine the stability range
of the electrode. Progressive scans with increments of 0.1 V were applied in both cathodic and anodic directions
originating from a potential of -0.35 V (vs RHE).
Appendix 9
CVs of 1 mM solutions of Im-diazo in acetonitrile (0.1 M TBAP) using pl-ITO as a working electrode. Scan rate =
50 mV/s.
170
Appendix 10
CVs in acetonitrile (0.1 M TBAP) of FePOMe immobilised on me-ATO modified using ‘post-coordination’ (method
A) (left) and ‘pre-coordination’ (method B) (right) at different scan rates.
Appendix 11
CVs in ph 7 (0.1 M PB) of FePOMe immobilised on me-ATO modified using (a) ‘post-coordination’ (method A)
and (b) ‘pre-coordination’ (method B) at different scan rates.
171
11 Bibliography
1. Bleich, K. & Guimaraes, R. D. Renewable Infrastructure Investment Handbook: A Guide for
Institutional Investors. (2016).
2. Tatin, A., Bonin, J. & Robert, M. A Case for Electrofuels. ACS Energy Lett. 1, 1062–1064
(2016).
3. Misson, M., Zhang, H. & Jin, B. Nanobiocatalyst advancements and bioprocessing
applications. J. R. Soc. Interface 12, 20140891–20140891 (2014).
4. Oh, H. S., Nong, H. N. & Strasser, P. Preparation of mesoporous Sb-, F-, and in-doped
SnO2bulk powder with high surface area for use as catalyst supports in electrolytic cells. Adv.
Funct. Mater. 25, 1074–1081 (2015).
5. Willkomm, J. et al. Dye-sensitised semiconductors modified with molecular catalysts for light-
driven H 2 production. Chem. Soc. Rev 45, 9–23 (2016).
6. Brennaman, M. K. et al. Finding the Way to Solar Fuels with Dye-Sensitized
Photoelectrosynthesis Cells. J. Am. Chem. Soc. 138, 13085–13102 (2016).
7. Materna, K. L., Crabtree, R. H. & Brudvig, G. W. Anchoring groups for photocatalytic water
oxidation on metal oxide surfaces. Chemical Society Reviews 46, 6099–6110 (2017).
8. Solymar, L. & Walsh, D. Lectures on the electrical properties of materials. (Oxford University
Press, 1993).
9. Parker, G. Introductory Semiconductor Device Physics. (Institute of Physics Publishing, 2004).
10. Grahn, H. T. Introduction to Semiconductor Physics. (World Scientific, 1999).
11. Krol, R. in Photoelectrochemical Hydrogen Production SE - 2 (eds. van de Krol, R. & Grätzel,
M.) 102, 13–67 (Springer US, 2012).
12. Gellings, P. J. & Bouwmeester, H. J. M. The CRC Handbook of Solid State Electrochemistry.
(Taylor & Francis, 1997).
13. Kröger, F. A. & Vink, H. J. Relations between the Concentrations of Imperfections in Crystalline
Solids. Solid State Phys. - Adv. Res. Appl. 3, 307–435 (1956).
14. González, G. B. Investigating the Defect Structures in Transparent Conducting Oxides Using
X-ray and Neutron Scattering Techniques. Materials (Basel). 5, 818–850 (2012).
15. Giménez, S. & Bisquert, J. Photoelectrochemical solar fuel production: from basic principles to
advanced devices. (Springer, 2016).
16. Schmickler, W. & Santos, E. Interfacial electrochemistry. (Springer, 2010). doi:10.1007/978-3-
642-04937-8
17. Bard, A. J. & Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications, 2nd
Edition. (John Wiley & Sons, 2001). doi:10.1016/j.aca.2010.06.020
18. Memming, R. Semiconductor Electrochemistry. Semiconductor Electrochemistry (Wiley-VCH
Verlag GmbH & Co. KGaA, 2015). doi:10.1002/9783527688685
19. Zoski, C. G. Handbook of electrochemistry. Handbook of Electrochemistry (Elsevier, 2007).
doi:10.1016/B978-0-444-51958-0.X5000-9
20. Shinagawa, T., Garcia-Esparza, A. T. & Takanabe, K. Insight on Tafel slopes from a
microkinetic analysis of aqueous electrocatalysis for energy conversion. Sci. Rep. 5, 13801
(2015).
21. Edwards, G. A., Bergren, A. J. & Porter, M. D. in Handbook of Electrochemistry 295–327
(2007). doi:10.1016/B978-044451958-0.50021-5
22. Scheres, L., Giesbers, M. & Zuilhof, H. Self-assembly of organic monolayers onto hydrogen-
terminated silicon: 1-alkynes are better than 1-alkenes. Langmuir 26, 10924–10929 (2010).
23. Li, Y. & Zuilhof, H. Photochemical grafting and patterning of organic monolayers on indium tin
oxide substrates. Langmuir 28, 5350–9 (2012).
172
24. Müller, V., Haase, F., Rathousky, J. & Fattakhova-Rohlfing, D. Surface functionalization of
mesoporous antimony doped tin oxide by metalorganic reaction. Mater. Chem. Phys. 137,
207–212 (2012).
25. Brennan, B. J. et al. Comparison of silatrane, phosphonic acid, and carboxylic acid functional
groups for attachment of porphyrin sensitizers to TiO2 in photoelectrochemical cells. Phys.
Chem. Chem. Phys. 15, 16605 (2013).
26. Zhang, L. & Cole, J. M. Anchoring groups for dye-sensitized solar cells. ACS Applied Materials
and Interfaces 7, 3427–3455 (2015).
27. Li, W. et al. What makes hydroxamate a promising anchoring group in dye-sensitized solar
cells? Insights from theoretical investigation. J. Phys. Chem. Lett. 5, 3992–3999 (2014).
28. McNamara, W. R. et al. Acetylacetonate anchors for robust functionalization of TiO2
nanoparticles with Mn(II)-terpyridine complexes. J. Am. Chem. Soc. 130, 14329–14338 (2008).
29. Takeshita, N., Okuno, M. & Ishibashi, T. Molecular conformation of DPPC phospholipid
Langmuir and Langmuir–Blodgett monolayers studied by heterodyne-detected vibrational sum
frequency generation spectroscopy. Phys. Chem. Chem. Phys. 19, 2060–2066 (2017).
30. Love, J. C., Estroff, L. A., Kriebel, J. K., Nuzzo, R. G. & Whitesides, G. M. Self-assembled
monolayers of thiolates on metals as a form of nanotechnology. Chem. Rev. 105, 1103–1169
(2005).
31. Pujari, S. P., Scheres, L., Marcelis, A. T. M. & Zuilhof, H. Covalent surface modification of
oxide surfaces. Angew. Chemie - Int. Ed. 53, 6322–6356 (2014).
32. Zaba, T. et al. Formation of Highly Ordered Self-Assembled Monolayers of Alkynes on Au(111)
Substrate. J. Am. Chem. Soc. 136, 11918–11921 (2014).
33. Guerrero, G., Alauzun, J. G., Granier, M., Laurencin, D. & Mutin, P. H. Phosphonate coupling
molecules for the control of surface/interface properties and the synthesis of nanomaterials.
Dalt. Trans. 42, 12569–12585 (2013).
34. Brumbach, M. & Armstrong, N. R. in Encyclopedia of Electrochemistry (Wiley-VCH Verlag
GmbH & Co. KGaA, 2007). doi:10.1002/9783527610426.bard100101
35. Li, J. et al. Characterization of transparent conducting oxide surfaces using self-assembled
electroactive monolayers. Langmuir 24, 5755–65 (2008).
36. Materna, K. L., Brennan, B. J. & Brudvig, G. W. Silatranes for binding inorganic complexes to
metal oxide surfaces. Dalt. Trans. 44, 20312–20315 (2015).
37. Péchy, P. et al. Preparation of phosphonated polypyridyl ligands to anchor transition-metal
complexes on oxide surfaces: application for the conversion of light to electricity with
nanocrystalline TiO 2 films. J. Chem. Soc., Chem. Commun. 0, 65–66 (1995).
38. Gillaizeau-Gauthier, I. et al. Phosphonate-based bipyridine dyes for stable photovoltaic
devices. Inorg. Chem. 40, 6073–6079 (2001).
39. Bae, E. et al. Effects of surface anchoring groups (Carboxylate vs Phosphonate) in ruthenium-
complex-sensitized TiO2 on visible light reactivity in aqueous suspensions. J. Phys. Chem. B
108, 14093–14101 (2004).
40. Bélanger, D. & Pinson, J. Electrografting: a powerful method for surface modification. Chem.
Soc. Rev. 40, 3995–4048 (2011).
41. Adenier, A., Chehimi, M. M., Gallardo, I., Pinson, J. & Vilà, N. Electrochemical oxidation of
aliphatic amines and their attachment to carbon and metal surfaces. Langmuir 20, 8243–53
(2004).
42. Palacin, S. et al. Molecule-to-metal bonds: Electrografting polymers on conducting surfaces.
ChemPhysChem 5, 1468–1481 (2004).
43. Allongue, P. et al. Covalent modification of carbon surfaces by aryl radicals generated from the
electrochemical reduction of diazonium salts. J. Am. Chem. Soc. 114, 5883–5884 (1992).
44. Downard, A. J. Electrochemically assisted covalent modification of carbon electrodes.
Electroanalysis 12, 1085–1096 (2000).
173
45. Mahouche-Chergui, S., Gam-Derouich, S., Mangeney, C. & Chehimi, M. M. Aryl diazonium
salts: a new class of coupling agents for bonding polymers, biomacromolecules and
nanoparticles to surfaces. Chem. Soc. Rev. 40, 4143–4166 (2011).
46. Tanaka, M., Sawaguchi, T., Sato, Y., Yoshioka, K. & Niwa, O. Surface modification of GC and
HOPG with diazonium, amine, azide, and olefin derivatives. Langmuir 27, 170–178 (2011).
47. Gross, A. J. et al. Nickel (II) tetraphenylporphyrin modified surfaces via electrografting of an
aryldiazonium salt. Electrochem. commun. 13, 1236–1239 (2011).
48. Li, Z., Yan, W. & Dai, S. Surface functionalization of ordered mesoporous carbons - A
comparative study. Langmuir 21, 11999–12006 (2005).
49. Griveau, S., Mercier, D., Vautrinul, C. & Chausse, a. Electrochemical grafting by reduction of
4-aminoethylbenzenediazonium salt: Application to the immobilization of (bio)molecules.
Electrochem. commun. 9, 2768–2773 (2007).
50. Ghilane, J. et al. Indirect reduction of aryldiazonium salts onto cathodically activated platinum
surfaces: Formation of metal-organic structures. Langmuir 21, 6422–6429 (2005).
51. Guselnikova, O. et al. Surface modification of Au and Ag plasmonic thin films via diazonium
chemistry: Evaluation of structure and properties. Colloids Surf., A 516, 274–285 (2017).
52. Mesnage, A., Lefèvre, X., Jégou, P., Deniau, G. & Palacin, S. Spontaneous grafting of
diazonium salts: chemical mechanism on metallic surfaces. Langmuir 28, 11767–78 (2012).
53. Chen, M., Kobashi, K., Chen, B., Lu, M. & Tour, J. M. Functionalized self-assembled
InAs/CaAs quantum-dot structures hybridized with organic molecules. Adv. Funct. Mater. 20,
469–475 (2010).
54. Agnès, C. et al. XPS study of ruthenium tris-bipyridine electrografted from diazonium salt
derivative on microcrystalline boron doped diamond. Phys. Chem. Chem. Phys. 11, 11647
(2009).
55. Maldonado, S. et al. Surface modification of indium tin oxide via electrochemical reduction of
aryldiazonium cations. Langmuir 22, 2884–91 (2006).
56. Adenier, A. et al. Study of the spontaneous formation of organic layers on carbon and metal
surfaces from diazonium salts. Surf. Sci. 600, 4801–4812 (2006).
57. Lamberti, F. et al. Surface functionalization of fluorine-doped tin oxide samples through
electrochemical grafting. ACS Appl. Mater. Interfaces 5, 12887–94 (2013).
58. Lund, T., Nguyen, P. T. & Ruhland, T. Electrochemical grafting of TiO2-based photo-anodes
and its effect in dye-sensitized solar cells. J. Electroanal. Chem. 758, 85–92 (2015).
59. Bureau, C. & Pinson, J. Method for modifying insulating or semi-conductive surfaces, and
resulting products. European Patent Application 1948720 (2006).
60. Combellas, C. et al. Surface modification of halogenated polymers. 4. Functionalisation of
poly(tetrafluoroethylene) surfaces by diazonium salts. Polymer (Guildf). 44, 19–24 (2003).
61. Chehimi, M. M. Aryl diazonium salts : new coupling agents in polymer and surface science.
(2012).
62. Laforgue, A., Addou, T. & Bélanger, D. Characterization of the Deposition of Organic
Molecules at the Surface of Gold by the Electrochemical Reduction of Aryldiazonium Cations.
Langmuir 21, 6855–65 (2005).
63. McKone, J. R., Lewis, N. S. & Gray, H. B. Will solar-driven water-splitting devices see the light
of day? Chemistry of Materials 26, 407–414 (2014).
64. McKone, J. R., Marinescu, S. C., Brunschwig, B. S., Winkler, J. R. & Gray, H. B. Earth-
abundant hydrogen evolution electrocatalysts. Chem. Sci. 5, 865–878 (2014).
65. Reisner, E., Windle, C. D. & Reisner, E. Heterogenised Molecular Catalysts for the Reduction
of CO 2 to Fuels. Chimia (Aarau). 69, 435–441 (2015).
66. Blakemore, J. D., Crabtree, R. H. & Brudvig, G. W. Molecular Catalysts for Water Oxidation.
Chemical Reviews 115, 12974–13005 (2015).
174
67. Haddad, A. Z., Garabato, B. D., Kozlowski, P. M., Buchanan, R. M. & Grapperhaus, C. A.
Beyond metal-hydrides: Non-transition-metal and metal-free ligand-centered electrocatalytic
hydrogen evolution and hydrogen oxidation. J. Am. Chem. Soc. 138, 7844–7847 (2016).
68. V. Dolganov, A. Metal-free Electrocatalyst for Hydrogen Production from Water. Int. J.
Electrochem. Sci. 1111, 9559–9565 (2016).
69. Reuillard, B. et al. Tuning Product Selectivity for Aqueous CO2Reduction with a Mn(bipyridine)-
pyrene Catalyst Immobilized on a Carbon Nanotube Electrode. J. Am. Chem. Soc. 139,
14425–14435 (2017).
70. Muresan, N. M., Willkomm, J., Mersch, D., Vaynzof, Y. & Reisner, E. Immobilization of a
molecular cobaloxime catalyst for hydrogen evolution on a mesoporous metal oxide electrode.
Angew. Chemie - Int. Ed. 51, 12749–12753 (2012).
71. DeKrafft, K. E. et al. Electrochemical water oxidation with carbon-grafted iridium complexes.
ACS Appl. Mater. Interfaces 4, 608–613 (2012).
72. Goff, A. Le et al. From hydrogenases to noble metal-free catalytic nanomaterials for
H2production and uptake. Science 326, 1384–1387 (2009).
73. Tong, L., Göthelid, M. & Sun, L. Oxygen evolution at functionalized carbon surfaces: a strategy
for immobilization of molecular water oxidation catalysts. Chem. Commun. 48, 10025 (2012).
74. Gentil, S. et al. Electrocatalytic O2Reduction at a Bio-inspired Mononuclear Copper Phenolato
Complex Immobilized on a Carbon Nanotube Electrode. Angew. Chemie - Int. Ed. 55, 2517–
2520 (2016).
75. Elgrishi, N., Griveau, S., Chambers, M. B., Bedioui, F. & Fontecave, M. Versatile
functionalization of carbon electrodes with a polypyridine ligand: metallation and
electrocatalytic H(+) and CO2 reduction. Chem. Commun. 51, 2995–8 (2015).
76. Polsky, R. et al. Diazonium-functionalized horseradish peroxidase immobilized via addressable
electrodeposition: Direct electron transfer and electrochemical detection. Langmuir 23, 364–
366 (2007).
77. Barbosa, O. et al. Glutaraldehyde in bio-catalysts design: a useful crosslinker and a versatile
tool in enzyme immobilization. RSC Adv. 4, 1583–1600 (2014).
78. Liu, F., Tao, G. & Zhuo, R. Synthesis of thermal phase separating reactive polymers and their
applications in immobilized enzymes. Polym. J. (Tokyo, Japan) 25, 561–567 (1993).
79. Putzbach, W. & Ronkainen, N. J. Immobilization techniques in the fabrication of nanomaterial-
based electrochemical biosensors: A review. Sensors 13, 4811–4840 (2013).
80. Guisan, J. M. Immobilization of Enzymes and Cells 3rd Edition. (Humana Press, 2013).
doi:10.1007/978-1-62703-550-7
81. Secundo, F. Conformational changes of enzymes upon immobilisation. Chem. Soc. Rev. 42,
6250 (2013).
82. Pace, C. N., Grimsley, G. R. & Scholtz, J. M. Protein ionizable groups: pK values and their
contribution to protein stability and solubility. Journal of Biological Chemistry 284, 13285–
13289 (2009).
83. Talasaz, A. H. et al. Prediction of protein orientation upon immobilization on biological and
nonbiological surfaces. Proc. Natl. Acad. Sci. U. S. A. 103, 14773–8 (2006).
84. Heidary, N. IR spectro-electrochemistry of an adsorbed oxygen-tolerant [NiFe] hydrogenase on
conductive surfaces. (2016).
85. Heidary, N. et al. Orientation-Controlled Electrocatalytic Efficiency of an Adsorbed Oxygen-
Tolerant Hydrogenase. PLoS One 10, e0143101 (2015).
86. Petek, M., Neal, T. E. & Murray, R. W. Spectroelectrochemistry. Application of Optically
Transparent Minigrid Electrodes under Semi-Infinite Diffusion Conditions. Anal. Chem. 43,
1069–1074 (1971).
87. Lindon, J. C., Tranter, G. E. & Koppenaal, D. W. Encyclopedia of spectroscopy and
spectrometry. (2016).
175
88. Valeur, B. & Service), W. I. (Online. Molecular fluorescence : principles and applications.
(Wiley-VCH, 2001).
89. Li, H. H. Refractive index of silicon and germanium and its wavelength and temperature
derivatives. J. Phys. Chem. Ref. Data 9, 561–658 (1980).
90. Li, H. H. Refractive Index of ZnS, ZnSe, and ZnTe and Its Wavelength and Temperature
Derivatives. J. Phys. Chem. Ref. Data 13, 103–150 (1984).
91. König, T. A. F. et al. Electrically tunable plasmonic behavior of nanocube-polymer
nanomaterials induced by a redox-active electrochromic polymer. ACS Nano 8, 6182–6192
(2014).
92. Mohamed, H. A. Properties of pure and light antimony-doped tin oxide thin films prepared by e-
beam technique. Optoelectron. Adv. Mater. Rapid Commun. 3, 936–941 (2009).
93. McQuillan, A. J. The discovery of surface-enhanced Raman scattering. Notes Rec. R. Soc. 63,
105–109 (2009).
94. Osawa, M. in Near-Field Optics and Surface Plasmon Polaritons (ed. Kawata, S.) 163–187
(Springer, 2001). doi:10.1007/3-540-44552-8_9
95. Chalmers, J. & Griffiths, P. R. Handbook of vibrational spectroscopy. (Wiley, 2002).
96. Vincent, K. A., Parkin, A. & Armstrong, F. A. Investigating and exploiting the electrocatalytic
properties of hydrogenases. Chemical Reviews 107, 4366–4413 (2007).
97. Mazurenko, I., Wang, X., de Poulpiquet, A. & Lojou, E. H2/O2 enzymatic fuel cells: from proof-
of-concept to powerful devices. Sustain. Energy Fuels 1, 1475–1501 (2017).
98. Wait, A. F., Parkin, A., Morley, G. M., Dos Santos, L. & Armstrong, F. A. Characteristics of
enzyme-based hydrogen fuel cells using an oxygen-tolerant hydrogenase as the anodic
catalyst. J. Phys. Chem. C 114, 12003–12009 (2010).
99. Kim, D.-H. & Kim, M.-S. Hydrogenases for biological hydrogen production. Bioresour. Technol.
102, 8423–8431 (2011).
100. Reisner, E. Solar hydrogen evolution with hydrogenases: From natural to hybrid systems.
European Journal of Inorganic Chemistry 2011, 1005–1016 (2011).
101. Sakai, T., Mersch, D. & Reisner, E. Photocatalytic hydrogen evolution with a hydrogenase in a
mediator-free system under high levels of oxygen. Angew. Chemie - Int. Ed. 52, 12313–12316
(2013).
102. Mersch, D. et al. Wiring of Photosystem II to Hydrogenase for Photoelectrochemical Water
Splitting. J. Am. Chem. Soc. 137, 8541–8549 (2015).
103. Lubitz, W., Ogata, H., Rüdiger, O. & Reijerse, E. Hydrogenases. Chemical Reviews 114,
4081–4148 (2014).
104. Murphy, B. J., Sargent, F. & Armstrong, F. A. Transforming an oxygen-tolerant [NiFe] uptake
hydrogenase into a proficient, reversible hydrogen producer. Energy Environ. Sci. 7, 1426–
1433 (2014).
105. Fritsch, J. et al. The crystal structure of an oxygen-tolerant hydrogenase uncovers a novel iron-
sulphur centre. Nature 479, 249–52 (2011).
106. Lenz, O. et al. H2 conversion in the presence of O2 as performed by the membrane-bound
[NiFe]-Hydrogenase of Ralstonia eutropha. ChemPhysChem 11, 1107–1119 (2010).
107. Evans, R. M. & Armstrong, F. A. in Methods in molecular biology (Clifton, N.J.) 1122, 73–94
(2014).
108. Cracknell, J. A., Wait, A. F., Lenz, O., Friedrich, B. & Armstrong, F. A. A kinetic and
thermodynamic understanding of O2 tolerance in [NiFe]-hydrogenases. Proc. Natl. Acad. Sci.
U. S. A. 106, 20681–20686 (2009).
109. Goris, T. et al. A unique iron-sulfur cluster is crucial for oxygen tolerance of a [NiFe]-
hydrogenase. Nat. Chem. Biol. 7, 310–318 (2011).
110. Frielingsdorf, S. et al. Reversible [4Fe-3S] cluster morphing in an O2-tolerant [NiFe]
176
hydrogenase. Nat. Chem. Biol. 10, 378–385 (2014).
111. Eckermann, A. L., Feld, D. J., Shaw, J. A. & Meade, T. J. Electrochemistry of redox-active self-
assembled monolayers. Coord. Chem. Rev. 254, 1769–1802 (2010).
112. Sezer, M., Millo, D., Weidinger, I. M., Zebger, I. & Hildebrandt, P. Analyzing the catalytic
processes of immobilized redox enzymes by vibrational spectroscopies. IUBMB Life 64, 455–
464 (2012).
113. Cracknell, J. a, Vincent, K. a & Armstrong, F. a. Enzymes as working or inspirational
electrocatalysts for fuel cells and electrolysis. Chem. Rev. 108, 2439–61 (2008).
114. Patil, S. A., Hasan, K., Leech, D., Hägerhäll, C. & Gorton, L. Improved microbial
electrocatalysis with osmium polymer modified electrodes. Chem. Commun. 48, 10183 (2012).
115. Plumeré, N. et al. A redox hydrogel protects hydrogenase from high-potential deactivation and
oxygen damage. Nat. Chem. 6, 822–827 (2014).
116. Zagal, J. H., Bedioui, F. & Dodelet, J.-P. N4-macrocyclic metal complexes. (Springer Science &
Business Media, 2006).
117. Zagal, J. H. Metallophthalocyanines as catalysts in electrochemical reactions. Coord. Chem.
Rev. 119, 89–136 (1992).
118. Shao, M. Electrocatalysis in fuel cells: A non- and low- Platinum approach. 9, (Springer-Verlag,
2013).
119. Blakemore, J. D., Crabtree, R. H. & Brudvig, G. W. Molecular Catalysts for Water Oxidation.
Chem. Rev. (2015). doi:10.1021/acs.chemrev.5b00122
120. Gersten, S. W., Samuels, G. J. & Meyer, T. J. Catalytic Oxidation of Water by an Oxo-Bridged
Ruthenium Dimer. J. Am. Chem. Soc. 104, 4029–4030 (1982).
121. Thomsen, J. M., Huang, D. L., Crabtree, R. H. & Brudvig, G. W. Iridium-based complexes for
water oxidation. Dalt. Trans. 44, 12452–12472 (2015).
122. Miyake, H., Ye, S. & Osawa, M. Electroless deposition of gold thin films on silicon for surface-
enhanced infrared spectroelectrochemistry. Electrochem. commun. 4, 973–977 (2002).
123. Angerstein-Kozlowska, H., Conway, B. E., Hamelin, A. & Stoicoviciu, L. Elementary steps of
electrochemical oxidation of single-crystal planes of Au Part II. A chemical and structural basis
of oxidation of the (111) plane. J. Electroanal. Chem. Interfacial Electrochem. 228, 429–453
(1987).
124. Chen, Z. et al. Fabrication of highly transparent and conductive indium-tin oxide thin films with
a high figure of merit via solution processing. Langmuir 29, 13836–42 (2013).
125. Kosynkin, D. V & Tour, J. M. Phenylene ethynylene diazonium salts as potential self-
assembling molecular devices. Org. Lett. 3, 993–995 (2001).
126. Broxton, T. J., Colton, R. & Traeger, J. C. Electrospray mass spectral studies of aromatic
diazonium salts and diazotates. J. Mass Spectrom. 30, 319–323 (1995).
127. Dogutan, D. K., Kwabena Bediako, D., Teets, T. S., Schwalbe, M. & Nocera, D. G. Efficient
synthesis of hangman porphyrins. Org. Lett. 12, 1036–1039 (2010).
128. Chang, C. J., Chng, L. L. & Nocera, D. G. Proton-coupled O-O activation on a redox platform
bearing a hydrogen-bonding scaffold. J. Am. Chem. Soc. 125, 1866–1876 (2003).
129. Yeh, C. Y., Chang, C. J. & Nocera, D. G. ‘Hangman’ porphyrins for the assembly of a model
heme water channel. J. Am. Chem. Soc. 123, 1513–1514 (2001).
130. Wisitruangsakul, N. et al. Monitoring catalysis of the membrane-bound hydrogenase from
Ralstonia eutropha H16 by surface-enhanced IR absorption spectroscopy. Angew. Chemie -
Int. Ed. 48, 611–613 (2009).
131. Becke, A. D. Density-functional exchange-energy approximation with correct asymptotic
behavior. Phys. Rev. A 38, 3098–3100 (1988).
132. Perdew, J. P. Density-functional approximation for the correlation energy of the
imhomogeneous electron gas. Phys. Rev. B 33, 8822–8824 (1986).
177
133. Frisch, M. J. et al. Gaussian 09, Revision D.01. (Gaussian, Inc., 2013).
134. Weigend, F. & Ahlrichs, R. Balanced basis sets of split valence, triple zeta valence and
quadruple zeta valence quality for H to Rn: design and assessment of accuracy. Phys. Chem.
Chem. Phys. 7, 3297–3305 (2005).
135. Rippers, Y., Utesch, T., Hildebrandt, P., Zebger, I. & Mroginski, M. A. Insights into the structure
of the active site of the O 2-tolerant membrane bound [NiFe] hydrogenase of R. eutropha H16
by molecular modelling. Phys. Chem. Chem. Phys. 13, 16146–16149 (2011).
136. Kozuch, J. et al. Voltage-dependent structural changes of the membrane-bound anion channel
hVDAC1 probed by SEIRA and electrochemical impedance spectroscopy. Phys. Chem. Chem.
Phys. 16, 9546–9555 (2014).
137. Rüdiger, O. et al. Enzymatic anodes for hydrogen fuel cells based on covalent attachment of
Ni-Fe hydrogenases and direct electron transfer to SAM-modified gold electrodes.
Electroanalysis 22, 776–783 (2010).
138. Liu, G., Böcking, T. & Gooding, J. J. Diazonium salts: Stable monolayers on gold electrodes for
sensing applications. J. Electroanal. Chem. 600, 335–344 (2007).
139. Shewchuk, D. M. & McDermott, M. T. Comparison of diazonium salt derived and thiol derived
nitrobenzene layers on gold. Langmuir 25, 4556–63 (2009).
140. D’Amours, M. & Belanger, D. Stability of Substituted Phenyl Groups Electrochemically Grafted
at Carbon Electrode Surface. J. Phys. Chem. B 107, 4811–4817 (2003).
141. Blankespoor, R., Limoges, B., Schöllhorn, B., Syssa-Magalé, J. L. & Yazidi, D. Dense
monolayers of metal-chelating ligands covalently attached to carbon electrodes
electrochemically and their useful application in affinity binding of histidine-tagged proteins.
Langmuir 21, 3362–3375 (2005).
142. Vaz-Dominguez, C. et al. Laccase electrode for direct electrocatalytic reduction of O2 to H2O
with high-operational stability and resistance to chloride inhibition. Biosens. Bioelectron. 24,
531–537 (2008).
143. Liu, G., Paddon-Row, M. N. & Justin Gooding, J. A molecular wire modified glassy carbon
electrode for achieving direct electron transfer to native glucose oxidase. Electrochem.
commun. 9, 2218–2223 (2007).
144. Rüdiger, O., Abad, J. M., Hatchikian, E. C., Fernandez, V. M. & De Lacey, A. L. Oriented
immobilization of Desulfovibrio gigas hydrogenase onto carbon electrodes by covalent bonds
for nonmediated oxidation of H2. J. Am. Chem. Soc. 127, 16008–16009 (2005).
145. Radi, A. E., Munoz-Berbel, X., Cortina-Puig, M. & Marty, J. L. Novel protocol for covalent
immobilization of horseradish peroxidase on gold electrode surface. Electroanalysis 21, 696–
700 (2009).
146. Pita, M. et al. High Redox Potential Cathode Based on Laccase Covalently Attached to Gold
Electrode. J. Phys. Chem. C 115, 13420–13428 (2011).
147. Vaz-Dominguez, C., Pita, M., De Lacey, A. L., Shleev, S. & Cuesta, A. Combined ATR-
SEIRAS and EC-STM study of the immobilization of laccase on chemically modified Au
electrodes. J. Phys. Chem. C 116, 16532–16540 (2012).
148. Yeşildağ, A. & Ekinci, D. Covalent attachment of pyridine-type molecules to glassy carbon
surfaces by electrochemical reduction of in situ generated diazonium salts. Formation of
ruthenium complexes on ligand-modified surfaces. Electrochim. Acta 55, 7000–7009 (2010).
149. Tsukahara, Y., Wada, T. & Tanaka, K. Redox Behavior of
Ruthenium(Bipyridine)(Terpyridine)(Carbonyl) Complex-modified Carbon Electrode and
Reactivity toward Electrochemical Reduction of CO2. Chem. Lett. 39, 1134–1135 (2010).
150. Cao, R. et al. Promotion of oxygen reduction by a bio-inspired tethered iron phthalocyanine
carbon nanotube-based catalyst. Nat. Commun. 4, 2076 (2013).
151. Wei, P. J., Yu, G. Q., Naruta, Y. & Liu, J. G. Covalent grafting of carbon nanotubes with a
biomimetic heme model compound to enhance oxygen reduction reactions. Angew. Chemie -
Int. Ed. 53, 6659–6663 (2014).
178
152. Maurin, A. & Robert, M. Catalytic CO 2 -to-CO conversion in water by covalently functionalized
carbon nanotubes with a molecular iron catalyst. Chem. Commun. 52, 12084–12087 (2016).
153. Harper, J. C., Polsky, R., Dirk, S. M., Wheeler, D. R. & Brozik, S. M. Electroaddressable
selective functionalization of electrode arrays: Catalytic NADH detection using aryl diazonium
modified gold electrodes. Electroanalysis 19, 1268–1274 (2007).
154. Gutierrez-Sanchez, C. et al. Efficiency of Enzymatic O 2 Reduction by Myrothecium verrucaria
Bilirubin Oxidase Probed by Surface Plasmon Resonance, PMIRRAS, and Electrochemistry.
ACS Catal. 6, 5482–5492 (2016).
155. Vincent, K. A. et al. Electrocatalytic hydrogen oxidation by an enzyme at high carbon monoxide
or oxygen levels. Proc. Natl. Acad. Sci. U. S. A. 102, 16951–16954 (2005).
156. Sensi, M., Del Barrio, M., Baffert, C., Fourmond, V. & Léger, C. New perspectives in
hydrogenase direct electrochemistry. Curr. Opin. Electrochem. 5, 135–145 (2017).
157. Menanteau, T., Levillain, E. & Breton, T. Electrografting via Diazonium Chemistry: From
Multilayer to Monolayer Using Radical Scavenger. Chem. Mater. 25, 2905–2909 (2013).
158. Hidalgo, R., Ash, P. A., Healy, A. J. & Vincent, K. A. Infrared spectroscopy during
electrocatalytic turnover reveals the Ni-L active site state during H2 oxidation by a NiFe
hydrogenase. Angew. Chemie - Int. Ed. 54, 7110–7113 (2015).
159. Reuillard, B. et al. Tuning Product Selectivity for Aqueous CO2 Reduction with a
Mn(bipyridine)-pyrene Catalyst Immobilized on a Carbon Nanotube Electrode. J. Am. Chem.
Soc. 139, 14425–14435 (2017).
160. Imabayashi, S.-I., Hobara, D., Kakiuchi, T., Knoll, W. & Plank-institut, M. Selective replacement
of adsorbed alkanethiols in phase-separated binary self-assembled monolayers by
electrochemical partial desorption. Langmuir 7463, 4502–4504 (1997).
161. Sumi, T. & Uosaki, K. Electrochemical Oxidative Formation and Reductive Desorption of a
Self-Assembled Monolayer of Decanethiol on a Au(111) Surface in KOH Ethanol Solution. J.
Phys. Chem. B 108, 6422–6428 (2004).
162. Chen, Y., Yang, C. & Wang, F. Bin. Electrochemical assessment of electrochemical oxidation
stability of self-assembled monolayers on gold and preparation of binary self-assembled
monolayers on gold. Electrochim. Acta 55, 3951–3956 (2010).
163. Hawecker, J., Lehn, J.-M. & Ziessel, R. Electrocatalytic reduction of carbon dioxide mediated
by Re (bipy)(CO) 3 Cl (bipy= 2, 2′-bipyridine). Chem. Commun. 328–330 (1984).
164. Ishida, H., Tanaka, K. & Tanaka, T. Electrochemical CO2 reduction catalyzed by ruthenium
complexes [Ru(bpy)2(CO)2]2+ and [Ru(bpy)2(CO)Cl]+. Effect of pH on the formation of CO
and HCOO-. Organometallics 6, 181–186 (1987).
165. Bourrez, M., Molton, F., Chardon-Noblat, S. & Deronzier, A. [Mn(bipyridyl)(CO)3Br]: An
abundant metal carbonyl complex as efficient electrocatalyst for CO2 reduction. Angew.
Chemie - Int. Ed. 50, 9903–9906 (2011).
166. Smieja, J. M. et al. Manganese as a substitute for rhenium in CO2 reduction catalysts: The
importance of acids. Inorg. Chem. 52, 2484–2491 (2013).
167. Walsh, J. J., Neri, G., Smith, C. L. & Cowan, A. J. Electrocatalytic CO2 reduction with a
membrane supported manganese catalyst in aqueous solution. Chem. Commun. (Camb). 50,
12698–701 (2014).
168. Sampson, M. D. & Kubiak, C. P. Electrocatalytic Dihydrogen Production by an Earth-Abundant
Manganese Bipyridine Catalyst. Inorg. Chem. 54, 6674–6676 (2015).
169. Sampson, M. D. et al. Manganese catalysts with bulky bipyridine ligands for the electrocatalytic
reduction of carbon dioxide: eliminating dimerization and altering catalysis. J. Am. Chem. Soc.
136, 5460–71 (2014).
170. Riplinger, C. & Carter, E. a. Influence of Weak Brønsted Acids on Electrocatalytic CO 2
Reduction by Manganese and Rhenium Bipyridine Catalysts. ACS Catal. 5, 900–908 (2015).
171. Sampson, M. D. & Kubiak, C. P. Manganese Electrocatalysts with Bulky Bipyridine Ligands:
Utilizing Lewis Acids to Promote Carbon Dioxide Reduction at Low Overpotentials. J. Am.
179
Chem. Soc. 138, 1386–1393 (2016).
172. Walsh, J. J. et al. Improving the efficiency of electrochemical CO 2 reduction using immobilized
manganese complexes. Faraday Discuss. 0, 1–14 (2015).
173. Rosser, T. E., Windle, C. D. & Reisner, E. Electrocatalytic and Solar-Driven CO2 Reduction to
CO with a Molecular Manganese Catalyst Immobilized on Mesoporous TiO 2. Angew. Chem.
Int. Ed. Engl. 55, 7388–7392 (2016).
174. Lukkari, J., Kleemola, K., Meretoja, M., Ollonqvist, T. & Kankare, J. Electrochemical post-self-
assembly transformation of 4-aminothiophenol monolayers on gold electrodes. Langmuir 14,
1705–1715 (1998).
175. Jiang, C., Elliott, J. M., Cardin, D. J. & Tsang, S. C. An electrochemical study of 4-
aminothiophenol/Pt nanoparticle multilayers on gold electrodes. Langmuir 25, 534–541 (2009).
176. Batz, V. et al. Electrochemistry and structure of the isomers of aminothiophenol adsorbed on
gold. J. Electroanal. Chem. 491, 55–68 (2000).
177. Jiang, D., Sumpter, B. G. & Dai, S. Structure and bonding between an aryl group and metal
surfaces. J. Am. Chem. Soc. 128, 6030–1 (2006).
178. Jiang, D. E., Sumpter, B. G. & Dai, S. How do aryl groups attach to a graphene sheet? J. Phys.
Chem. B 110, 23628–23632 (2006).
179. Le Floch, F., Simonato, J. P. & Bidan, G. Electrochemical signature of the grafting of diazonium
salts: A probing parameter for monitoring the electro-addressed functionalization of devices.
Electrochim. Acta 54, 3078–3085 (2009).
180. Menanteau, T., Levillain, E. & Breton, T. Spontaneous grafting of nitrophenyl groups on
carbon: Effect of radical scavenger on organic layer formation. Langmuir 30, 7913–7918
(2014).
181. Menanteau, T., Dias, M., Levillain, E., Downard, A. J. & Breton, T. Electrografting via
Diazonium Chemistry: The Key Role of the Aryl Substituent in the Layer Growth Mechanism. J.
Phys. Chem. C 120, 4423–4429 (2016).
182. Lehr, J., Williamson, B. E., Flavel, B. S. & Downard, A. J. Reaction of gold substrates with
diazonium salts in acidic solution at open-circuit potential. Langmuir 25, 13503–13509 (2009).
183. Jayasundara, D. R., Cullen, R. J., Soldi, L. & Colavita, P. E. In situ studies of the adsorption
kinetics of 4-nitrobenzenediazonium salt on gold. Langmuir 27, 13029–36 (2011).
184. Ricci, A. M., Méndez De Leo, L. P., Williams, F. J. & Calvo, E. J. Some evidence for the
formation of an azo bond during the electroreduction of diazonium salts on Au substrates.
ChemPhysChem 13, 2119–2127 (2012).
185. Yu, S. S. C., Tan, E. S. Q., Jane, R. T. & Downard, A. J. An electrochemical and XPS study of
reduction of nitrophenyl films covalently grafted to planar carbon surfaces. Langmuir 23,
11074–11082 (2007).
186. Gui, A. L. et al. A comparative study of electrochemical reduction of 4-nitrophenyl covalently
grafted on gold and carbon. Electroanalysis 22, 1824–1830 (2010).
187. Millo, D., Hildebrandt, P., Pandelia, M. E., Lubitz, W. & Zebger, I. SEIRA spectroscopy of the
electrochemical activation of an immobilized [NiFe] hydrogenase under turnover and non-
turnover conditions. Angew. Chemie - Int. Ed. 50, 2632–2634 (2011).
188. Saggu, M. et al. Spectroscopic insights into the oxygen-tolerant membrane-associated [NiFe]
hydrogenase of Ralstonia eutropha H16. J. Biol. Chem. 284, 16264–16276 (2009).
189. De Lacey, A. L., Fernández, V. M., Rousset, M. & Cammack, R. Activation and inactivation of
hydrogenase function and the catalytic cycle: Spectroelectrochemical studies. Chem. Rev. 107,
4304–4330 (2007).
190. Utesch, T. et al. Effect of the protonation degree of a self-assembled monolayer on the
immobilization dynamics of a [NiFe] hydrogenase. Langmuir 29, 673–682 (2013).
191. Breton, T. & Bélanger, D. Modification of carbon electrode with aryl groups having an aliphatic
amine by electrochemical reduction of in situ generated diazonium cations. Langmuir 24,
180
8711–8718 (2008).
192. Bryant, M. A. & Crooks, R. M. Determination of Surface pKa Values of Surface-Confined
Molecules Derivatized with pH-Sensitive Pendant Groups. Langmuir 9, 385–387 (1993).
193. Ludwig, M., Cracknell, J. A., Vincent, K. A., Armstrong, F. A. & Lenz, O. Oxygen-tolerant H2
oxidation by membrane-bound [NiFe] hydrogenases of ralstonia species coping with low level
H2 inair. J. Biol. Chem. 284, 465–477 (2009).
194. Murgida, D. H. & Hildebrandt, P. Electron-transfer processes of cytochrome c at interfaces.
New insights by surface-enhanced resonance Raman spectroscopy. Acc. Chem. Res. 37,
854–861 (2004).
195. Winkler, J. R. & Gray, H. B. Long-range electron tunneling. J. Am. Chem. Soc. 136, 2930–2939
(2014).
196. Nakamoto, K. in Infrared and Raman Spectra of Inorganic and Coordination Compounds 1–
273 (John Wiley & Sons, Inc., 2008). doi:10.1002/9780470405888.ch1
197. Barton, S. C., Gallaway, J. & Atanassov, P. Enzymatic Biofuel Cells for Implantable and
Microscal Devices. Chem. Rev. 104, 4867–4886 (2004).
198. Grieshaber, D., MacKenzie, R., Vörös, J. & Reimhult, E. Electrochemical Biosensors - Sensor
Principles and Architectures. Sensors 8, 1400–1458 (2008).
199. Zebda, A. et al. Single glucose biofuel cells implanted in rats power electronic devices. Sci.
Rep. 3, (2013).
200. Mazurenko, I., De Poulpiquet, A. & Lojou, E. Recent developments in high surface area
bioelectrodes for enzymatic fuel cells. Curr. Opin. Electrochem. (2017).
doi:10.1016/j.coelec.2017.07.001
201. Topoglidis, E., Astuti, Y., Duriaux, F., Grätzel, M. & Durrant, J. R. Direct electrochemistry and
nitric oxide interaction of heme proteins adsorbed on nanocrystalline tin oxide electrodes.
Langmuir 19, 6894–6900 (2003).
202. Frasca, S. et al. Mesoporous Indium Tin Oxide as a Novel Platform for Bioelectronics.
ChemCatChem 2, 839–845 (2010).
203. Kwan, P. et al. Spectroelectrochemistry of cytochrome c and azurin immobilized in nanoporous
antimony-doped tin oxide. Chem. Commun. 47, 12367 (2011).
204. Schaming, D. et al. Spectroelectrochemical characterization of small hemoproteins adsorbed
within nanostructured mesoporous ITO electrodes. Langmuir 28, 14065–14072 (2012).
205. Frasca, S. et al. Bioelectrocatalysis at mesoporous antimony doped tin oxide electrodes—
Electrochemical characterization and direct enzyme communication. Electrochim. Acta 110,
172–180 (2013).
206. Müller, V., Rathousky, J. & Fattakhova-Rohlfing, D. Covalent immobilization of redox protein
within the mesopores of transparent conducting electrodes. Electrochim. Acta 116, 1–8 (2014).
207. Mersch, D. et al. Wiring of Photosystem II to Hydrogenase for Photoelectrochemical Water-
splitting. J. Am. Chem. Soc. 150605124103001 (2015). doi:10.1021/jacs.5b03737
208. Reuillard, B. et al. High Performance Reduction of H2O2with an Electron Transport Decaheme
Cytochrome on a Porous ITO Electrode. J. Am. Chem. Soc. 139, 3324–3327 (2017).
209. González-Arribas, E. et al. Transparent, mediator- and membrane-free enzymatic fuel cell
based on nanostructured chemically modified indium tin oxide electrodes. Biosens.
Bioelectron. 97, 46–52 (2017).
210. Lund, T., Nguyen, P. T. & Ruhland, T. Electrochemical grafting of TiO2-based photo-anodes
and its effect in dye-sensitized solar cells. J. Electroanal. Chem. 758, 85–92 (2015).
211. Chen, X. et al. A Molecule with Dual Functionality 4-Aminophenylmethylphosphonic Acid: A
Comparison Between Layers Formed on Indium Tin Oxide by In Situ Generation of an Aryl
Diazonium Salt or by Self-Assembly of the Phosphonic Acid. Electroanalysis 23, 2633–2642
(2011).
212. Hebié, S., Devillers, C. H., Fournier, S. & Lucas, D. Direct Grafting of Free-Base meso-
181
Triarylporphyrins on Electrode Materials through Diazonium Reduction: Reversible Zinc(II)
Metallation of the Resulting Materials. ChemElectroChem 3, 45–50 (2016).
213. Rawson, F. J., Yeung, C. L., Jackson, S. K. & Mendes, P. M. Tailoring 3D single-walled carbon
nanotubes anchored to indium tin oxide for natural cellular uptake and intracellular sensing.
Nano Lett. 13, 1–8 (2013).
214. Lamberti, F. et al. Optoelectrochemical Biorecognition by Optically Transparent Highly
Conductive Graphene-Modified Fluorine-Doped Tin Oxide Substrates. ACS Appl. Mater.
Interfaces 6, 22769–77 (2014).
215. Charlton, M. R., Suhr, K. J., Holliday, B. J. & Stevenson, K. J. Electrochemical Modification of
Indium Tin Oxide Using Di(4-nitrophenyl) Iodonium Tetrafluoroborate. Langmuir 31, 695–702
(2015).
216. Finklea, H. O. et al. Characterisation of octadecanethiol-coated gold electrodes as microarray
electrodes by cyclic voltammetry and AC-impedance spectroscopy. Langmuir 9, 3660–3667
(1993).
217. Allongue, P. et al. Covalent modification of carbon surfaces by aryl radicals generated from the
electrochemical reduction of diazonium salts. J. Am. Chem. Soc. 119, 201–207 (1997).
218. Brooksby, P. A., Shields, J. D., Farquhar, A. K. & Downard, A. J. Reduction of Nitrophenyl
Films in Aqueous Solutions: How Many Electrons? ChemElectroChem 3, 2021–2026 (2016).
219. Brooksby, P. a & Downard, A. J. Multilayer nitroazobenzene films covalently attached to
carbon. An AFM and electrochemical study. J. Phys. Chem. B 109, 8791–8 (2005).
220. Ceccato, M. et al. Nitrophenyl groups in diazonium-generated multilayered films: which are
electrochemically responsive? Langmuir 26, 10812–21 (2010).
221. Menanteau, T., Levillain, E., Downard, A. J. & Breton, T. Evidence of monolayer formation via
diazonium grafting with a radical scavenger: electrochemical, AFM and XPS monitoring. Phys.
Chem. Chem. Phys. 17, 13137–13142 (2015).
222. Siebert, E. et al. Resonance Raman Spectroscopic Analysis of the [NiFe] Active Site and the
Proximal [4Fe-3S] Cluster of an O2-Tolerant Membrane-Bound Hydrogenase in the Crystalline
State. J. Phys. Chem. B 119, 13785–13796 (2015).
223. Liu, B. & Liu, J. DNA adsorption by indium tin oxide nanoparticles. Langmuir 31, 371–377
(2015).
224. Young, K. J. et al. Light-driven water oxidation for solar fuels. Coord. Chem. Rev. 256, 2503–
2520 (2012).
225. Bullock, R. M., Das, A. K. & Appel, A. M. Surface Immobilization of Molecular Electrocatalysts
for Energy Conversion. Chem. - A Eur. J. 23, 7626–7641 (2017).
226. Benck, J. D., Pinaud, B. A., Gorlin, Y. & Jaramillo, T. F. Substrate Selection for fundamental
studies of electrocatalysts and photoelectrodes: inert potential windows in acidic, neutral, and
basic electrolyte. PLoS One 9, e107942 (2014).
227. Rosser, T. E., Gross, M. A., Lai, Y.-H. & Reisner, E. Precious-metal free photoelectrochemical
water splitting with immobilised molecular Ni and Fe redox catalysts. Chem. Sci. 7, 4024–4035
(2016).
228. Hoertz, P. G., Chen, Z., Kent, C. A. & Meyer, T. J. Application of high surface area tin-doped
indium oxide nanoparticle films as transparent conducting electrodes. Inorg. Chem. 49, 8179–
8181 (2010).
229. Odrobina, J. et al. Backbone Immobilization of the Bis(bipyridyl)pyrazolate Diruthenium
Catalyst for Electrochemical Water Oxidation. ACS Catal. 7, 2116–2125 (2017).
230. Moore, G. F. et al. A visible light water-splitting cell with a photoanode formed by codeposition
of a high-potential porphyrin and an iridium water-oxidation catalyst. Energy Environ. Sci. 4,
2389 (2011).
231. Martini, L. A. et al. Modular assembly of high-potential zinc porphyrin photosensitizers attached
to TiO2 with a series of anchoring groups. J. Phys. Chem. C 117, 14526–14533 (2013).
182
232. Materna, K. L. et al. Heterogenized Iridium Water-Oxidation Catalyst from a Silatrane
Precursor. ACS Catal. 6, 5371–5377 (2016).
233. Li, F. et al. Organic Dye-Sensitized Tandem Photoelectrochemical Cell for Light Driven Total
Water Splitting. J. Am. Chem. Soc. 137, 9153–9159 (2015).
234. Chen, Z., Concepcion, J. J., Jurss, J. W. & Meyer, T. J. Single-site, catalytic water oxidation on
oxide surfaces. J. Am. Chem. Soc. 131, 15580–15581 (2009).
235. Sheehan, S. W. et al. A molecular catalyst for water oxidation that binds to metal oxide
surfaces. Nat. Commun. 6, 6469 (2015).
236. Negre, C. F. A. et al. Efficiency of interfacial electron transfer from Zn-porphyrin dyes into
TiO2correlated to the linker single molecule conductance. J. Phys. Chem. C 117, 24462–
24470 (2013).
237. Berisha, A., Chehimi, M., Pinson, J. & Podvorica, F. in Electroanalytical Chemistry 115–224
(CRC Press, 2015). doi:doi:10.1201/b19196-4
238. Bard, A. J. & Zoski, C. G. Electroanalytical Chemistry A Series of Advances: Volume 26. (CRC
Press, 2015).
239. Müller, V. et al. Highly Conducting Nanosized Monodispersed Antimony-Doped Tin Oxide
Particles Synthesized via Nonaqueous Sol−Gel Procedure. Chem. Mater. 21, 5229–5236
(2009).
240. Müller, V. et al. Transparent conducting films of antimony-doped tin oxide with uniform
mesostructure assembled from preformed nanocrystals. Small 6, 633–7 (2010).
241. Luo, L., Bozyigit, D., Wood, V. & Niederberger, M. High-Quality Transparent Electrodes Spin-
Cast from Preformed Antimony-Doped Tin Oxide Nanocrystals for Thin Film Optoelectronics.
Chem. Mater. 25, 4901–4907 (2013).
242. Oh, H.-S., Nong, H. N., Reier, T., Gliech, M. & Strasser, P. Oxide-supported Ir nanodendrites
with high activity and durability for the oxygen evolution reaction in acid PEM water
electrolyzers. Chem. Sci. 6, 3321–3328 (2015).
243. Ostermann, R., Zieba, R., Rudolph, M., Schlettwein, D. & Smarsly, B. M. Electrospun antimony
doped tin oxide (ATO) nanofibers as a versatile conducting matrix. Chem. Commun. 47, 12119
(2011).
244. Correa Baena, J. P. & Agrios, A. G. Antimony-doped tin oxide aerogels as porous electron
collectors for dye-sensitized solar cells. J. Phys. Chem. C 118, 17028–17035 (2014).
245. Auwärter, W., Écija, D., Klappenberger, F. & Barth, J. V. Porphyrins at interfaces. Nat. Chem.
7, 105–120 (2015).
246. Brumbach, M. T., Boal, A. K. & Wheeler, D. R. Metalloporphyrin assemblies on pyridine-
functionalized titanium dioxide. Langmuir 25, 10685–10690 (2009).
247. Willkomm, J., Muresan, N. M. & Reisner, E. Enhancing H 2 evolution performance of an
immobilised cobalt catalyst by rational ligand design. Chem. Sci. 6, 2727–2736 (2015).
248. Subbaiyan, N. K., Wijesinghe, C. A. & D’Souza, F. Supramolecular solar cells: Surface
modification of nanocrytalline TiO 2 with coordinating ligands to immobilize sensitizers and
dyads via metal-ligand coordination for enhanced photocurrent generation. J. Am. Chem. Soc.
131, 14645–14647 (2009).
249. Panagiotopoulos, A., Ladomenou, K., Sun, D., Artero, V. & Coutsolelos, A. G. Photochemical
hydrogen production and cobaloximes: the influence of the cobalt axial N-ligand on the system
stability. Dalt. Trans. 45, 6732–6738 (2016).
250. Chlistunoff, J. & Sansiñena, J. M. Effects of axial coordination of the metal center on the
activity of iron tetraphenylporphyrin as a nonprecious catalyst for oxygen reduction. J. Phys.
Chem. C 118, 19139–19149 (2014).
251. Ly, H. K. et al. 2 nd coordination sphere controlled electron transfer of iron hangman
complexes on electrodes probed by surface enhanced vibrational spectroscopy. Chem. Sci. 6,
6999–7007 (2015).
183
252. Kielb, P. Redox processes of heme proteins and metalloporphyrins studied by vibrational
spectroelectrochemistry. (Technische Universität Berlin, 2017).
253. Wang, D. & Groves, J. T. Efficient water oxidation catalyzed by homogeneous cationic cobalt
porphyrins with critical roles for the buffer base. Proc. Natl. Acad. Sci. U. S. A. 110, 15579–84
(2013).
254. Nakazono, T., Parent, A. R. & Sakai, K. Cobalt porphyrins as homogeneous catalysts for water
oxidation. Chem. Commun. Chem. Commun 49, 6325–6327 (2013).
255. Nakazono, T., Parent, A. R. & Sakai, K. Improving Singlet Oxygen Resistance during
Photochemical Water Oxidation by Cobalt Porphyrin Catalysts. Chem. - A Eur. J. 21, 6723–
6726 (2015).
256. Morlanés, N., Joya, K. S., Takanabe, K. & Rodionov, V. Perfluorinated cobalt phthalocyanine
effectively catalyzes water electrooxidation. Eur. J. Inorg. Chem. 2015, 49–52 (2015).
257. Cao, S. R. et al. Showcasing research from the laboratory of Professor Rui Cao at the As
featured in: The effect of the trans axial ligand of cobalt corroles on water oxidation activity in
neutral aqueous solutions. Phys. Chem. Chem. Phys. 19, 9755–9761 (2017).
258. Jia, H., Sun, Z., Jiang, D. & Du, P. Covalent Cobalt Porphyrin Framework on Multiwalled
Carbon Nanotubes for Efficient Water Oxidation at Low Overpotential. Chem. Mater. 27, 4586–
4593 (2015).
259. Joya, K. S., Morlanés, N., Maloney, E., Rodionov, V. & Takanabe, K. Immobilization of a
molecular cobalt electrocatalyst by hydrophobic interaction with a hematite photoanode for
highly stable oxygen evolution. Chem. Commun. 51, 13481–13484 (2015).
260. Barnett, S. M., Goldberg, K. I. & Mayer, J. M. A soluble copper–bipyridine water-oxidation
electrocatalyst. Nat. Chem. 4, 498–502 (2012).
261. Zhang, T., Wang, C., Liu, S., Wang, J. L. & Lin, W. A biomimetic copper water oxidation
catalyst with low overpotential. J. Am. Chem. Soc. 136, 273–281 (2014).
262. Zhang, T., Wang, C., Liu, S., Wang, J. L. & Lin, W. A biomimetic copper water oxidation
catalyst with low overpotential. J. Am. Chem. Soc. 136, 273–281 (2014).
263. Koepke, S. J., Light, K. M., VanNatta, P. E., Wiley, K. M. & Kieber-Emmons, M. T.
Electrocatalytic Water Oxidation by a Homogeneous Copper Catalyst Disfavors Single-Site
Mechanisms. J. Am. Chem. Soc. 139, 8586–8600 (2017).
264. Zhang, M. T., Chen, Z., Kang, P. & Meyer, T. J. Electrocatalytic water oxidation with a
copper(II) polypeptide complex. J. Am. Chem. Soc. 135, 2048–2051 (2013).
265. Shen, J., Wang, M., Zhang, P., Jiang, J. & Sun, L. Electrocatalytic water oxidation by copper(ii)
complexes containing a tetra- or pentadentate amine-pyridine ligand. Chem. Commun. 53,
4374–4377 (2017).
266. Fisher, K. J., Materna, K. L., Mercado, B. Q., Crabtree, R. H. & Brudvig, G. W. Electrocatalytic
Water Oxidation by a Copper(II) Complex of an Oxidation-Resistant Ligand. ACS Catal. 7,
3384–3387 (2017).
267. Luo, H. et al. High Surface Area Antimony-Doped Tin Oxide Electrodes Templated by Graft
Copolymerization. Applications in Electrochemical and Photoelectrochemical Catalysis. ACS
Appl. Mater. Interfaces 7, 25121–25128 (2015).
268. Odrobina, J. et al. Backbone Immobilization of the Bis(bipyridyl)pyrazolate Diruthenium
Catalyst for Electrochemical Water Oxidation. ACS Catal. 7, 2116–2125 (2017).
269. Chen, Z. & Meyer, T. J. Copper(II) catalysis of water oxidation. Angew. Chemie - Int. Ed. 52,
700–703 (2013).
270. Liu, C., Dubois, K. D., Louis, M. E., Vorushilov, A. S. & Li, G. Photocatalytic CO2 reduction and
surface immobilization of a tricarbonyl Re(I) compound modified with amide groups. ACS
Catal. 3, 655–662 (2013).
271. Schreier, M. et al. Covalent Immobilization of a Molecular Catalyst on Cu2O Photocathodes for
CO2 Reduction. J. Am. Chem. Soc. 138, 1938–1946 (2016).
184
272. Brushett, F. R. et al. A carbon-supported copper complex of 3,5-diamino-1,2,4-triazole as a
cathode catalyst for alkaline fuel cell applications. J. Am. Chem. Soc. 132, 12185–12187
(2010).
273. McCrory, C. C. L. et al. Electrocatalytic O 2 reduction by covalently immobilized mononuclear
copper(I) complexes: Evidence for a binuclear Cu 2O 2 intermediate. J. Am. Chem. Soc. 133,
3696–3699 (2011).
274. Kakuda, S., Peterson, R. L., Ohkubo, K., Karlin, K. D. & Fukuzumi, S. Enhanced catalytic four-
electron dioxygen (O2) and two-electron hydrogen peroxide (H2O2) reduction with a copper(II)
complex possessing a pendant ligand pivalamido group. J. Am. Chem. Soc. 135, 6513–6522
(2013).
275. Garrett, D. J. et al. Diazonium salt derivatives of osmium bipyridine complexes:
Electrochemical grafting and characterisation of modified surfaces. Electrochim. Acta 56,
2213–2220 (2011).
276. Müller, V. Mesoporous Transparent Conducting Films of Antimony Doped Tin Oxide as
Nanostructured Electrodes. (Ludwig-Maximilians-Universität München, 2013).
277. Tian, R. et al. The effect of annealing on a 3D SnO2/graphene foam as an advanced lithium-
ion battery anode. Sci. Rep. 6, 1–9 (2016).
278. Brinker, C. J. Hydrolysis and Condensation of Silicates: Effects on Structure. Journal of Non-
Crystalline Solids 100, 31–50 (1988).
279. Hanson, K. et al. Photostability of phosphonate-derivatized, Ru II polypyridyl complexes on
metal oxide surfaces. ACS Appl. Mater. Interfaces 4, 1462–1469 (2012).
280. Hyde, J. T. et al. Electrochemical instability of phosphonate-derivatized, ruthenium(III)
polypyridyl complexes on metal oxide surfaces. ACS Appl. Mater. Interfaces 7, 9554–9562
(2015).
281. Reuillard, B., Gentil, S., Carrière, M., Le Goff, A. & Cosnier, S. Biomimetic versus enzymatic
high-potential electrocatalytic reduction of hydrogen peroxide on a functionalized carbon
nanotube electrode. Chem. Sci. 6, 5139–5143 (2015).
282. Eckermann, A. L., Feld, D. J., Shaw, J. A. & Meade, T. J. Electrochemistry of redox-active self-
assembled monolayers. Coord. Chem. Rev. 254, 1769–1802 (2010).
283. Hoertz, P. G., Chen, Z., Kent, C. A. & Meyer, T. J. Application of high surface area tin-doped
indium oxide nanoparticle films as transparent conducting electrodes. Inorg. Chem. 49, 8179–
8181 (2010).
284. Galloni, P. et al. in Handbook of Porphyrin Science 33, 225–415 (World Scientific Publishing
Company, 2012).
285. Roth, K. M. et al. Measurements of electron-transfer rates of charge-storage molecular
monolayers on Si(100). Toward hybrid molecular/semiconductor information storage devices.
J. Am. Chem. Soc. 125, 505–517 (2003).
286. Laviron, E. The use of linear potential sweep voltammetry and of a.c. voltammetry for the study
of the surface electrochemical reaction of strongly adsorbed systems and of redox modified
electrodes. J. Electroanal. Chem. 100, 263–270 (1979).
287. Liau, Y.-H., Scherer, N. F. & Rhodes, K. Nanoscale Electrical Conductivity and Surface
Spectroscopic Studies of Indium−Tin Oxide. J. Phys. Chem. B 105, 3282–3288 (2001).
288. Laviron, E. General expression of the linear potential sweep voltammogram in the case of
diffusionless electrochemical systems. J. Electroanal. Chem. Interfacial Electrochem. 101, 19–
28 (1979).
289. Vitale, R. et al. Spectroelectrochemistry of FeIII- and CoIII-mimochrome VI artificial enzymes
immobilized on mesoporous ITO electrodes. Chem. Commun. 50, 1894–1896 (2014).
290. Renault, C. et al. Unraveling the mechanism of catalytic reduction of O 2 by microperoxidase-
11 adsorbed within a transparent 3D-nanoporous ITO film. J. Am. Chem. Soc. 134, 6834–6845
(2012).
291. Neumann, B. et al. Bioelectrocatalytic Reduction of Hydrogen Peroxide by Microperoxidase-11
185
Immobilized on Mesoporous Antimony-Doped Tin Oxide. ChemElectroChem 4, 913–919
(2017).
292. Hou, K. et al. Dye-anchored mesoporous antimony-doped tin oxide electrochemiluminescence
cell. Adv. Mater. 21, 2492–2496 (2009).
293. Renault, C., Harris, K. D., Brett, M. J., Balland, V. & Limoges, B. Time-resolved UV-visible
spectroelectrochemistry using transparent 3D-mesoporous nanocrystalline ITO electrodes.
Chem. Commun. 47, 1863–1865 (2011).
294. Ye, J. S. et al. Application of multi-walled carbon nanotubes functionalized with hemin for
oxygen detection in neutral solution. J. Electroanal. Chem. 562, 241–246 (2004).
295. Feng, Z., Sagara, T. & Niki, K. Application of potential-modulated UV-visible reflectance
spectroscopy to electron transfer rate measurements for adsorbed species on electrode
surfaces. Anal. Chem. 67, 3564–3570 (1995).
296. Picot, M., Nicolas, I., Poriel, C., Rault-Berthelot, J. & Barrière, F. On the nature of the electrode
surface modification by cathodic reduction of tetraarylporphyrin diazonium salts in aqueous
media. Electrochem. commun. 20, 167–170 (2012).
297. Elzing, A., van der Putten, A., Visscher, W. & Barendrecht, E. The cathodic reduction of
oxygen at cobalt phthalocyanine. J. Electroanal. Chem. Interfacial Electrochem. 200, 313–322
(1986).
298. Elzing, A., van der Putten, A., Visscher, W. & Barendrecht, E. The cathodic reduction of
oxygen at cobalt phthalocyanine. Influence of electrode preparation on electrocatalysis. J.
Electroanal. Chem. 200, 313–322 (1986).
299. Brown, A. P. & Anson, F. C. Cyclic and Differential Pulse Voltammetric Behavior of Reactants
Confined to the Electrode Surface. Anal. Chem. 49, 1589–1595 (1977).
300. Acevedo, D. & Abruña, H. D. Electron-transfer study and solvent effects on the formal potential
of a redox-active self-assembling monolayer. J. Phys. Chem. 95, 9590–9594 (1991).
301. Roth, K. M. et al. Comparison of electron-transfer and charge-retention characteristics of
porphyrin-containing self-assembled monolayers designed for molecular information storage. J.
Phys. Chem. B 106, 8639–8648 (2002).
302. Yasseri, A. A. et al. Structural and Electron-Transfer Characteristics of O-, S-, and Se-Tethered
Porphyrin Monolayers on Si(100). Journal of the American Chemical Society 127, 15603–
15612 (2005).
303. Jiao, J., Nordlund, E., Lindsey, J. S. & Bocian, D. F. Effects of counterion mobility, surface
morphology, and charge screening on the electron-transfer rates of porphyrin monolayers. J.
Phys. Chem. C 112, 6173–6180 (2008).
304. Eberspacher, T. A., Collman, J. P., Chidsey, C. E. D., Donohue, D. L. & Ryswyk, H. Van.
Modular Assembly and Air-Stable Electrochemistry of Ruthenium Porphyrin Monolayers.
Langmuir 19, 3814–3821 (2003).
305. Thorum, M. S., Hankett, J. M. & Gewirth, A. A. Poisoning the Oxygen Reduction Reaction on
Carbon-Supported Fe and Cu Electrocatalysts: Evidence for Metal-Centered Activity. J. Phys.
Chem. Lett. 2, 295–298 (2011).
306. Chen, R., Li, H., Chu, D. & Wang, G. Unraveling oxygen reduction reaction mechanisms on
carbon-supported fe-phthalocyanine and co-phthalocyanine catalysts in alkaline solutions. J.
Phys. Chem. C 113, 20689–20697 (2009).
307. Wang, N. et al. Oxygen electroreduction promoted by quasi oxygen vacancies in metal oxide
nanoparticles prepared by photoinduced chlorine doping. Chem. Commun. 51, 10620–10623
(2015).
308. Liu, B. et al. Improved Photoelectrocatalytic Performance for Water Oxidation by Earth-
Abundant Cobalt Molecular Porphyrin Complex-Integrated BiVO4 Photoanode. ACS Appl.
Mater. Interfaces 8, 18577–18583 (2016).
309. Dogutan, D. K., McGuire, R. & Nocera, D. G. Electocatalytic water oxidation by cobalt(III)
hangman β-octafluoro corroles. J. Am. Chem. Soc. 133, 9178–80 (2011).
186
310. Lee, C. H., Dogutan, D. K. & Nocera, D. G. Hydrogen generation by hangman
metalloporphyrins. J. Am. Chem. Soc. 133, 8775–8777 (2011).
311. Roubelakis, M. M., Bediako, D. K., Dogutan, D. K. & Nocera, D. G. Proton-coupled electron
transfer kinetics for the hydrogen evolution reaction of hangman porphyrins. Energy Environ.
Sci. 5, 7737 (2012).
312. Solis, B. H. et al. Theoretical analysis of cobalt hangman porphyrins: Ligand dearomatization
and mechanistic implications for hydrogen evolution. ACS Catal. 4, 4516–4526 (2014).
313. Terekhov, S. N., Galievsky, V. A., Chirvony, V. S. & Turpin, P. Y. Resonance Raman and
absorption characterization of cationic Co(II)-porphyrin in its complexes with nucleic acids:
Binding modes, nucleic base specificity and role of water in Co(II) oxidation processes. J.
Raman Spectrosc. 36, 962–973 (2005).
314. Kielb, P. et al. Spectroelectrochemistry of biomimetic cobalt hangman complexes during
electrocatalytic hydrogen formation and relevance of –COOH hanging group. in 16th European
Conference on the Spectroscopy of Biological Molecules (2015).
315. Siebert, F. & Hildebrandt, P. Vibrational spectroscopy in life science. (Wiley-VCH, 2008).
316. Parthasarathi, N., Hansen, C., Yamaguchi, S. & Spiro, T. G. Metalloporphyrin Core Size
Resonance Raman Marker Bands Revisited: Implications for the Interpretation of Hemoglobin
Photoproduct Raman Frequencies. J. Am. Chem. Soc. 109, 3865–3871 (1987).
317. Daniel, Q. et al. Re-Investigation of Cobalt Porphyrin for Electrochemical Water Oxidation on
FTO Surface: Formation of CoOx as Active Species. ACS Catal. 7, 1143–1149 (2017).
318. Surendranath, Y., Kanan, M. W. & Nocera, D. G. Mechanistic studies of the oxygen evolution
reaction by a cobalt-phosphate catalyst at neutral pH. J. Am. Chem. Soc. 132, 16501–16509
(2010).
319. Gerken, J. B. et al. Electrochemical Water Oxidation with Cobalt-Based Electrocatalysts from
pH 0–14: The Thermodynamic Basis for Catalyst Structure, Stability, and Activity. J. Am.
Chem. Soc. 133, 14431–14442 (2011).
320. Chen, W. et al. In situ electrochemical oxidation tuning of transition metal disulfides to oxides
for enhanced water oxidation. ACS Cent. Sci. 1, 244–251 (2015).
321. Swierk, J. R. & Tilley, T. D. Electrocatalytic Water Oxidation by Single Site and Small
Nuclearity Clusters of Cobalt. J. Electrochem. Soc. 165, H3028–H3033 (2018).
322. Garcia, E. M., Santos, J. S., Pereira, E. C. & Freitas, M. B. J. G. Electrodeposition of cobalt
from spent Li-ion battery cathodes by the electrochemistry quartz crystal microbalance
technique. J. Power Sources 185, 549–553 (2008).
323. Guerrero, G., Alauzun, J. G., Granier, M., Laurencin, D. & Mutin, P. H. Phosphonate coupling
molecules for the control of surface/interface properties and the synthesis of nanomaterials.
Dalt. Trans. 42, 12569–12585 (2013).
187
