Document [original]

Mesoporous oxides as efficient catalysts for the

electrocatalytic oxygen evolution reaction (OER)

vorgelegt von

Diplom-Chemiker Michael Bernicke

geb. in Berlin

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. Thomas Friedrich

Gutachter: Dr.-Ing. Ralph Krähnert

Gutachter: Prof. Dr. Peter Strasser

Gutachter: Prof. Dr. Michael Bron

Tag der wissenschaftlichen Aussprache: 7. September 2016

Berlin 2016

Wir sind gleichsam Zwerge, die auf den Schultern von Riesen sitzen, um mehr und

Entfernteres als diese sehen zu können - freilich nicht dank eigener scharfer Sehkraft oder

Körpergröße, sondern weil die Größe der Riesen uns zu Hilfe kommt und uns emporhebt.

- Bernhard von Chartres

Danksagung

Ich danke Dr.-Ing. Ralph Krähnert für die Aufnahme in den Arbeitskreis und die Möglichkeit

der Promotion. Durch anregende und strukturierte wissenschaftliche Diskussionen war

immer ein effizientes Arbeiten im Laufe meiner Promotion möglich. Für die unzähligen

Ratschläge und die fruchtbare Zusammenarbeit möchte ich mich bedanken.

Bei Herrn Professor Thomas Friedrich bedanke ich mich für die bereitwillige Übernahme der

Funktion des Vorsitzenden und bei den Herren Professoren Peter Strasser und Michael Bron

bedanke ich mich für die Erstellung des Gutachtens.

Mein weiterer Dank gilt Erik Ortel für die freundliche Einweisung in die Künste eines

Doktoranden. Bei Tobias Reier, Arno Bergmann und Nadine Menzel möchte ich mich für die

vielen Ratschläge zu Beginn und während meiner Promotion bedanken. Mein besonderer

Dank gilt hierbei Tobias Reier, der unzählige Stunden damit verbracht hat, mich in die

Geheimnisse der Elektrokatalyse einzuweihen. Des Weiteren bedanke ich mich bei Andreas

Lippitz für XPS- und Eicke Gericke für SAXS-Messungen. Besonders möchte ich Kornelia

Weh für zahlreiche ICP-Messungen danken. Bei Denis Bernsmeier bedanke ich mich für

bereitwilliges Messen von SAXS und BET. Björn Eckhardt und Katrin Schulz danke ich für

viele XRD-Messungen. Ich danke Jorge Araujo für DEMS-Messungen. Des Weiteren möchte

ich mich bei Erik Ortel, Benjamin Paul, Roman Schmack und allen Mitarbeitern des ZELMI

für unzählige elektronenmikroskopische Bilder bedanken. Nicolas Chaoui, Jérôme Roeser,

Nathaniel Leonard und vor allem Ebru Özer danke ich für sorgfältiges Korrekturlesen.

Bei meinen Kollegen Denis Bernsmeier, Björn Eckhardt, Roman Schmack, Katrin Schulz,

Benjamin Paul, Huan Wang und René Sachse möchte ich mich für ein angenehmes

Arbeitsklima und die gemeinsamen Abende mit viel Spaß bedanken.

Der Arbeitsgruppe von Herrn Professor Strasser danke ich für die angenehme

Arbeitsatmosphäre und die stundenlangen Freuden am Kickertisch. Besonders hervorheben

möchte ich hierbei Ebru Özer, Camillo Spöri, Hong Nhan, Tobias Reier, Vera Beermann,

Henrike Schmieß, Julian Steinberg, Malte Klingenhof, Thomas Merzdorf, Henner Heyen und

Elisabeth Hornberger.

Zu guter Letzt möchte ich mich bei meinen Eltern und meiner restlichen Familie für die

ständige Unterstützung in meinem Leben bedanken. Ihr wart immer an meiner Seite und

habt mich unterstützt, wo ihr nur konntet, dafür möchte ich euch von ganzem Herzen

danken!

Zusammenfassung

Wasserstoff ist ein wichtiger Energieträger der Zukunft und findet Verwendung in einer

Vielzahl industrieller Prozesse. Die elektrokatalytische Wasserspaltung stellt eine wichtige

Möglichkeit der Wasserstoffgewinnung dar und wird durch einen komplexen Mechanismus

an der sauerstofferzeugenden Anode limitiert. Die Erhöhung der Massenaktivität eingesetzter

Katalysatoren führt zur Verringerung der Investitionskosten von Elektrolyseuren.

Die vorliegende Doktorarbeit zeigt die Synthese mesoporöser Oxidfilme als saure oder

alkalische OER-Elektrokatalysatoren. Die Katalysatoren basieren auf unterschiedlichen

Metalloxiden, wie NiO, IrO2 und IrO2/TiO2. Die Herstellung oxidischer Filme erfolgte mittels

Tauchbeschichtung und basiert auf einer verdampfungsinduzierten Selbstanordnung eines

PEO-b-PB-b-PEO Blockcopolymers als Porentemplat und eines geeigneten

Metalloxidpräkursors sowie einer anschließenden thermischen Behandlung. Die jeweiligen

Synthesen und physikochemischen Charakterisierungen der einzelnen Verbindungen

werden im Diskussionsteil beschrieben.

Die Korrelation von Struktur und OER-Aktivität erlaubte das Ableiten von Struktur-Aktivitäts-

Beziehungen. Diese Beziehungen konnten genutzt werden, um OER-Einflussparameter zu

identifizieren. Die Verwendung verschiedener Systeme mit spezifischen Eigenschaften

erlaubte die gezielte Untersuchung der wichtigsten OER-Einflussparameter. Durch

Zusammenfassen der wichtigsten OER-Einflussparameter wurde ein mögliches OER-Model

vorgeschlagen. Das OER-Model beschreibt, bei welchen strukturellen Eigenschaften hohe

OER-Aktivitäten erwartet werden können. Zusätzlich wurde der Einfluss auf kinetische

Kenngrößen, wie etwa dem Tafelanstieg, berücksichtigt. Hohe OER-Aktivitäten und geringe

Tafelanstiege werden vor allem für Materialien mit geringer Kristallitgröße, hoher aktiver

Oberfläche, hoher Leitfähigkeit sowie bei elektrokatalytischen Messungen mit moderaten

Überspannungen (ca. η < 0.35 V) beobachtet.

Katalysatoren, welche nach dem postulierten OER-Model, hohe OER-Aktivitäten zeigen,

wurden mit kommerziellen Referenzsystemen verglichen. Es konnte gezeigt werden, dass

die in dieser Arbeit hergestellten Oxidfilme über eine 11- bis 24- fach höhere OER-Aktivität

bezogen auf die eingesetzte Masse an Iridium Metall im Vergleich zu kommerziell

verfügbaren Katalysatorpulvern verfügen.

Abstract

Hydrogen is considered an important energy carrier and feedstock for industrial applications.

The generation of hydrogen can be realised by electrocatalytic water splitting. However, the

efficiency of water electrolysers is limited by the slow kinetics of oxygen evolution reaction

(OER). An increase in mass based OER activity is mandatory for a lower capital cost of

electrolyser cells.

In this thesis, synthesis routes are presented for new mesoporous metal oxide films used as

catalytic layers for OER in alkaline or acidic media. The catalytic layers contain different

metal oxides, such as NiO, IrO2 and IrO2/TiO2. The synthesis of metal oxide coatings with

high accessible ordered mesopore structure was achieved via evaporation induced self

assembly. The synthesis succeeds by utilizing PEO-b-PB-b-PEO triblock copolymers as a

pore template, a suitable metal oxide precursor, and a final heat treatment under air. The

synthesis conditions and corresponding physicochemical characterisations are shown in the

discussion.

Correlation of structure and OER activity was used to deduce structure-activity relationships.

These relationships were then used to identify the most significant OER-controlling

parameters. All identified OER-controlling parameters were combined in order to develop an

OER-model describing which structural properties benefit OER activity the most.

Furthermore, the influence on kinetic parameters such as Tafel slope was investigated. High

OER-activities and low Tafel slopes were found for materials with low crystallinity, high

surface area, high conductivity and electrocatalytic measurements conducted at moderate

overpotentials (ca. η < 0,35 V).

In order to demonstrate the commercial relevance of the catalytic layers synthesised in this

work, their OER activity was compared with commercial reference catalysts. It was shown

that oxide layers of this work exhibit a 11 to 24 times higher iridium mass based OER activity

compared to commercial reference catalysts.

Published work

Note that parts of this thesis were already published and that reuse solely for this document

was permitted by John Wiley and Sons.

M. Bernicke, B. Eckhardt, A. Lippitz, E. Ortel, D. Bernsmeier, R. Schmack, R. Kraehnert,

ChemistrySelect 2016, 3, 1-9. http://dx.doi.org/10.1002/slct.201600110

M. Bernicke, E. Ortel, T. Reier, A. Bergmann, J. Ferreira de Araujo, P. Strasser, R.

Kraehnert, ChemSusChem 2015, 8, 1908-1915. http://dx.doi.org/10.1002/cssc.201402988

Contents

1 Motivation ........................................................................................................................... 1

2 State of the Art ................................................................................................................... 2

2.1 Electrocatalytic water splitting and electric properties .................................................. 2

2.2 Oxides with templated porosity ...................................................................................13

2.3 Synthesis routes for metal oxides ...............................................................................21

2.4 Deduced thesis aim and approaches ..........................................................................26

2.5 Thesis outline ..............................................................................................................27

3 Experimental .....................................................................................................................29

3.1 Synthesis of micelle-templated films ...........................................................................29

3.2 Analytical methods ......................................................................................................32

3.3 Electrocatalytic testing ................................................................................................35

4 Results and discussion ......................................................................................................37

4.1 Mesoporous templated NiO synthesized from Ni(NO3)2 and citric acid ........................38

4.2 Mesoporous templated IrO2 synthesized from Ir(OAc)3 ...............................................50

4.3 Mesoporous templated IrO2/TiO2 synthesized from Ir(OAc)3 and TiCl4 ........................65

4.4 Mesoporous templated IrO2/TiO2 synthesized from Ir(OAc)3 and TALH ......................78

4.5 Reference catalysts ....................................................................................................95

5 General discussion ............................................................................................................99

5.1 Influence of precursor on structure/morphology (TiO2, IrO2, IrO2/TiO2) ...................... 101

5.2 Influence of calcination temperature on morphology (TiO2, IrO2, IrO2/TiO2) ............... 104

5.3 Processes during synthesis and the resulting morphology ........................................ 107

5.4 Influence of calcination temperature on OER activity and ECSA ............................... 111

5.5 Influence of crystallinity on OER activity .................................................................... 112

5.6 Influence of electrical conductivity on OER activity .................................................... 114

5.7 Influence of layer thickness on gas removal rate ....................................................... 117

5.8 Tafel slope as a function of calcination temperature and potential ............................ 120

5.9 Investigation of Tafel slopes as a function of iridium loading and potential ................ 123

5.10 Deduced structure-activity relations ........................................................................ 124

5.11 Comparison of iridium-mass based OER activity ..................................................... 127

6 Conclusions and Outlook ................................................................................................. 130

References ......................................................................................................................... 132

Appendix ............................................................................................................................ 138

Acronyms ........................................................................................................................... 141

1 Motivation

Modern society consumes huge amounts of energy relying on the combustion of fossil

resources. The amount of fossil resources such as coal and oil is limited. Therefore, new

routes of energy conversion and storage have to be developed.

Molecular hydrogen (H2) has the highest mass specific energy density of all chemical fuels[1]

and thus appears to be a very attractive candidate for energy storage. The stored energy can

be regained as electrical energy in fuel cells[2] with water as a byproduct. However, depletion

of fossil resources[3] and competition with other industrial applications (e. g. Haber-Bosch-

process,[4] HCl-synthesis,[5] metal oxide reduction[6] and Fischer-Tropsch-Synthesis[7]) require

efficient and sustainable routes for hydrogen generation.

Many routes for generating molecular H2 are available, i.e. it is commonly generated by the

reaction of zinc and hydrochloric acid[8] for laboratory scale purposes. The reaction is

considered as uneconomic thus industrial generation of hydrogen relies on steam reforming[9]

and partial oxidation of methane[10] accompanied by the production of unwanted greenhouse

gases, i.e. CO2.

Several approaches for water splitting are available, such as photocatalysis,[11] thermal water

splitting[12] or water electrolysis.[13] Due to nonpolluting operation conditions and fairly high

efficiencies of 93 % (εΔH),[14] electrocatalytic water splitting is considered a promising

candidate for establishing a sustainable hydrogen economy. Clean energy economy without

indirect CO2 formation is established only when feeding the required energy for electrolysis

from renewable energy sources, such as solar or wind energy.

The efficiency and competiveness of water electrolysis is limited by a complex reaction

mechanism of the oxygen evolution reaction (OER) causing a high anodic overpotential. The

development and improvement of novel electrocatalysts is considered as the major task to

overcome these roadblocks in energy conversion.

2 State of the Art

The following chapter (2) introduces general aspects that are most relevant for the present

thesis. Chapter 2.1 presents principles of electrocatalytic water splitting in alkaline and acidic

electrolyte solutions. Moreover, the chapter contains literature reports of electrocatalysts

used for the oxygen evolution reaction in both media. The reported electrocatalysts usually

contain metal oxides. An increase in surface area is often used to enhance activity which can

be achieved by the introduction of porosity. Chapter 2.2 presents general synthesis concepts

for the preparation of oxides with templated porosity such as evaporation induced self

assembly (EISA). Chapter 2.3 introduces reports from literature exploiting EISA and polymer

template based synthesis routes for the preparation of different mesoporous metal oxides,

i.e. oxides of Al, Mg, Co, Zn, Ti and Ir. The presented synthesis routes are of particular

interest to achieve the specified goals in this thesis (chapter 2.4). A roadmap for the

successful realisation of the aims is presented in chapter 2.5.

2.1 Electrocatalytic water splitting and electric properties

The history of water electrolysis began in the beginning of the 19th century with the discovery

by Nicholson and Carlisle of electrocatalytic water splitting.[15] A century later more than 400

industrial water electrolysis devices were in operation. The first large water electrolysis plant

was deployed in 1939 and provided hydrogen production rates of 10 000 Nm³H2/h. The first

industrial electrolyser capable to work under pressurized conditions was developed in 1948

by Zdansky/Lonza. The development of water electrolysers was continued with the

construction of the first solid polymer electrolyte system by General Electric in 1966. In 1972

the first solid oxide water electrolysis units were introduced.

The following chapter 2.1 deduces the thermodynamic standard potential required for the

electrocatalytic water splitting (2.1.1). Advantages, disadvantages and typical OER catalysts

used for water splitting in alkaline (chapters 2.1.2 - 2.1.3) and acidic media (chapters 2.1.4 -

2.1.5) are presented. Finally, basic investigations reported in literature on structure-activity

relationships are shown for state of the art catalysts used in acidic water splitting, i.e. iridium

oxide. The presented publications underline effects on OER activity and electron transport

mechanism, i.e. layer thickness (chapter 2.1.6), calcination temperature (chapter 2.1.7) and

electrical conductivity (chapter 2.1.8).

2.1.1 Thermodynamic standard potential of water splitting

The net balance for electrochemical water splitting is:

H2O  ½ O2 + H2 | ΔH0 = 285.8 kJ/mol

The minimal required voltage E0 that has to be applied under standard conditions (p,T =

constant) in order to carry out electrolysis of water is expressed as:

𝐸0=−∆𝐺0

𝑛 ∙ 𝐹

Wherein ΔG0 is referred to as the change in Gibbs free energy under standard conditions, F

is related to the Faraday constant (96 485 C/mol) and n expresses the number of electrons

transferred during reaction (in this case n=2).

The fundamental Gibbs-Helmholtz equation is expressed as followed:

∆𝐺0=∆𝐻0− 𝑇∆𝑆0

For water electrolysis under standard conditions the standard reaction enthalpy ΔH0 amounts

to 285.8 kJ/mol, whereas the standard molar entropies ΔS0 are found to be:

∆𝑆0(𝐻2)=130.7 𝐽

𝐾 ∙ 𝑚𝑜𝑙 ; ∆𝑆0(𝑂2)=205.1 𝐽

𝐾 ∙ 𝑚𝑜𝑙 ; ∆𝑆0(𝐻2𝑂) = 70 𝐽

𝐾 ∙ 𝑚𝑜𝑙

The total change of molar entropy for the electrolysis of water is expressed as:

∆𝑆𝑡𝑜𝑡𝑎𝑙

0=∆𝑆0(𝐻2) + 0.5 ∙ ∆𝑆0(𝑂2)− ∆𝑆0(𝐻2𝑂)

And amounts to the following value:

∆𝑆𝑡𝑜𝑡𝑎𝑙

0=(130.6 + 0.5 ∙205.1 −70)𝐽

𝐾 ∙ 𝑚𝑜𝑙 =𝟏𝟔𝟑.𝟏𝟓 𝑱

𝑲 ∙ 𝒎𝒐𝒍

Inserting ΔS0 and ΔH0 in the Gibbs-Helmholtz equation allows the determination of Gibbs free

energy ΔG0 under standard conditions (298 K, 1 bar):

∆𝐺0=∆𝐻0− 𝑇∆𝑆0=285.8 𝑘𝐽

𝑚𝑜𝑙 −298 𝐾 ∙ 163.15 𝐽

𝐾 ∙ 𝑚𝑜𝑙 =𝟐𝟑𝟔.𝟖𝟖 𝒌𝑱

𝒎𝒐𝒍

Moreover the minimum required potential for the beginning of water electrolysis in an open

cell can be determined:

𝐸𝑐𝑒𝑙𝑙

0=−∆𝐺0

𝑛 ∙ 𝐹 =

−𝟐𝟑𝟔.𝟖𝟖 𝒌𝑱

𝒎𝒐𝒍

2∙96485 𝐶

𝑚𝑜𝑙

=−1.23 𝐽

𝐶=−𝟏.𝟐𝟑 𝑽

The negative value indicates that the reaction does not occur spontaneously thus requiring

energy to proceed. In conclusion, the determined voltage as high as 1.23 V is considered as

the thermodynamic minimum potential required for water electrolysis under standard

conditions (25 °C, 1 bar).

Water splitting usually is conducted in either acidic or alkaline media. The advantages and

disadvantages for electrolysers operating in various media are discussed in chapters 2.1.2 -

2.1.5.

2.1.2 General aspects of alkaline water splitting

Alkaline water splitting is characterized by two electrodes immersed in a solution typically

containing 20 % - 30 % KOH. Both electrodes are separated from each other by a diaphragm

ensuring physical segregation of evolving O2 and H2 at the anode and cathode, respectively.

The separation of both gases is necessary for the sake of safety and efficiency. The utilized

diaphragm is permeable for hydroxide and water molecules in order to provide transport of

reactants between both reaction chambers.[13]

Figure 1: Schematic illustration of an alkaline electro-catalytic water splitting cell. Anode and Cathode are

separated by a diaphragm to inhibit cross diffusion.

Anode

OER catalyst

(Ni/Co/Fe)

Cathode

HER catalyst

(Ni)

Diaphragm

½ O

The half-cell reactions for water splitting in alkaline media are:

anode: 4 OH- → 2 H2O + O2 + 4 e-

cathode: 4 H2O + 4 e- → 2 H2 + 4 OH-

overall: 2 H2O → 2 H2 + O2

A closer look at alkaline electro-catalytic water splitting cells reveals several disadvantages.

First, the diaphragm does not fully separate the evolving gases. As a result, evolving O2

diffuses from the anode to the cathode and H2 diffuses from the cathode to the anode. This

cross-diffusion leads to decreased efficiency by the reduction of enriched O2 at the cathode

and can be potentially dangerous due to the formation of a hazardous mixture of O2 and H2 in

the anode chamber. A shutdown and subsequent purge of the chamber with inert gas is

required in order to prevent damage.[13]

The cross diffusion rate is independent of current density. The contamination of both cell

chambers caused by crosss diffusion thus proceeds rather quickly at low production rates

and low current densities. As a result, hazardous working conditions are reached fast and

should be avoided by using high production rates in order to flush the contaminants from the

system.[13]

Many other problems are associated with the use of liquid electrolyte and diaphragms.

Notably, high ohmic losses occur throughout the liquid electrolyte and diaphragm which limits

the maximum achievable current density. Moreover, operation at high pressures is not

possible which results in a clunky stack design.[13]

2.1.3 OER catalysts used in alkaline media

Catalysts used for oxygen evolution in alkaline media are typically based on the oxides,

oxyhydroxides or hydroxides of metals such as CoOx,[16-19] CoFeOx,[16, 20-21] NiOx,[22-27]

NiCeOx,[22] NiCoOx,[28] NiCuOx,[28] NiFeOx,[16, 29-33] NiLaOx,[22] MnOx,[34-35] MnCoOx.[36]

Several reports in the literature show investigations on the electrocatalytic activity of nickel

based water splitting catalysts. The overpotential can be used as a measure of catalytic

activity. Yu et al.[23] prepared a nickel-based thin film (NiOx) on multi walled carbon nanotubes

deposited on ITO and observed an OER overpotential of 523 mV. Singh et al.[24]

electrodeposited layers of nickel oxide from [Ni(en)3]Cl2 (en= 1,2-diaminoethane) on glassy

carbon and reported 510 mV overpotential. Nardi et al.[25] deposited thin films of NiO Ni(Cp)2

(cp=cyclopentadienyl) on fluorine doped tin oxide (FTO) via atomic layer deposition and

achieved an overpotential of 400 mV. Lyons et al.[37] investigated the electrocatalytic activity

of polycrystalline nickel foils and reported an overpotential of 379 mV. Fominykh et al.[26]

prepared NiO nanoparticles using a solvothermal reaction in tert-butanol. Subsequently, NiO

nanoparticles were deposited on Au-coated QCM electrodes in order to access the OER

overpotential (280 mV). Trotochaud et al.[27] obtained a thin film of NiOx by spincoating a

solution containing Ni(NO3)2·6H2O on Au-coated QCM electrodes. Subsequent

electrocatalytic investigation in the OER regime revealed an overpotential of 279 mV.

McCrory et al.[38] published a comprehensive review on the electrocatalytic activity for

different metal oxide catalysts finding that the activity of NiOx catalysts can be further

enhanced by the addition of Fe. Gorlin et al.[30] synthesized Ni-Fe (62:38 at. %) catalysts with

a solvothermal approach and subsequently deposited them on carbon. Oxygen evolution

reaction activity was investigated in 0.1 M KOH revealing an overpotential of 0.25 V

(1 mA/cm²). Gong et al.[32] prepared a nickel-iron layered double hydroxide carbon nanotube

complex by a solvothermal method and showed an OER overpotential of 0.23 V (1 mA/cm²,

0.1 M KOH). Trotochaud et al.[33] electrodeposited thin films of Ni-Fe onto substrates and

subsequently investigated the OER activity revealing an overpotential of 0.27 V at 1 mA/cm²

(1 M KOH). A similar synthesis was conducted by Lu et al.[39] who electrodeposited Ni-Fe on

GC substrate and observed an OER overpotential of 0.31 V at 1 mA/cm² in 0.1 M KOH.

2.1.4 General aspects of acidic water splitting

Many problems arising for electrolysers working in alkaline media can be avoided when

operating under acidic conditions. For instance, the replacement of diaphragms with Nafion®

allows achieving higher current densities and better separation of product gases.[13]

Developed in the late 1960’s by Walter Grot of DuPont.[40-41] Nafion® is described as a teflon

derivate consisting of sulfonated tetrafluoroethylene polymer (PTFE) groups (Figure 2).

Figure 2: Nafion® was invented by Walter Grot of DuPont in the late 1960’s and is described as a polymer

typically deployed in electrolyser cells as a Proton Exchange Membrane (PEM). The polymer overcomes

typical problems observed for alkaline electrolysers, such as low pressure output, low partial load range

and low electric current densities.

x H

Nafion® is applicable in two major fields of electro-catalysis. It can be dissolved in lower

aliphatic alcohols and subsequently used as a binder for attaching a catalyst powder on

substrates or used as a Proton Exchange Membrane (PEM) in acidic electrolyser cells

(Figure 3)

Figure 3: Proton exchange membrane (PEM) electrolyser cell usually consist of OER and HER catalysts

separated by a proton conducting membrane (e.g. Nafion®) embedded in gas diffusion layers and bipolar

plates. Water is fed at the anode side, whereas evolving O2 and H2 is released of the compartment at the

anode and cathode side, respectively.

The following half reactions are observed for a PEM electrolyser cell:

anode: 2 H2O  O2 + 4 e- + 4 H+

cathode: 4 H+ + 4 e-  2 H2

overall: 2 H2O  2 H2 + O2

In general, PEM electrolyzer cells provide several advantages compared to alkaline

electrolysers. First of all, the Nafion membrane is as thin as 20 - 300 µm (commercially

available) and provides high proton conductivity between 0.083 and 0.16 S/cm.[42] This leads

to lower potential drop between the anode and cathode than for alkaline electrolysers.

Hence, current densities above 2 A/cm² can be achieved resulting lower operational costs.

Furthermore proton transport across the membrane responds very fast to the applied power

input, thus lowering the gas crossover rate and consequently increasing H2 purity in the

cathode chamber. In addition, no hazardous condition at low power input is reached in the

OER

catalyst

(Ir)

HER

catalyst

(Pt/C)

Membrane

Cathode

Anode

Gas

diffusion

layer

Bipolar plate

anode chamber. As a result, the operational range appears to be wider than for alkaline

electrolyser cells, whereby the nominal power density ranges between 10 - 100 %.

Electrolyser cells with a solid electrolyte allow a compact system architecture featuring

resistant structural properties allowing operation at high pressure.[43-44] This grants an

important benefit to the end user, who does not have to apply an additional amount of energy

in order to compress the evolved H2 gas. Furthermore, working under high pressure lowers

the specific volume of produced gas thus providing better gas removal.[13]

Even though acidic electrolyser cell provide plenty of advantages, some negative aspects

must be considered. Cross-diffusion of evolved gas in the respective opposite reaction

chambers is still present. Working pressures exceeding values higher than 100 bar further

enhance cross diffusion and usually requires the use of thicker membranes and internal gas

recombiners in order to prevent formation of hazardous gas mixtures. Furthermore, deployed

catalysts, current collectors and separator plates must withstand harsh working environment

characterized by low pH values, high applied potentials (>2 V) and high current densities.

Only a few materials can be applied to operate under such working conditions, e.g. noble

metal as catalysts as well as current collectors and separator plates based on titanium.

Water electrolysis is characterised by the Oxygen Evolution Reaction (OER) occuring at the

anode and the Hydrogen Evolution Reaction (HER) proceeding at the cathode. Typical

electrolysers are limited by the OER due to the fact that four electrons are needed to produce

one molecule of oxygen and that the OER proceeds by a complex reaction mechanism.

[13, 45-47] Active and stable catalysts with high accessible surface area can minimize the

overpotential required for both reactions[47] and potentially lower the amount of noble metal

thus decreasing investment cost of PEM electrolyser cells.

2.1.5 OER catalysts used in acidic media

A wide range of materials are used for the OER reaction, such as mixtures of IrO2/SnO2,[48]

IrO2/Ta2O5,[49] IrO2/Nb2O5,[50] IrO2/RuO2.[51] However, oxides of ruthenium and iridium show

the lowest OER overpotential in acidic media.[13, 52-53] Although ruthenium oxide shows higher

activity,[54] it typically corrodes during OER potential cycles.[55-56] Attempts were made in order

to stabilize RuO2, such as implementing 20 % IrO2 within RuO2. The corrosion rate of the

compound was successfully reduced to 4 % of the original value.[53] Iridium oxide is therefore

the best compromise for an active and stable OER catalyst.[13] Iridium oxide catalysts can be

prepared in different ways. Johnson et al.[57] dropcasted a solution containing iridium acetate

and isopropanol onto a titanium cylinder, followed by heat treatment at 480 °C.

Electrochemical testing in 0.1 M HClO4 indicated overpotentials of about 0.24 VRHE at a

current density of 1 mA/cm². Hu et al.[58] synthesized macroporous IrO2 utilizing colloidal SiO2

as pore template. Electrochemical testing in 0.5 M H2SO4 indicated 0.25 VRHE overpotential at

a current density of 1 mA/cm². Kushner-Lenhoff et al.[59] synthesized iridium oxide layers by

electrodepositing organic precursors Cp*Ir(H2O)3]2+ (Cp* = pentamethylcyclopentadienyl).

OER yielded overpotentials of about 0.267 V at 0.5 mA/cm². A similar synthesis by

Blakemore et al.[60] obtained a catalyst with approximately 0.270 V overpotential (at

0.5 mA/cm²). Oh et al.[61] prepared antimony doped tin oxide (ATO) with a sol-gel and

hydrothermal method. A solution containing H2IrCl6·xH2O, TTAB, water and NaOH was

heated to 70 °C and NaBH4 was added. The solution was cooled down and centrifuged in

order to obtain colloidal iridium nanodendrites. The as prepared colloids were washed,

dispersed in an ethanolic solution and finally deposited on an ATO support. Electrocatalytic

investigations were performed in 0.05 M H2SO4 observing an OER overpotential of 0.27 V at

1 mA/cm². A different approach was pursued by Nong et al.[62] who dealloyed a bimetallic

IrNix alloy precursor. The obtained IrNix@IrOx nanoparticles showed a core-shell architecture

and were deposited on mesoporous ATO. Electrochemical investigations were performed in

0.05 M H2SO4 showing an OER overpotential of 0.26 V at 1 mA/cm². Films of iridium oxide

and nickel oxide were synthesized by Reier et al.[63] via spincoating a solution containing

iridium acetate and nickel acetate tetrahydrate on titanium substrates. The as prepared

samples were heat treated at 450 °C and electrochemically investigated in 0.1 M HClO4. The

catalytic testing revealed an overpotential in the OER of 0.258 V at 1 mA/cm².

In general, iridium oxide catalysts with i) low crystallinity[64] and ii) high surface area[65] are

suggested to be more active for oxygen evolution. These characteristics can either be

adjusted by applying a moderate calcination temperature or by introduction of porosity into

the catalyst.

However, limited abundance and competition with other applications such as

supercapacitors,[66-67] stimulating neural electrodes[68-69] and microelectrodes for pH

sensing[70-71] require the most efficient utilization of IrO2 possible.[47] Many studies were

conducted on determine optimal synthesis conditions for iridium based OER catalysts. Some

of the most important findings are discussed in the chapters 2.1.6 - 2.1.8.

2.1.6 Basic investigations on iridium oxide - role of layer thickness

Johnson et al.[57] investigated the influence of layer thickness on electrochemical active

surface area and OER activity. They prepared, a solution containing Ir(OAc)3 and isopropyl

alcohol. Subsequently, a thin layer was prepared on a titanium substrate by spincoating,

whereas a thick layer was obtained by a dropcasting procedure. Afterwards, the samples

were heat treated at 480 °C in air. Catalyst layers were 17 -19 nm on silicon and even thinner

on titanium. An exact value for the titanium supported layer was not reported. However, the

layer thickness of dropcasted samples was not measured. XRD measurements were

conducted for thin and thick IrOx layers on titanium substrates. The thin IrOx layer showed

significant oxidation of the titanium susbtrate to TiO2, whereas no oxidized titanium was

observed for the thick layer. The authors conclude that, the IrOx layer is shielding the

underlying titanium susbtrate thus preventing oxidation. Investigations of the electrocatalytic

properties showed a 50-60 times higher electrochemical active surface area as well as higher

OER activity for the thicker IrOx layer. Furthermore, normalisation of the OER activity with

respect to the electrochemical active surface area showed a higher intrinsic OER activity for

the thicker IrOx layer related to the absence of an insulating TiO2 interlayer.

The work from Johnson et al. depicts two major advantages for thick IrOx layer on titanium

substrate. At first, they provide higher electrochemical surface area thus higher OER activity.

Furthermore, thick IrOx layers sufficiently prohibit oxidation from underlying titanium susbtrate

to TiO2 consequently providing a higher intrinsic OER activity.

2.1.7 Basic investigations on iridium oxide - role of calcination temperature

The impact of calcination temperature on structural properties of IrOx and electrocatalytic

OER activity was investigated by Reier et al.[64]. Similar to Johnson, a solution containing

Ir(OAc)3 and EtOH was spincoated on titanium coated silicon substrates. Subsequently, heat

treatment was conducted at temperatures ranging from 250 to 550 °C in air. Afterwards,

morphology, crystallinity and chemical state of IrOx was investigated as a function of

calcination temperature. A layer thickness of ~55 nm was found for samples calcined

between 250 and 450 °C, whereas calcination at 550 °C leads to thicker layers indicating

migration from titanium into the IrOx layer. Furthermore, XRD measurements were conducted

and showed higher IrOx crystallite size for samples calcined at higher temperaure. Samples

calcined at 350 °C exhibit the highest OER activity and electrochemical active surface area.

Calcination at 550 °C leads to oxidation of the underlying titanium substrate and enhances

diffusion of TiOx into the IrOx layer consequently leading to higher electrical resistivity

throughout the film volume and to reduced electrocatalytic OER activity.

The work from Reier et al. is of fundamental importance. It showed higher OER activity for

IrOx with smaller crystallite size. Furthermore, negative aspects of insulating TiO2 are

underlined and in good agreement with the observations made by Johnson et al. However,

the impact of TiO2 on electrical conductivity was not investigated in detail.

2.1.8 Basic investigations on iridium oxide - role of electrical conductivity

The influence of electrical conductivity on electrocatalytic activity was investigated by

Marshall et al.[72], by preparing a colloidal dispersion from metal precursors (i.e. H2IrCl6·4H2O

and SnCl2·2H2O) in ethylene glycol. After heating under reflux at a constant pH value of 2.5

the dispersion was centrifuged in order to obtain colloids. The colloids were then dried and

calcined in air at 500 °C. Subsequently, X-ray diffraction measurements revealed a rutile

structure with altering lattice parameters as a function of tin content indicating the formation

of a solid solution of iridium oxide and tin oxide. Electrical resistivity of the powder was

investigated and showed an increase of resistivity depending on the Sn content. Finally, the

electrochemical activity of the powder was investigated in a PEM electrolyser cell. The

activity normalized to the electrochemical active surface area remains high until a tin content

of 50 - 60 mol %, whereas a higher tin content leads to a remarkable drop of intrinsic activity.

The work from Marshall[72] thus underlines that highly active electrocatalysts are obtained

solely by providing a sufficient degree of conductivity.

Furthermore, Chen et al.[73] claims that high electrical conductivity is necessary in order to

provide a stable Ir-based catalyst on titanium substrate. Usually, low electrical conductivity

results in high electric fields throughout the coating causing quick migration of O2- species

through the substrate. Migrating O2- may then have a severe impact on the oxidation of the

underlying titanium substrate. Oxidized titanium leads to an additional ohmic loss thus

lowering the electro catalytic activity. The catalysts further lose activity with ongoing

passivation.

Comninellis et al.[74] studied the impact of electrical conductivity on electrocatalytic OER

activity. A metal salt solution containing at least one of the following compounds H2IrCl6,

RuCl3, TaCl5 or ZrOCl2 was coated on titanium and subsequently heat treated in air at

temperatures below 560 °C. The resulting microstructures of IrOx-TiO2, IrOx-ZrO2 and IrOx-

Ta2O5 were analyzed in terms of crystallinity, conductivity as well as OER activity. It was

found, that the obtained oxides showed low miscibility indicating a distribution of the

respective metal oxide clusters. Taking into account that TiO2, ZrO2 and Ta2O5 possess poor

electrical conductivity, whereas IrO2 shows metallic conductivity[54, 75-76] it was concluded that

electric conductivity is established by chains of conducting IrO2 clusters (percolation theory).

An insufficient amount of IrO2 in the catalyst leads to interruption of conducting chains,

causing a drop of electric conductivity as well as OER activity. Therefore, a significant

amount of metallically conductive metal oxide (i.e. IrO2) must be added in order to provide a

sufficient degree of conductivity that is beneficial for retaining highly active OER catalysts.

Another study investigated the mechanism of electron transport through a catalytic layer.

Oakton et al.[77] synthesized mixed metal oxides of IrO2/TiO2 by a simple one-pot Adams

method route. TiOSO4·0.6H2SO4·1.3H2O was dissolved in water along with H2SO4 and

IrCl3·3H2O. Afterwards, NaNO3 was added to the solution. The solution was dried in a rotary

evaporator and subsequently heat treated at 150 °C for 2h and 350 °C for 1 h. The obtained

powder was washed with water, dried and used for physicochemical and electrochemical

characterization. X-ray diffraction patterns of 40 % IrO2 - 60 % TiO2 revealed the presence of

three different oxides, i.e. TiO2 anatase, TiO2 rutile and IrO2 rutile. TEM images of the

corresponding sample in bright field mode showed IrO2 nanoparticles with 1 nm in diameter

deposited on TiO2 particles with a diameter of ca. 5 nm. The observed electrical conductivity

exponentially increased as a function of iridium content (TiO2: 2.3·10-8 S/cm, IrO2: 3.9 S/cm).

Percolation theory is a well known model to describe the electrical conductivity of randomly

packed particles with high and low electrical conductivity. The so called percolation limit

describes the minimal volume percentage that is required to form a conductive network.

Oakton determined a percolation limit of 35 % and observed an onset of electrical

conductivity for samples with iridium loadings higher than 35 mol % Ir (0.2 S/cm).

Further quantification of conductivity can be achieved by measuring sheet resistivity of

catalyst layers coated on insulating substrates such as glass. Furthermore, electrochemical

methods can be used in order to elucidate electron transport through bulk materials.

Stoerzinger et al.[78] deposited LaCoO3 on three substrates (with different conductivity) by

pulsed laser epitaxy in order to study the impact of conductivity on electron transport through

the layer by analyzing the peak positions of the [Fe(CN)6]3-/4- redox couple. They observed

enlarged distances of redoxpeaks spreading away from the equilibrium potential for samples

with low conductivity. Hence, samples with high resistivity cause a large voltage drop over

the layer leading to a lower voltage which promotes catalysis.

The literature survey from the chapters 2.1.6 - 2.1.8 underlines the influence of critical

reaction parameters that control the OER activity of iridium based catalysts. It can be

concluded that, the iridium based catalytic layer should preferentially be thick, amorphous,

electrically conductive and possess a high electrochemical active surface area in order to

operate as a highly active OER catalyst.

2.2 Oxides with templated porosity

Porous materials are used for different applications in industry and everyday life. Zeolites are

one of the most popular materials and used in washing powder. The properties of

mesoporous materials are presented in a very general manner in chapter 2.2.1. Different

synthesis approaches for the preparation of mesoporous materials are introduced in chapter

2.2.2 and the utilization of amphiphilic surfactants (chapter 2.2.3) i.e. amphiphilic block

copolymers (chapter 2.2.4) will be discussed in more detail. The synthesis of mesoporous

metal oxides with high pore ordering through self assembly of block copolymers (chapter

2.2.5) surrounded by suitable metal oxide precursors (chapter 2.2.6) will be described. The

self assembly of polymer templates can be triggered by exceeding the critical micelle

concentration achieved by ongoing solvent evaporation (EISA, chapter 2.2.7). A final

treatment decomposes the polymer template and converts the precursor into a mesoporous

metal oxide with a nanocrystalline structure (chapter 2.2.8).

2.2.1 Properties of mesoporous materials

Mesoporous systems are described as solid phase materials fully composed of pores. In the

case of connected pores the materials are considered as open porous materials providing

higher inner surface areas than bulk materials. This paves the way for different applications

in the field of catalysis and sorption. Zeolithes are the most prominent porous materials first

used by Henkel in 1977 for water softening. IUPAC classifies porous materials into 3 different

categories[79] according to their pore size.

microporous: < 2 nm

mesoporous: 2 - 50 nm

macroporous: > 50 nm

Microporous and mesoporous systems usually provide sufficient high surface areas for

catalytic processes. A major drawback for microporous and small mesoporous materials is

the occurrence of transport limitation processes during catalysis. Therefore, larger pore

diameters must be incorporated in order to provide sufficient product removal.[80] Mesoporous

systems are distinguished in unordered and ordered materials, both derived by different

synthesis approaches. The most relevant approaches for the introduction of mesoporosity in

materials are discussed in chapter 2.2.2.

2.2.2 Synthesis approaches for mesoporous oxides

Two different synthesis approaches are available in order to obtain mesoporous templated

materials. Endotemplating usually is characterized by a preliminary stage material that is

later surrounded by the mesoporous material precursor, whereas exotemplating first requires

a solid compound with cavities that can be later filled with precursor solution. A subsequent

template removal by etching or thermal treatment obtains the final porous material with a

high surface area (Figure 4).

Figure 4: The a) endo- and b) exotemplating approaches are characterized by the utilizaton of a template

(brown), the occurrence of intermediates (light blue) and the synthesis of a porous material (blue). a)

Endotemplating features the introduction of template in an arising solid material, whereas b)

exotemplating is characterised by the infiltration of a solid template with precursor solution.

Independently of the approach, both routes provide porous materials as a product. (Scheme based on the

work of Ortel[81])

The endotemplating approach is usually used for the synthesis of mesoporous oxides. Within

the approach, the preliminary stage material is a polymer that forms micelles within the metal

oxide precursor solution. After the condensation of the precursors into a “mesophase”, a

thermal treatment will remove the micelles and transform the precursor into a porous metal

oxide. The prior enclosed micelles provide an accessible pore system leading to a porous

metal oxide with high surface area.[82] The characteristics of the pore system such as pore

diameter and pore ordering can be easily tuned by an appropriate choice and concentration

of surfactant. Chapter 2.2.3 introduces the most common class of polymers (i.e. amphiphilic

block copolymers) used for the preparation of mesoporous metal oxides via endotemplating.

2.2.3 Amphiphilic surfactants

Amphiphilic surfactants exhibit a hydrophobic and hydrophilic part leading to versatile

reaction behaviours. The hydrophobic part consists of a large uncharged hydrocarbon

moiety, such as CH3(CH2)n with n > 4 and potentially contains alcohols or ethers, whereas

the hydrophilic part can be either ionic or non ionic. Ionic surfactants are subcategorised into

anionic (e.g. RCO2-) and cationic surfactants (e.g. RNH3+), where R presents the hydrophobic

part of the surfactant. Non ionic or uncharged surfactants usually contain alcohols with large

R groups. Amphiphilic surfactants contain a variety of subcategories, wherein the amphiphilic

a) Endotemplating

approach

b) Exotemplating

approach

block copolymers are from particular interest. Therefore, chapter 2.2.4 presents important

aspects related to amphiphilic block copolymers

2.2.4 Amphiphilic block copolymers

Amphiphilic block copolymers are the main building blocks for the preliminary stage materials

of the endotemplating approach. They are defined as an arrangement of covalent connected

polymers with different composition[83] and are divided in different types, such as diblock

copolymer and triblock copolymer. Figure 5 schematically shows the bonding characteristics

of diblock and triblock copolymers.

Figure 5: “Typical block copolymers” - Diblock and triblock copolymer with the main building block “A”

and “B”. Exemplarily, “A” might consist of a water soluble polymer such as polyethylene oxide (PEO),

whereas “B” is potentially composed of a water insoluble polymer like polypropylene oxide (PPO). BASF

developed such polymers in the 1950’s and still sell them under the tradename Pluronic®. For instance,

Pluronic® F127 is described as PEO106-PPO70-PEO106.

The amphiphilic block copolymers that are used for generating ordered mesoporous

materials mostly contain polyethylene oxide (PEO) as a hydrophilic part. Furthermore, a vast

amount of block copolymers are known in literature:

PS-b-PEO,[84-88] PI-b-PEO,[89-91] PMMA-b-PEO,[92] PHB-b-PEO (KLE),[93-95] PIB-b-PEO,[96-97]

PEO-b-PB)[97] and PEO-b-PB-b-PEO,[98] as well as commercially available block copolymers

like Pluronic® P123 (PEO20-PPO70-PEO20; Mw = 5800 g/mol) and Pluronic® F127 (PEO106-

PPO70-PEO106; Mw = 12600 g/mol).

All these amphiphilic block copolymers possess a variety of properties in solution according

to their structure and concentration, which directly influence the architecture of the

synthesized material. The most important characteristics of amphiphilic block copolymers

used for the synthesis of mesoporous materials will be highlighted in chapter 2.2.5.

AAB

B B B

Diblock copolymer

Triblock copolymer

AAB

B B B

AAAA

2.2.5 Critical micelle concentration and self assembly of block copolymers

Block copolymers in solution with a concentration below it’s critical micelle concentration

(CMC) are dissolved and randomly distributed. Exceeding the CMC significantly changes the

properties of the isotropic solution. The dissolved block copolymers form aggregates in order

to minimize their free energy. These aggregates usually shield the hydrophobic part in the

inside of the micelle, whereas the hydrophilic part keeps contact with the solution. A further

increase of polymer concentration caused by lower amounts of solvent between the micelles

results in rearrangement and various mesophases.[99] Figure 6 exemplarily shows different

mesophases potentially formed by amphiphilic block copolymers.

Figure 6: “Self assembly of block copolymers” - A large variety of different phases are formed by block

copolymers, e.g. spherical, cylindrical micelles with face centered cubic (fcc) and body centered cubic

packing (bcc). Image taken from Bucknall[100] with granted permission from The American Association for

the Advancement of Science.

Figure 6 shows the self organization of block copolymers possessing the capability to form

spherical and cylindrical micelles, vesicles, spheres with face-centered (fcc) cubic and body

centered cubic (bcc) packing, hexagonally packed cylinders, minimal surfaces (gyroid, F

surface, and P surface), simple lamellae, and modulated and perforated lamellae.[100]

The self organization is further exploited in the endotemplating approach by using polymers

as the preliminary stage material along with metal oxide precursor (see Figure 4). The

combination of a metal oxide precursor, and a polymer exhibiting the capability to form an

ordered mesophase is a well known technique in literature[94, 101-103] for the preparation of

mesoporous oxides. Chapter 2.2.6 describes fundamental aspects of one popular metal

oxide precursors, i.e. TiCl4. The rearrangement of polymers to highly ordered mesophases is

often achieved by exceeding the critical micelle concentration triggered by ongoing solvent

evaporation. The so called evaporation induced self assembly is an important tool for the

simple preparation of mesoporous metal oxides and explained in detail in chapter 2.2.7

2.2.6 Hydrolysis and condensation of metal oxide precursors

Suitable precursors must be utilized in order to obtain mesoporous templated compounds.

The metal oxide precursor usually is dissolved and subsequently interacts with solvent

molecules. A general reaction scheme is described as followed:

Hydrolysis: M-X + H2O  M-OH + HX

Condensation: M-OH + M-OH  M-O-M + H2O

Where M = metal centre and X = halide

First, hydrolysis takes place wherein the metal halide reacts with water to form a hydroxide

species under the release of a halide acid. Afterwards, the obtained metal hydroxide species

reacts with itself forming a metal oxide bridged network. Reports in literature describe TiCl4

as a suitable precursor for the synthesis of mesoporous templated TiO2.[99, 102, 104-105] A

general reaction scheme for TiCl4 performing hydrolysis in a solution of ethanol and water

was published for instance by Pan.[99]

TiCl4 + x EtOH + y H2O  TiCl4-x-y(OEt)x(OH)y + (x+y) HCl

Hydrolysis of TiCl4 takes place under the release of HCl triggering hydrolysis. Moreover, the

hydrolysed titanium species can undergo condensation reaction thus forming an oxygen

bridged metal network under the release of EtOH. The release of HCl during hydrolysis of

TiCl4 in solution of EtOH/H2O induces strong acidic environments. As a result, further

hydrolysis of TiCl4 or even polycondensation of partially oxygen bridged titanium species is

suppressed. The obtained species exhibit small molecular mass and are present as inorganic

oligomers in the sol dispersion. Titanium oligomers[106-107] under acidic conditions are

stabilized and terminated by hydroxide groups. They are referred to as nano building blocks

(NBB) due to the interaction of the hydroxide groups with the hydrophilic part of the block

copolymer micelles through hydrogen bonding.[99, 102, 108]

The last chapters pointed out typical characteristics of polymers (chapters 2.2.3 - 2.2.5) and

metal oxide precursors in solution (chapter 2.2.6). Moreover, the interaction between both

reactants was introduced (chapter 2.2.6). The following chapter (2.2.7) depicts the

combination of both reactants in an endotemplating approach for the preparation of a

mesoporous templated metal oxide.

2.2.7 Evaporation Induced Self Assembly (EISA) exemplarily explained with TiO2

The so called evaporation induced self assembly (EISA) requires a solution containing block

copolymers (e.g. PEO-PB-PEO[98, 109]) and metal oxide precursors (e.g. TiCl4,[98] TALH[109]).

The EISA process is described as a combination of the i) cooperative liquid crystal template

(CLCT) mechanism[99, 110] and the ii) true liquid crystal template (TLCT) mechanism[99, 110]

taking place at two different stages during film preparation i) initial sol solution preparation

and ii) film deposition and aging. Therefore, EISA is divided into two stages according to the

present mechanism.[99, 110]

i) Preparation of sol solution (cooperative liquid crystal template mechanism, CLCT)

A homogenous dispersion of block copolymer and inorganic precursor (e.g. TiCl4) is prepared

in a volatile solution (e.g. EtOH/H2O). In general, the inorganic precursor immediately reacts

in terms of hydrolysis and condensation.[99] Thus, leading to the formation of HCl and small

inorganic oligomers terminated by hydroxide groups referred to as nano building blocks

(NBBs, chapter 2.2.6).[106-107] The HCl formation causes an acidic environment consequently

suppressing further polycondensation of NBBs. Subsequently, the partially hydrolyzed

titanium species interact with hydrolytic regions of the block copolymers through hydrogen

bonding.[99, 102, 108] This cooperative liquid crystal template (CLCT) mechanism leads to the

formation of titanium oligomers attached to the block copolymers.

ii) Film deposition and ageing (true liquid crystal template mechanism, TLCT)

The as prepared solution is characterised by a dissolved polymer interacting with dissolved

metal oxide precursors through hydrogen bonding. The solution is now used for film

deposition on suitable substrates such as silicon or titanium. Dipcoating appears to be a

convenient method due to control of temperature, humidity and withdrawal rate.[111]

Immersion and consecutive dragging out of the substrate from the solution triggers several

processes. The continuous evaporation of the solvent induces an increase of polymer- and

metal oxide precursor concentration. At first, the enrichment of the block copolymer induces

phase segregation between the hydrophilic and hydrophobic part. After exceeding the critical

micelle concentration the block copolymers self assemble to micelles (TLCT).[99, 110] At

second, other volatile compounds besides EtOH evaporate (e.g. HCl). Thus, leading to an

increase of pH-value and consequently triggering the thermodynamically favourable

condensation and polycondensation reactions of the hydrolysed metal oxide precursor (sol-

gel process). Finally, a connected amorphous metal oxide network is formed around the

highly ordered self assembled micelles.[99]

Figure 7 schematically illustrates the evaporation induced self assembly of block copolymers

and metal oxide precursors during dipcoating.

Figure 7: Schematic illustration of the Evaporation Induced Self Assembly (EISA). A volatile solution

containing a block copolymer and metal oxide precursor is prepared. The substrate is immersed in

solution and consequently dragged out. Evaporation of solvent and HCl (in case of TiCl4) leads to

enrichment of the polymer and precursor concentration inducing self assembly of the polymers and

polycondensation of the oxide precursor. In the end, a connected amorphous metal oxide network is

formed around the ordered micelles.

A variety of different porous metal oxide compositions have been prepared by the

evaporation induced self assembly process. Chapter 2.2.8 presents the transformation of the

arranged spherical micelles surrounded by metal oxide precursor into a mesoporous

templated material.

Micelle-precursor-film

„mesophase“

Substrate

Evaporating

solvent

Polymer

Dissolvedprecursor

Withdrawal

direction

2.2.8 Transformation of ordered mesophases to mesoporous oxides

The micelle-precursor films prepared according to the procedure described in chapter 2.2.7

are subsequently dried at temperatures between 60 - 200 °C[99] in order to grant full

polycondensation of nano building blocks leading to a more stable mesophase. Afterwards,

heat treatment at higher temperatures is performed to transform micelle-precursors into a

mesoporous templated metal oxide with nanocrystalline structure. In case of Pluronic F127

(PEO106-PPO70-PEO106; Mw = 12600 g/mol) the decomposition temperature was found to be

340 °C,[112] whereas Pluronic P123 (PEO20-PPO70-PEO20; Mw = 5800 g/mol) decomposes

around 220 °C.[113] Thermally induced crystal growth of pore walls composed of TiO2 is

observed for calcination at 350 °C.[114] Decomposition of pore template and transformation of

amorphous pore wall into a crystalline material typically is accompanied by film shrinkage

perpendicular to the substrate. Due to strong attachment of metal oxide precursor on the

substrate surface the occurring contraction of the film is only present perpendicular to the

substrate thus prevent cracking of the layer.[81, 99] In case of TiO2, a lower crystallization

temperature accompanied by less film contraction was observed for slow heating ramps of

0.5 - 2 °C/min.[115] Figure 8 illustrates the heat treatment process of a micelle-precursor-film

mesophase into a crystalline oxide with an accessible porous system.

Figure 8: General scheme for the synthesis of mesoporous templated metal oxides. The micelle-

precursor-film obtained by dipcoating (described in chapter 2.2.7) is thermally treated in air. Thus,

decomposition of micelles and transformation of the precursor into a metal oxide with nanocrystalline

structures takes place. Furthermore, template removal is accompanied by film shrinkage perpendicular to

the substrate.

Polymer templating is a powerful route to introduce mesoporosity in various metal oxides, i.e.

TiO2,[98, 116] IrO2,[65] NiO,[117] RuO2,[118] SnO2,[105] Nb2O5,[105] MgO.[103] Thus, the next chapters

describe different synthesis approaches for the preparation of Al2O3, MgO, Co3O4, ZnO (all

chapter 2.3.1), TiO2 (chapter 2.3.2), and IrO2 (chapter 2.3.3).

Calcination

Layer thickness

Substrate

Micelle

Precursor

Micelle-precursor-film

„mesophase“

Mesoporous templated

metaloxide

Meso-

pore

Nano-

crystal

2.3 Synthesis routes for metal oxides

The following chapter (2.3) introduces synthesis concepts based on different metal oxide

precursor systems for the preparation of mesoporous metal oxides such as Al, Mg, Co, Zn, Ti

and Ir. Chapter 2.3.1 presents a very simple and versatile approach for the successful

preparation of metal oxides from aluminum, magnesium, cobalt and zinc. The synthesis

succeeds by employing a metal complex formed from citric acid and metal nitrate along with

a triblock copolymer PEO213-PB184-PEO213 employed as a mesopore template. Chapter 2.3.2

introduces exemplarily different precursor concepts reported in literature to obtain

mesoporous templated titania. Finally, chapter 2.3.3 shows to the best of our knowledge, the

only synthesis concepts based on polymer templating for the preparation of mesoporous

iridium oxide.

2.3.1 Synthesis concepts for mesoporous Al2O3, MgO, Co3O4, ZnO - The “citrate route”

Eckhardt et al.[103, 119] presented a synthesis route which combines three different strategies

such as i) complexing the metal ion (Pechini method), ii) decomposing the carbonate and iii)

pore templating with polymer micelles. The approach tackle typical problems associated with

the building mechanism of mesoporous oxides, thus granting access to various metal oxide

and even carbonate compounds of Al,[119] Mg,[103] Co,[119] and Zn.[119]

Figure 9: Scanning electron microscopy (SEM) images of accessible mesoporous templated metal

carbonates and oxides of aluminium, magnesium, cobalt and zinc. All compounds were synthesized via

citrate route. (SEM images taken from Eckhardt et al.[103, 119] and reprinted with permission from the

American Chemical Society)

Five different boundary conditions must be fulfilled in order to provide a generalized

synthesis approach for metal oxides. Figure 10 illustrates the different requirements usually

associated with the synthesis strategy:

Remaining template

polymer prevents

imaging by electron

microscopy due to

instability of

the sample.

50 nm

M(CO

)

Mg Co

a) b) c)

e) f) g)

50 nm

50 nm 50 nm 50 nm 50 nm

50 nm

1) Metal salts have to form a chemical complex with ligands containing carboxylic acid

functionality (e.g. citric acid). Moreover, chelating ligands are favored due to their higher

chemical stability.

2) The solution is coated onto a substrate leading to solvent evaporation. The containing

polymer should possess the capability of evaporation induced self assembly. Self assembly

is required for the successful formation of an ordered micelle-precursor-film mesophase.

3) The metal complex should show the ability to transfer into a carbonate at temperatures

below the descomposition temperature of the polymer template. Transformation of the metal

complex into a carbonate prior template removal is epecially important for preventing pore

structure collapsing thus providing stabilization of the micelle-precursor-film mesophase.

4) The pore template must be removed prior to metal carbonate decomposition in order to

obtain a mesoporous metal carbonate. Otherwise, the mesoporous structure of the metal

carbonate is not accessible due to the presence of pore template.

5) Finally, a calcination step is applied to decompose the metal carbonate consequently

leading to formation of nanocrystalline metal oxides with preserved pore structure.

Figure 10: The citric acid complexed metal nitrate synthesis route is shown together with the deduced

requirements (1-5). Necessary requirements for the successful preparation of mesoporous templated

metal oxides are as followed: 1) Formation of a soluble metal complex (i.e. metal nitrate and citric acid). 2)

Film deposition to induce self assembly of polymer template and obtain an ordered mesophase. 3)

Precursor decomposition and transformation of the complex into a structurally stable carbonate at low

temperatures under retaining the ordered mesophase. 4) Heat treatment in air to decompose polymer

template and generate mesoporous metal carbonate with accessible mesoporosity. 5) Heat treatment at

higher temperatures to transfer amorphous metal carbonates into nanocrystalline metal oxides under the

release of CO2 and preservation of the mesoporous system. Image taken from Eckhardt et al.[119] and

reused with the permission of the American Chemical Society.

Micelles

Mesoporous

metal oxide

Metal complex

M(CO3)x

PEO-b-PB-b-PEO

Mesoporous

metal carbonate

MOx

Micelle-structured

metal carbonate

Dip-coating solution

Metal nitrate

Polymer template

Citric acid

Micelle-structured

metal-complex film

Coating

T T T

Required:(1) Formation of a

soluble metal

complex

(2) Self-assembly into an

ordered mesophase

(4) Template removal without

mesopore collapse

(3) Transformation of the complex

into a structurally stable metal

carbonate at low temperatures

(5) Transformation of

carbonate into metal oxide

without mesopore collapse

Essential: Suitable thermal (in)stability of template,

metal complex and metal carbonate

In case of MgO[103] the metal complex is formed in solution by treating magnesium nitrate with

citric acid. The as prepared metal complex is transferred into a stable MgCO3 phase by

thermal treatment. Subsequently, a second calcination step at higher temperature leads to

decomposition of the carbonate and finally results in the formation of a mesoporous

templated MgO phase. The synthesis was conducted in ethanolic solution containing PEO-

PB-PEO triblock copolymer employed as a polymer template and a metal complex as an

oxide precursor. Dipcoating was performed on silicon substrate followed by calcination at 400

and 600 °C to obtain a mesoporous magesium carbonate and oxide. Another publication of

Eckhardt et al.[119] demonstrate the versatility of the citric acid approach by showing the

successful synthesis of micelle-templated oxides and carbonates of zinc, cobalt and

aluminium (see Figure 9 for corresponding SEM images).

2.3.2 Synthesis concepts for mesoporous TiO2

Several synthesis routes are available for the preparation of mesoporous templated TiO2.

Figure 11 exemplarily shows a selection of electron microscopy images of TiO2 obtained by

different metal oxide precursors such as a) Ti(iPrO)4, b) TiCl4, c) preformed TiO2 nanocrystals

and d) TALH.

Figure 11: SEM micrographs of mesoporous templated TiO2 synthesized from different precursors. a)

Antonelli et al.[120] prepared pre-stabilized titanium isopropylate with acetylacetonate and subsequently

introduced pore templating polymers. b) Yang et al.[105] achieved self-stabilization of titanium species by

providing intrinsically acidic environments through HCl, which is released by the reaction of TiCl4 in water

ethanol-rich solutions. c) Brezesinski et al.[95] first prepared TiO2 nanocrystals and subsequently added

KLE as a pore template. d) Ortel et al.[109] worked in non acidic conditions by employing lactic acid

complexed titanium (TALH) as a titania source and PEO-PB-PEO triblock copolymer as a mesopore

template. All pictures have been taken from the respective publications.

a) One of the first synthesis approaches to obtain mesoporous TiO2 was published by

Antonelli et al.[120] The authors prepared acetylacetonate stabilized titanium alkoxides such as

Ti(iPrO)4. The exclusion of the stabilizing agent (acetylacetonate) immediately led to

excessive hydrolysis and condensation reaction, thus precipitation of the oxide-alkoxide

aggregates. However, the stabilized titanium alkoxide was added to a solution containing a

Ti(

PrO)

+ stabilizer TiCl

TiO

nanocrystals TALH

a) b) c) d)

polymer template (e.g. tetradecylphosphate). After several synthesis steps including a final

heat treatment under air preserved the first hexagonally packed mesoporous TiO2 with a

surface area of 200 m²/g and a pore opening with 3.2 nm in diameter.

b) No stabilizing agent is necessary in the case of TiCl4 being utilized as a TiO2 source as it

provides self stabilization. The route was published by Yang et al.[105] and exploits the

formation of HCl during hydrolysis of TiCl4 in a water ethanol-rich solution. The formation of

HCl induces strong acidic environments thus suppressing further polycondensation of the

oxide-alkoxide aggregates. The stabilized titanium alkoxides are subsequently mixed with

suitable polymers (e.g. P123) and heat treated in air to obtain mesoporous templated TiO2 in

anatase phase with a crystallite size of 2.4 nm. Further physicochemical analysis revealed a

pore size of 6.5 nm and a BET surface area of 205 m²/g.

c) Brezesinski et al.[95] sythesized pre-formed TiO2 nanocrystals from TiCl4, EtOH and benzyl

alcohol. Afterwards, the obtained nanoparticles were redispersed in ethanol and

subsequently mixed with KLE as a pore template. Dipcoating and heat treatment in air at

600 °C leads to the formation of mesoporous templated TiO2 layers. The derived layers show

bimodal porosity with 1 - 4 and 20 - 25 nm diameter pores. Further physicochemical analysis

revealed a body centered cubic packing of the pore system and a BET surface area in the

range between 180 - 200 m²/g.

d) Moreover, Ortel et al.[109] prepared a solution of lactic acid complexed titanium (titanium(IV)

bis(ammonium lactato)dihydroxide abbreviated as TALH), polymer template (e.g PEO-PB-

PEO, F127), water and ethanol. Due to the high chemical stability of the metal complex no

hydrolysis and condensation reaction occur under non acidic conditions. Subsequently, the

solution was dipcoated on different substrates and heat treated in air. The prepared

mesoporous templated TiO2 showed high surface area, exhibited an anatase structure and

showed locally ordered mesopores with an opening of 13.5 nm (F127) and 29.4 nm (PEO-

PB-PEO) in diameter.

2.3.3 Synthesis concepts for mesoporous IrO2

Only a few precursor concepts for the preparation of polymer templated IrO2 is reported in

literature. Figure 12 shows different iridium oxide precursors used along with amphiphilic

block copolymers to obtain mesoporous IrO2.

Figure 12: SEM images of polymer templated IrO2 obtained by a) F127, K2IrCl6 and alkaline hydrolysis[121]

and b) PEO-PB-PEO and Ir(OAc)3. [65] Images were taken from the respective publications.

a) Chandra et al.[121] used PEO-PPO-PEO (F127) block copolymer as a pore template and a

monomeric [Ir(OH)6]2- complex obtained by alkaline hydrolysis of K2IrCl6 as an oxide

precursor. The solution was spincoated onto FTO substrate dried at 80 °C, annealed at

150 °C and finally heat treated at temperatures between 400 - 500 °C, respectively. Thicker

layers were obtained by multi spincoating with intermediate drying (80 °C) and annealing

(150 °C). Physicochemical analysis revealed a pore diameter of 7 nm, layer thickness of ca.

70 nm and BET surface areas between 65 m²/g (Tcalc. = 400 °C) and 105 m²/g (Tcalc. =

500 °C).

b) Ortel et al.[65] prepared an ethanol rich solution containing iridium acetate as a metal oxide

precursor and employed PEO-PB-PEO polymer as a pore template. The solution was

dipcoated on different substrates and consequently heat treated under air. The authors

reported a pore diameter of ~16 nm, rutile type structure, a locally ordered pore arrangement

and BET surface areas between 140 (Tcalc. = 300 °C) and 20 (Tcalc. = 700 °C) m² per m² of

geometric surface area.

Ir(OAc)3

K2IrCl6

2.4 Deduced thesis aim and approaches

The aim of this thesis is to obtain catalysts with higher electrocatalytic oxygen evolution

reaction activity compared to commercially available catalyst powders and to understand the

factors that control OER activity.

To achieve this goal the synthesis route presented for oxides of Al, Mg, Co, Zn (see chapter

2.3.1) is extended to Ni for electrochemical testing in alkaline media. The synthesis

approaches for IrO2 by Ir(OAc)3 (chapter 2.3.3) and TiO2 from TiCl4 and TALH (chapter 2.3.2)

are combined in order to obtain mesoporous templated layers of IrO2 and mixed metal oxides

IrO2/TiO2 with higher catalytic activity in acidic media compared to a commercial reference.

A controlled variation of synthesis parameters during the preparation of mesoporous

templated catalytic layers, i.e. NiO, IrO2, IrO2/TiO2 was performed. An extensive analysis for

the impact of synthesis parameters on structural properties and electrochemical activity was

performed. The obtained data were then evaluated in order to identify the optimal synthesis

conditions for the generation of highly active electrocatalysts. The varied parameters were: i)

calcination temperature, ii) iridium loading and iii) withdrawal rate.

i) Heat treatment conducted at different temperatures was used in order to study the effect of

crystallinity on intrinsic and overall OER activity. ii) A different amount of iridium oxide was

dispersed within TiO2 to obtain samples with different electrical conductivity. The different

electrical conductivity was then used to identify changes in electron transport mechanisms.

iii) By varying the withdrawal rate electrocatalytic coatings with different layer thicknesses

were obtained. The catalysts were used as a model type system to identify potential gas

removal limitation at high electrical current densities and potentials.

A comprehensive investigation and comparison of physicochemical and electrochemical

properties revealed that the overall OER activity is affected by several parameters such as

crystallinity, electrical conductivity and active surface area. However, no signs of transport

limitations were detected for systems with high layer thicknesses at high gas production

rates.

2.5 Thesis outline

Chapter 3 illustrates experimental data by offering information about the synthesis of

mesoporous templated catalytic coatings on different substrates (chapter 3.1). Moreover, the

applied physicochemical techniques (chapter 3.2) to determine structural properties and

electrocatalytic testing procedures (chapter 3.3) for quantifying the oxygen evolution reaction

activity and electrochemical accessible surface area are explained.

The results and discussion are presented in chapter 4. Chapter 4.1 describes a new

approach for the synthesis of NiO coatings with narrow pore size distribution and tunable

mesopore size. The synthesis is based on micelles of amphiphilic block copolymers

employed as a pore template and citric acid complexed metal compounds. The versatility of

the citrate based synthesis approach is confirmed by showing the accessibility of

homogeneous NiO coatings with templated and locally ordered mesoporous structure. The

electrochemical OER activity is investigated in alkaline media to deduce structure activity

relationships. Moreover, a comparison of OER activity with reports from literature is

provided.[122]

Chapter 4.2 exploits the potential of pore templating with polymer micelles in order to

produce model-type porous catalysts for the investigation of structure-activity relationships in

gas evolution reactions and for the optimization of the performance of IrO2-based OER

catalysts. The developed synthesis from Ortel[65] based on pore templating with micelles of

amphiphilic block copolymers PEO-PB-PEO[65, 98] is further improved to produce iridium oxide

films with controlled pore size, film thickness and crystallinity. The model systems are used to

study the influence of porosity and crystallinity on electrochemically accessible surface area

(ECSA), OER performance and gas transport. A controlled variation in thickness of the

porous catalysts explores, which parts of the catalyst can be utilized without transport

limitations during high-current OER. The combined knowledge is used to design a multilayer-

IrO2 catalyst that shows the lowest overpotential reported so far for monometallic oxide

compounds of Ir.[47]

Chapter 4.3 and 4.4 investigate two different synthesis routes for the preparation of mixed Ir

and Ti metal oxides. Both routes are based on a soft templating approach with PEO213-PB184-

PEO213 as a pore template, iridium acetate as an IrO2 source and either TiCl4 (chapter 4.3) or

TALH (chapter 4.4) as a titania precursor. The chemical properties of TiCl4 and TALH vary

from each other leading to different interaction with the dissolved polymer template thus

providing different morphological properties such as metal oxide distribution. The influence of

calcination temperature and iridium content for both synthesis routes is investigated in terms

of surface area, crystallinity, conductivity, morphology and OER activity. Correlations of the

obtained data are used to identify OER-controlling parameters for the successful preparation

of highly active OER electrocatalysts.

Chapter 4.5 illustrates the electrocatalytic performance of commercially available catalyst

powders. The purchased powders were deposited on the working electrode by a typical

“Nafion-ink” based synthesis approach. Electrocatalytic investigations were conducted in

order to identify the most active commercial powder. Moreover, the reproducibility of the ink-

based synthesis approach is shown by applying different iridium loadings on the working

electrode. The obtained data are used to determine the iridium mass based OER activity for

commercial reference catalyst powders.

Chapter 5 presents combined observations of systems used for electrochemical testing

under acidic condition, i.e. IrO2 and IrO2/TiO2. This general discussion is used to identify

general trends and deduce potential structure-activity relationships. Synthesis parameters

such as calcination temperature, crystallinity, electrical conductivity, electrochemical active

surface area (ECSA) and layer thickness were identified as the most relevant OER-

controlling parameters. Furthermore, the applied overpotential during electrocatalytic

investigation showed to possess an impact on kinetical aspects, i.e. Tafel slope. A

comprehensive study on observed structure-activity relations revealed high OER activities for

samples exhibiting low crystallinity, high electrical conductivity and large ECSA. Synthesized

catalytic coatings of IrO2 (chapter 4.2) and IrO2/TiO2 (chapter 4.3 (Ir(OAc)3 + TiCl4) and

chapter 4.4 (Ir(OAc)3 + TALH)) fulfilling the observed criteria are then used for comparison

with commercial catalytic powders (from chapter 4.5) in terms of iridium mass based OER

activity. The observed iridium mass normalized OER activity of polymer templated IrO2 and

IrO2/TiO2 appear to be at least 10 times higher than for commercially available catalysts.

3 Experimental

The present chapter (3) contains information about the synthesis of micelle-templated films

(chapter 3.1) and the applied analytical methods in order to determine morphology,

crystallinity, phase composition, surface area, pore ordering, surface composition and

electrical conductivity (chapter 3.2). Moreover, the electrocatalytic testing is described in

detail (chapter 3.3) offering information about the investigation of oxygen evolution reaction

(OER) activity and electrochemical active surface area (ECSA) in acidic and alkaline media,

respectively.

3.1 Synthesis of micelle-templated films

Dipcoating (Coater 5 AC, IdLabs Vesely) was used in order to deposit thin metal oxide films

on a suitable substrate. Different substrates were used for the investigation of certain

structural and catalytical properties: Physicochemical analysis was conducted for samples

coated on single side polished silicon wafers (SEM, XRD, TEM, SAED, XPS), double side

polished silicon wafers (physisorption), microscope slides (electrical conductivity), titanium

foil (SAXS) and titanium sheets/cylinders (SEM, XRD, TEM, SAED) Electrochemical

investigations were solely performed for layers coated on titanium sheets and cylinders

(OER, ECSA, DEMS). Prior to coating, silicon as well as titanium substrates passed through

different cleaning procedures. Silicon wafers, titanium foil and microscope slides were

washed with ethanol, whereas titanium substrates were firstly polished by SiO2-polishing

paste (Buehler, MasterMet 2, noncrystallizing colloidal silica suspension, 0.02 µm), then

ultrasonicated in water and eventually rinsed with ethanol (VWR Chemicals, 99.98 %

absolute).[122]

3.1.1 Mesoporous nickel oxide films

The solution for dipcoating was obtained by first dissolving 70 mg of a polymer template

poly(ethylene oxide)-b-poly(butadiene)-b-poly(ethylene oxide) (PEO213-PB184-PEO213,

containing 18 700 g mol-1 PEO and 10 000 g mol-1 PB, from Polymer Service Merseburg

GmbH)[65, 98] in 0.2 mL water (milliQ) and 2.8 mL ethanol. The solution was stirred for 12 h.

Subsequently, 144.1 mg of citric acid (Carl Roth, 99.5 % p.a.) and 436.0 mg of

Ni(NO3)2·6H2O (NeoLab, 98 %) were added.

The obtained green solution was transferred to a cuvette inside a dipcoater, where a

controlled atmosphere had been established with a temperature of 25 °C and a relative

humidity of 40 %. Dipcoating was performed at a withdrawal rate of 200 mm/min. Freshly

coated samples were dried for another 10 minutes under same conditions. Followed by two

subsequent heat treatments, the freshly synthesized samples were first thermally treated for

1 hour in a prehated drying oven at 250 °C and then calcined for 1 hour in a preheated muffle

furnace at 350, 400, 450 and 500 °C, respectively.

3.1.2 Mesoporous iridium oxide films synthesized from Ir(OAc)3

Mesoporous IrO2 films were synthesized on polished titanium cylinders following a procedure

established by Ortel et al.[65] Dipcoating solutions for calcination studies were obtained by

adding 225 mg Iridium(III) acetate (Heraeus, 48.76 % Ir content) and 45 mg of a polymer

template poly(ethylene oxide)-b-poly(butadiene)-b-poly(ethylene oxide) (containing

18 700 g/mol PEO and 10 000 g/mol PB, from Polymer Service Merseburg GmbH)[98] to a

volume of 1.5 mL ethanol. For thicker films and multilayers a volume of 1.3 mL EtOH and

0.1 mL H2O was used.

Dipcoating of single layer catalysts was performed under controlled atmosphere at a relative

humidity of 40 % and a temperature of 25 °C at a withdrawal rate of 100 mm/min in a

dipcoater. Films were calcined subsequently placing the samples for 5 min in a hot muffle

furnace with air atmosphere at temperatures between 325 and 625 °C, respectively. For the

controlled variation of film thickness the substrates withdrawal rate was adjusted between 10

and 150 mm/min using the same calcination routine (5 min, 375 °C). Multilayer catalysts

were obtained by repeated dipcoating (30 mm/min) and heat treating (30 min, 200 °C) of

individual layers followed by a final calcination step (5 min, 375 °C).[47]

3.1.3 Mesoporous mixed oxide films of Ir and Ti synthesized from Ir(OAc)3 and TiCl4

The dipcoating solution for the preparation of 30 wt. % Ir in Ir/TiO2 was obtained by dissolving

35 mg of a polymer template poly(ethylene oxide)-b-poly(butadiene)-b-poly(ethylene oxide)

(PEO213-PB184-PEO213, containing 18 700 g mol-1 PEO and 10 000 g mol-1 PB, from Polymer

Service Merseburg GmbH)[65, 98] in 1.25 mL ethanol (99.9 %, Roth) and 0.21 mL milliQ water.

Subsequently, a solution containing 1.25 mL EtOH and 93.3 µL of titanium(IV) chloride (TiCl4,

99.9 %, Sigma Aldrich) was added dropwise under stirring. Finally, 55.8 mg Ir(III) acetate

(Heraeus, 48.76 % Ir content) was added to the solution.

The obtained dark green solution was transferred to a cuvette inside a dipcoater, where a

controlled atmosphere has been established with a temperature of 25 °C and a relative

humidity of 40 %. Dipcoating was performed at a withdrawal rate of 200 mm/min on polished

titanium substrate. Freshly coated samples were dried for another 10 minutes under the

same conditions. Afterwards, the samples were calcined in a preheated muffle furnace for

10 minutes under air at temperatures between 200 to 600 °C, respectively.

3.1.4 Mesoporous mixed oxide films of Ir and Ti synthesized from Ir(OAc)3 and TALH

Solution for dipcoating a layer of 30 wt. % Ir in Ir/TiO2 was obtained by dissolving 35 mg of a

polymer template poly(ethylene oxide)-b-poly(butadiene)-b-poly(ethylene oxide) (PEO213-

PB184-PEO213, containing 18 700 g mol-1 PEO and 10 000 g mol-1 PB, from Polymer Service

Merseburg GmbH)[65, 98] and 55.8 mg Ir(III) acetate (Heraeus, 48.76 % Ir content) in 2.5 mL

methanol (Sigma Aldrich). Subsequently, 408.6 µL of Titanium(IV) bis (ammonium lactate)

dihydroxide solution (TALH, 50 wt. % in H2O, Sigma Aldrich) was added dropwise under

stirring.

Obtained dark green solution was transferred to a cuvette inside a dipcoater, where a

controlled atmosphere has been established with a temperature of 25 °C and a relative

humidity of 40 %. Dipcoating took place with a withdrawal rate of 200 mm/min on polished

titanium substrate. Freshly coated samples were dried for another 10 minutes under the

same conditions. Afterwards, the samples were calcined in a preheated muffle furnace for

10 minutes under air at temperatures between 200 to 600 °C, respectively.

3.1.5 Reference catalyst (IrOx/TiOx)

The dispersion for dropcasting a layer of commercial availabe IrOx/TiOx was derived by

suspending 5.4 mg of IrOx/TiOx powder (73.9 wt. % Ir in TiO2, “Elyst 750480”, Umicore) in

2.49 mL water (milliQ), 2.49 mL iPrOH (> 99.5 %, Roth) and 20 µL Nafion perfluorinated resin

solution (5 wt. % in lower aliphatic alcohols and water, contains 15-20 % H2O, Sigma

Aldrich). The dispersion was ultrasonicated for 15 minutes showing a black appearance

(“ink”). Subsequently, 5 µL of the ink was pipeted onto a polished titanium cylinder (A =

0.1963 cm²) under room temperature. The titanium cylinder was transferred into a preheated

furnace and dried at 60 °C for 5 minutes. The nominal loading of iridium metal on the titanium

cylinder amounts to ca. 4 µg (~20 µgIr/cm²). Higher loadings were achieved by multi

dropcasting (repeated pipeting of 5 µL several times) with intermediate drying steps at 60 °C

for 5 minutes. The procedure was used to produce nominal loadings (iridium metal) between

ca. 4 - 800 µgIr (20 - 4075 µgIr/cm²).

3.1.6 Reference catalyst (IrO2, Sigma Aldrich)

5 mg of commercially available IrO2 powder (99.9 %, Sigma Aldrich) was dispersed in

2.49 mL water (milliQ), 2.49 mL iPrOH (> 99.5 %, Roth) and 20 µL Nafion perfluorinated resin

solution (5 wt. % in lower aliphatic alcohols and water, contains 15-20 % H2O, Sigma

Aldrich). The suspension was ultrasonicated for 15 minutes showing black colour.

Subsequently, 3 µL of the dispersion was dropcasted onto a titanium cylinder followed by a

drying step in a preheated furnace at 60 °C for 5 minutes. The obtained IrO2 catalyst on a

titanium cylinder exhibit nominal loadings of 2.57 µgIr (13.1 µgIr/cm²).

3.2 Analytical methods

The present chapter (3.2) offers information about the spectrum of physicochemical

techniques used for determining the morphology (3.2.1: SEM, 3.2.2: TEM), crystallinity and

phase composition (3.2.2: SAED, 3.2.3: XRD), surface area (3.2.4: Kr-physisorption), locally

pore ordering (3.2.5: SAXS), surface composition (3.2.6: XPS), electrical conductivity (3.2.7),

iridium content (3.2.8: ICP-OES) and faraday efficiency (3.2.9: DEMS).

3.2.1 Scanning electron microscopy (SEM)

SEM images were obtained using a JEOL 7401F instrument with an accelerating voltage of

10 kV. To determine film thickness of mesoporous iridium oxide films, a ceramic knife was

used to scratch the layer followed by SEM imaging of tilted Ti cylinders. SEM images were

analyzed with ImageJ v1.43u[123] to derive film thickness and pore diameter.[47]

3.2.2 Transmission electron micrographs (TEM) and selected area diffraction (SAED)

TEM and SAED images were obtained with a FEI Tecnai G² 20 S-Twin at an accelerating

voltage of 200 kV. The film samples were scraped off from a titanium cylinder or sheet and

then collected with a TEM-grid.[47] TEM micrographs were obtained in brightfield and high

angle annular darkfield (HAADF, 3° tilt angle) mode. SAED images were recorded in

diffraction mode at a sample area of ca. 200 nm in diameter. During TEM operation the

electron dose was lowered by broadening the electron beam to prevent potential reduction of

metal oxide compounds.

3.2.3 X-ray diffraction (XRD)

XRD patterns were recorded with a Bruker D8 Advance instrument using Cu Kα radiation,

grazing incident for the incoming beam and a Goebel mirror. Scherrer equation was applied

to certain reflections in order to determine crystallite size for phases of IrO2 (110),[47] TiO2

rutile (110), TiO2 anatase (101) and NiO bunsenite (111) and (200).

3.2.4 Kr-physisorption

Surface area was measured with Kr physisorption at 77 K using the Autosorb-iQ automated

gas adsorption station from Quantachrome and samples coated on double side polished

silicon wafers. Prior to adsorption measurement the samples were degassed for 2 h at

150 °C in vacuum. The surface area was calculated via Brunauer-Emmett-Teller (BET)

method.[122]

3.2.5 Small angle X-ray scattering (SAXS)

NiO 2D small angle X-ray diffraction (SAXS) measurements were performed at BESSY PCB

beamline, using a wavelength of 1.5498 Å (8 keV), a 2D detector (981 x 1043 pixel) with a

sample to detector distance of 1920 mm. Samples were mounted on a moveable sample

holder, ensuring rotation of samples between 90° and 20° with respect to incident beam.

Investigated oxide and carbonate films were coated on thin silicon wafers (50 µm) and

analyzed in transmission mode at angles of β=90° or β=20° with respect to substrate

normal.[122]

IrO2 and IrO2/TiO2 2D SAXS patterns (small angle X-ray scattering) were recorded on a

Bruker Nanostar (three-pinhole collimation system) using a copper anode as X-ray source

(CuKα radiation), a 2D detector (2-D HI STAR 1024 x 1024 pixels) with a sample to detector

distance of 670 mm and employing a sample holder that enabled rotation of the sample

between 90° and 20° relative to incident beam. Respective oxide films were coated on thin

titanium foil and analyzed in transmission mode with a beam incident angle of β = 90° and

β = 20° with respect to the film surface.

3.2.6 X-ray photoelectron spectroscopy (XPS)

XPS measurements have been carried out with an ESCALab 200X photoelectron

spectrometer (VG Scientific, U.K.) for all samples. XPS were recorded at an angle of

emission of 0° using non-monochromatized Al Kα excitation. The Binding energy (BE) scale

was calibrated using the C1s component of aliphatic hydrocarbon at BE = 285.0 eV.[122]

Observed signals were finally associated with specific binding energies reported in literature

and listed in NIST Chemistry WebBook, NIST Standard Reference Database

3.2.7 Electrical conductivity

Electrical conductivity was measured for catalytic coatings on microscope slides with a

Keithley Model 6517B Electrometer employing a 8x8 pin probe head with an altering polarity

sequence of the pins. Light induced enhancement of conductivity observed for

semiconducting materials (e.g. TiO2) was avoided by measuring under dark environment.

Applied potential was varied between 1 V - 10 V and was increased by a factor of 1.2 for

each step. The recorded current response was plotted as a function of potential. The

observed slope corresponds to the electrical sheet conductivity and is expressed as

(Ohm/sq)-1. The volume conductivity of each catalytic layer can be derived by dividing sheet

conductivity by layer thickness (expressed as (Ohm·m)-1 or S·m-1).

3.2.8 Inductively coupled plasma - optical emission spectrometry (ICP-OES)

The iridium loading on titanium substrates (diameters: cylinder 5 mm; sheets 6 mm) was

determined by first suspending the sample in a mixture of 6 mL HCl (37 %, Roth), 2 mL

HNO3 (69 % Roth), 2 mL H2SO4 (96 %, Roth) and 40 mg NaClO3 (99.0 %, Alfa Aesar). The

obtained suspension was slowly stirred (130 rpm) for at least 3 hours under room

temperature and subsequently transferred into a microwave (Discover, SP-D, CEM

Corporation) for acid digestion at 180 °C (ramp 15 K/min, 40 min, 18 bar). 2 mL water (milliQ)

was added to the solution after completion of a cooling down period. The as prepared

solution was analyzed in a 715-ES-inductively coupled plasma system (ICP-OES, CCD

detector, Varian) in order to determine the iridium content.

3.2.9 Differential electrochemical mass spectrometry (DEMS)

The amount of formed gas products was assessed in separate DEMS experiments utilizing

mesoporous templated IrO2 coated onto titanium cylinders. The DEMS apparatus is partly

based on a design reported elsewhere[124] and consisted of an electrochemical flow cell

(0.5 M H2SO4, flow rate: 5 µL/s) connected via separation PTFE membrane to a PrismaTM

quadrupole mass spectrometer (QMS 200, Pfeiffer-Vacuum) equipped with two

turbomolecular pumps HiPace 80 operating the MS chamber at 10-6 mbar. The MS was

calibrated with a reference gas CO2 in N2/O2 (Linde, 75.04 % N2; 19.95 % O2; 5.01 % CO2).

Cyclovoltammetric measurements were conducted in the OER regime by cycling the

potential between 1.20 - 1.65 VRHE (6 mV/s) while recording the ion current for O2 (mass 32)

and CO2 (mass 44). The amount of evolved O2 and CO2 was calculated from the ion

currents.[47]

3.3 Electrocatalytic testing

The following chapter contains detailed information about the electrochemical setup (chapter

3.3.1) and measuring procedures to determine the OER activity in alkaline (chapter 3.3.2)

and acidic media (chapter 3.3.3), respectively. Moreover, the electrochemical active surface

area (ECSA) was assessed for samples tested under acidic conditions (chapter 3.3.3).

3.3.1 Electrochemical setup

Electrocatalytic testing was performed in a three electrode disc setup using a Biologic SP

200 potentiostat, a reversible hydrogen electrode (Gaskatel, HydroFlex®) as a reference and

a Pt-gauze (Chempur, 1024 mesh/cm², 0.06 mm wire diameter, 99.9 %) as a counter

electrode. All potentials in this work are referred to the reversible hydrogen electrode and are

iR corrected. The utilized electrochemical setup is schematically shown in Figure 13.

Figure 13: Electrochemical setup used for the determination of the oxygen evolution reaction (OER)

activity and electrochemical accessible surface area (ECSA). Catalytic layers of NiO, IrO2 and IrO2/TiO2

were deposited on titanium and served as the working electrode. Platinum gauze was used as a counter

electrode. The applied potential is referred to a reversible hydrogen electrode and corrected for iR drop.

Electrochemical testings were either performed with supporting electrolytes of KOH (alkaline, NiO) or

H2SO4 (acidic, IrO2, IrO2/TiO2).

All films coated on titanium cylinders or sheets (5 mm in diameter) were mounted on a

rotating disk shaft and served as a working electrode (n = 1600 rpm) using H2SO4 (Fixanal,

Reference

electrode

(RHE)

Luggin

capillary

Electrolyte

Counter

electrode

(Pt)

Working

electrode:

1600 rpm

Geometric surfacearea

duringelectrocatalytic

testing: 0.196 cm²

SP 200 potentiostat

3 Electrode set-up with

rotating disc electrode (RDE)

Fluka Analytical) or KOH (Sigma Aldrich, pellets, ≥85 %) as supporting electrolyte. The

electrolyte solution was purged with nitrogen prior to catalytic tests.

3.3.2 Alkaline procedure in KOH (NiO)

The OER activity was investigated by cyclic voltammetry in a potential window ranging

between 1.20 - 1.95 VRHE in 0.1 M KOH at a scan rate of 6 mV/s.[47, 122] Furthermore, NiO was

electrochemically activated[27] in 1.0 M KOH at a static applied potential of 1.75 VRHE for

5 hours. After every hour of chronoamperometric treatment an intermediated cyclic

voltammogram was conducted at a scan rate of 20 mV/s in order to assess increasing

oxygen evolution reaction performance as well as redox waves prior the OER onset. [47, 122]

3.3.3 Acidic procedure in H2SO4 (IrO2 and IrO2/TiO2)

The OER activity of mesoporous templated catalytic coating, i.e. IrO2 and IrO2/TiO2 was

investigated by cyclic voltammetry in a potential window ranging between 1.20 - 1.65 VRHE in

0.5 M H2SO4 at a scan rate of 6 mV/s.[47, 122] To analyze the electrochemically accessible

surface area (ECSA) the potential was sweeped between 0.40 and 1.40 VRHE at a scan rate

of 50 mV/s. Placing hydrous IrO2 in solution and alter the potential will change the valence

state of surface metal atoms. TiO2 shows no significant current response in the applied

potential range. Furthermore, a reversible proton inclusion mechanism can take place as

described by Trasatti et al.[125]:

IrOx(OH)y + δ H+ + δ e-  IrOx-δ(OH)y+δ

The ECSA of iridium oxide and mixed metal oxides of iridium and titanium was then

quantified by determining the mean value of the integrated anodic and cathodic scan of the

resulting cyclic voltammogram.[47, 55, 125-126]

4 Results and discussion

The following chapter (4) presents the synthesis approach for each mesoporous templated

system, i.e. NiO, IrO2 and mixed metal oxides of IrO2/TiO2. Subsequent to the film formation

all samples were characterized physicochemical. The resulting data were evaluated and

used for the creation of structure models containing detailed information on morphology,

crystallinity, phase composition and surface area. Moreover, the catalytic coatings were

electrochemically investigated to identify the performance in oxygen evolution reaction and

quantify the electrochemical accessible surface area. The combined physicochemical and

electrochemical data are then used to deduce structure-activity relationships. Exploiting this

knowledge allows the design of highly active OER-electrocatalysts with iridium mass based

OER-activities at least 10 times higher with respect to commercial available OER-catalysts.

Chapter 4.1 describes an approach for the synthesis of NiO by a citrate based synthesis

approach. The electrochemical OER activity is investigated for NiO layers on titanium

substrate in alkaline electrolyte solution to deduce structure activity relationships and conduct

a comparison with OER activities reported in literature.

Chapter 4.2 presents the potential of pore templating with polymer micelles for the

investigation of structure-activity relationships in gas evolution reactions and for the

optimization of the performance of IrO2-based OER catalysts. The chapter revealed that the

OER performance of mesoporous IrO2 is controlled by at least two independent factors, i.e.

the accessible surface area and the intrinsic activity per accessible site.

Chapter 4.3 shows an extended approach for the preparation of mixed metal oxide layers

with adjustable crystallinity and conductivity. The synthesis is based on micelles of triblock

copolymers and two metal oxide precursors such as Ir(OAc)3 and TiCl4. Chapter 4.4

demonstrates a synthesis approach for designing an almost throughout mixed metal oxide

comprising IrO2 and TiO2. The films were obtained by using a synthesis based on triblock

copolymers and two metal oxide precursors with similar chemical properties, i.e. Ir(OAc)3 and

TALH. The obtained layers of both chapters (4.3 and 4.4) were characterized

physicochemical and electrochemical. The obtained data were correlated to reveal structure-

activity relations indicating that the overall OER activity and intrinsic reactivity is affected by

the active surface area and electrical conductivity.

Chapter 4.5 illustrates a reproducible synthesis approach and the electrocatalytic

performance of commercial available catalyst powders.

4.1 Mesoporous templated NiO synthesized from Ni(NO3)2 and citric acid

Chapter 4.1 demonstrates the synthesis of homogeneous NiOx coatings with a templated and

locally ordered mesopore structure and their catalytic behaviour in alkaline OER.

Mesoporous NiCO3 and NiO films were obtained by deposition of Ni(NO3)2 in presence of

citric acid and the mesopore template poly(ethylene oxide)-b-poly(butadiene)-b-poly(ethylene

oxide) (PEO213-PB184-PEO213). Chapter 4.1.1 shows the thermal stability of the polymer

template and the nickel precursor investigated by thermogravimetric analysis (TGA) in order

to derive suitable parameters for calcination of deposited films. The influence of calcination

temperature on morphology, pore ordering and crystallinity of NiO and NiCO3 layers were

studied by scanning electron microscopy (SEM), transmission electron microscopy (TEM)

small angle X-ray scattering (SAXS), selected area electron diffraction (SAED) and X-ray

diffraction (XRD). The surface composition of the films was analyzed by X-ray photoelectron

spectroscopy (XPS) (chapter 4.1.2). Differently calcined catalysts were tested in alkaline

OER and studied by Kr-physisorption in order to deduce a direct relationship between OER

activity and accessible surface area (chapter 4.1.3). Electrochemical activation studies

similar to procedures reported in literature[27] resulted in further catalyst activation (chapter

4.1.4). Calcination at 350 °C produced NiO with low crystallinity yet high BET surface area

and resulted after further electrochemical activation in the highest OER activity. The most

active catalyst was further used for comparing the OER activity with reports from literature

(chapter 4.1.5).

4.1.1 Thermal behavior of pore template and nickel precursor

Mesoporous templated NiOx films were deposited onto cleaned substrates (Ti, Si) via

dipcoating performed at 25 °C and 40 % r.h. with 200 mm/min withdrawal rate. Dipcoating

solutions contained a polymer template poly(ethylene oxide)-b-poly(butadiene)-b-

poly(ethylene oxide) (PEO213-PB184-PEO213),[65, 98] 0.2 mL water (milliQ) and 2.8 mL ethanol,

to which 144.1 mg of citric acid and 436.0 mg of Ni(NO3)2·6 H2O had been added. Freshly

coated samples were dried for 10 minutes and further heat treated in air for (i) 1 hour at

250 °C and (ii) 1 hour at either 350, 400, 450 or 500 °C in a preheated muffle furnace (see

experimental part for further details).

The synthesis method relies on the thermal transformation of the deposited mesophase of

precursor complex and template polymer into the mesoporous carbonate and subsequent

conversion into nickel oxide. Thermogravimetric analysis (TGA) was therefore employed to

identify appropriate calcination conditions. Figure 14 shows the mass loss of a) template

polymer PEO-PB-PEO and b) precursor complex formed from nickel nitrate and citric acid

upon heating in air.

Figure 14: Thermogravimetric analysis (TGA) of a) template polymer PEO213-PB184-PEO213 and b) precursor

complex formed from nickel nitrate and citric acid. a) Polymer starts to decompose at ca. 250 °C and

rapidly decomposes between 350 °C and 425 °C. Nickel precursor complex shows a mass loss between

140 °C and 240 °C accompanied by release of CO2 (online-MS), indicating the formation of NiCO3. No

significant mass loss is observed in the range between 240 and 325 °C, indicating the presence of a

stable intermediate. Another rapid mass loss accompanied by the release of CO2 occurs for temperatures

above 325 °C, suggesting the transformation of NiCO3 to NiO.

Thermal degradation of the pore template during calcination is a commonly observed and

desirable feature of EISA based syntheses. The employed block copolymer PEO213-PB184-

PEO213 starts to decompose at temperatures around ~250 °C and rapidly decomposes

between 350 °C and 425 °C (Figure 14 a), indicating a better thermal stability than common

Pluronic type polymers.[127] The observed thermal stability of PEO-PB-PEO is similar to the

stability of poly(ethylene-co-butylene)-b-poly(ethylene oxide) (KLE), which is well-known for

its templating capabilities.[128]

The chemical complex consisting of Ni(NO3)2·6H2O and citric acid shows a distinctly different

thermal behaviour. A first mass loss occurs between 140 and 240 °C (Figure 14 b)

accompanied by release of CO2 as detected by online MS. In analogy to similar observation

reported for citric acid complexes of other metal nitrates [103, 119] the mass loss is interpreted

as a decomposition of the precursor complex into nickel carbonate.

The sample mass then remains constant between 240 °C and 325 °C indicating the presence

of a stable intermediate, i.e. nickel carbonate (Figure 14 b). Above 325 °C another rapid

mass loss of 25 % accompanied by CO2 release is observed. The formation of CO2 along

with the mass loss suggest in analogy to other systems [103, 119] that NiCO3 converts into

nickel oxide above 325 °C. Based on the obtained information, temperatures of (i) 250 °C

and (ii) 350 °C were selected to synthesize nickel carbonate and porous nickel oxide

samples for further analysis as presented in the next section.

100 200 300 400 500

100

a) Template

PEO-PB-PEO

mass / %

(i) 250 °C (ii) 350 °C

b) Complex

Ni(NO3)2+

citric acid

Template decomposition

Precursor

complex

NiCO

NiO

Temperature / °C

4.1.2 Physicochemical properties of coatings calcined at 250 °C and at 350 °C

Deposited Ni-based films calcined (i) for one hour at 250 °C and (ii) for one hour at 250 °C

and one more hour at 350 °C were investigated with respect to sample morphology,

crystalline phases and surface composition. Figure 15 presents results of the analysis by

scanning electron microscopy (SEM), transmission electron microscopy (TEM), small angle

X-ray scattering (SAXS) and selected area electron diffraction (SAED). Figure 16 displays

corresponding data obtained by a) X-ray diffraction analysis (XRD) and b) surface

composition from X-ray photoelectron spectroscopy (XPS).

Figure 15: Analysis of samples heat treated (i) for 1 h at 250 °C and (ii) for another hour at 350 °C by SEM

(a, f), TEM (b, g), SAXS at beam incident angles relative to the substrate of 90° (c, h) and 20° (d, i) as well

as SAED (e, j). Samples heat treated at 250 °C show complete penetration with templated mesopores

throughout the whole film volume (a, b) locally ordered parallel (c) and perpendicular (d) to the substrate

surface. SAED does not provide evidence for crystalline phase (e). Samples (ii) calcined at 350 °C

possess templated mesopores throughout the whole film volume (f, g) locally ordered parallel ((h) and

perpendicular (i) relative to the substrate orientation. SAXS obtained at 20 ° show a less defined pattern

indicating the onset of partial degradation of the ordered mesoporosity (i). SAED indicates crystallization

of the material into a NiO phase with rock salt structure (PDF: 00-047-1049).

Morphology. SEM images recorded for the cross section of sample (i) calcined at 250 °C

indicate the presence of mesopores in the materials cross section (Figure 15 a). The

abundant presence of mesopores throughout the bulk of the material is confirmed by TEM

images (Figure 15 b). SAXS recorded perpendicular to the substrate produces circular

diffraction patterns (Figure 15 c, β = 90°) indicative of local pore order with a d-spacing of

27.0 nm in the in-plane direction. Additional SAXS recorded at smaller angles relative to the

substrate (β = 20°) show an elliptically formed diffraction pattern with d-spacings

corresponding to 27.0 nm (in-plane) and 9.5 nm (out-of-plane, Figure 15 d). Such small-angle

diffraction patterns are typically observed for metal oxide films synthesized via EISA. The

elliptical rings can be attributed to a uniaxial and homogeneous shrinkage of film and their

100 nm 2 nm-1

100 nm

(111)

(200)

(220)

(311)

(222)

NiO

(PDF: 00 -047-104 9)

2 nm-1

SAXS 90°SEM TEM SAEDSAXS 20°

40 nm

(i) 250°C

(ii) 350°C

mesostructure in the direction perpendicular to the substrate during calcination,[98, 104, 129]

whereas no shrinkage occurs in the in-plane direction.

SAED patterns recorded for material (i) do not show any diffraction rings or spots, suggesting

the absence of crystalline phases (Figure 15 e). This observation is consistent with the

presence of an amorphous carbonate phase as previously observed for other metal oxides

and carbonates synthesized by a similar route.[103, 119]

Samples calcined by procedure (ii) for one hour at 250 °C and subsequently for one more

hour at 350 °C were analyzed in a similar fashion (Figure 15 f-j). SEM images recorded in

top-view mode show that the complete surface of the material is penetrated by mesopores

with a pore diameter of ca. 13.7 nm (Figure 15 f). The porosity extends throughout the

complete volume of the film (TEM, Figure 15 g). SAXS patterns recorded at β=90° still show

a circular shape with a d-spacing of 26.3 nm, which is almost identical to sample (i). Also the

SAXS patterns recorded at β=20° look similar to sample (i) with d-spacings corresponding to

26.3 nm (in-plane) and 6.8 nm (out-of-plane). The decrease from 9.5 nm (250 °C) to 6.8 nm

(350 °C) indicates further shrinkage of the film in the direction perpendicular to the substrate.

The less defined pattern of the ellipse (Figure 15 i) suggests also the onset of partial

degradation of the ordered mesopore system due to calcination at 350 °C. At the same time

SAED begins to evidence a crystalline phase (Figure 15 j). All of the observed diffraction

rings can be assigned to a Nickel(II) oxide with rock-salt structure (PDF: 00-047-1049). The

data suggest that an amorphous nickel carbonate formed during calcination (i) at 250 °C is

transformed into mesoporous nickel oxide during subsequent calcination (ii) at 350 °C.

Composition. Further evidence for this transformation can be obtained from analysis of the

bulk composition (XRD, Figure 16 a) and surface composition (XPS, Figure 16 b) of the Ni

containing films deposited on Si wafers.

Figure 16: Influence of calcination temperature on a) crystallinity and b) surface species. a) Samples

calcined at (i) 250 °C show no evidence for crystallinity, whereas a subsequent calcination (ii) at 350 °C

leads to appearance of signals at 37.5° and 43.0°, which can be attributed to (111) and (200) reflections of

crystalline NiO in rock salt structure (PDF: 00-047-1049) with crystallite sizes of ca. 5-6 nm (Debye-

Scherrer). The crystallite size increases from 6 to 11 to 17 to 27 nm (for 111 reflection) as a function of

calcination temperature. b) XPS shows a clear signal at a binding energy of 855.7 eV for samples (i)

calcined at 250 °C, which is assigned to NiCO3. Subsequent calcination at (ii) 350 °C results in a signal at

a binding energy of 855.7 eV as well as 854.4 eV. The signal at 854.4 eV is assigned to NiO, whereas the

signal at 855.7 eV is either related to surface NiCO3 or Ni(OH)2 .

X-ray diffraction confirms that calcination (i) at 250 °C does not result in crystalline phases

(Figure 16 a, i). Samples calcined at 350 °C show clear diffraction signals at positions of 2θ

equal to 37.5° (111) and 43.0° (200), which can be assigned to crystalline NiO with rock salt

structure (PDF: 00-047-1049, Figure 16 a, ii). Applying the Debye-Scherrer equation to the

(111) and (200) reflection reveals crystallite sizes of ca. 6 nm and 5 nm, respectively. The

crystallites grow further with increasing calcination temperature up to about 27 nm for

calcination at 500 °C. X-ray diffraction analysis thus supports the interpretation of initial

formation of an amorphous nickel carbonate and its subsequent transformation into

crystalline nickel oxide.

The surface composition of samples was analysed by XPS in the Ni 2p region (Figure 16 b).

Sample (i) calcined at 250 °C show one clear signal at a binding energy of 855.7 eV. Further

heating to 350 °C results in a spectrum that can be deconvoluted into two signals with

binding energies of 855.7 eV and 854.4 eV, respectively. The signal at 854.4 eV arising at

350 °C can be assigned to NiO[130] , which is consistent with results from XRD and SAED

analysis. Unfortunately, interpretation of the signal at 855.7 eV is complicated by the fact that

two different Ni species can contribute at this binding energy, i.e. NiCO3[131] and Ni(OH)2-

870 865 860 855 850

Intensity / a.u.

Binding Energy / eV

NiO

Ni(OH)2

b) XPS (Ni 2p3/2)

NiCO3

(i) 250 °C

(ii) 350 °C

30 35 40 45 50

Intensity / a.u.

2Θ / °

a) XRD

NiO (111) (200)

(i) 250 °C

(ii) 350 °C

400 °C

450 °C

500 °C

450 °C

500 °C

phase.[132] However, TGA (Figure 14) indicated the presence of a carbonate bulk phase up to

325 °C and the decomposition of the nickel carbonate accompanied by a mass loss and

release of CO2 at higher temperatures. It is therefore likely, that for sample (i) calcined at

250 °C the signal at 855.7 eV corresponds to NiCO3. For sample (ii) calcined at 350 °C, the

signal at 855.7 eV could be interpreted either as a surface carbonate or nickel hydroxide.

However, the fact that the carbonate decomposes already at 325 °C makes the assignment

to a surface hydroxide more likely. A further increase in calcination temperature to 450 and

500 °C show similar curves as for 350 °C but provide less pronounced signals at binding

energies of 855.7 eV (Ni(OH)2). The decrease in intensity is possibly related to a lower

content of Ni(OH)2 within NiO obtained for as prepared samples that were heat treated at

higher temperatures.

The combined data therefore suggest that calcination (i) of the deposited films at 250 °C

forms an amorphous Nickel carbonate phase structured by the locally ordered micelles.

Further calcination (ii) at 350 °C and above transforms this phase into crystalline nickel oxide

with templated mesopore structure. The content of Ni(OH)2 in NiO of as prepared samples

seems to lower as a function of calcination temperature. The specific surface area of the

Nickel oxide corresponds to approximately 300 m2/g as derived from Kr Physisorption

(surface area) and ICP-OES (mass).

4.1.3 OER activity vs. catalyst surface area

The OER activity of NiOx samples calcined at different temperatures was studied under

alkaline conditions and compared with the surface area of the respective catalysts as

obtained from Kr-physisorption experiments. Samples were calcined for 1 h at 250 °C

followed by 1 h at either 250, 350, 400, 450 or 500 °C to obtain NiOx with different crystallinity

and surface area. OER activity was assessed by cyclic voltammetric RDE measurements in

0.1 M KOH. Figure 17 presents a) the recorded cyclic voltammograms, b) pre-oxidation

waves prior to OER onset, c) the influence of calcination temperature on the OER current

density recorded at a potential of 1.85 VRHE and d) the influence of calcination temperature on

the surface area as derived from Kr-physisorption data in m² of film per m² of the substrates

planar dimensions.

Figure 17: Influence of calcination temperature of mesoporous templated NiO and NiCO3 on a) OER cyclic

voltammograms, b) pre-oxidation waves c) OER current density and d) BET surface area in m² of film per

m² of the substrates planar dimensions. OER activity was assessed in a RDE setup (n=1600 rpm) in

0.1 M KOH electrolyte sweeping the potential between 1.20 - 1.95 VRHE at a scan rate of 6 mV/s. a) All

catalysts are active in OER. b) Samples heat treated at 350 and 400 °C show weak pre-oxidation

characteristics prior OER onset, usually assigned to the reversible reactions of NiO to NiOOH and Ni(OH)2

to NiOOH. c) Geometric current density compared at 1.85 VRHE as a function of calcination temperature

reveals a drastic increase in OER activity when going from 250 to 300 °C, whereas higher calcination

temperature lead to continuous decrease of OER activity. This trend in OER activity closely correlates

with d) BET surface area presented in m2 film surface per m2 substrate. The correlation suggests that the

accessible surface is one of the most important activity determining parameters.

All catalysts show significant OER currents (Figure 17 a). However, the applied calcination

temperature has a severe influence on the achieved catalytic performance. Catalytic activity

drastically increases when going from 250 to 350 °C, whereas higher calcination

temperatures of 400, 450 and 500 °C lead to continuous decrease in activity (Figure 17 c).

The highest geometric current density is reached by the sample calcined at 350 °C with

4.3 mA/cm² observed at a potential of 1.85 VRHE.

Also the shape of the recorded CVs change with calcination temperature. The most active

films, i.e. samples calcined at 350 °C and at 400 °C, show a distinct oxidation wave at about

1.50 VRHE prior to the onset of the OER (Figure 17 b). This wave is not evident for samples

calcined at temperatures of 250, 450 and 500 °C. The signals can be assigned to the

following possible electrochemical reactions. [133]

1.3 1.4 1.5 1.6 1.7 1.8

0.0

0.1

0.2

0.3

0.4

0.5

j / mA cm

-2

geo

E vs. RHE / V

a) OER activity (3

CV)

1.4 1.6 1.8 2.0

j / mA cm-2

geo

E vs. RHE / V

500 °C

350 °C

450 °C

250 °C

400 °C

500 °C

350 °C 450 °C

250 °C

400 °C

b) pre-oxidation waves

250 300 350 400 450 500

j at 1.85 V vs. RHE / mA cm-2

geo

Tcalc. / °C

c) OER activity vs. T

calc.

d) BET surface area vs. T

calc.

250 300 350 400 450 500

BET surface area / m2m-2

Tcalc. / °C

pre-oxidation

wave

NiO + OH- - e-  NiOOH

Ni(OH)2 + OH- - e-  NiOOH + H2O

The impact of calcination temperature on the surface area of NiOx films is presented in Figure

17 d in terms of m2 of internal pore surface area of NiOx film per m2 of geometric surface area

of the planar substrate. Films calcined at 250 °C possess a very low surface area, suggesting

that this temperature is not sufficient to remove the template polymer. Calcination at 350 °C

produces the highest specific surface area, with a continuous decline at further increased

calcination temperatures. This trend in surface area as a function of calcination temperature

follows very closely the OER activity trend shown in Figure 17 c. It can be concluded that

within the range of studied synthesis parameters the accessible surface area has a

dominating impact on the resulting OER performance.

4.1.4 Electrochemical activation treatment

The overpotential observed in the present study in 1.0 M KOH at 1 mA/cm2 amounts to about

425 mV. This value is ca. 150 mV higher than the best reported Ni based catalysts based on

e.g. nanoparticle immobilization[26, 134] or nickel nitrate decomposition.[27] The lower

performance could be related to the fact that the synthesis in its present form produces

crystalline nickel oxide instead of the more active oxyhydroxides or hydroxides. According to

literature NiO is less active in the OER than nickel hydroxides or oxyhydroxides. However,

different reports suggest that nickel oxide can be transformed into a more active hydroxide

species by prolonged electrochemical treatments.[27, 135] Trotochaud et al.[27] performed a

galvanostatically anodic treatment of NiO at current densities of 10 mA/cm². After every hour

a CV was recorded to assess the resulting OER activity. The anodic treatment resulted in an

increase in OER activity from 0.2 mA/cm² to 1.0 mA/cm² (at 1.51 VRHE) after 6 hours of

treatment. Moreover, with extended treatments also the pre-oxidation wave at 1.40 VRHE

became more pronounced. Trotochaud[27] attributed the pre-oxidation/reduction wave to a

redox process of Ni(OH)2/NiOOH. The increased OER activity was rationalized by the better

ability of Ni(OH)2/NiOOH to intercalate water and anions. Furthermore, Klaus et al.[29]

reported in-situ Raman studies on electrodeposited Ni(OH)2. They observed a clear Raman

signal shift in the same potential region as the pre-oxidation waves reported by Trotochaud et

al.[27]. The Raman signals were assigned to Ni(OH)2 at lower potentials and NiOOH at higher

potentials, providing further evidence for a potential dependent transformation of Ni(OH)2 to

NiOOH.

We performed for the mesoporous NiO catalyst calcined at 350 °C a similar electrochemical

treatment as reported by Trotochaud et al. in order to increase the OER activity. The

treatment was performed with mesoporous templated NiO coated on titanium substrates with

5 mm in diameter and consisted of repeated potentiostatic treatments in 1.0 M KOH

(n=1600 rpm) at a potential of 1.75 VRHE for 5 hours. After every hour of potentiostatic

treatment an intermediate CV was recorded in the OER range by sweeping the potential with

20 mV/s from 1.75 VRHE to 1.20 VRHE and back to 1.75 VRHE in order to assess OER

performance as well as pre-oxidation and corresponding reduction waves. Alternating

potentiostatic treatments and CV measurement were repeated 5 times to follow the catalyst

activation. In order to assess also the subsequent OER activity 20 consecutive CVs in the

OER region were performed after 10 hours of potentiostatic treatment at 1.75 VRHE. Figure 18

displays a, b) the CVs recorded initially as well as after 1, 3 and 5 h of potentiostatic

treatment, c) the OER-related current density observed at 1.70 VRHE during the CV sweeps

and d) the 2nd, 5th, 10th and 20th CV after 10 hours of potentiostatic treatment.

Figure 18: Influence of potentiostatic activation treatment at 1.75 VRHE for NiO calcined at 350 °C on a,b)

CV shape and c) OER activity. Subsequent OER performance of modified NiO is investigated with d) 20

consecutive CVs after 10 hours of potentiostatic treatment. a, b) CVs in the OER range recorded with a

sweep rate of 20 mV/s from 1.75 VRHE to 1.20 VRHE before and after several electrochemical potentiostatic

treatments at a potential of 1.75 VRHE for 1 hour each. With each treatment OER activity increases (a) and

the pre-oxidation wave at 1.47 VRHE as well as the corresponding reduction wave at 1.28 VRHE become

more pronounced (b). c) Quantification of the activation in terms of observed geometric current density at

1.7 VRHE after repeated activation treatments. d) OER performance is further enhanced also by conducting

another 20 consecutive CVs after 10 hours of potentiostatic activation treatment.

1.2 1.4 1.6 1.8 2.0

j / mA cm

-2

geo

E vs. RHE / V

initial

after 1 h

after 3 h

after 5 h

iR corrected,

1.0 M KOH

a) CVs recorded after different

duration of activation

c) geometric OER current densities

compared at 1.7 V vs. RHE

0 1 2 3 4 5 6

j at 1.7 V vs. RHE

(anodic) / mA cm-2

geo

Duration of potentiostatic treatment / h

d) subsequent CVs in

the OER range

1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0

j / mA cm

-2

geo

E vs. RHE / V

iR corrected,

1.0 M KOH

CV2

CV5

CV20

CV10

b) magnification of CVs

prior OER onset

1.2 1.3 1.4 1.5 1.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

j / mA cm

-2

geo

E vs. RHE / V

initial

after 1 h

after 3 h

after 5 h

The potentiostatic treatment resulted in two major changes of the catalyst’s electrochemical

behaviour. With each hour of treatment the pre-oxidation wave observed at about 1.47 VRHE

and the corresponding reduction wave at 1.28 VRHE became more pronounced (Figure 18 b),

indicating an improved redox ability. Moreover, also the current density observed in each

subsequent CV became higher, indicating an increased OER performance (Figure 18 a).

Figure 18 c shows a quantitative evaluation of the increased activity by showing CVs

conducted between 1 and 5 h of potentiostatic treatment. The OER-related current density

obtained at a potential of 1.70 VRHE steadily increases from initially 2.7 mA/cm² to about

4.0 mA/cm² after 5 h of potentiostatic treatment. Figure 18 d depicts the OER activity during

subsequent repeated CVs. The measured OER activity at 1.70 VRHE further increases up to

5.1 mA/cm² accompanied by a small increase of the pre-oxidation wave, indicating the

opportunity for a further enhancement of OER activity. Both observed effects, the increase in

pre-oxidation wave currents as well as the increased OER activity, are in full agreement with

the report of Trotochaud et al.[27] It can be concluded, that the activity of the synthesized

mesoporous nickel oxide can be further increased by extended anodic treatments that

convert parts of the near-surface region of the catalyst into Ni(OH)2 or NiOOH

4.1.5 OER activity comparison with reported catalysts

The overpotential is often used as a measure of catalytic activity and was quantified for the

most active catalysts at 1 mA/cm² planar geometrical current density. The derived

overpotential is then related to values obtained in previous studies reported in literature.

Yu et al.[23] prepared a nickel-based thin film (NiOx) on multi walled carbon nanotubes