towards on-site production of hydrogen peroxide with gold
TRANSCRIPT
Towards On-Site Production of Hydrogen Peroxide with Gold-
Palladium catalysts in Electrocatalysis and Heterogeneous Catalysis
Dissertation
zur
Erlangung des Grades
Doktor-Ingenieur
der
Fakultät für Maschinenbau
der Ruhr-Universität Bochum
von
Enrico Pizzutilo
aus Zevio (VR), Italien
Bochum 2017
ii |
| iii
Towards On-Site Production of Hydrogen Peroxide with Gold-Palladium catalysts in
Electrocatalysis and Heterogeneous Catalysis
Dissertation
zur
Erlangung des Grades
Doktor-Ingenieur
der
Fakultät für Maschinenbau
der Ruhr-Universität Bochum
von
Enrico Pizzutilo
aus Zevio (VR), Italien
Bochum 2017
iv |
Dissertation eingereicht am : 10/08/2017
Tag der mündlichen Prüfung : 25/09/2017
Erster Referent : Prof. Dr. Gerhard Dehm
Zweiter Referent : Prof. Dr. Karl J. J. Mayrhofer
| v
ABSTRACT
Hydrogen Peroxide (H2O2) is considered among the hundred most important chemicals and
its industrial and domestic utilization is increasing constantly, with a market growing by
more than 5% a year to $ 6.4 billion by 2024. Despite being considered the greenest
chemical oxidant, with water being the only by-product, the current production method,
the anthraquinone process, can be considered less so. In a world looking for greener
solutions, two alternative synthesis routes, which potentially meet the industries green
goals as well as the consumers requirements, are currently being intensively investigated,
namely the electrocatalytic and the heterocatalytic synthesis. The first is based on the use
of electrochemical systems, like fuel-cells and electrolyzers, in which the H2O2 is produced
from the reduction of oxygen at the cathode. Electrochemical processes are receiving
increasing attention from both the scientific and industrial communities and are regarded
as promising for the future energy conversion. In particular, the so-called “Power-to-X”
field, to which the electrosynthesis of H2O2 belongs, is considered one of the first markets,
in which electrochemical reactors can be employed in large-scale. The second is based on
the heterogeneous process converting gaseous hydrogen and oxygen to H2O2 without energy
conversion. Both these alternative synthesis methods are considered atom efficient and
green. Furthermore, flexible system design could allow easy and scalable on-site
production of H2O2 meeting the end-user requirements.
Still, despite high expectations and interests, both technologies are under development and
not yet commercially available on a large scale, with most of the current efforts focusing on
research & development. In particular, the understanding of the role of active sites and of
the reaction mechanisms is fundamental for the design of selective and stable catalysts.
One of the most interesting materials, proposed almost 10 years ago, are Au-Pd bimetallic
catalysts. A key parameter for real applications is their durability, which is especially
critical under harsh electrochemical conditions; their stability will be therefore abundantly
addressed in this study. Moreover, while in the past years most of the advances were
obtained separately in electrocatalysis and heterogeneous catalysis, this thesis work aims
at the parallel study of H2O2 synthesis in both fields to define the synergies and contrasts
using the same catalyst materials. The results collected will culminate in a new catalytic
reaction mechanism, based on electron transfer during the heterocatalytic reaction. As the
proposed mechanism implies that electrochemical methods can be used to forecast the
catalytic behaviour, this PhD study concludes with a proposed electrochemical analysis of
the heterocatalytic reaction.
vi |
| vii
Contents ABSTRACT ............................................................................................................................. v
GLOSSARY ............................................................................................................................ x
- Introduction and State of the Art ......................................................................... 1 Chapter 1
1.1 H2O2: An Important Green Chemical Oxidant ............................................................................ 2
1.2 H2O2 Synthesis: Anthraquinone Process ................................................................................... 3
1.3 H2O2 Synthesis: Alternatives to the Anthraquinone Process ...................................................... 4
1.4 H2O2 Electrocatalytic Synthesis................................................................................................. 5
1.4.1 Fuel-cells and Electrolyzers Configurations ................................................................................................ 5
1.4.2 Oxygen Reduction Reaction (ORR) ............................................................................................................. 7
1.4.3 Au-Pd Catalyst ............................................................................................................................................ 8
1.4.4 Catalyst Stability ....................................................................................................................................... 11
1.5 H2O2 Heterocatalytic Synthesis ............................................................................................... 15
1.5.1 Catalytic Reactors ..................................................................................................................................... 15
1.5.2 Direct Synthesis ........................................................................................................................................ 16
1.5.3 Au-Pd Catalyst .......................................................................................................................................... 17
1.6 Electrocatalysis vs. Heterogeneous Catalysis: pro&contra ....................................................... 19
- Thesis Aims ....................................................................................................... 20 Chapter 2
- Experimentals ................................................................................................... 22 Chapter 3
3.1 Catalyst Synthesis .................................................................................................................. 23
3.2 Material Characterization ...................................................................................................... 24
3.2.1 Scanning Transmission Electron Microscopy (STEM) ............................................................................... 24
3.2.2 Elemental Analysis (ICPMS, XPS and EDS) ................................................................................................ 29
3.3 Electrocatalytic Measurements .............................................................................................. 32
3.3.1 Three-Electrode Electrochemical Cell and H2O2 Synthesis ....................................................................... 33
3.3.2 Scanning Flow Cell (SFC) ........................................................................................................................... 39
3.3.3 Floating Cell .............................................................................................................................................. 40
3.4 Heterocatalytic Measurement ............................................................................................... 43
3.4.1 H2O2 Catalytic Direct Synthesis and Degradation ..................................................................................... 43
3.4.2 Gas Chromatography for the Analysis of Exhaust Gas ............................................................................. 44
- Material Synthesis and Characterization ........................................................... 46 Chapter 4
4.1 Unsupported Au-Pd Catalyst .................................................................................................. 47
viii |
4.1.1 Particle Size - STEM .................................................................................................................................. 47
4.1.2 Composition – XPS and ICPMS ................................................................................................................. 48
4.1.3 Cyclic Voltammetry .................................................................................................................................. 49
4.2 Supported Au-Pd/C Catalyst .................................................................................................. 54
4.2.1 Particle size and Composition – STEM-EDS .............................................................................................. 54
4.2.2 Cyclic Voltammetry .................................................................................................................................. 56
- Palladium Electrodissolution from Model Surfaces and Nanoparticles ................57 Chapter 5
5.1 Poly-Pd Oxidation and Reduction in Different Acidic Media ................................................... 58
5.2 Poly-Pd Electrodissolution in Different Acidic Media: Influence of UPL ................................... 59
5.3 Poly-Pd Electrodissolution in Different Acidic Media: Slower Scan Rate .................................. 63
5.4 Comparison of Poly-Pd and Pd/C Electrodissolution ............................................................... 66
5.5 Discussion on Pd Oxidation/Dissolution ................................................................................. 69
5.6 Proposed Pd Dissolution Mechanism ..................................................................................... 74
5.7 Conclusion ............................................................................................................................ 76
- Addressing Stability of Bimetallic Electrocatalysts: the Case of Au-Pd AlloysChapter 6
...........................................................................................................................................77
6.1 Au and Pd Dissolution Onset Potentials ................................................................................. 78
6.2 Influence of Upper Potential Limit ......................................................................................... 81
6.3 Influence of Electrolytes ........................................................................................................ 82
6.4 Au-skin Formation Following Dealloying ................................................................................ 85
6.5 Influence of Gases ................................................................................................................. 86
6.6 Conclusion ............................................................................................................................ 87
- Electrocatalytic Peroxide Synthesis on Au-Pd Nanoparticles ..............................89 Chapter 7
7.1 Oxygen Reduction Reaction (ORR) ......................................................................................... 90
7.2 Composition/Ir,max/Selectivitymax Relationship ........................................................................ 94
7.3 Peroxide Reduction Reaction (PRR) ....................................................................................... 95
7.4 Potentiostatic H2O2 Production ............................................................................................. 97
7.5 Conclusion ............................................................................................................................ 99
- Au-Pd Bimetallic Catalyst Stability: Consequences for Peroxide Selectivity ........ 100 Chapter 8
8.1 Au and Pd Dissolution under ADPs ....................................................................................... 101
| ix
8.2 Evolution of Surface Composition: Cyclic Voltammetry in Ar ................................................. 102
8.3 Evolution of Catalyst Microscopic Structure: IL-STEM ........................................................... 103
8.4 Evolution of Composition: STEM-EDS and ICPMS .................................................................. 107
8.5 Evolution of H2O2 Selectivity: Cyclic Voltammetry in O2 ........................................................ 109
8.6 Composition/Ir,max/Selectivitymax Relationship after Degradation .......................................... 112
8.7 Conclusion .......................................................................................................................... 114
- On-demand H2O2 Production: a Parallel Study of Electro- and Heterogeneous Chapter 9
Catalysis ........................................................................................................................... 115
9.1 Common Goals of Electrocatalysis and Heterogeneous Catalysis ........................................... 116
9.2 Electrocatalytic vs. Heterocatalytic synthesis of H2O2: Related Properties .............................. 116
9.3 Electrocatalytic vs. Heterocatalytic synthesis of H2O2: Synergies and Differences ................... 117
9.3.1 Conversion vs. ORR Activity .................................................................................................................... 118
9.3.2 Catalytic vs. Electrocatalytic H2O2 Selectivity ......................................................................................... 119
9.3.3 H2O2 Degradation vs. PRR Activity .......................................................................................................... 120
9.3.4 H2O2 Productivity vs. H2O2 Current ......................................................................................................... 121
9.4 Discussion ........................................................................................................................... 121
9.4.1 Heterogeneous Catalysis of Electron-Transfer Reactions in Solution .................................................... 122
9.4.2 Electron Transfer in the H2O2 Catalytic Direct Synthesis? ...................................................................... 122
9.4.3 Floating Cell Study of the Coupled ORR/HOR Electrochemical Reactions .............................................. 124
9.4.4 Two Half Reactions in Catalysis and Electrocatalysis .............................................................................. 127
9.5 Conclusion .......................................................................................................................... 129
- Final Conclusions and Outlook ....................................................................... 130 Chapter 10
References ........................................................................................................................ 133
Articles and Conferences ................................................................................................... 151
Curriculum Vitae ............................................................................................................... 153
x |
GLOSSARY
Abbreviation Meaning
A Anode
Ametal Real Surface Area of the Metal
Ageo Geometrical Surface Area
ADP Accelerated Degradation Protocol
Au Gold
Au-O Gold Oxide
BF Bright Field
C Cathode
CE Counter Electrode
CFDE Channel Flow Double Electrode
CV Cyclic Voltammogram
D Diffusion Coefficient
DF Dark Field
DFT Density Functional Theory
E Potential
E° Standard Potential
E1/2 Half Wave Potential
Ecat Potential of Catalytic Reaction
Ed Disc Potential
Er Ring Potential
Emix Mixed Potential
Eonset Onset Potential
ECSA Electrocatalytic Surface Area
EDS Energy Dispersive X-Ray Spectroscopy
F Faraday Constant
GC Glassy Carbon
HAADF High Angle Annular Dark Field
HClO4 Perchloric Acid
H2SO4 Sulfuric Acid
| xi
HER Hydrogen Evolution Reaction
HOR Hydrogen Oxidation Reaction
Habs Absorbed Hydrogen
HUPD Hydrogen Under Potential Deposition
Hg Mercury
HR High Resolution
ICPMS Inductively Coupled Plasma Mass Spectrometer
I Current
Id Disc Current
IL Limiting Current
Ir Ring Current
Imix Mixed Current
Iper Peroxide Current
IL Identical Location
LPL Lower Potential Limit
MA Mass Activity
MEA Membrane Electrode Assembly
MeOH Methanol
N RRDE Collection Efficiency
n Number of Electrons
OCP Open Circuit Potential
ODH Oxidative Dehydrogenation
OER Oxygen Evolution Reaction
ORR Oxygen Reduction Reaction
OLEMS On-line Electrochemical Mass Spectrometer
PEMFC Proton Exchange Membrane Fuel Cell
Pd Palladium
Pd-O Palladium Oxide
Pt Platinum
PROR Peroxide Reduction and Oxidation Reaction
PRR Peroxide Reduction Reaction
POR Peroxide Oxidation Reaction
Poly- Polycrystalline
xii |
ppq Part Per Quadrillion
PVA Polyvinyl Alcohol
Q Charge
RE Reference Electrode
RHE Reversible Hydrogen Electrode
RDE Rotating Disc Electrode
RRDE Rotating Ring-Disc Electrode
S Area
SA Specific Activity
SEM Scanning Electron Microscopy
SFC Scanning Flow Cell
SHE Standard Hydrogen Electrode
STEM Scanning Transmission Electron Microscopy
SH2O2 Peroxide Selectivity
t Time
TEM Transmission Electron Microscopy
TCD Thermal Conductivity Detector
TF Thin-Film
UPL Upper Potential Limit
UPW Ultrapure Water
V Volume
WE Working Electrode
XPS X-Ray Photoelectron Spectroscopy
Greek Letters
π Greek Pi
ν Viscosity
cat Catalytic Reaction Rate
mix Mixed Reaction Rate
per Peroxide Reaction Rate at Mixed Potential
ρ Density
ω Angular Velocity
- Introduction and State of the Art
| 1
- Introduction and State of the Art1 Chapter 1
——————————————————————————————————————————
The following chapter is dedicated to a literature review and to the introduction to the
state of the art of H2O2 synthesis. An overview on the H2O2 application and on the
traditional anthraquinone synthesis route will be briefly presented, whereas the latest
advances in the alternative electrocatalytic and heterocatalytic synthesis, focus of the
present thesis work, will be discussed in more detail.
——————————————————————————————————————————
1 Parts of this chapter have been already published in:
E. Pizzutilo*, S.J. Freakley, S. Geiger, C. Baldizzone, A. Mingers, G.J. Hutchings, K.J.J. Mayrhofer, S. Cherevko
Catal. Sci. Technol. 2017, 7, 1848-1856.
E. Pizzutilo*, S. Geiger, S.J. Freakley, A. Mingers, S. Cherevko, G.J. Hutchings, K.J.J. Mayrhofer Electrochimica
Acta 2017, 229, 467–477.
E. Pizzutilo*, O. Kasian, C.H. Choi, S. Cherevko, G.J. Hutchings, K.J.J. Mayrhofer, , S.J. Freakley Chem. Phys.
Lett.. 2017, 683, 436-442.
E. Pizzutilo*, S.J. Freakley, S. Cherevko, S. Venkatesan, G.J. Hutchings, C.H. Liebscher, G. Dehm, K.J.J.
Mayrhofer, ACS catalysis.. 2017, 7, 5699-5705.
E. Pizzutilo*, On-demand H2O2 production: a study at the border between electro and heterogeneous catalysis (in
preparation)
There are therefore numerous verbal quotes from that publication. Some of the figures present in
the publication have been re-printed or modified.
Chapter 1
2 |
1.1 H2O2: An Important Green Chemical Oxidant
Hydrogen Peroxide (H2O2) is an inorganic molecule that is listed among the 100 most
important chemical compounds [1, 2]. It was first discovered in 1818 by Louis Jacques
Thénard as the product of the reaction of barium peroxide with nitric acid [3] and since
then it has become an extremely important commodity. Nowadays, it is commonly used in
a large number of both domestic and industrial applications (Figure 1.1) [4], which rely on
its strong oxidation properties. Indeed, H2O2 is a more versatile, efficient and
environmental friendly oxidizing agent than other oxidants (such as sodium hypochlorite
and nitric acid). Moreover, thanks to its high oxidation potentials, is effective over a large
pH range (pH=0 Eo=1.763 V, pH=14 Eo=0.878 V) [3]. Thus, it can be used for
instance in liquid-phase reactions to oxidize (H2O2+MMO+H2O) both organic and
inorganic substrates.
Figure 1.1 Representative schematic of the industrial and domestic applications of H2O2.
Annually more than 4.3 million tons are produced (data from 2015 [5]) and approximately
50 % of the global production is destined to pulp and paper industry as an alternative to
chlorine based oxidants [6, 7]. The remaining 50% is mainly used for textile bleaching
(∿10%), water treatment (∿ 10%) [8], synthesis of fine chemicals and of polyurethane (∿
20%) [9-11] and other smaller markets [3]. In a domestic environment, H2O2 is used for
bleaching (i.e. of textiles or air dyes) and at low concentrations (<5 wt%) for wound
disinfection and water purification. The growing demand of H2O2, in particular in certain
- Introduction and State of the Art
| 3
markets (as pulp and paper industry, chemical industry and domestic applications), is
boosting the global production. Indeed, the hydrogen peroxide market size is projected to
have a yearly growth of more than 5% till 2024, reaching approximately $6.4 billion from
the $3.9 billion registered in 2015 [12].
1.2 H2O2 Synthesis: Anthraquinone Process
Over 95% of H2O2 is manufactured by the so-called indirect anthraquinone process, also
known as auto-oxidation process [13]. This process is based on the observation of Manchot
who showed in 1901 that in alkaline conditions hydrobenzenes and hydroquinone undergo
auto-oxidation, producing peroxides [14]. The anthraquinone process, as it is known today,
was developed in 1939 by Ridel and Pfleiderer for the I.G. Farbenindustrie [15]. The
indirect synthesis of H2O2 consists of sequential hydrogenation and oxidation of alkyl
anthraquinones (Figure 1.2), which are aromatic organic compounds with formula C14H8O2.
The hydrogenation occurs on a substituted anthraquinone using a palladium or a nickel
catalyst, forming the diol. The latter is oxidized back to the original anthraquinone by O2,
leaving as a by-product H2O2 [13, 16].
Figure 1.2 Schematic of the anthraquinone process.
With this process, now highly energy efficient (high yield per cycle even at mild
temperatures of 30-60 °C) thanks to continuing optimization over the years, the produced
H2O2 can reach concentrations around 70 wt% (depending on the hydrogenation catalyst
and solvent used). Furthermore, the industrial production of peroxide can be achieved on a
very large scale with centralized plants capable of producing yearly up to 120,000 tons of
H2O2 [13]. In terms of safety, this process reduces the possibility of working in explosive
region as O2 and H2 are introduced in the reactor during separate steps, without direct
contact.
Chapter 1
4 |
1.3 H2O2 Synthesis: Alternatives to the Anthraquinone
Process
Despite the high efficiency and industrial capability, this traditional synthesis route is
affected by a series of drawbacks, which are difficult to overcome. i) First of all, it can only
be economically viable on a large production scale [17], which can be problematic as the
solutions of concentrated H2O2 need to be stored and transported, which can be hazardous
[18]. To improve storage of the aqueous solution containing peroxide, commonly the
addition of stabilizers like halide or acid (which has to be removed, depending on the
application) is required [3]. ii) The reaction cycles, inevitably cause the irreversible
formation of anthraquinone derivatives that do not participate further in the H2O2
synthesis. Thus, fresh anthraquinone has to be added constantly to the reactor in order to
regenerate the solution maintaining the efficiency over time [13]. iii) During the
hydrogenation, the anthraquinone reacts with the catalyst (i.e. palladium) resulting in its
decomposition (even though losses in modern reactors are minimized). iv) Finally, the
concentrated H2O2 has to be diluted after transport as most of the typical applications
require only concentrations in the range 3-5 wt% [19].
Considering these drawbacks, the efficiency of this process is questioned and researchers
are focusing on alternative, green and efficient synthesis routes to produce this
environmental friendly oxidant in a delocalized manner [20].
Figure 1.3 Schematic representing several synthesis pathways alternative to the anthraquinone
process.
Some of the alternative synthesis methods suggested in literature, are summarized in
Figure 1.3. The H2O2 electrocatalytic synthesis and heterocatalytic direct synthesis, being
the synthesis routes studied in this thesis work, will be described in more detail in the next
section. The other methods, instead, will be only mentioned here; for further details the
reader is invited to refer to the comprehensive review on the peroxide synthesis beyond the
anthraquinone process by Campos-Martin et al. [3].
- Introduction and State of the Art
| 5
The photocatalytic synthesis over semiconductor oxides (i.e. TiO2, SnO2) relies on the
formation of reactive oxygen species like OH, O2- and H2O2 at the surface, once the oxide is
irradiated by UV [21, 22]. The wet-chemical process uses superoxide compounds (i.e. BaO2,
NaO2) that react in liquid solvent to produce H2O2 (BaO2+H2SO4H2O2+BaSO4,
2NaO2+2H2O2NaOH+H2O2+O2) [23]. The oxidation of alcohols (i.e. Isopropanol,
Methylbenzylalcohol MBA) is a liquid-phase autoxidation process of alcohols which
produces H2O2 leaving as coproduct a ketone or aldehyde [24, 25]. Other processes are for
instance the enzymatic or plasma process. In the former, living organisms form H2O2
either as byproduct of an autoxidation reaction or through 1-e- reaction of oxygen [3, 26]. In
a plasma, H2O2 can be formed from H2 and O2 by electric discharges at atmospheric
pressure [27].
These synthesis methods are not as efficient as the anthraquinone process in the large
scale and/or they have rather low productivities [3]. However, the electrocatalytic and
heterocatalytic synthesis have attracted many research and development interest thanks
to their flexibility to be applied in a small-scale widespread manner with productivities
matching the demand of typical application [13, 19, 28].
1.4 H2O2 Electrocatalytic Synthesis
Hydrogen peroxide can be electrosynthesized in alkaline, acidic or even neutral media
through a two-electron reduction of O2, as an intermediate of the full four-electron oxygen
reduction reaction (ORR). Electrochemical reactors represent an attractive alternative, in
which H2 and O2 are provided separately to the electrodes, and are set to play a key role in
reaching energy conversion combined with chemical synthesis in a decentralized manner
[29]. The different configuration of electrocatalytic reactors proposed for the synthesis of
H2O2 as well as the fundamental background on the ORR and of catalyst stability, will be
introduced in the next sections.
1.4.1 Fuel-cells and Electrolyzers Configurations
Fuel cells and electrolyzers are open electrochemical devices in which the half-reactions of
a redox couple occur separately at the electrodes. The oxidation reaction (Red1Ox1+ne-)
occurs at the anode (A) and the reduction reaction (Ox2+ne-Red2) occurs at the cathode
(C). Each half-reaction is described by a standard potential (E°), as the potential of a
reversible electrode (measured under standard conditions) referred to the standard
hydrogen electrode (SHE, 2H++2e-H2) whose E° is conventionally 0. The half-reaction
standard potentials will determine i) whether a redox reaction is spontaneous or not and ii)
the cell voltage in case of half-reactions at the electrode. In a fuel cell, the energy released
Chapter 1
6 |
by a spontaneous process can be harvested, whereas an electrolyzer requires an energy
input (i.e. from a power supply) to drive a nonspontaneous redox reaction. Considering,
that H2O2 is synthesized from the ORR (O2H2O/H2O2, E°=1.229/0.69 V), different types of
electrochemical devices can be designed, depending on the anodic oxidation reaction chosen
(see Figure 1.4).
Figure 1.4 Exemplified design of electrocatalytic systems for the H2O2 production: fuel cell in a)
and electrolyzer in b). c) SEM image of a MEA with the main components highlighted and a TEM
image of the cathode catalyst (inset).
The H2O2 production by electrolysis have been introduced firstly in 1895 by the
Consortium für elektrochemische Industrie [3] and since then several cell designs have
been proposed since then [28, 30-43]. The anodic reaction in a H2O2 electrolyzer is the
oxygen evolution reaction (OER); at the anode site O2 is produced by oxidation of H2O or
OH- in acidic or alkaline media respectively (Figure 1.4b). More recently, also fuel cell
configurations have been investigated [44-55]. In this case, the anodic reaction is the
hydrogen oxidation reaction (HOR) (Figure 1.4a). Due to its standard potential (E°= 0 V)
that is lower than the cathodic ORR standard potential, the reactions in the fuel cell
configuration are spontaneous. Thus, both electricity and chemical production can be
obtained from a single fuel cell system. On the other hand, electrolyzers are simpler (no
need of H2 supply) and electrical current being the only requirement, they can be easily
commercialized as table-top systems (i.e. in hospital and home-based applications). The O2
required at the cathode site is either provided by air or better recycled directly from the
anode [32]. In electrolyzers, however, a concurring reaction is taking place at low
potentials that should be avoided, namely the hydrogen evolution reaction (HER).
The cathode/membrane/anode structure is generally known as membrane electrode
assembly (MEA) (Figure 1.4c). MEAs are usually encased between bipolar plates,
- Introduction and State of the Art
| 7
designed to provide electrical contact, gas supply through the gas diffusion layer (GDL)
and heat distribution/cooling.
On the fundamental level, the understanding of the cathodic ORR is a key factor for
the industrial success of such systems, as discussed in the next section.
1.4.2 Oxygen Reduction Reaction (ORR)
The ORR on a catalyst surface is a fundamental process in electrocatalysis. The complete
4-electron reduction of O2 is rather sluggish, thus limiting the commercialization of fuel
cell systems on a large scale [56, 57]. The mechanism of such multi-electron reaction is still
controversially discussed, especially concerning the elementary steps and reaction
intermediates. However, it is commonly accepted that O2 can be reduced to H2O2 through a
2-electron reduction process in both acidic
O2+2H++2e-H2O2 Equation 1.1
and alkaline media
O2+2H2O+2e-HO2-+OH- Equation 1.2
Where the hydroperoxyl radical (HO2-), is the radical form of H2O2 in a basic solution. The
production in alkaline medium has been widely investigated. As electrocatalyst, metal-free
carbon is commonly used, thus reducing the costs considerably [44, 45, 50, 52, 58].
Nevertheless, the presence of hydroxyl ions can facilitate the degradation of H2O2.
Furthermore, the low membrane efficiency of an alkaline fuel cell (AFC) is also limiting the
technology. Therefore, the stability of H2O2 being of primary importance, the
electrochemical synthesis in acidic medium (i.e. in a proton exchange membrane fuel cell,
PEMFC [59]) appears to be a more promising route [60].
ORR mechanism in acidic media
Among the different mechanisms proposed [61], the scheme suggested by Wroblowa et al.
is most commonly used [62] (Figure 1.5a). Following this mechanism, H2O2 is formed as an
intermediate product of the ORR [63-65] in the so called “peroxo mechanism”.
Chapter 1
8 |
Figure 1.5 a) Oxygen reduction reaction mechanisms and b) Schematic diagram of the H2O2
electrocatalytic production following the peroxo mechanism on a nanoparticle.
Note that also other pathways may lead to a complete 4-electron ORR to H2O (Table 1.1): i)
the “dissociative mechanism”, where the O2 reduction and protonation follows the O-O
bond breaking; ii) the “associative mechanism” in which the bond breaking follows a first
reaction step; and iii) the “peroxide decomposition mechanism”, in which the adsorbed
H2O2 is dissociate in OH.
Table 1.1 Mechanisms of Electrocatalytic Synthesis (ref. [57]).
H2O2 Formation
(“peroxo mechanism”)
O2 intermediate
dissociation to H2O
(“dissociative
mechanism”)
Hydroperoxo
intermediate
dissociation path to
H2O (“associative
mechanism”)
H2O2 decomposition
O2(g)O2*
O2*+H++e-OOH*
OOH*+H++e-H2O2*
H2O2*H2O2(g)
O2(g)O2*
O2*2O*
2O*+2H++2e-2OH*
2OH*+2H++2e-2H2O*
O2(g)O2*
O2*+H++e-OOH*
OOH*OH*+O*
OH*+O*+ H++e- 2OH*
O2(g)O2*
O2*+H++e-OOH*
OOH*+H++e-H2O2*
H2O2*2OH*
1.4.3 Au-Pd Catalyst
To form H2O2 in the ORR, the initial O-O bond should not break throughout the reaction,
which otherwise would yield H2O.
The state of the art ORR electrocatalysts are Pt-group noble metals (i.e. Pt, Pd) [56, 66,
67]. Pd stands out as the metal with the smallest overpotential, i.e. highest activity, for the
ORR [65, 68-73] after Pt [69, 74]. Moreover, it costs around 50% less than Pt [70] and
binary Pd-M (M=Cu, Co, Ni, Fe) and ternary alloys showed even higher activities [72, 75-
80]. However, Pt and Pd mainly reduce O2 in a 4-electron pathway [73, 81-83]. Indeed,
H2O2 production on these metals is observed below 0.3 VRHE, in the so-called hydrogen
under potential deposition (HUPD) region [82, 84, 85], or above 0.3 VRHE in the presence of
organic or anionic impurities [86-88]. The adsorbed species modify the reaction mechanism
- Introduction and State of the Art
| 9
as they block active sites for the oxygen dissociation, and the ORR can proceed there only
through the peroxo mechanism (without O-O bond breaking). Recently, Choi et al. showed
enhanced ORR selectivity to H2O2 due to the suppression of O-O bond breaking also in (i)
carbon-coated Pt surface [89] and (ii) atomically dispersed Pt [90].
On the other hand, Au has been commonly used as electrode for various applications and
fundamental studies, thanks to its chemical inertness in the stability potential window of
water and resistance to oxide formation. The “revival” of gold has attracted the attention in
the electrocatalysis community, as it reveals interesting activities for carbon monoxide
oxidation, alcohol oxidation and ORR [91]. Au electrodes show remarkable variation in the
ORR kinetics and the mechanism varying between 2- and 4-electrons process, depending
on support, crystallographic orientation, size and pH [91-95]. However, Au electrodes are
generally affected by a low activity and an onset potential much lower than the standard
potential (E°= 0.69 V) [21, 95].
Based on density functional theory (DFT) calculation, Viswanathan et al. [21] described
the different ORR behavior of metals using the free energy of adsorbed species (i.e. OH*,
OOH*) as descriptor. Considering the peroxo mechanism (Table 1.1), in materials that bind
O2 intermediates too strongly (i.e. Pd, Pt) the H2O2 formation is limited by the OOH*
protonation and the breaking of the O-O bond is favored. On the other hand, in materials
that bind O2 intermediates weakly (i.e. Au) the limiting step is the O2 adsorption and the
formation of OOH*. Adsorption, can vary with experimental condition and this gives rise to
the variable behavior between 2- and 4- electron process characteristic of all electrodes
that interact weakly with O2 (i.e. Au, Ag, Hg) [91, 96]. The representative volcano plot is
shown in Figure 1.6. The peak of the 2-electron plot (red) coincides with the Nerstian
standard potential of the ORR [97]. On the contrary, even the optimal catalyst for the H2O
formation in a 4-electron ORR requires a certain overpotential (η) to overcome the
difference in adsorption energies between intermediated.
Figure 1.6 Vulcano plot of activity for the 2- and 4-electron ORR (from ref. [21]).
Chapter 1
10 |
An ideal electrocatalyst for the 2-electron ORR should facilitate the O2 reduction already at
the standard potential (0.69 V); thus, it should be positioned at the top of the red volcano
plot in Figure 1.6. A common strategy to modify the adsorption energies of a catalyst is
through alloying. Indeed, it is known that bimetallic catalysts often provide enhanced
activity compared to their pure counterparts. The superior activity can result from three
different effects. i) The electronic or ligand effect which causes changes in the band
structure, thus influencing the strength of the binding between the metal surface and
adsorbate molecules [98-100]. ii) The geometric effect produces surface strain as a
consequence of the atomic arrangement of surface atoms to reduce the lattice mismatch
[101, 102]. iii) An ensemble effect arises when individual or small groups (ensembles) of
different metal atoms on the surface act as preferential active sites available to adsorbates
[103, 104]. The co-presence of both metals and their surface atomic arrangement impacts
the reaction rates and kinetics [105-108] and this is particularly relevant for applications
such as CO oxidation (AuPd [109]), ethanol oxidation (AuPd [108, 110-112]), methanol
oxidation (Pt-M [113, 114]), formic acid oxidation (Pt-M [115, 116]), ORR [117-120] and
also H2O2 synthesis and reduction (AuPd [105, 121-124], PtHg [97], PdHg [125]).
Figure 1.7 Overview of different activities for the production of H2O2 from various experimental
studies (from ref. [97]).
Therefore, the alloy of materials with such different adsorption energies like Au and Pd
(right and left leg of the volcano plot), will inevitably influence ORR activity and selectivity
[105, 122, 126, 127]. Jirkovsky et al. showed that the addition of a small fraction (8%) of Pd
in an Au matrix corresponds to an increase in H2O2 selectivity [105]. This was attributed to
the alloying effect, in particular the ensemble (or geometric effect) caused by the presence
of finely dispersed Pd in Au, which influences how the oxygen molecule is adsorbed on the
catalyst surface [104]. Other studies suggest that Pd improves the oxygen adsorption,
while Au avoids the breaking of the O-O bond, resulting in an increase activity while
maintaining a high selectivity [128-130] even though other works disagree [127].
- Introduction and State of the Art
| 11
Recently, the addition of mercury to Pt or Pd led to the discovery of unprecedentedly active
(see Figure 1.7) electrocatalysts for the H2O2 synthesis [97, 125, 131]. However, their
application in real systems might be limited for safety reasons due to the toxicity of Hg.
Thus, this thesis work will focus on the study of Au-Pd based catalysts which showed the
highest electrocatalytic activity after the Hg-based catalysts.
1.4.4 Catalyst Stability
Fuel cells and electrolyzers are nowadays considered as efficient and attractive energy-
conversion devices for future emission-free mobile and stationary applications. However,
beside activity, catalyst stability is essential to meet industrial and economical
requirements. Indeed, catalysts have to maintain their performances over a long time
range under the harsh environment in electrochemical systems. Thus, the understanding
of catalyst stability is a prerequisite for any catalytic study: indeed, the catalyst activity
alone can be ambiguous or meaningless if the structure is changing under operational
condition. In recent years, the group of Prof. Mayrhofer contributed significantly to the
understanding of the degradation mechanism of noble metals like Pt. In particular, metal
dissolution (along with the eventual successive re-deposition) was demonstrated to be of
primary importance in the course of catalyst degradation [132]. The recent implementation
of an electrochemical scanning flow cell (SFC) combined with time-resolved monitoring of
the dissolved species present in the electrolyte by using an on-line inductively coupled
plasma mass spectrometer (ICPMS) provided new insights on the degradation/dissolution
of noble metals like Pt [132-141], Ir [142-146] (which are relevant for fuel cell/electrolyzer
applications) and also Au [147, 148]. Nevertheless, even if perceived to be of paramount
importance [149], a detailed study of the dissolution mechanism of Pd and Au-Pd
bimetallic catalyst has not been done with this technique yet.
Palladium
The mechanisms of Pd metal dissolution processes are still largely unknown, and
contradictory results on the amount of dissolved metal under various operation conditions
and on the exact metal dissolution onset potentials are often reported [139, 150]. The
Pourbaix diagram (Figure 1.8) suggests that Pd can be thermodynamically oxidized and
even dissolved at pH values and potentials relevant for ORR in acidic environment [151].
However, despite the similarity with Pt, Pd exhibits important differences in its
electrochemical behavior. Indeed, at high anodic potentials it is more prone to the
formation of higher oxides (i.e. PdO2), hydrous oxide growth and oxygen absorption into the
outer layers of the Pd lattice, thus resulting in a higher dissolution rate compared to Pt
[152, 153]. The nature of the oxide species formed on the Pd surface and the relation to its
dissolution are still under debate [150].
Chapter 1
12 |
Figure 1.8 Potential-pH equilibrium diagram for the system Pd-water at 25 °C (from ref. [151]).
Contradictory results are also reported on the Pd dissolution mechanism. Rand and Woods,
studying the dissolution by cyclic voltammetry and calculating the difference between the
charge associated with anodic oxidation and cathodic reduction, firstly concluded that Pd
dissolution is mainly an anodic mechanism [153], which was successively supported by
other authors [154, 155]. Vracar et al. proposed that the anodic dissolution is determined
by the transfer of a second electron to Pd(OH) species yielding PdO/Pd(OH)2 [156]. Many
authors, argue instead that the Pd electrodissolution is mainly a consequence of reduction
of Pd oxides [150], such as Pd(OH) [154, 157-159], PdO and PdO2 [149, 150, 157, 160-163],
thus resulting in a dominant cathodic process. The electrodissolution of Pd is influenced by
several factors, including: (i) nature of anions and cations [150, 164, 165], (ii) H absorption
accompanied with formation of α and β hydrides [166, 167], (iii) pH and the electrolyte
concentration [150, 168, 169], (iv) high temperature by influencing the solubility product
[153], (v) scan rate, applied potential protocol [150, 169] and surface
morphology/composition [169].While most of these studies suggest Pd dissolution under
potential cycling, only few works report time-resolved data on dissolution of Pd, which can
provide a better insight on the dissolution mechanisms by relating the dissolution rates
with the surface oxidation state. Cadle [163] and Bolzàn et al. [154] used a rotating ring
disk electrode (RRDE) to collect the dissolved Pd species (Pd2+ was suggested), thus they
were the first to study the time-resolved anodic and cathodic Pd dissolution in sulfuric
acid. Recently, Shrestha et al. [149] used a channel flow double electrode (CFDE) to study
Pd dissolution. CFDE is in principle similar to RRDE: gold collectors in a flow
configuration follow a Pd working electrode. Their system is efficacious in relating the
surface transitions with the dissolution in a time-resolved manner; however, the direct
- Introduction and State of the Art
| 13
quantitative measurement of the dissolved mass was not done. Furthermore, studies at
high potentials, where higher oxidation states might occur, are challenging since the
oxygen evolved at the working electrode causes a high reduction current at the working
electrode [149], thus masking the contribution of dissolution. The use of a quartz crystal
microbalance is also a useful approach to relate online surface processes like also Pd
dissolution in a quantitative way as shown by Grdeń et al. and Łukaszewski et al. [170,
171]. Nevertheless, a study of the dissolution also in the oxygen evolution potential region
has not been done despite its fundamental interest, as it is known that these two processes
are closely related [172]. Additionally, the vast majority of these works only deal with bulk
Pd, whereas the stability of high-surface-area catalysts used in real applications has not
been studied thoroughly so far [173, 174].
Gold
Au is often used as electrode for many reactions, thanks to its chemical inertness and
resistance to oxide formation in the stability window of H2O. In general, Au is considered
an inert electrode and dissolution is neglected [148, 175]. However, the Pourbaix diagram
(Figure 1.9) already suggests that in acidic condition, Au might not be stable at high
potentials, relevant in the range of OER.
Figure 1.9 Potential-pH equilibrium diagram for the system Au-water at 25 °C (from ref. [151]).
Recently, Cherevko et al. [147, 148, 175, 176] shed new light on the Au dissolution
mechanism studied with the SFC-ICPMS. Even though the exact oxide nature is still
unknown, Au dissolution is strictly related with the oxide formation (same onset potential
Chapter 1
14 |
∿ 1.3 VRHE). At very positive potential, low dissolution of Au3+ species is due to surface
passivation by oxide, which is then “disturbed” during place-exchange between Au and
oxygen ions. Thus, Au ions are exposed to the electrode/electrolyte interface and
dissolution is observed (depending on the competition between re-passivation, re-
deposition and ion diffusion into bulk liquid). For more details on the mechanism of Au
dissolution readers can refer to the work of Cherevko [148].
Stability of common bimetallic catalyst and Au-Pd for the H2O2
As already discussed earlier, in bimetallic nanoparticles (as Au-Pd) the surface
compositions along with the arrangement play a key role and are often crucial for high
reactivity [105-108]. However, despite the great excitement around these catalysts, it is a
great challenge to control activity over extended times only by tuning composition and
structure during synthesis. Indeed, the harsh reaction environment and the applied
operational conditions [177-179] play a key role in the stability and thus the success and
future application of bimetallic catalyst. This is especially relevant during start-stop when
the potential can reach 1.5 VRHE [180]. Under such condition, metal migration and surface
segregation [110], as well as dissolution and dealloying [66, 181, 182] can induce
alterations in the surface composition and consequently in the activity over time.
Particularly dissolution is an important factor that needs to be considered in studies of
solid-liquid interfaces, although typical rates of noble-metal dissolution are relatively low.
However, they can be relevant over long periods (years) that these catalysts are in
operation and as of a consequence in economic considerations [183]. Moreover, dealloying,
which is the faster dissolution of one of the alloyed metal in a bimetallic catalyst, can have
a severe impact on the catalyst activity and selectivity that rely on the surface composition.
Considering possible electrocatalysts for the H2O2 synthesis in an acidic environment,
Yamanaka et al. investigated transition metal-based catalysts such as Co- and Mn-based
catalysts [51, 53, 54, 184]. However, it is reasonable to assume that these transition metals
in acidic media will likely dissolve in the potential of operational interest, thus inducing
activity changes during long operation [89]. Also catalysts obtained by the addition of
mercury to Pt or Pd can be considered unstable due to the low dissolution onset potential of
Hg [97, 125, 151]. Au and Pd, in comparison, are not only safe, but being both noble metals
with high dissolution onset-potential in acidic media [85, 144, 147], are promising stable
candidates for long term applications. However, Au and Pd have different dissolution onset
potentials and dissolution rates under electrochemical environment; thus, the study of Au-
Pd surface evolution is important to define their stable operational range.
- Introduction and State of the Art
| 15
1.5 H2O2 Heterocatalytic Synthesis
Hydrogen peroxide can be synthesized through a heterocatalytic reaction starting from O2
and H2 [3, 20, 185]. This synthesis route, also known as direct synthesis, is an elegant and
efficient alternative to the classical anthraquinone method. Furthermore, the removal of
anthraquinone as intermediate in the reaction provides a much greener route to H2O2
production, as the reaction would have 100% atomic efficiency. Furthermore, thanks to the
relative small size of the reactors, these systems can be employed for a distributed
production where low peroxide concentrations are required [29].
A basic overview of the direct synthesis and the reactor configuration, as well as of the
traditional description of the reaction mechanism, will be provided in the following
paragraphs.
1.5.1 Catalytic Reactors
In heterogeneous catalysis, there are two main reactor designs, namely the batch and the
flow reactor (Figure 1.10). These systems differ on how the product are produced and
collected. The batch autoclave reactor (Figure 1.10a) is the most common method for the
catalytic synthesis of H2O2 and it consists in a closed pressurized system in which the
produced H2O2 remains in the reactor for the entire reaction time. Since the formed
peroxide is in contact with the catalyst and can further react, it is very difficult to measure
absolute synthesis rates. On the other hand, in the flow fixed bed reactor (Figure 1.10b) it
is possible to control the residence time and thus the contact of reactants with the catalyst
allowing the study of reaction rates. An alternative, recently developed type of reactor is
the so-called “catalytic membrane”, whose advantage is to avoid contact between O2 and H2
thus yielding high H2O2 concentration [186, 187].
Figure 1.10 Schematic setup of a) batch autoclave reactor and b) flow reactor.
Chapter 1
16 |
1.5.2 Direct Synthesis
One of the first patents claiming the H2O2 formation by the combination of molecular O2
and H2 was granted in 1914 to Henkel and Weber [188]. Later on in 1939 another patent
on a small scale production of H2O2 with Pd through heterogeneous catalysis [15] opened
the way to the field of “direct synthesis” in a catalytic reactor [3, 20, 185, 189]. Despite 100
years having passed, still no commercial available system is on the market today. One of
the biggest challenges is that catalysts which are active for the H2O2 synthesis, are also
generally active for the further H2O2 hydrogenation as well as for the direct combustion of
O2 and H2 [190]. These two reactions, together with the H2O2 chemical decomposition, form
H2O thus decreasing the overall efficiency of the catalytic process.
Even though the exact mechanism is not yet completely resolved and still debated [57, 191]
diverse descriptions are known, depending on the adsorption mechanism [192]: the Eley–
Rideal and the Langmuir-Hinshelwood mechanisms. The latter has been proposed for the
H2O2 direct synthesis [130, 191, 193], consisting in the sequential hydrogenation of
molecular oxygen on the catalyst surface following the dissociative adsorption of hydrogen
yielding adsorbed H atoms (Figure 1.11). Such mechanism, known also as Horiuti-Polanyi
hydrogenation mechanism [194], was elaborated almost 100 years ago and it is still widely
accepted, even though recently also non Horiuti-Polanyi hydrogenation mechanisms have
been hypothesized [195].
Figure 1.11 a) Direct synthesis mechanisms and b) Schematic diagram of the H2O2 heterocatalytic
formation on a nanoparticle.
Note that also other non-selective alternative pathways may result in the production of
H2O (Table 1.2) [130]: i) the “dissociative mechanism”, where the O-O bond breaks; ii) the
“intermediate dissociative mechanism” in which the bond breaking follows a first
hydrogenation step; and iii) the “peroxide decomposition mechanism”, in which the
adsorbed H2O2 is dissociate in OH.
- Introduction and State of the Art
| 17
Table 1.2 Mechanisms of Direct Synthesis (ref. [130]).
H2O2 Formation O2 intermediate
dissociation to H2O
Hydroperoxo
intermediate
dissociation path to
H2O
H2O2
decomposition
H2(g)+O2(g)2H*+O2*
2H*+O2*H*+OOH*
H*+OOH*H2O2*
H2O2*H2O2(g)
H2(g)+O2(g)2H*+O2*
2H*+O2*2H*+2O*
2H*+2O*OH*+H*+O*
OH*+H*+O*H2O*+O*
H2(g)+O2(g)2H*+O2*
2H*+O2*H*+OOH*
H*+OOH*OH*+H*+O*
OH*+H*+O* H2O*+O*
H2(g)+O2(g)2H*+O2*
2H*+O2*H*+OOH*
H*+OOH*H2O2*
H2O2*2OH*
Considering all these side reactions, it is clear that H2O2 is also an unstable compound,
therefore the experiments must be designed carefully to overcome the unselective route to
H2O. For instance, one can decrease the reaction temperature or add chemicals that
stabilize the peroxide preventing its decomposition [13].
1.5.3 Au-Pd Catalyst
For more than 90 years, the state of the art direct synthesis catalysts are based on Pd.
Early studies used gas compositions in the reactors that were in the explosive region [5-
95%] at elevated pressure reaching high production rates and concentrations of H2O2 (35
wt%). However this of course is not a viable commercial process in terms of safety [196]. In
1961 a study by Pospelova showed that the addition of cyanides in acidic solution (HCl,
HNO3) improves the H2O2 productivity with Pd catalysts by suppressing the H2O2
degradation [197-199]. Choudhary et al. suggested that the anions in the solution block
active sites; this could lead either to increased selectivity or to a decreased decomposition
reaction [200]. For instance, in an acid solution containing halides (in particular Br- and Cl-
) the hydrogenation reaction is suppressed while the selectivity increases [13, 201, 202].
These improvements were achieved for both halides in the reaction media and halides
incorporated in the catalyst during preparation. The nature of the anions strongly
influences the selectivity (decreasing Br->Cl->no halide>F-) even with the addition of a
small amount (though an optimum amount is crucial) [203]. The Pd oxidation state plays
also an important role as it was demonstrated that reduced Pd is highly active for the H2
conversion but inactive for the H2O2 production [201]. On the other hand, after treatment
with oxidizing species the reaction selectivity increases as surface PdO is thought to
decrease the decomposition rate [201], even though other reports disagree [204].
In a highly acidic solution (i.e. 0.1-1 M HCl), the metal catalyst can dissolve from its
support; Lunsford et al. showed that colloidal Pd formed by dissolution was also active as
homogeneous catalyst [205]. However, as dissolution in the reactor should be avoided,
acidic properties can be introduced differently, for instance by using acidic modified
supports. Common supports like C, SiO2 and TiO2 can be modified by group such as SO3H
to achieve higher selectivity and yields [206].
Chapter 1
18 |
From the literature discussed so far it appears clear that Pd, despite being active, requires
the use of halides and acid to achieve high selectivity. However, from the commercial point
of view, this is undesirable when highly pure H2O2 needs to be produced.
In 2002 a first study on Au/Al2O3 by the group of Prof. Hutchings showed that H2O2
production could be achieved also with an Au based catalyst, even though at lower rates
compared to Pd [207, 208]. Further studies on Au/SiO2 [209, 210] suggested that the
activity was related to the support and to the size of the nanoparticles, with larger Au
being less active.
A few years later, a break-through in the direct synthesis field was the discovery of
synergistic effect in Pd-based bimetallic catalysts [201]. In a first study, Choudary et al.
compared the performances of Pd-M (M=Ru, Rh, Pt, Au) catalyst and showed that the
presence of Pt and Au enhanced the yield of H2O2, although the enhancement with Pt was
limited [201]. The presence of Au in the catalyst matrix could simply act as stabilizer for
hydroperoxy species which are thought to be intermediate in the production [211].
Since the discovery of such a synergic effect, the Au-Pd system has been widely
characterized under different operational conditions [212] and with different synthesis
routes. In particular, the role of the support (SiO2 [213], Al2O3 [214], TiO2 [215], Fe2O3
[216] and carbon [20]) was thoroughly investigated [217, 218]. These studies showed that
carbon supported Au-Pd catalysts have the highest activity followed in order by SiO2 >
TiO2 > Al2O3 – Fe2O3. The acidity and isoelectric point of the support were shown to be
crucial: indeed, acidic supports like C and SiO2 had the highest productivity. Therefore,
one of the successful synthesis approaches to improve catalysts performances consists in
an acid pre-treatment. Edwards et al. [20] showed that the hydrogenation activity
decreased (and at the same time selectivity increased up to 95%) once carbon was pre-
treated in 2 wt% HNO3, prior to metal impregnation. Beside the support, also the catalyst
morphology, the homogeneity of the alloy, the composition and the particle size play a role
[219]. Au-Pd catalysts are now considered standard catalysts and many theoretical [128,
130, 220-222] and experimental [105, 126, 127, 219, 223-226] studies were carried out in
recent years. The catalyst performance is related to the different oxygen binding energies
of the selective Au and active Pd, resulting in increased productivity for Au-Pd bimetallic
(highest productivity between 50 and 75 at% of Pd) [123, 128]. Recently other alloys and
supports have been proposed successfully (i.e. PdSn) and might open new possibilities in
the field of direct synthesis [227].
Hutchings et al. studied these catalysts in a solvent free of halides and acids (66% MeOH
and 34% H2O at 2 °C); gases were diluted with CO2 in order to avoid explosive mixtures
but also to promote the reaction itself. It was indeed shown that CO2 in the solvent mixture
forms carbonic acid which acts as a stabilizer for H2O2 [19].
- Introduction and State of the Art
| 19
1.6 Electrocatalysis vs. Heterogeneous Catalysis:
pro&contra
Before entering in detail into the results of my thesis work, I would like to shortly
summarize the advantages and disadvantages of the eterocatalytic/heterocatalytic
synthesis route:
Table 1.3 Advantages and disadvantages of using the electrocatalysis of heterogeneous catalysis
synthesis route.
Electrocatalysis Heterogeneous Catalysis
Advantages Green
Atom efficient
On-site decentralized synthesis
Pure H2O2 in neutral media
No safety issues
Energy conversion
Green
Atom efficient
On-site decentralized synthesis
Pure H2O2 in neutral media
Costs and scalability
Disadvantages Membrane degradation
H2O2 degradation in the system
H2O2 separation in acidic and
alkaline media
Safety issues
H2O2 dissolution in the system
H2O2 separation in acidic media and/or
in MeOH
No energy conversion
Chapter 2
20 |
- Thesis Aims Chapter 2
Despite the differences described so far, nowadays scientists in both fields are making
special efforts towards the understanding of the reaction mechanisms and of the role of
catalysts active and selective sites. Moreover, catalysts used for the reactions are often the
same (i.e. in the case of Au-Pd); however, these fields are generally regarded as distinct
without much exchange of information and knowledge. Therefore, a parallel study on the
electrocatalytic and heterocatalytic synthesis of H2O2 using the same catalyst materials,
like in this thesis work, is still missing and would be of interest for scientists of both fields.
Among others the aims of this thesis are:
1. It has been shown that most of the catalysts for the synthesis of H2O2 are Pd based
catalysts. Pd, being a noble metal, should be stable in acidic condition over a wide
potential range. However, in the harsh electrochemical environment of a fuel cell
and/or an electrolyzer the cathodic potentials can well exceed 1.0 VRHE causing
substantial catalyst degradation. Compared to the Pt-based catalysts, much less
information and studies on Pd dissolution are available in literature, none of these
on its on-line dissolution. In Chapter 5, the gap in the knowledge on bulk and
nanoparticulate Pd dissolution under various electrochemical condition will be
filled;
2. Thanks to the synergic alloying effect in bimetallic catalysts, the activities are often
enhanced compared to the pure metals. However, the different dissolution onset
potentials of the alloyed metals can result in a faster dissolution of one of the two
components. In Au-Pd catalysts, even small dissolution of Pd (dealloying) can
modify dramatically the nanoparticle surface composition and thus the adsorption
properties and the catalytic activities towards H2O2. The dealloying of Au-Pd
unsupported nanoparticles under various simulated degradation conditions will be
therefore fundamentally addressed in Chapter 6.
3. Compared to the direct synthesis on Au-Pd catalysts, the numbers of papers on the
H2O2 electrosynthesis in acidic media are limited. Most of the studies focus on
carbon supported particles with low Pd-content. Instead, in this thesis work, the
ORR and H2O2 degradation behavior of a whole compositional spectrum from pure
Pd to pure Au are investigated on non-supported particles (to avoid any influence
from the support). The resulting potential-selectivitymax-H2O2activitymax will be
presented in Chapter 7.
4. Considering dealloying (which was shown to be also triggered simply by the
reaction environment and gas mixtures used) it is legitimate to raise the question
on whether such bimetallic catalysts are appropriate for applications where long-
term stability is required or not. Three scenarios can be envisaged, one in which no
- Thesis Aims
| 21
dissolution is occurring and the other two where only Pd dealloying is observed or
both Au and Pd dissolution occurs. The consequent evolution of the structure and of
the ORR response will be summarized in a potential-selectivitymax-H2O2activitymax
schema in Chapter 8.
5. The final Chapter 9 will be dedicated to the comparison of the results collected with
the same Au-Pd/C catalysts in electrocatalysis and in heterogeneous catalysis. The
goal is to define common points and differences in the two systems and to describe
how this information can be implemented in the understanding of the catalytic
reaction mechanism. A conclusive discussion will examine how electrochemical
methods could be employed to study catalytic rates of reactions involving electron
transfers.
Please note that the results outlined in the following thesis yield from an accurate
selection of different projects (within the Max Planck Society as well as with external
partner institutions) in which the author was involved during the PhD. In order to not
divert from the main scope of the thesis, i.e. the study of Au-Pd catalyst employed in the
H2O2 synthesis, scientific contributions, involving different catalytic systems or co-
authorships, have not been included in this dissertation. However, the author would like to
vividly encourage the reader to refer to these studies, which are already published or
under revision process. These works include stability and dissolution investigations of pure
Pt materials2 as well as Pt-alloy systems (PtNi3, PtCo4, PtRu5), Ir materials for oxygen
evolution reaction (OER)6, and non-noble catalysts for ORR7. Additionally, the author
contributed to a review on the experimental methodologies employed to study the
degradation in PEM fuel cells8.
2 Pizzutilo, E., et al., On the Need of Improved Accelerated Degradation Protocols (ADPs):
Examination of Platinum Dissolution and Carbon Corrosion in Half-Cell Tests. Journal of The
Electrochemical Society, 2016. 163(14): p. F1510-F1514. 3 Mezzavilla, S., et al., Structure–Activity–Stability Relationships for Space-Confined PtxNiy
Nanoparticles in the Oxygen Reduction Reaction. ACS Catalysis, 2016. 6(12): p. 8058-8068. 4 Pizzutilo, E., Knossalla, J., et al., The Space Confinement Approach Using Hollow Graphitic
Spheres to Unveil Activity and Stability of Pt-Co Nanocatalysts for PEMFC, Adv. En. Mat., 2017
(accepted). 5 Hengge, K., et al., Accelerated fuel cell tests of anodic Pt/Ru catalyst via identical location TEM:
new aspects of degradation behavior, ACS Applied Materials & Interfaces, 2017 (submitted). 6 Geiger, S., et al., The Stability-number as new metric for electrocatalyst stability benchmarking – a
case study of iridium-based oxides towards acidic water splitting, JACS, 2017 (submitted). 7 Choi, C.H., et al., Minimizing operando demetallation of Fe-NC electrocatalysts in acidic medium,
ACS Catalysis, 2016, 6(5), 3136-3146. 8 Mezzavilla, S., et al., Experimental methodologies to understand the degradation of nanostructured
electrocatalysts for PEM fuel cells: advances and opportunities. ChemElectroChem, 2016, 3(10),
1524-1536.
Chapter 3
22 |
- Experimentals9 Chapter 3
——————————————————————————————————————————
In the following chapter, all the methodologies employed throughout this study will be
introduced. The chapter is divided in four distinct sections, namely catalyst synthesis,
initial material characterization, electrocatalytic characterization and heterocatalytic
characterization. This work being done within a collaboration between the Max-Planck-
Institut für Eisenforschung (MPIE) and the Cardiff University, parts of the measurements
have been conducted at the Cardiff Catalysis Institute (CCI)10.
——————————————————————————————————————————
9 Parts of this chapter have been already published in:
E. Pizzutilo*, S.J. Freakley, S. Geiger, C. Baldizzone, A. Mingers, G.J. Hutchings, K.J.J. Mayrhofer, S. Cherevko
Catal. Sci. Technol. 2017, 7, 1848-1856.
E. Pizzutilo*, S. Geiger, S.J. Freakley, A. Mingers, S. Cherevko, G.J. Hutchings, K.J.J. Mayrhofer Electrochimica
Acta 2017, 229, 467–477.
E. Pizzutilo*, O. Kasian, C.H. Choi, S. Cherevko, G.J. Hutchings, K.J.J. Mayrhofer, , S.J. Freakley Chem. Phys.
Lett.. 2017, 683, 436-442.
E. Pizzutilo*, S.J. Freakley, S. Cherevko, S. Venkatesan, G.J. Hutchings, C.H. Liebscher, G. Dehm, K.J.J.
Mayrhofer, ACS catalysis. 2017, 7, 5699-5705.
E. Pizzutilo*, On-demand H2O2 production: a study at the border between electro and heterogeneous catalysis (in
preparation)
There are therefore numerous verbal quotes from that publication. Some of the figures present in
the publication have been re-printed or modified.
10 The Au-Pd catalysts were synthesized at the CCI by Dr. Simon J. Freakley. Initial material
characterization was done in both institutes: TPR, BET, XRD at the CCI by Dr. Simon Freakley
while STEM-EDS, XPS at the MPIE. Finally, the electrocatalytic measurement were done fully at
the MPIE while the heterocatalytic measurement at CCI.
- Experimentals
| 23
3.1 Catalyst Synthesis
A schematic representation of the Au-Pd catalysts synthesis route is presented in the
following Figure 3.1. The colloidal synthesis followed by nanoparticle immobilization onto a
support is a well-established method that has been often utilized for preparing
nanocatalysts employed in the direct synthesis of hydrogen peroxide [123].
Figure 3.1 Representative schematic of the synthesis route employed for both the colloidal
unsupported and the carbon supported Au-Pd catalysts.
Initially the Au and Pd metal precursors, respectively stock solution of PdCl2 (6 mgPd ml-1,
Sigma Aldrich, Reagent Plus® 99%) and HAuCl4 . 3H2O (12.5 mgAu ml-1 Sigma Aldrich, Au
assay ≥49.0%), are mixed together in UPW, forming an aqueous solution with 20 mg of
metal in 800 ml. The desired final theoretical Au:Pd ratio is determined by adjusting the
respective relative concentration of Au and Pd precursors used in the solution. Separately,
Chapter 3
24 |
also aqueous solutions of 0.1M NaBH4 (Sigma Aldrich) and 1 wt% PVA (Polyvinyl alcohol,
Sigma Aldrich, MW=10,000, 80% hydrolyzed) are also prepared. The 1 wt% PVA is then
added to the aqueous metal solution (PVA/(Au + Pd) (w/w)=1.2). After vigorous stirring at
room temperature, the freshly prepared 0.1M NaBH4 is also added (0.1 M, NaBH4/(Au +
Pd) (mol/mol)=5); after approximately 30 min a dark-brown (the color varies slightly with
the composition) sol is generated. So far, the described synthesis is common for both the
unsupported nanoparticles and the supported ones. The formers are obtained by
concentrating the generated sol to 0.1 mgmetal ml-1 with a rotary evaporator; the obtained
colloidal solution is stable in time and can be used to immobilize the nanoparticles onto the
desired GC electrode. To obtain the carbon supported nanoparticles, instead, the generated
sol is immobilized by adding activated carbon (Vulcan XC72R, acidified at pH 1 by sulfuric
acid) under vigorous stirring. A precise amount of carbon is added to have a metal loading
of 10 wt% on the support. After approximately 2 h, the slurry is filtered, washed
thoroughly with deionized water and a black catalyst powder is finally obtained after
drying at 120 °C for 16 h. The powder can be used directly for the heterocatalytic
measurements or can be dispersed in UPW (0.1 mgmetal ml-1) forming a black ink that can
be used to prepare the electrodes for the electrocatalytic measurements.
3.2 Material Characterization
3.2.1Scanning Transmission Electron Microscopy (STEM)
Transmission electron microscopy (TEM) is employed to image particles in the nm scale
(i.e. the nanoparticles used in this study are below 10 nm). Indeed, as when particles are
much smaller than the wavelength of light a standard optical microscope cannot be used.
The TEM image is obtained from the interaction between the sample and a beam of
electrons transmitted through the specimen. In the scanning transmission electron
microscopy (STEM) the electron beam is focused to a small spot and then scanned over the
area of interest.
The electrons that pass straight through the specimen can be collected by a detector and
the image generated, consisting in a two-dimensional projection of the scanned area, is
known as bright field (BF). Contrast is obtained when transmitted light is attenuated by
the specimen (i.e. in dense or thick areas). Diffracted electrons can also be collected by an
annular dark field detector positioned slightly off angle to the incident beam. The image
generated is known as dark field (DF) and can be used to image crystalline structures.
Finally, the diffracted electrons collected at very high angle (high angle annular dark field,
HAADF) are caused by the presence of heavy elements (as supported metal particles). The
different detection regions are schematized in the following Figure 3.2.
- Experimentals
| 25
Figure 3.2 Schematic of the detection regions of a STEM microscope.
Microscopes
All STEM measurements reported in this work have been carried out at the MPIE (Figure
3.3) by means of a JEM-2200FS (from Jeol, Japan) instrument operating at 200 kV and by
means of a FEI TITAN aberration-corrected (CEOS) electron microscope operating at 300
kV. The former is mainly operated in BF mode for particle counting and determination of
statistical size distribution, whereas the TITAN is used for high resolution images
collected by means of the HAADF detector (73-352 mrad).
Figure 3.3 The JEM-2200FS (left) and FEI TITAN (right) electron microscope in use at MPIE.
Sample preparation
To avoid the formation of large aggregates of particles that might be too thick for the
electron beam to be transmitted, the sample preparation for STEM imaging requires the
use of ultra-low amounts of material (sub µg cm-2 range). Therefore, the Au-Pd colloidal
suspension is diluted by a factor of 1:10 in ultrapure water (UPW, PureLab Plus system,
Elga, 18 MΩ·cm), whereas a small amount of Au-Pd/C catalyst powder (<100 mg) is
dispersed in 1 ml of UPW by means of an ultrasound bath. Following on, approximately 5
µl of the so prepared catalyst inks are dispersed onto a Lacey carbon film supported by gold
coated TEM grid (PLANO GmbH). After being dried in air, the grids are ready to be
inserted in the electron microscope for characterization.
Chapter 3
26 |
IL-TEM
To monitor the evolution of the catalyst structure throughout the electrochemical stability
treatment, an interesting method consists in the so-called “Identical Location” or IL-TEM
approach that was firstly developed by Mayrhofer et al. [228]. For this study, this
procedure has been employed for the understanding of the degradation mechanisms of
AuPd/C catalyst on the nanoscale (see in particular Chapter 8). Compared to standard
TEM post mortem analysis of catalysts (i.e. from RDE films after treatment) [229-231], IL-
TEM is not destructive, as the materials do not need to be mechanically scraped from the
electrode. Thus, such an approach allows the sequential characterization of the catalysts
structural evolution at various stages of the electrochemical treatment.
Catalyst degradation is often the result of a complex superposition of several competing
mechanisms and pathways; over the last years, many studies based on the IL-TEM
approach contributed to their partial elucidation. In particular, Pt and Pt-based alloys (e.g.
PtNi, PtCo) were investigated and the following degradation mechanisms have been
proposed [135, 181, 232-236]: dissolution (e.g. of Pt), Ostwald ripening, dealloying (faster
dissolution of the alloyed element), agglomeration, particle detachment and carbon
corrosion (see Figure 3.4).
Figure 3.4 Schematic representation of the degradation mechanisms.
Despite the amelioration generated by IL-TEM a true real-time separation of the processes
acting at the nanoscale is still out of reach; ideally, this could be achieved by an in-situ
approach [237, 238]. However, an actual in-situ electrochemical system is still under
development and several obstacles (e.g. radiolytic effects), are yet to be overcome [239].
- Experimentals
| 27
To perform IL-TEM a special “finder” grid (NH7 S 166A9, Plano) provided with
alphabetical and/or numerical code (Figure 3.5), that is used to track back the specific
location identified (red rectangles) during the initial characterization at the microscope.
After recording several locations (for statistical reason), the grid is transferred to the
standard three-electrode rotating disc electrode (RDE) setup and used as WE after
immobilizing it on top of a standard glassy carbon (GC) RDE tip (using a Teflon lit as in
Figure 3.5). Once that the chosen accelerated degradation protocol (ADP) has been applied,
the grid is rinsed with UPW. After drying, the sample can be reinserted in the TEM for
further analysis. This whole process can be repeated several times, however, particular
care needs to be taken while transferring the grid from the TEM to the RDE to avoid
mishandling and impurities that might introduce artifacts in the measurement, especially
for elemental analysis (EDS).
Figure 3.5 Graphical representation of the tracking process standardly used for IL-TEM by means
of letters and numbers on the finder grid. The TEM micrographs represents two different
magnification used to find the same identical catalyst area.
Methods
In this PhD thesis work, carbon supported Au-Pd/C catalysts have been investigated with
the IL approach. To the author’s knowledge, this is the first study on such catalysts, as
most of the studies in literature focus on Pt based catalysts. Nevertheless, the
understanding of structural evolution of these catalysts is important to understand their
performances under electrochemical condition.
The IL results are presented in the following Chapter 8. Three ADPs with a scan rate of 1
V s-1 are considered:
ADP-0.8 consisting in 1000 CVs in the range [0.1-0.8] VRHE with IL after 100 and
1000 cycles;
Chapter 3
28 |
ADP-1.2 consisting in 1000 CVs in the range [0.1-1.2] VRHE with IL after 100 and
1000 cycles;
ADP-1.6 consisting in 100 CVs in the range [0.1-1.6] VRHE with IL after 100 cycles;
The potential ranges are chosen in order to study the catalyst evolution under different
dissolution/dealloying regimes (see Chapter 6 and Chapter 8). Furthermore, ADP-1.6 is
chosen beyond the typical potential windows of steady-state operation in order to i)
fundamentally understand the degradation also beyond Au oxidation and ii) to figure out
limitations in operation conditions when reactors are not run in steady state conditions.
Indeed, in a fuel cell during start-up/shut-down procedures potentials at the cathode can
rise up to ca. 1.5 VRHE [240].
Statistical particle size evaluation
The surface area of supported and unsupported catalysts is estimated from TEM average
sizes following the calculation described in [95, 150]. The area of each single particle in the
TEM micrographs is determined using ImageJ; the average size is then calculated
assuming a spherical geometry (circle in the 2D image). The volume of each particle is:
V=4/3 π r3 Equation 3.1
Where r is the radius (half of the mean particle size). The surface area is:
A=4 π r2 Equation 3.2
Thus, the specific surface area is:
ECSA=3/(r*ρ) Equation 3.3
Where ρ is the crystallographic density of the alloy assumed to linearly depend on the
composition of the crystallographic densities of gold and palladium (ρAu=19.3 g cm-3 and
ρPd=12.02 g cm-3):
ρ(AuxPdy)=(x * ρAu + y * ρPd)/(x+y) Equation 3.4
The total metal surface area was calculated as follows:
Ametal=ECSA*m Equation 3.5
Where m is the mass of metal (see the specification in the single chapter. Typically: SFC -2
ngmetal and RDE 2 µgmetal)
- Experimentals
| 29
Other surface area estimation
To validate this method the surface area is also estimated from the double layer capacity
(as for AuPd unsupported nanoparticles in SFC Figure 3.6).
Figure 3.6 Recorded cyclic voltammetry for AuPd catalyst (1 layer) at 10 and 200 mV s-1. From the
slope of the current values is derived the capacity (inset).
The surface area is around 0.005 cm2 (considering a value of 44.5 μF cm-2 [241]), in the
same range of the estimated TEM value (around 0.003 cm2). The small difference could be
considered within the margin of error due to i) the ambiguity in the evaluation of the
capacity, ii) the difficulty in estimating precisely the exact loading in the SFC and iii) the
overestimation of the capacity that includes also the glassy carbon support.
3.2.2 Elemental Analysis (ICPMS, XPS and EDS)
The performances of bimetallic catalysts are not only determined by their structure and
size. Indeed, their composition and metals distribution play an important role, especially
for reactions relying on the alloying effect. Therefore, it is of pivotal importance to have at
disposal appropriate tools to investigate the composition both on the macroscale i.e. by
inductively coupled plasma mass spectrometry (ICPMS) or X-ray photoelectron
spectroscopy (XPS) and on the nanoscale i.e. by STEM-energy dispersive X-ray
spectroscopy (-EDS).
Inductively coupled plasma mass spectrometry (ICPMS)
ICPMS is a highly sensitive and precise type of mass spectrometer that can detect metals
at very low concentration (part per quadrillion, ppq) by simply analyzing the metal ions
obtained by ionizing the sample with inductively coupled plasma. Thanks to this
technique, the metal ratio of the investigated catalyst can be estimated with extreme
precision. The disadvantage of such technique is that it is destructive to the catalyst.
Thus, the Pd content in the catalyst has to be statistically determined from the post
mortem analysis of the catalyst (i.e. as-received and at different stages of the ADPs). After
Chapter 3
30 |
the desired RDE electrochemical treatment, the catalyst films are scraped mechanically
from the GC tips and dissolved in aqua regia (4 ml prepared from Merck, Suprapur acids).
After boiling for approximately one hour, the solutions are diluted (1:10) and then
measured by ICPMS.
ICPMS analyses are performed at the MPIE using an ICPMS, NexION 300X from Perkin
Elmer (Figure 3.7). The quantitative determination of 197Au and 106Pd content is obtained
by comparison to calibrated internal standard solutions of 187Re and 103Rh respectively.
Figure 3.7 The NexION 300X in use at the MPIE.
ICPMS can be operated not only in steady-state mode, but also in online mode i.e. for
detecting dissolved elements during electrochemical tests. The online operation will be
illustrated in the section related to the electrochemical measurement.
X-Ray Photoelectron Spectroscopy (XPS)
XPS is a technique based on the photoelectric effect which can give information such as
oxidation state and composition on the catalysts’ surface (sensitive depth of around 10 nm).
When atoms absorb high energy X-ray radiation, core electrons can be ejected, each with a
characteristic kinetic energy (Ek) depending on their binding energy (Eb), on the energy of
the incident X-rays and on the work function of the spectrometer (, energy required to
eject a core electron into the vacuum from the Fermi level). The binding energy for each
core electron is related to the specific element and to the oxidation state of the sample
(higher oxidation states correspond to higher binding energy). To detect an electron, the
energy of the incident X-ray photons (h) needs to be higher than both the binding energy
and the work function. The spectrometer measures the remaining kinetic energy:
Ek = h -Eb - Equation 3.6
- Experimentals
| 31
Maintaining constant the energy of the incident X-ray photons, XPS spectra can be
obtained as intensity of detected photoelectrons vs kinetic energy. A simplified energy level
diagram is shown in Figure 3.8.
Figure 3.8 Energy level diagram to illustrate the energy barriers associated with the photoelectric
effect, where Eb is the binding energy of the core electron, φ is the work function and Ek is the
residual kinetic energy associated with the electron.
XPS measurements are performed applying a monochromatic Al Kα X-ray source (1486.6
eV) operating at 15 kV and 25 W (Quantera II, Physical Electronics, Chanhassen).
Analysis of the spectra and of the Au:Pd molar ratios was carried out by Dr Olga Kesian
using Casa XPS (http://www.casaxps.com/).
Energy Dispersive X-Ray spectroscopy (EDS)
EDS spectroscopy is an analytical technique based on the emission of X-rays from the
specimen and is commonly employed for the elemental analysis. When the primary
electron beam ejects core electrons (i.e. from an inner shell), an electron from an outer shell
can fill the vacancy and a characteristic X-ray is emitted (Figure 3.9).
Figure 3.9 Energy level diagram to illustrate the emission of the characteristic X-rays and a typical
EDS spectrum of an AuPd/C catalyst.
The energy of such X-rays is related to the difference between the outer shell energy
(higher) and the inner shell energy (lower) and is characteristic of the material analyzed.
Chapter 3
32 |
EDS is typically used in electron microscopy as STEM, where the electron beam scanning
over the sample is used as excitation source for the X-rays that are detected in an energy-
dispersive spectrometer. EDS is limited mainly to heavy elements and the spectrum can be
affected by overlapping of elements. Therefore, for this study the Au:Pd molar ratio is
estimated also from the Au Mα (∿ 2.1 keV) and Pd Lα and Lβ (∿ 2.8 keV and 3 keV,
respectively) that do not overlap (see green peaks in the EDS spectrum of Figure 3.9). The
values are also compared to the ICPMS data. Apart from composition, when using an
atomic resolution TEM, it is possible to spatially resolve the elemental distribution within
the nanoparticles to determine whether the elements are homogeneously distributed or not
(i.e. in a core shell).
The chemical composition and the elemental mapping investigation of single nanoparticles
of the as-received AuPd/C catalyst as well as after ADPs are performed by EDS at the
MPIE. The Bruker Super-X windowless 4 quadrant silicon drift detector with a solid angle
> 0.7 srad fitted in the Cs probe corrected FEI Titan Themis STEM operated at 300kV was
used. The EDS measurements were performed in the STEM mode with a probe size of
about 1 Å. A probe current of 70 pA for imaging and 0.5 pA for EDS, probe semi-
convergence angle of 23.8 mrad, as well as inner- and outer semi-collection angles of the
high angle annular dark-field (HAADF) detector of 73-352 mrad were used for imaging and
STEM-EDS measurements.
3.3 Electrocatalytic Measurements
As already discussed in the introduction, H2O2 can be produced in an electrocatalytic
process from the reduction of O2. Therefore, it can be obtained as a product in the cathode
of a fuel cell or an electrolyser (depending on the anodic process). Even though in situ fuel
cell and/or electrolyser tests are the ultimate step in catalysts characterization, typically
half-cell thin-film (TF) electrochemical experiments are used to fast scan novel materials.
Indeed, the preparation of electrodes for thin-film measurements is rather simple and their
characterization and interpretation straightforward compared to complex multicomponent
membrane electrode assemblies (MEA). Therefore, ORR activity/selectivity and catalysts
stability have been carried out in half-cell experiments following the thin-film rotating disc
electrode (TF-RDE) method described by Schmidt et al. [242, 243], whereas their
dissolution has been studied with the scanning flow cell (SFC) [244, 245] and the mixed
potential with the floating cell [246]. This section contains the description of
electrochemical set-ups and experimental details for the determination of catalyst
selectivity, productivity and stability.
- Experimentals
| 33
3.3.1 Three-Electrode Electrochemical Cell and H2O2 Synthesis
Electrochemical cell- Set-up
A three compartment, three electrodes Teflon cell prepared in-house is used for Ar
background (with RDE tip), activity/selectivity and stability (with RRDE tip) experiments
(shown in Figure 3.10). All electrochemical measurements are performed at room
temperature (∿ 25 °C). The RDE tip, used in Ar background experiments, consists in an in-
house manufactured Teflon tip with a glassy carbon disc (GC, Sigradur), whose geometric
area is 0.196 cm2. The RRDE commercial tip (AFE6R1PT, Pine Research Instrumentation),
instead, consists in a glassy carbon disc (0.196 cm2) and a Pt ring (1 mm thickness)
embedded in a Peek tip. The electrodes are usually obtained by dropcasting catalyst ink
onto the GC; an homogeneous catalyst film should be obtained after drying. In case of a
powder Au-Pd/C catalyst the ink is prepared by its dispersion in UPW (18 MΩ, TOC < 3
ppb, ELGA) and sonication for approximately 15 min. The composition of the ink can be
adjusted depending on the catalyst; thus, for some suspensions also a small amount of
isopropanol is added to improve the dispersion. Electrode loading is always maintained
constant at 10 µgmetal cm-2 for all RDE/RRDE measurements.
Figure 3.10 Teflon three-compartment cell (from ref. [247]). The RDE/RRDE is housed in the main
compartment, while the CE and RE are housed in separate compartments.
The RDE/RRDE tips served as working electrodes (WE), which are then connected to their
rotator shaft (Radiometer Analytical and Pine rotator, Pine, respectively). The counter
electrode (CE) consisting in a graphite rod and the reference electrode (RE) consisting in a
saturated Ag/AgCl electrode (Metrohm) are both hosted in separate compartments of the
electrochemical cell. In order to prevent chloride contamination to the main chamber, the
RE is always separated by a Nafion membrane [247]. For the measurement of productivity
Chapter 3
34 |
(Chapter 7), also the CE is separated by Nafion to avoid H2O2 degradation at this electrode.
Gamry reference 600 potentiostats are employed (two synchronized ones in the case of the
RRDE measurements). Bipotentiostat measurements are controlled with the commercial
Gamry software, whereas all the other measurements are recorded with an in-house
developed LabVIEW software that controls also the gas system and the rotator [248]. For
all RDE/RRDE tests, 0.1 M HClO4, prepared from UPW and concentrated suprapure acid
(Merck, Suprapur), is used as electrolyte.
RRDE- ORR activity measurement
The ORR activity of Pt-based catalysts is generally normalized by the electrochemically
active surface area (ECSA) and it is obtained from the effective separation between kinetic
and diffusion currents by means of Levich-Koutecky equation. Exemplar polarization
curves for a Pt/C nanocatalyst (3 nm, Tanaka), recorded in O2 saturated electrolyte at
different rotation rates, before and after subtraction of the capacitive currents are shown
in Figure 3.11.
Figure 3.11 a) CV of Pt/C (3 nm, Tanaka) in Ar (black) and O2 saturated electrolyte at different
rotation rates (blue). b) Respective CV after subtraction of capacitive currents from the Ar
background.
Under convection regime, these curves always show a constant current between 0.2 and 0.6
VRHE. In this potential range called diffusion limited region, the current depends only on
the intrinsic O2 mass transport to the electrode. In the so-called kinetic region at higher
potential, instead, the current is only determined by the charge transfer at the catalyst
surface. In between, in the mixed kinetic region, the current rapidly changes and it is
controlled by both kinetic and mass transport limitation.
- Experimentals
| 35
The theoretical diffusion limited current IL is calculated from the Levich equation:
𝑰𝑳 = 𝟎. 𝟔𝟐 × 𝒏 × 𝑭 × 𝑨𝒈𝒆𝒐 × 𝑫𝑶𝟐
𝟐𝟑 × 𝝂−
𝟏𝟔 × 𝒄𝑶𝟐
× 𝝎𝟏𝟑
Equation 3.7
with n the number of electrons transferred, F the Faraday constant (96485 C mol-1), Ageo
the geometric surface area (0.196 cm-2), D as the diffusion coefficient of O2 in the electrolyte
(1.93 .10-5 cm2 s-1), ν as the kinematic viscosity of the electrolyte (1.26 . 10-6 mol cm-3) and ω
the angular velocity (2πf). Assuming a rotation of 900 rpm, the limiting current results
4.54 and 2.27 mA cmgeo-2 for a 4-electron and a 2-electron process. From this value,
applying the Levich-Koutecký equation [249], it is possible to extrapolate the kinetic
values and thus the specific activity (SA).
In case of bimetallic material (i.e. Au-Pd) for the ORR to H2O2, little is known about the
real active/selective sites and for some composition no diffusion limited current is reached
even at low potentials (see Figure 7.2). As the determination of SA is rather dubious in this
case, the evaluation of performances can be achieved by the direct comparison of the
polarization curves obtained in O2 saturated electrolytes. Conventionally, either the half
wave potential (E1/2) or the onset potential (Eonset) are used to compare such catalysts. In
this PhD work, the onset potentials (which vary of several hundreds of mV with
composition) are generally used as a first term of comparison between catalysts. These first
observations are then completed with analytical data obtained with the RRDE (see next
section).
Unless stated otherwise, all the activity/selectivity measurements are carried out at a scan
rate of 0.05 V s-1, rotation rate of 900 rpm and employing IR compensation by positive
feedback [250].
RRDE- selectivity
RRDE is an elegant in situ technique that makes use of a second working electrode to
study multi-electron electrochemical reaction mechanisms or to detect reaction byproducts.
In our case, RRDE is used to monitor simultaneously the H2O2 production which occurs
during the O2 reduction at the disc of the RRDE. Indeed, the H2O2 obtained at the disc is
directly detected on the Pt ring where it is oxidized (H2O2O2+2H++2e-) originating a
positive current (see Figure 3.12).
Chapter 3
36 |
Figure 3.12 Schematic of the RRDE working principle for the measurement of H2O2. After reduction
of O2 to H2O2 at the disc, the H2O2 is being collected and oxidized at the ring of the RRDE tip.
From the O2 reduction at the disc and the H2O2 oxidation at the ring, two currents are
detected: disc current (Id) and ring current (Ir) or peroxide current (Iper=Ir/N).
From these values, it is possible to derive the number of exchanged electrons (n) per
molecule of O2 and also the selectivity for the H2O2 formation (SH2O2) defined as follows:
𝒏 = 𝟒 ×𝑰𝒅
𝑰𝒓 𝑵⁄ − 𝑰𝒅
Equation 3.8
and
𝑺𝑯𝟐𝑶𝟐= 𝟐 ×
𝑰𝒓 𝑵⁄
𝑰𝒓 𝑵⁄ − 𝑰𝒅= 𝟏𝟎𝟎 ×
𝟒 − 𝒏
𝟐
Equation 3.9
where N is the collection efficiency of the RRDE. The latter is independent of the reaction
studied and is only a function of geometric parameters. However, often the theoretical
value does not coincide with the real collection efficiency, i.e. if the RRDE is not perfectly
even. Thus, the collection efficiency is determined experimentally by calibration by a
reversible and simple system as the 1-electron redox couple ferrocyanide([Fe(CN)6]4-
)/ferricyanide([Fe(CN)6]3-):
[Fe(III)(CN)6]3- + e- [Fe(II)(CN)6]4-
E° = 0.361 VSHE (25 °C, 1atm)
Equation 3.10
At negative potential sweeps, the ferricyanide is reduced to ferrocyanide at the disc and
some of the ferrocyanide reaches the ring (whose potential is above E°) where it is oxidized
back to ferricyanide. From the Id and Ir measured (Figure 3.13) the collection efficiency
(N=Ir/Id) is evaluated to be ∿0.22 and constant at different rotation rates.
- Experimentals
| 37
Figure 3.13 Sweep at different rotation rates in Ar saturated 0.1 M HClO4 + 0.01 M K3[Fe(CN6)].
Scan rate: 50 mV s-1.
During all RRDE measurements for the H2O2 detection, the ring potential (Er) is set
constant at 1.0 VAg/AgCl (approximately 1.28 VRHE) where the rate of H2O2 oxidation is
diffusion limited in a Pt electrode[57] and O2 reduction is negligible.
Figure 3.14 PROR reactions on a poly-Pt electrode in Ar saturated 0.1 M HClO4 + 10 mM H2O2.
Scan rate: 50 mV s-1.
RRDE- H2O2 productivity measurement
H2O2 electrocatalytic productivity measurements (Chapter 7) have been performed with
the unsupported Au-Pd catalyst, as potentiostatic measurement for 2 and 30 min in O2
saturated electrolyte and at a rotation rate of 900 rpm. In this case the potential is held
constant at the desired value (see details in 7.4) and the produced H2O2 is calculated with
the following two methods: i) by integrating the Iper measured and calculating the
equivalent moles with Faraday’s law; ii) by measuring the final content of H2O2 in 100 ml
electrolyte after the test. In this study, the latter is used if not else specified. The final
H2O2 concentration is measured electrochemically from the diffusion current at the Pt ring.
1.28
VRHE
Chapter 3
38 |
Four solutions with increasing H2O2 concentrations are used for initial calibration,
resulting in the following concentration/current curve (Figure 3.15).
Figure 3.15 Calibration curve for the Pt ring electrode with different concentrations of H2O2.
RRDE- Stability test
Considering that, if to be used in PEM fuel cells, these catalysts must be stable for long
time in a harsh electrochemical environment, it is necessary to investigate their durability.
Despite the substantial differences between a real fuel cell and a liquid electrochemical
cell, fast accelerated degradation protocols can be designed and used in an electrochemical
cell to predict the long-term behavior of such a catalyst. The understanding of the
degradation mechanism on the nanoscale (Figure 3.4) achieved with the IL approach
combined with the evolution of the electrochemical behavior, can provide a complete
picture of the degradation mechanism. The combination of IL approach and
electrochemical method is fully presented in Chapter 8. The chosen ADPs (ADP-0.8, ADP-
1.2 and ADP1.6) have been already introduced in the section dedicated to the IL-approach.
The UPL are chosen to study the different impacts of dissolution/dealloying on the catalyst
morphology and this will have, as expected, also a strong influence on their electrochemical
response. In order to follow the evolution of the electrochemical behavior, the disc and ring
currents (Id and Ir) are monitored at regular intervals of ADPs (i.e. after 1-10-50-100-1000
degradation CVs).
RDE- Peroxide Reduction Reaction (PRR)
As the H2O2 is being produced, at the same potential it is also likely to be reduced, as it is
also shown for Pt in Figure 3.14 (for more details the reader is also invited to refer to the
work of Katsounaros et al. [57, 251-253]). H2O2 can be “degraded” through the following
processes: i) the electrochemical peroxide reduction reaction (PRR) (H2O2+2e-+2H+2H2O),
ii) the electrochemical peroxide oxidation reaction (POR) (H2O2O2+2e-+2H+) and iii) the
chemical disproportionation reaction (2H2O22H2O+O2). This last is directly influenced by
the reaction environment as H2O2 is highly reactive, whereas the first two processes are
- Experimentals
| 39
directly related mainly to the catalyst surface status and presence of impurities that poison
active sites [57, 251-253]. Therefore, it is also necessary to study the PRR activity. In this
case, 10 mM H2O2 are added to the Ar purged electrolyte (100 ml) and the PRR curves are
recorded in the range [0.1-1.0] VRHE.
3.3.2 Scanning Flow Cell (SFC)
The SFC [245] is a three-electrode electrochemical cell in which the electrolyte is flowing in
from an external reservoir (where normally it is saturated with the gas of interest) to the
working electrode (with the 1 mm aperture of the cell positioned on top of the catalyst spot)
and out. Such a configuration is more flexible than the standard electrochemical cell, as it
allows fast electrochemical screening of several catalyst spots in a short time, which is for
example desirable when studying material libraries [254-256]. Furthermore, it can be
combined with other analytical techniques, as a mass spectrometer (i.e. ICPMS, OLEMS).
For instance, SFC combined with ICPMS (Figure 3.16) is an extremely powerful tool to
study the online metal dissolution triggered by electrochemical experiments. Pt and Au
dissolution has been thoroughly characterized with online SCF-ICPMS [132, 134, 137, 147,
148, 176, 257], whereas few data have been published on Pd dissolution [144] and no
studies have so far been done with Au-Pd bimetallic catalysts. However, there is much
interest in understanding the dissolution/dealloying behavior of such materials, under
different electrochemical conditions (electrolyte, UPL, gases, scan rates…).
Figure 3.16 Schematic representation of the SFC setup. The WE is either a polycrystalline foil or a
GC foil with deposited catalyst spots. The electrolyte carries the dissolved metals from the WE to
the ICPMS where they are detected.
Chapter 3
40 |
As WE either pure polycrystalline materials are used (i.e. poly-Pd) or nanocatalysts are
deposited on GC electrodes. The deposit consists in circular spots that are printed onto GC
plates from the prepared catalyst ink using a drop-on-demand printer (Nano-PlotterTM
2.0, GeSim). The nanoplotter prints 200 drops (each of volume 150–200 pl; a precise
estimation of the volume is done at every printing with a fast camera) in rapid succession
using a piezoelectric pipette. Several layers can be printed on the same spot (whose size is
slightly smaller than the size of the SFC aperture), however, if nothing else is specified
normally only one layer has been used for the measurements. Once the spots have dried
the GC plates can be transferred to the SFC and used as a WE. The RE and CE are an
Ag/AgCl electrode and a graphite rod, respectively. The measurements are performed at
room temperature with a flow of 193 μl min−1. The dissolution is studied under several
conditions, specified in the relative Chapters. For example, the influence of scan rates and
UPL is studied by cyclic voltammetry or linear sweeps in Ar or O2 purged 0.1 M HClO4 and
in 0.1 M H2SO4, prepared by dilution of concentrated acid (Suprapur®, Merck) in ultrapure
water (PureLab Plus system, Elga, 18 MΩ). The detection of the dissolved Au and Pd at the
ICPMS is done by comparison with an internal standard as described earlier in this
chapter.
3.3.3 Floating Cell
The floating cell has been recently developed by Polymeros et al. [246] as a technique
bridging the gap between fuel cell measurement and TF-RDE. This innovative technique
allows making half-cell studies as in RDE, but without limitation due to the diffusion of
gases in the electrolyte. Indeed, gases are provided directly to the catalyst layer and the
reactions occur at the so-called three-phase boundary (solid, liquid and gas), as in the fuel
cell. This technique is called “floating cell” as it consists in a house-manufactured Teflon
cell with gas inlet/outlet and a small aperture on the bottom that can host a TEM grid
which floats on the electrolyte (Figure 3.17). The WE is prepared by depositing a certain
amount of catalyst (10 µg) onto the TEM grid that is inserted into the cell and contacted
with a gold wire. Thereafter, the floating cell is carefully approached to the liquid meniscus
of the underlying electrolyte (4 M HClO4). This procedure is particularly critical, as a
flooding of the WE must be avoided, since it would prevent the formation of the three-
phase boundary. The RE and CE are an Ag/AgCl electrode and a graphite rod, respectively.
Once ready, gases (O2 or H2) are dispensed from the gas inlet channel and standard cyclic
voltammetry can be performed. The current reached with this technique are of one to two
orders of magnitude greater than the one obtained with RDE as the reactant is not limited
by its solubility in the electrolyte. Please note that this technique is still under
development and that the liquid-solid interface and thus the catalyst utilization need to be
better understood.
- Experimentals
| 41
Figure 3.17 a) Schematic representation of a measurement with the floating cell. Gases (i.e. O2) are
provided through a gas channel, whereas the protons are provided by the electrolyte. The reaction
occurs in the three-phase boundary at the interface between the catalyst (on a TEM grid), the
electrolyte and the gases. b) 3D representation of the floating cell with the inserted TEM grid.
The measurements are always performed at room temperature and an in-house developed
LabVIEW software is used to control both the potentiostat (Gamry reference 600) and the
gas flow.
Mixed Potential Theory
Often, in real electrochemical applications (i.e. metal corrosion, electroless plating,
extraction of minerals) more than one redox couple can be present. For instance, a mixed
redox system with two couples can be expressed as follow:
Red1 + Ox2 --> Ox1 + Red2 Equation 3.11
where the anodic reaction is:
Red1 --> Ox1 + ne- , E1° Equation 3.12
and the cathodic reaction is:
Ox2 + ne- --> Red2 , E2°>E1° Equation 3.13
All these chemical processes can be explained accordingly to the additivity principle
proposed by Wagner and Traud in 1938 [258] and reformulated by Spiro et al. [259, 260].
This principle, known also as the mixed potential theory, states that “the current-potential
curve of a mixture of couples can be obtained by adding algebraically, at each potential, the
currents given by each of the couple present”. And also “an electrode in a system containing
two redox couple automatically adopts a mixed potential (or more correctly, a mixture
potential) Emix”. At this potential, the current of the anodic process (the redox couple of
Chapter 3
42 |
lower Nerst potential as the HOR) and the cathodic process (the redox couple of higher
Nerst potential as the ORR) exactly balance each other and the current can be defined as a
mixed current, Imix (see the schematic representation in Figure 3.18).
Figure 3.18 Schematic representation of the redox couple acting on the catalyst.
From the polarization curves and the relative Imix, it is possible to derive also the rate of
the reactions with Faraday’s law:
𝒎𝒊𝒙 =𝑰𝒎𝒊𝒙
𝒏𝑭
Equation 3.14
where n is the number of exchanged electrons and F is the Faraday constant (96485 C
mol-1). If the additivity principle is applicable, then the rate at which the reaction is
proceeding under catalytic condition (cat) should be the same as the one derived from the
mixed potential approach (mix) and the potentials should correspond as well (Ecat=Emix).
However, it should be underlined that cat equals to mix only if the catalytic and
electrocatalytic measurements occur under comparable conditions. In other words, it is
hard to compare the mix obtained from standard RDE studies with the cat measured in a
batch catalytic reactor for the simple fact that different hydrodynamics (and thus reactant
diffusions) are established. Therefore, to generalize the concept of the mixed potential in
the case of the H2O2 synthesis, the catalysts are characterized electrochemically under
non-diffusion-limited conditions (Chapter 9) using the newly developed floating cell. For
more information about the mixed potential theory please consult the work of Spiro et al.
[259, 261, 262].
- Experimentals
| 43
3.4 Heterocatalytic Measurement
As already discussed in the introduction, hydrogen peroxide can be produced in a
heterocatalytic process starting from a mixture of H2 and O2. Typically, the direct
heterogeneous synthesis can be obtained either in a pressurized batch reactor or in a flow
reactor configuration [19]. In this PhD work a batch reactor consisting in a stainless-steel
autoclave (nominal volume 100 ml) from Parr Instrument (Figure 3.19) equipped with
pressure/temperature sensors and an overhead stirrer (0-2000 rpm) is used.
Figure 3.19 The Parr Instrument system used for the batch reactor measurement.
The measurements, consisting in H2O2 catalytic synthesis and its
hydrogenation/degradation, have been performed at the CCI. The final H2O2 content is
estimated by titration, while analysis of the exhaust gases with chromatography is used to
calculate the reaction selectivity.
3.4.1H2O2 Catalytic Direct Synthesis and Degradation
For the standard direct synthesis test, a solution of methanol and water (5.6 g of MeOH
and 2.9 g of UPW) and 10 mg of supported Au-Pd/C catalyst are mixed together. Once the
autoclave is charged with the prepared solvent/catalyst solution and sealed, it is purged
with 5% H2/CO2 (0.7 MPa) for three times to eliminate ambient air. Next, it is pressurized
with 1.1 MPa 25% O2/CO2 and 2.9 MPa 5% H2/CO2. This gives a 2:1 ratio of oxygen to
hydrogen at a total pressure of 4.0 MPa (note that the maximum working pressure is 14
MPa). Thereafter, the temperature is decreased to 2 °C and stirring at 1200 rpm is
initiated. After 30 min of direct synthesis, the solution is filtered from the catalyst and the
amount of H2O2 produced is determined by titration of ∿0.25 g aliquots.
The H2O2 hydrogenation/degradation activity is tested similarly to the direct synthesis
activity. This time the starting solvent consists in a ∿4 wt% H2O2 solution in ethanol and
water (5.6 g of MeOH, 2.23 g of UPW and 0.67 g 50 wt% H2O2). The precise peroxide
concentration in the initial solution is accurately determined prior to any experiment by
titration (a few drops of the initial solution are used at this point). As we are only
Chapter 3
44 |
interested in the degradation, this time the autoclave is only pressurized with 5% H2/CO2
(again to 2.9 MPa) and the reaction is run for 30 min at 2 °C under stirring conditions.
Finally, the remaining H2O2 present in the filtered solution is determined by titration of
∿0.04 g aliquots.
Titration
Titration is a common volume-based method used in chemistry to determine concentration
of an analyte or titrand in a solution from the reaction with a certain volume (titre) of
prepared standard solution of reagent or titrant. H2O2 concentration in a solution can be
determined using several titrants like Iodine (I2), Permanganate (MnO4-) or Ceric sulfate
(Ce(SO4)2). In this PhD work the solvents are titrated against an acidified Ce(SO4)2
solution using ferroin as an indicator. The precise concentration of the Ce(SO4)2 which is
approximately 8 x 10-3 mol l-1) is determined by standardization against
(NH4)2Fe(SO4)2·6H2O and ferroin again as indicator. The overall chemical reaction
occurring is: H2O2 + 2 Ce(SO4)2 Ce2(SO4)3 + H2SO4 + O2.
From the titre the amount of H2O2 is calculated as follows:
𝑽𝒕𝒐𝒕 𝐂𝐞(𝐒𝑶𝟒)𝟐 =𝒕𝒊𝒕𝒓𝒆 × 𝟖. 𝟓
𝒔𝒐𝒍𝒗𝒆𝒏𝒕 𝒂𝒍𝒊𝒒𝒖𝒐𝒕
Equation 3.15
Where Vtot Ce(SO4)2 is the total volume that would be required to titre the 8.5 g of solvent.
𝐌𝒐𝒍𝒆𝒔 𝐂𝐞(𝐒𝑶𝟒)𝟐 =𝑽𝒕𝒐𝒕 𝐂𝐞(𝐒𝑶𝟒)𝟐 × [𝐂𝐞(𝐒𝑶𝟒)𝟐]
𝟏𝟎𝟎𝟎
Equation 3.16
And finally, the moles of H2O2
𝐌𝒐𝒍𝒆𝒔 𝑯𝟐𝑶𝟐 =𝐌𝒐𝒍𝒆𝒔 𝐂𝐞(𝐒𝑶𝟒)𝟐
𝟐
Equation 3.17
From the moles of H2O2 the productivity value, conventionally presented as molH2O2 kgmetal-1
h-1, is calculated. In this work, however, in order to compare the values to the
electrochemical data the productivity is shown as molH2O2 µgmetal-1. The peroxide
degradation is, instead, obtained by calculating the ratio between the final and initial
moles of H2O2.
3.4.2 Gas Chromatography for the Analysis of Exhaust Gas
Gas chromatography (schematic in Figure 3.20) is an analytical technique used commonly
to analyse qualitatively as well as quantitatively a mixture of gases. The mixture of gases
is injected together with a carrier gas or mobile phase (typically inert gases He or N2) in a
column filled by a stationary phase (typically a liquid or polymer) where the separation
- Experimentals
| 45
occurs as each gas is carried at different speed through the column. The speed and rates
depends on the physical and chemical properties of the gases or component to be separated
and on their interaction/adsorption with the column wall and/or the stationary phase.
Finally, the components are identified and quantitatively detected as they leave the
column at different times. Commonly a thermal conductivity detector (TCD) is used, which
relies on the different thermal conductivities of the component being measured and the
carrier gas.
Figure 3.20 Schematic of a gas chromatography system.
During the analysis of exhaust gases from the direct H2O2 synthesis in a batch reactor, Ar
is used as a carrier gas and remaining H2, O2 and CO2 (retention times 1.4, 2.7 and 7.2 min
respectively) are detected quantitatively. To the total gas flow approximately 1% of N2 is
added as internal standard to separate peaks easily to which the areas of O2 and H2 are
compared. The hydrogen conversion (in %) is determined from the difference between the
initial and final H2:N2 ratios, which are proportional to the moles of H2 before and after the
reaction:
𝐂𝐨𝐧𝐯𝐞𝐫𝐬𝐢𝐨𝐧 =𝐅𝐢𝐧𝐚𝐥 𝐦𝒐𝒍𝒆𝒔 𝑯𝟐
𝐈𝐧𝐢𝐭𝐢𝐚𝐥 𝐦𝐨𝐥𝐞𝐬 𝑯𝟐
Equation 3.18
The reaction selectivity is determined from the amount of H2O2 produced and H2 converted:
𝐒𝐞𝐥𝐞𝐜𝐭𝐢𝐯𝐢𝐭𝐲 =𝐌𝒐𝒍𝒆𝒔 𝑯𝟐𝑶𝟐
𝐂𝐨𝐧𝐯𝐞𝐫𝐭𝐞𝐝 𝐦𝐨𝐥𝐞𝐬 𝑯𝟐
Equation 3.19
A Varian 3800 gas chromatograph equipped with a TCD is employed and the reactor is
fitted with a 4 m molecular sieve 5 Å column. The column is held for 6 min at 40 ºC to
allow the separation of the gases and then it is increased to 200 ºC (25 ºC min-1) to
remove all CO2 and moisture from the column.
Chapter 4
46 |
- Material Synthesis and Chapter 4
Characterization11
——————————————————————————————————————
In the following chapter, the Au-Pd materials employed for the electrocatalytic and
heterocatalytic studies will be presented and their initial characterization described. The
particles’ average sizes and size distributions are obtained from statistical analysis with
STEM, whereas their composition is estimated from ICPMS analysis of the dissolved
catalysts, as well as with XPS and/or EDS measurements of the freshly prepared catalysts.
Furthermore, initial CVs in Ar purged electrolyte are also recorded to study the influence
of the composition on the oxidation and reduction potentials. For sake of simplicity, the
unsupported nanoparticles, obtained directly in the colloidal solutions, and the carbon
supported nanoparticles, obtained via colloidal immobilization onto Vulcan XC72R, are
hereafter described separately.
——————————————————————————————————————
11 Parts of this chapter have been already published in:
E. Pizzutilo*, S.J. Freakley, S. Geiger, C. Baldizzone, A. Mingers, G.J. Hutchings, K.J.J. Mayrhofer, S. Cherevko
Catal. Sci. Technol. 2017, 7, 1848-1856.
E. Pizzutilo*, S. Geiger, S.J. Freakley, A. Mingers, S. Cherevko, G.J. Hutchings, K.J.J. Mayrhofer Electrochimica
Acta 2017, 229, 467–477.
E. Pizzutilo*, O. Kasian, C.H. Choi, S. Cherevko, G.J. Hutchings, K.J.J. Mayrhofer, , S.J. Freakley Chem. Phys.
Lett.. 2017, 683, 436-442.
E. Pizzutilo*, S.J. Freakley, S. Cherevko, S. Venkatesan, G.J. Hutchings, C.H. Liebscher, G. Dehm, K.J.J.
Mayrhofer, ACS catalysis.. 2017, 7, 5699-5705.
E. Pizzutilo*, On-demand H2O2 production: a study at the border between electro and heterogeneous catalysis (in
preparation)
There are therefore numerous verbal quotes from that publication. Some of the figures present in
the publication have been re-printed or modified
- Material Synthesis and Characterization
| 47
4.1 Unsupported Au-Pd Catalyst
4.1.1 Particle Size - STEM
Figure 4.1 a) Representative bright field TEM micrographs and statistic particle size distributions
showing a) Au, b) Au9Pd, c) Au3Pd, d) AuPd, e) AuPd3 and f) Pd sol gel nanoparticles deposited on a
lacey carbon TEM grid.
The Au-Pd catalysts are synthesized, following previous literature on the direct synthesis
of H2O2 [123, 263], through the sol-immobilization method described in the previous
Chapter 4
48 |
Chapter 3. The various colloidal solutions obtained with different compositions from pure
Au to pure Pd (Au, Au9Pd, Au3Pd, AuPd, AuPd3, Pd) are deposited onto a lacey carbon grid
for STEM characterization with the JEOL 2200FS microscope.
From the acquired bright field STEM micrographs (Figure 4.1), the particle size
distribution and the average size of the different catalysts are estimated statistically (see
details of the measurement in the dedicated Chapter 3). The number of analyzed
nanoparticles is around 300-400 particle and their average size (with relative standard
deviation) are summarized in Table 4.1. The size resulted in the range 3-4 nm for all the
considered samples. Previous HR-TEM characterizations of the so-synthesized catalysts
reported the presence of a face-centered cubic (fcc) structure [123].
Table 4.1 Particle size and specific surface area of the prepared materials investigated in this
study.
median /
nm st. dev.
ECSA* / m2
g-1
Au
Au9Pd
Au3Pd
AuPd
AuPd3
Pd
4.3
2.7
3.5
4.1
3.7
3.2
±1.3
±1.0
±1.1
±0.8
±1.1
±1.2
73
121
98
92
120
150
*ECSA refers to the catalyst specific surface area, which was calculated from the particle mean
size.
Note that, in the micrographs, enhanced contrast in some region could indicate the
presence of overlapped nanoparticles. Therefore, also in the catalyst film formed on the
electrodes particles could overlap, depending on the loading and on drying condition.
4.1.2 Composition – XPS and ICPMS
Few drops of the colloidal nanoparticles with various composition are deposited onto a GC
support and analyzed with XPS. The obtained XPS spectra are displayed in Figure 4.2. For
all the samples containing Pd, a distinct shift in the binding energy of the Au(4f)
photoelectron peak is observed towards lower binding energies compared to pure Au
(whose peak is at 84.2 eV). Such a shift indicates the alloying of Au with Pd and it is
typically observed elsewhere in literature for Au-Pd alloys [123, 264, 265]. The analysis of
the Pd(3d) peak (binding energy ∿336 eV, for pure Pd), is rather difficult due to the
overlapping with the Au(4d) component (∿334.5 eV for pure Au). For low Pd composition,
the nanoparticles comprise Pd0 species predominantly, whereas a second peak
corresponding to the presence of Pd2+ species is observed for AuPd3 and Pd.
- Material Synthesis and Characterization
| 49
Figure 4.2 XPS spectra for the series of freshly prepared Au-Pd catalysts.
From the relative intensities of the Au and Pd peaks in the XPS spectra, the Au:Pd molar
ratio can be calculated and compared to the values obtained by ICPMS analyses. The
results summarized in Table 4.2 show a substantial consistency between the values
obtained with the two techniques, thus proofing their validity. Only the Au molar ratio of
the Au9Pd obtained via ICPMS resulted slightly lower compared to the one obtained via
XPS. However, note that for low Pd content the estimation of the molar ratio with XPS is
less accurate owing to direct overlap between Pd(3d) and Au(4d) peaks.
Table 4.2 Particle size and Au:Pd molar ratios estimated via ICP-MS and XPS of the prepared
catalysts.
Au:Pd ratio
ICPMS
Au:Pd ratio
XPS
Au
Au9Pd
Au3Pd
AuPd
AuPd3
Pd
1:0
91:9
77:23
46:54
26:74
0:1
1:0
98:2
74:26
48:52
27:73
0:1
4.1.3 Cyclic Voltammetry
Polycrystalline
Prior to any other measurement, CVs of polycrystalline poly-Pd and poly-Au are collected
in Ar saturated electrolyte (0.1M HClO4) as a reference (Figure 4.3). The values are
normalized to the geometric surface areas (0.196 cm-2 for the 5mm polycrystalline
electrodes). In the typical profiles (recorded in Figure 4.3a,b), different surface processes
can be observed at different potential regions.
Chapter 4
50 |
Figure 4.3 CVs of poly-Pd (a) and poly-Au (b) electrodes showing the different electrode processes
occurring in Ar purged 0.1M HClO4. The dependence of the reduction peaks from the applied UPL
is shown for poly-Pd (c) and poly-Au (d). Scan rate: 200 mV s-1.
The Pd electro-oxidation commences at approximately 0.7 VRHE in the anodic scan. In the
cathodic direction, a well-defined oxide reduction peak is typically visible below 0.8 VRHE
with a maximum around 0.6-0.7 VRHE [150, 167, 266, 267]; also a second peak for oxide-
reduction is reported in literature around 1.2-1.3 VRHE. The latter is reported to correspond
to the reduction of Pd(IV)-oxide formed at high potentials [150, 267]. In the CVs in Figure
4.3a this broad peak is not visible due to the low upper potential limit (UPL) applied;
however, it will be discussed with more detail in the following Chapter 5. Proceeding
cathodically, after the so-called double layer region (where no faradaic process occurs), at
potentials lower than 0.3 VRHE the concurrent H adsorption and bulk absorption (with the
formation of Pd hydride) originate a large cathodic current. It is reported in literature that
Pd, unlike Pt, absorbs hydrogen in the potential range where the under potential
deposition of H (HUPD) as well as the hydrogen evolution (HER) occur [150, 268]. At higher
potentials than HER, the absorbed hydrogen (Habs) in the poly-Pd bulk structure is
desorbed resulting in a large anodic current. A more detailed discussion on Pd oxidation
will be provided in Chapter 5; the interested reader is also referred to the review on Pd
electro-oxidation of Grdeń et al. [150].
- Material Synthesis and Characterization
| 51
Unlike Pd, Au does not show any activity for the hydrogen absorption/desorption (Figure
4.3b) at low potentials. This region is dominated by the charging of the double layer
capacitance[269]. At potential higher than 1.0 VRHE the oxidation and reduction of the Au
surface occur. Despite uncertainties and still ongoing debate around Au oxidation
mechanism, it is believed to be initiated by coverage of adsorbed OH/O species (at potential
above 1.2-1.3 VRHE) and incorporation into the subsurface layers [147, 148, 175, 270-272].
At higher potential, also hydroxide in bulk phase can be formed [148]. The oxide
formation/reduction depends on various parameters as on the crystalline orientation, grain
size, pH [91].
Increasing the UPL the amount of oxide being formed increases. Thus, the charge
associated to the reduction peaks increases gradually (Figure 4.3c,d). Furthermore, the
hysteresis between anodic and cathodic scan increases and the reduction peak shifts to
lower potentials. This hysteresis is more marked for poly-Pd, which behaves similarly to Pt
[134] and its origin is not fully understood at present [150].
To better compare the different processes in the poly-Pd and poly-Au, CVs with a similar
amount of reduction charge are shown in the same following graph (Figure 4.4).
Figure 4.4 CVs of poly-Au and poly-Pd electrodes in Ar purged 0.1M HClO4. Scan rate: 200 mV s-1.
Unsupported nanoparticles
Initially, Ar background CVs of the as prepared catalysts are recorded 0.1M HClO4 (Figure
4.5).
Chapter 4
52 |
Figure 4.5 initial cyclic voltammogramms [0.1-1.6] VRHE for the series of freshly prepared Au-Pd
catalysts in Ar purged 0.1M HClO4. Scan rate: 200 mV s-1.
The curves are normalized by the total surface area (Ametal in Table 4.3) estimated from the
TEM particle size following Equation 3.1-v and considering a loading of 1.96 µgmetal (10
µgmetal cm-2 onto the GC electrode of the RDE). Note that these values are obtained with
several approximations and that the real surface areas might slightly be different.
However, the same method has been applied for all the catalysts; therefore, in a first
approximation, these values can be here used for catalysts comparison.
Table 4.3 Metallic surface area calculated considering a loading of 1.96 µg.
Ametal (cm2)
Au
Au9Pd
Au3Pd
AuPd
AuPd3
Pd
1.43
2.36
1.93
1.80
2.31
2.42
The shape of the CVs mirrors the nanocatalyst surface state that only depends on the alloy
composition. An UPL of 1.6 VRHE is chosen in order to measure also the oxide reduction
peak of Au which is only visible for UPL > 1.5 VRHE as shown in Figure 4.3d [148]. The
typical features of Pd and Au observed in the polycrystalline electrodes (Figure 4.3 and
Figure 4.4) are present and variate with the composition. At potentials lower than 0.4 VRHE
hydrogen is adsorbed on the Pd active sites, resulting in the well-known HUPD [150]. The
nanoparticles do not show any large cathodic current corresponding to hydrogen bulk
absorption. The uptake of hydrogen is limited with the nano-size and therefore only a
distinct HUPD is recorded in accordance with literature [85, 150, 273, 274]. As expected
from the previous measurement on poly-Au (Figure 4.3b), HUPD is not observed in pure or
low Au samples and is only present from the composition 1:1 molar (Au:Pd). This
- Material Synthesis and Characterization
| 53
observation is in well agreement with the work of Lukaszewski et al. showing that
hydrogen adsorption is visible only when Pd content exceeds 30 at.% [275]. The HUPD
features are influenced by alloying. For instance, two desorption peaks are observed for
alloys with Pd content higher than 65 at% (as AuPd3 in our case) [275].
Proceeding anodically, the onset potentials for the formation of Au oxide (Au-O) and Pd
oxide (Pd-O) are around 1.3 VRHE and 0.7 VRHE respectively. These values are in accordance
with the ones obtained for polycrystals. Note that, both the oxide coverages as well as the
reduction peaks depend on the applied UPLs. The Au-O and Pd-O reduction peaks
dominate the cathodic scan. Their maxima are comprised between 0.6 VRHE and 1.1 VRHE,
i.e. between the maxima of reduction peaks of pure Pd and pure Au, respectively. Among
the alloyed samples, one or two reduction peaks can be observed and their presence and
amplitude reflect the composition: i) while for low Pd content (Au9Pd) only Au-O reduction
is recorded (as in [105]), ii) for high Pd content (AuPd3) only Pd-O reduction is present (as
in [241]); iii) for intermediate compositions (Au3Pd and AuPd), instead, both reduction
peaks are observed. Interestingly, the Pd-O reduction peak position is shifting significantly
towards lower potentials when the Pd content increases. Such a shift has been attributed
to alloying of Au and Pd during the synthesis [105, 126, 241], achieved also thanks to the
high diffusivity of Pd in Au [276]. A less pronounced shift can be also observed for the Au-O
reduction and it depends on the composition [105, 241]. Analyzing the CVs, it is in
principle possible to derive the real electroactive surface area of Au and Pd separately, i.e.
by using i) the charge under the oxide reduction peak [241], ii) the charge under the oxide
formation peak, iii) the charge of the HUPD [277] or iv) the linear dependence of Pd-O
reduction shift with the composition as exploited by Rand and Woods [278]. However, i-ii)
the estimation of the Pd-O oxidation/reduction peaks is challenging when different
oxidized states are present [150] and in general for alloys [126, 241]. Indeed, in Au-Pd
alloys the “so-defined” Pd-O reduction can be linked to the oxygen desorption from a new
surface phase, rather than simply from a Pd surface. Such a new phase results from the
atomic interaction occurring between Pd and Au [241]. iii) Concerning the HUPD, Au is
reported to hinder the hydrogen adsorption/desorption [105]. Moreover, in both cases, the
integration of the charge is often affected by systematic errors depending on the definition
of the baseline. Finally, iv) it was shown that the approach of Rand and Woods is limited
only to alloys with Pd content higher than 40 at% [241, 279] and for Au-Pd non polarized to
the region where hydrogen absorption takes place [241]. Despite these limitations, the
reduction charges and positions are good indicators for a rough estimation of catalysts
surface composition. The active sites, for the reaction we are interested in, are not yet fully
understood. Thus, the normalization for the area of a single metallic phase is meaningless.
Therefore, in this study the CVs are mostly normalized either to the mass or to the total
surface area Ametal calculated from the statistical particle size and loading.
Chapter 4
54 |
4.2 Supported Au-Pd/C Catalyst
4.2.1 Particle size and Composition – STEM-EDS
Figure 4.6 Bright Field (BF) scanning transmission electron microscopy (STEM) micrographs and
relative particle size distributions of a) Au/C, b) Pd/C and c) AuPd/C. d) Dark-field high-resolution
(HR) STEM investigation of as synthesized AuPd/C with relative EDS spectra (red for Au peak and
green for Pd peak).
- Material Synthesis and Characterization
| 55
The carbon supported Au-Pd/C catalysts are synthesized via immobilization of the Au-Pd
colloidal solutions (previously characterized) onto Vulcan XC72R carbon, following
previous literature [123, 263]. Three catalysts with a metal loading of 10 wt% have been
synthesized, namely Au/C, AuPd/C and Pd/C. The support is necessary for understanding
and studying the catalysts stability during electrochemical ADPs as in the following
Chapter 8 and for heterocatalytic studies as in the following Chapter 9. The obtained
catalyst powders are dispersed in UPW via ultrasonication bath; once a homogeneous ink
is obtained few drops are deposited onto a grid for STEM characterization with the JEOL
2200FS and the TITAN microscopes. The acquired bright field micrographs of the fresh
catalysts are used, as described before, to estimate their size (Figure 4.6a,b,c).
The particle size distribution is relatively homogeneous and narrow around ca. 3-4 nm for
the considered samples (see also Table 4.4). EDS mapping and spectra are acquired with
the TITAN STEM microscope in high-angle annular dark-field (HAADF) mode for the sole
AuPd/C sample (Figure 4.6d). EDS spectra obtained from individual particles showed
characteristic Au M lines and Pd L lines confirming the presence of intimately mixed AuPd
alloys rather than of physical mixtures of pure Au and Pd particles. Furthermore, Au and
Pd appear to be randomly and homogeneously distributed in the nanoparticles as shown in
Figure 4.7.
Figure 4.7 Dark-field high-resolution(HR) STEM micrograph of a single AuPd/C nanoparticle and
EDS mapping of Au (red) and Pd (green).
The composition estimated from the EDS spectra indicate that Au average molar ratio is
around 44 at%. Thus, it is close to the theoretical composition as calculated from the
amount of precursors used during the synthesis. Further ICPMS analysis of the fresh
catalyst dissolved in aqua regia shows a similar Au:Pd ratio of 46:54 (Table 4.4). Please
note that the composition of a single particle can show even important deviation depending
on size as previously observed for such alloys [123].
Chapter 4
56 |
Table 4.4 Particle size and specific surface area of the prepared materials investigated in this
study.
median /
nm st. dev.
ECSA* / m2
g-1
Au:Pd ratio
ICPMS
Au:Pd ratio
EDS
Au/C
AuPd/C
Pd/C
3.8
3.1
4.1
±1.2
±1.2
±1.3
82
120
122
1:0
46:54
0:1
1:0
44:56
0:1
*ECSA refers to the catalyst specific surface area, which was calculated from the particle mean
size.
4.2.2 Cyclic Voltammetry
As already observed for the unsupported nanoparticles, the presence of both Pd and Au on
the AuPd/C catalyst surface is confirmed by the Pd-O and Au-O reduction peaks observed
in the initial CV of the as-prepared catalysts (Figure 4.8). The shift of Pd-O reduction peak
to higher potentials indicates the presence of alloying [105, 126, 127, 241]. Furthermore,
also the Pd oxidation wave in the anodic scan is shifting with alloying. As expected, the
presence of the high surface area Vulcan support results in the higher double layer
capacitances observed here (compared to the ones shown in Figure 4.5).
Figure 4.8 Initial cyclic voltammograms [0.1-1.6] VRHE of fresh Au/C, AuPd/C and Pd/C catalyst
recorded in Ar purged 0.1M HClO4. Scan rates: 0.2 V s-1.
- Palladium Electrodissolution from Model Surfaces and Nanoparticles
| 57
- Palladium Electrodissolution from Chapter 5
Model Surfaces and Nanoparticles12
——————————————————————————————————————
A comprehensive knowledge of the catalysts’ stability is necessary when considering their
applications in electrocatalytic systems which require high durability over the years of
operation. Dealloying, caused by the faster dissolution of one of the alloyed elements, can
be crucial for the catalyst choice as the compositional and structural changes influence the
catalyst performances, like activity and selectivity to H2O2. However, prior to any study
dealloying study on Au-Pd catalysts, the fundamental dissolution of the single Au and Pd
metals should be first addressed. While the dissolution of Au-poly has been studied
thoroughly in the past years, the Pd stability is still not fully understood and Pd is one of
the few noble metals whose dissolution has not been studied with the SFC-ICPMS yet.
Therefore, the following chapter will be dedicated to the understanding of the
oxidation/dissolution behaviors in acidic media (sulfuric and perchloric acids). Crucial
parameters influencing dissolution like potential, scan rate, UPL and electrolyte
composition will be introduced. In addition, a comparison between poly-Pd and the
supported high-surface area Pd/C catalyst is carried out. The results evidence that three
main contributions (one anodic and two cathodic) promote the transient dissolution. At
potentials below 1.5 VRHE the anodic dissolution is the dominating mechanism, whereas at
higher potentials the cathodic mechanisms prevail. Based on the experimental outcome of
this comprehensive study a mechanism for Pd dissolution is suggested in the conclusion.
——————————————————————————————————————
12 Parts of this chapter have been already published in:
E. Pizzutilo*, S. Geiger, S.J. Freakley, A. Mingers, S. Cherevko, G.J. Hutchings, K.J.J. Mayrhofer Electrochimica
Acta 2017, 229, 467–477.
There are therefore numerous verbal quotes from that publication. Some of the figures present in the publication
have been re-printed or modified.
Chapter 5
58 |
5.1 Poly-Pd Oxidation and Reduction in Different Acidic
Media
Pd voltammogramms in deaerated solution (Figure 5.1) are recorded using the SFC with
0.1M HClO4 and 0.1M H2SO4 as electrolytes. Typical profiles for poly-Pd in aqueous acidic
solutions are observed. The Pd oxidation in 0.1M HClO4 commences at approximately 0.7
VRHE (here A1 peak) in the anodic scan. In the cathodic scan direction, a well-defined Pd-O
reduction peak is visible below 0.8 VRHE (here C1 peak) [150, 167, 266, 267], while a second
poorly defined peak for Pd-O reduction is reported in literature around 1.2-1.3 VRHE (here
C2 peak). The latter is thought to correspond to the reduction of Pd(IV)-oxide formed at
high potentials [150, 267]. In Figure 5.1 this broad peak is labelled, even though is not
clearly visible as the applied UPL is too low. However, it will be discussed in the following
sections, where UPLs up to 1.8 VRHE are considered in the study of the dissolution.
Figure 5.1 CVs recorded for a poly-Pd electrode in the SFC setup in 0.1M HClO4 and in 0.1M H2SO4.
Scan rate: 200 mV s-1. The position of the anodic oxidation peak and two cathodic reduction peaks
are indicated with A1, C2 and C1 respectively. The complete Pd CV (including hydrogen
adsorption/absorption and desorption) is described in Figure 4.3.
Interestingly, in the two considered electrolytes the onset potentials for the Pd electro-
oxidation slightly differ: in H2SO4 the onset potential is higher (ca. 0.75 VRHE). This can be
attributed to the difference in the anion adsorption strength, which can have a strong
influence in the Pd electro-oxidation behavior [150]. Solomun et al. claimed that
perchlorate anions (ClO4-) do not undergo specific adsorption and that only weak
(electrostatic) interactions occur between the Pd surface and the anions of the electrolyte.
On the other hand, the interaction of other anions such as the (bi-)sulfate anion (HSO4-
/SO42-) can be stronger [160-162]. Furthermore, in H2SO4, the charge associated to Pd-O
- Palladium Electrodissolution from Model Surfaces and Nanoparticles
| 59
reduction is slightly higher and the peak is also slightly shifted. A more detailed discussion
on Pd oxidation and anion influence will be provided in the final part of the present
chapter. However, the reader is also referred to the critical review on Pd literature of
Grdeń et al. [150].
5.2 Poly-Pd Electrodissolution in Different Acidic Media:
Influence of UPL
Potential sweeps to increasing UPL in the two considered acidic media are applied to poly-
Pd electrode. The potential program and the corresponding dissolution profile are
presented in Figure 5.2a,b, respectively. The cleaning cycles (30 CVs at 200 mV s-1) are
characterized by an initially higher Pd dissolution signal, which is probably due to the
contribution of initially present surface defects. After approximately 10 CVs a constant Pd
dissolution signal and a stable CV is measured, indicating that a clean, steady surface
state for this potential window is obtained.
Figure 5.2 a) Potential program applied to the poly-Pd electrode consisting of 30 scans (200 mV s-1)
for cleaning, an open circuit potential (OCP) phase and several scans (10 mV s-1) with increasing
UPL. The measured poly-Pd dissolution profiles are shown in b). The inset in b) corresponds to the
integrated dissolved mass of Pd per cycle at different UPL. The corresponding CVs are shown in
Figure 5.3.
During the slower CV, Pd dissolution is observed at potentials where Pd oxidizes (E > 0.7
VRHE) and a small deviation from the background signal is observable first with an UPL
above 0.8-0.85 VRHE, in line with the onset potential shown by Łukaszewski et al. obtained
with the quartz microbalance [171]. The amount of formed Pd oxide and thus the
dissolution increases with the applied UPL. In fact, the charge associated to both the
Pd(II)-O reduction peak (C1) and the Pd(IV)-O reduction peak (C2) increases gradually with
Chapter 5
60 |
potential (Figure 5.3). In literature it is known that increasing the UPL the reduction peak
of Pd is shifting to lower potentials as a direct consequence of the different amount of oxide
formed [150] (as observed also for Pt [134]), even though a clear explanation is not
available at present.
Figure 5.3 Cathodic sweeps of the slow scan (10 mV s-1) for different UPL in 0.1M H2SO4 and HClO4
are shown in a) and b) respectively. The charges corresponding to the Pd-O reduction peaks are
also plotted: c) charge density QC1 of the Pd(II)-O reduction PeakC1 as a function of the UPL and d)
charge density QC2 of the Pd(II)-O reduction PeakC2 as a function of the UPL.
Note that the position of the Pd(II)-oxide reduction peak (C1) in HClO4 is shifting more to
lower potentials compared to the shift in H2SO4 (up to 50 mV difference). Similarly, the
associated Pd(II)-oxide reduction (C2) charge is initially the same, while at higher
potentials a difference up to ca. 20% in the reduction charge (higher in H2SO4) was
measured (Figure 5.3). This is probably due to the different interaction of the electolytes
anions with the Pd electrode (see discussion). At different UPL up to three different peaks
in the Pd dissolution profile (corresponding to the peak anodic A1 and cathodic C2, C1
respectively) can be observed. A comparison of the mass dissolved during the anodic and
the two cathodic contributions to the transient dissolutions in the two acids are shown in
Figure 5.4. Note that until 1.1 VRHE only a single dissolution peak is visible in both
electrolytes. However, the applied scan rate (10 mV s-1) does not allow a clear separation
between the individual dissolution peaks. Therefore, similar measurements at selected
UPLs with a slower scan rate (2 mV s-1) are presented in the next paragraph.
- Palladium Electrodissolution from Model Surfaces and Nanoparticles
| 61
Figure 5.4 Integrated mass of dissolved Pd corresponding to the anodic dissolution peak (A1) and
the two cathodic peaks (C2 and C1) are reported in HClO4 (a) and H2SO4 (b) at different UPL during
the protocol shown in Figure 5.2. c) Comparison of the anodic (A1) dissolution peak in the two
acids.
The cathodic dissolution peaks (C2 and C1) increase constantly (Figure 5.4a,b) with
increasing UPL as more oxide is formed, with C2 becoming the dominating contribution at
high potential. The quantitative dissolution values per cycle are reported in Table 5.1,
along with the measured dissolution of Pt and Au under similar conditions.
Table 5.1 Comparison of the amount of Au, Pt [147], Pd in 0.1M H2SO4 and Pd* in 0.1M HClO4
dissolved per cycle depending on the applied UPL as derived from potential sweep experiments at
10 mV s-1. BDL stands for below the detection limit.
UPL / VRHE Au / ng cmgeo
-2
cycle-1
Pt / ng cmgeo
-2
cycle-1
Pd / ng cmgeo
-2
cycle-1
Pd* / ng cmgeo
-2
cycle-1
0.9
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
BDL
BDL
BDL
BDL
BDL
1.6
4.4
7.4
12.5
20
BDL
BDL
0.4
1.3
2.7
4.4
5.8
7.0
8.0
9.0
0.36
5.1
21.3
51.5
83.6
114.2
149.9
185.8
224.4
271.7
0.02
0.8
4.8
12.9
18.9
22.3
26.6
32
39
50
Chapter 5
62 |
Comparing the dissolution in the same medium (sulfuric acid), it turns out that Pd is
dissolving at a much higher rate than the other noble metals considered. Furthermore, Pd
in H2SO4 dissolves 5 times more than in perchloric acid. Similar trends were reported in
other works [150, 153, 171] and it was attributed to the formation of different complexes
between the dissolved species and the anion in the electrolyte (see discussion).
In Figure 5.4c it is shown the anodic contribution (A1) to the transient dissolution; unlike
the other to contribution it does not increase steadily with the UPL. Indeed, it is possible to
identify different stages in the anodic transient dissolution behavior: i) A first immune
region at potentials lower than Pd oxidation; ii) a region between 0.8 and 1.4 VRHE where
the transient anodic dissolution is increasing with the UPL; iii) a region in the 1.4-1.7 VRHE
potential range, where the transient anodic dissolution is constant (independently of the
UPL), due probably to the oxide coverage that lead to passivation and iv) a region for
potential higher than 1.7-1.8 VRHE, where the transient anodic dissolution increases again
and could be attributed to the surface change in the OER potentials. Anodic passivation is
also confirmed by the decay in the dissolution signal during potentiostatic (steady-state)
experiment (Figure 5.5).
Figure 5.5 Potentiostatic dissolution of poly-Pd at different applied potential during a potential
step experiment (time of each potential step: 300 s).
The potential program applied consisted in 30 activation cycles followed by open circuit
potential (OCP) and a series of potential steps of 300 s each with increasing potential from
0.6 to 1.6 VRHE (0.2 V for each step). Dissolution is observed starting from 1.0 VRHE. For
each step is observed a jump in dissolution, followed by a fast decay, indicating that there
is no continue steady-state dissolution. Indeed, with time the oxide is covering and thus
passivating the metal surface, resulting in the observed decrease in the dissolution.
- Palladium Electrodissolution from Model Surfaces and Nanoparticles
| 63
5.3 Poly-Pd Electrodissolution in Different Acidic Media:
Slower Scan Rate
Potential sweeps to increasing UPL (0.9, 1.2, 1.5, 1.8 VRHE) in the two different acidic
media with a 2 mV s-1 scan rate are applied to a poly-Pd electrode (Figure 5.6, Figure 5.7
and mass cycles in Figure 5.8). At this slow scan rate the different dissolution processes
occurring during cyclic voltammetry can be clearly distinguished.
Figure 5.6 a) 4 slow scans (2 mV s-1) with increasing UPL (0.9, 1.2, 1.5, 1.8 VRHE) and b) the
corresponding measured poly-Pd dissolution profiles in 0.1M HClO4 and H2SO4. The position of the
first anodic (A1) dissolution peak and the two cathodic (C2 and C1) are marked by red, grey and
blue arrows respectively.
As expected, due to the slower scan rate, the dissolution per cycle is higher; furthermore,
the observed quantitative difference between dissolution in perchloric and sulfuric acid is
confirmed (see values in Table 5.2). However, in the case of slower scan rates the difference
appears to be slightly reduced (the dissolution in H2SO4 is here only almost 3 times than in
HClO4, while at faster scan rate is 5 times).
Table 5.2 The comparison of amount of Pd in 0.1M H2SO4 and Pd* in 0.1M HClO4 dissolved per cycle
depending on the applied UPL as derived from potential sweep experiments at 2 mV s-1.
UPL /
VRHE
Pd / ng cmgeo
-2
cycle-1
Pd* / ng cmgeo
-2
cycle-1
0.9
1.2
1.5
1.8
1.5
98.8
259.6
429.6
0.06
36.9
80.8
106.3
Chapter 5
64 |
Colored arrows in Figure 5.6 mark the positions of the peaks: red corresponding to the
anodic oxidation/dissolution (A1), grey and blue corresponding to the two cathodic
reduction/dissolution peaks (C2 and C1 respectively). For the sake of clarity, the single
dissolution profiles are shown separately in Figure 5.7.
Figure 5.7 Poly-Pd dissolution profiles in 0.1M HClO4 and H2SO4 with a) UPL= 0.9 VRHE (comparison
of the poly-Pd dissolution onset potential in the inset), b) UPL= 1.2 VRHE, c) UPL= 1.5 VRHE, d) UPL=
1.8 VRHE. Scan rate: 2 mV s-1.
The dissolution onset potential can be evaluated as the deviation from the background
signal. In the two electrolytes, the measured dissolution onsets appear shifted of
approximately 50 mV (inset of Figure 5.7a). This could also be caused by the difference in
the dissolution rates of Pd in the two analyzed electrolytes. Indeed, Pd in HClO4might also
dissolve earlier than measured, but just being below the ICP-MS detection limit.
The maximum of the anodic dissolution peak in the two electrolytes matches very well for
all the different UPL, whereas the maximum of the cathodic dissolution peaks, in
particular the peaks C1 are delayed with increasing applied UPL in HClO4. This delay
mirrors the greater shift of the Pd(II)-oxide reduction peaks with UPL observed in the CVs
(Figure 5.3).
- Palladium Electrodissolution from Model Surfaces and Nanoparticles
| 65
The UPLs are chosen to distinguish dissolution processes occurring at different potentials.
i) At a potential lower than 1.1 VRHE only one peak is present as a combined minor anodic
and cathodic peak. ii) In the potential range between 1.1 and 1.4 VRHE a shoulder peak
related to the cathodic dissolution due to the C1 reduction starts to appear (blue arrow).
With the measured UPL of 1.2 VRHE the maximum of this second peak is measured at 0.8
VRHE during the cathodic scan, which well corresponds to the C1 peak observed in CV with
the same UPL. (iii) At more positive potentials, a third dissolution peak between the two is
appearing (gray arrow) and is increasing dramatically. With an UPL of 1.5 VRHE the
maximum of this third peak is measured at 1.1-1.2 VRHE during the cathodic scan, which
matches the broad reduction peak C2 observed in the CVs. The corresponding mass cyclic
voltammograms acid are shown in Figure 5.8, indicating the trend of the three different
contributions (one anodic and two cathodic) to the dissolution more clearly.
Figure 5.8 Mass cyclic voltammograms in a) 0.1M H2SO4 and b) HClO4 corresponding to the
dissolution profiles in Figure 5.6. The percentage of anodic (A1) and cathodic dissolution (C2 and
C1) are shown for the respective acid in the insets.
At UPLs up to 1.5 VRHE the three peaks are not perfectly separated, despite the very low
scan rate (2 mV s-1), while at 1.8 VRHE the anodic dissolution and the first cathodic
dissolution peaks appear nicely distinguished. Furthermore, the anodic dissolution
maxima appear to be at the same potential for all the four cycles, whereas the cathodic
dissolution maxima shift to lower potentials in accordance with the shifts of the reduction
peaks (Figure 5.3), due to the irreversibility of the oxide formation [150] as reported also
for other noble metals [134, 280]. Interestingly, in H2SO4 the dissolution maximum appears
to be before the reduction maximum (the former is approximately 30 mV higher; see Figure
5.9). Similar findings were also obtained for Pt cathodic dissolution in H2SO4 [136]. In
HClO4, instead, the two peak potentials correspond well. This difference is not well
Chapter 5
66 |
understood at present and it might derive from the different interactions of the electrolyte
anion with Pd.
Figure 5.9 Correlation between cathodic dissolution and Pd(II)-oxide reduction signals in 0.1M
H2SO4 (a) and HClO4 (b). UPL: 1.5 VRHE. Scan rate: 2 mV s-1.
These results allow us already to dissipate some controversy about the nature of Pd
dissolution. As discussed in the introduction, there is an ongoing debate whether it is an
anodic process or not. The relative contribution to the dissolution of the three different
peaks is shown in the inset of Figure 5.8. At low UPL the process is predominantly anodic
(note that however below 1.1 VRHE only one peak is appearing and is not possible to
distinguish between anodic and cathodic dissolution). Increasing the UPL it first appears
the peak C1 and above 1.4 VRHE the peak C2. In HClO4 with an UPL of 1.8 VRHE the anodic
contribution is reduced to around 37% (A1) and the cathodic rises to 63% (52 and 11% for
C2 and C1 respectively). Thus, with increasing UPL the transient dissolution of Pd switches
from an anodic process to a process dominated by Pd-oxide reduction. Moreover, at
potentials where the OER becomes relevant the C2 reduction/dissolution process becomes
dominant.
5.4 Comparison of Poly-Pd and Pd/C Electrodissolution
In order to estimate the value of the previous results obtained on poly-Pd for real
application, carbon supported Pd nanoparticles (Pd/C) are synthesized and analyzed (see
materials in Chapter 4). The Pd/C ink is printed on a GC plate obtaining spots that are
measured using the SFC-ICPMS. The initial activation CVs and the associated dissolution
are shown in Figure 5.10.
- Palladium Electrodissolution from Model Surfaces and Nanoparticles
| 67
Figure 5.10 a) Dissolution profiles of poly-Pd and supported Pd/C nanoparticles during 30
activation cycles with a scan rate of 200 mV s-1 and b) the corresponding cyclic voltammograms of
the Pd/C electrode in SFC. The Pd/C dissolution signal was normalized with the surface area after
activation.
In contrast with poly-Pd the dissolution rate of Pd/C is steadily decreasing during
activation. This is because, unlike for bulk material, the dissolution of nanoparticles along
with other degradation mechanisms lead to a decrease in surface area, evident from a
comparison between the first and the last CVs of the activation protocol (Figure 5.10b).
Table 5.3 Calculated charges of Pd-oxide reduction peaks and corresponding calculated areas
using 424 µC cm-2 as reduction charge per unit area.
CV Q /
µC
Ametal /
cm2
Initial Pd/C
Activated Pd/C
Activated poly-Pd
2.28
1.48
4.6
0.0054
0.0035
0.0109
The Pd/C dissolution measurement (Figure 5.11) follows the same protocol reported in
Figure 5.6 and is performed only in HClO4. The electrochemical determination of the
surface area through Pd-oxide reduction is convenient but not straightforward as it
requires a precise knowledge of the potential formation of 1 oxide monolayer (ML) reported
to be around 1.4-1.5 VRHE [150, 281]. To compare the electrochemical dissolution, the data
shown in Figure 5.11 are normalized by the Ametal, which is 0.0109 and 0.0035 cm-2 for poly-
Pd and Pd/C respectively. This is determined from the Pd-oxide reduction of the last
activation cycle ([0.1-1.4] VRHE), which directly precede the dissolution measurement. Pd-
oxide reduction and thus Ametal during activation of Pd/C decrease by ca. 35% indicating a
surface area change due to catalyst degradation.
Chapter 5
68 |
Figure 5.11 a) slow scans (2 mV s-1) with increasing UPL and b) the corresponding measured poly-
Pd and Pd/C dissolution profiles normalized by the real surface are (for Pd/C estimated after the
activation). Normalized CVs (a-inset) and the normalized mass dissolved per cycle (b-inset) are also
shown. Electrolyte: 0.1M HClO4.
At low potentials, CVs show one interesting difference between catalysts: unlike poly-Pd,
Pd/C does not show a large cathodic current and anodic peak corresponding to the H bulk
absorption (inset in Figure 5.11a). This behavior was already known in literature and was
reported to be size dependent [273, 274].
Potential sweeps to increasing UPL (0.9, 1.2, 1.5 VRHE) in HClO4 with a 2 mV s-1 scan rate
are applied (Figure 5.11). The same feature for poly-Pd, namely the presence of up to three
peaks in the dissolution profile is also observed for Pd/C. While the anodic peak A1 and
cathodic peak C2 well correspond, the cathodic dissolution C1 is shifted for Pd/C to lower
potentials (time delay in Figure 5.11b). The peak position generally depends on different
parameters such as the mass transfer of dissolved species out of the carbon matrix, the
flow rate and scan rate. While the last two are the same in both measurements, the
amount of printed Pd/C catalyst is so low that the mass transfer limitation can be
neglected. A more valuable explanation relates to the shift of the reduction to lower
potential for Pd/C (see inlet CVs in Figure 5.11a).
Considering the quantitative dissolution, it is observed a slightly higher dissolution per
electrochemical real surface area in the case of nanoparticulate Pd/C catalyst at all
considered potentials. Only few works are reported in the literature of nanoparticulate Pd
dissolution and to the knowledge of the author no on-line detection of dissolved Pd from
nanoparticles is reported. Generally they indicate influence of surface morphology,
geometry and particle sizes [173, 174]. Kumar et al. studying the anodic oxidation onset
potential in presence of chlorides suggested a size dependent destabilization of the
- Palladium Electrodissolution from Model Surfaces and Nanoparticles
| 69
nanoparticles compared to bulk Pd [174]. In our case, we do not see any significant
difference in dissolution onset potential between the two electrode systems, but the
dissolution profiles suggest a small difference in their behavior. Note however, that a
precise quantitative evaluation is rather challenging especially when dealing with
nanoparticulate catalyst. Indeed, i) the Ametal is determined with the same electrochemical
method for both catalyst even though the precise potential of formation of 1 oxide ML can
slightly change with surface morphology and geometry. Indeed, with same UPL the oxide
formation and reduction might be different from nanoparticles and bulk Pd [174]. ii) Ametal
of Pd/C might change during measurement in consequence of catalyst degradation (even
though measurements are limited to 3 cycles to minimize degradation). Furthermore, iii)
remaining PVA from synthesis might influence the dissolution (even though the washing
step is expected to remove it). Finally, iv) for carbon supported nanoparticles the catalyst
loading in the experiment might also play a role as shown recently by Keeley et al. [139].
Indeed, the authors showed for Pt/C that the specific dissolution (normalized per surface
area) is decreasing when the loading increases. This phenomenon was attributed to the
decreased diffusion of Pt ions into bulk solution as ions remain trapped in the porous
catalyst deposit when loading is higher.
5.5 Discussion on Pd Oxidation/Dissolution
The major experimental findings of this work can be summarized as follow:
The Pd dissolution is strictly correlated to the oxide formation and reduction.
However, no simple correlation could be established between the two processes.
Indeed, the dissolution onset potential in HClO4 appears to be around 50 mV higher
than in H2SO4, whereas the oxidation onset potential in HClO4 is slightly lower
(Figure 5.1, Figure 5.2, Figure 5.7);
Below 1.1 VRHE it is not possible to differentiate between anodic and cathodic
processes. Between 1.1 and 1.4 VRHE a cathodic dissolution related to the C1
reduction is observed. At more positive potentials a third dissolution peak,
corresponding to the broad C2 reduction, appears and it increases dramatically with
the UPL (Figure 5.2 and Figure 5.6).
Increasing the UPL, the oxide coverage increases. Therefore, while transient anodic
dissolution initially increases with UPL, in the 1.4-1.7 VRHE potential range the
formed oxide protects Pd from increasing dissolution. Beyond 1.7-1.8 VRHE anodic
dissolution increases again in correspondence to the OER region (Figure 5.4);
Unlike for anodic dissolution the cathodic dissolution increases almost linearly with
UPL (Figure 5.4), becoming the dominant process for potential higher than 1.7
VRHE. Furthermore, its onset and maxima shifts to lower potentials with increasing
Chapter 5
70 |
UPL (Figure 5.8), in accordance with the shift of the cathodic C1 and C2 reduction
peaks (Figure 5.3);
Pd dissolves much more than Pt and Au and the dissolution depends strongly on the
scan rate. Pd dissolution in sulfuric acid was found to be 5 times higher than in
HClO4 (Table 5.1 and Table 5.2);
In potentiostatic experiments below 1.6 VRHE the dissolution rate decreases with
time, indicating the passivation of the surface (Figure 5.5);
All these findings are additionally validated for a carbon supported high-surface
area Pd/C nanocatalyst, which is more interesting for application. A slightly small
increase in dissolution per real surface are is observed for Pd/C (Figure 5.10 and
Figure 5.11).
With our findings we confirmed the close connection between the Pd oxidation states and
its transient dissolution, which was already observed for Au and Pt electrode materials
[148]. Indeed, the electrochemical oxidation and the dissolution of Pd have similar
standard potentials. Pourbaix expressed the oxidation of Pd as [151]:
𝑷𝒅 + 𝑯𝟐𝑶 → 𝑷𝒅𝑶 + 𝟐𝑯+ + 𝟐𝒆− 𝑬𝒐 = 𝟎. 𝟗𝟏𝟕 + 𝟎. 𝟎𝟓𝟗𝟏𝐥𝐨𝐠 [𝑯+] Equation 5.1
or 𝑷𝒅 + 𝟐𝑯𝟐𝑶 → 𝑷𝒅(𝑶𝑯)𝟐 + 𝟐𝑯+ + 𝟐𝒆− 𝑬𝒐 = 𝟎. 𝟖𝟗𝟕 + 𝟎. 𝟎𝟓𝟗𝟏𝐥𝐨𝐠 [𝑯+] Equation 5.1b
𝑷𝒅𝑶 + 𝑯𝟐𝑶 → 𝑷𝒅𝑶𝟐 + 𝟐𝑯+ + 𝟐𝒆− 𝑬𝒐 = 𝟏. 𝟐𝟔𝟑 + 𝟎. 𝟎𝟓𝟗𝟏𝐥𝐨𝐠 [𝑯+] Equation 5.2
or 𝑷𝒅(𝑶𝑯)𝟐 + 𝟐𝑯𝟐𝑶 → 𝑷𝒅(𝑶𝑯)𝟒 + 𝟐𝑯+ + 𝟐𝒆− 𝑬𝒐 = 𝟏. 𝟐𝟖𝟑 + 𝟎. 𝟎𝟓𝟗𝟏𝐥𝐨𝐠 [𝑯+] Equation 5.2b
And the dissolution of Pd can be described as [151]:
𝑷𝒅 → 𝑷𝒅𝟐+ + 𝟐𝒆− 𝑬𝒐 = 𝟎. 𝟗𝟖𝟕 + 𝟎. 𝟎𝟐𝟗𝟓𝐥𝐨𝐠 [𝑷𝒅𝟐+] Equation 5.3
𝑷𝒅𝑶𝟐 + 𝟒𝑯+ + 𝟐𝒆− → 𝑷𝒅𝟐+ + 𝑯𝟐𝑶
𝑬𝒐 = 𝟏. 𝟏𝟗𝟒 + 𝟎. 𝟏𝟏𝟖𝟐 𝐥𝐨𝐠[𝑯+] − 𝟎. 𝟎𝟐𝟗𝟓[𝑷𝒅𝟐+]
Equation 5.4
𝑷𝒅𝑶 + 𝟐𝑯+ → 𝑷𝒅𝟐+ + 𝑯𝟐𝑶 𝐥𝐨𝐠[𝑷𝒅𝟐+] = −𝟑. 𝟎𝟐 + 𝟐 𝐥𝐨𝐠[𝑯+] Equation 5.5
or 𝑷𝒅(𝑶𝑯)𝟐 + 𝟐𝑯+ → 𝑷𝒅𝟐+ + 𝟐𝑯𝟐𝑶 𝐥𝐨𝐠[𝑷𝒅𝟐+] = −𝟐. 𝟑𝟓 + 𝟐𝐥𝐨𝐠 [𝑯+] Equation 5.5b
As anticipated, despite the large amount of literature and the variety of methods applied,
several aspects of Pd electro-oxidation are still poorly understood, such as the chemical
composition, thickness and adsorption behavior of Pd oxide layers [150]. In particular,
- Palladium Electrodissolution from Model Surfaces and Nanoparticles
| 71
there are some relevant issues in the literature that require additional research [150]: i)
the first product formed during oxidation was considered by several authors to be
Pd(OHads) [154, 158, 159, 167, 267, 282], while other authors suggested the formation of
Pd(II)-oxide/hydroxide species such as PdO [150, 283] or Pd(OH)2 [151]; ii) the potential for
the formation of an oxide monolayer (generally reported to occur around 1.45-1.5 VRHE) is
unclear, as is iii) the onset potential for the formation of higher oxidation species (i.e
Pd(IV)-oxide, thicker β Pd(IV)-oxide in the OER region); iv) the presence of subsurface
oxygen is claimed by some groups to play an important role in the reactivity and stability
of the metal [152, 161, 284] and v) it is not obvious if anhydrous/hydrous oxide is present or
not at different potentials. Concerning the last point, we will consider both reactions
(Equations 5.1-5.1b and 5.2-5.2b), but in the following discussion we will rather talk of
Pd(II)- and Pd(IV)-oxides.
In the literature two Pd reduction peaks are reported: i) a well-defined reduction peak at
lower potentials labeled here as C1 (Figure 5.1) that corresponds to the reduction of Pd(II)-
oxide (Equation 5.1-5.1b) and ii) a second broad reduction peak around 1.2-1.3 VRHE (here,
C2), which is thought to correspond to the reduction of Pd(IV)-oxide formed at high
potentials (>1.3 VRHE), slightly below the OER onset [150, 267] (Equation 5.2-5.2b). This
higher oxidation state was confirmed with XPS measurement by Chausse et al. [285].
Zhang et al. and Birrs et al. showed independently that a thick Pd “β hydrous oxide” [266,
280, 286] is only formed at very large anodic polarization (higher than the OER onset) and
its reduction is correlated to several peaks in the low potential region, around HUPD [150,
152, 280]. In our experimental results no peaks of this kind are observed up to 1.8 VRHE,
therefore the presence of a thicker hydrous oxide layer (elsewhere referred as β oxide [150])
can be safely excluded from the following considerations, at least for potentials up to 1.7
VRHE.
According to the literature and to CVs one would expect already some Pd dissolution in
parallel with the initial Pd oxidation, namely around 0.7 VRHE and 0.75 VRHE in HClO4 and
H2SO4 respectively (Figure 5.1). However, a small deviation from the background signal is
only observable first with an UPL above 0.8-0.85 VRHE, close to the thermodynamically
predicted standard potential for Pd metal electro-dissolution (E0(Pd/Pd2+) = 0.987 V +
0.0295 log(Pd2+)), which assuming a reasonable Pd2+ concentration of 1 nmol dm-3 would be
approximately 0.72 VSHE (0.78 VRHE at pH=1). Experimentally, there is a more than 100 mV
shift for the dissolution onset in comparison to oxidation. A similar difference was already
observed for Pt dissolution and it was tentatively related to the ICPMS detection limit.
Recently, a modified SFC configuration allowed the accumulation of dissolved Pt. Thus,
dissolution was measured also at potential, close to the Pt oxidation onset [133]. In the
case of Pd this difference could be attributed either to the ICPMS detection limit (as for Pt)
or to the higher standard potential of the Pd electro-dissolution compared to the Pd
Chapter 5
72 |
oxidation. Furthermore, the dissolution onset potential in HClO4 appeared to be around 50
mV higher than in H2SO4. This is somehow contradictory with the oxidation onset
potential which in HClO4 is lower (Figure 5.1). Therefore, no simple correlation between
oxidation and dissolution is established, as previously observed for Au in both acids [175].
Interestingly, despite exhibiting similar features, the actual measured Pd dissolution in
the two electrolytes is quantitatively very different. Indeed, Pd in H2SO4 is dissolving at
rates approximately 5 times higher than in HClO4 (Table 5.1). Furthermore, comparing the
dissolution in the same medium (H2SO4), it turns out that Pd is dissolving at a much
higher rate than other noble metals like Pt and Au. Already Rand and Woods [153]
reported Pd dissolution to be approximately 30 times higher than Pt in H2SO4, in good
agreement to our results. Much higher dissolution of Pd compared to Pt was also observed
by Łukaszewski et al. [171]. Burke et al. [152] affirmed that this marked behavior is
related to the ionic radii difference of the respective cations. In fact, the electrostatic field
around smaller Pd cations is stronger, which leads to more stable Pd complexes and a
stronger solvation shell [150], resulting in the observed enhancement in Pd electro-
dissolution. The observed dissolution difference in the two electrolytes could be attributed
to a difference in the amount of oxide formed. Effectively, the UPL being equal, the
measured Pd reduction charge in H2SO4 is visibly higher (Figure 5.3), suggesting more
oxide formation. However, the difference in the reduction charges is only up to ca. 20%
(Figure 5.3c). Therefore, different dissolution behavior could be originated by the different
nature of the anions in the electrolyte. In literature, many works reported enhanced
electro-dissolution in presence of chlorides and iodides [150, 157, 158], however only few
works reported differences between HClO4 and H2SO4, the latter being the sole choice of
electrolyte for most of the experimental studies. Recently, Grdeń et al. [150] reviewed
several Pd studies and classified anions on the basis of their Pd electro-dissolution
promotional effect as follows: ClO4- < HSO4
-/SO42- < Cl- < I-. Anions like Cl- and I- form
stable Pd-anion complexes that can lead to an increase in dissolution [150]. Solomun,
studying the role of anions in H2SO4 and HClO4 with LEED and XPS, suggested that the
adsorbed anion can weaken the Pd-Pd surface bonds [150, 162]. They also proposed that
the adsorption of HSO4-/SO4
2- in the early stages of surface oxidation facilitates the
interfacial place exchange [160-162], thus resulting in enhanced Pd dissolution in the case
of HSO4-/SO4
2-, as confirmed with our experimental findings. Furthermore, the dissolved
Pd2+ can form in acidic electrolytes stable complexes, that if on the one hand can explain
the enhanced electro-dissolution of Pd compared to Au and Pt, on the other hand can be at
the origin of the different electro-dissolution in HClO4 and H2SO4.
Even though the absolute amount of dissolved Pd per cycle is quite different in the two
electrolytes (see Table 5.1), the percentage contribution of the different dissolution peaks
follows qualitatively the same trend. i) Below 1.1 VRHE only one peak is present, as at these
- Palladium Electrodissolution from Model Surfaces and Nanoparticles
| 73
low potentials it is not possible to distinguish between anodic and cathodic dissolution. ii)
Between 1.1-1.4 VRHE a dissolution peak corresponding to the cathodic reduction C1 is
appearing and becoming more and more important. This dissolution peak is observed in
literature with different techniques as RRDE [154, 163], CFDE [149] and quarzt
microbalance [170, 171] and is assigned to Pd(II)-oxide reduction [150]. However, the
Pd(II)-oxide can only undergo chemical dissolution (Equations 5.5-5.5b), which is generally
disregarded for other metals like Pt. The Pd solubility is higher than that of Pt and this
could mean that, unlike for Pt, the chemical dissolution might play a role for Pd.
Nevertheless, the experimental results indicate an existing correlation between the Pd(II)-
oxide reduction (C1) and the dissolution peak, which cannot be easily explained only with
chemical dissolution. Therefore, the dissolution can be attributed to the reduction and
desorption of adsorbed oxygen species that causes de-passivation (Equation 5.3). iii) At 1.4
VRHE a second cathodic dissolution peak corresponding to the broad Pd(IV)-oxide reduction
(C2) is observed (Equations 5.2-5.2b). Even though the integration of such a broad peak is
not easy, we can safely say that even at high UPL the amount of Pd(IV)-oxide formed is
less than the amount of Pd(II)-oxide formed (C2 reduction charge density is much smaller
than C1 reduction charge density as shown in Figure 5.3). On the other hand, the amount
of dissolved Pd related to Pd(IV)-oxide reduction (C2) is much larger than the dissolved Pd
related to Pd(II)-oxide reduction (C1) (Figure 5.4). (iv) Above 1.7 VRHE the cathodic
dissolution overall exceeds the anodic dissolution. In particular, at 1.8 VRHE the cathodic
dissolution associated to the Pd(IV)-oxide reduction (C2) becomes the dominant dissolution
mechanism.
Interesting is the trend of the transient anodic Pd dissolution with different UPLs as
shown in Figure 5.4, where different potential regions can be observed in the two
electrolytes. Above 0.9 VRHE Pd oxidation to Pd(II)-oxide (Equations 5.1-5.1b) and Pd metal
dissolution to Pd2+ (Equation 5.3) are proceeding in parallel and upon an increase in UPL
the transient anodic dissolution increases. Between 1.4 and 1.7 VRHE no increase in
transient anodic dissolution is observed. This can have two reasons: i) Around 1.3-1.4 VRHE
a complete monolayer of Pd(II)-oxide is formed, thus preventing further Pd metallic
dissolution (through Equation 5.3). In the literature, different studies generally agree that
the complete formation of a monolayer occurs between 1.4-1.5 VRHE [150]. However, in this
case the chemical dissolution of Pd(II)-oxide (Equations 5.5-5.5b) would still be present in
contrast to the observed passivation. Therefore, either the chemical dissolution can be
disregarded (as for Pt), or the passivation arises from ii) the formation of a top layer of
chemically stable Pd(IV)-oxide, which is reported to start, as mentioned above, also around
1.3-1.4 VRHE. However, if the Pd(IV)-oxide would cover completely the Pd surface one would
expect much higher Pd(IV)-oxide reduction charges (peak C2). Therefore, we suggest that
the kinetics of the Pd(II)-oxide chemical dissolution (Equations 5.5-5.5b) is too slow and the
Chapter 5
74 |
associated dissolution products are below the ICPMS detection limit. In this sense, the
contribution of Equations 5.5-5.5b is neglected in the following mechanistic discussion and
the observed passivation between 1.4 and 1.7 VRHE can be explained with the formation of a
complete monolayer of Pd(II)-oxide. At more positive potentials, the amount of anodically
dissolved Pd increases again. The origin of this behavior is not clear yet and should be
further investigated. However, this could be attributed to i) evolution of oxygen (as
observed for different metals [144]) and/or to ii) changes in the oxide structure from a thin
α Pd oxide to a thick, hydrous, porous β Pd oxide [157, 280, 286] and/or to iii) formation of
Pd(VI)-oxides [150, 151]. Indeed, the last two are reported to take place above the OER in
acidic media.
5.6 Proposed Pd Dissolution Mechanism
Even though the precise nature of Pd oxide is still unresolved, we showed that its
dissolution process can be safely ascribed to surface processes involving different oxidation
states and the changes between them. Additional work needs to be done to describe
precisely the transient Pd dissolution. Nevertheless, a tentative mechanism can be derived
from our experimental observations (Figure 5.12).
Figure 5.12 Proposed model of the transient dissolution of Pd. A1: from double layer region to Pd-
oxide (both Pd(II) and Pd(IV) oxidation states depending on the UPL). C2: reduction of Pd(IV)-
oxide to Pd(II)-oxide and Pd metal (with dissolution). C1: reduction of Pd(II)-oxide to Pd metal.
- Palladium Electrodissolution from Model Surfaces and Nanoparticles
| 75
The main contribution to the anodic dissolution (related to oxidation peak A1) comes
from metal Pd dissolution to Pd2+ (Equation 5.3), that is proceeding in parallel with surface
oxidation to Pd(II)-oxide (Equations 5.1-5.1b). The formed Pd(II)-oxide can be chemically
dissolved (Equations 5.5-5.5b), yielding other Pd2+, however as discussed earlier its
contribution is neglected. As the potential increases Pd(II)-oxide oxidizes to Pd (IV)-oxide
(Equations 5.2-5.2b). Pd passivates (geometrically and/or electrochemically) once the first
oxide monolayer is formed (no increase in transient anodic dissolution). The formed Pd-
oxide film is rather complex and depends strongly on the UPL. Nevertheless, we suggest a
possible general composition. For UPLs in the 0.7-1.4 VRHE potential range, the formation
of more Pd(II)-oxide (Equations 5.1-5.1b) is favored over the formation of Pd(IV)-oxide and
a monolayer Pd(II)-oxide is obtained around 1.4 VRHE. Once the potential is raised above
1.4 VRHE the formation of Pd(IV)-oxide becomes thermodynamically favorable and a layer of
surface Pds(IV)-oxide forms on top.
During the cathodic scan, first the Pds(IV)-oxide is reduced back to Pd(II) (Equations 5.2-
5.2b) (C2 reduction peak) or dissolved to Pd2+ through the electrochemical reaction
(Equation 5.4) yielding the first cathodic dissolution peak. This peak is only obtained
when the UPL is high enough that Pds(IV)-oxide is formed (Equations 5.2-5.2b).
Furthermore, Equation 5.4 is dependent on both the pH and the amount of oxide formed.
Thus, it can nicely explain the steep increase with the UPL of the amount of dissolved Pd
related to this first cathodic dissolution peak. Indeed, it becomes the dominant dissolution
mechanism above 1.7 VRHE, where more Pd(IV)-oxide is formed.
A second cathodic dissolution (related to the reduction peak C1) is observed at lower
potentials where Pd(II)-oxide reduction (Equation 5.1-5.1b) takes place. During transient
conditions, the mechanism of Pd ions production is not well understood. In many past and
recent works, this dissolution was related to Pd(II)-oxide reduction yielding Pd2+ [149].
Based on electrochemical equilibria [151] Pd(II)-oxide could dissolve in a chemical
reduction, which as discussed earlier can be disregarded. It has been suggested elsewhere
for Au and other noble metals that the dissolution during the negative direction scan is due
to the de-passivation of the oxide, resulting in the dissolution of the exposed metal ion
[148]. Effectively, assuming a reasonable Pd2+ concentration of 1 nmol dm-3, from the
dissolved amount of Pd, the equilibrium potential for the Pd metal electro-oxidation
(E0(Pd/Pd2+) = 0.987 V + 0.0295 log(Pd2+), in Equation 5.3) would be approximately 0.72
VSHE (0.78 VRHE). At such potential of the Pd(II)-oxide would be already partially reduced
and thus free Pd metal would be exposed to the electrolyte and be available for dissolution.
Still, the estimated equilibrium potential of Equation 5.3 is higher compared to the lowest
potential at which dissolution was detected. This could be simply an effect of i) mass
Chapter 5
76 |
transport limitation and/or ii) due to the presence of defects and adatoms formed during
the oxide reduction whose equilibrium potential can differs from that of the bulk. As
another possible contribution to this second cathodic dissolution peak we suggest that
some remaining small amount of bulk Pdb(IV)-oxide embedded in the Pd(II)-oxide layer
might play a role. Indeed, when Pd(II)-oxide is reduced back to Pd metal the remaining
Pdb(IV)-oxide can dissolve in a non-reversible process through Equation 5.4. In summary,
this second cathodic dissolution peak can be explained by assuming a direct dissolution of
the Pd metal and/or a dissolution of a remaining Pd(IV)-oxide. Both explanations well
match the correspondence of the Pd(II)-oxide reduction peak C1 and the dissolution
measured with ICPMS.
5.7 Conclusion
In conclusion, despite the uncertainty and complexity of the Pd oxidation states and
mechanism, in this Chapter a model for the transient Pd dissolution based on the unique
SFC-ICPMS experimental results has been proposed. This model is not only suitable for
ideal poly- Pd, as experimental results confirm its validity also for supported high-surface-
area catalysts, which despite their major interests for application were not studied
previously. Therefore, these findings will be of interest for future studies on Pd and Pd-
based alloys degradation in real applications.
While the proposed mechanism well explains the observed dissolution trends, still some
unresolved questions remain open and will need further investigations. First, the lack of a
precise knowledge of the chemical species formed at the Pd surface represents an obstacle
for a full understanding of Pd dissolution. Secondly, the role of the transition between thin
α oxide and thick β hydrous oxide formed at very high anodic polarization or the formation
of Pd(VI) oxide and its relevance for the transpassive region could not be fully clarified.
Finally, the influence of parameters such as temperature, the presence of anions and
cations in different electrolytes and the nanoparticle size needs further investigation.
- Addressing Stability of Bimetallic Electrocatalysts: the Case of Au-Pd Alloys
| 77
- Addressing Stability of Bimetallic Chapter 6
Electrocatalysts: the Case of Au-Pd Alloys13
——————————————————————————————————————————
Bimetallic catalysts are known to often provide enhanced activities compared to the pure
metals, due to electronic, geometric and ensemble effects. Au-Pd catalysts in particular are
known for higher selectivity towards H2O2 compared to the more active Pd catalyst.
However, applied catalytic reaction conditions may induce re-structuring, metal diffusion
and dealloying. These can result in a drastic change in surface composition, thus limiting
the applicability of bimetallic catalysts in real systems as a consequence of performance
degradation (i.e. decrease in selectivity towards H2O2). Following the work on Pd
dissolution (Chapter 5), this Chapter is dedicated to the study of dealloying using an Au-Pd
bimetallic nanocatalyst as a model system. The changes in surface composition over time
are monitored in-situ by cyclic voltammetry while Au and Pd dissolutions are measured
on-line with ICPMS. It is demonstrated how experimental conditions such as different
acidic media (HClO4 and H2SO4), different gases (Ar and O2), UPL and scan rate
significantly affect the partial dissolution rates and consequently the surface composition.
The understanding of these alterations is crucial for the determination of fundamental
catalyst activity and selectivity (see also the following Chapter 8), and plays an essential
role for real applications, where long term stability is a key parameter.
——————————————————————————————————————————
13
Parts of this chapter have been already published in:
E. Pizzutilo*, S.J. Freakley, S. Geiger, C. Baldizzone, A. Mingers, G.J. Hutchings, K.J.J. Mayrhofer, S. Cherevko
Catal. Sci. Technol. 2017, 7, 1848-1856.
There are therefore numerous verbal quotes from that publication. Some of the figures present in
the publication have been re-printed or modified.
Chapter 6
78 |
6.1 Au and Pd Dissolution Onset Potentials
The transient dissolution of Au, Pd and AuPd alloy is studied by utilizing the SFC-ICPMS
(Figure 6.1). As discussed also for poly-Pd in the previous Chapter, the Au and Pd
dissolution during dynamic potential operation are initiated by the formation of the
respective surface oxides (Au-O and Pd-O) as confirmed from the initial CVs (Figure 4.5).
Figure 6.1 a) Pd and Au dissolution profiles of 4 printed layers of AuPd catalyst during a CV [0.05-
1.5] VRHE in Ar purged 0.1 M HClO4 with a scan rate of 2 mV s-1. Separately, the Au and Pd
dissolution of the alloyed AuPd catalyst are shown in comparison with that of the pure metals (Pd
in b) and Au in c) respectively). The dotted lines represent the pure metal. The corresponding
integrated dissolution is shown in the respective inset. Flow rate is 193 μL min−1.
The prepared catalyst colloidal ink is printed on a GC plate, resulting in an array of
samples that can be measured using the SFC with an opening of around 1 mm in diameter.
For this measurement 4 printed catalyst layers are used to better identify the dissolution
onset potentials, since the deviation from the background signal is easier to be observed
- Addressing Stability of Bimetallic Electrocatalysts: the Case of Au-Pd Alloys
| 79
when more catalyst is used. Note that the number of layers can influence the specific
dissolution [139], whereas it does not influence the onset potential. The loading of a single
printed layer is estimated from the droplet size and is approximately 2 ng. From the
statistical average particle size and the loading, the total surface area per deposited layer
is calculated (Table 6.1; see calculation from Chapter 3). The Ametal is used to normalize the
dissolution during the first cycle (Figure 6.1) and not for the other figures, since during a
degradation measurement the nanoparticle surface area and its composition are steadily
changing.
Table 6.1 Particle size and specific surface area of the prepared materials investigated in this
study.
Ametal (1l)** /
mm2
Au
AuPd
Pd
0.14
0.18
0.31
*ECSA refers to the catalyst specific surface area, which was calculated from the particle mean
size; **At refers to the total surface area of per deposited layer (≈2 ng).
For a better comparison with the pure metal counterparts in Figure 6.1b,c Au and Pd
dissolutions are shown separately: the full and the dotted lines represent the alloy and the
pure metal, respectively. The measured dissolution onset potentials are defined as the
deviation from the background signal in the positive scan (see Figure 6.2). These are
respectively ≈0.78 VRHE for pure Pd and ≈1.3 VRHE for pure Au. The value for pure Pd is in
accordance with measurements on poly-Pd, while a previous study on poly-Au in HClO4
showed values slightly higher than our Au nanoparticles [144].
Figure 6.2 Comparison of dissolution onset potentials of a) Pdpure and Pdalloyed nanoparticles and b)
Aupure and Aualloyed nanoparticles.
In Figure 6.1c the Au profile presents the typical two peaks corresponding to dissolution
during anodic and cathodic scan. Several mechanisms of Au-O formation and dissolution
have been already thoroughly described, although the exact reaction pathway is still not
clarified [148, 175].
Chapter 6
80 |
As already observed for poly-Pd, the dissolution of nanoparticulate Pd per cycle (≈117 ng
cm-2) is more extensive than Au and other noble metals [85, 153]. Moreover, instead of two
separate peaks resulting from oxidation and reduction the presence of a third and
sometimes fourth peaks indicates that additional processes, clarified in Chapter 5 [85],
play an important role. These can be related to the complex structure of Pd oxides, the
oxidation state of Pd and the oxide’s chemical composition (Pd(II)-oxide [150] and at higher
anodic potentials Pd(IV)-oxide [150, 287]).
Concerning the behavior of the alloyed metals, controversial results regarding increased
stability of metals in alloyed nanoparticles compared to the pure metals are reported [67,
288-291]. In fact, some theoretical DFT calculations claim that the presence of an alloying
element would induce a shift in the oxidation and dissolution potential, thus leading to a
stabilization of the alloy. This is possibly related to a delayed coverage of O* and OH*
intermediates. Often, the doping of Au is reported to have a positive effect in the
stabilization of other noble metals such as Pt [292]. Recently, however, Cherevko et al.
showed that a Pt sub-monolayer on poly-Au is not stable, but rather shows significant
dissolution of both Au and Pt similar to the pure polycrystalline elements [293]. In our
case, the Au and Pd dissolved masses in the alloy normalized by the Ametal (insets in Figure
6.1) are in absolute terms approximately half of those for the pure metals (for Pd is ≈60%,
for Au ≈50%). Considering that the nominal stoichiometry is 50% Au and 50% Pd, and
assuming that the initial surface composition does not differ significantly, this suggests
that the dissolution normalized by the respective surface area in the alloys is
approximately in line with the pure metals. Nevertheless, a possible non-homogeneity of
the alloys and the difficulty in estimating the real surface composition make the
interpretation of the results rather challenging. A study on a model surface would be
therefore recommended to confirm/exclude the effect of Au on the overall dissolution per
cycle.
Interestingly, however, the dissolution onset potential of Pd in the alloyed catalyst (Figure
6.2a) is slightly higher (approximately 30 mV higher around 0.81 VRHE) compared to pure
Pd. Similarly, Cherevko et al. showed that the onsets of Pt and Au dissolution after
intermixing shift to slightly higher potential than the pure elements [293]. Furthermore, in
line with oxide reduction peak shift (Figure 4.5), the cathodic Pd dissolution of the alloyed
material ends significantly earlier, one more time confirming a correlation between Pd-
oxide reduction and cathodic dissolution processes. Therefore, alloying influences clearly
the dissolution onset and final potentials, whereas no significant effect in the quantitative
dissolution is observed.
- Addressing Stability of Bimetallic Electrocatalysts: the Case of Au-Pd Alloys
| 81
6.2 Influence of Upper Potential Limit
Changing the UPL or the scan rate has no influence on the dissolution onset potentials.
Nevertheless, as shown for polycrystalline metals, the rate of dissolution and the shape of
the dissolution peaks and profiles for the Au-Pd alloy are strictly related to the UPL (see
Figure 6.3) and scan rate. Significant dissolution signal is observed only when the
potential is above 1.0 VRHE and 1.4 VRHE for Pd and Au respectively. Below a certain
potential only a single peak is discernible, while at higher potentials two distinct peaks
(corresponding to anodic and cathodic dissolution) are present. This was already observed
in the case of poly-Au dissolution [147] and it is probably due to the enhancement of anodic
dissolution. The amount of dissolved Au and Pd in every cycle, which corresponds to the
area under the dissolution profiles, is increasing with the UPL (inset of Figure 6.3b). Note
that at higher scan rates, it is not possible to distinguish the two cathodic peaks even at
higher UPL as visible already from the cycles at 10 mV s-1 in Figure 6.4. The overlap
between dissolution peaks at higher scan rates was previously observed for poly-Pt and
related to the technical limitations of the setup [132].
Figure 6.3 a) Several cycles to different upper limit potentials (ULP) with 10 mV s-1 for AuPd
catalyst (1 layer) in 0.1M HClO4. b) Corresponding Au and Pd dissolution profiles and the amount
of dissolved metal per cycle (inset).
Considering the dissolution onset potentials, it is possible to define a stability window for
bimetallic nanoparticles like AuPd catalyst: below the Pd onset potential (≈0.8 VRHE)
virtually no metal is being leached out from the catalyst surface, so that the composition
remains unchanged. Above 0.8 VRHE severe dissolution of Pd and Au occurs, which leads to
changes in surface composition and long-term degradation of the catalyst. These
considerations of course are not taking in account surface restructuring and metal
migration, which might occur even at low potentials.
Chapter 6
82 |
6.3 Influence of Electrolytes
Online dissolution under cycles in the range [0.1-1.6 VRHE] are recorded in Ar purged 0.1 M
HClO4 and H2SO4 (Figure 6.4) to characterize the changes in dissolution rate and surface
composition of the particles over time. Such a high overpotential is chosen in order to (i)
accelerate the degradation and to (ii) follow the Au reduction peak that is visible only with
scans to high potentials. In Figure 6.4 the dissolution is not normalized to the Ametal, since
area and surface composition change during the measurement due to dissolution.
Figure 6.4 Dissolution profiles of AuPd nanoparticles (1 layer) in Ar purged (b) 0.1M HClO4 and (c)
0.1M H2SO4 during 50 cyclic voltammograms between 0.1 and 1.6 VRHE with a scan rate of 200 mV s-1;
some CVs at slower scan rate (10 mV s-1) were recorded to plot the dissolution cycle profiles with
time (insets in b-c).
Comparing the Pd dissolution profiles during the first cycle in HClO4 and H2SO4 (Figure
6.5a) it is possible to clearly conclude that the second is promoting the dissolution more,
whereas Pd dissolution onset potential is approximately the same in both electrolytes. The
behavior of Pd seems to be similar to the behavior of Pt. Indeed, no significant variation in
the onset potential with pH or amount of sulfate or perchlorate anions was found
- Addressing Stability of Bimetallic Electrocatalysts: the Case of Au-Pd Alloys
| 83
previously also for poly-Pt [132]. On the other side, our group showed that for poly-Au
there is a shift of almost 100 mV in the Au dissolution onset potential (≈1.3 VRHE in H2SO4
and ≈1.4 VRHE in HClO4) [175]. The Au in the alloyed AuPd (Figure 6.5b) exhibits also a
small shift in dissolution onset potential of approximately 50 mV. However, the Au
dissolution in both cases seems to start slightly earlier (≈ 1.25-1.30 VRHE) than poly-Au.
Figure 6.5 Comparison of a) Pd and b) Au dissolution profiles of AuPd nanoparticles (1 layer) in Ar
purged electrolyte 0.1M HClO4 (full line) and 0.1M H2SO4 (dotted line) during the first cyclic
voltammogram between 0.1 and 1.6 VRHE with a scan rate of 10 mV s-1 (see Figure 6.4).
It seems that in HClO4 for the first cycle anodic dissolution is more relevant, while in
H2SO4 is more important the cathodic dissolution. However, since the first cycle is the as
prepared catalyst without any activation it is difficult to draw any conclusion. In both
cases after a determined number of cycles Pd is completely dissolved and we have a Au
enriched surface as also the trend in Pd dissolution suggest (Table 6.2).
Table 6.2 dissolved Pd and Au during the slow CVs (10 mV s-1) shown in Figure 6.4.
NCV Pd / ng Au / ng
HClO4 H2SO4 HClO4 H2SO4
1
10
30
50
0.122
0.034
0.013
0.006
0.233
0.011
0.002
0.001
0.027
0.014
0.013
0.012
0.068
0.024
0.019
0.018
CV profiles corresponding to the slower cycles (1, 5, 10, 20, 30 and 50) of the protocol
showed in Figure 6.4 are reported in Figure 6.6. The surface and its composition are
changing rapidly: the Pd-O reduction peak decreases, while the Au-O increases in
magnitude during CVs. At the same time, the amount of dissolved Pd is constantly
dropping as a consequence of the decrease in surface Pd. The charges associated with the
characteristic Au-O and Pd-O reduction peaks in the profile are proportional to Au and Pd
surface areas, respectively. However, as previously discussed, the extrapolation of surface
area in alloys is ambiguous, therefore we simply report the associated reduction charges
(insets in Figure 6.6), which in any case are represent clearly the trend of Pd and Au
surface areas.
Chapter 6
84 |
Figure 6.6 RDE CVs in Ar purged 0.1M HClO4 and 0.1M H2SO4 corresponding to the measurement in
Figure 6.4 are shown in a) and b) respectively. The relative Pd and Au oxide reduction charges are
displayed in the insets.
Interestingly, the Pd dissolution throughout the measurement is not simply decreasing
quantitatively, but also the profile is changing as shown in the comparison of the cycles
with slower scan rates (inset Figure 6.4b,c). Indeed, the anodic dissolution onset potential
is shifting positively from 0.9 VRHE of the first cycle to approximately 1.0 VRHE. Similar
positive shift was observed for sub-monolayer of Pt@Au dissolution [293]. During the
cathodic scan, the dissolution maxima (only one peak is distinguishable at this scan rate)
as well as the dissolution final potential are slightly shifting to higher potentials. This is
correlated to the decrease in Pd content with dissolution, which produces a more “intimate”
mixed alloy with finely dispersed Pd in the Au matrix. Indeed, the Pd-O reduction peak
potential in AuPd alloys is strictly correlated to the Pd content (Figure 6.6 and Figure 4.5):
the less Pd is present in the alloy, the higher the potential for Pd-O reduction is [105],
which explains the change in dissolution maxima potentials.
The amount of dissolved metal per cycle is changing significantly with the electrolytes.
Indeed, during the first cycle Pd is dissolved more in H2SO4 (≈0.24 ng) compared to HClO4
(≈0.13 ng). This difference in Pd removal is mirrored in the recorded CVs (Figure 6.6) by a
faster decrease of the Pd-O reduction peak [110]: in H2SO4 Pd disappears after 10 CVs
- Addressing Stability of Bimetallic Electrocatalysts: the Case of Au-Pd Alloys
| 85
whereas in HClO4 it is still detectable after 50 CVs. Therefore the enhanced
electrodissolution of Pd depends on the nature of anions present in the electrolyte, as they
facilitate the formation of products and/or intermediates [85, 150]. In particular, based on
the results reported in literature, H2SO4 promotes Pd electrodissolution more [85].
For sake of comparison, the same dissolution protocol has been performed also with the
pure metal counterparts.
Figure 6.7 Dissolution profiles of a) pure Pd and b) pure Au nanoparticles (1 layer) in Ar purged
0.1M HClO4 during 50 cyclic voltammograms between 0.1 and 1.6 VRHE with a scan rate of 200 mV
s-1; corresponding CVs are shown in c) and d), respectively.
In Figure 6.7 are shown the dissolution profile and CVs of pure Au and pure Pd in 0.1M
HClO4 (with the same protocol shown in Figure 6.4 for AuPd nanoparticles). In both cases
during the degradation protocol the oxide reduction peaks are decreasing as well as the
dissolution rates (in particular for Pd which is dissolving more). This is because the total
surface area is decreasing due to the dissolution.
6.4 Au-skin Formation Following Dealloying
Concluding our observation, the CVs of the Au-Pd alloys change significantly during
continuous potential cycling in acidic media to sufficiently high potentials, as described in
earlier reports [110, 126, 241, 278, 294]. In literature this was attributed to i) Au migration
to the surface [110], to ii) potential dependent Pd surface segregation [122] or to iii)
selective Pd removal [241]. Our results show that the main reason for surface Au
Chapter 6
86 |
enrichment is dealloying. Of course the other two, especially surface diffusion of Au atoms
[295], cannot be completely excluded, however their role in this process is considered minor
compared to dissolution. Indeed, Pd is dissolving at a much higher rate than Au; thus, Au
is being increasingly exposed to the surface, hence forming a gold “skin” (Figure 6.8).
Though, some isolated Pd atoms might still be present on the catalyst surface. Indeed,
ICPMS measurements of the degraded catalyst show a final Pd/Au ratio of 30/70 mol%
after 50CVs to 1.6 VRHE in HClO4 (much lower in H2SO4), thus confirming the presence of
Pd in the core even after the degradation measurement.While this might be positive for
materials like PtM used for ORR leading to an increased Pt surface, in applications where
the surface metal composition is crucial for activity and selectivity, dealloying needs to be
avoided to retain the desired initial properties. Detailed information about the application
are therefore required and these need to be compared to the bimetallic stability window.
Nevertheless, dealloying can cause the formation of porous bimetallic structure (as shown
from dealloyed PtNi nanocatalyst [135]) that can lead to new interesting perspective as
shown for gold nanoporous catalyst [296-298].
Figure 6.8 Schematic representation of selective palladium dealloying, yielding a gold enriched
surface composition. a) Fresh as-prepared catalyst, b) Pd dissolution and c) Au-enriched surface
after potential cycling.
6.5 Influence of Gases
Online dissolution of AuPd nanoparticles is recorded also in O2 purged 0.1M HClO4 (Figure
6.9). While no significant differences with Ar purged electrolyte are observed during
potential cycling below the dissolution onset, the presence of O2 leads to a shift in the open
circuit potential (OCP). Namely, the OCP in O2 purged electrolyte reaches approximately
0.9 VRHE, slightly above the measured dissolution onset potential, while in Ar purged
electrolyte it remains below 0.8 VRHE. Therefore, Pd is being dissolved at OCP in the
presence of O2. Gas induced changes in alloys surface composition are already reported in
heterogeneous catalysis literature [110, 299], which is commonly attributed to metal
migration. According to our results, however, we suggest that also selective dissolution in
the presence of different gases plays an important role in determining the surface
a) b) c)
- Addressing Stability of Bimetallic Electrocatalysts: the Case of Au-Pd Alloys
| 87
composition. This is particularly relevant for the long-term stability of bimetallic
nanoparticles in reactions that requires gases such as O2, CO2, O3 that can cause high OCP
values, or in reactors that are shut down frequently and air is able to diffuse in. Therefore,
gas induced dealloying has also to be taken in consideration in heterogeneous catalysis,
where no potential control is applied.
Figure 6.9 Pd dissolution profile of AuPd nanoparticles (1 printed layer) in O2 (full line) and Ar
(dotted line) purged 0.1M HClO4. 10 cyclic voltammograms between 0.1 and 0.6 VRHE (below the
dissolution potential) with a scan rate of 50 mV s-1 followed by OCP.
6.6 Conclusion
In summary, this Chapter dealt with an extensive dissolution study on alloyed AuPd
nanoparticles supported directly on the electrode. The reaction environment is strongly
influencing metal dissolution and dealloying: i) different electrolytes cause a significant
variation in the dissolution rate depending on the nature of anions and/or cations present
in the solution, and ii) dissolved O2 plays a key role in enhancing the dissolution rates by
shifting the OCP. Even though the interpretation of the results is challenging due to the
difficulty in estimation of the precise surface composition, the quantitative normalized
dissolution indicates that the dissolution of Au and Pd in the alloy is approximately half of
the normalized dissolution of the metal counterparts. Considering the nominal
composition (1:1 molar ratio), no major stabilization of Pd is therefore observed. On the
other hand, the measured dissolution profiles differ for alloyed Pd compared to the pure
metal.
A well-defined stability window can be defined: no dissolution/dealloying of Pd occurs
below ≈0.8 VRHE and in the presence of a gas with low OCP. In such cases, changes in
Chapter 6
88 |
surface composition are only assignable to metal migration or segregation, which were not
addressed in here. Above potentials of 0.8 VRHE, In contrast with other works, the results
show that the main contribution to changes in surface composition is coming from
selective dissolution rather than from metal migration. The faster dissolution rate of Pd
results in Au surface enrichment.
AuPd catalyst are used here as a model system, however, the results and implications can
be extended to any bimetallic system. Structure, surface composition, and thus activity, can
change over time under reaction environments due to selective dissolution. In turn,
dealloying could be also exploited positively with selective dissolution by subjecting
particles to electrochemical conditions, in order to control and tune the catalyst surface
composition. With this “activation” the bimetallic effects could be optimized to achieve and
maintain enhanced catalytic activity.
On the other hand, in real applications it is difficult to avoid dissolution and thus to control
the bimetallic surface composition over the long reaction times. In fact, in fuel cells it is
likely to have potential spikes which exceed the stability window during start and stop
condition, while in heterogeneous catalysis mixtures of gases might lead to dealloying
through changes in the potential of the system. In both cases this is detrimental for
application based on reactions, where the coexistence of both metals on the surface is
necessary (i.e. peroxide synthesis, alcohols oxidation, formic acid oxidation).
In conclusion, it should be note to the reader that having exhaustive
dissolution/dealloying data combined with precise information about the reaction
environment are of crucial importance to guarantee the performance and stability of all
materials that rely on ensemble effects. Indeed, if potential fluctuations occur, the resulting
dealloying can change in a short time dramatically their surface composition and therefore
their activity.
- Electrocatalytic Peroxide Synthesis on Au-Pd Nanoparticles
| 89
- Electrocatalytic Peroxide Synthesis on Chapter 7
Au-Pd Nanoparticles 14
——————————————————————————————————————————
As discussed in the introduction, the electrochemical synthesis represents a promising and
attractive alternative to the traditional anthraquinone process, as it combines both on-site
chemical and electrical production. However, compared to the direct heterocatalytic
synthesis, the electrocatalytic synthesis is much less studied. The design of novel selective
electrocatalyst as well as the understanding of active sites is challenging. Commonly, the
alloying of elements with different reaction
intermediate binding energies is employed to
tune the selectivity. In the present Chapter the
H2O2 electrochemical production on Au-Pd
unsupported nanocatalysts with compositions
ranging from pure Au to pure Pd (Au, Au9Pd,
Au3Pd, AuPd, AuPd3, Pd) will be presented. In
particular, in the first part, the change in the
ORR mechanism with composition will be shown,
followed by the change in the PRR activity.
Finally, potentiostatic conditions at potentials of
maximum peroxide current for each catalyst will be used to simulate a possible application;
the H2O2 productivities are evaluated after 2 and 30 min of measurement and provide
additional information on the catalysts behaviors in real systems.
——————————————————————————————————————————
14 Parts of this chapter have been already published in:
E. Pizzutilo*, O. Kasian, C.H. Choi, S. Cherevko, G.J. Hutchings, K.J.J. Mayrhofer, , S.J. Freakley Chem. Phys.
Lett.. 2017, 683, 436-442.
There are therefore numerous verbal quotes from that publication. Some of the figures present in
the publication have been re-printed or modified.
Chapter 7
90 |
7.1 Oxygen Reduction Reaction (ORR)
Initially, as reference for the following measurements on different Au-Pd compositions,
ORR polarization curves are collected only for pure polycrystalline metals (Figure 7.1).
Figure 7.1 a) Anodic disc polarization currents obtained for poly-Au and poly-Pd bulk electrodes in
O2 saturated 0.1M HClO4. Scan rate: 50 mV s-1. b) Anodic ring current (Ir) obtained for poly-Au and
poly-Pd bulk electrodes during disc polarization in O2 saturated 0.1M HClO4. Scan rate: 50 mV s-1. c)
Selectivity obtained from Ir and Id for poly-Au and poly-Pd bulk electrodes during disc polarization
in O2 saturated 0.1M HClO4. Scan rate: 50 mV s-1.
As expected from literature, poly-Pd behaves as a 4-electron catalyst, whereas the poly-Au
only reduces O2 following the 2-electrons pathway, resulting in a SH2O2 which is close to
100%. On the other hand, the SH2O2 of poly-Pd deviates from zero only at very low potential
- Electrocatalytic Peroxide Synthesis on Au-Pd Nanoparticles
| 91
corresponding to HUPD. Similar behavior is also observed for poly-Pt and, although not fully
clarified yet, it is attributed to a change in reduction mechanism caused by H adsorption.
The non-supported Au-Pd electrocatalysts (Au, Au9Pd, Au3Pd, AuPd, AuPd3 and Pd) are
prepared through a sol-immobilization method described in Chapter 3 [123, 263] and their
ORR behavior is studied with the RRDE [300] (Figure 7.2).
Figure 7.2 RRDE results obtained for ORR on different Au-Pd catalyst compositions in O2
saturated 0.1M HClO4. Rotation: 900 rpm. Scan rate : 10 mV s-1. Er: 1.0 VAg/AgCl. The colors (green for
Id, violet for Ir, brown for SH2O2) are graded with the change in composition. Only the anodic sweep
is shown. a) Disc polarization current (Id), b) Ring current (Ir) profiles during one cycles [0.1-1.0]
VRHE and c) Calculated values of SH2O2.
Their size distribution and average particle size estimated by STEM are summarized in
Table 4.1 and the real Au:Pd molar ratios are close to the nominal as confirmed with both
Chapter 7
92 |
ICP-MS and XPS measurements (Figure 4.2 and Table 4.2) and the initial alloy surface
composition and surface status is investigated with cyclic voltammetry in the previous
Figure 4.5. As the particle size is for all the nanoparticles in the range of 3-4 nm, the ORR
polarization curves are normalized to the Ageo.
Figure 7.2a show the ORR disc currents (Id). The data show at a glance, how the
composition significantly affects the ORR half-wave and onset potentials. Indeed, from 0.4
VRHE for Au the onset is shifting positively, until it reaches 0.9 VRHE for Pd nanoparticles.
The comparison with theoretical diffusion limited currents (calculated from the Levich
equation [249]) suggests already that the mechanisms is switching from a 2- (where H2O2
production is dominant) to a 4-electrons process.
To confirm the presence of H2O2 the corresponding measured ring current (Ir) profiles are
shown in Figure 7.2b. Correcting the latter by the collection efficiency N the total H2O2
current (Iper) can be estimated and compared to Id (Figure 7.3).
Figure 7.3 Disc polarization current (Id in green) and corresponding H2O2 current (Iper in violet) calculated
from the ring current (Ir) corrected with the collection efficiency N.
The lower limit of 0.1 VRHE is chosen above the hydrogen evolution (HER) region to avoid
high cathodic currents. As expected from the Id, also the Ir, and thus the H2O2 production
(Iper), is considerably influenced by the composition. In particular, the highest Ir current is
measured for pure Au nanoparticles (∿0.3 mA cmgeo-2 @0.1 VRHE) and the lowest for Pd
nanoparticles (<0.05 mA cmgeo-2). At intermediate compositions, Ir decreases and the H2O2
production onset is also shifting. For low Pd content catalysts (Au, Au9Pd, Au3Pd) the onset
potential at the ring and the disc coincide. Interestingly, for the remaining composition
(AuPd, AuPd3, Pd) the onset potential at the ring does not follow any longer the shift in the
onset potential at the disc but it is constant at a value close to the standard potential. Note
- Electrocatalytic Peroxide Synthesis on Au-Pd Nanoparticles
| 93
that, as already discussed in the introduction, in a 2-electron process the peak of the
volcano plot coincides with the Nerstian potential for the reaction. In other words, while
the best ORR catalyst requires a large overpotential (∿ 0.4 V) for its reduction to H2O, an
ideal catalyst for the H2O2 production can have zero overpotential [131]. The total H2O2
produced in a cycle ([0.1-1.0] VRHE, 10 mV s-1) is highest for the Au3Pd sample (area under
Ir polarization curves of Figure 7.2a). Interestingly, for AuPd3 and in particular for Pd a
peak of H2O2 is observed at low potential (<0.2-0.3 VRHE). This can be attributed to a
change in ORR mechanism once the Pd is covered with hydrogen in the HUPD region, as is
well known also for Pt-based catalysts and observed also for poly-Pd (Figure 7.1) [82].
Another peak in the Ir at higher potential is also often observed in both nanoparticles Pd
and poly-Pd that corresponds to the Pd-O reduction, however its interpretation will require
further studies.
From the Id and Ir the SH2O2 has been evaluated (Figure 7.2c) and as expected it is
decreasing with the Pd content from ∿95% to less than 10% for pure Au and Pd
respectively. High SH2O2 of Au nanoparticles is confirmed also with poly-Au (see Figure
7.4): the latter is slightly more active, but both electrodes show similar onset potential and
SH2O2 (around 95-98%, confirmed also by the match between Id and Iper for the Au
nanocatalyst in Figure 7.4).
Figure 7.4 Comparison of RRDE results between poly-Au and Au nanoparticles during disc
polarization in O2 saturated 0.1M HClO4. Scan rate: 50 mV s-1.
Several studies on Au (prevalently polycrystalline) indicate a remarkable influence on the
kinetics and mechanisms of the ORR of the reaction condition, size, support and
crystallographic orientation [91, 93, 94, 301]. Recently Jirkovsky et al. showed that the
SH2O2 of carbon supported gold nanoparticles strongly depends on the average particle size:
the smaller it is the higher the SH2O2 [95]. However, they showed a potential dependence of
Chapter 7
94 |
Au SH2O2, not observed in our case. This difference might be due to the influence of the
presence of a carbon support. Indeed, they showed a selectivity decreases with layer
thickness, indicating that further reduction or degradation of H2O2 might take place within
the film. However, this difference might also be due to the different synthesis protocol and
should need future investigation.
In another publication, Jirkovsky et al. observed an enhancement of SH2O2 for low Pd
concentration (<15%) in Au-Pd catalyst with a maximum observed for Pd molar content of
8% [105], whose selectivity approached 95%. This was attributed to the presence of Pd
monomers surrounded by Au at the surface of the alloy. Our Au9Pd does not show any
improvement in terms of SH2O2 compared to the pure Au sample, whose SH2O2 is already
close to 95%. As discussed before this requires more investigation also in terms of how the
support influences the H2O2 production and fine tuning. Indeed, first principles calculation
suggests a strong influence on SH2O2 due to the large activity difference of Au and Pd and to
geometric effect [128]. In practice, this could correspond to an increase in activity while
maintaining a high SH2O2, which however is not observed with our samples.
7.2 Composition/Ir,max/Selectivitymax Relationship
From all the RRDE results for the ORR on Au-Pd catalysts it is possible to draw a picture
to predict the behavior of such catalysts for future applications. As observed also
elsewhere, both SH2O2 and Ir exhibit a maximum [97, 105, 125] (see resuming Figure 7.5).
Figure 7.5 SH2O2 and Ir maxima vs. disc potentials corresponding to the measurement in Figure 7.2.
Tentatively, this maximum can be attributed to a change in mechanisms at large
overpotential, at which H2O formation is favored rather than H2O2. The maxima of Ir and
SH2O2 are shifting with the composition (Figure 7.2d) as also observed for the H2O2
production onset potential.The highest Ir and SH2O2 are observed for pure Au nanoparticles,
- Electrocatalytic Peroxide Synthesis on Au-Pd Nanoparticles
| 95
whereas increasing the Pd content the overpotential for H2O2 evolution is decreasing up to
value close to the nominal for Au3Pd at the price of a decreased SH2O2 and of maximum Ir
(Figure 7.5). In a real application, this could mean less energy lost in the production
process due to a lower overpotential. Considering, that most application requires only H2O2
around 2-8 wt% this could be an acceptable compromise.
7.3 Peroxide Reduction Reaction (PRR)
It is of utmost importance for application to answer the question whether the produced
H2O2 would be further reduced in the fuel cell at the potentials at which the H2O2 is
produced. Therefore, PRR of freshly prepared Au-Pd electrodes are studied in Ar-saturated
0.1M HClO4 containing 10 mM of H2O2. The corresponding anodic polarization curves are
shown in Figure 7.6.
Figure 7.6 Anodic sweep of the PRR on Au-Pd catalyst compositions in Ar saturated 0.1M HClO4 +
10mM H2O2. Rotation: 900 rpm. Scan rate: 50 mV s-1.
As expected, an increase in Pd content corresponds to an increase in PRR currents, with
the highest diffusion limited current obtained for Pd. On the other hand, Au is not active
for the PRR (only a very small cathodic current is observed below 0.2 VRHE). These results
are also confirmed by the measurement on poly-Au and poly-Pd performed at different
rotation rates (see Figure 7.7) and by previous literature on Au-Pd codeposited
nanoparticles [126, 302] and Pd electrodeposited on a poly-Au electrode [303].
Interestingly, all the polarization curves match around 0.82 VRHE, indicating both a change
in the mechanism (from reduction to oxidation) and surface state. Katsounaros et al. found
that whether H2O2 will be reduced or oxidized on poly-Pt is determined only by the surface
state, and thus by the applied potential [57]. They showed also that the PRR proceeds on
reduced surface sites initially producing adsorbed OH groups, which are thereafter
Chapter 7
96 |
reduced, whereas POR is initiated by reducing an oxidized surface. Therefore, the resulting
current is the sum of the two partial PRR and POR currents; the potential at which PRR
and POR are equal (where actually the polarization curves match) is therefore associated
to the transition between a reduced and an oxidized state. The Pd oxidation onset in HClO4
is above 0.7 VRHE [85, 150], that is indeed when the PRR current starts to decrease (Figure
7.6). The dependence of this transition from the surface state is also confirmed in a
comparison between poly-Au, poly-Pd and poly-Pt (Figure 7.7a).
Figure 7.7 a) PROR comparison with different polycrystalline metals. PROR at different rotation
rate with b) poly-Au and c) poly-Pd in Ar saturated 0.1 M HClO4 + 1 mM H2O2. Scan rate: 50 mV s-1.
Comparing for poly-Pd and poly-Pt the potential at which PRR and POR are equal in the
positive going sweep, the measure on poly-Pt shows a shift of about 70 mV to higher
potentials in compared to poly-Pd. This result is in accord with the shift in the respective
oxidation onset potential [65]. Furthermore, while the PRR diffusion limited current on
poly-Pd and poly-Pt equals (Figure 7.7), the POR on poly-Pd seems to proceed differently
(the same is also observed for poly-Au). Both the POR behavior of Pd and Au will require
further investigation.
- Electrocatalytic Peroxide Synthesis on Au-Pd Nanoparticles
| 97
7.4 Potentiostatic H2O2 Production
To finally bridge the gap between possible “real application” and our fundamental studies
we need to understand how the catalysts behave during demanding continuous
(potentiostatic) H2O2 production. At the end of each potentiostatic experiment (2 and 30
min), the H2O2 concentration in the electrolyte is evaluated from the POR limiting current
measured with the Pt ring electrode. Prior to any measurement, a calibration curve (POR
limiting current at different H2O2 molar concentration) is obtained (see Figure 3.15).
Figure 7.8 Measured a) Ir, b) Id and c) Selectivity curves during 2 min potentiostatic experiment in
O2 saturated 0.1M HClO4. Rotation: 900 rpm. Ed corresponds to the potential of Ir,max (see Figure
7.5). Er: 1.28 VRHE.
RRDE data (Figure 7.8) during 2 min potentiostatic condition (potential of Ir,max for each
composition, in order to maximize the productivity) confirms the activity and SH2O2 trend
observed in the previous sections. Interestingly, the Ir (and thus SH2O2) slightly increases
with time for Pd rich catalysts (AuPd3, Pd). This is even more evident during the 30
Chapter 7
98 |
minutes measurements (see Figure 7.9) and can be tentatively attributed to the presence of
either impurities or spectator species [3, 19] that initially poison active sites for the 4-
electron pathway. As a consequence in some site O2 is not being separated into oxygen
atoms any longer, thus favoring the H2O2 production. This was also observed for Pt [38].
Such spectator species can be easily removed with a simple CV [13]; indeed, the ORR
behavior before and after the measurement is not affected. In case of Au rich catalyst, the
opposite behavior is observed, namely a decrease in the Ir which can also be attributed to
impurities blocking partially the Au surface. The understanding of the role of spectator
species and impurities requires further investigation.
Figure 7.9 Selectivity curves during 30 min potential hold for different Au-Pd catalyst
compositions in O2 saturated 0.1M HClO4. Rotation: 900 rpm. Er: 1.28 VRHE.
The productivity of the catalyst after 2 and 30 minutes is reported in Table 7.1. Note that
the final value is corrected with the H2O2 lost to the detection at the ring. In line with the
high SH2O2 and low PRR activity of Au, its productivity is the highest. Increasing the Pd
content, the SH2O2 decreases and PRR activity increases, resulting in a proportional
productivity decrease. Despite the low productivity expected for 4-electron catalysts with
low SH2O2 as Pd, it yields ∿2.40 mol gmetal-1 cm-2
geo after 30 min. This can be again attributed
to a poisoning of the active sites for complete reduction, favoring a 2-electron process.
Table 7.1 H2O2 productivity after 2 and 30 mins of measurement and the average selectivity during a 30
min measurement
Potential hold /
VRHE
Productivity
2 min /
mol gmetal-1 cm-2
geo
Productivity
30 min /
mol gmetal-1 cm-2
geo
Average
selectivity /
%
Au
Au9Pd
Au3Pd
AuPd
AuPd3
Pd
0.10
0.15
0.25
0.40
0.45
0.50
0.46
0.39
0.40
0.29
0.24
0.14
6.79
5.46
5.87
3.37
3.42
3.13
89
80
74
52
48
39
- Electrocatalytic Peroxide Synthesis on Au-Pd Nanoparticles
| 99
From the productivities, an evaluation of the effect of PRR is not straightforward.
However, comparing the measured H2O2 concentration after 30 min with the total H2O2
produced evaluated from the integration of the Ir, it is possible to get at least some
information. For instance, the total productivity from integration of Ir resulted 3.39 mol
gmetal-1 cm-2
geo for Pd. Therefore, it is higher than the productivity measured from the final
H2O2 in the solution, meaning that in 30 min part of the produced H2O2 has been further
reduced to H2O (PRR) by the Pd catalyst itself. Similar observations are obtained for all
the Pd rich alloys, for which the PRR is more relevant,
As a final proof, being Au the most productive, the potentiostatic experiments on Au are
also extended to 1 and 2 h yielding 12.3 and 22.9 mol gmetal-1 cm-2
geo. Therefore, it seems
that the H2O2 is increasing linearly with the time.
7.5 Conclusion
In this Chapter, unsupported Au-Pd catalysts are used to study the electrocatalytic
behaviour for H2O2 electrocatalytic synthesis from fundamental perspectives till continuous
H2O2 production. Composition affects significantly ORR and PRR activities: upon an
increase in Au content, Ir and SH2O2 increase, whereas the ORR onset is shifting to lower
potentials. Thus, the SH2O2 of Au is the highest (95%) at the price of low activity. Unlike in
previous theoretical as well as experimental works no strong geometrical effects (triggered
by the presence of atomically dispersed Pd in Au) enhancing the activity while maintaining
high SH2O2, are observed in this work. For Au9Pd and Au3Pd, the activity is indeed higher
compared to Au, but the SH2O2 decreases (to 80 and 60%). Considering applications, these
compositions might still be interesting if a compromise between SH2O2 and energy output is
allowed. On the other hand, if high H2O2 concentration and high SH2O2 is required, pure Au
would potentially be the catalyst of choice, as its productivity during a potentiostatic
measurement suggests (highest productivity that steadily increases up to 2h).
These observations on a spectrum of compositions between elements with contrasting O2
adsorption energy can be extended to other alloys with similar characteristic. This
fundamental study can help to forecast the ORR behaviors and reaction selectivities during
applications. Still the stability of catalysts and the consequences of their degradation to the
electrocatalytic performances, are important parameters to influence the catalysts choice
and will therefore be investigated in the following Chapter 8
.
Chapter 8
100 |
- Au-Pd Bimetallic Catalyst Stability: Chapter 8
Consequences for Peroxide Selectivity 15
——————————————————————————————————————————
The degradation of Au-Pd bimetallic catalysts during H2O2 electrocatalytic production from
ORR is addressed in this Chapter. To the author’s knowledge it is the first study on the
degradation of candidate catalysts for H2O2 synthesis. Degradation can occur in real
application (i.e. in PEMFC) as the catalysts are exposed to harsh electrochemical
conditions, especially during start and stop. Particularly critical are the changes in surface
composition, which are likely to occur if the dissolution of the two metals occurs at
different potentials inducing dealloying. It is therefore of primary interest to forecast the
chemical and structural changes and the consequent evolution of electrocatalytic
performances under different operational conditions. In this context, Au and Pd
dissolutions/degradation are studied under three different ADPs with SFC-ICPMS.
Furthermore, the evolution of size and composition is monitored with IL-STEM and EDS
and is correlated to changes in electrochemical performance measured with RRDE.
——————————————————————————————————————————
15 Parts of this chapter have been already published in:
E. Pizzutilo*, S.J. Freakley, S. Cherevko, S. Venkatesan, G.J. Hutchings, C.H. Liebscher, G. Dehm, K.J.J.
Mayrhofer, ACS catalysis,. 2017, 7, 5699-5705.
There are therefore numerous verbal quotes from that publication. Some of the figures present in
the publication have been re-printed or modified.
- Au-Pd Bimetallic Catalyst Stability: Consequences for Peroxide Selectivity
| 101
8.1 Au and Pd Dissolution under ADPs
In previous Chapters, the difference in Pd and Au dissolution onset potentials (ca. 0.8 and
1.3 VRHE respectively) was addressed. Starting from this study, the stability of supported
Au-Pd/C is here studied under ADPs with UPL varying between 0.8 VRHE (below Pd
dissolution) and 1.6 VRHE (above Au dissolution onset). The three considered ADPs with a
scan rate of 1 V s-1 are:
ADP-0.8 consisting in 1000 CVs in the range [0.1-0.8] VRHE with IL after 100 and
1000 cycles;
ADP-1.2 consisting in 1000 CVs in the range [0.1-1.2] VRHE with IL after 100 and
1000 cycles;
ADP-1.6 consisting in 100 CVs in the range [0.1-1.6] VRHE with IL after 100 cycles;
The potential ranges are chosen in order to study the catalyst evolution under different
dissolution/dealloying regimes (see Chapter 6). The latter is stopped after a lower amount
of degradation cycles as the structure and electrochemical behavior is changing drastically
in few cycles under such harsh potential condition.
Figure 8.1 reports the online potential dissolution profiles of Pd and Au measured by
means of SFC-ICPMS during the first 800 s of the respective ADPs on the AuPd/C catalyst.
Figure 8.1 Online dissolution profiles of a) Pd and b) Au recorded for AuPd/C means SFC/ICPMS
technique during degradation CVs [0.1-UPL] VRHE.
As expected from the relative onset potentials (in Chapter 6), Pd dissolution occurs during
the ADP-1.2, while Au dissolution is visible only under ADP-1.6. Interestingly, the Pd
dissolution profiles after ca. 800 s of ADP-1.2 approaches the background signal, while the
signal for ADP-1.6 is still high, as Au is also dissolving exposing fresh Pd on the surface of
the nanoparticles.
Chapter 8
102 |
8.2 Evolution of Surface Composition: Cyclic Voltammetry
in Ar
Evidence of the surface change induced by dissolution is supported by the CVs recorded in
Ar saturated electrolytes (Figure 8.2). Indeed, once the UPL is sufficiently high (above Au
oxide formation [263], ca. 1.5 VRHE) two distinct peaks are observable on the reverse scan:
Pd oxide and Au oxide reduction (ca. 0.55 and 1.15 VRHE for Pd/C and Au/C respectively).
Figure 8.2 CV [0.1-1.6] VRHE recorded with AuPd/C in Ar purged 0.1M HClO4, before as well as after
ADPs (1000CV 0.8 VRHE, 1000CV 1.2 VRHE, 100CV 1.6 VRHE). Au/C and Pd/C CVs are also shown as
reference. Scan rates: 0.2 V s-1.
The positions and relative charges (Table 8.1) of these two peaks are associated with the
surface alloy composition [105, 126, 127, 150, 241, 275].
Table 8.1 Potential and charge associated to the Pd-O and Au-O reduction peaks corresponding to
the Ar background CVs recorded after degradation.
AuPd/C
Pd-O reduction Au-O reduction
Ed /
VRHE
Q /
mC
Ed /
VRHE
Q /
mC
Initial
ADP-0.8
ADP-1.2
ADP1.6
0.75
0.76
0.77
0.17
0.19
0.06
<0.01
1.05
1.05
1.07
1.09
0.29
0.27
0.5
0.36
For example, when the UPL is kept lower than the Pd dissolution onset potential, i.e.
under ADP-0.8, the shape of the CV is maintained even after 1000 degradation cycles.
Nevertheless, the charge associated to Pd reduction peak appears to be slightly larger after
degradation ADP-0.8. As no significant dissolution is occurring during such ADP, the
- Au-Pd Bimetallic Catalyst Stability: Consequences for Peroxide Selectivity
| 103
difference can be tentatively attributed to a dynamic change in the nanoparticle structure
during potential cycling. Indeed, DFT and experimental studies suggest catalyst surface
rearrangement or “breathing”, i.e. Pd surface segregation with absorbed H2 [304, 305], O2
[122] or CO [306-308] in addition to Au migration towards the surface during catalytic
[299] and electrocatalytic [110] measurements.
As the potential exceeds the Pd dissolution, the charge associated with the Pd reduction is
decreasing during CVs under both ADP-1.2 and ADP-1.6 (Figure 8.4e and Figure 8.5d
respectively). After only 100 CV under ADP-1.6 the reduction peak fully disappears,
whereas after 1000 CV under ADP-1.2 still the reduction peak is observed (see CV in
Figure 8.2). This suggests that Pd is still present on the surface despite 1000 CV cycles at
ADP-1.2. However, it needs to be considered that this reduction peak was suggested to be
associated to a new surface phase of alloyed Au and Pd [241]. With this consideration, Pd
could still be present under a surface consisting of mainly Au. In this case, the observed
reduction peak might be attributed to surface Au alloyed with underlying Pd. However, we
would exclude this as such feature is not observed in core-shell configuration (with Au-
shell) [241]. We suggest here that there is some dispersed Pd on the surface which is not
further dissolving thanks to the presence of Au that might stabilize it, as elsewhere
suggested for Pt [292]. As expected during Pd leaching (Figure 8.1), the surface
concentration of Au increases [110, 241, 278] as observed from the associated reduction
peak.
The Au reduction peak after ADP-1.6 is lower than after ADP-1.2 as Au is dissolving under
ADP-1.6 (see also the Au reduction peak evolution under ADP-1.6 in Figure 8.5d). For a
more detailed investigation on Au-Pd catalyst surface changes (and the relative changes in
CVs) we invite the reader to refer to the work of Lukaszewski et al. [241]. In their study,
the influence of the electrochemical protocol, of the initial alloy composition as well of H2
absorption or O2 electrochemisorption is discussed.
8.3 Evolution of Catalyst Microscopic Structure: IL-STEM
Electrochemistry is a powerful tool to study the macroscopic changes of the catalyst
changes in the argon background CVs, as discussed in the previous section, that correlate
to changes in the catalyst surface state and composition. Microscopic characterization
during different steps of the ADPs with IL-STEM can provide further insight on the
catalyst structural evolution. For instance, particle size changes (in Figure 8.3, Figure 8.4
and Figure 8.5 for ADP-0.8, ADP-1.2 and ADP-1.6 respectively) and of composition (Figure
8.6) can be correlated with the evolution of the CVs at different stages of the relative ADP.
Chapter 8
104 |
Figure 8.3 Collection of Comparison of bright-field IL-STEM micrographs recorded at different
degradation stages during ADP-0.8: a) initial, b) after 100 CVs and c) after 1000 CVs. Corresponding
d) statistical particle size distributions and e) cyclic voltammograms.
In Figure 8.3 STEM micrographs of the same AuPd/C catalyst spot after 1, 100 and 1000
CVs of ADP-0.8 shows no remarkable changes in particle size distribution and a small
decrease in number of particle (Table 8.2) probably due to agglomeration. The macroscopic
Ar voltammographs show as expected also no significant changes.
Table 8.2 N of particle measured for the statistical analysis and the calculated average size.
ADP-0.8 N of particles Mean St. dev
Initial
100
1000
391
381
355
3.7
3.7
4.0
±1.4
±1.3
±1.4
- Au-Pd Bimetallic Catalyst Stability: Consequences for Peroxide Selectivity
| 105
Figure 8.4 Collection of Comparison of bright-field IL-STEM micrographs recorded at different
degradation stages during ADP-1.2: a) initial, b) after 100 CVs and c) after 1000 CVs. Corresponding
d) statistical particle size distributions and e) cyclic voltammograms.
In Figure 8.4 STEM micrographs of the same AuPd/C catalyst spot after 1, 100 and 1000
CVs of ADP-1.2 shows changes in particle size distribution and a decrease in number of
particle down to 50% of the initial already after 100 CVs (Table 8.3) probably due to
superposition of different degradation mechanism such as agglomeration, detachment,
dissolution of Pd causing dealloying and Ostwald ripening. The macroscopic Ar
voltammographs show as expected also a decrease in the reduction peak, however still
some Pd might be present after 1000 CVs as evidenced by the CV at higher UPL in Figure
8.2.
Table 8.3 N of particle measured for the statistical analysis and the calculated average size.
ADP-1.2 N of particles Mean St. dev
Initial
100
1000
215
119
115
3.8
4.4
4.3
±1.0
±1.1
±1.2
Chapter 8
106 |
Figure 8.5 Collection of Comparison of bright-field IL-STEM micrographs recorded at different
degradation stages during ADP-1.6: a) initial and b) after 100 CVs. Corresponding c) statistical
particle size distributions and d) cyclic voltammograms.
In Figure 8.5 STEM micrographs of the same AuPd/C catalyst spot after 1, 100 CVs of
ADP-1.6 shows changes in particle size distribution and a decrease in number of particle
after 100 CVs (Table 8.4) probably due to superposition of different degradation
mechanism such as agglomeration, detachment, dissolution of Pd causing dealloying and
Ostwald ripening. The macroscopic Ar voltammographs show a dramatic decrease in the
Pd reduction peak within few CVs (unlike under ADP-1.2 here it is completely disappeared
after 100 CVs).
Table 8.4 N of particle measured for the statistical analysis and the calculated average size.
ADP-1.6 N of particles Mean St. dev
Initial
100
169
116
3.4
4.4
±1.0
±1.3
- Au-Pd Bimetallic Catalyst Stability: Consequences for Peroxide Selectivity
| 107
8.4 Evolution of Composition: STEM-EDS and ICPMS
Giving the relevance of knowing the spatial distribution of the alloying elements of
AuPd/C, additional high-resolution STEM-EDS analysis of single nanoparticles are
acquired on some sample degraded nanoparticles after ADP-1.2 and ADP-1.6 (Figure 8.6).
Figure 8.6 a) High angle annular dark-field scanning transmission electron microscopy (HAADF-
STEM) investigation and STEM-EDS elemental maps of AuPd/C after ADP-1.2 and ADP-1.6 of Au.
Corresponding catalyst EDS line scan after degradation are shown in b) and c), respectively.
The macroscopic changes of the overall catalyst film are, instead, characterized by post-
mortem ICPMS analysis (Figure 8.7b) and compared to EDS spectra (Figure 8.7a).
Figure 8.7 a) EDS spectra normalized to the Au-M peak of AuPd/C catalyst and b) Pd% molar ratio
(molPd/(molPd+molAu)) before and after ADPs measured by post-mortem analysis with the ICPMS.
Chapter 8
108 |
The initial theoretical composition of AuPd/C (Figure 8.7) is confirmed both by EDS
elemental map (47 Pd mol%) and ICPMS bulk analysis (46±1 Pd mol%). STEM-EDS
mapping of the as synthesized catalyst indicates homogeneous Pd and Au distribution
within the nanoparticles (Figure 4.7). These high resolution analysis can be combined with
the previous IL-STEM data, thus allowing a clearer description of the catalyst degradation
mechanisms on the nanoscale [236]. Once again, after ADP-0.8 (Figure 8.3) no major
change in the particle size distribution and in the overall number of particles are observed
(the small increase in mean particle size from 3.7 to 4.0 nm can be observed due to minor
agglomeration). For both ADP-1.2 and ADP-1.6 the number of particle counts dropped and
the average particle size increases from 3.8 to 4.4 nm and from 3.5 to 4.4 nm for ADP-1.2
and ADP-1.6 respectively due to a decrease in number of particles with sizes below ∿3 nm
(Figure 8.4, Figure 8.5). For such ADPs, this increase, as well as the particle rounding, can
be attributed to potential induced metal dissolution/dealloying and consequent Ostwald
ripening forming rounder nanoparticles [309, 310]. Interestingly, comparing the IL-TEM
micrographs after 100 CVs and 1000 CVs of ADP-1.2, the number of particle counts as well
as the average size remains unchanged. In this case, clearly, the Au dissolution is excluded
whereas Pd dissolution/dealloying decreases below the ICP-MS detection limit (Figure 8.1)
as the surface is enriched in Au. Therefore, either Au might stabilize the remaining Pd as
claimed in literature [290, 292] or a protective Au shell is preventing further dealloying as
already observed in the case of Pt based alloys [311-313].
Coming to the compositional changes, in agreement with the online ICP-MS results, upon
ADP-0.8 the metal atoms are still homogeneously distributed within the catalyst and the
Pd molar ratio (defined as molPd/(molPd+molAu)) is 45±1 mol%, within the error compared to
the initial composition. Increasing the UPL, the Pd molar ratio (Figure 8.7b) decreases
after 100 CVs to 41±1 and 30±3 mol% for ADP-1.2 and ADP-1.6 respectively (a trend also
confirmed by the decrease in Pd Lα and Lβ of the respective EDS spectra in Figure 8.7a).
Upon additional 900 cycles of ADP-1.2 the measured Pd molar ratio is 39±2 mol% (average
after 3 separate degradation measurements), indicating that the further dealloying is
almost negligible. Clearly, under ADP-1.2 a quasi-stable configuration, for which no
changes in composition and structure occurs, is obtained after few potential cycles.
On the nanoscale, the EDS line scans after ADP-1.2 (Figure 8.6b) indicate that while the
core is still homogeneous (Pd and Au intensity heights are equal as expected by the initial
composition), the Pd content at the surface is lower but seemingly still present. Indeed, in
this case an evident formation of core-shell nanoparticles was not observed (as also
indicated by the small Pd-O reduction peak in Figure 8.1b). On the other hand, the EDS
line scans after ADP-1.6 (Figure 8.6c) clearly establish the presence of a AuPd core (similar
Pd and Au intensities) surrounded by a Au shell (0.5-1 nm thickness). When the UPL
exceeds the threshold for significant Au dissolution this shell is destabilized and the
- Au-Pd Bimetallic Catalyst Stability: Consequences for Peroxide Selectivity
| 109
dissolution of Pd from the core of the nanoparticles is initiated, causing further de-alloying
as observed also for Pt based catalyst [311]. Note however, that a precise visualization of
the molar surface composition is rather challenging due to the small size of the
nanoparticle, which can be susceptible to beam damage. Some monomer or small Pd
clusters stabilized by the surrounding Au might be therefore still present on the surface
but the surface composition is very Au rich
8.5 Evolution of H2O2 Selectivity: Cyclic Voltammetry in O2
Figure 8.8 Collection of the ORR results (Id in a), Ir in b) and selectivity in c)) obtained with RRDE
in O2 saturated 0.1M HClO4 for the AuPd/C before as well as after ADPs: a) Ir, b) Id and c)
selectivity. As a reference, the Au/C and Pd/C ORR are also shown. Rotation: 900 rpm. Scan rate:
0.05 V s-1. Er: 1.0 VAg/AgCl.
In the previous sections, the surface composition and nanostructural changes due to the
dissolution and catalyst degradation under the three considered ADPs was described. This
section, is dedicated to the study of how these changes are actually influencing the ORR
Chapter 8
110 |
performance. In particular, the focus is on the H2O2 production and SH2O2 collected with
RRDE (Figure 8.8). The measured Id, Ir as well as the calculated SH2O2 are shown in Figure
8.8a,b,c respectively. As a reference, also the RRDE results for Au/C and Pd/C are reported
here. The estimated mean particle diameters (determined by TEM) of the as prepared
catalysts are 4.5±1 nm, 3.7±1.4 nm, 4.0±0.8 nm for Au/C, AuPd/C and Pd/C respectively.
Owing to the similar average particle size and a narrow distribution, the total surface area
is expected to be in the same range.
Nevertheless, currents here are normalized to the geometric surface area (0.196 cm2), as
the real surface area of AuPd is expected to change significantly following catalyst
degradation. The pure metal catalysts (Au/C and Pd/C) provide frames between an almost
pure 2-electron behavior of Au [91, 94, 127, 301] and a 4-electron behavior of Pd [73, 81,
314, 315]. In our previous Chapter [127] the SH2O2 of both unsupported Au nanoparticles
and poly-Au was 95% and independent of the applied potential, whereas the SH2O2 of
Vulcan supported Au/C increases from around 80-85% at 0.1 VRHE to almost 100% at
higher potentials (see also a comparison in Figure 8.9).
Figure 8.9 Comparison of poly-Au, Au and Au/C selectivity calculated from the respective Id and Ir
obtained at 900 rpm rotation rate. Scan rate: 0.05 V s-1. Er: 1.0 VAg/AgCl.
This potential dependence was observed also by Jirkovski et al. [95, 105] and it might be
related to the presence of the carbon support, as they showed a selectivity decrease with
layer thickness. The initial ORR behavior of the alloyed catalyst (AuPd/C) was also
described earlier [126, 127, 316]. The shift in ORR onset potential and the change in H2O2
current with respect to the pure metal counterparts was attributed to a change in
mechanism, as the electronic structure changes in the alloy in dependence of the spatial
Au and Pd atom distribution [105]. In this case, a maximum of 40-45% in SH2O2 of the
initial AuPd/C is observed at ca. 0.5 VRHE. Despite the lower SH2O2 compared to pure Au, the
- Au-Pd Bimetallic Catalyst Stability: Consequences for Peroxide Selectivity
| 111
alloyed metal is much more active and the overpotential for the ORR to H2O2 is therefore
significantly lower. Thus, it can be considered as a good candidate for the electrocatalytic
on-site production for those applications for which low H2O2 concentration in H2O is
required.
The following Figures show the evolution of the ORR behavior under the three ADPs.
Figure 8.10 Evolution of the ORR results obtained with RRDE in O2 saturated 0.1M HClO4 for the
AuPd/C under ADP-0.8: a) Ir and b) Id. Rotation: 900 rpm. Scan rate: 0.05 V s-1. Er: 1.0 VAg/AgCl.
Figure 8.11 Evolution of the ORR results obtained with RRDE in O2 saturated 0.1M HClO4 for the
AuPd/C under ADP-1.2: a) Ir and b) Id. Rotation: 900 rpm. Scan rate: 0.05 V s-1. Er: 1.0 VAg/AgCl.
Chapter 8
112 |
Figure 8.12 Evolution of the ORR results obtained with RRDE in O2 saturated 0.1M HClO4 for the
AuPd/C under ADP-1.6: a) Ir and b) Id. Rotation: 900 rpm. Scan rate: 0.05 V s-1. Er: 1.0 VAg/AgCl.
Under ADP-0.8, no relevant changes, besides a negligible shift of the onset potential, are
observed (Figure 8.10), whereas significant changes occur under ADP-1.2 and ADP-1.6. In
Figure 3 the last positive sweeps after degradation (1000 CVs of ADP-0.8 and ADP-1.2 and
100 CVs of ADP-1.6) are shown and intermediate positive sweeps are illustrated in the SI.
When only Pd is dissolving (ADP-1.2) the onset potential is shifted by ∿200 mV after only
10 CVs (Figure 8.11). Upon further potential cycling the ORR onset potential as well as the
H2O2 current onset potential stabilize around 0.7 VRHE and the Id and Ir do not change
significantly between 50 and 1000 CVs. It seems that under such conditions the initial
performance degradation is followed by a “stable” state, which is still more active than
pure Au and more selective than the initial AuPd/C (70% at the potential of maximum Ir).
From this state on no further degradation is observed, unless the UPL is raised. For
instance, under ADP-1.6 already after 10 CVs the onset potential is shifted by ∿300 mV
(Figure 8.12). After 100 CVs the degraded catalyst behaves very similar to Au/C (Figure
8.8), while the onset potential still remains slightly higher than pure Au. This can be
attributed to the alloying effect of Pd that is still present in the core, as it was abundantly
reported for dealloyed Pt-M alloys with a Pt skin [106, 312, 317-319].
8.6 Composition/Ir,max/Selectivitymax Relationship after
Degradation
From all the RRDE results for the ORR on AuPd/C catalysts it is possible to draw a picture
to predict the behavior of such catalysts during the degradation. As observed also
elsewhere, both SH2O2 and Ir exhibit a maximum [97, 105, 125] (see resuming Figure 8.13).
- Au-Pd Bimetallic Catalyst Stability: Consequences for Peroxide Selectivity
| 113
Figure 8.13 Ir maxima vs. disc potential corresponding to the measurement in Figure 8.8.
The trend of the SH2O2 in Figure 8.8c and of the Ir,max in Figure 8.13 visually summarize the
results: the potential shift can be attributed to the change in surface composition described
in the previous sections. Indeed, a similar shift with composition is also obtained by
directly tuning the catalyst composition (and thus its surface composition) during
synthesis, as shown previously in Figure 7.5. The UPL-dependent evolution of the AuPd
alloy surface composition under the considered ADPs is schematically represented in
Figure 8.14, indicating the formation of a core shell depending on the applied UPL. The
surface reactivity and H2O2 selectivity at the potential of maximum Ir (indicated by the
percentage values) follow the surface composition evolution. The change in particle size is
also reported and it is attributed to the re-deposition of dissolved metals (Ostwald
ripening) as well as to particle agglomeration.
Figure 8.14 Representative evolution of the surface composition and of the H2O2 selectivity (%
values) during the ADPs.
Chapter 8
114 |
8.7 Conclusion
In this Chapter, the structural changes of carbon supported Au/C, Pd/C, and AuPd/C
catalysts were investigated. Three different potential degradation conditions are chosen
with the aim to distinguish the degradation processes that might affect performances,
namely below the Pd/Au dissolution onset potentials (ADP-0.8), above the Pd but below the
Au dissolution potentials (ADP-1.2), and above the Au/Pd dissolution potentials (ADP-1.6).
The associated chemical and structural changes are then related to the H2O2
electrochemical production and SH2O2. While under ADP-0.8 the catalyst and its
electrochemical behavior are unchanged, above the Pd dissolution potential (ADP-1.2) the
surface composition changes, becoming enriched in Au. In this case, Pd surface de-alloying
did not result in an evident core-shell configuration. As the Au does not dissolve, further
Pd dissolution is prevented and a “stable” state is obtained, for which no further
electrochemical changes are observed. As a result, the catalyst still remains more active
than pure Au/C and more selective than the initial AuPd/C. When the UPL is high enough
to induce Au dissolution and redeposition (ADP-1.6) in parallel to a significant Pd
dissolution, the catalyst degradation results in an alloyed core surrounded by an Au shell.
Thus, the behavior approaches the one of pure Au, although the alloying effect of Pd still
present in the core induces a small shift in the ORR onset potential.
For the first time, this study correlates the catalyst degradation with the H2O2 selectivity
using AuPd as a model catalyst. The conclusions are particularly important and can be
extended to similar bimetallic catalysts (as Au-Pd, Hg-Pd, HgPt…), which are considered
promising for the on-site electrocatalytic H2O2 production in fuel cells or other related
electrochemical reactions. Indeed, sudden potential changes or spikes, recurring during
start-stop of fuel cells, can induce significant surface composition changes within few cycles
that might substantially change the electrochemical behavior. Dedicated strategies either
in catalyst material design or in control of operational conditions have to be considered for
an effective employment of a direct H2O2 synthesis approach. More generally, this work
emphasizes the importance of fundamental long-term stability investigations for complex
electrochemical reactions where selectivity is a crucial performance indicator.
- On-demand H2O2 Production: a Parallel Study of Electro- and Heterogeneous Catalysis
| 115
- On-demand H2O2 Production: a Chapter 9
Parallel Study of Electro- and Heterogeneous
Catalysis 16
——————————————————————————————————————————
Electrocatalysis and heterogeneous catalysis are often progressing separately, despite
working towards the same goals. One example is the on-site production of H2O2, which is
highly desired as an alternative to the costly and inefficient centralized production.
Efficient catalyst design for the on-site synthesis by either small heterogeneous (i.e. in the
direct synthesis) or electrocatalytic (i.e. fuel cells/electrolyzers) reactors requires the
understanding of reaction mechanisms as well as of the role of the active sites in order to
design better catalysts. This Chapter is therefore dedicated to the comparison of the
heterocatalytic and electrocatalytic synthesis of H2O2 and it intends to highlight the
communalities and differences of the two processes by using the same carbon supported
Au-Pd/C nanocatalyst. This combined approach can open a new perspective for the future
studies in these two fields. For instance, by studying separately the half reactions
occurring simultaneously (HOR and ORR) using a new electrocatalytic system, the “floating
cell”, it is possible to explain the better performances of AuPd/C in terms of heterocatalytic
H2O2 productivity.
——————————————————————————————————————————
16 Parts of this chapter have been already published in:
E. Pizzutilo*, On-demand H2O2 production: a study at the border between electro and heterogeneous catalysis (in
preparation)
There are therefore numerous verbal quotes from that publication. Some of the figures present in
the publication have been re-printed or modified.
Chapter 9
116 |
9.1 Common Goals of Electrocatalysis and Heterogeneous
Catalysis
Electrocatalysis and liquid phase heterogeneous catalysis are two connected branches of
catalysis; in fact, electrocatalysis can be regarded as the ‘interface’ between
electrochemistry and liquid phase heterogeneous catalysis. A liquid phase heterogeneous
catalytic system in the presence of charged components that accumulate at the liquid-solid
interface forming an electrified interphase can be considered as an electrochemical system
[320]. However, despite this general idea, researches in these fields often work separately
to one another with limited exchange of information, from which both fields would highly
take advantage. Papers dealing with synergies and contrasts between the fields are limited
[320-324]. Nevertheless, considering the existing literature, it is evident that scientists are
often looking at the same processes [324] such as i) adsorption/desorption processes, ii) CO
oxidation [325, 326], iii) alcohol (ethanol, methanol, glycerol and other polyols) oxidation to
chemical intermediates in heterogeneous catalysis [327-330] or total oxidation to electrical
energy in electrocatalysis [331-334] and iv) processes involving O2 such as for the ORR,
oxidative dehydrogenation (ODH) [321] of hydrocarbons and the formation of H2O2 [3] via
direct heterocatalytic synthesis [19, 20, 227, 335] or electrocatalytic synthesis [89, 90, 95,
97, 105, 125, 131].
9.2 Electrocatalytic vs. Heterocatalytic synthesis of H2O2:
Related Properties
This study will compare and contrast these H2O2 production methods using the same
catalyst materials, namely carbon supported AuPd/C homogeneous alloys as well as Au/C
or Pd/C particles between 3-5 nm. Catalyst characterization is reported in Chapter 4.
Several studies using AuPd as a catalyst have been published over the last years for both
the H2O2 electrocatalytic synthesis [105, 122, 126, 127, 226] and the direct heterocatalytic
synthesis [20, 189, 336, 337]. These experimental works were also supported by
independent first principle studies [128-130, 220-222, 338]. From these comparative
results, we hope to start a valuable discussion that can be the basis of future positive
collaboration and reciprocal influences between these two fields.
Despite the different described mechanisms (see the dedicated Electrocatalytic and
Heterocatalytic sections in Chapter 1), experimental conditions and setups, it is possible to
identify related properties and parameters. For instance, i) in the direct synthesis the
amount of reagent (H2 and/or O2) consumed to yield H2O/H2O2 in a sequential
hydrogenation reaction (O2+2H22H2O / O2+H22H2O2) is described in terms of
conversion; in the electrochemical synthesis, the ORR activity can be regarded ability of
- On-demand H2O2 Production: a Parallel Study of Electro- and Heterogeneous Catalysis
| 117
a catalyst to reduce/consume O2 in a 4-/2-electron process to H2O/H2O2 under applied
potential conditions (O2+4e-+4H+2H2O / O2+2e-+2H+2H2O2). In a comparison between
different catalysts, the activity can be defined as the measured reduction current (Id) at a
defined fixed potential (fixed Ed). ii) For both systems, the peroxide selectivity (SH2O2)
represents the amount of produced H2O2 relative to the amount of H2/O2 and electrons
consumed (the remaining being converted to H2O). The selectivity is defined in terms of
converted moles (molesH2O2/molesH2/O2) in catalysis and of partial peroxide current (Iper/Itot;
the currents are proportional to the converted moles too) in electrocatalysis. iii) In the
electrocatalytic synthesis the peroxide reduction reaction (PRR) activity indicates
the rate at which H2O2 is reduced/consumed in a 2-electron process to H2O under applied
potential conditions (H2O2+2e-+2H+2H2O); similarly, the chemical peroxide
degradation expresses the percentage amount of initial H2O2 consumed during the
reaction to yield H2O (H2O2+H22H2O) in a hydrogenation process. The PRR/degradation
reactions are particularly critical for application. Indeed, the stability of H2O2 in the
reaction environment is, together with the selectivity, crucial for an efficient catalyst
utilization production in real systems. Finally, (iv) the overall amount of H2O2 synthesized
during the electrochemical and chemical reaction can be identified respectively with the
peroxide current (Iper) and the peroxide productivity. All these identified parameters
are summarized in the following Table 9.1.
Table 9.1 Traditional schematization of the electrocatalytic and catalytic reaction involved in the
H2O2 synthesis
Direct synthesis
Electrocatalytic
Synthesis
(i) Oxygen Hydrogenation
O2+2H22H2O O2+H22H2O2
ORR O2+4e-+4H+2H2O O2+2e-+2H+2H2O2
-Conversion % 𝑓𝑖𝑛𝑎𝑙 𝑚𝑜𝑙𝐻2
𝑖𝑛𝑖𝑡𝑖𝑎𝑙 𝑚𝑜𝑙𝐻2 -activity Onset and/or Id @ fixed Ed
(ii) Peroxide selectivity % 𝑚𝑜𝑙𝐻2𝑂2
𝑐𝑜𝑛𝑠𝑢𝑚𝑒𝑑 𝑚𝑜𝑙𝐻2 Peroxide selectivity % 𝑆𝐻2𝑂2
= 200 ×𝐼𝑟 𝑁⁄
𝐼𝑟 𝑁⁄ − 𝐼𝑑
(iii)
Peroxide degradation/ Hydrogenation
H2O2+H22H2O PRR H2O2+2e-+2H+2H2O
-Degradation % 𝑓𝑖𝑛𝑎𝑙 𝑚𝑜𝑙𝐻2𝑂2
𝑖𝑛𝑖𝑡𝑖𝑎𝑙 𝑚𝑜𝑙𝐻2𝑂2 -activity Id @ fixed Ed
(iv) Peroxide production [molH2O2 µgmetal-1] Peroxide current Iper = Ir/N [mA µgmetal-1]
9.3 Electrocatalytic vs. Heterocatalytic synthesis of H2O2:
Synergies and Differences
Having these comparative metrics in mind, the actual behavior of the same catalysts will
be compared in the two systems (collection of results is shown in Figure 9.1 to Figure 9.4 in
such a way that a direct comparison of related properties can be easily visualized).
Chapter 9
118 |
9.3.1 Conversion vs. ORR Activity
Figure 9.1 Comparison of a) conversion during 30 min of direct synthesis and b) ORR
electrocatalytic activity obtained for Au/C (red), AuPd/C (blue) and Pd/C (green).
i) The catalyst conversion rate and the ORR activity are shown in Figure 9.1a and Figure
9.1b respectively. In both cases, Pd/C is the catalyst that exhibits better performances than
that of Au/C and AuPd/C. Indeed, after 30 min of direct synthesis with Pd/C approximately
23±3% of the reagent initially present in the catalytic system has been converted. In a
similar experiment, Au/C appears to be a bad catalyst for such reaction as only 3±1% of the
reagent is converted. The comparison of the ORR activity is slightly less straightforward,
especially for those readers who are less familiar with electrochemistry. As a convention,
ORR catalyst are compared on the basis of their specific activity (SA, mA cm-2) or mass
activity (MA, mA g-1) at 0.9 VRHE [311]. However, this convention is not applicable with
such catalysts as for Au/C no reduction current is observed above ∿0.55 VRHE. Therefore, to
compare the catalysts performances, it is possible either to use the onset potentials, or the
measured current at a specific potential: the more active a catalyst is the higher the onset
potential and measured current. Following such criteria, is evident that Pd/C is the
catalyst with the highest activity as its onset potential is ∿0.95 VRHE. For potentials lower
than 0.7 VRHE, the polarization curve of Pd/C shows the typical plateau-like behavior
characteristic of the diffusion limiting condition (for a 4-electron process) that occurs in the
RDE due to the cell hydrodynamics and O2 concentration in the electrolyte. Indeed, around
0.4 VRHE the specific current is ∿4.4 mA cmgeo-2 within the error of theoretical limiting
current (4.5 mA cmgeo-2). At the same potential neither Au/C nor AuPd/C reach the
diffusion limited condition, as the reduction mechanism changes. Please note that, even if
Au/C is not so active in our system, it has been shown that the ORR kinetics and
mechanism of Au can vary significantly with support, crystallographic orientation, size and
pH [91, 93, 94]. The alloyed AuPd/C catalyst shows in both systems intermediate
performances between those of the pure metals, as the reduction mechanism and the
number of exchanged electrons is changing proportionally to the composition [127].
- On-demand H2O2 Production: a Parallel Study of Electro- and Heterogeneous Catalysis
| 119
9.3.2 Catalytic vs. Electrocatalytic H2O2 Selectivity
Figure 9.2 Comparison of peroxide selectivities in the a) direct synthesis and b) electrocatalytic
synthesis obtained for Au/C (red), AuPd/C (blue) and Pd/C (green).
ii) In Figure 9.2a and Figure 9.2b is shown the SH2O2 in the direct and electrocatalytic
synthesis respectively. Here, the highest SH2O2 is obtained for Au/C. In the direct synthesis,
it is 75±15%, whereas for Pd/C is only 21±3%. The high error bar obtained for Au/C is due
to the very low conversion of such catalyst (Figure 9.1a). Also in this case, the alloy shows
intermediate selectivity between the pure metals.
While in the direct synthesis only the catalyst selectivity during the overall reaction (here,
30 min) can be derived, with electrochemical characterization it is possible to have “on-
line” measurement. This results in a selectivity curve that is depending primarily on the
applied potential. The most selective material among the considered is also in the
electrochemical system the pure Au/C with SH2O2 varying between ∿80-85% and ∿100% in
the considered potential range (above 0.55 VRHE selectivity is not considered as Au/C is no
longer active for the ORR) and thus is in line with the results observed in the catalytic
system. Similar trends are also to be find in literature for carbon supported Au [95].
Nevertheless, it is shown in the previous Chapter 7 that the SH2O2 of both unsupported Au
and poly-Au is constant around ∿95%. This apparent discrepancy can be tentatively
adduced to the presence of the carbon layer where the H2O2 can be further reduced (see
Chapter 8). Pd/C on the other hand has selectivity close to 0% since Pd is reducing O2 in a
full 4-electron pathway [73, 81] as observed already from the reduction current in Figure
9.1b. This is true for almost all potentials, except below ∿0.25 VRHE, at which SH2O2 is
increasing to ∿10% (due to hydrogen coverage in the HUPD region as discussed in Chapter
7) [85, 150]. Finally, the AuPd/C SH2O2 is intermediate with a maximum (∿40-45%) around
0.55-0.6 VRHE. The maximum observed for different composition (Chapter 7) could be
related to the simultaneous reduction of O2 to H2O2 and reduction of H2O2 to H2O, whose
rate depend on the potential [32].
Chapter 9
120 |
Interestingly, the SH2O2 of both AuPd/C and Pd/C appears to be higher in the catalytic
system. This is especially evident for Pd/C catalyst that in the electrocatalytic system show
no selectivity at all. Note, however, that Pd is likely to be poisoned and thus if some active
sites are being blocked the SH2O2 increases with the time as we have shown in a
potentiostatic experiments in Chapter 7 [127], probably due to spectator species [88, 253],
or also to methanol present in the direct synthesis. Indeed it was shown in electrocatalysis
that according to the mechanism proposed by Breiter at al. [339-341] during methanol
oxidation the formation of CO* intermediates can lead to poisoning of noble metals like Pt
and Pd [342]. Note, however, that in the direct synthesis high SH2O2 is also obtained in pure
H2O.
9.3.3 H2O2 Degradation vs. PRR Activity
Figure 9.3 Comparison of a) H2O2 degradation during 30 min of direct synthesis and b) PRR
electrocatalytic activity obtained for Au/C (red), AuPd/C (blue) and Pd/C (green).
iii) The H2O2 degradation and the PRR activity are shown in Figure 9.3a and Figure 9.3b
respectively. Such properties are critical for future application, as a catalyst should not
only be active and selective, but it should avoid further H2O2 consumption under
catalytic/electrocatalytic environment. Generally, catalysts with high conversion rate or
ORR activity (4-electron path) show higher H2O2 degradation rate [123] or PRR activity
[57, 343, 344]. No surprise, therefore, that after 30 min of catalytic reaction almost 71% of
the initial H2O2 is degraded once Pd/C is used as a catalyst. Instead, only 33% and almost
no H2O2 is degraded when AuPd/C and Au/C are tested. Similarly, in the electrocatalytic
system, Au/C is not active at any potential for the PRR, whereas Pd/C is the most active
material with the reduction current reaching the diffusion limited condition, this time due
to the concentration of H2O2 in the electrolyte. Such currents are not reached by the
AuPd/C catalyst that once again shows an intermediate behavior [126, 127, 302]. Note that
for the PRR activity we only currents are compared and not onset potentials. Indeed, the
onsets for both AuPd/C and Pd/C are around 0.8 VRHE, at which, the PRR switches to POR.
- On-demand H2O2 Production: a Parallel Study of Electro- and Heterogeneous Catalysis
| 121
9.3.4 H2O2 Productivity vs. H2O2 Current
Figure 9.4 Comparison of a) peroxide productivity during 30 min of direct synthesis and b)
peroxide current obtained for Au/C (red), AuPd/C (blue) and Pd/C (green).
iv) So far, direct synthesis and the electrocatalytic synthesis exhibits similar trends for
related properties. However, when it comes to what is really the key point of such process,
namely the H2O2 production, the behavior changes dramatically. The H2O2 production is
expressed by the productivity (Figure 9.4a) and by the H2O2 current Iper (Figure 9.4b). As
expected, the most selective material is also the most electrochemically productive: Iper of
Au/C tops around 0.2 VRHE with ∿0.16 mA µgmetal-1, double than the ∿0.08 mA µgmetal
-1
measured for AuPd/C around 0.55 VRHE. Note, however, that the onset potential of AuPd/C
is much higher than Au/C. Finally, on fresh clean Pd/C no Iper is observed (except around
HUPD) as O2 is getting reduced directly to H2O. This trend is also maintained in long
measurements. Indeed, after 30 min of potentiostatic measurement (see Chapter 7) Au still
has the highest productivity (1.33 *10-6 mol µgmetal-1).
As anticipated, the picture changes totally when analyzing the catalytic productivity.
Indeed, this time the less active material, with ∿0.06*10-6 mol µgmetal-1, is Au/C, while the
most active material, with ∿0.33*10-6 mol µgmetal-1, is AuPd/C. The pure Pd/C catalyst
instead shows intermediate productivity. In literature [123], it was shown that the highest
productivity is been reached around the Au:Pd composition 1:1-1:3. It has been attributed
to the high activity of Pd and the high selectivity of Au which are balanced in the alloy
leading to an overall synergetic effect and improved performance for the direct synthesis
of H2O2 [222, 337, 345].
9.4 Discussion
In this section, we will start a discussion around the observed discrepancy between the
H2O2 productions in the two different systems.
Chapter 9
122 |
9.4.1 Heterogeneous Catalysis of Electron-Transfer Reactions in
Solution
Prior to further considerations, it should be clear to the reader that a liquid phase
heterogeneous catalytic system can be regarded as an electrochemical system. Indeed,
between the two phases (solid and liquid) accumulation and/or depletion of charges occurs
in an electrified interphase [346]. Furthermore, catalytic reactions involving a change in
the oxidation state must involve an electron transfer either between adsorbed reactants
ions or through the metal catalyst. Spiro et al. published several papers on the
investigation of heterogeneous oxidation-reduction reactions catalyzed by electron transfer
through a solid material [261, 347-350]. His work on the ferricyanide-ferrocyanide
([Fe(CN)6]3-+e-[Fe(CN)6]3-) and iodine-iodide (3I-I3-+2e-) system catalyzed by Pt
revealed that the rates of the two separate electrochemical reactions are equal at the
“mixture potential” in the reaction mixture where only the catalytic reaction ([Fe(CN)6]3-
+3I-[Fe(CN)6]3- +I3-) is occurring [349]. With this in mind, the catalytic process has to be
regarded as the sum of coupled redox-reaction in equilibrium with no net charge transfer
(i.e. the anodic and cathodic partial contributions are equal). In electrochemical terms, this
implies that the reaction is occurring at the OCP or “mixed potential” and all
heterogeneous reactions can be considered to occur at the OCP set by the reaction
environment. The concept of mixed potential is commonly used in corrosion [351] and other
industrial processes like froth floating, mineral extraction and electroless plating [262]. As
for corrosion, a consequence of the realization that a certain reaction involves electron
transfer through the metal implies that the behavior of redox reactions in heterogeneous
catalysis can be predicted from electrochemical experiments alone [349].
Despite the differences and the limited exchanges between electrocatalysis and catalysis
community, the reader should be at this point aware that any consideration about a
heterogeneous catalytic system cannot disregard the electrochemical contribution. Besides
the initial studies in the 60s, to the authors knowledge only a handful of studies aimed at
bridging the gap between these two fields [320-324], and none of them specifically
addressed the case of the H2O2 synthesis.
9.4.2 Electron Transfer in the H2O2 Catalytic Direct Synthesis?
As mentioned in Chapter 1, the mechanism for the direct synthesis has been always
described as a sequence of hydrogenation steps without involving electron transfer through
the metal catalyst. However, revising literature on the heterocatalytic H2O2 synthesis, it is
possible to find some works indicating that the mechanism could be different than so far
believed, even though the related mechanistic discussions are limited. As an example,
Choudary et al. published several studies on the H2O2 synthesis in acidic environment and
in the absence of H2, by using hydrazine as a reduction agent over Pd and hydroxylamine
- On-demand H2O2 Production: a Parallel Study of Electro- and Heterogeneous Catalysis
| 123
over Au catalysts [352-354]. Furthermore, they claim that in an acidic environment the
SH2O2 increases from 10% to about 60% compared to neutral media [355]. In another
heterogeneous process involving the catalytic reduction of nitrite (NO2-) on AuPd catalysts,
Seraj et al. explicitly considered H2 as an electron donor [356]. Continuing, Abate et al.
suggested that the presence of protons is hindering the breaking of the O-O bond thus
favoring the H2O2 formation over H2O [357]. However, only recently an alternative reaction
mechanism to the well accepted hydrogenation mechanism has been firmly advanced by
Wilson and Flaherty [191]. Studying the steady-state H2O2 formation rates using a Pd
catalyst they found that protic media was needed to produce H2O2 and that rates increased
with increasing [H+]. Based on this observation the Wilson-Flaherty mechanism for the
direct synthesis is a non-Langmuirian mechanism where the H2O2 is formed in a pathway
that involves a water mediated proton-electron transfer. It consists of a decoupled redox
reaction with a short-range electron transfer within the metal Pd catalyst. The electrons
for the ORR (O2+2H++2e- H2O2) are provided by the HOR (H2 2H++2e-); thus, the two
reactions can occur at two different sites provided that the catalyst is conductive.
Despite their enormous contribution to the advances in the knowledge around the direct
synthesis, their analysis was limited to the heterogeneous catalytic study and to only a
pure Pd catalyst. However, i) if electron transfer in the metal is involved, an
electrochemical study can undoubtedly add value to understanding of the reaction
mechanism; furthermore, ii) it is important also to include in the mechanistic discussion
bimetallic catalysts (i.e AuPd) that compared to pure Pd exhibit much higher productivity.
In the following Figure 9.5 and Table 9.2 are reported the proposed reaction mechanism for
the formation of H2O2 and H2O thorugh the ORR occurring in the direct synthesis.
Figure 9.5 Proposed mechanism fort the ORR in the direct synthesis.
This schematic (similar to Figure 1.5 and Figure 1.11) was inspired from the elementary
steps on supported Pd cluster proposed by Wilson and Flaherty [191]. However, in their
Chapter 9
124 |
analysis they considered electrochemical steps only for the formation of H2O2 and only
chemical steps without electron transfer through the metal for the further formation of
H2O. However, if the first reaction has to occur with electron transfer, we suggest that also
the other reactions involving oxygen/hydroperoxo intermediate dissociation might involve
electron transfer as for the electrochemical ORR. Of course, compared to the pure
electrochemical ORR, the presence of adsorbed H* from the dissociation on molecular
hydrogen (H2) should be also considered in parallel (see Figure 9.5 and Table 9.2) [358].
Note that the adsorption/desorption of H* (H2->2H*->2H+) depends on the applied or
operation potential [359, 360].
Table 9.2 Proposed series of elementary steps involved in the formation of H2O2 and H2O during the
direct synthesis on a metallic catalyst. *is a pure chemical process.
H2O2 Formation O2 intermediate
dissociation to H2O
Hydroperoxo intermediate
dissociation path to H2O
H2O2 decomposition
H2(g)2H* H*H++e-+* O2(g)O2* O2*+H++e-OOH* OOH*+H++e-H2O2* H2O2*H2O2(g)
O2(g)O2* O2*2O* O*+H++e-OH*/ O*+H*OH* OH*+H++e-
H2O*/OH*+H*H2O* H2O*H2O(g)
O2(g)O2* O2*+H++e-OOH* OOH*OH*+O* O*+ H++e- OH*/ O*+ H* OH* OH*+H++e- H2O*/OH*+H*H2O* H2O*H2O(g)
H2O2*2OH*
H2O2
disproportionation*
H2O2(g)H2O(g)+O2(g)
9.4.3 Floating Cell Study of the Coupled ORR/HOR Electrochemical
Reactions
One of the main benefits of using electrochemical methods to analyze reactions involving
electron transfers is the possibility to study independently the different involved reactions.
In the specific case of the direct synthesis of H2O2, this means that the HOR and the ORR
occurring simultaneously in the heterogeneous direct synthesis can be measured
separately in two consecutive electrochemical experiments. Some clear advantages of such
approach are i) the possibility of understanding the active sites for such reaction and
guiding the design of catalysts that are active for both reactions and as well selective to
H2O2; ii) the possibility of studying the mixed potential following the mixed potential
theory. In the case of H2O2 synthesis, the catalyst electrochemical behaviour under non-
diffusion-limited condition is recorded (Figure 9.6 and Figure 9.7) using the newly
developed floating cell [246].
In this case, H2 or O2 are provided directly onto the catalyst surface without being limited
by their solubility in the electrolyte and/or by the electrode rotation. Thus, much higher
current can be obtained: the ORR polarization curves obtained in the floating cell are more
than ten-fold higher than that obtained in RDE under diffusion-limited condition (see
Figure 9.6), whereas the onset potentials are comparable. Such technique opens the way
to collecting kinetic information; however, it is still under development and, for instance,
the effective catalyst utilization has yet to be clarified.
- On-demand H2O2 Production: a Parallel Study of Electro- and Heterogeneous Catalysis
| 125
Figure 9.6 HOR (green) and ORR (red) polarization curves of a) Au/C, b) AuPd/C and c) Pd/C
measured in the floating cell and the ORR (black) measured in the RDE at 900 rpm. Magnification
of the ORR onset area (inset). Scan rate: 50 mV s-1.
The comparison of ORR and HOR behaviours for different catalysts (depicted in Figure 9.7)
confirm once again the highest activity of the Pd/C sample, whereas the Au/C is the least
active catalyst for both reactions (no HOR activity observed).
Figure 9.7 HOR (dotted line) and ORR (full line) polarization curves of Au/C (red), AuPd/C (blue)
and Pd/C (green) measured in the floating cell. Scan rate: 50 mV s-1.
Chapter 9
126 |
When the H2O2 is produced electrocatalytically, the electrons and the protons are provided
directly through the electrode and the acidic supporting electrolyte respectively. Thus, the
higher SH2O2 results in higher productivity. Instead, when the H2O2 is produced
catalytically, the HOR is the source of electrons while the partially acidic environment
(CO2 dissolved in the solvent result in carbonic acid formation with a pH around 3-4)
provide the protons. Wilson and Flaherty showed indeed that H2O2 is being formed on
Pd/SiO2 catalyst only in protic solvents [191]. Continuing, as the source of electrons from
HOR is lacking, the catalytic productivity of Au/C is close to zero, despite its SH2O2. In
terms of mixed potential theory, being the anodic reaction (H2H++2e-) close to zero at all
potentials, the mixed current Imix must be also close to zero (see Figure 9.8) and so does
also the mixed reaction rate mix. Concerning the other two catalysts, their mixed
potentials are very similar, around 0.4-0. 5 VRHE, slightly higher for Pd/C. The mixed
currents are ∿0.9+0.4 mA µgmetal and ∿2.3+1.4 mA µgmetal for AuPd/C and Pd/C
respectively. Despite the high error bars related to the undetermined catalyst utilization in
the floating cell, it is anyway possible to trace the trend of the mixed current with the
composition (Imix increasing with the Pd content). This trend does not take in account the
SH2O2; however, to estimate the productivity rate (per=Iper/nF) it is necessary to take it in
account (i.e. SH2O2 from RRDE measurement ∿29% at Emix) when calculating the H2O2
current (Iper=SH2O2*Imix/(2-SH2O2)). The so estimated per show a volcano-like behaviour
(Figure 9.8), with the mixed alloy having the highest productivity around ∿0.85±0.22 *10-9
mol µgmetal-1 s-1.
Figure 9.8 SH2O2 (from the RRDE measurement), Imix and the estimated per plotted vs. Emix.
The heterocatalytic productivity (of 30 min reaction) in Figure 9.4a shows exactly the same
trend with a maximum for AuPd/C whose catalytic rate cat is ∿0.19±0.01 *10-9 mol µgmetal-1
s-1. As a proof of concept, both per and cat are of the same order of magnitude; the
difference in the values could be explained as follow: i) the per estimation did not consider
- On-demand H2O2 Production: a Parallel Study of Electro- and Heterogeneous Catalysis
| 127
the reduction/degradation of accumulated H2O2 that can occur in the 30-min reaction
under catalytic condition. Reducing the reaction time to 2 min was shown to limit also the
degradation, resulting in a three-/four- fold productivity increase [123]. ii) The reaction
conditions are slightly different in the two systems and this of course can have some
impact; for instance, the dissolved CO2 forming carbonic acid or other spectator species can
poison catalysts active sites thus changing the overall activity and selectivity. Indeed,
electrochemical studies suggest that the SH2O2 of Pd can even change due to surface
oxygenated-species and adsorbed hydrogen [88, 127]. The polarization curves shown in
Figure 9.1 to Figure 9.4 are obtained in suprapure environment after initial cleaning
cycles. Clean Pd, as discussed until now, reduces O2 in a 4-electron pathway with a SH2O2
close to zero like Pt. However, when holding the potential (i.e. at Emix) the Pd/C SH2O2 is
also changing (Figure 7.9 and Figure 9.9). This can explain the significant catalytic
productivity of Pd/C observed in Figure 9.4a despite its very low SH2O2.
Figure 9.9 Measured Pd/C selectivity during 30 min potentiostatic experiment (@0.5 VRHE) in O2
saturated 0.1M HClO4. Rotation: 900 rpm. Er: 1.28 VRHE.
9.4.4 Two Half Reactions in Catalysis and Electrocatalysis
So far, we have shown that electrochemical methods can be used to describe similarities
and differences between electro- and heterocatalytic processes involving electron transfer.
It is in practice possible to exploit electrochemical method further than for a purely
descriptive aim; ideally, electrochemistry can provide fast and clean tools to predict the
reaction rates and heterocatalytic behavior of a set of catalysts. For instance, one could
create a material library for bimetallic catalyst, scan the electrochemical
composition/activity/SH2O2 trend and from this predict the composition that will have the
best performances in the heterocatalytic process. This initial study described the
productivity trend of a bimetallic Au-Pd based catalyst in the direct synthesis of H2O2. The
Chapter 9
128 |
highest reaction rate (per) predicted for the bimetallic AuPd/C results from the interplay of
the Pd high HOR/ORR activity and of the Au high SH2O2. Li et al. suggested that the role of
the catalyst in the direct synthesis might just be to cleave the molecular H2 and supply H
atoms for the O2 hydrogenation [222]. Wilson et al. added that the catalyst need to be
active for the HOR to oxidize H2, providing protons and electrons [191]. The following ORR
occurs at a different site of the Pd. Extending this concept to a bimetallic catalyst, our
results and the literature discussed so far suggest that the HOR occurs in an active site
like Pd cluster, whereas the ORR to H2O2 should occur in a more selective active site (i.e.
pure Au, Pd monomer) that prevent the O2 cleavage which is supported by the observations
of Hutchings [20] that it is possible to synthesize catalysts with 100% hydrogen utilization
suggesting the sites for H2O2 synthesis are different than the sites for H2O2 hydrogenation.
Actually, Pd dimers or bigger cluster are required to have hydrogen adsorption [104]; thus
a catalyst with atomically dispersed Pd that showed interesting electrochemical behavior
would not be an efficient catalyst for the direct synthesis [105].
The synergetic ensemble effect showed by the AuPd/C can be in principle obtained by
combining any selective metal with an active metal for HOR as shown in Figure 9.10a.
Nevertheless, the ideal heterocatalyst should have active sites for the HOR that are
however not active for the 4-electrons ORR. The risk otherwise is to have, yes, the highest
productivity, but a selectivity far less than the ideal 100%. For instance, the AuPd/C
catalyst, despite being a standard catalyst with high productivity for the direct synthesis,
shows only around 50% selectivity (Figure 9.2a). It would be more reasonable if the two
metals would be exploited separately to maximize their overall efficiency, for example in
an electrocatalytic reactor. In this case, the reactions of interest can be separated and the
appropriate catalyst chosen to maximize both the activity for the electron source at one
electrode and the SH2O2 at the other electrode (see Figure 9.10b).
Figure 9.10 a) Schematic representation of the two half reactions in a catalytic reactor. b)
Schematic representation of the two half reactions in a an electrocatalytic reactors.
- On-demand H2O2 Production: a Parallel Study of Electro- and Heterogeneous Catalysis
| 129
Moreover, the electrochemical approach has several other advantages: i) it combines
chemical and electrical production; ii) ambient pressure; iii) increased safety as gases are
provided separately to the electrodes; iv) flexibility of media and of cell configuration. Still
the electrochemical reactor concept requires improvements due to limiting factors as
temperature, high H2O2 reactivity in presence of impurities (i.e. Fe) causing membrane
degradation and the diffusion limitation when using O2 saturated catholyte.
9.5 Conclusion
This Chapter described comprehensively the H2O2 production in a heterogeneous catalytic
reactor and in an electrochemical cell. The following four parameters that similarly
describes the reactions in the two systems were characterized: i) O2 conversion/ORR
activity, ii) H2O2 chemical/electrochemical selectivity, iii) H2O2 degradation/PRR activity,
iv) H2O2 chemical productivity/ peroxide electrochemical productivity (expressed by the
H2O2 current). Experimental results indicate that in terms of the first three parameters
(activity, SH2O2 and degradation) the different catalysts exhibit similar trends with the
composition, whereas their productivity differs strikingly in the considered system. Indeed,
while in the electrochemical cell the highest SH2O2 of Au/C corresponds to the highest H2O2
current, in the chemical reactor the highest productivity is only achieved with the
bimetallic AuPd/C, whose SH2O2 is intermediate between the pure metals. This discrepanc
has been addressed by considering the chemical reaction as resulting from two
electrochemical reactions (HOR and ORR) with electron transfer. Therefore, an
electrochemical method, like the newly developed floating cell, can be used of advantage to
study separately these electrochemical reactions in a non-diffusion limited manner. The
collected HOR and ORR polarization curves were finally elaborated in the frame of the
mixed potential theory and correlated with the results obtained in the chemical reactor.
The estimated H2O2 formation reaction rate (per) resulted in the same order of magnitude
of the measured chemical reaction rate (cat). Furthermore, the productivity trend with the
maximum obtained for AuPd/C was confirmed with our electrochemical approach.
In conclusion, we showed the important and complementary information that can be
obtained by combining chemical with electrochemical methods and that can be exploited in
future studies also to investigate and understand the catalytic reaction mechanisms and
the role of active catalyst sites. Even though is out of the focus of this work, it can be
suggested that this approach is valid also for other electro-/heterocatalytic studies, such as
glycerol electro- [361, 362]/hetero- [363, 364] oxidation, ethanol electro- [110, 334,
365]/hetero- [366] oxidation and other more.
Chapter 10
130 |
- Final Conclusions and Outlook Chapter 10
As discussed already in the introduction, the anthraquinone process is considered the
standard method for the current industrial H2O2 production. Despite its high selectivity,
this process is uneconomic at small scales especially if the final application requires low
content of H2O2. Therefore, the development of an alternative small-scale on-site
production method is highly desired. Both the catalytic direct synthesis process from
molecular O2 and H2 as well as the electrocatalytic synthesis process from O2 as an
intermediate product of the ORR, are considered promising from both the academic and
industrial sectors. Among the different candidates, in recent years Pd-based alloys have
clearly gained the attention in both the electrocatalysis and heterogeneous catalysis
communities. To meet the industrial requirements, the undesired, unselective H2O
production (resulting from the H2O2 hydrogenation or reduction) should be avoided and the
selectivity maximized. Nevertheless, the design of new selective catalysts cannot disregard
the actual reaction mechanism.
The comparison of electrocatalytic and the catalytic H2O2 synthesis using the same catalyst
material as in Chapter 9 is one of the major contributions of this thesis work. Although
different nomenclatures are conventionally used, similar properties can be identified: i)
catalytic conversion and ORR activity, ii) catalytic/electrocatalytic selectivity, iii) H2O2
degradation and PRR activity and iv) catalytic/electrocatalytic productivity. These are
affected by the catalyst composition, with pure Au being i) the least active for conversion
and ORR, ii) the most selective and iii) the least active for H2O2 degradation and PRR.
Nevertheless, while behaving similarly the considered Au-Pd catalyst showed a
remarkable difference in the productivity (expressed in electrocatalysis by the H2O2
current, Iper) with AuPd showing higher productivity in the catalytic system and Au in the
electrocatalytic system. Following the recent work of Wilson et al. on the direct H2O2
synthesis using a Pd nanocatalyst, a catalytic reaction mechanism on an Au-Pd
nanocatalyst based on the electrochemical ORR and HOR reactions have been proposed.
The oxidation of H2 may act as a source of electrons which are afterwards used for the
reduction of O2 to H2O2. The mechanism is consistent with the observation that Au is the
least productive heterocatalyst while having a selectivity close to 90%. Indeed, as it is not
active for the HOR, electrons are not available for the ORR. On the other hand, in an
electrocatalytic system electrons are provided externally to the electrodes and Au, being
the most selective, results also to be the most productive. An ideal catalyst for the direct
catalytic synthesis should exhibit high HOR activity as well as ORR selectivity to H2O2;
generally, this can be achieved by combining (in the right ratio) elements with opposite
absorption properties (i.e. as in the case of Au-Pd). On the other hand, for an ideal
electrocatalyst a user should focus on the selectivity to maximize the chemical conversion;
- Final Conclusions and Outlook
| 131
this would be the case using Au. Nevertheless, in applications where a compromise
between energy conversion and chemical synthesis is possible, other composition showing
an intermediate activity and selectivity can of course be advantageous (as is the case with
using Au3Pd, AuPd and AuPd3 as shown in Chapter 7).
Recognizing that H2O2 is not only produced through chemical steps, but also through
electron transfer steps, is a huge advance in our comprehension of the role of catalyst
active sites. This can open new ways of characterizing these materials, which can help in
future catalyst development. Indeed, by electrochemical methods it is in principle possible
to study separately the single electrochemical reactions (i.e. ORR and HOR in this case)
and describe the overall catalytic process with the additivity principle of the mixed
potential theory. Even though a precise description of the whole catalytic reaction (taking
in account pressure, mass transport and influence of the reaction media) is still out of
reach, in this thesis, the use of a newly developed electrochemical system, the floating cell,
based on triple phase boundary is proposed. Thanks to the recorded ORR/HOR kinetic
curves it was possible to derive the peroxide formation reaction rates with the additivity
principle (per), which is in the same order of magnitude of the measured chemical rate
(cat). This new approach will lay the foundations for future collaborations between the
electrocatalysis and heterocatalysis communities aiming at catalyst development for new
chemical synthesis applications.
The determination of whether one of these two alternative H2O2 syntheses is more
advantageous than the other is out of the scope of this thesis. However, during the past
three years of hands-on experience in both fields, the author has gained an overview of the
critical issues (regardless of catalyst activity/selectivity) that should be addressed to
achieve commercial viability of one or both processes. In particular, the safety of catalytic
reactors needs to be improved as the mixture of H2 and O2 under pressure can reach a
critical explosive concentration if experimental design is not strictly addressed. This can
also limit the amount of H2O2 that can be produced. Small flow membrane reactors could
actually prevent the mixing of the two gases and should be further investigated. On the
other hand, while safety is not an issue for an electrocatalytic reactor, the durability of
both the reactor and the catalyst should not be neglected as it can have important impact
on the running costs of the system. Indeed, H2O2 can degrade the membrane when metallic
impurities are present and the catalyst itself can change behavior completely if it
degrades. The degradation behavior under various simulated electrochemical treatments
has been shown in Chapter 5, Chapter 6 and Chapter 8. To the reader, it must be clear
that high activity and selectivity are alone pointless if the catalyst itself is not stable
enough for a long-time application. This is especially important for those bimetallic
catalysts that rely on the ensemble effect generated by the co-presence of both metals on
the surface. For this, even mild dissolution conditions can result in a rapid evolution of the
Chapter 10
132 |
surface composition with dramatic performance changes. A very small amount of
dissolution is also observed at OCP in the presence of O2 (Chapter 6) implying that the
composition could change also in a catalytic system, although rates are much lower
compared to a start and stop condition that occurs at the cathode. Continuing, a common
obstacle to both systems derives from the H2O2 degradation; indeed, it is desirable to
produce neutral H2O2, without acid or basic solution and without additive, which could
prevent its degradation and decomposition.
The results presented in this work require further experimental validation of the
electrochemical approach, based on the floating cell, to study chemical reactions involving
electron transfers. Furthermore, the authors opinion is that experimental measurements
on real electrochemical system (fuel cell or electrolyzer) using for example Au/C as catalyst
should follow this fundamental study.
References
| 133
References
1. Pesterfield, L., The 100 Most Important Chemical Compounds: A Reference Guide (by Richard L.
Myers). Journal of Chemical Education, 2009. 86(10): p. 1182.
2. Zellmer, L., The 100 most important chemical compounds: A reference guide. Library Journal, 2007.
132(20): p. 158-158.
3. Campos-Martin, J.M., G. Blanco-Brieva, and J.L.G. Fierro, Hydrogen peroxide synthesis: An outlook
beyond the anthraquinone process. Angewandte Chemie-International Edition, 2006. 45(42): p. 6962-
6984.
4. http://h2o2.evonik.com/product/h2o2/en/application-areas/pages/default.aspx.
5. http://www.prweb.com/releases/hydrogen_peroxide/bleaching_pulp_paper/prweb8067430.htm.
6. http://www.essentialchemicalindustry.org/chemicals/hydrogen-peroxide.html.
7. Walsh, P.B., Hydrogen-Peroxide - Innovations in Chemical Pulp Bleaching. Tappi Journal, 1991. 74(1):
p. 81-83.
8. Jones, C.W., Application of Hydrogen Peroxide and Derivates. Royal Society of Chemistry, London,
1990.
9. Klemm, E., et al., Direct Gas-Phase Epoxidation of Propene with Hydrogen Peroxide on TS-1 Zeolite in
a Microstructured Reactor. Industrial & Engineering Chemistry Research, 2008. 47(6): p. 2086-2090.
10. Shin, S.B. and D. Chadwick, Kinetics of Heterogeneous Catalytic Epoxidation of Propene with
Hydrogen Peroxide over Titanium Silicalite (TS-1). Industrial & Engineering Chemistry Research,
2010. 49(17): p. 8125-8134.
11. Robinson, D.J., et al., Oxidation of thioethers and sulfoxides with hydrogen peroxide using TS-1 as
catalyst. Physical Chemistry Chemical Physics, 2000. 2(7): p. 1523-1529.
12. https://www.gminsights.com/industry-analysis/hydrogen-peroxide-market.
13. Edwards, J.K. and G.J. Hutchings, Palladium and Gold-Palladium Catalysts for the Direct Synthesis
of Hydrogen Peroxide. Angewandte Chemie-International Edition, 2008. 47(48): p. 9192-9198.
14. Manchot, W., Ueber Sauerstoffactivirung. Justus Liebigs Annalen der Chemie, 1901. 314(1-2): p. 177-
199.
15. Riedl, H. and G. Pfleiderer, Production of hydrogen peroxide. US Patent 2158525. US Patent 2158525,
1939.
16. Pfleiderer, G., H. Riedl, and W. Deuschel, Process for the preparation of selective nickel catalysts,
DE801840C. 1950.
17. Edwards, J.K., et al., The Direct Synthesis of Hydrogen Peroxide Using Platinum-Promoted Gold-
Palladium Catalysts. Angewandte Chemie-International Edition, 2014. 53(9): p. 2381-2384.
18. Astbury, G.R., Safe Scale-Up of Oxidation by Hydrogen Peroxide in Flammable Solvents. Organic
Process Research & Development, 2002. 6(6): p. 893-895.
19. Edwards, J.K., et al., Advances in the direct synthesis of hydrogen peroxide from hydrogen and oxygen.
Catalysis Today, 2015. 248: p. 3-9.
20. Edwards, J.K., et al., Switching Off Hydrogen Peroxide Hydrogenation in the Direct Synthesis Process.
Science, 2009. 323(5917): p. 1037-1041.
21. Viswanathan, V., H.A. Hansen, and J.K. Norskov, Selective Electrochemical Generation of Hydrogen
Peroxide from Water Oxidation. Journal of Physical Chemistry Letters, 2015. 6(21): p. 4224-4228.
134 |
22. Kormann, C., D.W. Bahnemann, and M.R. Hoffmann, Photocatalytic Production of H2o2 and Organic
Peroxides in Aqueous Suspensions of Tio2, Zno, and Desert Sand. Environmental Science &
Technology, 1988. 22(7): p. 798-806.
23. Jakob, H., et al., Peroxo Compounds, Inorganic, in Ullmann's Encyclopedia of Industrial Chemistry.
2000, Wiley-VCH Verlag GmbH & Co. KGaA.
24. Charles, R.H., Production of hydrogen peroxide by the partial oxidation of alcohols. 1949, Google
Patents.
25. Leyshon, D.W., R.J. Jones, and R.N. Cochran, Production of hydrogen peroxide. 1993, Google Patents.
26. Sanderson, W.R., Cleaner industrial processes using hydrogen peroxide. Pure and Applied Chemistry,
2000. 72(7): p. 1289-1304.
27. Venugopalan, M. and R.A. Jones, Chemistry of Dissociated Water Vapor and Related Systems.
Chemical Reviews, 1966. 66(2): p. 133-+.
28. Yamanaka, I. and T. Murayama, Neutral H2O2 Synthesis by Electrolysis of Water and O2.
Angewandte Chemie International Edition, 2008. 47: p. 3.
29. Siahrostami, S., et al., Enabling direct H2O2 production through rational electrocatalyst design (vol
12, pg 1137, 2013). Nature Materials, 2013. 12(12).
30. Ando, Y. and T. Tanaka, Proposal for a new system for simultaneous production of hydrogen and
hydrogen peroxide by water electrolysis. International Journal of Hydrogen Energy, 2004. 29(13): p.
1349-1354.
31. Drogui, P., et al., Oxidising and disinfecting by hydrogen peroxide produced in a two-electrode cell.
Water Research, 2001. 35(13): p. 3235-3241.
32. Drogui, P., et al., Hydrogen peroxide production by water electrolysis: Application to disinfection.
Journal of Applied Electrochemistry, 2001. 31(8): p. 877-882.
33. Jasinski, R.J. and C.G. Kuehn, Method for the electrolytic production of hydrogen peroxide. 1983,
Google Patents.
34. Sudoh, M., H. Kitaguchi, and K. Koide, ELECTROCHEMICAL PRODUCTION OF HYDROGEN
PEROXIDE BY REDUCTION OF OXYGEN. Journal of Chemical Engineering of Japan, 1985. 18(5):
p. 409-414.
35. Kuehn, C.G. and F. Leder, Three compartment electrolytic cell method for producing hydrogen peroxide.
1982, Google Patents.
36. Gyenge, E.L. and C.W. Oloman, Electrosynthesis of hydrogen peroxide in acidic solutions by mediated
oxygen reduction in a three-phase (aqueous/organic/gaseous) system Part II: Experiments in flow-by
fixed-bed electrochemical cells with three-phase flow. Journal of Applied Electrochemistry, 2003. 33(8):
p. 665-674.
37. Kastening, B. and W. Faul, Herstellung von Wasserstoffperoxid durch kathodische Reduktion von
Sauerstoff. Chemie Ingenieur Technik, 1977. 49(11): p. 911-911.
38. Alvarez-Gallegos, A. and D. Pletcher, The removal of low level organics via hydrogen peroxide formed
in a reticulated vitreous carbon cathode cell, Part 1. The electrosynthesis of hydrogen peroxide in
aqueous acidic solutions. Electrochimica Acta, 1998. 44(5): p. 853-861.
39. Oloman, C. and A.P. Watkinson, The electroreduction of oxygen to hydrogen peroxide on fixed bed
cathodes. The Canadian Journal of Chemical Engineering, 1976. 54(4): p. 312-318.
40. Oloman, C. and A.P. Watkinson, Hydrogen peroxide production in trickle-bed electrochemical reactors.
Journal of Applied Electrochemistry, 1979. 9(1): p. 117-123.
41. Anna Da Pozzo, L.D.P., Carlo Merli, Elisabetta Petrucci, An experimental comparison of a graphite
electrode and a gas diffusion electrode for the cathodic production of hydrogen peroxide. Journal of
Applied Electrochemistry, 2005. 35(4): p. 6.
42. Qiang, Z., J.-H. Chang, and C.-P. Huang, Electrochemical generation of hydrogen peroxide from
dissolved oxygen in acidic solutions. Water Research, 2002. 36(1): p. 85-94.
References
| 135
43. Yamada, N., et al., Development of Trickle‐Bed Electrolyzer for On‐Site Electrochemical Production of
Hydrogen Peroxide. Journal of The Electrochemical Society, 1999. 146(7): p. 2587-2591.
44. Alcaide, F., E. Brillas, and P.L. Cabot, Electrogeneration of hydroperoxide ion using an alkaline fuel
cell. Journal of the Electrochemical Society, 1998. 145(10): p. 3444-3449.
45. Brillas, E., F. Alcaide, and P.L. Cabot, A small-scale flow alkaline fuel cell for on-site production of
hydrogen peroxide. Electrochimica Acta, 2002. 48(4): p. 331-340.
46. Xu, F., et al., A new cathode using CeO2/MWNT for hydrogen peroxide synthesis through a fuel cell.
Journal of Rare Earths, 2009. 27(1): p. 128-133.
47. Lobyntseva, E., et al., Electrochemical synthesis of hydrogen peroxide: Rotating disk electrode and fuel
cell studies. Electrochimica Acta, 2007. 52(25): p. 7262-7269.
48. Murayamaa, T., et al., Catalytic neutral hydrogen peroxide synthesis from O2 and H2 by PEMFC fuel.
Catalysis Today, 2011. 164: p. 6.
49. Otsuka, K. and I. Yamanaka, One step synthesis of Hydrogen Peroxide through fuel cell reaction.
Elerrrochmica Aero, 1990. 35(2): p. 4.
50. Yamanaka, I., et al., A Fuel-Cell Reactor for the Direct Synthesis of Hydrogen Peroxide Alkaline
Solutions from H-2 and O-2. Chemsuschem, 2011. 4(4): p. 494-501.
51. Yamanaka, I., et al., Electrocatalysis of heat-treated Mn-porphyrin/carbon cathode for synthesis of
H2O2 acid solutions by H-2/O-2 fuel cell method. Chemistry Letters, 2006. 35(12): p. 1330-1331.
52. Yamanaka, I., et al., Direct and continuous production of hydrogen peroxide with 93% selectivity using
a fuel-cell system. Angewandte Chemie-International Edition, 2003. 42(31): p. 3653-3655.
53. Yamanaka, I., et al., Catalytic Synthesis of Neutral H2O2 Solutions from O-2 and H-2 by a Fuel Cell
Reaction. Chemsuschem, 2008. 1(12): p. 988-992.
54. Yamanaka, I., et al., Catalytic Synthesis of Neutral Hydrogen Peroxide at a CoN2Cx Cathode of a
Polymer Electrolyte Membrane Fuel Cell (PEMFC). Chemsuschem, 2010. 3(1): p. 59-62.
55. Sethuraman, V.A., et al., Importance of catalyst stability vis-à-vis hydrogen peroxide formation rates in
PEM fuel cell electrodes. Electrochimica Acta, 2009. 54(23): p. 5571-5582.
56. Katsounaros, I., et al., Oxygen Electrochemistry as a Cornerstone for Sustainable Energy Conversion.
Angewandte Chemie-International Edition, 2014. 53(1): p. 102-121.
57. Katsounaros, I., et al., Hydrogen peroxide electrochemistry on platinum: towards understanding the
oxygen reduction reaction mechanism. Physical Chemistry Chemical Physics, 2012. 14(20): p. 7384-
7391.
58. Sarapuu, A., et al., Electrochemical reduction of oxygen on anthraquinone-modified glassy carbon
electrodes in alkaline solution. Journal of Electroanalytical Chemistry, 2003. 541: p. 23-29.
59. Steele, B.C.H. and A. Heinzel, Materials for fuel-cell technologies. Nature, 2001. 414(6861): p. 345-352.
60. Duke, F.R. and T.W. Haas, Homogeneous Base-Catalyzed Decomposition of Hydrogen Peroxide. Journal
of Physical Chemistry, 1961. 65(2): p. 304-&.
61. Gnanamuthu, D.S. and J.V. Petrocelli, A Generalized Expression for Tafel Slope and Kinetics of
Oxygen Reduction on Noble Metals and Alloys. Journal of the Electrochemical Society, 1967. 114(10):
p. 1036-+.
62. Wroblowa, H.S., Y.C. Pan, and G. Razumney, Electroreduction of Oxygen - New Mechanistic Criterion.
Journal of Electroanalytical Chemistry, 1976. 69(2): p. 195-201.
63. Walch, S., et al., Mechanism of molecular oxygen reduction at the cathode of a PEM fuel cell: Non-
electrochemical reactions on catalytic Pt particles. Journal of Physical Chemistry C, 2008. 112(22): p.
8464-8475.
64. Jacob, T. and W.A. Goddard, Water formation on Pt and Pt-based alloys: A theoretical description of a
catalytic reaction. Chemphyschem, 2006. 7(5): p. 992-1005.
136 |
65. Norskov, J.K., et al., Origin of the overpotential for oxygen reduction at a fuel-cell cathode. Journal of
Physical Chemistry B, 2004. 108(46): p. 17886-17892.
66. Koh, S. and P. Strasser, Electrocatalysis on bimetallic surfaces: Modifying catalytic reactivity for
oxygen reduction by voltammetric surface dealloying. Journal of the American Chemical Society, 2007.
129(42): p. 12624-12625.
67. Stamenkovic, V.R., et al., Trends in electrocatalysis on extended and nanoscale Pt-bimetallic alloy
surfaces. Nature Materials, 2007. 6(3): p. 241-247.
68. Chen, L.Y., et al., Nanoporous PdNi Bimetallic Catalyst with Enhanced Electrocatalytic Performances
for Electro-oxidation and Oxygen Reduction Reactions. Advanced Functional Materials, 2011. 21(22): p.
4364-4370.
69. Shao, M., Palladium-based electrocatalysts for hydrogen oxidation and oxygen reduction reactions.
Journal of Power Sources, 2011. 196(5): p. 2433-2444.
70. Meng, H., D.R. Zeng, and F.Y. Xie, Recent Development of Pd-Based Electrocatalysts for Proton
Exchange Membrane Fuel Cells. Catalysts, 2015. 5(3): p. 1221-1274.
71. Erikson, H., et al., Enhanced electrocatalytic activity of cubic Pd nanoparticles towards the oxygen
reduction reaction in acid media. Electrochemistry Communications, 2011. 13(7): p. 734-737.
72. Wang, W.M., et al., Carbon-supported Pd-Co bimetallic nanoparticles as electrocatalysts for the oxygen
reduction reaction. Journal of Power Sources, 2007. 167(2): p. 243-249.
73. Antolini, E., Palladium in fuel cell catalysis. Energy & Environmental Science, 2009. 2(9): p. 915-931.
74. Vracar, L.M., D.B. Sepa, and A. Damjanovic, Palladium Electrode in Oxygen-Saturated Aqueous-
Solutions - Reduction of Oxygen in the Activation-Controlled Region. Journal of the Electrochemical
Society, 1986. 133(9): p. 1835-1839.
75. Savadogo, O., et al., New palladium alloys catalyst for the oxygen reduction reaction in an acid
medium. Electrochemistry Communications, 2004. 6(2): p. 105-109.
76. Shao, M.H., et al., Palladium monolayer and palladium alloy electrocatalysts for oxygen reduction.
Langmuir, 2006. 22(25): p. 10409-10415.
77. Zhao, J., A. Sarkar, and A. Manthiram, Synthesis and characterization of Pd-Ni nanoalloy
electrocatalysts for oxygen reduction reaction in fuel cells. Electrochimica Acta, 2010. 55(5): p. 1756-
1765.
78. Kariuki, N.N., et al., Colloidal Synthesis and Characterization of Carbon-Supported Pd-Cu
Nanoparticle Oxygen Reduction Electrocatalysts. Chemistry of Materials, 2010. 22(14): p. 4144-4152.
79. Shao, M.H., K. Sasaki, and R.R. Adzic, Pd-Fe nanoparticles as electrocatalysts for oxygen reduction.
Journal of the American Chemical Society, 2006. 128(11): p. 3526-3527.
80. Raghuveer, V., P.J. Ferreira, and A. Manthiram, Comparison of Pd-Co-Au electrocatalysts prepared by
conventional borohydride and microemulsion methods for oxygen reduction in fuel cells.
Electrochemistry Communications, 2006. 8(5): p. 807-814.
81. Antolini, E., et al., Palladium-based electrodes: A way to reduce platinum content in polymer electrolyte
membrane fuel cells. Electrochimica Acta, 2011. 56(5): p. 2299-2305.
82. Markovic, N.M., H.A. Gasteiger, and P.N. Ross, Oxygen Reduction on Platinum Low-Index Single-
Crystal Surfaces in Sulfuric-Acid-Solution - Rotating Ring-Pt(Hkl) Disk Studies. Journal of Physical
Chemistry, 1995. 99(11): p. 3411-3415.
83. Gasteiger, H.A. and N.M. Markovic, Just a Dream-or Future Reality? Science, 2009. 324(5923): p. 48-
49.
84. Damjanovic, A., M.A. Genshaw, and J.O. Bockris, Role of Hydrogen Peroxide in Oxygen Reduction at
Platinum in H2so4 Solution. Journal of the Electrochemical Society, 1967. 114(5): p. 466-+.
85. Pizzutilo, E., et al., Palladium electrodissolution from model surfaces and nanoparticles.
Electrochimica Acta, 2017. 229: p. 467-477.
References
| 137
86. Markovic, N.M. and P.N. Ross, Surface science studies of model fuel cell electrocatalysts. Surface
Science Reports, 2002. 45(4-6): p. 121-229.
87. Markovic, N.M., et al., Oxygen reduction reaction on Pt(111): effects of bromide. Journal of
Electroanalytical Chemistry, 1999. 467(1-2): p. 157-163.
88. Rahul, R., et al., The role of surface oxygenated-species and adsorbed hydrogen in the oxygen reduction
reaction (ORR) mechanism and product selectivity on Pd-based catalysts in acid media. Physical
Chemistry Chemical Physics, 2015. 17(23): p. 15146-15155.
89. Choi, C.H., et al., Hydrogen Peroxide Synthesis via Enhanced Two-Electron Oxygen Reduction Pathway
on Carbon-Coated Pt Surface. Journal of Physical Chemistry C, 2014. 118(51): p. 30063-30070.
90. Choi, C.H., et al., Tuning selectivity of electrochemical reactions by atomically dispersed platinum
catalyst. Nature Communications, 2016. 7.
91. Rodriguez, P. and M.T.M. Koper, Electrocatalysis on gold. Physical Chemistry Chemical Physics, 2014.
16(27): p. 13583-13594.
92. Blizanac, B.B., et al., Anion adsorption, CO oxidation, and oxygen reduction reaction on a Au(100)
surface: The pH effect. Journal of Physical Chemistry B, 2004. 108(2): p. 625-634.
93. Strbac, S., N.A. Anastasijevic, and R.R. Adzic, Oxygen Reduction on Au(100) and Vicinal Au(910) and
Au(11, 1, 1) Faces in Alkaline-Solution - a Rotating-Disk Ring Study. Journal of Electroanalytical
Chemistry, 1992. 323(1-2): p. 179-195.
94. Markovic, N.M., R.R. Adzic, and V.B. Vesovic, Structural Effects in Electrocatalysis - Oxygen Reduction
on the Gold Single-Crystal Electrodes with (110) and (111) Orientations. Journal of Electroanalytical
Chemistry, 1984. 165(1-2): p. 121-133.
95. Jirkovsky, J.S., M. Halasa, and D.J. Schiffrin, Kinetics of electrocatalytic reduction of oxygen and
hydrogen peroxide on dispersed gold nanoparticles. Physical Chemistry Chemical Physics, 2010.
12(28): p. 8042-8052.
96. Sanchez-Sanchez, C.M. and A.J. Bard, Hydrogen Peroxide Production in the Oxygen Reduction
Reaction at Different Electrocatalysts as Quantified by Scanning Electrochemical Microscopy.
Analytical Chemistry, 2009. 81(19): p. 8094-8100.
97. Siahrostami, S., et al., Enabling direct H2O2 production through rational electrocatalyst design.
Nature Materials, 2013. 12(12): p. 1137-1143.
98. Kitchin, J.R., et al., Role of strain and ligand effects in the modification of the electronic and chemical
properties of bimetallic surfaces. Physical Review Letters, 2004. 93(15).
99. Rodriguez, J.A. and D.W. Goodman, The Nature of the Metal Metal Bond in Bimetallic Surfaces.
Science, 1992. 257(5072): p. 897-903.
100. Stamenkovic, V., et al., Changing the activity of electrocatalysts for oxygen reduction by tuning the
surface electronic structure. Angewandte Chemie-International Edition, 2006. 45(18): p. 2897-2901.
101. Mavrikakis, M., B. Hammer, and J.K. Norskov, Effect of strain on the reactivity of metal surfaces.
Physical Review Letters, 1998. 81(13): p. 2819-2822.
102. Kibler, L.A., et al., Tuning reaction rates by lateral strain in a palladium monolayer. Angewandte
Chemie-International Edition, 2005. 44(14): p. 2080-2084.
103. Chen, M.S., et al., The promotional effect of gold in catalysis by palladium-gold. Science, 2005.
310(5746): p. 291-293.
104. Maroun, F., et al., The role of atomic ensembles in the reactivity of bimetallic electrocatalysts. Science,
2001. 293(5536): p. 1811-1814.
105. Jirkovsky, J.S., et al., Single Atom Hot-Spots at Au-Pd Nanoalloys for Electrocatalytic H2O2
Production. Journal of the American Chemical Society, 2011. 133(48): p. 19432-19441.
106. Stamenkovic, V.R., et al., Effect of surface composition on electronic structure, stability, and
electrocatalytic properties of Pt-transition metal alloys: Pt-skin versus Pt-skeleton surfaces. Journal of
the American Chemical Society, 2006. 128(27): p. 8813-8819.
138 |
107. Hong, J.W., et al., Atomic-Distribution-Dependent Electrocatalytic Activity of Au-Pd Bimetallic
Nanocrystals. Angewandte Chemie-International Edition, 2011. 50(38): p. 8876-8880.
108. Cui, C.H., et al., Remarkable Enhancement of Electrocatalytic Activity by Tuning the Interface of Pd-Au
Bimetallic Nanoparticle Tubes. Acs Nano, 2011. 5(5): p. 4211-4218.
109. Liu, P. and J.K. Norskov, Ligand and ensemble effects in adsorption on alloy surfaces. Physical
Chemistry Chemical Physics, 2001. 3(17): p. 3814-3818.
110. Brodsky, C.N., et al., Electrochemically Induced Surface Metal Migration in Well-Defined Core-Shell
Nanoparticles and Its General Influence on Electrocatalytic Reactions. Acs Nano, 2014. 8(9): p. 9368-
9378.
111. Nie, M., et al., Highly efficient AuPd-WC/C electrocatalyst for ethanol oxidation. Electrochemistry
Communications, 2007. 9(9): p. 2375-2379.
112. Lee, Y.W., et al., Synthesis and Electrocatalytic Activity of Au-Pd Alloy Nanodendrites for Ethanol
Oxidation. Journal of Physical Chemistry C, 2010. 114(17): p. 7689-7693.
113. Suzuki, S., et al., Effect of Surface Composition of Platinum-ruthenium Nanoparticles on Methanol
Oxidation Activity. Polymer Electrolyte Fuel Cells 10, Pts 1 and 2, 2010. 33(1): p. 321-332.
114. Lang, X.Y., et al., Novel Nanoporous Au-Pd Alloy with High Catalytic Activity and Excellent
Electrochemical Stability. Journal of Physical Chemistry C, 2010. 114(6): p. 2600-2603.
115. Ghosh, T., et al., Pt Alloy and Intermetallic Phases with V, Cr, Mn, Ni, and Cu: Synthesis As
Nanomaterials and Possible Applications As Fuel Cell Catalysts. Chemistry of Materials, 2010. 22(7):
p. 2190-2202.
116. Leonard, B.M., et al., Facile Synthesis of PtNi Intermetallic Nanoparticles: Influence of Reducing Agent
and Precursors on Electrocatalytic Activity. Chemistry of Materials, 2011. 23(5): p. 1136-1146.
117. Bliznakov, S.T., et al., Pt Monolayer on Electrodeposited Pd Nanostructures: Advanced Cathode
Catalysts for PEM Fuel Cells. Journal of the Electrochemical Society, 2012. 159(9): p. F501-F506.
118. Fernandez, J.L., et al., Pd-Ti and Pd-Co-Au electrocatalysts as a replacement for platinum for oxygen
reduction in proton exchange membrane fuel cells. Journal of the American Chemical Society, 2005.
127(38): p. 13100-13101.
119. Kuttiyiel, K.A., et al., Gold-promoted structurally ordered intermetallic palladium cobalt nanoparticles
for the oxygen reduction reaction. Nature Communications, 2014. 5.
120. Xing, Y.C., et al., Enhancing Oxygen Reduction Reaction Activity via Pd-Au Alloy Sublayer Mediation
of Pt Monolayer Electrocatalysts. Journal of Physical Chemistry Letters, 2010. 1(21): p. 3238-3242.
121. Freakley, S.J., et al., Gold Catalysis: A Reflection on Where We are Now. Catalysis Letters, 2015.
145(1): p. 71-79.
122. Jirkovsky, J.S., et al., Potential-Dependent Structural Memory Effects in Au-Pd Nanoalloys. Journal of
Physical Chemistry Letters, 2012. 3(3): p. 315-321.
123. Pritchard, J., et al., Direct Synthesis of Hydrogen Peroxide and Benzyl Alcohol Oxidation Using Au-Pd
Catalysts Prepared by Sol Immobilization. Langmuir, 2010. 26(21): p. 16568-16577.
124. Yang, F., et al., Au-Pd nanoparticles supported on carbon fiber cloth as the electrocatalyst for H2O2
electroreduction in acid medium. Journal of Power Sources, 2013. 233: p. 252-258.
125. Verdaguer-Casadevall, A., et al., Trends in the Electrochemical Synthesis of H2O2: Enhancing Activity
and Selectivity by Electrocatalytic Site Engineering. Nano Letters, 2014. 14(3): p. 1603-1608.
126. Erikson, H., et al., Oxygen Electroreduction on Electrodeposited PdAu Nanoalloys. Electrocatalysis,
2015. 6(1): p. 77-85.
127. Pizzutilo, E., et al., Electrocatalytic synthesis of hydrogen peroxide on Au-Pd nanoparticles: from
fundamentals to continuous production. Chemical Physics Letters, 2017. In Press.
128. Ham, H.C., et al., Geometric Parameter Effects on Ensemble Contributions to Catalysis: H2O2
Formation from H-2 and O-2 on AuPd Alloys. A First Principles Study. Journal of Physical Chemistry
C, 2010. 114(35): p. 14922-14928.
References
| 139
129. Ham, H.C., et al., Importance of Pd monomer pairs in enhancing the oxygen reduction reaction activity
of the AuPd(100) surface: A first principles study. Catalysis Today, 2016. 263: p. 11-15.
130. Todorovic, R. and R.J. Meyer, A comparative density functional theory study of the direct synthesis of
H2O2 on Pd, Pt and Au surfaces. Catalysis Today, 2011. 160(1): p. 242-248.
131. Siahrostami, S., et al., Activity and Selectivity for O-2 Reduction to H2O2 on Transition Metal
Surfaces. Electrochemical Synthesis of Fuels 2, 2013. 58(2): p. 53-62.
132. Topalov, A.A., et al., Towards a comprehensive understanding of platinum dissolution in acidic media.
Chemical Science, 2014. 5(2): p. 631-638.
133. Cherevko, S., et al., Dissolution of Platinum in the Operational Range of Fuel Cells. Chemelectrochem,
2015. 2(10): p. 1471-1478.
134. Topalov, A.A., et al., Dissolution of Platinum: Limits for the Deployment of Electrochemical Energy
Conversion? Angewandte Chemie-International Edition, 2012. 51(50): p. 12613-12615.
135. Baldizzone, C., et al., Stability of Dealloyed Porous Pt/Ni Nanoparticles. Acs Catalysis, 2015. 5(9): p.
5000-5007.
136. Cherevko, S., N. Kulyk, and K.J.J. Mayrhofer, Durability of Platinum-Based Fuel Cell Electrocatalysts:
Dissolution of Bulk and Nanoscale Platinum. Nano Energy.
137. Cherevko, S., et al., Temperature-Dependent Dissolution of Polycrystalline Platinum in Sulfuric Acid
Electrolyte. Electrocatalysis, 2014. 5(3): p. 235-240.
138. Galeano, C., et al., Toward Highly Stable Electrocatalysts via Nanoparticle Pore Confinement. Journal
of the American Chemical Society, 2012. 134(50): p. 20457-20465.
139. Keeley, G.P., S. Cherevko, and K.J. Mayrhofer, The Stability Challenge on the Pathway to Low and
Ultra‐Low Platinum Loading for Oxygen Reduction in Fuel Cells. ChemElectroChem, 2015.
140. Mayrhofer, K.J.J., et al., Fuel cell catalyst degradation on the nanoscale. Electrochemistry
Communications, 2008. 10(8): p. 1144-1147.
141. Pizzutilo, E., et al., On the Need of Improved Accelerated Degradation Protocols (ADPs): Examination
of Platinum Dissolution and Carbon Corrosion in Half-Cell Tests. Journal of The Electrochemical
Society, 2016. 163(14): p. F1510-F1514.
142. Cherevko, S., et al., Oxygen evolution activity and stability of iridium in acidic media. Part 1. - Metallic
iridium. Journal of Electroanalytical Chemistry, 2016. 773: p. 69-78.
143. Cherevko, S., et al., Stability of nanostructured iridium oxide electrocatalysts during oxygen evolution
reaction in acidic environment. Electrochemistry Communications, 2014. 48: p. 81-85.
144. Cherevko, S., et al., Dissolution of Noble Metals during Oxygen Evolution in Acidic Media.
Chemcatchem, 2014. 6(8): p. 2219-2223.
145. Geiger, S., et al., Activity and Stability of Electrochemically and Thermally Treated Iridium for the
Oxygen Evolution Reaction. Journal of the Electrochemical Society, 2016. 163(11): p. F3132-F3138.
146. Cherevko, S., et al., Oxygen and hydrogen evolution reactions on Ru, RuO2, Ir, and IrO2 thin film
electrodes in acidic and alkaline electrolytes: A comparative study on activity and stability. Catalysis
Today, 2016. 262: p. 170-180.
147. Cherevko, S., et al., Electrochemical dissolution of gold in acidic medium. Electrochemistry
Communications, 2013. 28: p. 44-46.
148. Cherevko, S., et al., Gold dissolution: towards understanding of noble metal corrosion. Rsc Advances,
2013. 3(37): p. 16516-16527.
149. Shrestha, B.R., A. Nishikata, and T. Tsuru, Channel flow double electrode study on palladium
dissolution during potential cycling in sulfuric acid solution. Electrochimica Acta, 2012. 70: p. 42-49.
150. Grdeń, M., et al., Electrochemical behaviour of palladium electrode: Oxidation, electrodissolution and
ionic adsorption. Electrochimica Acta, 2008. 53(26): p. 7583-7598.
140 |
151. Pourbaix, M., Atlas of Electrochemical Equilibria in Aqueous Solutions, 2nd ed., Nat. Assoc. of
Corrosion Engineers, Houston, Texas, 1974, Chapter IV, 17.1. 1974.
152. Burke, L.D. and J.K. Casey, An Examination of the Electrochemical-Behavior of Palladium Electrodes
in Acid. Journal of the Electrochemical Society, 1993. 140(5): p. 1284-1291.
153. Rand, D.A.J. and R. Woods, Study of Dissolution of Platinum, Palladium, Rhodium and Gold
Electrodes in 1 M Sulfuric-Acid by Cyclic Voltammetry. Journal of Electroanalytical Chemistry, 1972.
35(Nmar): p. 209-&.
154. Bolzan, A.E., M.E. Martins, and A.J. Arvia, The Electrodissolution of Base Palladium in Relation to
the Oxygen Electroadsorption and Electrodesorption in Sulfuric-Acid-Solution. Journal of
Electroanalytical Chemistry, 1984. 172(1-2): p. 221-233.
155. Juodkazis, K., et al., Anodic dissolution of palladium in sulfuric acid: An electrochemical quartz crystal
microbalance study. Russian Journal of Electrochemistry, 2003. 39(9): p. 954-959.
156. Vracar, L.M., D.B. Sepa, and A. Damjanovic, Palladium Electrode in Oxygen Saturated Solutions -
Rest Potentials in Solutions of Different Ph. Journal of the Electrochemical Society, 1987. 134(7): p.
1695-1697.
157. Bolzan, A.E. and A.J. Arvia, The Electrochemical-Behavior of Hydrous Palladium Oxide Layers
Formed at High Positive Potentials in Different Electrolyte-Solutions. Journal of Electroanalytical
Chemistry, 1992. 322(1-2): p. 247-265.
158. Bolzan, A.E. and A.J. Arvia, Effect of the Electrolyte-Composition on the Electroreduction of Palladium
Oxide-Films. Journal of Electroanalytical Chemistry, 1993. 354(1-2): p. 243-253.
159. Perdriel, C.L., E. Custidiano, and A.J. Arvia, Modifications of Palladium Electrode Surfaces Produced
by Periodic Potential Treatments. Journal of Electroanalytical Chemistry, 1988. 246(1): p. 165-180.
160. Solomun, T., Initial-Stages of Electrooxidation of Pd (100) Surfaces in Sulfuric-Acid-Solution - an Xps
Study. Journal of Electroanalytical Chemistry, 1987. 217(2): p. 435-441.
161. Solomun, T., Electro-Oxidation of the Pd(100) Surface - Potential Dependence of Oxygen Incorporation
into the Substrate. Journal of Electroanalytical Chemistry, 1988. 255(1-2): p. 163-177.
162. Solomun, T., The Role of the Electrolyte Anion in Anodic-Dissolution of the Pd(100) Surface. Journal of
Electroanalytical Chemistry, 1991. 302(1-2): p. 31-46.
163. Cadle, S.H., Ring-Disk Electrode Study of Palladium Dissolution. Journal of the Electrochemical
Society, 1974. 121(5): p. 645-648.
164. Harrison, J.A. and T.A. Whitfield, The Dissolution of Palladium in Various Electrolytes.
Electrochimica Acta, 1983. 28(9): p. 1229-1236.
165. Llopis, J.F., L. Victori, and J.M. Gamboa, Radiochemical Study of Anodic Behavior of Palladium.
Electrochimica Acta, 1972. 17(12): p. 2225-&.
166. Ibl, N., G. Gut, and M. Weber, Electrodeposition and Catalytic Activity of Palladium Powders.
Electrochimica Acta, 1973. 18(4): p. 307-314.
167. Burke, L.D. and M.B.C. Roche, An Electrochemical Investigation of Monolayer and Multilayer Oxide-
Films on Palladium in Aqueous-Media. Journal of Electroanalytical Chemistry, 1985. 186(1-2): p. 139-
154.
168. Wagner, C., Kinetik Und Mechanismus Von Umsetzungen Zwischen Flussigen Legierungen Und
Schlacken. Zeitschrift Fur Elektrochemie, 1958. 62(3): p. 386-389.
169. Schumacher, R., et al., The application of fast potential steps to noble metal electrodes: a correlation of
electrodissolution with changes in the surface morphology and composition. Journal of
Electroanalytical Chemistry, 1993. 354(1–2): p. 59-70.
170. Grdeń, M., J. Kotowski, and A. Czerwiński, Study of electrochemical palladium behavior by the quartz
crystal microbalance. I. Acidic Solutions. Journal of Solid State Electrochemistry, 1999. 3(6): p. 348-
351.
References
| 141
171. Łukaszewski, M. and A. Czerwiński, Dissolution of noble metals and their alloys studied by
electrochemical quartz crystal microbalance. Journal of Electroanalytical Chemistry, 2006. 589(1): p.
38-45.
172. Kolotyrkin, Y.M., V.V. Losev, and A.N. Chemodanov, Relationship between Corrosion Processes and
Oxygen Evolution on Anodes Made from Noble-Metals and Related Metal-Oxide Anodes. Materials
Chemistry and Physics, 1988. 19(1-2): p. 1-95.
173. Maksimov, Y.M., A.V. Smolin, and B.I. Podlovchenko, On the ratio of processes of adsorbed oxygen
layer formation and palladium surface layer dissolution at linear anodic potential sweep. Russian
Journal of Electrochemistry, 2007. 43(12): p. 1412-1417.
174. Kumar, A. and D.A. Buttry, Size-Dependent Anodic Dissolution of Water-Soluble Palladium
Nanoparticles. Journal of Physical Chemistry C, 2013. 117(50): p. 26783-26789.
175. Cherevko, S., et al., A Comparative Study on Gold and Platinum Dissolution in Acidic and Alkaline
Media. Journal of the Electrochemical Society, 2014. 161(12): p. H822-H830.
176. Cherevko, S., et al., Effect of Temperature on Gold Dissolution in Acidic Media. Journal of the
Electrochemical Society, 2014. 161(9): p. H501-H507.
177. Arends, I.W.C.E. and R.A. Sheldon, Activities and stabilities of heterogeneous catalysts in selective
liquid phase oxidations: recent developments. Applied Catalysis a-General, 2001. 212(1-2): p. 175-187.
178. Borup, R., et al., Scientific aspects of polymer electrolyte fuel cell durability and degradation. Chemical
Reviews, 2007. 107(10): p. 3904-3951.
179. Tao, F., et al., Evolution of Structure and Chemistry of Bimetallic Nanoparticle Catalysts under
Reaction Conditions. Journal of the American Chemical Society, 2010. 132(25): p. 8697-8703.
180. Kocha, S.S., Electrochemical Degradation: Electrocatalyst and Support Durability. Polymer Electrolyte
Fuel Cell Degradation, 2012: p. 89-214.
181. Mezzavilla, S., et al., Structure–Activity–Stability Relationships for Space-Confined PtxNiy
Nanoparticles in the Oxygen Reduction Reaction. ACS Catalysis, 2016. 6(12): p. 8058-8068.
182. Rudi, S., et al., Electrochemical Dealloying of Bimetallic ORR Nanoparticle Catalysts at Constant
Electrode Potentials. Journal of the Electrochemical Society, 2015. 162(4): p. F403-F409.
183. DoE, U., Hydrogen, Fuel Cells and Infrastructure Technologies Programm: Multiyear Research,
Development and Demonstration Plan
2011.
184. Yamanaka, I., et al., Study of Direct Synthesis of Hydrogen Peroxide Acid Solutions at a Heat-Treated
MnCl-Porphyrin/Activated Carbon Cathode from H-2 and O-2. Journal of Physical Chemistry C, 2012.
116(7): p. 4572-4583.
185. Choudhary, V.R., A.G. Gaikwad, and S.D. Sansare, Activation of supported Pd metal catalysts for
selective oxidation of hydrogen to hydrogen peroxide. Catalysis Letters, 2002. 83(3-4): p. 235-239.
186. Abate, S., et al., Performances of Pd-Me (Me=Ag, Pt) catalysts in the direct synthesis of H2O2 on
catalytic membranes. Catalysis Today, 2006. 117(1): p. 193-198.
187. Choudhary, V.R., A.G. Gaikwad, and S.D. Sansare, Nonhazardous Direct Oxidation of Hydrogen to
Hydrogen Peroxide Using a Novel Membrane Catalyst. Angewandte Chemie, 2001. 113(9): p. 1826-
1829.
188. Henkel, H. and W. Weber, Manufacture of hydrogen peroxid. 1914, Google Patents.
189. Enache, D.I., et al., Solvent-free oxidation of primary alcohols to aldehydes using Au-Pd/TiO2
catalysts. Science, 2006. 311(5759): p. 362-365.
190. Chinta, S. and J.H. Lunsford, A mechanistic study of H2O2 and H2O formation from H2 and O2
catalyzed by palladium in an aqueous medium. Journal of Catalysis, 2004. 225(1): p. 249-255.
191. Wilson, N.M. and D.W. Flaherty, Mechanism for the Direct Synthesis of H2O2 on Pd Clusters:
Heterolytic Reaction Pathways at the Liquid-Solid Interface. Journal of the American Chemical
Society, 2016. 138(2): p. 574-586.
142 |
192. Knözinger, H. and K. Kochloefl, Heterogeneous Catalysis and Solid Catalysts, in Ullmann's
Encyclopedia of Industrial Chemistry. 2000, Wiley-VCH Verlag GmbH & Co. KGaA.
193. Ford, D.C., et al., Partial and complete reduction of O2 by hydrogen on transition metal surfaces.
Surface Science, 2010. 604(19–20): p. 1565-1575.
194. Horiuti, I. and M. Polanyi, Exchange reactions of hydrogen on metallic catalysts. Transactions of the
Faraday Society, 1934. 30: p. 1164-1172.
195. Yang, B., et al., Evidence To Challenge the Universality of the Horiuti-Polanyi Mechanism for
Hydrogenation in Heterogeneous Catalysis: Origin and Trend of the Preference of a Non-Horiuti-
Polanyi Mechanism. Journal of the American Chemical Society, 2013. 135(40): p. 15244-15250.
196. Zhou, B. and L.K. Lee, Catalyst and process for direct catalystic production of hydrogen peroxide,
(H2O2). 2001, Google Patents.
197. Pospelova, T.A. and N.I. Kobozev, Catalytic Synthesis of Hydrogen Peroxide from the Elements on
Palladium .3. The Active Centers of Hydrogen Peroxide Decomposition on Palladium. Zhurnal
Fizicheskoi Khimii, 1961. 35(6): p. 1192-1197.
198. Pospelova, T.A. and N.I. Kobozev, Palladium Catalyzed Synthesis of Hydrogen Peroxide from the
Elements .2. The Active Centers of Palladium in the Synthesis of H2o2. Zhurnal Fizicheskoi Khimii,
1961. 35(3): p. 535-542.
199. Pospelova, T.A., N.I. Kobozev, and E.N. Eremin, Palladium Catalyzed Synthesis of Hydrogen Peroxide
from the Elements .1. Conditions for the Formation of Hydrogen Peroxide. Zhurnal Fizicheskoi Khimii,
1961. 35(2): p. 298-305.
200. Choudhary, V.R., C. Samanta, and T.V. Choudhary, Factors influencing decomposition of H2O2 over
supported Pd catalyst in aqueous medium. Journal of Molecular Catalysis A: Chemical, 2006. 260(1): p.
115-120.
201. Choudhary, V.R., C. Samanta, and T.V. Choudhary, Direct oxidation of H2 to H2O2 over Pd-based
catalysts: Influence of oxidation state, support and metal additives. Applied Catalysis A: General, 2006.
308: p. 128-133.
202. Choudhary, V.R., C. Samanta, and P. Jana, Decomposition and/or hydrogenation of hydrogen peroxide
over Pd/Al2O3 catalyst in aqueous medium: Factors affecting the rate of H2O2 destruction in presence
of hydrogen. Applied Catalysis A: General, 2007. 332(1): p. 70-78.
203. Samanta, C. and V.R. Choudhary, Direct oxidation of H2 to H2O2 over Pd/CeO2 catalyst under
ambient conditions: Influence of halide ions. Chemical Engineering Journal, 2008. 136(2): p. 126-132.
204. Burch, R. and P.R. Ellis, An investigation of alternative catalytic approaches for the direct synthesis of
hydrogen peroxide from hydrogen and oxygen. Applied Catalysis B: Environmental, 2003. 42(2): p. 203-
211.
205. Dissanayake, D.P. and J.H. Lunsford, The direct formation of H2O2 from H2 and O2 over colloidal
palladium. Journal of Catalysis, 2003. 214(1): p. 113-120.
206. Park, S., et al., Direct synthesis of hydrogen peroxide from hydrogen and oxygen over palladium
catalyst supported on SO3H-functionalized mesoporous silica. Journal of Molecular Catalysis A:
Chemical, 2010. 319(1): p. 98-107.
207. Landon, P., et al., Direct synthesis of hydrogen peroxide from H2 and O2 using Pd and Au catalysts.
Physical Chemistry Chemical Physics, 2003. 5(9): p. 1917-1923.
208. Landon, P., et al., Direct formation of hydrogen peroxide from H2/O2 using a gold catalyst. Chemical
Communications, 2002(18): p. 2058-2059.
209. Mitsutaka, O., et al., Direct Production of Hydrogen Peroxide from H2 and O2 over Highly Dispersed
Au catalysts. Chemistry Letters, 2003. 32(9): p. 822-823.
210. Ishihara, T., et al., Synthesis of hydrogen peroxide by direct oxidation of H2 with O2 on Au/SiO2
catalyst. Applied Catalysis A: General, 2005. 291(1): p. 215-221.
211. Hutchings, G.J., Nanocrystalline gold and gold-palladium alloy oxidation catalysts: a personal
reflection on the nature of the active sites. Dalton Transactions, 2008(41): p. 5523-5536.
References
| 143
212. Freakley, S.J., et al., Effect of Reaction Conditions on the Direct Synthesis of Hydrogen Peroxide with a
AuPd/TiO2 Catalyst in a Flow Reactor. ACS Catalysis, 2013. 3(4): p. 487-501.
213. Edwards, J.K., et al., Au-Pd supported nanocrystals as catalysts for the direct synthesis of hydrogen
peroxide from H2 and O2. Green Chemistry, 2008. 10(4): p. 388-394.
214. Solsona, B.E., et al., Direct Synthesis of Hydrogen Peroxide from H2 and O2 Using Al2O3 Supported
Au−Pd Catalysts. Chemistry of Materials, 2006. 18(11): p. 2689-2695.
215. Edwards, J.K., et al., Direct synthesis of hydrogen peroxide from H2 and O2 using TiO2-supported Au–
Pd catalysts. Journal of Catalysis, 2005. 236(1): p. 69-79.
216. Edwards, J.K., et al., Direct synthesis of hydrogen peroxide from H2 and O2 using Au-Pd/Fe2O3
catalysts. Journal of Materials Chemistry, 2005. 15(43): p. 4595-4600.
217. Edwards, J.K., et al., Comparison of supports for the direct synthesis of hydrogen peroxide from H2 and
O2 using Au–Pd catalysts. Catalysis Today, 2007. 122(3): p. 397-402.
218. Ntainjua N, E., et al., The role of the support in achieving high selectivity in the direct formation of
hydrogen peroxide. Green Chemistry, 2008. 10(11): p. 1162-1169.
219. Edwards, J.K., et al., Au-Pd supported nanocrystals as catalysts for the direct synthesis of hydrogen
peroxide from H-2 and O-2. Green Chemistry, 2008. 10(4): p. 388-394.
220. Ham, H.C., et al., On the Role of Pd Ensembles in Selective H2O2 Formation on PdAu Alloys. Journal
of Physical Chemistry C, 2009. 113(30): p. 12943-12945.
221. Ham, H.C., et al., Pd ensemble effects on oxygen hydrogenation in AuPd alloys: A combined density
functional theory and Monte Carlo study. Catalysis Today, 2011. 165(1): p. 138-144.
222. Li, J. and K. Yoshizawa, Mechanistic aspects in the direct synthesis of hydrogen peroxide on PdAu
catalyst from first principles. Catalysis Today, 2015. 248: p. 142-148.
223. Edwards, J.K., et al., Direct synthesis of hydrogen peroxide from H-2 and O-2 using supported Au-Pd
catalysts. Faraday Discussions, 2008. 138: p. 225-239.
224. Edwards, J.K., et al., Strategies for Designing Supported Gold-Palladium Bimetallic Catalysts for the
Direct Synthesis of Hydrogen Peroxide. Accounts of Chemical Research, 2014. 47(3): p. 845-854.
225. Edwards, J.K., et al., Direct synthesis of hydrogen peroxide from H-2 andO(2) using Au-Pd/Fe2O3
catalysts. Journal of Materials Chemistry, 2005. 15(43): p. 4595-4600.
226. Staszak-Jirkovsky, J., et al., The bifurcation point of the oxygen reduction reaction on Au-Pd
nanoalloys. Faraday Discussions, 2016. 188(0): p. 257-278.
227. Freakley, S.J., et al., Palladium-tin catalysts for the direct synthesis of
H<sub>2</sub>O<sub>2</sub> with high selectivity. Science, 2016. 351(6276): p. 965-968.
228. Mayrhofer, K.J.J., et al., Non-destructive transmission electron microscopy study of catalyst
degradation under electrochemical treatment. Journal of Power Sources, 2008. 185(2): p. 734-739.
229. Ferreira, P.J., et al., Instability of Pt/C electrocatalysts in proton exchange membrane fuel cells - A
mechanistic investigation. Journal of the Electrochemical Society, 2005. 152(11): p. A2256-A2271.
230. Yasuda, K., et al., Platinum dissolution and deposition in the polymer electrolyte membrane of a PEM
fuel cell as studied by potential cycling. Physical Chemistry Chemical Physics, 2006. 8(6): p. 746-752.
231. Akita, T., et al., Analytical TEM study of Pt particle deposition in the proton-exchange membrane of a
membrane-electrode-assembly. Journal of Power Sources, 2006. 159(1): p. 461-467.
232. Baldizzone, C., et al., Confined-Space Alloying of Nanoparticles for the Synthesis of Efficient PtNi Fuel-
Cell Catalysts. Angewandte Chemie-International Edition, 2014. 53(51): p. 14250-14254.
233. Meier, J.C., et al., Degradation Mechanisms of Pt/C Fuel Cell Catalysts under Simulated Start-Stop
Conditions. Acs Catalysis, 2012. 2(5): p. 832-843.
234. Meier, J.C., et al., Design criteria for stable Pt/C fuel cell catalysts. Beilstein Journal of
Nanotechnology, 2014. 5: p. 44-67.
144 |
235. Meier, J.C., et al., Stability investigations of electrocatalysts on the nanoscale. Energy &
Environmental Science, 2012. 5(11): p. 9319-9330.
236. Mezzavilla, S., et al., Experimental methodologies to understand the degradation of nanostructured
electrocatalysts for PEM fuel cells: advances and opportunities. ChemElectroChem, 2016. 3(10): p.
1524-1536.
237. Yaguchi, T., et al., Development of a technique for in situ high temperature TEM observation of
catalysts in a highly moisturized air atmosphere. Journal of Electron Microscopy, 2012. 61(4): p. 199-
206.
238. Simonsen, S.B., et al., Direct Observations of Oxygen-induced Platinum Nanoparticle Ripening Studied
by In Situ TEM. Journal of the American Chemical Society, 2010. 132(23): p. 7968-7975.
239. Hodnik, N., G. Dehm, and K.J.J. Mayrhofer, Importance and Challenges of Electrochemical in Situ
Liquid Cell Electron Microscopy for Energy Conversion Research. Accounts of Chemical Research,
2016. 49(9): p. 2015-2022.
240. Ohma, A., et al., Membrane and Catalyst Performance Targets for Automotive Fuel Cells by FCCJ
Membrane, Catalyst, MEA WG. Polymer Electrolyte Fuel Cells 11, 2011. 41(1): p. 775-784.
241. Lukaszewski, M. and A. Czerwinski, Electrochemical behavior of palladium-gold alloys.
Electrochimica Acta, 2003. 48(17): p. 2435-2445.
242. Paulus, U.A., et al., Oxygen reduction on a high-surface area Pt/Vulcan carbon catalyst: a thin-film
rotating ring-disk electrode study. Journal of Electroanalytical Chemistry, 2001. 495(2): p. 134-145.
243. Schmidt, T.J., et al., Characterization of high-surface area electrocatalysts using a rotating disk
electrode configuration. Journal of the Electrochemical Society, 1998. 145(7): p. 2354-2358.
244. Klemm, S.O., et al., Time and potential resolved dissolution analysis of rhodium using a
microelectrochemical flow cell coupled to an ICP-MS (vol 677, pg 50, 2012). Journal of
Electroanalytical Chemistry, 2013. 693: p. 127-127.
245. Klemm, S.O., et al., Coupling of a high throughput microelectrochemical cell with online
multielemental trace analysis by ICP-MS. Electrochemistry Communications, 2011. 13(12): p. 1533-
1535.
246. Polymeros, G., PhD Thesis. 2017.
247. Mayrhofer, K.J.J., et al., An Electrochemical Cell Configuration Incorporating an Ion Conducting
Membrane Separator between Reference and Working Electrode. International Journal of
Electrochemical Science, 2009. 4(1): p. 1-8.
248. Topalov, A.A., et al., Development and integration of a LabVIEW-based modular architecture for
automated execution of electrochemical catalyst testing. Review of Scientific Instruments, 2011. 82(11):
p. 114103.
249. Bard, A.J. and L.R. Faulkner, Electrochemical Methods. Wiley, New York, 2001. 2nd edition.
250. van der Vliet, D., et al., On the importance of correcting for the uncompensated Ohmic resistance in
model experiments of the Oxygen Reduction Reaction. Journal of Electroanalytical Chemistry, 2010.
647(1): p. 29-34.
251. Katsounaros, I. and K.J.J. Mayrhofer, The influence of non-covalent interactions on the hydrogen
peroxide electrochemistry on platinum in alkaline electrolytes. Chemical Communications, 2012. 48(53):
p. 6660-6662.
252. Katsounaros, I., J.C. Meier, and K.J.J. Mayrhofer, The impact of chloride ions and the catalyst loading
on the reduction of H2O2 on high-surface-area platinum catalysts. Electrochimica Acta, 2013. 110: p.
790-795.
253. Katsounaros, I., et al., The impact of spectator species on the interaction of H2O2 with platinum -
implications for the oxygen reduction reaction pathways. Physical Chemistry Chemical Physics, 2013.
15(21): p. 8058-8068.
254. Grote, J.-P., et al., Screening of material libraries for electrochemical CO2 reduction catalysts –
Improving selectivity of Cu by mixing with Co. Journal of Catalysis, 2016. 343: p. 248-256.
References
| 145
255. Schuppert, A.K., et al., Potential-resolved dissolution of Pt-Cu: A thin-film material library study.
Electrochimica Acta, 2014. 144: p. 332-340.
256. Schuppert, A.K., et al., A Scanning Flow Cell System for Fully Automated Screening of Electrocatalyst
Materials. Journal of the Electrochemical Society, 2012. 159(11): p. F670-F675.
257. Topalov, A.A., et al., The impact of dissolved reactive gases on platinum dissolution in acidic media.
Electrochemistry Communications, 2014. 40: p. 49-53.
258. Wagner, C. and W. Traud, Über die Deutung von Korrosionsvorgängen durch Überlagerung von
elektrochemischen Teilvorgängen und über die Potentialbildung an Mischelektroden. Zeitschrift für
Elektrochemie und angewandte physikalische Chemie, 1938. 44(7): p. 391-402.
259. Spiro, M., A critique of the additivity principle for mixed couples
Modern aspects of electrochemistry: No.34. 2001, luwer Academic / Plenum Publishers. p. 1-11.
260. Creeth, A.M. and M. Spiro, A re-formulation of the Wagner and Traud additivity principle: catalytic
and electrochemical experiments with iodide modified surfaces. Journal of Electroanalytical Chemistry
and Interfacial Electrochemistry, 1991. 312(1): p. 165-174.
261. Spiro, M., Heterogeneous Catalysis in Solution .17. Kinetics of Oxidation-Reduction Reactions
Catalyzed by Electron-Transfer through the Solid - Electrochemical Treatment. Journal of the
Chemical Society-Faraday Transactions I, 1979. 75: p. 1507-1512.
262. Spiro, M., Polyelectrodes: the behaviour and applications of mixed redox systems. Chemical Society
Reviews, 1986. 15(2): p. 141-165.
263. Pizzutilo, E., et al., Addressing stability challenges of using bimetallic electrocatalysts: the case of gold-
palladium nanoalloys. Catalysis Science & Technology, 2017. 7: p. 1848-1856.
264. Li, Z.J., et al., Formation and characterization of Au/Pd surface alloys on Pd(111). Surface Science,
2007. 601(8): p. 1898-1908.
265. Hufner, S., G.K. Wertheim, and J.H. Wernick, Xps Core Line Asymmetries in Metals. Solid State
Communications, 1975. 17(4): p. 417-422.
266. Birss, V.I., et al., An electrochemical study of the composition of thin, compact Pd oxide films. Journal
of the Chemical Society-Faraday Transactions, 1996. 92(20): p. 4041-4047.
267. Zhang, A.J., V.I. Birss, and P. Vanysek, Impedance Characterization of Thin Electrochemically Formed
Palladium Oxide-Films. Journal of Electroanalytical Chemistry, 1994. 378(1-2): p. 63-76.
268. Lewis, F.A., The Palladium Hydrogen System. Academic Press, 1967.
269. Hamelin, A., Cyclic voltammetry at gold single-crystal surfaces. Part 1. Behaviour at low-index faces.
Journal of Electroanalytical Chemistry, 1996. 407(1): p. 1-11.
270. Tian, M., W.G. Pell, and B.E. Conway, Nanogravimetry study of the initial stages of anodic surface
oxide film growth at Au in aqueous HClO4 and H2SO4 by means of EQCN. Electrochimica Acta, 2003.
48(18): p. 2675-2689.
271. Angersteinkozlowska, H., et al., Elementary Steps of Electrochemical Oxidation of Single-Crystal
Planes of Au .1. Chemical Basis of Processes Involving Geometry of Anions and the Electrode Surfaces.
Electrochimica Acta, 1986. 31(8): p. 1051-1061.
272. Angersteinkozlowska, H., et al., Elementary Steps of Electrochemical Oxidation of Single-Crystal
Planes of Au .2. A Chemical and Structural Basis of Oxidation of the (111) Plane. Journal of
Electroanalytical Chemistry, 1987. 228(1-2): p. 429-453.
273. Zalineeva, A., et al., Electrochemical Behavior of Unsupported Shaped Palladium Nanoparticles.
Langmuir, 2015. 31(5): p. 1605-1609.
274. Cherevko, S., N. Kulyk, and C.H. Chung, Nanoporous palladium with sub-10 nm dendrites by
electrodeposition for ethanol and ethylene glycol oxidation. Nanoscale, 2012. 4(1): p. 103-105.
275. Lukaszewski, M., et al., Electrosorption of hydrogen into palladium-gold alloys. Journal of Solid State
Electrochemistry, 2003. 7(2): p. 69-76.
146 |
276. Gossner, K. and E. Mizera, Alloy Formation at the Deposition of Palladium on Gold at Room-
Temperature. Journal of Electroanalytical Chemistry, 1981. 125(2): p. 359-366.
277. Henning, S., J. Herranz, and H.A. Gasteiger, Bulk-Palladium and Palladium-on-Gold Electrocatalysts
for the Oxidation of Hydrogen in Alkaline Electrolyte. Journal of the Electrochemical Society, 2015.
162(1): p. F178-F189.
278. Rand, D.A.J. and R. Woods, Determination of Surface Composition of Smooth Noble-Metal Alloys by
Cyclic Voltammetry. Journal of Electroanalytical Chemistry, 1972. 36(1): p. 57-69.
279. Gossner, K. and E. Mizera, The Anodic-Oxidation of Gold + Palladium Alloys. Journal of
Electroanalytical Chemistry, 1982. 140(1): p. 47-56.
280. Zhang, A.J., M. Gaur, and V.I. Birss, Growth of Thin, Hydrous Oxide-Films at Pd Electrodes. Journal
of Electroanalytical Chemistry, 1995. 389(1-2): p. 149-159.
281. Łukaszewski, M., M. Soszko, and A. Czerwiński, Electrochemical Methods of Real Surface Area
Determination of Noble Metal Electrodes - an Overview. International Journal of Electrochemical
Science, 2016. 11(6): p. 4442-4469.
282. Seo, M. and M. Aomi, Piezoelectric Response to Surface Stress Change of a Palladium Electrode in
Sulfate Aqueous-Solutions. Journal of the Electrochemical Society, 1992. 139(4): p. 1087-1090.
283. Tian, M. and B.E. Conway, Phenomenology of oscillatory electro-oxidation of formic acid at Pd: role of
surface oxide films studied by voltammetry, impedance spectroscopy and nanogravimetry. Journal of
Electroanalytical Chemistry, 2005. 581(2): p. 176-189.
284. Gossner, K. and E. Mizera, The Anodic Behavior of Pd Electrodes in 1-M H2so4. Journal of
Electroanalytical Chemistry, 1981. 125(2): p. 347-358.
285. Chausse, V., P. Regull, and L. Victori, Formation of a Higher Palladium Oxide in the Oxygen Evolution
Potential Range. Journal of Electroanalytical Chemistry, 1987. 238(1-2): p. 115-128.
286. Birss, V.I., et al., Properties of thin, hydrous Pd oxide films. Journal of Electroanalytical Chemistry,
1997. 429(1-2): p. 175-184.
287. Correia, A.N., et al., Active surface area determination of Pd-Si alloys by H-adsorption. Electrochimica
Acta, 1997. 42(3): p. 493-495.
288. Ramirez-Caballero, G.E., et al., Surface segregation and stability of core-shell alloy catalysts for oxygen
reduction in acid medium. Physical Chemistry Chemical Physics, 2010. 12(9): p. 2209-2218.
289. Sasaki, K., et al., Core-Protected Platinum Monolayer Shell High-Stability Electrocatalysts for Fuel-
Cell Cathodes. Angewandte Chemie-International Edition, 2010. 49(46): p. 8602-8607.
290. Sasaki, K., et al., Highly stable Pt monolayer on PdAu nanoparticle electrocatalysts for the oxygen
reduction reaction. Nature Communications, 2012. 3: p. 1115-1124.
291. Greeley, J. and J.K. Norskov, Electrochemical dissolution of surface alloys in acids: Thermodynamic
trends from first-principles calculations. Electrochimica Acta, 2007. 52(19): p. 5829-5836.
292. Gatalo, M., et al., Positive Effect of Surface Doping with Au on the Stability of Pt-Based
Electrocatalysts. ACS Catalysis, 2016. 6(3): p. 1630-1634.
293. Cherevko, S., et al., Pt Sub-Monolayer on Au: System stability and insights into Platinum
electrocheical dissolution. Journal of the Electrochemical Society, 2016. 163(3): p. H228-H233.
294. Woods, R., Electrolytic Co-Deposited Palladium-Gold Electrodes - Effect of Potential Cycles on Surface
Properties. Electrochimica Acta, 1969. 14(7): p. 632-&.
295. Dona, J.M. and J. Gonzalezvelasco, Mechanism of Surface-Diffusion of Gold Adatoms in Contact with
an Electrolytic Solution. Journal of Physical Chemistry, 1993. 97(18): p. 4714-4719.
296. Fujita, T., et al., Atomic origins of the high catalytic activity of nanoporous gold. Nat Mater, 2012.
11(9): p. 775-780.
297. Asao, N., et al., Nanostructured Materials as Catalysts: Nanoporous-Gold-Catalyzed Oxidation of
Organosilanes with Water. Angewandte Chemie International Edition, 2010. 49(52): p. 10093-10095.
References
| 147
298. Biener, J., et al., Nanoporous Gold: Understanding the Origin of the Reactivity of a 21st Century
Catalyst Made by Pre-Columbian Technology. ACS Catalysis, 2015. 5(11): p. 6263-6270.
299. Alayoglu, S., et al., Surface Composition and Catalytic Evolution of Au (x) Pd1-x (x=0.25, 0.50 and
0.75) Nanoparticles Under CO/O-2 Reaction in Torr Pressure Regime and at 200 A degrees C.
Catalysis Letters, 2011. 141(5): p. 633-640.
300. Prater, K.B. and A.J. Bard, Rotating Ring-Disk Electrodes .1. Fundamentals of Digital Simulation
Approach Disk and Ring Transients and Collection Efficiencies. Journal of the Electrochemical Society,
1970. 117(2): p. 207-&.
301. Guerin, S., et al., A combinatorial approach to the study of particle size effects on supported
electrocatalysts: Oxygen reduction on gold. Journal of Combinatorial Chemistry, 2006. 8(5): p. 679-686.
302. Nagaiah, T.C., et al., Electrochemically Deposited Pd-Pt and Pd-Au Codeposits on Graphite Electrodes
for Electrocatalytic H2O2 Reduction. Analytical Chemistry, 2013. 85(16): p. 7897-7903.
303. Kang, M., et al., Simple Electrodeposition of Dendritic Pd Without Supporting Electrolyte and Its
Electrocatalytic Activity Toward Oxygen Reduction and H2O2 Sensing. Electroanalysis, 2013. 25(12):
p. 2691-2699.
304. Okube, M., et al., Topologically Sensitive Surface Segregations of Au-Pd Alloys in Electrocatalytic
Hydrogen Evolution. Chemelectrochem, 2014. 1(1): p. 207-212.
305. Venkatachalam, S. and T. Jacob, Hydrogen adsorption on Pd-containing Au(111) bimetallic surfaces.
Physical Chemistry Chemical Physics, 2009. 11(17): p. 3263-3270.
306. Gao, F., Y.L. Wang, and D.W. Goodman, CO Oxidation over AuPd(100) from Ultrahigh Vacuum to
Near-Atmospheric Pressures: CO Adsorption-Induced Surface Segregation and Reaction Kinetics.
Journal of Physical Chemistry C, 2009. 113(33): p. 14993-15000.
307. Hahn, C., et al., Synthesis of thin film AuPd alloys and their investigation for electrocatalytic CO2
reduction. Journal of Materials Chemistry A, 2015. 3(40): p. 20185-20194.
308. Soto-Verdugo, V. and H. Metiu, Segregation at the surface of an Au/Pd alloy exposed to CO. Surface
Science, 2007. 601(23): p. 5332-5339.
309. Rasouli, S., et al., Surface area loss mechanisms of Pt3Co nanocatalysts in proton exchange membrane
fuel cells. Journal of Power Sources, 2017. 343: p. 571-579.
310. Ahluwalia, R.K., et al., Dynamics of Particle Growth and Electrochemical Surface Area Loss due to
Platinum Dissolution. Journal of The Electrochemical Society, 2014. 161(3): p. F291-F304.
311. Mezzavilla, S., et al., Structure–Activity–Stability Relationships for Space-Confined PtxNiy
Nanoparticles in the Oxygen Reduction Reaction. ACS Catalysis, 2016: p. 8058-8068.
312. Gan, L., et al., Core-Shell Compositional Fine Structures of Dealloyed PtxNi1-x Nanoparticles and
Their Impact on Oxygen Reduction Catalysis. Nano Letters, 2012. 12(10): p. 5423-5430.
313. Rasouli, S., et al., On the Degradation of PtNi nanocatalysts for PEM Fuel Cells: An Identical Location
Aberration-corrected STEM Study. Microscopy and Microanalysis, 2016. 22(S3): p. 1358-1359.
314. Jiang, L., et al., Size-Dependent Activity of Palladium Nanoparticles for Oxygen Electroreduction in
Alkaline Solutions. Journal of the Electrochemical Society, 2009. 156(5): p. B643-B649.
315. Kondo, S., et al., Active Sites for the Oxygen Reduction Reaction on the Low and High Index Planes of
Palladium. Journal of Physical Chemistry C, 2009. 113(29): p. 12625-12628.
316. Damjanovic, A. and V. Brusic, Oxygen Reduction at Pt-Au and Pd-Au Alloy Electrodes in Acid Solution.
Electrochimica Acta, 1967. 12(9): p. 1171-1184.
317. Hasche, F., M. Oezaslan, and P. Strasser, Activity, Stability, and Degradation Mechanisms of
Dealloyed PtCu3 and PtCo3 Nanoparticle Fuel Cell Catalysts. Chemcatchem, 2011. 3(11): p. 1805-
1813.
318. Mani, P., R. Srivastava, and P. Strasser, Dealloyed binary PtM3 (M = Cu, Co, Ni) and ternary PtNi3M
(M = Cu, Co, Fe, Cr) electrocatalysts for the oxygen reduction reaction: Performance in polymer
electrolyte membrane fuel cells. Journal of Power Sources, 2011. 196(2): p. 666-673.
148 |
319. Strasser, P., et al., Lattice-strain control of the activity in dealloyed core-shell fuel cell catalysts. Nature
Chemistry, 2010. 2(6): p. 454-460.
320. Horanyi, G., Heterogeneous Catalysis and Electrocatalysis. Catalysis Today, 1994. 19(2): p. 285-311.
321. Ozkan, U., Bridging Heterogeneous Catalysis and Electro-catalysis: Catalytic Reactions Involving
Oxygen. Topics in Catalysis, 2013. 56(18-20): p. 1603-1610.
322. Over, H., Surface Chemistry of Ruthenium Dioxide in Heterogeneous Catalysis and Electrocatalysis:
From Fundamental to Applied Research. Chemical Reviews, 2012. 112(6): p. 3356-3426.
323. Stonehart, P. and P.N. Ross, The Commonality of Surface Processes in Electrocatalysis and Gas-Phase
Heterogeneous Catalysis. Catalysis Reviews, 1975. 12(1): p. 1-35.
324. Wieckowski, A. and M. Neurock, Contrast and Synergy between Electrocatalysis and Heterogeneous
Catalysis. Advances in Physical Chemistry, 2011. 2011: p. 18.
325. Masatake, H., et al., Novel Gold Catalysts for the Oxidation of Carbon Monoxide at a Temperature far
Below 0 °C. Chemistry Letters, 1987. 16(2): p. 405-408.
326. Rodriguez, P., N. Garcia-Araez, and M.T.M. Koper, Self-promotion mechanism for CO electrooxidation
on gold. Physical Chemistry Chemical Physics, 2010. 12(32): p. 9373-9380.
327. Dimitratos, N., et al., Oxidation of glycerol using gold-palladium alloy-supported nanocrystals.
Physical Chemistry Chemical Physics, 2009. 11(25): p. 4952-4961.
328. Hughes, M.D., et al., Tunable gold catalysts for selective hydrocarbon oxidation under mild conditions.
Nature, 2005. 437(7062): p. 1132-1135.
329. Hutchings, G.J., et al., New approaches to designing selective oxidation catalysts: Au/C a versatile
catalyst. Topics in Catalysis, 2006. 38(4): p. 8.
330. Zope, B.N., et al., Reactivity of the Gold/Water Interface During Selective Oxidation Catalysis. Science,
2010. 330(6000): p. 74-78.
331. Yan, S., et al., Electrocatalytic Performance of Gold Nanoparticles Supported on Activated Carbon for
Methanol Oxidation in Alkaline Solution. The Journal of Physical Chemistry C, 2011. 115(14): p. 6986-
6993.
332. Spendelow, J.S. and A. Wieckowski, Electrocatalysis of oxygen reduction and small alcohol oxidation in
alkaline media. Physical Chemistry Chemical Physics, 2007. 9(21): p. 2654-2675.
333. Shen, P.K. and C. Xu, Alcohol oxidation on nanocrystalline oxide Pd/C promoted electrocatalysts.
Electrochemistry Communications, 2006. 8(1): p. 184-188.
334. Xu, C., et al., Methanol and ethanol electrooxidation on Pt and Pd supported on carbon microspheres in
alkaline media. Electrochemistry Communications, 2007. 9(5): p. 997-1001.
335. Garcia-Serna, J., et al., Engineering in direct synthesis of hydrogen peroxide: targets, reactors and
guidelines for operational conditions. Green Chemistry, 2014. 16(5): p. 2320-2343.
336. Menegazzo, F., et al., Effect of the addition of Au in zirconia and ceria supported Pd catalysts for the
direct synthesis of hydrogen peroxide. Journal of Catalysis, 2008. 257(2): p. 369-381.
337. Bernardotto, G., et al., New Pd–Pt and Pd–Au catalysts for an efficient synthesis of H2O2 from H2 and
O2 under very mild conditions. Applied Catalysis A: General, 2009. 358(2): p. 129-135.
338. Li, J., T. Ishihara, and K. Yoshizawa, Theoretical Revisit of the Direct Synthesis of H2O2 on Pd and
Au@Pd Surfaces: A Comprehensive Mechanistic Study. Journal of Physical Chemistry C, 2011.
115(51): p. 25359-25367.
339. Breiter, M.W., Nature of Strongly Adsorbed Species Formed on Platinized-Platinim after Addition of
Methanol Formic Acid and Formaldehyde. Journal of Electroanalytical Chemistry, 1967. 15(2-3): p.
221-&.
340. Breiter, M.W., A Study of Intermediates Adsorbed on Platinized-Platinum during Steady-State
Oxidation of Methanol Formic Acid and Formaldehyde. Journal of Electroanalytical Chemistry, 1967.
14(4): p. 407-&.
References
| 149
341. Breiter, M.W., Role of Adsorbed Species for Anodic Methanol Oxidation on Platinum in Acidic
Electrolytes. Discussions of the Faraday Society, 1968(45): p. 79-&.
342. Zhao, X., et al., Recent advances in catalysts for direct methanol fuel cells. Energy & Environmental
Science, 2011. 4(8): p. 2736-2753.
343. Li, X., D. Heryadi, and A.A. Gewirth, Electroreduction activity of hydrogen peroxide on Pt and Au
electrodes. Langmuir, 2005. 21(20): p. 9251-9259.
344. Cao, D.X., et al., Kinetics of hydrogen peroxide electroreduction on Pd nanoparticles in acidic medium.
Journal of Electroanalytical Chemistry, 2008. 621(1): p. 31-37.
345. Staykov, A., et al., Theoretical Study of the Direct Synthesis of H2O2 on Pd and Pd/Au Surfaces. The
Journal of Physical Chemistry C, 2008. 112(49): p. 19501-19505.
346. Trasatti, S. and R. Parsons, Interphases in Systems of Conducting Phases (Recommendations 1985).
Pure and Applied Chemistry, 1986. 58(3): p. 437-454.
347. Ravno, A.B. and M. Spiro, Heterogeneous Catalysis in Solution .3. Heterogeneous Catalysis and Other
Types of Interaction between Metals and Oxidation-Reduction Reactions. Journal of the Chemical
Society, 1965(Jan): p. 97-&.
348. Spiro, M., The Heterogeneous Catalysis by Metals of Electron-Transfer Reactions in Solution. Journal of
the Chemical Society, 1960(Sep): p. 3678-3679.
349. Spiro, M. and P.W. Griffin, Proof of an Electron-Transfer Mechanism by Which Metals Can Catalyse
Oxidation-Reduction Reactions. Journal of the Chemical Society D-Chemical Communications,
1969(6): p. 262-&.
350. Spiro, M. and A.B. Ravno, Heterogeneous Catalysis in Solution .2. Effect of Platinum on Oxidation-
Reduction Reactions. Journal of the Chemical Society, 1965(Jan): p. 78-&.
351. Power, G.P. and I.M. Ritchie, Mixed Potential Measurements in the Elucidation of Corrosion
Mechanisms .1. Introductory Theory. Electrochimica Acta, 1981. 26(8): p. 1073-1078.
352. Choudhary, V.R., P. Jana, and S.K. Bhargava, Reduction of oxygen by hydroxylammonium salt or
hydroxylamine over supported Au nanoparticles for in situ generation of hydrogen peroxide in aqueous
or non-aqueous medium. Catalysis Communications, 2007. 8(5): p. 811-816.
353. Choudhary, V.R., P. Jana, and C. Samanta, Generation of hydrogen peroxide via the selective reduction
of oxygen by hydrazine sulfate over Br-promoted Pd/Al2O3 catalyst in an aqueous medium at ambient
conditions. Applied Catalysis A: General, 2007. 323: p. 202-209.
354. Choudhary, V.R., C. Samanta, and P. Jana, A novel route for in-situ H2O2 generation from selective
reduction of O2 by hydrazine using heterogeneous Pd catalyst in an aqueous medium. Chemical
Communications, 2005(43): p. 5399-5401.
355. Choudhary, V.R. and P. Jana, Direct H2-to-H2O2 oxidation over highly active/selective Br–F–
Pd/Al2O3 catalyst in aqueous acidic medium: Influence of process conditions on the H2O2 formation.
Applied Catalysis A: General, 2009. 352(1–2): p. 35-42.
356. Seraj, S., et al., PdAu Alloy Nanoparticle Catalysts: Effective Candidates for Nitrite Reduction in
Water. ACS Catalysis, 2017: p. 3268-3276.
357. Abate, S., et al., The issue of selectivity in the direct synthesis of H2O2 from H2 and O2: the role of the
catalyst in relation to the kinetics of reaction. Topics in Catalysis, 2006. 38(1): p. 181-193.
358. Almora-Barrios, N., et al., Electrochemical Effects at Surfactant–Platinum Nanoparticle Interfaces
Boost Catalytic Performance. ChemCatChem, 2017. 9(4): p. 604-609.
359. Pletcher, D. and S. Sotiropoulos, Hydrogen adsorption-desorption and oxide formation-reduction on
polycrystalline platinum in unbuffered aqueous solutions. Journal of the Chemical Society, Faraday
Transactions, 1994. 90(24): p. 3663-3668.
360. Zhan, D., J. Velmurugan, and M.V. Mirkin, Adsorption/Desorption of Hydrogen on Pt Nanoelectrodes:
Evidence of Surface Diffusion and Spillover. Journal of the American Chemical Society, 2009. 131(41):
p. 14756-14760.
150 |
361. Zhang, J.-h., et al., A remarkable activity of glycerol electrooxidation on gold in alkaline medium.
Electrochimica Acta, 2012. 59: p. 156-159.
362. Su, L., et al., Palladium/titanium dioxide nanofibers for glycerol electrooxidation in alkaline medium.
Electrochemistry Communications, 2009. 11(11): p. 2199-2202.
363. Carrettin, S., et al., Selective oxidation of glycerol to glyceric acid using a gold catalyst in aqueous
sodium hydroxide. Chemical Communications, 2002(7): p. 696-697.
364. Porta, F. and L. Prati, Selective oxidation of glycerol to sodium glycerate with gold-on-carbon catalyst:
an insight into reaction selectivity. Journal of Catalysis, 2004. 224(2): p. 397-403.
365. Camara, G., R. De Lima, and T. Iwasita, Catalysis of ethanol electrooxidation by PtRu: the influence of
catalyst composition. Electrochemistry Communications, 2004. 6(8): p. 812-815.
366. Christensen, C.H., et al., Formation of Acetic Acid by Aqueous‐Phase Oxidation of Ethanol with Air in
the Presence of a Heterogeneous Gold Catalyst. Angewandte Chemie International Edition, 2006.
45(28): p. 4648-4651.
Articles and Conferences
| 151
Articles and Conferences
List of Publications (First Autorship)
1. Minimizing operando demetallation of Fe-NC electrocatalysts in acidic medium.
ACS Catalysis, 2016, 6(5), 3136-3146.
2. Experimental methodologies to understand the degradation of nanostructured
electrocatalysts for PEM fuel cells: advances and opportunities. ChemElectroChem,
2016, 3(10), 1524-1536.
3. Structure–Activity–Stability Relationships for Space-Confined PtxNiy Nanoparticles
in the Oxygen Reduction Reaction. ACS Catalysis, 2016, 6(12), 8058-8068.
4. On the Need of Improved Accelerated Degradation Protocols (ADPs): Examination of
Platinum Dissolution and Carbon Corrosion in Half-Cell Tests. Journal of The
Electrochemical Society, 2016, 163(14), F1510-F1514.
5. Palladium electrodissolution from model surfaces and nanoparticles. Electrochimica
Acta 2017, 229, 467–477.
6. Addressing stability challenges of using bimetallic 1 electrocatalysts: the case of
gold-palladium nanoalloys. Catal. Sci. Technol. 2017, 7, 1848-1856.
7. Electrocatalytic synthesis of hydrogen peroxide on Au-Pd nanoparticles: from
fundamentals to continuous production. Chem. Phys. Lett. 2017, 683, 436-442.
8. The Space Confinement Approach Using Hollow Graphitic Spheres to Unveil
Activity and Stability of Pt-Co Nanocatalysts for PEMFC. Adv. En. Mat., 2017
(accepted).
9. Gold-Palladium Bimetallic Catalyst Stability: Consequences for Hydrogen Peroxide
Selectivity. ACS catalysis, 2017, 7, 5699-5705.
10. Accelerated fuel cell tests of anodic Pt/Ru catalyst via identical location TEM: new
aspects of degradation behavior. International Journal of Hydrogen Energy, 2017
(submitted).
11. The Stability-number as new metric for electrocatalyst stability benchmarking – a
case study of iridium-based oxides towards acidic water splitting, Nature Energy,
2017 (submitted).
12. On-demand H2O2 production: a study at the border between electro and
heterogeneous catalysis. (in preparation).
152 |
List of Conferences
Orals
1. On-site production of hydrogen peroxide in electrocatalysis and in heterogeneous
catalysis. 2017, Electrochemistry 2017, Berlin, Germany
2. Hydrogen peroxide on-site production: a fundamental study on the direct synthesis
and electrocatalytic synthesis using Au-Pd catalysts. 2017, ISE Topical Meeting,
Buenos Aires, Argentina
3. Activity and stability investigation of PtCo@HGS. 2016, ERTL Symposium, Berlin,
Germany
4. Gold-Palladium Catalysts: Towards H2O2 Production: Direct Synthesis and/or
Electrocatalytic Synthesis?. 2016, ECS, San Diego, California
5. H2O2 synthesis: comparative study of electrocatalysis and heterogeneous catalysis
with gold-palladium catalysts. 2016, MaxNet Meeting, Berlin, Germany
6. Gold-Palladium catalysts for the electrocatalytic production of H2O2. 2015, CCI,
Cardiff, UK
Posters
1. Increased Stability of Pore Confined Pt-Co Electrocatalyst for PEMFC. 2017, IRES,
Düsseldorf, Germany.
2. Improving the stability of PEMFC catalyst by space confinement. 2016, Ph.D.
program evaluation, Düsseldorf, Germany.
Curriculum Vitae
| 153
Curriculum Vitae
Enrico Pizzutilo
Oberbilker Allee 88, 40227 Düsseldorf, Germany
+49 15201443786
Place of birth:
Nationality:
Date of birth:
Verona
Italian
26.08.1989
Education
10/2014-10/2017
Ruhr Universität Bochum, Germany
PhD in Mechanical Engineering
International Max Planck Research School for Interface
Controlled Materials for Energy Conversion (IMPRS-
SurMat)
Visiting Student:
University of Copenhagen, Denmark (2 weeks)
Cardiff University, UK (2 weeks)
Jülich Forschungszentrum, Germany (3 months)
10/2011 - 03/2014
Università di Bologna, Italy
Master in Energy Engineering
110/110 e lode
(with honors)
09/2013 – 03/2014
Université Paul Sabatier, France
Master thesis abroad
03/2012 – 08/2012
Technische Universität Berlin, Germany
Erasmus Semester
09/2008 – 10/2011
Università di Bologna, Italy
Bachelor in Energy Engineering
110/110
09/2003 – 07/2008
Liceo Scientifico „Angelo Messedaglia“, Verona, Italy
High-School Diploma
98/100
Work experience
10/2014-10/2017
Max-Planck-Institut für Eisenforschung GmbH, Düsseldorf, Germany
Research associate
11/2016 – 01/2017
AIESEC and Ministry of Women and Vulnerable Populations, Ica, Peru
Volonteering project with children
04/2014 – 07/2014
Toulouse Tech Transfer, Toulouse, France
Development engineer
09/2013 – 03/2014
LAPLACE and CIRIMAT, Toulouse, France
Master thesis
04/2013 – 06/2013
Institute for Electrical and Information Engineering, Università di
Bologna, Italy
Intern
154 |
04/2011 – 09/2011
Chemical Department „G. Ciamician”, Università di Bologna, Italy
Bachelor thesis
03/2011 – 04/2011
Italian National Agency for New Technologies, Energy and Sustainable
Economic Development, Bologna, Italy
Intern
Others Background actor in commercials, school tutoring, inventory, steward in
football stadium (3 years) and in fairs, waiter during summer seasons,
volunteer in children summer camps (4 years)
Further education
Computer skills
Office
Origin
Blender
JAVA JDK und Eclipse
Very good
Very good
Basic
Basic
University course: Matlab, Python, Solid Edge (CAD 3D),
Thermoflex, PSCAD, Comsol Multiphysics
Language skills
Italian
English
German
French
Spanish
Native language
Very good, work language
Good, C1.2 course
Good, work language in
2014
Basic, B1 course
Other
Seminars i.e. “Programming with Java”, “Summer School on
Instrumental Methods in Electrochemistry”, „Presentation Skills“,
„Creative Problem Solving“, “Negotiation Skills”, “Self- and Time
Management”, “Creative Scientific Writing”.