| 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.

| 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


| 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].


| 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,


| 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


| 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].


| 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


| 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.


| 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.


| 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].


| 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.

——————————————————————————————————————————





Acta 2017, 229, 467–477.


Lett.. 2017, 683, 436-442.


Mayrhofer, ACS catalysis. 2017, 7, 5699-5705.


preparation)



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 |


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.

——————————————————————————————————————





Acta 2017, 229, 467–477.


Lett.. 2017, 683, 436-442.


Mayrhofer, ACS catalysis.. 2017, 7, 5699-5705.


preparation)


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.


| 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].


| 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


| 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).


| 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 |


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


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.

——————————————————————————————————————



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


| 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.


| 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.


| 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).


| 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.


| 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


| 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,


| 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


| 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.


| 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:





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


| 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.


study.

Ametal (1l)** /

mm2

Au

AuPd

Pd

0.14

0.18

0.31


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.


| 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


| 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


| 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)


| 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.

——————————————————————————————————————————



Lett.. 2017, 683, 436-442.



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


| 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


| 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,


| 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.


| 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


| 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.

——————————————————————————————————————————



Mayrhofer, ACS catalysis,. 2017, 7, 5699-5705.



- 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:


1000 cycles;


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


| 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


| 105


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.



Initial

100

1000

215

119

115

3.8

4.4

4.3

±1.0

±1.1

±1.2

Chapter 8

106 |


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).



Initial

100

169

116

3.4

4.4

±1.0

±1.3


| 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


| 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


| 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.



Chapter 8

112 |



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).


| 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.

——————————————————————————————————————————



preparation)



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


| 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].


| 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.


| 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


| 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.


| 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


| 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.


| 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.

http://h2o2.evonik.com/product/h2o2/en/application-areas/pages/default.aspx

http://www.prweb.com/releases/hydrogen_peroxide/bleaching_pulp_paper/prweb8067430.htm

http://www.essentialchemicalindustry.org/chemicals/hydrogen-peroxide.html

http://www.gminsights.com/industry-analysis/hydrogen-peroxide-market

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


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


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


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.


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.


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.


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


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.


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.


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


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


284. Gossner, K. and E. Mizera, The Anodic Behavior of Pd Electrodes in 1-M H2so4. Journal of


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.


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.


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.


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.


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.


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.


362. Su, L., et al., Palladium/titanium dioxide nanofibers for glycerol electrooxidation in alkaline medium.


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

[email protected]

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

mailto:[email protected]

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”.

towards on-site production of hydrogen peroxide with gold

Documents