Catalysis - Volume 28

17:23:09.
Published on 19 February 2016 on http://pubs.rsc.org | doi:10.1039/9781782626855-FP001
Catalysis
Volume 28
17:23:09.
View Online
View Online
A Specialist Periodical Report
17:23:09.
Catalysis
Volume 28
A Review of Recent Literature
Editors
James J. Spivey, Louisiana State University, USA
Yi-Fan Han, East China University of Science and Technology,
Shanghai, China
K. M. Dooley, Louisiana State University, USA
Authors
Erfan Behravesh, bo Akademi University, Turku, Finland
Yuxiang Chen, Carnegie Mellon University, Pittsburgh, PA, USA
Marc-Olivier Coppens, University College London, UK
Wei-Lin Dai, Fudan University, Shanghai, China
Jing Ding, Fudan University, Shanghai, China
Angelos M. Efstathiou, University of Cyprus, Nicosia, Cyprus
Xiangchen Fang, Fushun Research Institute of Petroleum and
Petrochemicals and East China University of Science and Technology,
Shanghai, China
Ruihua Gao, Fudan University, Shanghai, China
Yun Hang Hu, Michigan Technological University, MI, USA
Leena Hupa, bo Akademi University, Turku, Finland
Rongchao Jin, Carnegie Mellon University, Pittsburgh, PA, USA
Phumelele E. Kleyi, Nelson Mandela Metropolitan University,
Port Elizabeth, South Africa
Frederic C. Meunier, CNRS, Universite
Lyon, France
Alina Moscu, CNRS, Universite
Lyon, France
Dmitry Murzin, bo Akademi University, Turku, Finland
Adeniyi S. Ogunlaja, Nelson Mandela Metropolitan University,
Chong Peng, Fushun Research Institute of Petroleum and Petrochemicals,
China and East China University of Science and Technology,
Shanghai, China
Ayomi Sheamilka Perera, University College London, UK
Tapio Salmi, bo Akademi University, Turku, Finland
Yves Schuurman, CNRS, Universite
Lyon, France
Zenixole R. Tshentu, Nelson Mandela Metropolitan University,
Ryan S. Walmsley, Rhodes University, Grahamstown, South Africa
View Online
17:23:09.
Wei Wei, Michigan Technological University, MI, USA

Xinli Yang, Fudan University, Shanghai, China
Chenjie Zeng, Carnegie Mellon University, Pittsburgh, PA, USA
Ronghui Zeng, Fushun Research Institute of Petroleum and
Petrochemicals, China
Quanjing Zhu, Fudan University, Shanghai, China
17:23:09.
View Online
ISBN: 978-1-78262-427-1
PDF ISBN: 978-1-78262-685-5
EPUB eISBN: 978-1-78262-805-7
DOI: 10.1039/9781782626855
ISSN: 0140-0568
A catalogue record for this book is available from the British Library
r The Royal Society of Chemistry 2016
All rights reserved
Apart from any fair dealing for the purpose of research or private study
for non-commercial purposes, or criticism or review, as permitted under
the terms of the UK Copyright, Designs and Patents Act, 1988 and the
Copyright and Related Rights Regulations 2003, this publication may not be
reproduced, stored or transmitted, in any form or by any means, without the
prior permission in writing of The Royal Society of Chemistry, or in the case of
reprographic reproduction only in accordance with the terms of the licences
issued by the Copyright Licensing Agency in the UK, or in accordance with
the terms of the licences issued by the appropriate Reproduction Rights
Organization outside the UK. Enquiries concerning reproduction outside the
terms stated here should be sent to The Royal Society of Chemistry at the
address printed on this page.
Published by The Royal Society of Chemistry,
Thomas Graham House, Science Park, Milton Road,
Cambridge CB4 0WF, UK
Registered Charity Number 207890
For further information see our web site at www.rsc.org
Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY
17:23:09.
View Online
Preface
17:23:13.
DOI: 10.1039/9781782626855-FP007
Chapter 1: Tungsten containing materials as heterogeneous catalysts for

green catalytic oxidation process
This review is given by Wei-Lin Dai, Jing Ding, Quanjing Zhu, Ruihua
Gao, Xinli Yang from Fudan university, Shanghai, China. It aims to
provide a comprehensive description of the recent advances in the field
of tungsten-containing heterogeneous catalyst for green catalytic oxidation process. This review collects more than 90 literatures and consists
of three sections. The first part exhibits the advances in the pristine
tungsten-based catalysts for the green catalytic oxidation process; the
second one highlights various green catalytic oxidation reactions with
tungsten-based catalysts supported on dierent carriers; the last one illustrates the existing problems and outlook for the tungsten-based
catalysts applied to the green catalytic oxidation reactions. The examples
discussed in this review highlight the need to design and synthesis of
tungsten-based catalysts. Perhaps more importantly, they also are of
value for researchers in the area of heterogeneous catalysis to develop
highly ecient green oxidation catalytic systems.
Chapter 2: Alumina ceramic foams as catalyst supports
In the next review, Alumina ceramic foams as catalyst supports is reviewed by Erfan Behravesh, Leena Hupa, Tapio Salmi, Dmitry Yu. Murzin
from bo Akademi University, Finland. Ceramic foams have a wide range
of potential applications in biomedicine, thermal insulation, filtration of
molten metal alloys, absorption of environmental pollutants, catalyst
supports, etc. Herein, three main methods of manufacturing ceramic
foams are introduced with the main emphasize on the replica technique.
Furthermore, dierent techniques for improving structural properties of
ceramic foams are reviewed. The focus of this review is on fabrication of
macro-porous alumina foams with high interconnected porosity. In
addition, experimental data for manufacturing of ceramic foams via the
replica technique is represented along with literature surveys. Slurries
consisted of alumina powder mixed in aqueous solutions of polyvinyl
alcohol (PVA) and magnesia and titania as sintering aids. The foams were
produced by tuning dierent processing parameters to give properties
suited for catalyst supports. These parameters included pore size of the
polyurethane (PU) foam used as a template, parameters in the PU foam
pretreatment, particle size of alumina powder in the slurry, slurry loading
and drying of the green alumina coated PU foam. Finally, the key factors
for optimizing ceramic foams in terms of mechanical strength and
interconnectivity are introduced together with an outlook for future advances in ceramic foams as catalyst supports.
Catalysis, 2016, 28, viix | vii

c
The Royal Society of Chemistry 2016
17:23:13.
View Online
Chapter 3: Recent advances in the synthesis and catalytic applications of

atomically precise gold nanoclusters
This review is contributed by Yuxiang Chen, Chenjie Zeng, and Rongchao
Jin from Carnegie Mellon University. This review summaries the recent
advances in the synthesis and catalytic application of atomically precise
Aun(SR)m nanoclusters. Structurally characterized nanoclusters can serve
as new model catalysts for obtaining atomic/molecular level insights into
the catalytic processes, including the precise size-dependent catalytic
reactivity and how molecules are adsorbed and activated on the catalytic
active sites, as well as the structural sensitivity of the catalyst to the reactions. While this area is still in its infancy, promising work has been
reported and demonstrated the catalytic power of atomically precise
nanoclusters. Such reactions include catalytic oxidation, chemoselective
catalytic hydrogenation, catalytic semihydrogenation, etc. In addition,
precisely doped nanoclusters provide a unique opportunity to tune the
catalytic reactivity on a truly atom-by-atom basis. Overall, atomically
precise nanoclusters hold great promise in the discovery of unique
catalytic processes as well as in advancing the fundamental understanding of catalytic mechanisms at the atomic/molecular level.
Chapter 4: Research and Development of Hydrocracking Catalysts and
technology
Research and Development of Hydrocracking Catalysts and technology is
reviewed by Chong Peng, Xiangchen Fang and Ronghui Zeng from both
Fushun Research Institute of Petroleum and Petrochemicals, SINOPEC,
China and East China University of Science and Technology, Shanghai,
China. Hydrocracking (HCK), one of the main approaches to deep process heavy oil, is a catalytic conversion process where feedstock undergoes hydrogenation, S/N removal, molecular restructuring, cracking, and
other reactions. It can process straight-run gasoline/diesel, vacuum gas
oil, and other secondary processing fractions such as fluid catalytic
cracking (FCC) diesel, FCC clarified oil, coker diesel, coker gas oil, and
deasphalted oil and produce various quality clean fuels such as liquefied
gas, gasoline, kerosene, jet fuel, diesel, and various quality petrochemical
materials such as light/heavy naphtha and tail oil. In this chapter, the
research progress on commercial HCK technology and its relative catalysts are discussed. The typical technical characteristics and the representative processes from dierent corporations, such as Universal Oil
Products, Albemarle, Criterion, Haldor Topsoe, and SINOPEC, are also
presented. The development trend of HCK technology in the future is
outlined.
Chapter 5: Titano-silicates: Their history, evolution and scope of
application
Ayomi Perera and Marc-Olivier Coppens of University College London
review the rapidly evolving catalysis of titano-silicates, especially as applied to selective oxidations. There have been significant advances in
understanding how these materials function, and in the synthesis of new
Ti-silicate structures. A key application is in epoxide manufacture, and
viii | Catalysis, 2016, 28, viix
View Online
17:23:13.
this application is reviewed here. But several potential green chemical

processes are under evaluation, and these are discussed and critiqued
as well.
Chapter 6: Nanofiber-supported metal-based catalysts
A group headed by Zenixole Tshentu from Nelson Mandela Metropolitan
and Rhodes Universities have reviewed nanofiber-supported metal catalysts. The review covers most of the conventional catalytic transition
metal/metal oxides supported on dierent types of electrospun nanofibers. Catalytic metal ion complexes supported on electrospun nanofibers, via coordination to the desired functional groups of polymer chains,
have also been discussed. While the use of electrospun nanofibers as
catalyst support is still at its infancy stage, several application studies
have shown that the use of nanofiber-based catalytic materials exhibited
good catalytic activity as a result of the increased surface area-to-volume
ratio. There is discussion of catalyst reusability and challenges associated
with the use of electrospun nanofibers in catalysis.
Chapter 7: Elucidation of Mechanistic and Kinetic Aspects of Water-Gas
Shift Reaction on Supported Pt and Au Catalysts via Transient Isotopic
Techniques
Angelos Efstathiou of the University of Cyprus comprehensively summarizes recent work from the low-temperature watergas shift catalysis
literature. He places special emphasis on steady-state isotopic transient
kinetic analysis and other transient isotopic techniques to probe mechanism and determine important kinetic parameters for supported Pt and
Au catalysts on reducible and non-reducible metal oxides. These results
are extended and put into context by comparing to predictions of recent
computational (DFT) studies.
Chapter 8: Recent progresses on the use of supported bimetallic catalysts
for the preferential oxidation of CO (PROX)
Alina Moscu, Yves Schuurman, and Frederic Meunier (Institut de
Recherches sur la Catalyse et lEnvironnement de Lyon) report on current
progress and prospects on PROX, an essential step in fuel processing of
hydrocarbons to produce CO-free dihydrogen. This can be converted into
clean energy, particularly in PEM fuel cells. The CO just downstream of
WGS must be reduced to low levels, often to ppm concentrations. Improvements must be addressed at low temperatures. Recent studies have
shown progress in bimetallic compounds compared to monometallic
counterparts. The improved activity of alloys and bimetallic appear to be
to geometric or electronic eects. The review here pays particular attention to in situ IR-based studies realized over Pt-based formulations, since
CO is both a reactant and a molecular probe enabling the determination
of the state of metals under reaction conditions. Reaction results at
conditions at the working catalyst are essential. In situ and operando
conditions enable can be used to probe the true active phases, because
alloy segregation can readily occur even due to minor modification of the
experimental conditions.
Catalysis, 2016, 28, viix | ix
17:23:13.
View Online
Chapter 9: 3D MoS2/Graphene Hybrid Layer Materials as Counter Electrodes for Dye-Sensitized Solar Cells
Wei Wei and Yun Hang Hu (Michigan Tech) discuss the development of
dye-sensitized solar cells (DSSCs), which have attracted considerable attention as an alternative to conventional silicon based solar cells because
of their low cost, low energy consumption, simple fabrication process,
and high power conversion eciency. Typically, DSSCs are composed of a
photoelectrode (a transparent conducting subtract with a dye coated TiO2
film), an electrolyte, and a counter electrode (CE). The synthesis and
DSSC counter electrode applications of graphene sheets and MoS2 materials are briefly reviewed. Furthermore, in this chapter, they report a
new method to synthesize 3D MoS2/graphene hybrid layer materials as
counter electrode catalysts for DSSCs.
x | Catalysis, 2016, 28, viix
CONTENTS
17:23:16.
Cover
Image provided courtesy of
computational science company
Accelrys (www.accelrys.com). An
electron density isosurface mapped
with the electrostatic potential for an
organometallic molecule.
This shows the charge distribution
across the surface of the molecule
with the red area showing the
positive charge associated with the
central metal atom. Research carried
out using Accelrys Materials
Studioss.
Preface
vii
Tungsten containing materials as heterogeneous catalysts for

green catalytic oxidation process
Wei-Lin Dai, Jing Ding, Quanjing Zhu, Ruihua Gao and

Xinli Yang
1 Introduction
2 Pristine W-based catalyst for green catalytic oxidation
3 W-based catalyst supported on dierent carriers for
green catalytic oxidation
4 Conclusion and outlook
Acknowledgements
References
Alumina ceramic foams as catalyst supports
1
2
6
23
24
24
28
Erfan Behravesh, Leena Hupa, Tapio Salmi and

Dmitry Yu. Murzin
1 Introduction
2 Experimental
3 Results and discussion
4 Conclusions and outlook
Acknowledgements
References
28
32
35
46
47
47
Catalysis, 2016, 28, xixiii | xi
View Online
17:23:16.
Recent advances in the synthesis and catalytic applications

of atomically precise gold nanoclusters
Yuxiang Chen, Chenjie Zeng and Rongchao Jin
1 Introduction
2 Synthesis, structure, and properties of gold nanoclusters
3 Catalytic properties of Aun(SR)m nanoclusters
4 Summary
Acknowledgements
References
Research and development of hydrocracking catalysts and

technology
51
51
53
65
82
83
83
86
Chong Peng, Xiangchen Fang and Ronghui Zeng

1 Introduction
2 History
3 Hydrocracking catalyst
4 Hydrocracking processes
5 Kinetic models of hydrocracking
6 Consideration for hydrocracking technology development
References
Titano-silicates: highlights on development, evolution and

application in oxidative catalysis
86
87
88
94
110
116
117
119
Ayomi Sheamilka Perera and Marc-Olivier Coppens

1 Introduction
2 Synthesis and characterisation of titano-silicates
3 Application of titano-silicates as oxidative catalysts
4 Outlook: is the future bright for titano-silicate catalysts?
5 Summary
Acknowledgements
References
Nanofiber-supported metal-based catalysts

Adeniyi S. Ogunlaja, Phumelele E. Kleyi, Ryan S. Walmsley and
Zenixole R. Tshentu
1 Introduction
2 Nanofiber-supported metal/metal oxide catalysts
3 Nanofiber-supported metal complexes
4 Challenges associated with electrospun nanofibers
5 Conclusions
References
xii | Catalysis, 2016, 28, xixiii
119
122
124
134
137
138
138
144
144
146
156
165
169
169
View Online
Elucidation of mechanistic and kinetic aspects of watergas

shift reaction on supported Pt and Au catalysts via transient
isotopic techniques
175
17:23:16.
Angelos M. Efstathiou
1
2
3
4
Introduction
Watergas shift reaction mechanisms
SSITKAoperando methodology
Application of SSITKA and other transient isotopic
techniques towards elucidation of WGS reaction
mechanisms
5 Conclusions
Acknowledgements
References
175
180
187
206
Recent progresses on the use of supported bimetallic catalysts

for the preferential oxidation of CO (PROX)
237
230
231
231
Alina Moscu, Yves Schuurman and Frederic C. Meunier

1 Benefits and challenges associated with the use of
multi-metallic materials as catalysts and electrodes
2 Hydrogen as energy carrier in fuel cell applications
3 Main features of the preferential CO oxidation (PROX)
4 Platinum-tin-based catalysts
5 Surface analyses by in situ infrared spectroscopy
6 Conclusions
References
3D MoS2/Graphene hybrid layer materials as counter electrodes

for dye-sensitized solar cells
237
238
240
251
254
262
262
268
Wei Wei and Yun Hang Hu

1 Introduction
2 Preparation of DSSCs
3 Characterization of DSSCs
4 Conclusions
Acknowledgements
References
268
270
272
278
278
278
Catalysis, 2016, 28, xixiii | xiii
17:23:16.
View Online
17:23:19.
Published on 19 February 2016 on http://pubs.rsc.org | doi:10.1039/9781782626855-00001
Tungsten containing materials as

heterogeneous catalysts for green
catalytic oxidation process
Wei-Lin Dai,* Jing Ding, Quanjing Zhu, Ruihua Gao and
Xinli Yang
DOI: 10.1039/9781782626855-00001
This chapter provides a comprehensive description of the recent advances in the field
of tungsten-containing heterogeneous catalysts for green catalytic oxidation processes.
This chapter contains three sections. The first exhibits the advances in the pristine
tungsten-based catalysts for the green catalytic oxidation process. The second highlights
various green catalytic oxidation reactions with tungsten-based catalysts supported on
dierent carriers. The third illustrates the existing problems and outlook for the tungstenbased catalysts applied in the green catalytic oxidation reactions. All these contributions
provide a proper guide and overview to tungsten-based catalysts for the sake of the better
development of highly ecient green oxidation catalytic systems.
Introduction
Green Chemistry is now a central issue in both academia and industry

due to global contamination and other environmental and health risks of
chemicals. Oxidation is a core technology for converting petroleum-based
materials to useful chemicals of a higher oxidation state. However, the
traditional methods of oxidation always require the use of stoichiometric
levels of oxidant such as chromates, permanganates and periodates,
which could result in complex heat management needs and by-products
that are harmful to the environment.1 As the increasing emerge of
the environmental issue, searching for a green synthesis method of
the catalysts and further for an environmentally friendly process of the
catalytic oxidation are urgently needed to meet the challenging environmental demands and industrialization requirements. Nowadays, a clean
and environmentally friendly process using environmental benign
oxidants such as air,2 oxygen3 or hydrogen peroxide4 are preferred.
In recent years, there has been a growing interest in tungsten-based
catalyst due to their properties and wide potential applications. This kind
of materials exhibits unique acidity that improves the activity and
selectivity, chemical and thermal stability and environmental protection,
thus being regarded as a kind of promising catalyst owing to its versatile
application, including the metathesis and isomerization of alkenes,5 selective oxidation of unsaturated compounds,6 hydrodesulfurization and
hydrocracking of heavy fractions in petroleum chemistry,7,8 and dehydrogenation of alcohols.9 Moreover, the tungsten-based catalyst can be
applied to the photocatalytic oxidation reaction, especially tungsten
Department of Chemistry & Shanghai Key Laboratory of Molecular Catalysis and
Innovative Material, Fudan University, Shanghai 200433, P. R. China.
E-mail: wldai@fudan.edu.cn
Catalysis, 2016, 28, 127 | 1

c
17:23:19.
View Online
trioxide. Tungsten trioxide, with a band gap between 2.4 and 2.8 eV, is a
visible light response catalyst with stability in acidic reaction conditions.10
The deeply positive level of the valence band (VB) makes it suitable for
achieving the ecient oxidative decomposition of the organic compounds
under solar or visible irradiation. In addition, tungsten-based catalysts also
have been widely used in the field of electrocatalysis. For instance, tungsten trioxide, dependent on its own special electrochemical properties, can
be used as an anode catalyst for hydrogen oxidation in proton exchange
membrane fuel cells (PEMFCs), promoting the development of PEMFCs.
In this chapter, we aim to provide a comprehensive description of
the recent advances in the field of tungsten-containing heterogeneous
catalyst for green catalytic oxidation process. Though significant
advances have been achieved in the development of the tungstencontaining heterogeneous catalyst in green catalytic oxidation process,
there is rarely relevant review focusing on the aspect. Hence, this is our
main driving force to contribute to the fundamental exemplifications
in this field. This review collects more than 90 literatures and consists
of three sections. The first part exhibits the advances in the pristine
tungsten-based catalysts for the green catalytic oxidation process; the
second one highlights various green catalytic oxidation reactions with
tungsten-based catalysts supported on dierent carriers; the last one
illustrates the existing problems and outlook for the tungsten-based
catalysts applied to the green catalytic oxidation reactions. Herein, we
sincerely wish to give colleagues a proper guide and overview to tungstenbased catalysts for the sake of the better development of highly ecient
green oxidation catalytic systems.
Pristine W-based catalyst for green catalytic oxidation
Pristine tungsten-based catalyst mainly includes tungsten trioxide,

tungstic acid, tungsten carbide, etc. Recently, a great deal of research
based on pristine tungsten-based catalysts was focused on various green
catalytic oxidation reactions, which exhibited excellent catalytic performance and chemical stability.
2.1 Tungsten trioxide catalyst for green catalytic oxidation
Tungsten trioxide (WO3) is an important rare earth oxide and can be used
in the field of gas sensors,11 catalysis,12 electrochemistry13 and photocatalysis14 because of its unique gasochromic, acidity, electrochromic
and photochromic properties. Tungsten trioxide, a very ecient catalyst
for various acid catalyzed reactions due to its solid acidic and redox
properties, is applied to some oxidation reactions, such as the selective
oxidation of alcohols or aldehydes and the epoxidation of alkenes.
Su et al. synthesized hexagonal single crystal WO3 nanorods with the aid
of p-aminobenzoic acid through a facile hydrothermal method.15 The
catalyst showed good performance for oxidation of cyclohexene to adipic
acid compared to commercial WO3. Meanwhile, WO3, with a band gap
between 2.4 and 2.8 eV, is a visible light response catalyst with stability
under acidic conditions, which makes it a suitable choice for the
2 | Catalysis, 2016, 28, 127
View Online
17:23:19.
photocatalytic oxidation of organic pollutants under solar irradiation.

The possible photocatalytic oxidation process using WO3 as a photocatalyst can be expressed by the following set of equations.
WO3 hv-e h1
(1)
H2O h1-H2O1-HO H1
(2)
O2 e-O2
(3)
H2O-HO H1
(4)
HO h1-HO
(5)
O2 H1-HOO
(6)
O2 H1-HO2
(7)
1
HO2 HO2 -H2 O2 O2
2
(8)
H2O2 H1 e-HO H2O
(9)
When photocatalyst WO3 is irradiated with light energy above 2.7 eV,
electron (e) and hole (h1) pairs are generated (eqn (1)). Then, the
photogenerated holes can react with water hydroxide ions to produce
hydroxyl radicals (eqn (2) and (4)). Meanwhile, the reduction of dissolved
oxygen by the photogenerated electrons results in the generation of
superoxide anion radicals. Meanwhile, radicals can be initiated, generating a variety of charged neutral and ionic species (eqn (59)).
Villa et al. reported that mesoporous WO3 was prepared by replicating
technique using ordered mesoporous silica KIT-6 as the template, whose
surface area reached up to 151 m2 g1.16 The photocatalytic oxidation of
CH4 into CH3OH over this catalyst using electron scavengers and H2O2
was studied and the photocatalytic activity of WO3 toward methanol
production could be enhanced by a factor of 2.5 and 1.7 when adding
Fe31 and Cu21, respectively. Meanwhile, the photocatalytic activity is
significantly aected by the particle morphology (including shape and
size) of WO3. Hameed et al. further claimed that the photocatalytic
activity for the degradation of phenols could be enhanced by controlling
the morphology of the photocatalyst (disc-shaped WO3).17 The possible
route of degradation/mineralization of phenolic substrates was further
investigated in Fig. 1. WO3 NPs are also shown to possess excellent
photocatalytic degradation of pharmaceutical compounds, like Lidocaine
from water.18 The study suggests that photocatalytic property of WO3 NPs
Catalysis, 2016, 28, 127 | 3
17:23:19.
View Online
Fig. 1 The possible route of degradation/mineralization of phenolic substrates. Reprinted

with M. Aslam et al., Morphology controlled bulk synthesis of disc-shaped WO3 powder
and evaluation of its photocatalytic activity for the degradation of phenols, J. Hazard.
Mater., 276, 120128. Copyright (2014), with permission from Elsevier.17
can be taken for large-scale application in pharmaceutical industries for

the ecient removal of pharmaceutical stu from euents.
2.2 Tungstic acid catalyst for green catalytic oxidation
Tungstic acid, as a well-known and widely used solid acid catalyst,
has received a considerable attention due to its non-toxicity, cost
eectiveness, ease of handling and good selectivity and activity for a
wide variety of reactions. Our group firstly reported a green procedure for
O-heterocyclization of 1,5-cyclooctadiene (COD) by catalytic oxidation
with aqueous H2O2.19 The eect of the amount of the tungstic acid was
well studied as well as the molar ratio of COD and H2O2 and the volume
ratio of the solvent to reactant. The green catalytic oxidation process
had a lot of advantages compared with the other traditional chemical
methods and the tungstic acid catalyst could also be easily recovered.
Meanwhile, our group also developed a new economic and green route to
synthesize phthalide from 1,2-benzenedimethanol using aqueous
hydrogen peroxide as the oxidant and tungstic acid as the catalyst under
organic solvent-free conditions.20 The desired product with high purity
and good yield was obtained. The tungstic acid catalyst could be easily
filtrated after reaction and reused for more than 6 times. The above two
green catalytic oxidation processes are firstly reported and can open up a
new prospect for the application of tungsten-based catalysts.
2.3 Tungsten carbide catalyst for green catalytic oxidation
Tungsten carbide (WC), as a green catalyst for electron-transfer reactions,
has been extensively investigated in various applications, such as
hydrogenolysis and isomerization reactions, fuel cells, hydrogen evolution and catalytic oxidation, because it exhibits catalytic properties
similar to those of noble metals as well as the chemical and thermal
stability. WC has been reported to possess three dierent crystalline
phases (b-W2C, a-WC and b-WC1x) depending on the reaction conditions and the synthetic route. New methods of thermo-programmed
4 | Catalysis, 2016, 28, 127
17:23:19.
View Online
elaboration of carbide synthesis have led to significant improvement in

the catalytic oxidation application, due to WC with relatively high specific
surface areas. Meanwhile, tungsten carbide also has a well known
catalytic action on the process of oxidation of hydrogen in acidic media
where it has a particular chemical stability and it is more resistant to
poisoning agents than noble metals. Although the catalytic eciency of
WC is much lower than that of noble metals, a considerable interest has
been taken in the field of fuel cells for the development of inexpensive
systems. More recently, WC has been taken into consideration as a
catalyst that could convert H2 and O2 to water without the drawbacks of
noble metals which can contaminate the active materials and enhance
discharge and sulphation. Many investigations have been carried out to
obtain highly active tungsten carbide and determine its electrochemical
behavior. Penazzi et al. studied electrochemical oxidation of hydrogen
using WC as catalyst.21 The final results suggested that the rate of
chemical oxidation of hydrogen was proportional to PH2 only. Meanwhile,
the eect of WC on electrochemical oxidation of hydrogen increased
with the applied potential and was proportional to the specific BET
area of the catalyst. Wang et al. demonstrated for the first time that the
intrinsic catalytic activity of WC nanorods (WC NRs) towards typical
peroxidase substrate, such as 3,3 0 ,5,5 0 -tetramethyl-benzidine (TMB)
and o-phenylenediamine (OPD) in the presence of hydrogen peroxide
(H2O2).22 Compared to natural enzyme HRP, WC NRs exhibited superior
catalytic activity and good reutilization. The schematic illustration of
oxidation reaction of TMB and OPD for the intrinsic peroxidase-like
activity of WC NRs was further studied, as demonstrated in Fig. 2.
Although homogeneous tungsten-based catalysts exhibit superior
catalytic performance in the field of traditional catalytic oxidation,
photocatalytic oxidation and electrochemical oxidation, they still restrict
further applications in industry due to the diculty in separation
Fig. 2 The schematic illustration of oxidation reaction of TMB and OPD for the intrinsic
peroxidase-like activity of WC NRs. Reprinted from N. Li et al., Novel tungsten carbide
nanorods: an intrinsic peroxidase mimetic with high activity and stability in aqueous and
organic solvents, Biosens. Bioelectron., 54, 521527. Copyright (2014), with permission
from Elsevier.22
Catalysis, 2016, 28, 127 | 5
View Online
and reuse. Hence, it is urgent to design new kinds of tungsten-based

heterogeneous catalysts.
17:23:19.
3 W-based catalyst supported on dierent carriers for

green catalytic oxidation
Tungsten-based materials, as an outstanding catalyst, have been applied
to many green catalytic oxidation reactions, such as conventional green
oxidation including selective oxidation of unsaturated compounds, selective oxidation of saturated hydrocarbons, etc.; photocatalytic oxidation
for degradation of organic pollutants and unsaturated compounds, etc.;
electrocatalytic oxidation, for instance, methanol electrochemical
oxidation, etc. However, the homogeneous catalytic system restricts its
further application in industry for the dicult separation and reuse of
the homogeneous tungsten-based catalyst. One of the most promising
ways to accomplish this aim is to design the W-containing heterogeneous
catalysts. Hence, many heterogeneous tungsten-based catalysts supported on dierent carriers (such as metal oxide, siliceous mesoporous
molecular sieves and carbon materials) have been designed for green
catalytic oxidation that show good performance in the target reaction.
3.1 W-based catalyst supported on metal oxides
The metal oxide, dependent on its simple preparation, strong surface
acidic sites, high thermal stability and good catalytic activities, has been
considered as a kind of useful material in the catalytic oxidation field.23
Tin dioxide (SnO2), by virtue of its unique chemical and mechanical
stabilities as well as two specific characteristics: variation in valence state
and oxygen vacancies defects,24 has been widely used as one of the
most important smart and functional materials for technological and
industrial applications, such as transparent conductive electrodes, anode
materials for lithium-ion batteries, solar energy conversion, electrochemical devices, antistatic coatings and catalysis.2531 So far, there have
been several studies on the WO3/SnO2 composites applied to the green
catalytic oxidation reactions. Our group reported the eect of calcination
temperature of the support and the catalyst of WO3/SnO2 on the catalytic
oxidation of 1,2-benzenedimethanol by H2O2 as shown in Scheme 1.32 A
series of WO3/SnO2 composite catalysts for the catalytic oxidation were
prepared by co-precipitation-impregnation method and characterized
with various techniques. The characterization results showed that the
calcination temperature of the catalyst and the support was essential to
the structural evolution of the WO3/SnO2. The studies of the catalytic
behavior of WO3/SnO2 catalyst in the selective oxidation of 1,2-benzenedimethanol to phthalide using aqueous H2O2 as the oxidant suggested
that the optimized calcination temperature of the support and the
catalyst was 1023 and 823 K, respectively, which was mainly attributed to
the high dispersion of tungsten species and few W61 ions entered
into SnO2 lattice. The latter would decrease inevitably the amount of
surface active tungsten species. Meanwhile, Kamata et al. demonstrated a
6 | Catalysis, 2016, 28, 127
17:23:19.
View Online
Scheme 1 The selective oxidation of 1,2-benzenedimethanol by H2O2. Reprinted from

Z. Y. Zhang et al., Eect of calcination temperature of the support and the catalyst of
WO3/SnO2 on the catalytic oxidation of 1,2-benzenedimethanol by H2O2, Appl. Catal. A:
Gen., 482, 171178. Copyright (2014), with permission from Elsevier.32
heterogeneous catalyst with tungsten and zinc oxides supported on SnO2

(WZn/SnO2), which was proved ecient for the selective oxidation of
various alkenes, amines, silanes, and sulfides with aqueous H2O2 as the
green oxidant.33 It was notable that the catalyst could be recovered from
the reaction mixture and reused for several times without any obvious
loss of catalytic performance.
Titania (TiO2), a widely used catalyst support, is known to enhance the
activity in many cases due to the strong interaction between the active
phase and the support.34 Titania has three crystalline phases: rutile,
anatase, and brookite. Rutile is the thermodynamically stable state,
whereas the other two phases are metastable.35 Due to the fact that the
crystalline state and structure of the support strongly aect the catalytic
activity and selectivity, the design and selection of novel, highly active
catalysts place many requirements on the supports used. Our group
reported that the as-prepared WO3/TiO2SiO2 sample was employed as
the heterogeneous catalyst in the liquid-phase cyclopentene (CPE) oxidation by aqueous H2O2, which exhibited higher selectivity (about 75%)
to glutaraldehyde (GA) and, in turn, higher GA yield than the WO3/SiO2
heterogeneous catalyst and even the homogeneous tungstic acid catalyst
under the same reaction conditions.36 The amorphous WO3 phase was
identified as the active sites and the loss of the active sites was proved to
be unimportant. Meanwhile, the eects of various factors on the catalytic
behaviors, such as the WO3 loading, the calcination temperature, the
surface acidity and the reaction media, were also investigated and
discussed. Additionally, our group found that a novel thermally stable
coreshell structured WO3/TiO2 catalyst shown in Fig. 3 was applied to
the catalytic oxidation of CPE to GA by aqueous H2O2.37 The results
revealed that WO3 species were highly dispersed in the interspace of
titania spheres to form the novel WO3/TiO2 catalyst, and there was strong
interaction between WO3 and the TiO2 support. The high dispersion
of the tungsten species achieved on the coreshell structured titania
Catalysis, 2016, 28, 127 | 7
17:23:19.
View Online
Fig. 3 SEM and TEM images of the mesoporous titania microspheres (a and b) and the
20 wt% WO3/TiO2 catalyst (c and d). Reprinted from X.-L. Yang et al., Synthesis of novel
coreshell structured WO3/TiO2 spheroids and its application in the catalytic oxidation of
cyclopentene to glutaraldehyde by aqueous H2O2, J. Catal., 438450. Copyright (2005),
with permission from Elsevier.37
spheres, the strong interaction, and the medium-strong acidity of the

catalyst were all considered to contribute to the superior catalytic
behavior of the WO3/TiO2 catalyst in the selective oxidation of CPE to GA.
The optimal tungsten content was 20 wt%, and the GA yield over this
catalyst reaches 69%.
In addition, WO3/TiO2 is also a popular photocatalyst for degradation
of organic pollutants that has already been well studied by several
research groups. Zhang et al. developed a hierarchically nanostructured
WO3/TiO2 photocatalyst via the subsequent hydrothermal treatment of
electrospun TiO2 nanofibers in the presence of tungstic acid.38 As shown
in Fig. 4, the WO3 seed layer on the TiO2 nanofibers provided growth
sites, facilitating the nucleation and growth in the solution, thus
TiO2 nanofibers and WO3 nanorods were closely combined together. The
WO3/TiO2 demonstrated enhanced visible light absorption and increased
photocatalytic degradation of organic pollutants, due to the migration of
electronhole pairs between TiO2 and WO3, thus increasing the lifetime
of the charge carriers, large surface areas and light utilization and
the unobstructed migration of electrons to the surface. The energy band
structure and electronhole pair separation in the hierarchically
structured TiO2/WO3 nanofibers were described in Fig. 5. Wang et al.
investigated a low temperature peptization process to construct a
8 | Catalysis, 2016, 28, 127
17:23:19.
View Online
Fig. 4 Schematic illustration showing the formation of the hierarchically structured TiO2/
WO3 nanofibers. Reprinted from ref. 38 with permission from The Royal Society of
Chemistry.
Fig. 5 Schematic diagram showing the energy band structure and electronhole pair
separation in the hierarchically structured TiO2/WO3 nanofibers. Reprinted from ref. 38,
with permission from The Royal Society of Chemistry.
uniform composite structure containing amorphous WO3 and anatase

TiO2.39 Visible-light photocatalytic activity of the as-prepared WO3/TiO2
composite was induced by UV pre-irradiation. The whole consecutive
procedure could be defined as a consecutive photocatalytic process
as shown in Fig. 6, which was beneficial to the ecient use of solar
energy.
Zirconia (ZrO2)-supported tungsten oxide catalysts have been extensively studied in recent years because of their ability to catalyze a wide
range of reactions such as selective reduction of NOx to N2, oligomerization of oC20 alkanes to gasoline, diesel and lubricants (C301), and
isomerization of alkanes.40,41 However, the oxidation properties of the
zirconia-supported tungsten oxides were rarely reported. Our group
firstly reported that a series of tungsten oxide supported on commercial
ZrO2 that was synthesized via a traditional impregnation method were
Catalysis, 2016, 28, 127 | 9
17:23:19.
View Online
Fig. 6 Schematic illustration of the consecutive photocatalytic process. Reprinted from

ref. 39 with permission from The Royal Society of Chemistry.
Fig. 7 Schematic illustration of the selective oxidation of 1,2-benzenedimethanol.

Reprinted from Q. J. Zhu et al., Eect of tungsten precursor on the high activity of the
WO3/ZrO2 catalyst in the oxidative lactonization of 1,2-benzenedimethanol, Appl. Catal. A:
Gen., 435, 141147. Copyright (2012), with permission from Elsevier.42
applied to the oxidative lactonization of 1,2-benzenedimethanol to

phthalide with H2O2, as shown in Fig. 7.42 The tungsten precursor and
the calcination temperature were crucial to the dispersion and the nature
of the tungsten species on ZrO2 and the excellent catalytic performance
of the catalyst prepared by calcination at 823 K after using tungstic
acidoxalic acid complex as the tungsten precursor was attributed to
the presence of polymeric WO6 units. Galano et al. experimentally
and theoretically explored the dependence between catalytic activity of
WOxZrO2 system in the oxidation of dibenzothiophene (DBT) and its
relationship with local acidity.43 The structural requirements indicated
that the oxidative eciency (per W-atom) increased as the WOx surface
density become larger, up to 7 W nm2. These results strongly suggested
that n-meric domains and/or WO3x nanoparticles (NPs) anchored on the
surface were more reactive than monomeric and three-dimensional
structures and WOx domains of intermediate size provided a better
compromise between surface acidity and catalytic eciency. The
10 | Catalysis, 2016, 28, 127
17:23:19.
View Online
theoretical analysis revealed that the combined presence of Lewis and

Bronsted sites energetically favored the formation of peroxometallate
complexes and that OOH addition reached its maximum for the minimum (non-zero) number of Lewis sites and maximum number of
Bronsted sites.
Aluminum oxides, especially g-Al2O3, have been widely employed as
supports for catalytic oxidation reactions because of their high surface
area, thermal and chemical stability and low cost. As previous literature
reported, aluminum oxides (Al2O3)-supported tungsten based catalysts
have been widely applied to the green catalytic oxidation process. Lu et al.
developed a novel oxidation method of alcohols to the corresponding
aldehydes and ketones with high eciency under molecular oxygen
in the presence of the 1,1-diphenyl-2-picrylhydrazyl (DPPH) catalyst
and the tungsten oxide/alumina (WO3/Al2O3) co-catalyst.44 Various
aromatic, alicyclic, and aliphatic alcohols could be converted to their
corresponding carbonyl compounds in excellent yields. Suzuki et al.
reported an ecient and selective aerobic oxidative transformation of
primary amines to oximes proceeded with high eciency under mild
conditions in the presence of the DPPH catalyst and WO3/Al2O3 cocatalyst.45 Various alicyclic and aliphatic amines could be converted to
their corresponding oximes in excellent yields. In the DPPHWO3/Al2O3
system, DPPH acted as an electron transfer mediator, and an alkylhydroperoxide intermediate was transformed into an oxime by the WO3/
Al2O3 co-catalyst. Meanwhile, the reaction could be rationalized by assuming the mechanism, as depicted in Scheme 2.
Vanadium oxide (V2O5) is an attractive material due to its good catalytic, electrical and optical properties. The good catalytic activity is the
result of easy reduction and oxidation between the multiple oxidation
states of vanadium in V2O5.46 Meanwhile, mesoporous vanadium oxide
solid, by virtue of thermal and mechanical stability, plays a key role as
support in the catalytic oxidation.47 Makgwane et al. studied an ecient
room temperature oxidation process of cyclohexane to cyclohexanone (K)
and cyclohexanol (A) by highly active nanostructured WO3/V2O5 composite as the catalyst.48 The catalyst exhibited high catalytic activity with
up to 90% conversion and excellent recyclability and stability. The catalytic results appeared to suggest a possible existence of a strong interaction eect between the two combined metal oxides that enhanced their
performance when compared to individual oxides. The enhanced performance of WO3/V2O5 was partly attributed to the redox and possible
structural modifications due to the strong metal-support interaction. The
possible mechanism route of cyclohexane oxidation over WO3/V2O5
catalyst was shown in Scheme 3.
The metal oxides as the support of W-based catalyst have performed
satisfactorily in green catalytic oxidation processes. Nevertheless, there
are still many problems in the practical applications of these reported
catalysts, including the needs of expensive raw materials of the metal
oxides and the poor recycling eciency of the catalyst. In order to work
out the drawbacks mentioned above, siliceous mesoporous supports
have been widely investigated.
Catalysis, 2016, 28, 127 | 11
17:23:19.
View Online
Scheme 2 Proposed mechanism for DPPHWO3/Al2O3 catalyzed aerobic oxidation of

primary amines. Reprinted with permission from K. Suzuki, T. Watanabe and S. I.
Murahashi, J. Org. Chem., 2013, 78, 2301. Copyrights (2013), American Chemical
Society.45
3.2 W-based catalyst supported on siliceous mesoporous molecular

sieves
Silicon materials with unique physicochemical properties have received
tremendous attention in many fields, such as solar energy conversion,
optoelectronics, biological sensors and supports.49 Porous silicon, with
the intrinsic properties of silicon and the unique features of a porous
structure, has been considered as one of the most attractive materials for
use in catalysis,50 luminescence,51 supercapacitors52 and batteries.53
Until now, many eorts have been made in the preparation of porous
silicon. A variety of mesoporous silicon materials have been extensively
studied, such as mesoporous MCM-41, MCM-48, HMS, MCF, SBA-15
and so on.
Since the first invention of MCM-41 in 1992 by the Mobile researchers,
MCM-41 material has attracted much interest because of their potential
applications in many fields of science and engineering, such as
adsorption, separation, and catalysis involving bulky molecules.54 In
particular, its remarkable textural properties, such as high surface area
12 | Catalysis, 2016, 28, 127
17:23:19.
View Online
Scheme 3 Plausible mechanism route to cyclohexane liquid-phase oxidation products

formation. Reprinted from P. R. Makgwane et al., Ecient room temperature oxidation of
cyclohexane over highly active hetero-mixed WO3/V2O5 oxide catalyst, Catal. Commun.,
54, 118123. Copyrights (2014), with permission from Elsevier.48
and large pore volume, make it very suitable for application as catalyst
supports. Li et al. utilized the hydrolysis of tetraethylortho silicate and
ammonium tungstate in the presence of cetylpyridinium bromide as
template in strongly acidic medium to synthesize a novel tungstencontaining MCM-41 mesoporous molecular sieve.55 The tungstencontaining MCM-41 mesoporous molecular sieve was applied to the
cyclohexene conversion, which strongly suggested that the isolated W
sites incorporated in the framework of W-MCM-41 aorded the mesostructure with high catalytic activity with respect to the crystalline WO3 in
promoting the hydroxylation of cyclohexene using dilute H2O2 as
terminal oxidant. Our group reported the W-MCM-41 catalyst for the
selective oxidation of CPE to GA with aqueous hydrogen peroxide.56 It was
found that tungsten species stably existed in the silica-based matrix of
MCM-41 up to a Si/W molar ratio of 40. The W-MCM-41 catalyst exhibited
the highest activity and selectivity in the selective oxidation of CPE to GA.
The convenient separation of the W-MCM-41 catalyst from the reaction
products mixture and its longer lifetime made it more feasible than the
corresponding homogenous catalysts when applied in industrial use.
MCM-48 with a 3D cubic mesostructure consists of two interpenetrating continuous networks of chiral channels. These enantiomeric
pairs of porous channels are separated by an inorganic wall that followed
exactly the gyroid (G-surface) infinite periodic minimal surface (IPMS).
This unique 3D channel network is thought to provide a highly opened
porous host that provides easy and direct access for guest species, thus
facilitating inclusion or diusion throughout the pore channels without
pore blockage.5759 Hence, the structural peculiarities of the MCM-48
make it a potentially very interesting and promising catalyst/catalyst
Catalysis, 2016, 28, 127 | 13
17:23:19.
View Online
support, adsorbent, sensor and an excellent inorganic template for the

synthesis of nanostructures.60 Zheng et al. reported a series of MCM-48
supported 12-tungstophosporic acid mesoporous materials that were
synthesized via a wet impregnation method. The characterization results
indicated that the mesoporous phase of MCM-48 supports remained
almost unchanged upon the HPW loading, while the long-range order
decreased noticeably. Furthermore, HPW/MCM-48 was proved to be an
ecient catalyst for the green synthesis of benzoic acid with aqueous
hydrogen peroxide as oxidant and the 35 wt% loading of HPW proved to
be optimal under the system.61 Meanwhile, our group firstly developed a
novel W-containing MCM-48 catalyst under hydrothermal conditions via
pH adjustment using tetraethoxysilane (TEOS) as Si source, Na2WO4 as W
source and cetyltrimethylammonium bromide (CTAB) as the structuredirecting template.62 The as-synthesized material showed the typical
structure of MCM-48. WO3 species were highly dispersed into the lattice
of the bulk and might be imbedded separately, which could be served as
the active centers for the selective oxidation of CPE to GA. The FT-IRpyridine adsorption confirmed the presence of strong Brnsted acid sites
and Lewis acid sites upon incorporating tungsten oxide species into the
MCM-48 materials, which were beneficial to the catalytic performance.
The optimal tungsten content was 20 wt% and the GA yield over this
catalyst exceeded 66%, higher than those over commercial and xerogel
silica supported WO3 catalyst or WO3/TiO2-SiO2 catalyst, suggesting its
promising potential use in industry.
Siliceous mesocellular foams (MCFs) with well-defined ultra large
mesopores and hydrothermally robust frameworks were first described
through an oil-in-water microemulsion method by Winkel et al.63,64 MCF
materials have gained many benefits from such a facilitated synthesis
method such as well defined pore and wall structure, thick walls, and
highly hydrothermal stability, which give MCFs unique advantages as
catalyst supports. Cao et al. developed high-quality mesoporous W-MCF
materials featuring a well-defined three dimensional (3D) mesoporosity
and ultra large mesopores with dierent Si-to-W ratios via a direct
hydrothermal method.65 A high tungsten content up to 20 wt.% could be
well incorporated into the framework of the MCF material. The W-doped
MCF materials appeared to be suitable as catalysts in the selective
oxidation of 1,3-butanediol to 4-hydroxy-2-butanone in the hydrogen
peroxide system, as shown in Scheme 4. The catalyst exhibited a high
activity and an extremely high stability as a function of the test and reuse
due to the presence of three dimensional mesocellular networks with
ultralarge mesopores which favors the diusion of reactants and
products. Our group also studied WO3-containing mesocellular silica
foam catalysts.66 These catalysts were synthesized via an in situ method
by using TEOS and sodium tungstate as precursors via a traditional
impregnation method. Both catalysts exhibited excellent performance for
the target reaction (O-heterocyclization). The ultra large mesopores of the
catalysts were helpful for the transport of the large raw material and
products during the reaction. Meanwhile, the recycling experiment
results indicated that the in situ method-derived catalyst showed much
14 | Catalysis, 2016, 28, 127
17:23:19.
View Online
Scheme 4 The oxidation process of 1,3-butanediol by H2O2. Reprinted from Y. Su et al.,

Tungsten-containing MCF silica as active and recyclable catalysts for liquid-phase
oxidation of 1,3-butanediol to 4-hydroxyl-2-butanone, Appl. Catal. A: Gen., 315,
91100. Copyrights (2006), with permission from Elsevier.65
Scheme 5 Incorporating catalytic oxotungsten tetrahedra centers into the framework of

MCF. Reprinted from R. H. Gao et al., High-activity, single-site mesoporous WO3-MCF
materials for the catalytic epoxidation of cycloocta-1,5-diene with aqueous hydrogen
peroxide, J. Catal., 256, 259267. Copyrights (2008), with permission from Elsevier.67
better stability than the impregnation one, suggesting that the tungsten
oxide on the catalyst synthesized by the impregnation method was more
easily aggregated after the reaction than the one on the catalyst synthesized by the in situ method. Whats more, our group presented that
the ammonium acetate-treated WO3-MCF (AMA-treated WO3-MCF) catalysts exhibited good performance and retained the special structure of
the supports under special pre-treatment conditions.67 The recycling
experiment demonstrated the excellent stability of the AMA-treated
WO3-MCF catalyst. Clearly, the tungsten clusters on the surface were
single-site {WO4} species bound strongly to silica through WOSi
covalent bonds, as illustrated in Scheme 5.
Hexagonal mesoporous silica (HMS), commonly synthesized by the
assembly pathway of hydrogen-bonding interactions between neutral
primary alkylamine and neutral inorganic precursors at room temperature,68 has high surface area and large and uniform pore size.
Catalysis, 2016, 28, 127 | 15
17:23:19.
View Online
These features extend its applications in the field of catalysis, molecular

sieving, and supports. Moreover, HMS possesses a much thicker framework wall, smaller domain size with short channels, and larger textual
mesoporosity.6971 These properties are distinguishable from those of
MCM-41, and provide better transport channels for reactants to access
the active centers and better diusion channels for products to move out
than their MCM-41 analogs. Our group studied the tungsten-containing
HMS via dodecylamine as template at room temperature.72 The asprepared material was very active as a catalyst for the selective oxidation
of CPE to GA with environmentally benign hydrogen peroxide as the
oxidant. Complete conversion of CPE and very high yield of GA (76%)
were obtained over the W-HMS catalyst with a Si/W molar ratio at 30.
Furthermore, almost no tungsten species were leached into the reaction
solution, enabling the catalyst to be employed for many reaction
cycles without dramatic deactivation. Meanwhile, our group reported a
novel and green route for the selective oxidation of CPE oxide to GA by
using aqueous H2O2 as the oxidant and WS2@hexagonal mesoporous
silica (WS2@HMS) material as the catalyst, which displayed a very large
surface area, high eciency, excellent selectivity and outstanding
reusability.73
SBA-15, a type of ordered mesoporous material achieved by using a
triblock copolymer as template under strongly acidic conditions, is a
promising candidate in catalysis since it possesses a high surface area
(6001000 m2 g1) and uniform tubular channels with tunable pore
diameters in the range of 530 nm.74 Our group firstly demonstrated the
use of W-doped SBA-15 catalyst prepared by a novel in situ synthesis
method as highly ecient catalyst for the direct production of GA via
selective oxidation of CPE by using non-aqueous hydrogen peroxide as
the green oxidant.75 Obviously, the morphology and structure still remained unchanged after WOx species doping. The heterogeneous
W-doped SBA-15 catalyst exhibited an excellent activity and selectivity for
the selective transformation of CPE to GA and the presence of a high
surface concentration of WOx species dispersed on well ordered
hexagonal pore walls of SBA-15 support was essential to the superior
performance of the catalyst for the selective oxidation of CPE. Jia et al.
developed mesoporous SBA-15 materials modified with oxodiperoxo
tungsten complexes via a post-grafting route as ecient catalysts for the
epoxidation of olefins with hydrogen peroxide.76 The preparation of
hybrid mesoporous SBA-15 materials was presented in Scheme 6.
Compared with the catalytic properties of hybrid SBA-15 materials
containing dierent ligands, the catalyst bearing pyrazolylpyridine ligand
exhibited relatively high recoverability, stability and very high eciency
of H2O2 utilization under optimized conditions. The catalytic activity
could be further improved by using solvent mixtures of CH3CN and
CH3COOH at a temperature as low as 35 1C. Furthermore, reasons for
improving catalytic performance of the hybrid material were discussed
and this phenomenon could be attributed to a suitable coordination
interaction between the chelate ligand and the WO(O2)2 fragment and the
structure and surface properties of SBA-15 support.
16 | Catalysis, 2016, 28, 127
17:23:19.
Published on 19 February 2016 on http://pubs.rsc.org | d
Catalysis, 2016, 28, 127 | 17
Scheme 6 Schematic representation for the preparation of hybrid mesoporous SBA-15 materials. Reprinted from J. Y. Tang et al., Mesoporous SBA-15 materials
modified with oxodiperoxo tungsten complexes as ecient catalyss for the epoxidation of olefins with hydrogen peroxide, J. Mol. Catal. A: Chem., 313, 3137.
Copyrights (2009), with permission from Elsevier.76
17:23:19.
View Online
In this section, several heterogeneous silica-supported WO3 catalysts

have been designed that show good performance to the target green
oxidation reaction. However, due to the complexity of the preparation
procedures, the need for expensive raw materials and the diculties of
large-scale production, the synthesis of new support materials is still
under development. During the past few years, novel carriers such as
sustainable carbon materials (including polymeric graphitic carbon
nitride, graphene, ordered mesoporous carbons) have attracted considerable attention. So far, there are no reports on the application of
silica-supported WO3 catalysts to the photocatalytic or electrochemical
applications.
3.3 W-based catalyst supported on sustainable carbon materials

Carbon material remains one of the most attractive and well-studied
material systems in the scientific community due to its amazing variety
and versatility in combination with low cost, availability, and wide ranging properties. The physical chemical, optical and electronic properties
of carbon materials vary according to its allotropic form and also greatly
depend on its structure, morphology and surface composition. High
surface area carbon materials have been extensively used for sorption,
sensing, photovoltaic, catalysis and storage applications.7781
Polymeric graphitic carbon nitride (g-C3N4), an appealing and potential
material, has recently received considerable attention. It has been widely
used in photocatalysis, fuel cells and gas storage, by virtue of its various
physicochemical properties, such as excellent thermal and chemical
stability, nontoxic, electrical conductivity, energy storage and gas
adsorption. Simultaneously, g-C3N4 is a p-conjugated polymer semiconductor with a layered structure formed by tri-s-triazine construction
unit and an optical band gap of 2.7 eV, which has shown some photocatalytic activity under visible light. Nevertheless, its specific surface area
is small, and the recombination of photo-generated electron-hole pairs is
high, which restrict the further improvement of the catalytic activity.82
WO3 is also a semiconductor material with an optical band gap of 2.7 eV,
and it exhibits photocatalytic activity under visible light.83 Since the
optical band gap of WO3 is almost the same as that of g-C3N4, WO3 and
g-C3N4 are simultaneously excited. Therefore, it is expected that the
composite structure may improve the photocatalytic activity of g-C3N4
by increasing the number of photogenerated electronhole pairs.84
WO3/g-C3N4 composites have been attracting much attention in recent
years, especially in the field of photocatalysis. Li et al. reported a novel
WO3/g-C3N4 composite photocatalyst via a calcination method.85 The
highest MB (methylene blue) and 4-CP (p-chlorophenol) degradation
eciency for the WO3/g-C3N4 (9.7%) composite respectively reached up
to 97% and 43% under visible light irradiation, while pure g-C3N4 only
induced 81% degradation of MB within 3 h and 3% degradation of 4-CP
within 6 h. A possible mechanism of the visible light activity of a
WO3/g-C3N4 catalyst was proposed and illustrated in Fig. 8. Katsumata
et al. prepared g-C3N4/WO3 composites by a physical mixing method and
18 | Catalysis, 2016, 28, 127
17:23:19.
View Online
Fig. 8 Proposed mechanism for the photodegradation of MB on WO3/g-C3N4 composites. Reprinted with permission from ref. 85, with permission from The Royal Society of
Chemistry.
the photocatalytic degradation of CH3CHO using the composite was

investigated as well. Under visible light irradiation, the CH3CHO photodegradation by the g-C3N4/WO3 composite was greater than that for
g-C3N4 alone. With increasing WO3 content of the composites, the
photodegradation activity increases, indicating that WO3 is strongly
associated with the photodegradation of CH3CHO. The optimum mixture
of g-C3N4 and WO3 for photocatalysis corresponded to the composition
of G2W8 (g-C3N4 : WO3 2 mol : 8 mol). A CH3CHO photodegradation
mechanism was proposed and discussed in terms of energy band positions in Scheme 7.86
Our group first successfully applied the WO3/g-C3N4 nanocomposites
in the selective oxidation of cyclohexene oxide to adipaldehyde using
aqueous H2O2 as the oxidant. The composites were prepared by a green
and facile ultrasonic-assisted method and exhibited high activity and
excellent recycling capacity. In addition, the formation of WO3 nanorods
with the help of g-C3N4 nanosheets was evidenced and the possible
catalytic mechanism was investigated, as described in Fig. 9.87
Graphene, as a novel single-layer and two-dimensional material with
a large theoretical specific surface area, high chemical and thermal
stability and high intrinsic electron mobility, is a very promising candidate for electrochemistry, solar cells, nanoelectronics, catalysis, etc. To
date, many graphene-based composites containing tungsten trioxide
have been reported for the improved eciency in photocatalysis.
Yu et al. presented a systematic investigation on the incorporation of
WO3 nanorods and graphene for high-eciency visible-light-driven
photocatalysis and NO2 gas sensing.88 This novel composite showed
remarkably enhanced photocatalytic performance compared to pure
Catalysis, 2016, 28, 127 | 19
17:23:19.
View Online
Scheme 7 Possible degradation mechanism of CH3CHO over a g-C3N4/WO3 composite

photo catalyst under visible light irradiation. Reprinted from K. Katsumata, Preparation of
graphitic carbon nitride (g-C3N4)/WO3 composties and enhanced visible-light-driven
photodegeneration of acetaldehyde gas, J. Hazard. Mater., 260, 475482. Copyrights
(2013), with permission from Elsevier.86
Fig. 9 Possible reaction mechanism of cycloalkene oxides over WO3/g-C3N4 composites. Reprinted from J. Ding, Carbon nitride nanosheets decorated with WO3 nanorods:
Ultrasonic-assisted facile synthesis and catalytic application in the green menufacture of
dialdehydes, Appl. Catal. B: Environ., 165, 511518. Copyrights (2015), with permission
from Elsevier.87
20 | Catalysis, 2016, 28, 127
17:23:19.
View Online
Scheme 8 Proposed photodegradation mechanism of RhB over WO3 nanorods/

graphene composites. Reprinted from ref. 88, with permission from The Royal Society
of Chemistry.
WO3 nanorods. The high photocatalytic activity of the WO3/graphene

nanocomposite was found to be related to the increased adsorption
toward chemical species, enhanced light absorption and ecient
charge separation and transfer. Meanwhile, the possible mechanism of
the significantly enhanced photocatalytic activity, superior sensing sensitivity and selectivity was thoroughly studied, as described in Scheme 8,
indicating the potential applications of graphene-based nanomaterials in
the high performance visible-light-driven photocatalysts and gas sensors.
Luo et al. synthesized WO3/graphene (WO3/GN) via an easy and ecient
electrochemical method with a one-step treatment and the assistance of
twelve tungsten phosphate. The WO3 nanoparticles were distributed on
the surfaces of graphene sheets and facilitated the electron transfer and
light absorption. The UV light photocatalytic performance had been
improved up to 2 and 2.2 times respectively compared with GN and WO3
nanoparticles.89 WO3/RGO materials were also synthesized via the one
step hydrothermal method. The incorporation of RGO sheets with WO3
could facilitate the electron transfer and light absorption and electrical
conductivity of photocatalyst.90 The WO3-RGO materials exhibited better
photocatalytic oxidation activity of methylene blue (MB) and water
splitting performance than that of pure WO3 nanoparticle.
Due to its unique atom-thick 2D structure and remarkably physicochemical properties, graphene-based materials have been making a
profound impact in photoelectrochemical applications, especially WO3/
graphene system. Huang et al. developed a kind of uniform reduced
graphene oxide-wrapped WO3 nanowire nanocomposite.91 The composites possess increased surface area with more reaction sites for
interaction, thus resulting in a significantly improved current value of
oxidation peaks for honokiol. Zheng et al. demonstrated a novel 2D2D
Catalysis, 2016, 28, 127 | 21
17:23:19.
View Online
WO3RGO (WO3-reduced graphene oxide) nanocomposite by the direct

hydrothermal growth of WO3 nanoplates on FTO substrates and
subsequent in situ photo reduction to deposit RGO layers on the WO3
nanoplate surface.92 Photo anodes made of the WO3RGO nanocomposites showed an enhanced photocurrent of 2.0 mA cm2 at a bias of
1.23 V vs. RHE, which was higher than that of the pristine WO3 nanoplates and among the best values reported for the hydrothermally
synthesized WO3-based photoanodes. This improved photocurrent was
attributed to both the reduced charge recombination in ultrathin WO3
nanoplates and enhanced charge transfer at the electrode/electrolyte
interface. Devadoss et al. presented the fabrication of grapheneWO3
membranes as a potential alternative photoanode in photoelectronchemical glucose sensing applications and the possibility of fortifying its
photocatalytic activity using plasmonic gold nanoparticles.93 The photoelectron chemical tests illustrated that the performance of AuNPs
supported on grapheneWO3 membrane was superior to other systems
without backbone conducting channel. The mechanism of glucose oxidation at the grapheneWO3Au hybrid membrane modified with glucose oxidase (GOD) enzyme was depicted in Fig. 10.
Ordered mesoporous carbons (OMCs), by virtue of their appealing
structural characteristics, such as periodic and uniform mesopores and
Fig. 10 Schematic representation of (a) grapheneWO3Au triplet junction for glucose

sensing, (b) energy levels at grapheneWO3Au photoelectrode under light illumination
and glucose oxidation mechanism. Reprinted from A. Devadoss, P. Sudhagar, S. Das, S. Y.
Lee, C. Terashima, K. Nakata, A. Fujishima, W. Choi, Y. S. Kang and U. Paik, ACS Appl. Mater.
Interfaces, 2014, 6, 4864. Copyright (2014), with permission from American Chemical
Society.93
22 | Catalysis, 2016, 28, 127
17:23:19.
View Online
Fig. 11 Schematic illustration of the formation process of highly ordered mesoporous

C-WO3 films. Reprinted from ref. 94, with permission from The Royal Society of
Chemistry.
large surface area, have been extensively studied in various applications,

such as electrochemistry, photocatalysis, catalysis and so on. Meanwhile,
mesoporous carbons with 3D pore structures can also promote the mass
transport of both reactants and by-products. He et al. developed a simple
self-assembly route to prepare ordered mesoporous C-WO3 nanocomposites with large surface area and excellent corrosion resistance.94 The
above detailed preparation process for OMC/WC was schematically described in Fig. 11. Tungsten oxide grew in the form of rods as a result of
the confinement eect of the porous structure, improving graphitization
degree and hydrophilicity. Meanwhile, the ordered mesoporous CWO3
nanocomposites, as electrocatalyst carriers for PEMFC, displayed superior electrocatalytic activities compared with pure ordered mesoporous
carriers. Wang et al. firstly presented a facile soft-template synthesis of
order 2D hexagonal mesoporous tungsten carbide (OMC/WC) composite
nanomaterials with a surface area of 538 m2 g1.95 The Pt nanoparticles
supported on OMC/WC revealed a better performance than that of the
commercial PtRu@C catalyst for methanol electro-oxidation.
4 Conclusion and outlook

In this contribution, some recent promising examples among tungsten
containing materials, which have already been extensively applied in the
Catalysis, 2016, 28, 127 | 23
17:23:19.
View Online
field of green catalytic oxidation, have been presented. Tungsten-based

catalyst, as a promising material for green catalytic oxidation process,
such as selective oxidation of unsaturated compounds, photocatalytic
oxidation of organic pollutants and electrochemical catalytic oxidation in
the field of proton exchange membrane fuel cell, exhibits excellent
catalytic activity, chemical and thermal stability and ecient separation
ability from the solvent mixture. However, considerable eorts are still
required to solve the following issues.
(i) More ecient, stable and cheap catalysts should be exploited to
fulfill the catalytic oxidation processes. Extensive studies have been
conducted to investigate metal oxides-supported WO3 and found that the
synthesis of these materials required expensive raw materials, similar to
the silica-supported WO3 catalysts. Meanwhile, the complexity of the
preparation procedures and the diculties of large-scale productions of
silica-supported WO3 catalysts have limited their wide applications.
As novel supports, porous carbon materials will be an alternative and
interesting candidate in the applications of catalytic oxidation process,
especially in the conventional green catalytic oxidation processes.
(ii) The application scope limits of some tungsten containing
materials are still far from being reached. For example, few literatures
have been reported on tungsten-based catalyst supported on sustainable
carbon materials in conventional catalytic oxidation including selective
oxidation of unsaturated organic substrates.
(iii) In the presence of hydrogen peroxide oxidant, the leaching of
active tungsten species and deactivation of the W-based catalyst in the
oxidation process are big challenges for industrial application. Till now,
no commercial tungsten-based heterogeneous catalysts are found, thus
research on tungsten-based material with high activity and stability and
its industrial application may be the hot point in this area.
In summary, it is reasonable to believe that new catalysts (besides
carbon materials) and/or green catalytic processes based on tungstenbased material will continue to be explored in the near future.
Acknowledgements
This work was financially supported by the Major State Basic Resource
Development Program Grant No. 2012CB224804) and NNSFC (Project
20973042, 21173052, 21373054).
References
1
2
3
4
5
S. D. Bishopp, J. L. Scott and L. T. Murciano, Green Chem., 2014, 16, 3281.

S. Biella and M. Rossi, Chem. Commun., 2003, 3, 378.
D. I. Enache, J. K. Edwards, P. Landon, E. B. Solsona, A. F. Carley,
A. A. Herzing, M. Watanabe, C. J. Kiely, D. W. Knight and G. J. Hutchings,
Science, 2006, 311, 362.
P. McMorn, G. Roberts and G. J. Hutchings, Catal. Lett., 1999, 63, 193.
R. D. Wilson, D. G. Barton, C. D. Baertsch and E. Iglesia, J. Catal., 2000,
194, 175.
24 | Catalysis, 2016, 28, 127
View Online
6
7
17:23:19.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
B. F. Sels, D. E. De Vos and P. A. Jacobs, Angew. Chem., Int. Ed., 2005, 44, 310.
T. Kabe, W. H. Qian, A. Funato, Y. Okoshi and A. Ishihara, Phys. Chem. Chem.
Phys., 1999, 1, 921.
Y. Rezgui and M. Guemini, Appl. Catal., A, 2005, 282, 45.
C. D. Baertsch, K. T. Komala, Y. H. Chua and E. Iglesia, J. Catal., 2002,
205, 44.
B. Weng, J. Wu, N. Zhang and Y. J. Xu, Langmuir, 2014, 30, 5574.
J. Polleux, A. Gurlo, N. Barsan, U. Weimar, M. Antonietti and
M. Niederberger, Angew. Chem., Int. Ed., 2006, 45, 261.
E. Skliri, I. N. Lykakis and G. S. Armatas, RSC Adv., 2014, 4, 46170.
A. Phuruangrat, D. J. Ham, S. J. Hong, S. Thongtem and J. S. Lee, J. Mater.
Chem., 2010, 20, 1683.
R. Abe, T. Takata, H. Sugihara and K. Domen, Chem. Commun., 2005,
30, 3829.
Q. Sun, F. Xiao, S. Y. Ren, Z. J. Dong, J. Wang and X. T. Su, Ceram. Int., 2014,
40, 11447.
N. Li, W. Y. Lu, K. M. Pei, Y. Y. Yao and W. X. Chen, Appl. Catal., B, 2015,
163, 150.
M. Aslam, I. M. I. Ismail, S. Chandrasekaran and A. Hameed, J. Hazard.
Mater., 2014, 276, 120.
A. Fakhri and S. Behrouz, Sol. Energy., 2015, 112, 163.
R. H. Gao, W. L. Dai, Y. Y. Le, X. L. Yang, Y. Cao, H. X. Li and K. N. Fan, Green
Chem., 2007, 9, 878.
Q. J. Zhu, W. L. Dai and K. N. Fan, Green Chem., 2010, 12, 205.
S. Bodoardo, A. M. Maja, A. N. Penazzi and F. E. G. Henn, Electrochim. Acta,
1997, 42, 2603.
N. Li, Y. Yan, B. Y. Xia, J. Y. Wang and X. Wang, Biosens. Bioelectron., 2014,
54, 521.
P. S. Reddy, P. Sudarsanam, G. Raju and B. M. Reddy, Catal. Commun., 2010,
11, 1224.
T. Toupance, H. E. Hamzaoui, B. Jousseaume, H. Riague, I. Saadeddin,
tz, Chem. Mater., 2006, 18, 6364.
G. Campet and J. Bro
T. Toupance, O. Babot, B. Jousseaume and G. Vilaa, Chem. Mater., 2003,
15, 4691.
R. Das, T. Jana and S. Ray, Sol. Energy. Mater. Sol. Cells, 2005, 86, 207.
H. Liu, D. Long, X. Liu, W. Qiao, L. Zhan and L. Ling, Electrochim. Acta, 2009,
54, 5782.
L. Q. Zhang, X. H. Liu, Y. C. Perng, J. Cho, J. P. Chang, S. X. Mao, Z. Z. Ye and
J. Y. Huang, Micron, 2012, 43, 1127.
S. Jiang, W. Yue, Z. Gao, Y. Ren, H. Ma, X. Zhao, Y. Liu and X. Yang, J. Mater.
Sci., 2013, 48, 3870.
D. Su, H. J. Ahn and G. Wang, Chem. Commun., 2013, 49, 3131.
G. B. Liu, Q. D. Zhang, Y. Z. Han, N. Tsubaki and Y. S. Tan, Green Chem.,
2013, 15, 1501.
Z. Y. Zhang, Q. J. Zhu, J. Ding, X. Liu and W. L. Dai, Appl. Catal., A, 2014,
482, 171.
K. Kamata, K. Yonehara, Y. Sumida, K. Hirata, S. Nojima and N. Mizuno,
Angew. Chem., Int. Ed., 2011, 50, 12062.
A. A. Ismail and D. W. Bahnemann, J. Mater. Chem., 2011, 21, 11686.
M. Gopal, W. J. M. Chan and L. C. DeJonghe, J. Mater. Sci., 1997, 32, 6001.
R. H. Jin, X. Xia, W. L. Dai, J. F. Deng and H. X. Li, Catal. Lett., 1999, 62, 201.
X. L. Yang, W. L. Dai, C. W. Guo, H. Chen, Y. Cao, H. X. Li, H. Y. He and
K. N. Fan, J. Catal., 2005, 234, 438.
Catalysis, 2016, 28, 127 | 25
View Online
38
39
40
17:23:19.
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
L. Zhang, Y. G. Li, Q. H. Zhang and H. Z. Wang, CrystEngComm, 2013,

15, 5986.
D. Su, J. Y. Wang, Y. P. Tang, C. Liu, L. F. Liu and X. J. Han, Chem. Commun.,
2011, 47, 4231.
Q. Zhu, X. Chu, Z. Zhang, W. L. Dai and K. N. Fan, Appl. Catal., A, 2012,
435, 141.
J. G. Santiesteban, J. C. Vartuli, S. Han, R. D. Bastian and C. D. Chang,
J. Catal., 1997, 168, 431.
Q. J. Zhu, X. F. Chu, Z. Y. Zhang, W. L. Dai and K. N. Fan, Appl. Catal., A, 2012,
435436, 141.
G. R. Gattorno, A. Galano and E. T. Garcia, Appl. Catal., B, 2009, 92, 1.
Y. Q. Zhu, J. Xu and M. Lu, Catal. Commun., 2014, 48, 78.
K. Suzuki, T. Watanabe and S. I. Murahashi, J. Org. Chem., 2013, 78, 2301.
Y. Miseki, H. Kusama and K. Sayama, Chem. Lett., 2012, 41, 1489.
E. Skliri, I. N. Lykakis and G. S. Armatas, RSC Adv., 2014, 4, 46170.
P. R. Makgwane and S. S. Ray, Catal. Commun., 2014, 54, 118.
J. F. Wang, K. X. Wang, F. H. Du, X. X. Guo, Y. M. Jiang and J. S. Chen, Chem.
Commun., 2013, 49, 5007.
N. Megouda, Y. Cofininier, S. Szunerits, T. Hadjersi, O. ElKechai and
R. Boukherroub, Chem. Commun., 2011, 47, 991.
L. Vaccaro, A. Morana, V. Radzig and M. Cannas, J. Phys. Chem. C, 2011,
115, 19476.
Q. Liu, M. H. Nayfeh and S. T. Yau, J. Power Sources, 2010, 195, 3956.
H. P. Jia, P. F. Gao, J. Yang, J. L. Wang, Y. Nuli and Z. Yang, Adv. Energy
Mater., 2011, 1, 1036.
M. E. Davis, Nature, 2002, 417, 813.
Z. R. Zhang, J. S. Suo, X. M. Zhang and S. B. Li, Appl. Catal., A, 1999,
179, 11.
H. Chen, W. L. Dai, J. F. Deng and K. N. Fan, Catal. Lett., 2002, 81, 131.
C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli and J. S. Beck, Nature,
1992, 359, 710.
J. S. Beck, J. C. Vartuli, W. J. Roth, M. E. Leonowicz, C. T. Kresge,
K. D. Schmitt, C. T. W. Chu, D. H. Olson, E. W. Sheppard, S. B. McCullen,
J. B. Higgins and J. L. Schlenker, J. Am. Chem. Soc., 1992, 114, 10834.
T. W. Kim, P. W. Chung and V. S. Y. Lin, Chem. Mater., 2010, 22, 5093.
U. S. Taralkar, M. W. Kasture and P. N. Joshi, J. Phys. Chem. Solids, 2008,
69, 2075.
H. Y. Wu, X. L. Zhang, X. Chen, Y. Chen and X. C. Zheng, J. Solid State Chem.,
2014, 211, 51.
X. L. Yang, W. L. Dai, R. H. Gao, H. Chen, H. X. Li, Y. Cao and K. N. Fan,
J. Mol. Catal. A: Chem., 2005, 241, 205.
P. S. Winkel, W. W. Lukens, D. Y. Zhao, P. D. Yang, B. F. Chmelka and
G. D. Stucky, J. Am. Chem. Soc., 1999, 121, 254255.
P. S. Winkel, W. W. Lukens, P. D. Yang, D. I. Margolese, J. S. Lettow, J. Y. Ying
and G. D. Stucky, Chem. Mater., 2000, 12, 686.
Y. Su, Y. M. Liu, L. C. Wang, M. Chen, Y. Cao, W. L. Dai, H. Y. He and
K. N. Fan, Appl. Catal., A, 2006, 315, 91.
B. M. Reddy, M. K. Patil, G. K. Reddy, B. T. Reddy and K. N. Rao, Appl. Catal.,
A, 2007, 332, 138.
R. H. Gao, X. L. Yang, W. L. Dai, Y. Y. Le, H. X. Li and K. N. Fan, J. Catal.,
2008, 256, 259.
P. T. Tanev, M. Chibwe and T. J. Pinnavaia, Nature, 1994, 368, 321.
R. T. Yang, T. J. Pinnavaia, W. Li and W. Zhang, J. Catal., 1997, 172, 488.
26 | Catalysis, 2016, 28, 127
View Online
70
71
72
17:23:19.
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
T. R. Pauly, Y. Liu, T. J. Pinnavaia, S. J. L. Billinge and T. P. Rieker, J. Am.

Chem. Soc., 1999, 121, 8835.
A. Tuel, Microporous Mesoporous Mater., 1999, 27, 151.
X. L. Yang, W. L. Dai, H. Chen, J. H. Xu, Y. Cao, H. X. Li and K. N. Fan, Appl.
Catal., A, 2005, 283, 1.
Q. J. Zhu, X. F. Chu, Z. Y. Zhang, W. L. Dai and K. N. Fan, RSC Adv., 2013,
3, 1744.
Z. H. Luan, E. M. Maes, H. P. Van Der, D. Y. Zhao, R. S. Czernuszewicz and
L. Kevan, Chem. Mater., 1999, 11, 3680.
X. L. Yang, W. L. Dai, H. Chen, Y. Cao, H. X. Li, H. Y. He and K. N. Fan,
J.Catal., 2005, 229, 259.
J. Y. Tang, L. Wang, G. Liu, Y. Liu, Y. Z. Huo, W. X. Zhang, M. J. Jia and
W. R. Thiel, J. Mol. Catal. A: Chem., 2009, 313, 31.
K. Kinoshita, Carbon: electrochemical and physicochemical properties, John
Wiley Sons, New York, NY, 1AD, 1988.
S. Maldonado, S. Morin and K. J. Stevenson, Carbon, 2006, 44, 1429.
D. Jana, C. L. Sun, L. C. Chen and K.-H. Chen, Prog. Mater. Sci., 2013, 58, 565.
S. Maldonado and K. J. Stevenson, J. Phys. Chem. B, 2005, 109, 4707.
P. Ayala, R. Arenal, M. Rummeli, A. Rubio and T. Pichler, Carbon, 2010,
48, 575.
Y. P. Zang, L. P. Li, Y. Zuo, H. F. Lin, G. S. Li and X. F. Guan, RSC Adv., 2013,
3, 13646.
D. Chen and J. H. Ye, Adv. Funct. Mater., 2008, 18, 1922.
K. Sayama, H. Hayashi, T. Arai, M. Yanagida, T. Gunji and H. Sugihara, Appl.
Catal., B, 2010, 94, 150.
L. Y. Huang, H. Xu, Y. P. Li, X. N. Cheng, J. X. Xia, Y. G. Xu and G. B. Cai,
Dalton. Trans., 2013, 42, 8606.
K. Katsumata, R. Motoyoshi, N. Matsushita and K. Okada, J. Hazard. Mater.,
2013, 260, 475.
J. Ding, Q. Q. Liu, Z. Y. Zhang, X. Liu, J. Q. Zhao, S. B. Cheng, B. N. Zong and
W. L. Dai, Appl. Catal., B, 2015, 165, 511.
X. Q. An, J. C. Yu, Y. Wang, Y. M. Hu, X. L. Yu and G. J. Zhang, J. Mater. Chem.,
2012, 22, 8525.
Y. M. Zhao, X. Y. Wei, Y. Y. Wang and F. Luo, Chem. Phys. Lett., 2014, 607, 34.
L. Fu, T. Xia, Y. H. Zheng, J. Yang, A. W. Wang and Z. Wang, Ceram. Int., 2015,
41, 5903.
M. Huang, Y. Wu and W. B. Hu, Ceram. Int., 2014, 40, 7219.
H. Y. Wu, M. Xu, P. M. Da, W. J. Li, D. S. Jia and G. F. Zheng, Phys. Chem.
Chem. Phys., 2013, 15, 16138.
A. Devadoss, P. Sudhagar, S. Das, S. Y. Lee, C. Terashima, K. Nakata,
A. Fujishima, W. Choi, Y. S. Kang and U. Paik, ACS Appl. Mater. Interfaces,
2014, 6, 4864.
T. Wang, J. Tang, X. L. Fan, J. H. Zhou, H. R. Xue, H. Guo and J. P. He,
Nanoscale, 2014, 6, 5359.
Y. Wang, C. X. He, A. Brouzgou, Y. Liang, R. W. Fu, D. C. Wu, P. Tsiakaras
and S. Q. Song, J. Power Sources, 2012, 200, 8.
Catalysis, 2016, 28, 127 | 27
Alumina ceramic foams as catalyst

supports
17:23:28.
Erfan Behravesh,* Leena Hupa, Tapio Salmi and

Dmitry Yu. Murzin*
DOI: 10.1039/9781782626855-00028
Ceramic foams have a wide range of potential applications in biomedicine, thermal insulation, filtration of molten metal alloys, adsorption of environmental pollutants, catalyst
supports, etc. Since the physical properties of the foams do not fully meet the requirements in some applications, improvement of conventional fabrication methods or totally
new techniques are of interest. Herein, three main methods of manufacturing ceramic
foams are introduced with the main emphasize on the replica technique. Furthermore,
dierent techniques for improving structural properties of ceramic foams are reviewed.
The focus of this review is on fabrication of macro-porous alumina foams with high
interconnected porosity. In addition, experimental data for manufacturing of ceramic
foams via the replica technique are presented along with literature surveys. Slurries
consisted of alumina powder mixed in aqueous solutions of polyvinyl alcohol (PVA) and
magnesia and titania as sintering aids. The foams were produced by tuning dierent
processing parameters to give properties suited for catalyst supports. These parameters
included pore size of the polyurethane (PU) foam used as a template, parameters in the PU
foam pretreatment, particle size of alumina powder in the slurry, slurry loading and drying
of the green alumina coated PU foam. Finally, the key factors for optimizing ceramic
foams in terms of mechanical strength and interconnectivity are introduced together with
an outlook for future advances in ceramic foams as catalyst supports.
Introduction
Ceramic materials have distinct advantages over other materials such as

polymers and metals. Among the ceramic materials, those containing
porosity have attracted the attention during the last decades due to their
special properties and features compared to the conventional dense
counterparts. Traditionally, porosity has been avoided because it may
lead to catastrophic failure of brittle ceramics during loading. Today,
porous materials find a broad range of applications. In moderately
loaded applications, the porous ceramics typically fail locally and crack
propagation may be avoided.
Porous ceramic foams are generally classified into two types: closed
pore foams which are like a network of soap bubbles, and open pore
foams (reticulated foams) with porosity exceeding 60%. Any change in
closed and open porosity, pore size distribution and pore morphology
might lead to a significant change in the properties of the final material.1
One of the advantages of open pore foams is the flow-through capability enabling liquids and gases to flow through the foam easily. In
producing porous ceramics as a catalyst support, mechanical strength
and interconnectivity are the main properties which should be optimized
bo Akademi University, Faculty of Science and Engineering, Johan Gadolin Process
Chemistry Centre, Biskopsgatan 8, Turku 20500, Finland.
E-mail: erfan.behravesh@abo.fi; dmurzin@abo.fi
28 | Catalysis, 2016, 28, 2850

c
View Online
17:23:28.
regardless the preparation method. It should be noted that the microstructural features are influenced by the processing route for manufacturing of the porous material.
Cellular ceramics are divided into foams, honeycombs, connected
rods, connected fibers, hollow spheres and bio-template structures.3
According to the unrivaled properties of ceramic foams like high permeability, high external surface, high temperature stability and good
thermal shock resistance, they have a wide range of applications. The
applications of cellular ceramics include filtration of molten metals and
hot gases, thermal and acoustic insulation, light-weight structural components, support for fuel cells, electrodes, sensors, bioreactors, radiant
burners as well as scaolds for bone replacement.47 In addition, ceramic
foams have broad applications in dierent catalytic processes as catalyst
supports. They have crucial benefits including high yield and selectivity,
improved temperature control and heat management. Ceramic foams
have shown several advantages over packed beds of catalyst particles
which have made them an interesting field of study for dierent reactions. The main advantages of ceramic foams are (1) the ability to
match the shape and size of the reactor for easier loading of long and
narrow tubes; (2) reduced pressure drop relative to packed beds, saving
energy costs; (3) higher eectiveness factor because of higher external
surfaces; and (4) enhanced heat transfer, avoiding hot spots and allowing
better reactor stability for highly exothermic reactions.8 In addition to
these advantages, many other benefits of structured catalyst supports to
overcome technical problems are reviewed by Moulijn et al.9
Pressure drop and heat transfer are the main transport properties of
the foams. Many eorts have been made aiming to study the pressure
drop in ceramic foams such as investigation of the eect of dierent
parameters on pressure drop.10,11 Basically, pressure drop in ceramic
foams depends on the fraction of the pore volumes between the struts.
Richardson et al. however, showed that changes in texture of the surface
e.g., by washcoating the surface, influence the pressure drop in ceramic
foams implying the pressure drop in foams also depends on the physical
characteristics of the surface.12
Enhanced convection in ceramic foams due to turbulence in tortuous
pores and radiation between the struts leads to a better heat transfer
in foams compared to fixed bed reactors. For instance, Peng and
Richardson developed a suitable radial heat transfer correlation for one
dimensional reactor model.13 The reactor contained 30 PPI a-Al2O3 ceramic foam. The authors compared the heat transfer between equivalent
ceramic foam and particle beds. It was shown that the ratio of heat transfer
coecient to particles increased. The increase is even higher at higher
Reynolds numbers. In addition, they also found that washcoating of the
foams with 30 PPI of pore density increases the heat transfer coecient.
There are many heat transfer limited operations including both
endothermic and exothermic reactions. Twigg and Richardson compiled
a list of these reactions including partial oxidations, alkylations, oxychlorination, hydrogenation and dehydrogenations, for which ceramic
foams as catalyst supports could be beneficial (Table 1).14
Catalysis, 2016, 28, 2850 | 29
View Online
Table 1 Classification of industrial catalytic processes for which ceramic
foams could be beneficial.14 Reproduced from M. V. Twigg and J. T.
Richardson, Theory and Applications of Ceramic Foam Crystals. Chem.
Eng. Res. Des., 80. Copyright (2002) with permission from Elsevier.
17:23:28.
Type
Exothermic nature
Partial oxidation
Partial oxidation
Partial oxidation
Partial oxidation
Partial oxidation
Alkylation
Alkylation
Oxidative rearrangement
Oxychlorination
Oxidation
Hydrogenation
Hydrogenation
Hydrogenation
Endothermic nature
Dehydrogenation
Dehydrogenation
Oxidation
Dehydrogenation
Steam reforming
Steam reforming
Example
Ethylene to ethylene oxide
o-Xylene to phthalic anhydride
Propene to acrylic acid
Butane to maelic anhydride
Methanol to formaldehyde
Benzene to ethyl benzene
Diethylbenzene or cumene
Water gas shift
Acetic acid to vinyl acetate
Ethylene to ethylene dichloride
Methanol synthesis
Methanation of CO/CO2
FischerTropsch synthesis
Ethylbenzene to styrene
Cyclohexane to benzene
Cyclohexane to cyclohexanone
Butanol to methyl ethyl ketone
Natural gas to synthesis gas
Naphtha to synthesis gas
FischerTropsch synthesis is one example of complex and exothermic

reactions which can be successfully done by using foams. Brown et al.
compared Co/g-Al2O3 foam catalyst with powders and pellets.15 A higher
eectiveness factor and reaction rate were achieved for the catalyst
foams. Other catalytic applications of ceramic foams have been reported
for exothermic reactions such as steam reforming, especially methane
steam reforming and ethylene epoxidation using Rh/g-Al2O3 and Ag/gAl2O3 for which a higher reaction rate and turnover frequency compared
to catalysts in powder form were achieved.13,15,16
Apart from applications dealing with transport properties, most applications of foam catalysts are related to chemistry that occurs on the
surface of the struts. These applications include: selective CO oxidation,17 solar methane reforming with CO2,18 methanol to olefins,19
catalytic combustion20 and oxidative dehydrogenation for which desirable intermediates are produced in consecutive reactions.21
Ceramic foams with high porosity are produced using one of the three
main processing routes: replica technique, direct foaming and sacrificial
template. The manufacturing methods dier in terms of processing
features and final properties of the foams. Majority of ceramic foams is
produced by the replica technique. The replica method was reported by
Schwartswalder and Somers in 196322 and since then it has been utilized
in several studies devoted to the preparation of ceramic foams. This
method consists of slurry preparation, pretreatment of the polyurethane
(PU) foam in order to adhere more slurry to the surface, impregnation of
30 | Catalysis, 2016, 28, 2850
View Online
17:23:28.
23
PU foam by dipping or even spraying with ceramic slurry drying of the

foam and finally sintering of the ceramic particles. In addition, during
sintering the polymer template is burnt out by applying a high temperature to get the final material with the desired properties.3
There are dierent templates which can be used as cellular structure in
replica technique. These templates can be either synthetic including
polymeric24 and carbonic foams25 or even natural ones such as coral26 or
wood.27
Replica technique is a well-established method in order to produce
cellular ceramics. With this technique open porosity up to 90% with cell
sizes ranging from a few hundred micrometers to several millimeters is
achievable. This method is simple and flexible because it is applicable
for any ceramic material which can be dispersed into a suspension.
Moreover, this method is cost eective without toxic emissions during
PU foam burn out. The main outstanding feature of this technique is
high open porosity achievable for the ceramic foams which makes them
suitable to be used as catalyst supports. Obtaining ceramic foams with
high mechanical strength is, however, a crucial step in this method. It
can be either because of low adhesion of suspension to the polymeric
foam as template or because of formation of cracks in very thin struts
during sintering while the polymeric foam is burnt out.23 However,
there are several ways to reduce this shortcoming. The properties of
alumina slurries can be adjusted, e.g. by using additives such as MnO,
TiO2, MgO, Y2O3 and ZrO2. Pretreating of PU foam before slurry
impregnation, doing a second impregnation step in order to fill the
cracks in the struts and using fibers or reactive compounds to improve
the integrity of the materials have been reported as means to increase
the mechanical strength.24,2834 Wash-coating as a secondary coating
also helps to get higher mechanical strength because it can fill the
cracks and flaws in the struts. In addition, the slurry viscosity plays a
crucial role in determining the mechanical strength of the alumina
foams.35
Direct foaming is another method of manufacturing porous ceramics
used for fabrication of structures containing both open and closed porous structures.3638 This method is based on incorporation of air into a
suspension of ceramic powders or into a ceramic precursor suspension.
The slurry is then set to keep the air bubbles structure in the following
drying and sintering steps. The ceramic therefore retains the resulting
spherical pores. There are dierent ways for incorporating gas bubbles
such as by mechanical frothing, injection of gas stream, gas releasing
chemical reactions or solvent evaporation.1 Direct foaming is a low-cost
and eective way of making ceramic foams but stabilization of the air
bubbles within the initial suspension is a crucial step in this method. The
large interfacial area of the mixture of the air bubbles and the suspension
causes the bubbles to change from the spherical shape to polyhedral
cells. The spherical shape of the bubbles has been stabilized by adding
into the solution solid particles having a suitable surface chemistry39 or
surfactants which adsorb at the gaswater interface and control surface
tension and the interface stability.40
Catalysis, 2016, 28, 2850 | 31
17:23:28.
View Online
Sacrificial template technique is another strategy for producing

cellular ceramics. This technique consists of a preparation of ceramic
precursor and dispersion of a solid sacrificial organic template into the
precursor. Later on during heating, the solid materials are extracted to
form pores within the microstructure. Dierent kinds of sacrificial materials such as synthetic41 and natural42 ones are used in this method
which influence the way of removing these materials via template burn
out. The choice of the sacrificial material which decomposes during
thermal treatment is based on the desired porosity, pore size distribution
and pore morphology of the final ceramic foam. Releasing of excess
gasses during decomposition of the sacrificial template is a crucial step
in this technique. Thus, this step should be done with slow rate which
consequently increases the processing time.1
In hollow building blocks technique ceramic hollow spheres are produced by a sacrificial coating method.4244 In this technique, first the
sacrificial cores with dierent diameters are coated and packed in a mold
with slurry and then the cores are attached to each other using a second
slurry followed by calcination and sintering. This technique is however
suitable for producing cellular structures with mainly closed pores so it is
not an appropriate method for making catalyst foams.
This work discusses the optimization of replica technique in preparation of ceramic foams with high open porosity and appropriate mechanical strength to be used as catalyst supports. Dierent methods to
fabricate ceramic foams with hierarchically porosity and suitable mechanical strength have been reported.1,3,45 However, improving the final
structural properties of the foams by tuning the processing parameters
had not been done before. Thus, this work along with overviewing relevant literature addressed the processing parameters including PU foam
pore size, silica sol concentration in the foam pretreatment, particle size
of the alumina powder, amount of slurry loading on the surface of the PU
foam and the influence of drying on the final foam properties. The results give guidelines for manufacturing of porous alumina based ceramic
catalyst supports.
Experimental
2.1 Materials and methodology

Polyurethane foams (Recticel Oy) with 10, 15 and 20 PPI (pore per inch)
were used as templates for the ceramic foams. The fabrication process of
the ceramic foams in this work is illustrated in flow chart below (Fig. 1).
The hydrophobic PU foams were pretreated in two steps in order to increase the adhesion of the slurry. First, the foams were etched in a 1 wt%
sodium hydroxide solution for 24 h to improve the impregnation of the
PU foam in the slurry.46 After drying the foams were immersed in 20, 25
or 30 wt% silica sol solutions. The slurry was prepared by mixing aalumina (Termorak Oy), titania (Aldrich) and magnesia (Fluka) in 5 wt%
polyvinyl alcohol (Acros Organics) solution as a binder. Slurries were
mixed using 10 g of a dry batch (Al2O3, MgO and TiO2) and were milled
(Philips Minimill) for 15, 45 or 75 minutes.
32 | Catalysis, 2016, 28, 2850
View Online
Etching
1 wt% NaOH
24 h
Heating
65C
Slurry of Al2O3, MgO,

TiO2 (99:0.5:0.5 wt%)
and 5 wt% PVA
Impregnating
with silica sol
17:23:28.
(70 wt% solid content)
Ball milling
15, 45 & 75
min
Drying
65C, 48 h
Immersion of
the PU foam
in the ceramic
slurry
Sintering
1500C
6h
Drying
65C
2.5 h & 23 h
Blowing with
air to remove
excess slurry
Fig. 1 Fabrication procedure of ceramic foams via replica technique.
Table 2 Processing parameters in preparation of green slurry coated foams.

Pore size of
the foam (PPI)
Silica sol
concentration (wt%)
Number of
immersions
Foam 1
20
30
Foam 2
15
20
Foam 3
10
20
Foam 4
10
20
Foam 5
Foam 6
20
20
30
30
2
4
Drying
After second and
last immersion
After second and
last immersion
After second and
last immersion
After second and
last immersion
After last immersion
After last immersion
The preparation of green (before sintering) ceramic foams was done

using six dierent combinations of processing parameters (Table 2).
The drying time after immersions was 2.5 and 23 h after the second and
the last immersions, respectively. The foams were squeezed to remove
trapped air only after the first immersion. After each immersion the excess slurry was removed by air blowing.
During sintering the temperature was slowly increased from room
temperature to 1500 1C.47 The sintering temperature should be optimized
since increasing the sintering temperature enhances the mechanical
strength but simultaneously it could reduce the permeability and surface
area and also might lead to exaggerated grain growth.48,49 Scheme 1
shows the sintering schedule of the ceramic foams. The cooling step was
Catalysis, 2016, 28, 2850 | 33
17:23:28.
View Online
Scheme 1 Sintering schedule for manufacturing of the ceramic foams.
slow in order to avoid formation of stresses that might impair the

mechanical strength of the ceramic foams. Generally, both rapid heating
and cooling might lead to crack formation in the ceramic structure.
Zhang et al. studied the eect of the thermal shock resistance on the
structure of ceramic foams.50 It was concluded that in the case of cold
shock, the maximum tensile stress occurs at the surface of the ceramic
foam while for hot shock it occurs at the centre of the ceramic foam.
The thermal stress of the ceramic foam depends on the time, location,
and relative density.
Dierent sintering schedules have been used for producing ceramic
foams via the replica technique. One of the most common ones for
adapting a foam with a desired microstructure is two-step sintering.51
This procedure is optimal for achieving well sintered struts and suppressing grain growth in the final stage of sintering. In this procedure,
the temperature is increased first to a peak value T1 and then reduced to a
dwell temperature T2. A high relative density (around 70%) is usually
achieved at T1 as an intermediate density while a full density is obtained
at T2.52 In addition, since diusion of the atoms and grain boundary
migration are assumed to take place independently during sintering, the
two-step sintering is thought as an eective way to control the pore size or
porosity.53 On the other hand, in sintering of pure alumina well sintered
struts were not achieved with this technique.54
The composition and properties of the slurries and foams were studied
using dierent methods. The particle size distribution of alumina was
measured after dierent milling times using laser diraction (Malvern
AEROS). The sintering properties of the dried slurries were measured
with hot stage microscopy (Misura HSM ODHT 1600-50-08). The critical
temperatures for burning of the PU foam were determined with dierential thermal and thermogravimetric analysis (TA instrument-SDT
Q600).
SEM (Gemini, LEO 1530) images were taken of the ceramic foams in
order to observe the sintering degree and to measure the thickness of the
struts. In addition, the eect of pretreatment steps on the PU foam was
observed from SEM images. Finally, SEM images of the cross-sections
were utilized to measure the macroporosity of the ceramic foams. XRD
34 | Catalysis, 2016, 28, 2850
17:23:28.
View Online
(Philips, PW 1830) analysis of dried slurry and sintered ceramic foam was
done to investigate whether the sintering introduced changes in the
crystallite size. Mechanical strength of the foams was measured using
crush tester (Lorentzen & Wettre, SE 048). Mercury porosimetry (Pascal
140 and 440 series, Thermo Electron Corporation) was used for measuring microporosity of the foam which showed the highest mechanical
strength and macroporosity.
Results and discussion
3.1 Foam pretreatment

Pretreatment of the PU foam is essential for achieving ceramic foams of
good strength. Since the PU foam surface is hydrophobic, adhesion of a
uniform ceramic powder layer using an aqueous slurry is dicult. In
addition, removing the organic template during firing may cause flaws
and thus decreases the mechanical strength of the final ceramic foams.46
In order to avoid these drawbacks dierent PU foam pretreatment
methods have been reported to promote the adhesion of the slurry to the
organic foam and also to reduce crack growth during sintering. Binders
and fibers are used in most of the cases to ensure the adhesion of the
slurry to the ceramic foam and also to inhibit the crack growth during
sintering.55 Silica sol,33,46,5558 polyvinyl alcohol,5962 polysaccharide2
and bentonite63,64 are the most common binder additives used to increase the adhesion of the slurry to the PU foam. Application of an adhesive coating to the flexible foam prior to slurry impregnation has been
reported.46 Such adhesive coating can even improve the elasticity of the
PU foam and thus make the impregnation easier. The adhesion of the
slurry may also be aected by suitable flocculation of ceramic particles or
by formation of a gel precipitate which does not redisperse into the
slurry.65 Flocculation of the ceramic powders in the immediate vicinity of
the foam surface may be achieved e.g. by synthetic high molecular weight
polymer such as polyacrylamide or by cellulose derivative such as polysaccharide gum.2,65 If a gel precipitate on the template surface is desired,
usually a base such as monethanolamine which is adsorbed on the foam
surface is utilized. After drying and dipping into a solution of aluminum
hydroxychloride, aluminum hydroxide gel precipitates on the foam
surface.
Pu et al. reported a method for pretreating polyurethane foams to form
a hydrophilic silica gel on the template surface.46 The foams were first
immersed in 1 wt% sodium hydroxide and then dipped in silica sol and
dried. Since the silica gel made the PU surface coarser the slurry spread
smoothly. SEM images of the surface condition after the silica sol pretreatment and mechanical strength of the final foams verified that the
silica sol enhanced formation of a uniform coating with the slurry and
also significantly increased the slurry loading.
One important parameter in increasing the slurry adhesion to the
surface is the concentration of the binders or fibers. For example, Kim
et al., used 2 to 5 wt% polyvinyl alcohol as a binder when manufacturing
alumina foams via the replica method.62 For concentrations higher
Catalysis, 2016, 28, 2850 | 35
17:23:28.
View Online
than 5 wt% the PVA powders remained undissolved. The authors fixed
the PVA concentration at 3 wt% since at concentrations lower than
2 wt% the binding strength was low. Adding fibers was found to aect
the slurry adhesion: if used in concentrations 1 to 5 wt% a good adhesion was achieved. At higher concentrations, the fibers clump and
consequently do not give a proper dispersion in the slurry. The fiber
concentration should be at least 1 wt% to give a significant increase of
the strength of the foams.64
Apart from binders, gelcasting is another way of avoiding the
crack formation during drying and thus improving the mechanical
strength.6668 The pores in the gel network are much smaller than the
pores between the ceramic particles. In addition, the curvature of
the meniscus is smaller than the spaces between particles because
the polymer network is smaller than the powder network. Thus, the capillary pressure which causes drying consolidation will be greater.
Chuanuwatanakul et al. studied the influence of the amount of gelcasting
additives on the mechanical strength of the ceramic foam green bodies.69
They compared the eect of cross-linking agent versus the addition of just
a binder on tensile strength, compressive strength and elastic modulus.
Using PVA as a binder and 2,5-dihydrofurane (DHF) as a cross-linking
agent together increased the strength and elastic modulus, significantly.
Authors of this work studied the influence of pretreatment of the PU
foam as one parameter aecting the strength. In the pretreatment the
concentration of silica sol was varied. Figure 2 shows SEM images of the
polyurethane foam surface after 24 hours of etching with 1 M sodium
hydroxide and the surface after coating with 30 wt% silica sol.
The SEM images show large cracks in the PU foam surface after the
pretreatment. Cracks in the silica sol layer are likely to form when the
sample is dried and water evaporates. The influence of the silica sol
concentration on the crack formation after etching with 1 M NaOH and
coating with 30, 25 and 20 wt% silica sol is illustrated in SEM micrographs in Fig. 3.
Figure 3 suggests that less cracks formed at lower silica concentration.
The mechanical strength of the final ceramic foams indicated that the
foam produced by using 20 wt% silica sol in the pretreatment step had a
Fig. 2 Surface of the PU foam after; (a) etching with 1 M sodium hydroxide and (b)
impregnation with 30 wt% silica sol.
36 | Catalysis, 2016, 28, 2850
17:23:28.
View Online
Fig. 3 Surface of PU foam after etching with 1 M NaOH and coating with (a) 30 wt%,
(b) 25 wt% and (c) 20 wt% silica sol.
higher strength than the one with 30 wt% (discussed in 3.5). It was assumed that a more flexible silica layer formed when using silica sols with
low concentrations. Similar results were reported by Pu et al.: silica sol gel
made the template surface coarser and finally improved the mechanical
strength by enabling better slurry adherence.46 It should be emphasized,
however, that the concentration of the silica sol must be precisely
optimized to give a desired mechanical strength.
3.2 Milling time vs. particle size and sintering degree
Alumina particles can be sintered into ceramic monoliths at temperatures around 1700 1C. High sintering temperature causes abnormal
grain growth resulting in low strength. Sintering at lower temperatures
is, however, possible. In order to have an ecient process at lower
sintering temperatures, additives such as magnesium oxide30 are used
to inhibit excessive grain growth. Additives might also have detrimental
eects on the hardness and inertness of alumina.70 Besides using
additives, some other approaches to achieve desired structures such as
microwave sintering have been reported.7174 Chinelatto et al., reported
heating curve control as a simple and ecient method which helps to
maximize the final density and minimize the grain growth of the particles.75 However, the main parameter, which can lower the sintering
temperature considerably, is utilization of initial alumina powders with
smaller particle sizes.70,76 Furthermore, finer grain sizes of alumina
enhances particle packing,7779 flexural strength and wear resistance.80,81 Particle size is also directly related with the pore size between
Catalysis, 2016, 28, 2850 | 37
the particles present in the final sintered alumina. The finer the particle, the smaller is the pore size.82
Contamination is an obstacle associated with using very fine particles.
Since the surface area is large, removal of impurities from the particle
surfaces is dicult. Contamination during ball milling typically takes
place due to abrasion of hard mill materials.70 De-contamination of the
powders can be done e.g. by boiling in 35% HCl solution.83 Careful
selection of milling media and choice of milling parameters are essential
for achieving a powder with the controlled particle size and composition.
Influence of the particle size and particle size distribution on dryingshrinkage behavior of alumina was studied by Tari et al.84 They found
that in low shear rates, as the average particle size decreases, surface
forces dominate over the hydrodynamic forces, which lead to an increase
in shear thinning behavior. Moreover, an improvement in particle
packing is achievable by suitable size distributions of particles in suspension which also leads to an increase in apparent density in green
state.85
In the experiments performed by the authors the influence of the
milling time on the alumina particle size to enable sintering at lower
temperatures was elaborated. The eect of particle sizes on degree of
sintering was also studied. Alumina particles were milled for 15, 45 and
75 minutes. The volume density as a function of average particle sizes,
and the average particle size as a function of milling time are shown in
Figs. 4 and 5, respectively.
The particle size distribution became broader and shifted to finer sizes
with a prolonged milling time (Fig. 4). The average particle size decreased
from the original average (B50 mm) to 12 mm at 75 min (Fig. 5).
10
Volume density [%]
17:23:28.
View Online
Without milling
15 min milling
45 min milling
75 min milling
5
4
3
2
1
0
0.2
20
200
Size classes [m]

Fig. 4 Volume density versus particle size distribution of alumina (before and after 15, 45
and 75 minutes of milling).
38 | Catalysis, 2016, 28, 2850
View Online
Average particle size (m)
51.8
50
40.1
40
30
24.1
20
12.7
10
0
0
10
20
30
40
50
60
70
80
Milling time (min)
Fig. 5 Average particle size of alumina as a function of milling time (before and after 15,
45 and 75 minutes of milling).
104
Sintering %
17:23:28.
60
(1) Al2O3+MgO+PVA
(15 min milling)
102
(1)
100
(2)
(2) A2O3+MgO+PVA
(45 min milling)
98
(3)
(3) Al2O3+MgO+TiO2+PVA
(75 min milling)
96
94
0
500
1000
1500
Temperature C
Fig. 6 Influence of particle size on commencement of sintering.
The eect of the milling time on the commencement of sintering was

measured by hot stage microscope using dried slurries (Al2O3, MgO, TiO2
and PVA) milled for 15, 45 and 75 min (Fig. 6).
The particles which had been milled for 15 min started to sinter at
1370 1C. Sintering of the powders milled for 45 and 75 min started at 1300
and 1170 1C, respectively. The maximum height decrease at 1500 1C was
5% for the powders milled for 75 min while the height of the samples
milled for 45 and 15 min decreased up to 4 and 1%, respectively. Thus,
the sintering progress was slightly aected by the milling time. The difference in the sintering degree at 1500 1C was minor for the slurries
milled for 45 and 75 min. Moreover, by using shorter milling times,
contamination from the milling materials was minimized. Accordingly,
ball milling for 45 min was selected for further experiments.
3.3 Thermal analysis of PU foam
Analyzing the burning behavior of PU foam is crucial since it gives data
about critical temperatures during burning out the organic foam and also
about shrinkage of the coated foam during heating. Dressler et al.
studied the behavior of PU foam during burning.86 They correlated the
Catalysis, 2016, 28, 2850 | 39
17:23:28.
View Online
Fig. 7 DTA/TG curves of polyurethane foam.
PU decomposition and resulting shrinkage of the ceramic coated foam.

The results suggest that the PU sponge thermally expands below 167 1C.
Thereupon, the slurry coated foam starts to shrink simultaneously with
PU decomposition and by the time the PU foam is completely burnt out
the shrinkage is finished.
DTA/TG for PU foam in the experiments reported in this work within
the temperature range of interest for the burning of the PU foam are
shown in Fig. 7.
The peak value at 380 1C in the DTA curve is due to burning of the foam
starting at around 300 1C. The PU foam is combusted completely at ca.
600 1C which is in good agreement with the values reported by Noudem
et al.,87 and Callcut et al.88 Therefore, the particle layer which mimics the
PU foam structure should achieve an acceptable sintering degree at
temperatures 300600 1C in order not to collapse at higher temperatures.
3.4 Microstructure of the foams
Figure 8 shows SEM micrographs of the prepared ceramic foams using
the parameters given in Table 2.
No clear dierence in the sintering degree of the foams was observed.
Neck growth of some of the particles can be seen in the micrographs
albeit the grain growth and densification were minor.
SEM images of foams 1 and 5 show the influence of the slurry loading
on the thickness of the struts (Fig. 9). Foam 1 had been immersed four
times in the slurry, while only two immersions had been done for foam 5.
Photographs of these two foams manufactured using 20 PPI PU foam as
template are shown in Fig. 10.
The strut diameters in Figs. 9 and 10 suggest that several dipping steps
with drying in between assist in achieving thicker struts of the final
ceramic foams.
40 | Catalysis, 2016, 28, 2850
17:23:28.
View Online
Fig. 8 SEM micrographs of the produced foams: (a) foam 1, (b) foam 2 and (c) foam 3.
Fig. 9 Diameter of the struts: (a) foam 1 dipped four times in the ceramic slurry with
drying after the second and the last immersions, (b) foam 5 dipped twice with drying after
the last immersion.
3.5 Mechanical strength

Replica technique is a suitable method in terms of controlling the
mechanical strength of the foams which are abated by hollow struts.
Nonetheless, there are dierent methods in order to overcome this
shortcoming e.g. by providing solutions for achieving a better uniform
coating, reducing the flaws and pores caused by the template burnout
and increasing loading content and strut thickness.
In addition to treating the PU foam before impregnating with ceramic
slurry, viscosity of the slurry plays a decisive role in controlling the
strength of ceramic foams. Ceramic slurry must possess shear thinning
behavior.89 Rheological behavior varies with using dierent additives and
with solid content of the slurry which consequently aects the loading
Catalysis, 2016, 28, 2850 | 41
17:23:28.
View Online
Fig. 10 Visual appearance of ceramic foams manufactured using 20 PPI PU template:

(a) foam 1 (four immersions) and (b) foam 5 (two immersions).
content of the slurry.23 Voigt et al. used dierent thickeners to prepare

thickened alumina slurries for foam production by replica technique.90
They evaluated the influence of viscosity, thixotropy and the level of shear
thinning behavior on coating behavior. It was observed that both quantity and quality of the coating were changed markedly by varying thickener and solid contents. It was found also that slurries with shear
thinning index of no0.85 exhibited a good coating behavior.
In the fabrication of silicon carbide ceramics using the polymer sponge
replication technique, an increase in the slurry viscosity gave higher
material loading, thicker struts and increased strength. In contrast,
the cell size of the foam decreased with increasing slurry viscosity.89
Optimizing an appropriate solid content is challenging; very high solid
concentrations give too viscous slurries to give a proper coat especially of
the core of the PU foam.91 Moreover, high viscosity thickens the struts
and consequently abates the permeability of the foam.92 The common
solid concentrations of slurries used in replication method are between
50 to 70%.
Besides using slurries with higher viscosities, multi-impregnating of
the template in the slurry has been suggested as a means to increase the
mechanical strength of the struts. Usually, an additional layer above the
previous layers has been applied using drying step after each dipping.60
For instance, a ceramic foam was made by impregnating of urethane
template with a ceramic slurry containing alumina and zirconia for three
times with drying in between, followed by sintering at 1500 1C.59,93
The mechanical strength of the resulting foams was 510 MPa compared to 13 MPa for the foams impregnated only once in the slurry. It
should be noted that multiple impregnating is helpful insofar as the
open-cells are not clogged or filled with the slurry; otherwise the permeability of the open-celled structure is diminished. Although multiple
dipping is successful in increasing the strength, recoating the structure
after the first coating layer is dicult even if the sample is dried. In
recoating, the previously deposited layer can partly be washed away thus
42 | Catalysis, 2016, 28, 2850
17:23:28.
View Online
decreasing the strength. In addition, burn out of the template can give
rise to micro pores also at recoated structures. Zhu et al., tried to minimize micro pores by using a double firing process in which the polyurethane template was removed after the first coating at 800 1C, followed
by recoating with a thinner layer covering the micro pores and final
sintering at higher temperatures.56 However, with multi-impregnating
technique usually triangular channels remain within the struts. To
overcome this defect, infiltration technique has been used to fill up the
hollow struts formed after the pre-sintering of ceramic foams.92,94 After
the infiltration under vacuum the resulting foams are sintered for the
second time. By using this technique, a two-fold increase in mechanical
strength compared to multi impregnating technique was reported by
Vogt et al.,92 and Jun et al.,94 who used alumina and zirconia slurries,
respectively.
The mechanical strength of ceramic foams has been increased by using
reaction bonded aluminum oxide (RBAO) instead of conventional alumina.95 RBAO has high mechanical strength because of better wetting
behavior which is due to finer grain structure and absence of glassy
phases at grain boundaries. This material is made by milling of Al/Al2O3
mixture followed by oxidizing, shaping, sintering and finally machining.
Oxidation of Al is, however, challenging since the reaction is exothermic
and may cause cracks in the structure.
One important parameter in achieving high mechanical strength is
the thickness uniformity of the struts especially when the foams are
coated several times with the slurry. Removing the slurry from the inner
part of the PU foams is dicult; the larger the diameter of the foam, the
more dicult is the removal of the excess slurry. Large foams have been
fabricated by attaching foams of smaller diameters to uniform foam of a
desirable size. Pearson et al., proposed a method for adhering two
ceramic foams to each other with a liquid reinforced bonding resin
made of a liquid resin and reinforcement fibres.96 The liquid was cured
to form a solid resin followed by pyrolyzing and densifying. It should be
mentioned that the reinforcement fibres were formed of a ceramic
material matching the chemical composition of the foams. This method
can also be used to repair damaged foams by applying another
ceramic layer.
Uniform strut thickness is usually ensured by removing excess ceramic
slurry from the coated organic templates by the roll-pressing56,89,97,98 and
centrifuging.33 Roll-pressing is based on passage of the coated ceramic
foam through preset rolls. In addition to controlling uniformity, it aects
the permeability of the foam. However, this technique is challenging
since usually flaws such as sharp corners at the strut edges remain in the
structure which deteriorates the mechanical properties of the foams and
limits their applications.99,100
In contrast, in centrifuging technique the centrifugal forces are equally
applied to the entire body thus enabling more uniformly coated structures. Pu et al., successfully prepared ceramic foams with uniform strut
thickness via the replica technique by using two centrifuging steps.33 The
second step was done after the slurry was dried to improve the adhesion
Catalysis, 2016, 28, 2850 | 43
to the first layer. Centrifugal technique has been reported useful also in
fabrication of foams with irregular shapes.
Figure 11 demonstrates the compression strength of the foams fabricated in this work. The mechanical strength of the foams varied from
around 70 to 1500 kPa suggesting that the number of immersions and
drying in-between had remarkable eects on the mechanical strength.
The highest mechanical strength in this work is higher compared to the
value achieved in another study in which the PU foams were only dipped
once into the slurry.101 The strength remarkably increased from three
dippings (foam 4) to four dippings (foam 3). The only dierence in the
manufacturing parameters between these two samples was the number
of dippings causing a twofold increase in the mechanical strength.
This observation is in accordance with the results of Pu et al.,33 and
Zhu et al.56
Drying also aected the strength of the foams. Foam 1 and foam 6 were
manufactured using the same parameters except drying steps. About a
tenfold increase in the mechanical strength was achieved by adding one
more drying step.
The cell size of the template PU foam also aected the mechanical
strength. Foam 2 made of 15 PPI PU template had a higher mechanical
strength than the foam 3 made of 10 PPI foam. Similar results were reported by Brezny et al.,102 and Rehorek et al.103 They studied the eect of
the template cell size on the final strength of the ceramic foam and found
that the compressive and tensile strength varied inversely with the cell
size. Basically, the PU foam with higher porosity (15 PPI) has higher
surface area than the PU foam with lower porosity (10 PPI). Therefore, a
1600
Compression strength / kPa
17:23:28.
View Online
1200
1490
1180
1080
800
570
400
68
95
Foam 5
Foam 6
0
Foam 1
Foam 2
Foam 3
Foam 4
Dipping
Drying
2nd+4th
2nd+4th
2nd+4th
2nd+4th
2nd
4th
Fig. 11 Mechanical strengths of the manufactured foams using parameters given in

Table 2: Foams 1 (20 PPI), 2 (15 PPI), 3 (10 PPI), 4 (10 PPI), 5 (20 PPI) and 6 (20 PPI). Number
of dippings in the slurry and steps after which the foams were dried are indicated below
the graph. Standard deviation for foam 2 was 2.19% (seven parallel foams with the same
procedure were made).
44 | Catalysis, 2016, 28, 2850
17:23:28.
View Online
larger amount of slurry adhered to the PU template with higher porosity

and consequently higher mechanical strength was achieved. However,
Acchar et al., reported that the cell size did not aect the mechanical
strength.104
A comparison between foams 1 (20 PPI) and 2 (15 PPI) shows, however,
contradictory results. Although the foam 1 is more porous than the foam
2, it demonstrated a lower mechanical strength. This dierence may
depend on silica concentration in the pretreatment of foam 1. The cracks
formed in the silica-coated foam are likely to aect the final strength. The
eect of the silica concentration in the primary coating on the mechanical strength is not clear and should be studied in more detail. Another
possibility might be that if the cracks in the inner primary silica coating
are not properly covered by the ceramic slurry, they might decrease the
strength of the final foam.
3.6 Porosity measurements
High porosity increases the permeability of the ceramic foam.105,106
In addition, pore size and also the strut diameters play a crucial role in
heat transfer.107 Herein, the influence of PU foam pore size on the
macroporosity of the final alumina foam was measured for the samples
prepared using 20, 15 and 10 PPI PU templates. Figure 12 displays the
SEM images of the cross-sections of the foams.
According to the image analysis, the final foams had approximately the
same macroporosity (64%) albeit the pore sizes of the PU templates were
dierent. Thus, the PU template pore size did not aect the macroporosity of the final foam. According to Guzman, a porosity of 6075% is
considered as high value for alumina foams.108 Thus, these foams have
proper macroporosity in order to be used as catalyst supports.
As additional data, cumulative and relative pore volumes as a function
of the pore diameter were measured using mercury porosimetry for foam
2 having the highest mechanical strength (Fig. 13). It should be noted
that here the pores do not refer to the cell size but porosity of the struts.
The graph (Fig. 13) shows that the pore diameter varied between 4 and
120 mm. In addition, the pore size distribution showed two maxima: the
pores with the d 16 mm diameter comprised of 5.5% of the total pore
Fig. 12 SEM images of the cross-section of the porous alumina foams. The black parts
show the open pores; (a) foam 1 (20 PPI), (b) foam 2 (15 PPI) and (c) foam 3 (10 PPI).
Catalysis, 2016, 28, 2850 | 45
View Online
Pore size distribution
Cumulative pore volume (mm3/g)
10
9
200
8
7
6
150
5
4
100
3
2
50
Relative pore volume (%)
17:23:28.
250
1
0
0.001
0.01
0.1
10
100
0
1000
Pore diameter (mm)

Fig. 13
Cumulative and relative pore volume as a function of the micro pore diameter.
volume, and the pores with d 100 mm corresponded to 9% of the total

pore volume. In order to get a high mechanical strength the struts should
have high density which means that they should be less porous.
3.7 Shrinkage and weight loss
Shrinkage and weight loss were measured for the foam 2 using seven
parallel samples. The shrinkage and the weight loss were (14 0.34)%
and (6 0.09)%, respectively. These values are considered small and
therefore; do not aect the performance of the foams as catalyst supports. However, the shrinkage must be controlled since high shrinkage
abates the interconnectivity of the cells but on the other hand it should
be high enough to avoid high pressure drop.
4 Conclusions and outlook

There are three main fabrication methods of ceramic foams which are
reviewed in this work; replica technique, direct foaming and sacrificial
template. The replication procedure consists of several steps; slurry
preparation, dipping of the foam into slurry, drying and sintering. There
are also other manufacturing parameters which should be taken into
account including foam pretreatment, composition and viscosity of the
ceramic slurry and alumina particle size. In addition, the pore size and
shrinkage of the ceramic foams should be controlled.
The most important properties of the ceramic foams used for preparation of primary catalyst support are open porosity and mechanical
strength. In this work, dierent approaches for improving mechanical
features of primary supports based on alumina foams manufactured via
replica technique were reviewed. The replica technique was selected as a
cost eective method to manufacture the catalyst supports with acceptable mechanical strength and pore interconnectivity. The open porosity
and mechanical strength of the foams were controlled by the PU template
46 | Catalysis, 2016, 28, 2850
17:23:28.
View Online
pore size, pretreatment of the PU foam with silica sol, particle size of
alumina powder in the slurry and impregnation and drying steps for
achieving an optimal adhesion of the particles on the struts of the PU
template.
Particle loading was the key factor to improve the mechanical strength
of the foams which was mainly increased by more dippings and drying
steps. However, high number of dippings increases the struts diameter
and consequently reduces the interconnectivity of the cells. Thus, the
amount of dippings should be optimized. Herein, in order to get an
appropriate mechanical strength other parameters such as the cell size of
PU foam or pretreatment steps can help to adhere an acceptable amount
of slurry to the struts surfaces.
Mechanical strength varied inversely with cell size of the PU foam.
Using silica sol with low concentrations in the pretreatment step such as
less than 30 wt% avoided crack formation in struts and helped to adhere
required amount of the slurry to the PU foam. By reducing the alumina
particle sizes to 12.7 mm via milling, sintering was only slightly improved.
The pore size of the PU template did not aect the macroporosity of the
final foams, markedly.
Novel fabrication techniques should be developed allowing manufacturing foams with specific structural properties for specific applications.109 In addition, there would be a possibility to extend the features
of the ceramic foams e.g., by wash-coating a-alumina as primary support
with g-Al2O3 in order to enhance the catalyst surface area. Thereafter,
an active phase can be introduced to the surface of the foams. Several
examples of utilization of catalytic foams prepared in such way have been
already reported in the literature.109111
Acknowledgements
The work is a part of the activities of Johan Gadolin Process Chemistry
Centre, a centre of excellence in scientific research financed by bo
Akademi University, Finland.
References
1
2
3
4
5
6
7
8
A. R. Studart, U. T. Gonzenbach, E. Tervoort and L. J. Gauckler, J. Am. Ceram.

Soc., 2006, 89, 1771.
hler, U. Hopfner, J. Aigner and E. Wintermantel,
H. Haugen, J. Will, A. Ko
J. Eur. Ceram. Soc., 2004, 24, 661.
P. Colombo, Philos. Trans. R. Soc., 2006, 364, 109.
S. Barg, C. Soltmann, M. Andrade, D. Koch and G. Grathwohl, J. Am. Ceram.
Soc., 2008, 91, 2823.
ttcher, P. M. Scho
neich, S. Haase, S. Hempel and R. Lange,
F. Lali, G. Bo
Chem. Eng. Res. Des., 2015, 94, 365.
M. Twigg and J. Richardson, Stud. Surf. Sci. Catal., 2006, 162, 135.
A. C. Basagiannis and X. E. Verykios, Catal. Today, 2007, 127, 256.
M. V. Twigg and J. T. Richardson, Ind. Eng. Chem. Res., 2007, 46, 4166.
Catalysis, 2016, 28, 2850 | 47
View Online
17:23:28.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
J. A. Moulijn, I. M. Makkee and A. E. van Diepen, Chemical Process Technology, 2nd edn, Wiley, New York, 2013.
ski
W. Regulski, J. Szumbarski, L. aniewski-Wo""k, K. Gumowski, J. Skibin
and M. Wichrowski, Chem. Eng. Sci., 2015, 137, 320.
D. Edouard, M. Lacroix, C. P. Huu and F. Luck, Chem. Eng. J., 2008, 144, 299.
J. T. Richardson, Y. Peng and D. Remue, Appl. Catal., A, 2000, 204, 19.
Y. Peng and J. T. Richardson, Appl. Catal., A, 2004, 266, 235.
M. V. Twigg and J. T. Richardson, Chem. Eng. Res. Des., 2002, 80, 183.
S. A. Brown, PhD dissertation, Department of Chemical Engineering, University of Houston, Houston, TX, 2001.
R. Faure, F. Rossignol, T. Chartier, C. Bonhomme, A. Maitre and
G. Etchegoyen, J. Eur. Ceram. Soc., 2011, 31, 303.
E. I. Tsyganova, I. I. Didenkulova, V. M. Shekunova and Y. A. Aleksandrov,
Russ. J. Gen. Chem., 2005, 75, 1599.
R. Buck, J. F. Muir, R. H. Hogan and R. D. Skocypec, Sol. Energy Mater., 1991,
24, 449.
T. Kodama, A. Kiyama, T. Moriyama, T. Yokoyama, K. I. Shimizu and
H. Andou, Energy Fuels, 2003, 17, 914.
E. I. Tsyganova, I. I. Didenkulova, V. M. Shekunova and Y. A. Aleksandrov,
Russ. J. Gen. Chem., 2005, 75, 1354.
W. Flick and M. Hu, Appl. Catal., 1999, 187, 13.
K. Schwartswalder and A.V. Somers, US Pat., 3 090 094 (A), 1963.
ckel, C. G. Aneziris and J. Huba
lkova
, Ceram. Interfaces, 2013,
C. Voigt, E. Ja
39, 2415.
J. Luyten, I. Thijs, W. Vandermeulen, S. Mullens, B. Wallaeys and
R. Mortelmans, Adv. Appl. Ceram., 2005, 104, 4.
A. Sherman, R. Tuas and R. Kaplan, Am. Ceram. Soc. Bull., 1991, 70, 1025.
R. A. White, N. J. Weber and E. W. White, Science, 1972, 176, 922.
H. Sieber, C. Homann, A. Kaindl and P. Greil, Adv. Eng. Mater., 2000,
2, 105.
M. Sathiyakumar and F. D. Gnanam, Ceram. Interfaces, 2002, 28, 195.
H. Erkalfa, Z. Misirli and T. Baykara, Ceram. Interfaces, 1995, 21, 345.
nyova
, J. Sedla
nkova
and P.
D. Galusek, K. Ghilla
cek, J. Koza
Sajgalk, J. Eur.
Ceram. Soc., 2012, 32, 1965.
V. Viswabaskaran, F. D. Gnanam and M. Balasubramanian, J. Mater. Process.
Technol., 2003, 142, 275.
P. M. Hargus, J. A. Mula and M. K. Redden, US Pat., 4 866 011 (A), 1989.
X. Pu, X. Liu, F. Qiu and L. Huang, J. Am. Ceram. Soc., 2004, 87, 1392.
A. E. M. Paiva, P. Sepulveda and V. C. Pandolfelli, J. Mater. Sci., 1999,
34, 2641.
S. Dhara and P. Bhargava, Int. J. Appl. Ceram. Technol., 2006, 3, 382.
F. K. Juillerat, U. T. Gonzenbach, P. Elser, A. R. Studart and L. J. Gauckler,
J. Am. Ceram. Soc., 2011, 94, 77.
L. Y. Zhang, D. Zhou, Y. Chen, B. Liang and J. B. Zhou, J. Eur. Ceram. Soc.,
2014, 34, 2443.
Y. Li, Y. L. X. Cheng, L. Gong, J. Feng, W. Cao and R. Zhang, J. Eur. Ceram.
Soc., 2015, 35, 267.
E. Dickinson, R. Ettelaie, T. Kostakis and B. S. Murray, Langmuir, 2004,
20, 8517.
P. Sepulveda and J. G. P. Binner, J. Eur. Ceram. Soc., 1999, 19, 2059.
R. A. Lopes and A. M. Segadaes, Mater. Sci. Eng., A, 1996, 209, 149.
J. Luyten, S. Mullens, J. Cooymans, A. M. De Wilde and I. Thijs, Adv. Eng.
Mater., 2003, 14, 715.
48 | Catalysis, 2016, 28, 2850
View Online
43
44
45
46
17:23:28.
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
I. Thijs, J. Luyten and S. Mullens, J. Am. Ceram. Soc., 2004, 87, 170.
F. Qi, X. Xu, J. Xu, Y. Wang and J. Yang, J. Am. Ceram. Soc., 2014, 97, 3341.
P. Colombo, C. Vakifahmetoglu and S. Costacurta, J. Mater. Sci., 2010,
45, 5425.
X. Pu, L. Jia, D. Zhang, C. Su and X. Liu, J. Am. Ceram. Soc., 2007, 90,
2998.
Y. S. Han, J. B. Li, Y. J. Chen and Q. M. Wei, J. Mater. Process. Technol., 2002,
128, 313.
Y. Han, J. Li, Q. Wei and K. Tang, Ceram. Interfaces, 2002, 28, 755.
W. Xue, Y. Huang, Z. Xie and W. Liu, J. Eur. Ceram. Soc., 2012, 32, 3151.
Y. Z. Zhang, B. L. Wang and J. E. Li, J. Am. Ceram. Soc., 2013, 96, 2615.
I. W. Chen and X. H. Wang, Nature, 2000, 404, 168.
I. M. Chen, Interface Sci., 2000, 8, 147.
T. Isobe, A. Ooyama, M. Shimizu and A. Nakajima, Ceram. Interfaces, 2012,
38, 787.
, P.
K. Bodisova
Sajgalk, D. Galusek and P.
Svanc
arek, J. Am. Ceram. Soc.,
2007, 90, 330.
X. Zhu, D. Jiang, S. Tan and Z. Zhang, J. Am. Ceram. Soc., 2001, 84, 1654.
T. Senguttuvan, H. Kalsi, S. Sharda and B. Das, Mater. Chem. Phys., 2001,
67, 146.
Y. Kato and M. Fujimoto, GB Pat., 2 168 336 (A), 1986.
P. N. Kesten and K. Schwartzwalder, US Pat., 3 451 841 (A), 1969.
J. Han, L. Hu, Y. Zhang and Y. Zhou, J. Am. Ceram. Soc., 2009, 92, 2165.
H. D. Kim, T. Nakayama, B. J. Hong, K. Imaki, T. Yoshimura and T. Suzuki,
J. Eur. Ceram. Soc., 2010, 30, 2735.
A. Sarmiento, A. M. Montes, D. Hotza, A. P. Novaes de Oliviera and
J. A. Escobar, Suplemento Rev. LatinAm. Metal. Mater., 2009, 1, 353.
J. C. Blome, US Pat., 4 265 659 (A), 1981.
F. Ernest and G. Ravault, US Pat., 4 004 933 (A), 1977.
E. S. Carreras, F. Chabert, D. E. Dunstan and G. V. Franks, J. Colloid Interface
Sci., 2007, 313, 160.
F. S. Ortega, F. A. O. Valenzuela, C. H. Scuracchio and V. C. Pandolfelli,
J. Eur. Ceram. Soc., 2003, 23, 75.
F. Chabert, D. E. Dunstan and G. V. Franks, J. Am. Ceram. Soc., 2008,
91, 3138.
C. Chuanuwatanakul, C. Tallon, D. E. Dunstan and G. V. Franks, J. Am.
Ceram. Soc., 2013, 96, 1407.
J. S. Forrester, H. J. Goodshaw, E. H. Kisi, G. J. Suaning and J. S. Zobec,
J. Aust. Ceram. Soc., 2008, 44, 47.
J. Samuels and J. R. Brandon, J. Mater. Sci., 1992, 27, 3259.
J. Cheng, D. Agrawal, R. Roy and P. S. Jayan, J. Mater. Process. Technol., 2000,
108, 26.
D. D. Upadhyaya, A. Ghosh, G. K. Dey, R. Prasad and A. K. Suri, J. Mater. Sci.,
2001, 36, 4707.
R. R. Menezes and R. H. G. A. Kiminami, J. Mater. Process. Technol., 2008,
203, 513.
A. S. A. Chinelatto, E. M. J. A. Pallone, A. M. Souza, M. K. Manosso, A. L.
Chinelatto and R. Tomasi, Sintering of Ceramics-New Emerging Techniques,
ed. A. Lakshmanan, Intech, 2012, ISBN: 978-953-51-0017-1.
P. Figiel, M. Rozmus and B. Smuk, J. Achievements Mater. Manuf. Eng., 2011,
48, 29.
Catalysis, 2016, 28, 2850 | 49
View Online
77
17:23:28.
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
Y. Chaiyabutr, R. Giordano and R. Pober, J. Biomed. Mater. Res., Part B, 2009,

88, 502.
H. X. Peng, Z. Fan and J. R. G. Evans, Ceram. Interfaces, 2000, 26, 887.
L. De Jonghe and M. Rahaman, Handbook of Advanced Ceramics, ed.
S. Somiya, Elsevier, 2003, ch. 4, p. 284.
P. Rao, M. Iwasa and I. Kondoh, J. Mater. Sci. Lett., 2000, 19, 543.
Z. He and J. Ma, Ceram. Interfaces, 2001, 27, 261.
N. Das and H. S. Maiti, J. Eur. Ceram. Soc., 1999, 19, 341.
G. R. Karagedov and N. Z. Lyakhov, Nanostruct. Mater., 1999, 11, 559.
G. Tar`, J. M. F. Ferreira and A. T. Fonseca, Ceram. Interfaces, 1999, 25, 577.
K. Mori, M. Ohashi and K. Osakada, Int. J. Mech. Sci., 1998, 40, 989.
M. Dressler, S. Reinsch, R. Schadrack and S. Benemann, J. Eur. Ceram. Soc.,
2009, 29, 3333.
J. G. Noudem, S. Lemonnier, M. Prevel, E. S. Reddy, E. Guilmeau and
C. Goupil, J. Eur. Ceram. Soc., 2008, 28, 41.
S. Callcut and J. C. Knowles, J. Mater. Sci. Mater. Med., 2002, 13, 485.
X. Yao, S. Tan, Z. Huang and D. Jiang, Ceram. Interfaces, 2006, 32, 137.
C. Voigt, C. G. Aneziris and J. Hubalkova, J. Am. Ceram. Soc., 2015, 98, 1460.
L. Montanaro, Y. Jorand, G. Fantozzi and A. Negro, J. Eur. Ceram. Soc., 1998,
18, 1339.
U. F. Vogt, M. Gorbar, P. Dimopoulos-Eggenschwiler, A. Broenstrup,
G. Wagner and P. Colombo, J. Eur. Ceram. Soc., 2010, 30, 3005.
Y. Kato and M. Fujimoto, GB Pat., 2 168 337 (A), 1986.
I. K. Jun, Y. M. Kong, S. H. Lee, H. E. Kim, H. W. Kim and K. C. Goretta,
J. Am. Ceram. Soc., 2006, 89, 2317.
N. Claussen, S. Wu and D. Holz, J. Eur. Ceram. Soc., 1994, 14, 97.
W. R. Pearson, WO Pat., 9959819 (A1), 1999.
B. B. Burton, US Pat., 3 111 396 (A), 1963.
J. C. Yarwood, J. E. Dore and R. K. Preuss, US Pat., 4 075 303 (A), 1978.
J. Saggio-Woyaansky and C. Scottetal, Am. Ceram. Soc. Bull., 1992, 71, 1674.
R. Brezny and D. J. Green, J. Am. Ceram. Soc., 1989, 72, 1145.
B. Ma, Y. Li, G. Liu and D. Liang, Ceram. Interfaces, 2015, 41, 3237.
R. Brezny and D. J. Green, Acta Metall. Mater., 1990, 38, 2517.
L. Rehorek, I. Dlouhy and Z. Chlup, Ceram.Silik., 2009, 53, 237.
W. Acchar, F. B. M. Souza, E. G. Ramalho and W. L. Torquato, Mater. Sci.
Eng., A, 2009, 513514, 340.
M. D. M. Innocentini, P. Sepulveda, V. R. Salvini, V. C. Pandolfelli and
J. R. Coury, J. Am. Ceram. Soc., 1998, 81, 3349.
E. A. Moreira, M. D. M. Innocentini and J. R. Coury, J. Eur. Ceram. Soc., 2004,
24, 3209.
C. C. Tseng, R. L. Sikorski, R. Viskanta and M. Y. Chen, J. Am. Ceram. Soc.,
2012, 95, 2015.
I. Y. Guzman, Glass Ceram., 2003, 60, 280.
J. Luyten, S. Mullens, J. Cooymans, A. M. De Wilde, I. Thijs and R. Kemps,
J. Eur. Ceram. Soc., 2009, 29, 829.
A. Bos, E. B. M. Doesburg and C. W. R. Engelen, Eurogel-91, ed. S. Vilminot,
R. Nass and H. Schmidt, Elsevier, 1992, p. 103.
M. D. Gasparyan, V. N. Grunskii, A. V. Bespalov, N. A. Popova,
T. A. Bagramyan, N. S. Grigoryan, A. A. Abrashov, M. B. Rozenkevich, Yu.
S. Pak, A. N. Bukin and A. S. Sumchenko, Glass Ceram., 2015, 71, 396.
50 | Catalysis, 2016, 28, 2850
Recent advances in the synthesis and

catalytic applications of atomically precise
gold nanoclusters
17:23:38.
Yuxiang Chen, Chenjie Zeng and Rongchao Jin*

DOI: 10.1039/9781782626855-00051
This chapter summaries the recent advances in the synthesis and catalytic application of
atomically precise Aun(SR)m nanoclusters. Structurally characterized nanoclusters can
serve as new model catalysts for obtaining atomic/molecular level insights into the
catalytic processes, including the precise size-dependent catalytic reactivity and how
molecules are adsorbed and activated on the catalytic active sites, as well as the structural
sensitivity of the catalyst to the reactions. While this area is still in its infancy, promising
work has been reported and demonstrated the catalytic power of atomically precise
nanoclusters. Such reactions include catalytic oxidation, chemoselective catalytic hydrogenation, catalytic semihydrogenation, etc. In addition, precisely doped nanoclusters
provide a unique opportunity to tune the catalytic reactivity on a truly atom-by-atom
basis. Overall, atomically precise nanoclusters hold great promise in the discovery of
unique catalytic processes as well as in advancing the fundamental understanding of
catalytic mechanisms at the atomic/molecular level.
Introduction
Catalysis based upon metal nanoparticles is of major importance in the

chemical industry. Precious metals such as platinum (Pt), palladium (Pd)
and recently gold (Au) have been widely studied.16 While metal nanoparticles can be used as unsupported (or free) catalysts in solution phase
reactions, the gas phase reactions requires supported catalysts. The
supported nanocatalysts are typically made by an impregnation method
(i.e. oxide supports soaked in a solution of metal salt, followed by drying
and high temperature calcination). The calcination is necessary to convert metal salt into metal particles, but unfortunately this thermal step
leads to polydisperse nanoparticles on the supports, and the polydispersity poses major challenges in fundamental studies of catalysis.2
First of all, polydisperse nanoparticles can only give an averaged catalytic
reactivity and the information of the precise size dependence cannot be
obtained. Second, the surface of the nanocatalystwhich is critically
importantis hard to be mapped out and thus remains poorly defined.
Third, the metal particles are inherently attached on the support and the
separate roles of metal particles and the support cannot be obtained,
especially for those cases that the metal/support interface is critical for
the catalytic reactions.
The colloidal deposition approach to metal nanoparticles has long
been utilized to prepare relatively monodisperse nanoparticles for catalytic investigation. But such an approach only partially solves the issues
Department of Chemistry, Carnegie Mellon University, 4400 Fifth Ave, Pittsburgh,
PA 15213, USA. E-mail: rongchao@andrew.cmu.edu
Catalysis, 2016, 28, 5185 | 51

c
17:23:38.
View Online
from the wet impregation and other conventional methods. The recent
development in colloid chemistry has oered excellent control over
particle size, shape and composition, but generally it is not completely
clear about what capping agents are present on the particle surfaces and
how they are bonded to the particle surfaces. Without knowing the surface structure and composition at the atomic level, it is hard to rationalize the catalytic mechanism and perform in-depth experimental and
theoretical studies on fundamental catalysis.
To overcome the major obstacles of nanoparticle catalysts in fundamental research, it is of paramount importance to creat well-defined
nanocatalysts with control at the atomic level, and this is indeed the
major goal in current catalytic research (e.g. see the strategic goals of the
U.S. Department of Energy). This major goal requies breakthroughs in
the chemical synthesis of nanocatalysts, the develepment of surface and
in situ characterization tools, and the computational eorts.
Opportunities have come up with the recent success in synthesizing
atomically precise gold nanoparticles.79 These unique nanoparticles
(often called nanoclusters) permit the crystallization (like oragnometallic
compounds in homogeneous catalysis) and structure determination by
X-ray crystallography. Among the nanoclusters, the thiolate or phosphine
protected gold nanoclusters are particularly promising in serving as the
model catalysts for achieving some fundamental understanding of
heterogeneous catalysis. Below we shall focus our discussion on thiolateprotected Aun(SR)m nanoclusters, where n refers to the precise number of
gold atoms and m to the number of ligands. A wide range of sizes with n
ranging from about ten to a few hundred atoms (equivalent size: subnanometer to 23 nm) have become available. These nanoparticles are
unique in that both the formula and atomic structure are well defined,
while conventional nanoparticles can only be represented by an average
size (e.g. 5 0.3 nm) and no exact formula can be determined. By single
crystal X-ray crystallography, every atom of the nanocluster (i.e. gold and
surface ligands) can be precisely pinned down. Based upon such total
structures of nanoclusters, a combined eort of experiment and theory
may achieve atomic/molecular level understanding of nanocatalysis.2
In this chapter, we first briefly discuss the synthesis of Aun(SR)m
nanoclusters, structural determination, and some general properties
relevant to catalysis, and then focus on the catalytic reactions that have
been reported to be catalyzed by such nanoclusters (with ligand-on or
o nanoclusters). The Aun(SR)m nanoclusters are expected to become a
promising class of model catalysts, in particular the attainment of X-ray
structures of nanoclusters permits precise reactivity-structure correlation. Although much work remains to be carried out, the catalysis
research of Aun(SR)m nanoclusters hold great promise in revealing the
atomic/molecular level details and significantly advance fundamental
understanding of catalytic mechanism, e.g. achieving the insight into
the size dependence and deep understanding of the molecular activation, catalytic active sites, and catalytic mechanism by correlation with
the structures of gold nanoclusters. Future research on atomically precise nanocluster catalysts will contribute to the fundamental catalysis
52 | Catalysis, 2016, 28, 5185
View Online
and the new design of highly selective catalysts for specific chemical
processes.
17:23:38.
2 Synthesis, structure, and properties of gold

nanoclusters
2.1 Synthetic strategies for obtaining atomically precise Aun(SR)m
nanoclusters and X-ray crystallographic determination of their total
structures
Herein we briefly discuss several synthetic methods of obtaining atomically precise Aun(SR)m nanoclusters.1013 The first ecient synthetic
approach to obtain gold nanoclusters with true monodispersity is called
size focusing methodology. In this method, an appropriate distribution of size-mixed nanoclusters is first prepared by kinetic control over
the reduction reaction of gold precursor (typically Au(I)-SR complexes or
polymers) with NaBH4 or other reducing agents, which is the key step for
the attainment of single-sized nanoclusters as the final product. Then,
these size-mixed nanoclusters are subjected to size-focusing under harsh
conditions (e.g., at 8090 1C and in the presence of excess thiol); under
such conditions the unstable nanoclusters are decomposed or converted
to the more stable ones, and eventually only the most stable size in the
initial size distribution survives the focusing process. The initial size
distribution should be properly controlled (for example, not too broad
otherwise, two or multiple stable sizes would be resulted after size
focusing). Another method to create atomically precise gold nanoclusters
is called ligand induced size transformation.12 In this method, an
existing magic sized gold nanocluster is used as a precursor and subsequently transformed to a new magic-sized nanocluster by reacting
with excess amount of dierent type of thiols (e.g., transformation
of Au38(SC2H4Ph)24 to Au36(SPh-t-Bu)24 by 4-tertbutybenzenthiol).12 Up-todate, a series of size-discrete, robust Aun(SR)m nanoclusters have been
available through intense research in the past years, such as Au25(SR)18,14
Au36(SR)24,15 Au38(SR)24,16,17 Au64(SR)32,18 Au102(SR)44,19 Au130(SR)50,13,20
Au133(SR)52,21 Au144(SR)60,17,22 Au333(SR)79,23 etc. A more complete account of the Aun(SR)m sizes can be found in a recent review.24 Herein, we
choose a few sizes to be used in catalytic tests for a more detailed
discussion.
2.1.1 The Au25(SR)18 nanocluster. The size-focusing phenomenon
was first observed in the synthesis of Au25(SR)18 in our early work.25,26
In this synthesis, the first step involves the spontaneous reduction of
Au(III) to Au(I) by thiol (typically 35 equivalents per mole of gold),
forming Au(I)-SR complexes or polymers (i.e. the intermediates). Subsequently, NaBH4 is added to further reduce Au(I)-SR to Au(0), producing
polydisperse nanoclusters protected by thiolate. Over the prolonged
aging period, a spontaneous size-focusing process was observed.
Initially, the polydisperse product gave a decay-like spectrum (note: the
absorption peaks of individual sizes of nanoclusters were washed out
due to polydispersity), but with time, we observed several prominent
Catalysis, 2016, 28, 5185 | 53
View Online
Fig. 1 Electrosray ionization (ESI) mass spectrum of the [Au25(SC2H4Ph)18] nanocluster.

The experimental formula weight is 7394 Da (the most abundant peak in the isotope
pattern; theoretical value: 7394). Reproduced with permission from ref. 25. M. Zhu,
E. Lanni, N. Garg, M. E. Bier and R. Jin, J. Am. Chem. Soc., 2008, 130, 1138. Copyright
(2008) American Chemical Society.
0.6
Abs (AU)
17:23:38.
peaks in the UV-vis spectrum of the crude product, indicating a gradual

growth of monodisperse nanoclusters. Three absorption peaks were
observed at 670, 450, and 400 nm, and the final product consists of molecularly pure Au25(SR)18 nanoclusters (Fig. 1 and 2).25,26 In recent synthetic eort, Yuan et al. reported a method for gram scale synthesis of
0.3
0.0
400
500
600
700
800
900
1000
1100
Wavelength (nm)
Fig. 2 Optical absorption spectrum of the Au25(SR)18 nanocluster.
54 | Catalysis, 2016, 28, 5185
17:23:38.
View Online
Au25 nanoclusters, even with adjustable ratios of two or three types of

thiolate ligands on the cluster surface.27 The surface functional groups
such as carboxyl, amine, and hydroxyl may be investigated in future
catalytic work.
The case of the Au25(SR)18 synthesis illustrates the basic principle of
the size-focusing method. This process is based upon the stability
property of dierent-sized nanoclusters.10 In order to attain atomic
monodispersity, we found that it is critical to control the size distribution
of the starting Aux(SR)y mixture so that only one specific size of
nanoclusters survives the size-focusing process. The importance of the
initial size distribution for the final molecular purity of nanoclusters is
also demonstrated in the case of Au38(SR)24 synthesis.28
The molecularly pure nanoclusters permit the growth of single crystals
for total structure determination of the nanoclusters. X-ray crystallography reveals that the Au25(SR)18 nanocluster comprises a Au13 icosahedral core (Fig. 3a, magenta) and a Au12(SR)18 shell (Fig. 3a, cyan and
yellow),14,29 consistent with the theoretical structure of Au25(SCH3)18.30
It is worth pointing out that some of the surface atoms are indeed
exposed, such as the triangular Au3 site as shown in the space-filling
Fig. 3 X-ray crystal structure of Au25(SCH2CH2Ph)18 nanoclusters, (a) ball-stick model,

(b) space-fill model (magenta: gold atoms of the core, cyan: gold atoms of the surface),
(c) dissection of the cluster into an Au13 core and six surface units. Adapted with
permission from ref. 9. H. Qian, M. Zhu, Z. Wu and R. Jin, Acc. Chem. Res., 2012, 45,
1470. and ref. 14. M. Zhu, C. M. Aikens, F. J. Hollander, G. C. Schatz and R. Jin, J. Am. Chem.
Soc., 2008, 130, 5883. Copyright (2012) American Chemical Society.
Catalysis, 2016, 28, 5185 | 55
17:23:38.
View Online
model (Fig. 3b). These sites (resembling pockets) may become the
catalytically active sites since they are readily accessible by reactant
molecules in catalytic reactions.
The electronic structure of Au25(SR)18 exhibits discrete energy levels
caused by the quantum-size eect.14 The molecular orbitals can be
roughly divided into the Au-core orbitals (i.e., primarily contributed by
the Au13 core) and the shell orbitals (i.e., primarily contributed by the
Au12(SR)18 part);14,31 this coreshell electronic picture is in line with the
geometric coreshell structure. Apparently, the core and surface Au
atoms possess dierent oxidation states, i.e. the surface Au atoms are
relatively electron-deficient and are formally Au(I), while the core Au
atoms are relatively electron-rich. Such a feature has been found to be
important in catalytic processes. The coreshell structure of Au25(SR)18
constitutes the origin of many interesting properties. For example,
MacDonald et al.32 observed an interesting interplay between the
metallic behavior of the Au13 inner core and the molecular behavior
of the six RSAuS(R)AuSR staple motifs.
2.1.2 The Au38(SR)24 nanocluster. The synthesis of Au38(SR)24 also
followed the size-focusing methodology.28 The first step was to obtain
a crude mixture (in this case, glutathionate (SG)-protected Aun(SG)m
nanoclusters). Then the size-mixed nanoclusters were subjected to a
thermal thiol etching process28,33 (e.g., 80 1C) in a two-phase (water/
toluene) system. During the harsh size focusing process of the starting
polydisperse nanoclusters are gradually converted to the most stable
product, i.e. the Au38(SC2H4Ph)24 nanocluster, and molecular purity
was achieved (Fig. 4).28 The optical spectrum of Au38(SR)24 is highly
Fig. 4 ESI-MS spectrum of the charge-neutral Au38(SC2H4Ph)24 nanoclusters. The

experimental formula weight of Au38(SC2H4Ph)24 is 10777.5 Da (theoretical value: 10778.05).
In mass spectrometric analysis, CsOAc was added to form positively charged [cluster-Cs]1
adducts. The small peak at 10777.5 (i.e. from the [Au38(SCH2CH2Ph)24]1) was due to an
induced ionization by Cs1 salt. Adapted with permission from ref. 28. H. Qian, Y. Zhu and
R. Jin, ACS Nano, 2009, 3, 3795. Copyright (2009) American Chemical Society.
56 | Catalysis, 2016, 28, 5185
17:23:38.
View Online
Fig. 5 UV-vis spectrum of phenylethanethiolate-protected neutral Au38(SR)24. Adapted

with permission from ref. 28. H. Qian, Y. Zhu and R. Jin, ACS Nano, 2009, 3, 3795.
Copyright (2009) American Chemical Society.
structured, and multibands are found at 1050, 745, 620, 560, 520, and
490 nm, with EgB0.9 eV, Fig. 5.
A key condition to obtain single-sized Au38(SC2H4Ph)24 nanoclusters is
to control the size distribution of the Aun(SG)m mixture prior to the sizefocusing step. In this case, we found that the solvent played an important
role in controlling the size range of the Aun(SG)m starting mixture.28
Acetone was found to produce an appropriate size distribution of Aun(SG)m
(that permitted high-yielding synthesis of Au38(SR)24). Comparison between acetone and methanol (solvents) showed that the acetone-mediated
synthesis of Aux(SG)y (the starting mixture) had a dominant size range
from 8 to 18 kDa, while the methanol system produced a dominant size
range below 8 kDa, i.e. smaller than Au38. The acetone-mediated high yield
synthesis of Au38(SC2H4Ph)24 is attributed to the down-conversion of those
higher-mass Aux(SG)y nanoclusters (38ono100), while such components
were missing in the methanol-mediated synthesis, hence, low yield of
Au38. It is worth noting that if the initial size range of Aux(SG)y should not
be too large, Au144(SR)60 nanoclusters could be produced, leading to a
mixture of Au38(SR)24 and Au144(SR)60 nanoclusters.
The atomic structure of the Au38(SCH2CH2Ph)24 nanocluster comprises
a biicosahedral Au23 core and six dimeric staples SRAuSRAuSR
(Au2(SR)3 for short) as well as three monomeric staples SRAuSR
(Au(SR)2 for short) protecting the Au23 core (Fig. 6).16 Apparently, the gold
atoms are distributed in two kinds of chemical environments (surface
and core), which benefits the catalytic reactions which often involve
adsorption and activation of more than one type of reactant molecules.
Catalysis, 2016, 28, 5185 | 57
17:23:38.
View Online
Fig. 6 Total structure of Au38(SC2H4Ph)24. The carbon tails (SC2H4Ph) are omitted for
clarity. (A) Au23 biicosahedral kernel; (B) Position of dimeric staples [Au2(SR)3] and monomeric staple [Au(SR)2]. (C) Side view and top view of the Au38(SR)24 total structure.
Losovyj et al.34 investigated the photoemission properties of Au38

clusters and found a strong resonant enhancement in photoemission
at the gold 4f(7/2) threshold and a weaker enhancement at the 4f(5/2)
threshold in the open (undressed) Au38 cluster, indicating an f-to-d
Coster-Kronig resonance transition but also s-d hybridization; note that
this resonant transition is forbidden in atomic gold due to the filled 5d10
system. The photoemission results indicate that the unusual catalytic
properties of ultrasmall gold nanoclusters such as Au38 are not a result of
localized orbitals.
2.1.3 The Au99(SR)42 nanocluster. This nanocluster was synthesized
via the size-focusing method. In the first step, polydispersed Au
nanoclusters protected by thiolate ligands (SPh) were prepared at the
thiol/Au ratio of 2.0 in dichloromethane. These polydispersed nanoclusters were then used as the precursor for the second step and the
polydispersed Au nanoclusters were converted to Au99(SPh)42 after size
focusing in the presence of excess thiol (HSPh) at 80 1C for overnight.
The UV-vis spectrum of Au99(SPh)42 shows absorption bands at 730,
600, 490, and 400 nm (Fig. 7A).35 Since no surface plasmon resonance
(SPR) band was found (typically 520 nm for spherical metallic Au
nanoparticles), the Au99(SPh)42 nanocluster is still in the non-metallic
regime. By extrapolating the optical absorbance to zero, the bandgap of
58 | Catalysis, 2016, 28, 5185
17:23:38.
View Online
Fig. 7 (A) Optical spectrum of the Au99(SPh)42 nanocluster dissolved in dichloromethane

(inset: the photon energy plot). (B) ESI mass spectrum of the Au99(SPh)42 product (denoted M).
(C) ESI mass spectrum of Au99(SPhCH3)42 (denoted M 0 ). (D) TGA curve (in N2 atmosphere)
of the Au99(SPh)42 nanocluster. In (B) and (C), k 1000. Adapted with permission from
ref. 35. G. Li, C. Zeng and R. Jin, J. Am. Chem. Soc., 2014, 136, 3673. Copyright (2014)
American Chemical Society.
Au99(SR)42 (R Ph, PhCH3) was determined to be B0.7 eV. The cluster

formula was determined by ESI-MS and is confirmed by the mass
dierence of 590.4 Da between the SPh and the SPh-Me protected
product (i.e. 590.4/14 42.1). The total structure of Au99(SPh)42 has not
been solved. The catalytic activity of Au99(SPh)42 in chemoselective
hydrogenation was investigated and compared with Au25(SPh)18 and
Au36(SPh)24 (see Section 3.2.3).
2.1.4 The Au144(SR)60 nanocluster. Both the Au25(SR)18 and
Au38(SR)24 cases explicitly demonstrate the importance of the size
distribution of the starting nanoclusters (i.e., prior to the size-focusing
step) for the eventual formation of monodisperse nanoclusters. The
first step is essentially kinetically controlled, while the second step (i.e.
size focusing) is essentially controlled by thermodynanics.11 To obtain
the proper size-distribution of the starting polydisperse Aux(SR)y
for the synthesis of monodisperse Au144(SR)60 nanoclusters,22 we
adjusted the gold salt to thiol ratio (Au/SR) to 1 : 3 in the first step and
as such, the size range of the starting Aux(SR)y mixture was controlled
to be around 2436 kDa. In the subsequent size-focusing step, the
initial polydispersed nanoclusters were incubated in large amounts of
thiol at 80 1C, and after 24 h of size focusing pure Au144(SC2H4Ph)60
nanoclusters were obtained (Fig. 8). Before size focusing, the initial
polydispersed nanoclusters showed a decay-like UV-vis spectrum; after
Catalysis, 2016, 28, 5185 | 59
17:23:38.
View Online
Fig. 8 ESI-MS spectrum of Au144(SC2H4Ph)60 nanoclusters. CsOAc was added to form

positively charged adducts for ESI analysis; note that the metal core charge is changing in
this case, i.e. the number of Cs1 ions adducted is not equal to the charge number (c.f. the
Au38(SC2H4Ph)24 case). The experimental molecular weight is 18299.12 36598.2
(theoretical: 36596.5). Adapted with permission from ref. 22. H. Qian and R. Jin, Nano
Lett., 2009, 9, 4083. Copyright (2009) American Chemical Society.
size focusing, step-like bands were observed at 510 and 700 nm and
these absorption bands are characteristic of Au144(SC2H4Ph)60
nanoclusters.22 The atomic structure of the Au144(SC2H4Ph)60 nanocluster has not been solved, as the larger nanoclusters pose major
challenges in crystallization. The optical spectrum of Au144(SC2H4Ph)60
shows bands at 400, 520 and 700 nm, Fig. 9.
2.1.5 Nanoclusters with face-centered cubic (FCC) structure. The
FCC structure was first discovered in the Au36(TBBT)24 nanocluster,15
where TBBT is the abbreviation of 4-tert-butylbenzenethiol. It has a 28gold-atom kernel and is based on cuboctahedral building blocks and
thus can be viewed as a fragment of the FCC structure. One can identify
the abca cubic-close-packing layers (Fig. 10A). In contrast, the
icosahedron-based Au23 kernel in the Au38(SR)24 nanocluster is of nonFCC structure. Alternatively, one may view the kernel as as a two-shelled
tetrahedral structure, i.e. an Au4 tetrahedron inside the Au24 cage. The
Au28 gold kernel exposes four {111} and six {100} facets, among which
the four {111} facets constitute the four faces of the truncated tetrahedral Au28 kernel, and six {100} facets on the six edges of the tetrahedron. The Au28 kernel is protected by four dimeric staples (SRAuSR
AuSR) and twelve bridging thiolates (SR). Each of the four dimeric
staples protects one {111} facet (Fig. 10B). On each {100} facet (comprising
two squares), two bridging thiolates can be identified (Fig. 10B). The
ESI-MS is shown in Fig. 11A and the optical spectrum in Fig. 11B.
Zeng et al. further synthesized the Au20(TBBT)16,36 Au28(TBBT)20,37
and Au44(SR)2838 nanoclusters. Together with Au36(TBBT)24, these four
60 | Catalysis, 2016, 28, 5185
17:23:38.
View Online
Fig. 9 UV-vis (A) and UV-vis-NIR spectrum of phenylethanethiolate-protected

Au144(SR)60. Redrawn with permission from ref. 22. H. Qian and R. Jin, Nano Lett., 2009,
9, 4083. Copyright (2009) American Chemical Society.
clusters form a magic series with a uniform size progression of

Au8(TBBT)4 and all of them adopt the FCC structure.
2.2 Thermal stability of Aun(SR)m nanoclusters
The thermal stability of Aun(SR)m nanoclusters is important for catalytic
applications as many reactions are run under thermal conditions.
Generally, the thiolate-protected gold nanoclusters are thermally robust
since they are synthesized by size focusing under harsh conditions. Our
motivation is to utilize the ligand-protected nanoclusters as catalysts so
Catalysis, 2016, 28, 5185 | 61
17:23:38.
View Online
Fig. 10 Total structure of Au36(SPh-t-Bu)24. The carbon tails (Ph-t-Bu) are omitted for
clarity. (A) The view of kernel structure based on polyhedron fusion and shell-by-shell
mode, respectively. (B) Surface thiolate protecting modes and the total structure of
Au36(SR)24.15
Fig. 11 (Left) ESI-MS and (Right) UV-vis spectrum of tert-butylbenzenethiolate-protected

Au36(SR)24. Adapted with permission from ref. 15. C. Zeng, H. Qian, T. Li, G. Li, N. L. Rosi,
B. Yoon, R. N. Barnett, R. L. Whetten, U. Landman and R. Jin, Angew. Chem., Int. Ed., 2012,
51, 13114. r Wiley-VCH.
that the atomic structures of nanoclusters can be correlated with the

catalytic properties, even though the presence of ligands may reduce the
catalytic activity. On the other hand, it is also worth comparing ligand-on
catalysts with the ligand-o ones. Thus, it is meaningful to find out the
temperature at which the surface ligands will be desorbed.
Thermogravimetric analysis (TGA) shows that Aun(SR)m (e.g., Au25(SR)24,
Au38(SR)24, and Au144(SR)60) starts to lose ligands at B200 1C and the ligand loss is complete at B250 1C (Fig. 12a).39 The ligand-loss temperature
was found to be unaected by the atmosphere (e.g., N2, air, O2, and H2).40
The isothermal stability of nanoclusters was further tested (i.e.,
maintained at 150 1C for 60 min in air atmosphere). Using Au25(SR)18 as a
62 | Catalysis, 2016, 28, 5185
17:23:38.
View Online
Fig. 12 (a) Thermogravimetric analysis of Aun(SR)m nanoclusters (under N2 atmosphere).

(b) Isothermal stability analysis of unsupported Au25(SR)18 nanoclusters (maintained at
150 1C, in air atmosphere for 60 mins). No discernable loss of ligands (starting: 100%) was
observed. The inset shows a zoom-in of the TGA curve at constant 150 1C. Reproduced
with permissions from ref. 39. Y. Zhu, H. Qian and R. Jin, Chem. Eur. J., 2010, 16, 11455,
and of Elsevier Inc. from ref. 41. G. Li, D.-e. Jiang, C. Liu, C. Yu and R. Jin, J. Catal., 2013,
306, 177. r Wiley-VCH.
typical example, we found no weight loss (i.e. no weight loss) during the
isothermal process (Fig. 12b).41 NMR, optical spectroscopy and mass
spectrometry analyses all confirm no change caused by the 150 1C
thermal treatment,42 which indicates the high stability of the gold
nanoclusters. This is reasonable since the structure of Au25(SR)18 is
thermodynamically stable according to density functional theory (DFT)
analysis, thus, structural isomerization should not occur, unless ligands
are lost at higher temperatures.39
2.3 Reactivity of Aun(SR)m nanoclusters with O2
Our interest in exploring Aun(SR)m for catalysis was inspired by the
observation of reversible conversion between [Au25(SR)18]0 and
[Au25(SR)18].43,44 The redox properties of nanoclusters are important
for those catalytic reactions that involve electron transfer in activating
reactants.
The native Au25(SR)18 nanoclusters from the synthesis are anionic
(i.e.,
[Au25(SR)18],
counterion tetraoctylammonium
(TOA1)).14
Interestingly, we found that when a solution of [Au25(SR)18] was
exposed to air, the nanoclusters were gradually converted to chargeneutral [Au25(SR)18]0 clusters,43 which was first discovered by the color
change of the solution (see also the UV-vis spectral changes in Fig. 13).
Single crystal X-ray crystallography36 revealed that the product is
[Au25(SCH2CH2Ph)18]0 and its structure shows the same framework as
that of [Au25(SCH2CH2Ph)18] but without the presence of tetraoctyl
ammonium (i.e. the counterion of the anionic cluster).
The attainment of [Au25(SCH2CH2Ph)18]0 and [Au25(SCH2CH2Ph)18]
provides an opportunity for deep understanding of the electronic
properties. Further work44 revealed that the negative charge of
[Au25(SCH2CH2Ph)18] surprisingly resides in the Au13 core of the cluster,
rather than on the Au12(SR)18 surface or on ligands. This explains why
Catalysis, 2016, 28, 5185 | 63
17:23:38.
View Online
Fig. 13 Optical spectra of [Au25(SR)18]q. Reproduced with permission from ref. 9. H. Qian,
M. Zhu, Z. Wu and R. Jin, Acc. Chem. Res., 2012, 45, 1470. Copyright (2012) American
Chemical Society.
the gold core, instead of the thiolate ligands, was first oxidized (i.e.
one-electron loss to yield neutral [Au25(SCH2CH2Ph)18]0). Normally, one
would expect that O2 would first oxidize the thiolate ligands on the
Au surface, which is indeed the case in bulk gold surfaces functionalized
by thiolates. The redox process between [Au25(SCH2CH2Ph)18]0 and
[Au25(SCH2CH2Ph)18] is completely reversible.44 While the interaction
of O2 with [Au25(SCH2CH2Ph)18] is quite slow at room temperature, it
speeds up at higher temperatures, and if using peroxide as the oxidant,
the anion-to-neutral conversion is much faster that using O2. At room
temperature, other sizes of nanoclusters such as Au36(SR)24, Au38(SR)24,
Au99(SR)42, Au144(SR)60 were not found to interact with O2.
Later work has also succeeded in the preparation of positively charged
[Au25(SR)18]1 nanoclusters. Its optical absorption spectrum is shown in
Fig. 13. The availability of various charge states of Au25(SR)18 permits the
investigation of the charge state eect (see Section 3.4).
Recent work by Kawasaki et al.45 reported the nanocluster-induced
generation of highly reactive singlet oxygen (1O2), which can be utilized in
catalytic oxidation. The 1O2 was eciently produced through the direct
photosensitization by Au25(SR)18 nanoclusters (SR phenylethanethiol,
SC2H4Ph, or captopril) via visible or near-IR irradiation (e.g. 532, 650, and
808 nm). 1O2 was successfully detected by direct observation of its characteristic emission around 1276 nm as well as by indirect methods of
using three dierent 1O2-selective probes.
64 | Catalysis, 2016, 28, 5185
View Online
17:23:38.
Catalytic properties of Aun(SR)m nanoclusters
3.1 Catalytic oxidation

The reactivity of Au25(SR)18 nanoclusters with O2 inspired us to pursue
the catalytic application of Au25(SR)18 and other Aun(SR)m nanoclusters in
catalytic oxidation.46,47 The activation of oxygen (O2) plays an important
role in the gold nanocluster-catalyzed selective oxidation processes.
Below we discuss the catalytic properties of nanoclusters in carbon
monoxide oxidation, styrene oxidation and sulfide oxidation.
3.1.1 Catalytic oxidation of carbon monoxide. The CO oxidation reaction has been extensively studied in the field of gold catalysis.1 An
interesting observation was the distinct eect of the support; among
the common supports, TiO2 was generally identified to be the best support for nanogold catalysts in the CO oxidation reaction.
Using the ligand-protected Aun(SR)m nanoclusters as catalysts for CO
oxidation to CO2, we found some phenomena32,40 that are distinctly
dierent from the case of bare nanogold catalysts. Starting with the
atomically precise, molecularly pure Au25(SR)18 nanoclusters, Nie et al.40
prepared several oxides as the supports for Au25(SR)18. Surprisingly, the
Au25(SR)18/TiO2 catalyst was found to have no catalytic activity even up to
200 1C (Fig. 14a), which is in direct contrast with the performance of
conventional Au(bare)/TiO2 catalysts.1 For other oxide supports, the
Au25(SR)18/Fe2O3 catalyst showed a moderate activity (onset temperature
100 1C), and CeO2 was found to be the best support (onset temperature
60 1C, 62% conversion at 160 1C). Without any activation, the activities
of the Au25(SR)18/MxOy catalysts are overall not high compared to the
conventional nanogold catalysts.
Interestingly, pretreatment of the Au25(SR)18/CeO2 catalyst in O2 for
1.52 h at 150 1C (denoted as Tpre) led to a drastic increase in catalytic
activity (onset temperature shifted to r.t.), (Fig. 14b), and 100% CO
conversion was reached at 100 1C. Prolonged O2 pretreatment at 150 1C
(e.g., 42 h) was also tested, but it was found not to further enhance the
catalytic activity (Fig. 14b). In addition, further increasing the Tpre to
250 1C (above the thiolate desorption temperature) did not show appreciable improvement in activity. Apparently, O2 thermal pretreatment had
activated the cluster catalyst. In terms of the activation mechanism, one
possibility would be the removal of ligands during the thermal pretreatment process, but we did not find evidence of thiolate desorption
through in situ mass spectrometry monitoring. Recent work by Wu and
coworkers48 reported partial removal of ligands through X-ray absorption
and IR analyses, but there is still a debate on whether the ligands are
removed or retained before/during the catalytic process.
It is worth noting that the drastic eect of thermal O2 pretreatment was
not observed in the Au25(SR)18/TiO2 system, nor in the Au25(SR)18/oxide
catalysts pretreated in N2 (as opposed to O2).40 Theoretical simulations
indicated that chemical adsorption of O2 on ligand-protected Au25(SR)18
does not occur;48,49 if so, the O2 interaction with Au25(SR)18 in solution
phase conversion of the anion to the charge neutral state should proceed
Catalysis, 2016, 28, 5185 | 65
17:23:38.
66 | Catalysis, 2016, 28, 5185

Fig. 14 (a) Catalytic activity of dierent oxide-supported Au25(SR)18/MOx catalysts for CO oxidation. Catalyst pretreatment condition: N2 at room temperature (r.t.)
for 0.5 h; reaction conditions: GHSVB7500 mL g1 h1, catalyst: 0.1 g. (b) Catalytic activity of Au25(SR)18/CeO2 after dierent pretreatments. (c,d) Eect of water
vapor on CO conversion over various pretreated Au25(SR)18/CeO2 catalyst under dierent pretreatment temperatures: (c) feed gases with water vapor and (d) without
vapor. Adapted with permission from ref. 40. X. Nie, H. Qian, Q. Ge, H. Xu and R. Jin, ACS Nano, 2012, 6, 6014. Copyright (2012) American Chemical Society.
17:23:38.
View Online
via a physical adsorption pathway, and therefore O2 activation on the

Au25(SR)18/oxide catalysts is expected to involve the oxide support.
The presence of water vapor in the feed gases exhibited a promotional
eect on the catalyst performance (Fig. 14c and d). For example, in
the case of feed-gas containing vapor, O2 pretreatment at even lower
temperature (e.g., Tpre 100 1C) could lead to the same drastic enhancement in activity.40 The result indicates that the surface OH species may
be involved in the CO oxidation.
The above results imply that the interface between Au25(SR)18 and CeO2
is critical and the perimeter sites should be the catalytic active-site.40 We
speculate that O2 should first convert to O2 by withdrawing an electron
from the Au25(SR)18 nanocluster, then migrates to the nanocluster/CeO2
interface and converts to hydroperoxide species on CeO2; the activated
CO should be oxidized at the perimeter sites of the catalyst (Scheme 1).
The mechanism of the catalytic CO oxidation by Au25(SR)18/CeO2 catalyst
still needs more detailed investigation in further work.
To rule out the possibility that O2 would burn o thiolate ligands at the
pretreatment temperatures, Nie et al.50 performed O2-TPO experiments
over Au38(SR)24/CeO2, in which the signals of O2 consumption, H2O and
CO2 generation were monitored by mass spectrometry (Fig. 15). A peak of
CO2 generation was observed at B100 1C, but this is due to the fact that
CeO2 powders capture CO2 in air, rather than burning of the thiolate
ligands on Au38(SR)24 clusters since no corresponding H2O peak was
observed at the same temperature, Fig. 15. A distinct O2 consumption
peak occurred at B230 oC, Fig. 15. The H2O generation signal (i.e.
positive peak) was observed at B175 1C (onset) and above was attributed
to H2O desorption from the catalyst surface and also possibly due to the
H2O product from O2 burning of gaseous, desorbed ligands at 4175 1C,
evidenced by a concurrent CO2 peak centered at B230 1C. Overall, the
H2O and CO2 generation process above 175 1C can be attributed to O2
burning of desorbed gaseous ligands, consistent with TGA results.
However, no burning of ligands was detected during the mild O2
pretreatment process since the pretreatment temperature is much
lower than 175 1C.
Ma et al.51 investigated the CO oxidation reaction using ligando Au25 and Au144 nanoclusters as catalysts. The ligands on
Au25(SCH2CH2Ph)18 and Au144(SCH2CH2Ph)60 nanoclusters were
Scheme 1 Proposed CO oxidation at the perimeter sites of Au25(SR)18/CeO2. Adapted

with permission of the American Chemical Society from ref. 40. X. Nie, H. Qian, Q. Ge,
H. Xu and R. Jin, ACS Nano, 2012, 6, 6014.
Catalysis, 2016, 28, 5185 | 67
View Online
MS signals (a.u.)
17:23:38.
H2 O
O2
CO2
100
200
300
400
500
Temperature (C)
Fig. 15 Mass spectrometric signals of O2-TPO profiles of Au38(SR)24/CeO2 catalyst. From
top to bottom: H2O generation (positive peak at 4175 1C), O2 consumption (negative peak
between 175310 1C), and CO2 generation (two peaks at 50150 1C and 175320 1C,
respectively). Adapted from ref. 50. X. Nie, C. Zeng, X. Ma, H. Qian, Q. Ge, H. Xu and R. Jin,
Nanoscale, 2013, 5, 5912 with permission of the Royal Society of Chemistry.
removed via thermal treatment at 300 1C in air (note: the ligand desorption temperature B200 1C, Fig. 12a). No sintering of the Au25 and
Au144 nanoclusters was found, as the ordered heterostructured mesoporous material supports (e.g., CuO-EP-FDU-12 and Co3O4-EPFDU-12)
can eectively stabilize the ligand-o Au25 and Au144 nanoclusters. The
ligand-o Au25 and Au144 clusters were characterized by Z-contrast
scanning transmission electron microscopy (STEM, where Z refers to the
atomic number of element) and no size growth was found compared to
the initial core size before ligand desorption. The average size of ligando Au144 clusters is determined to be 1.67 0.2 nm which is consistent
with the expected size (metal core) of the Au144(SCH2CH2Ph)60
nanoclusters, Fig. 16a. The ligand-o Au25 and Au144 nanoclusters
supported on both CuO-EP-FDU-12 and Co3O4-EP-FDU-12 are more active
than ligand-on catalysts in CO oxidation at relatively low temperatures
(e.g., 56% CO conversion at 20 1C and 100% conversion at 70 1C for the
Au25 cluster supported on Co3O4-EP-FDU-12, Fig. 16b). Thus, this thermal
treatment method can be applied when ligand-o gold cluster catalysts
are preferred.
Ligand-o Au38 clusters supported on oxides were also investigated for
CO oxidation.50 Gaur et al.52 observed that high activity wasnt observed
until surface thiolate ligands were fully removed by a high temperature
treatment. In contrast, work by Nie et al. showed that a mild pretreatment
(e.g. 130 1C) without removing ligands can largely increase the activity.50
The catalyst after complete removal of thiolate ligands by O2 pretreatment at 250 1C indeed gave a somewhat lower catalytic activity and lower
stability in long-time reaction process.50 We point out that the oxide
68 | Catalysis, 2016, 28, 5185
17:23:38.
View Online
Fig. 16 (a) Z-contrast STEM image of Au144 nanoclusters supported CuO-EP-FDU-12

after calcination at 300 1C. (b) Oxidation of CO as a function of temperature over
supported gold cluster catalysts. Adapted from ref. 51. G. Ma, A. Binder, M. Chi, C. Liu,
R. Jin, D.-e. Jiang, J. Fan and S. Dai, Chem. Commun., 2012, 48, 11413 with permission of
the Royal Society of Chemistry.
CH CH2
Au 25(SR)18
O
CH CH2
O2 or/and TBHP
Benzaldehyde Styrene epoxide
majority
O
CH3
Acetophenone
minority
Scheme 2 Selective oxidation of styrene catalyzed by Au25(SR)18 nanocluster catalysts.
support should play a major role and the dierent results between Nie
et al. and Gaur et al. may be caused by the dierent type of oxide supports
used in the tests. The Au38(SR) catalyst also exhibited some eects of
water vapor added into the feed gas, similar to the case of Au25(SR)18
catalysts.
3.1.2 Selective oxidation of styrene. We investigated solution-phase
styrene oxidation catalyzed by Aun(SR)m nanocluster catalysts (free and
SiO2-supported gold nanoclusters) using O2 as the oxidant, Scheme 2.39
The catalytic reaction was run at 80100 1C for 1224 h, which gave rise
to benzaldehyde as the major product (up to B70% selectivity) and
styrene epoxide (B25% selectivity) and acetophenone (o5% selectivity)
using Au25(SR)18 nanoclusters.
Using free (i.e. unsupported) Au25, Au38 and Au144 nanocluster catalysts, we observed a strong size-dependence in the catalytic oxidation
reaction of styrene, and the smaller Aun(SR)m nanocluster catalyst
exhibited much higher catalytic activity.39 The SiO2-supported Au25(SR)18
catalyst gave rise to comparable performance as that of the free
nanocluster catalyst, but the advantage of Au25(SR)18/SiO2 catalyst is its
recyclability for re-use in catalytic reaction; no apparent deterioration in
activity and selectivity of the reused supported catalyst was observed after
multiple cycles.39
Catalysis, 2016, 28, 5185 | 69
17:23:38.
View Online
Furthermore, Zhu et al. compared three oxidant systems using

Au25(SR)18/SiO2 catalyst: (A) tert-butyl hydroperoxide (TBHP) as the
oxidant; (B) TBHP as the initiator and O2 as the main oxidant; (C) O2 as
the oxidant (without initiator).47 System-A gave a high conversion of
styrene (e.g., 86%) with 100% selectivity for benzaldehyde since TBHP is a
highly reactive oxidant and easy to activate; system-B gave rise to a much
lower activity (25% conversion with 100% selectivity for benzaldehyde)
since O2 is much more dicult to activate than TBHP, and system-C was
even lower (18% conversion with 80% selectivity for benzaldehyde). The
catalytic results were consistent with the fact that TBHP is much more
reactive than O2. A similar phenomenon on the major eect of the
oxidant was also observed in the Au38(SR)24/SiO2 and Au144(SR)60/SiO2
catalysts.39 The results indicate that the activation of the oxidant is a key
step for achieving high conversion of styrene. The results also imply that
the active oxygen species in the catalytic cycles may be peroxide-like or
hydroperoxide-like species.39 Using CeO2 of dierent morphology (e.g.
spherical versus rod), Huang et al. observed that Au25 on the CeO2
nanorods showed higher selectivity towards styrene epoxide than Au25
on the surface of CeO2 nanoparticles,53 This morphology eect was
attributed to the dierent charge distribution and electron transfer
between Au25 and the surfaces of CeO2.
Liu et al. used synchrotron radiation-based X-ray absorption finestructure spectroscopy (XAFS) and ultraviolet photoemission spectroscopy (UPS) to characterize the Au25(SCH2CH2Ph)18 catalyst for
achieving correlation of the styrene-oxidation catalytic results with the
electronic properties of the cluster.54 Their results revealed that the
ligands act as not only stabilizers but also as d-band electron acceptor for
Au atoms. They proposed that the d-band unoccupied state population is
important for the catalytic activity and selectivity.54
Tsukuda and coworkers55 investigated Au25(SG)18/HAP catalyst for
selective oxidation of styrene, where HAP refers to hydroxyapatite
(Ca10(PO4)6(OH)2), Scheme 3. The ligand-o Au25/HAP catalyst was
Scheme 3 Epoxidation of styrene catalyzed by ligands-o Au25/HAP catalyst. Adapted

from ref. 55. Y. Liu, H. Tsunoyama, T. Akita and T. Tsukuda, Chem. Commun., 2010, 46,
550 with permission of the Royal Society of Chemistry.
70 | Catalysis, 2016, 28, 5185
17:23:38.
View Online
obtained by removing the ligands via 300 1C thermal treatment. The

catalyst was explored in styrene oxidation (with anhydrous tert-butyl
hydroperoxide (TBHP) as the oxidant) and the reaction yielded styrene
oxide with a selectivity of B92%, in contrast with the results obtained by
Zhu et al.39 in which the benzaldehyde (as opposed to the epoxide)
constitutes the major product.
Toward mechanistic understanding of the catalytic active site of the
thiolate-protected gold clusters in the styrene oxidation, Liu et al.56
performed DFT simulations on Au25(SR)18 as a benchmark model
system. They found that the Au atom in the staple motifs is the major
active site for the styrene oxidation in presence of TBHP as oxidant or
initiator. The formal oxidation state of the staple Au atom is Au(I). When
Au(I) (bi-coordinated) to Au(III) (tetra-coordinated), the O2 activation is
achieved during this cycle.
In terms of structure dependence, Zhu et al.57 compared the
catalytic activities of spherical Au25(SC2H4Ph)18 and rod-shaped
Au25(PPh3)10(SC2H4Ph)5Cl2 and found that the spherical cluster was
superior to the rod cluster for styrene oxidation. The higher catalytic
performance of the Au25(SC2H4Ph)18 was attributed to the coreshell
geometric structure that has an open exterior atomic shell and to the
electronic structure of an electron-rich Au13 core and an electrondeficient Au12 shell.
3.1.3 Selective oxidation of benzyl alcohol. Tsukuda and coworkers
probed the ligand eect of thiolated nanoclusters in the aerobic oxidation of benzyl alcohol (Scheme 4).58 They found that the presence of
thiolate ligands on Au25(SC12H25)18 (supported on porous carbon) increases the selectivity for benzaldehyde formation, albeit the activity
was reduced. The ligand-o cluster catalyst alters the product selectivity
for the benzoic acid product.58
Tuel and coworkers59 deposited Au25(SPh-p-NH2)17 clusters into the
mesopores of a silica support (SBA-15). Ligand removal by calcination at
400 1C led to very small gold nanoparticles (1.8 nm) with a narrow
distribution. They found that the ligand-on catalyst could not dissociate
O2, but exceptional activity was observed when radical initiators (e.g. alkyl
hydroperoxides) were added.
3.1.4 Selective oxidation of sulfides. Selective oxidation of sulfides
to sulfoxides (Scheme 5) is an important organic transformation and
the resulting sulfoxides are versatile intermediates for the preparation
of biologically and medicinally important products. The side-reaction
of sulfoxidation is the formation of sulfone. It is highly desirable to
achieve selective oxidation of sulfides to sulfoxides under mild
O
Au25/CNT
Ph
OH
O2
Benzyl Alcohol
Scheme 4
Ph
O
H
O
+
Ph
OH
Ph
Benzaldehyde Benzoic Acid
Ph
Benzoate
Benzyl alcohol oxidation catalyzed by nanoclusters.

Catalysis, 2016, 28, 5185 | 71
View Online
S
R1
O
O O
Au 25(SR)18/TiO2
S
S
PhIO
+
+
R2
R1 R2
DCM, 40C, 12h R1 R2
Sulfide
Sulfoxide
Sulfone
17:23:38.
Scheme 5 Oxidation of sulfides by iodosylbenzene using Au25(SR)18/TiO2 catalyst.
CH3
Au 25(SCH2CH2Ph)18
HCCl3,532 nm
Sulfide
O
S
O
CH3 +
Sulfoxide
O
S
Me
Sulfone
1
Scheme 6 Photocatalytic oxidation of sulfides to sulfoxides by O2 generated on the

photo-exited Au25(PET)18 catalyst.
conditions. In terms of bonding, sulfide has a much less binding

anity toward gold than thiolate, but sulfides can still be chemisorbed
onto the surface of gold nanoclusters via the interaction between the
sulfur atom (S) and the gold surface. Compared to thiolate and
sulfide, while sulfoxide (S(O)) can only weakly bind to the gold
surface; hence the conversion from sulfide to sulfoxide on the clusters
surface should be feasible.
Li et al.60 investigated the oxide-supported B1 wt% Au25(SR)18
nanocluster catalyst for selective oxidation of sulfides to sulfoxides
using iodosylbenzene (PhIO) as the oxidant (Scheme 5). The TiO2supported Au25(SR)18 nanocluster catalyst gave rise to a high catalytic
activity (e.g., B97% conversion of sulfide with B92% selectivity for the
sulfoxide product). Dierent types of sulfides have been investigated in
order to test the eects of electronic and steric factors in the selective
oxidation. The sulfides with electron-rich groups were oxidized with
higher conversions than those bearing electron-deficient groups, as well
as the sulfides with greater steric groups, showed lower conversions. The
support eect was also investigated, and the catalyst activity of
Au25(SR)18/oxide followed the order of TiO24Fe2O34CeO24MgO. The
oxide-supported Au25(SR)18 catalysts showed excellent recyclability in the
sulfoxidation process.60
Catalytic selective oxidation of sulfides to sulfoxides by 1O2 (Scheme 6)
was further demonstrated on the photoexcited Au25(SC2H4Ph)18
nanoclusters.45 When the reactions were carried out in open vessels in
dark with O2 bubbling, almost no sulfoxidation of sulfides was observed.
With the assistance of photoexcited Au25(SR)18 at 532 nm irradiation,
a B100% selective catalytic oxidation of sulfides to sulfoxides was
observed. The conversion increased with higher reaction temperatures or
longer reaction times. The B100% selectivity achieved in the sulfoxidation of sulfides was in contrast to the use of other oxidant such as PhIO
(87% only). Kawasaki et al.45 also examined the catalytic oxidation of
sulfides to sulfoxides by 1O2 in the presence of Au25(SR)18 using dierent wavelengths of photo-excitation (532, 650, and 808 nm). Nearly 100%
selectivity was achieved in the sulfoxidation of sulfide, irrespective of the
72 | Catalysis, 2016, 28, 5185
View Online
17:23:38.
excitation wavelength and the conversion is the largest for photoexcited

Au25(SR)18 upon 650 nm photo excitation, which is very close to the
absorbance at the HOMO-LUMO peak of Au25(SR)18.
3.2 Catalytic hydrogenation
3.2.1 Hydrogenation of nitrophenol or nitrobenzene. In addition to
catalytic oxidation processes, gold nanoclusters have also been explored
in hydrogenation reactions. Generally, gold is not active in activating
H2, but the gold core can store electrons (e.g. from NaBH4) and hence
catalyzes hydrogenation.
Yamamoto et al.61 compared the catalytic properties of Au25(SG)18
and dimethylformamide (DMF)-stabilized gold nanoclusters (note:
not of atomic precision) in the reduction reaction of 4-nitrophenol to
4-aminophenol by NaBH4 in water at room temperature, Scheme 7. The
reduction process was monitored by UV-vis spectroscopy. The Au25(SG)18
nanoclusters exhibited higher catalytic activity than DMF-capped clusters
in the nitrophenol reduction; the pseudo-first-order rate constant was
8103 s1 for Au25(SG)18 vs 3103 s1 for DMF-capped Au clusters.
High catalytic activity of Au25(SG)18 catalyst can be achieved at a very low
catalyst concentration, even at 1.0 mM. No induction time was observed in
the case of Au25(SG)18 despite SG being a strongly binding ligand due to
less steric hinderance, while the DMF-stabilized Au clusters exhibited an
induction time (B4000 s), which was attributed to the impedance from
the surface DMF layer on the clusters when reactants access the cluster
surface.61 For Au25(SG)18 nanoclusters, the unique coreshell structure
has less steric hinderance and thus renders the catalytic active sites more
accessible to reactants.
Shivhare et al.62 demonstrated intact Au25(SR)18 (where, R C6H13,
C8H17, and C12H25) nanoclusters for the reduction of nitrophenol by
NaBH4 in mixed THF/water and under N2 atmosphere. The Au25(SR)18
nanoclusters were stable and retained their structural integrity towards
NaBH4 treatment, while the larger AuB180(SC6H13)B100 nanoclusters were
not. Song et al. reported the use of selenolate-protected Au25(SePh)18 for
catalytic reduction of 4-nitrophenol to 4-aminophenol using NaBH4CN as
the reducing agent.63
Li et al. reported high activity of a Au44(PET)32 nanocluster at low
temperature (e.g. 0 1C).64 The new Au44(PET)32 cluster was converted from
the Au25(PET)18 and its catalytic activity was compared with other clusters
including Au25(PET)18, Au38(PET)24, Au44(TBBT)24, and Au144(PET)60. The
Au44(SC2H4Ph)32 was found to exhibit the highest catalytic activity at
room temperature (i.e. nearly twice that of Au25(PET)18). More importantly, even at 0 1C Au44(PET)24 still shows high activity (i.e. 97% yield of
4-aminophenol with 0.1 mol% (Au atom base)), while all other clusters
HO
NO2
Au 25(SG)18
NaBH4
HO
NH2
Scheme 7 Hydrogenation reaction of 4-nitrophenol to 4-aminophenol catalyzed by

Au25(SR)18 nanocluster catalysts.
Catalysis, 2016, 28, 5185 | 73
17:23:38.
View Online
exhibit essentially no activity at this low temperature. Since both

Au44(TBBT)24 and Au44(PET)32 are of the same metal core size, their large
dierence in activity indicates that the structure of Au44(PET)32 may be
critical. The Au44(PET)32 nanocluster was robust and showed no
degradation in the catalytic reaction, evidenced by the unchanged UV-vis
spectrum, and the clusters could be readily recovered by column chromatography for reuse in catalytic cycles without losing the activity. Future
eort of solving the structure of Au44(PET)32 will reveal the catalytic
dierence between the two types of Au44 clusters.
A closely relevant reaction, i.e. nitrobenzene hydrogenation to aniline
at 80 1C in a toluene solution with H2 as the reducing agent, has also been
investigated with gold nanoclusters as catalysts.65 Fang et al. focused
their investigation on the gold size and oxide support eects using
Au25 NCs as the precursor. The Au25(SR)18 nanoclusters protected by
6-mercaptohexanoic (MHA) were first made and then wet deposited onto
various inorganic supports, including hydroxyapatite (HAP), TiO2
(Degussa P25), activated carbon (AC), pyrolyzed graphene oxide (PGO),
and fumed SiO2. The catalysts were then calcined at 300 1C under a N2
stream to remove the surface ligands. The supports, HAP and P25, were
found to retain the Au25 size without obvious size growth owing to the
strong interactions between gold and supports, while the other three
supports led to severe aggregations of Au NCs, and thus the Au/HAP and
Au/P25 exhibited higher catalytic activity than the other supported catalysts. Fang et al. found that the removal of surface thiolate ligands from
the Au NCs is crucial for catalytic nitrobenzene hydrogenation. This is
owing to the fact that gold particles are not active in adsorbing and
activating H2 compared to other precious metals such as Pd, Pt and Rh.
3.2.2 Selective hydrogenation of aldehydes and ketones. In terms
of chemoselective hydrogenation of dierent functional groups (e.g.
aldehyde group versus nitro group or CC group), the Aun(SR)m
nanocluster catalysts were demonstrated to be capable of the selective
hydrogenation of a,b-unsaturated aldehydes and ketones to a,bunsaturated alcohols under mild conditions (60 1C, in mixed solvents
(1 : 1 toluene/acetonitrile)).46 Under the investigated reaction conditions, the Au25(SR)18 nanoclusters were found to preferentially hydrogenate the CO bond against the CC bond, and the product (i.e.,
unsaturated alcohol) was obtained with 76% selectivity, whereas 14%
selectivity was observed for unsaturated ketone, and 10% for saturated
alcohol.46 Further exploration of the catalytic reaction led to a nearly
complete selectivity for a, b-unsaturated alcohol in a mixed toluene/
ethanol (1 : 1) solvent at lower reaction temperatures (e.g. room temperature), albeit the conversion was not high (o20%).66 The Au25(SR)18
nanocluster catalyst was confirmed to remain intact after the hydrogenation reaction based upon the UV-vis and mass spectrometry analyses.
To enhance the conversion of a,b-unsaturated aldehydes and ketones,
oxide-supported nanocluster catalysts were investigated (including
Au25(SR)18/Fe2O3, Au25(SR)18/TiO2, and Au25(SR)18/SiO2), Table 1.66 Nearly
B100% selectivity for a,b-unsaturated alcohols was retained. A range of
74 | Catalysis, 2016, 28, 5185
View Online
Table 1 Chemoselective hydrogenation reaction of a,b-unsaturated ketones to a,bunsaturated alcohols catalyzed by Au25(SR)18 nanocluster catalysts. Adapted from ref. 66.
Y. Zhu, H. Qian, B. A. Drake and R. Jin, Angew. Chem., Int. Ed., 2010, 49, 1295 with
permission of r WILEY-VCH.
R3
R1
17:23:38.
Entry
Substrates
CHO
1
H3C
CHO
H3C
Au25(SR)18, r.t.
R2
R3
R1
H2, toluene/ethanol,
R2 UA
OH
Selectivity
for UA
Conversion (%)
Au25
Au25/Fe2O3
Au25/TiO2
100
38
49
46
100
43
53
51
100
44
54
53
100
39
52
52
100
29
45
44
46
47
47
CHO
H3C
CH3
H3C
CHO
H2C
H3C
H3C
H2C
CH3
O
CHO
91
The selectivity for unsaturated alcohol (UA) is 90% (free Au25(SR)18 catalyst, Au25), 92%
(Au25(SR)18/Fe2O3 catalyst, Au25/Fe2O3), and 91% (Au25(SR)18/TiO2 catalyst, Au25/TiO2),
respectively.
OHC
Au 99(SPh)42
NO2
HOH2C
80C, 20 bar H2
NH2 + HOH2C
NO2 + OHC
NH2
Scheme 8 Hydrogenation of 4-nitrobenzaldehyde to 4-nitrobenzyl alcohol using the

Au99(SPh)42 nanocluster as a catalyst.
substrates were tested and all gave rise to high selectivity for the
unsaturated alcohol product. It was found that the catalytic activity was
largely improved from less than 20% to 4050%. The catalytic activity of
the Au25(SR)18/Fe2O3 and Au25(SR)18/TiO2 catalysts was almost equivalent
(Table 1).
3.2.3 Selective hydrogenation of nitrobenzaldehyde derivatives.
In recent work, Li et al. synthesized thermally robust Au99(SPh)42
nanoclusters and investigated catalytic application in chemoselective
hydrogenation of nitrobenzaldehyde derivatives (Scheme 8).35 The
Au99(SPh)42 nanocluster is not yet in the metallic state and its
HOMO-LUMO gap (or band-gap) was determined to be B0.7 eV.35
Ceria-supported Au99(SPh)42 was utilized as a catalyst for chemoselective hydrogenation of nitrobenzaldehyde to nitrobenzyl alcohol in
water using H2 gas as the hydrogen source.35 The catalytic hydrogenation reaction was carried out under mild conditions (80 1C for 12 h).
The B100% selective hydrogenation of the aldehyde group came as a
Catalysis, 2016, 28, 5185 | 75
View Online
17:23:38.
67
surprise, because conventional nanogold catalysts instead give rise to

the reduction product of the nitro group.
A range of nitrobenzaldehyde derivatives were tested on the
Au99(SPh)42/CeO2 catalyst and high activity and complete selectivity for
the reduction of the aldehyde group were obtained (Table 2). From the
substrates, the catalytic activity of the Au99(SPh)42/CeO2 catalyst is
considerably influenced by the electrophilic aromatic substitution,
instead of the electron-rich/deficient side groups. The supported catalysts
show excellent recyclability in the reaction.35
We further test the size-dependent catalytic performance of Au25, Au36,
Au38 and Au99 nanoclusters.35,68 High activity and selectivity were
obtained. Based on the structures of Au25(SPh)18 and Au36(SPh)24, DFT
modeling of the 4-nitrobenzaldehyde adsorption68 shows that both the
CHO and NO2 groups are adsorbed on the SAuS staples of the gold
nanocluster (Fig. 17). The adsorption of the 4-nitrobenzaldehyde
molecule on the dierent-sized Aun(SR)m nanoclusters are moderately
strong and similar in strength. The DFT results suggest that the catalytic
activity of the Aun(SR)m nanoclusters is primarily determined by the
surface areas of the Au nanoclusters, consistent with the observed trend
of the conversion of 4-nitrobenzaldehyde versus the cluster size.
Table 2 Chemoselective hydrogenation of a range of substrates with nitro group and
aldehyde group using Au99(SPh)42/CeO2 as the catalyst. Reaction conditions: same as
noted in Table 3. Adapted with permission from ref. 35. G. Li, C. Zeng and R. Jin, J. Am.
Chem. Soc., 2014, 136, 3673. Copyright (2014) American Chemical Society.a,b
R'
CHO
O 2N
Entry
Substrate
OHC
pyridine, 80 o C, 12h
CH2OH
O 2N
Product
NO 2
HOH2C
OHC
R'
Au 99(SPh)42/CeO 2, H2O
NO 2
O 2N
3
HOH2C
NO 2
OH
4
5
OHC
100
73.2
100
98.5
100
84.1
100
90.0
100
98.9
100
99.5
100
NO 2
NO 2
HOH2C
NO 2
NO 2
NO 2
HOH2C
H3C
H3C
6
NO 2
Cl
HOH2C
NO 2
Cl
7
OHC
93.1
OH
OHC
OHC
Select. (%)
HOH2C
O2N
OHC
Conv. (%)
NO 2
HOH2C
NO 2
a
Reaction conditions: Au99(SPh)42/oxide (100 mg, 1 wt% loading) catalysts in 1 mL H2O,
0.05 mmol 4-nitrobenzaldehyde, 0.1 mmol pyridine, 20 bar H2, 80 1C, 12 h.
b
The conversion (Conv.) of 4-nitrobenzaldehyde and selectivity for 4-nitrobenzyl alcohol
were determined by NMR analysis. n.d. not detected; The conversion and selectivity values
are averaged over three measurements (error bar: B0.5%).
76 | Catalysis, 2016, 28, 5185
17:23:38.
View Online
Fig. 17 The 4-nitrobenzaldehyde molecule adsorbed on Au25(SCH3)18 nanocluster in

top view (a) and side view (b), and on Au38(SCH3)24 nanocluster in top view (c) and side
view (d). Reproduced with permission from ref. 68. G. Li, D.-e. Jiang, S. Kumar, Y. Chen and
R. Jin, ACS Catal., 2014, 4, 2463. Copyright (2014) American Chemical Society.
Overall, experiment combined with theory oers the molecular

insight into the hydrogenation of 4-nitrobenzaldehyde and the active site
structure on gold nanocluster catalysts.
3.2.4 Semihydrogenation of alkynes. The oxide-supported spherical
Au25(SR)18 nanocluster was explored for the catalytic semihydrogenation of alkynes to alkenes using H2 under relatively mild conditions
(100 oC, 20 bar H2).69 The terminal chain alkynes (e.g. 1-hexyne,
1-octyne, and 1-decyne) and terminal aromatic alkyne (e.g. phenylacetylene) all gave high conversion (4 99%) and high selectivity for the
alkene product (yield499%), Table 3. In contrast, the Au25(SR)18/TiO2
catalyst exhibited no activity for internal alkynes and the conversion for
all the tested internal alkynes was below 1% (Table 4, entries 24).
We proposed a deprotonation activation mechanism for alkynes on the
ligand-protected nanocluster catalyst.69 The terminal alkynes possess a
terminal hydrogen (HCC), but internal alkynes have no terminal
hydrogen and thus cannot be activated by the Au25(SR)18 nanoclusters.
We furthermore tested the ligand-o catalyst (Au25/TiO2),69 which was
prepared by thermally treating the Au25(SR)18/TiO2 powder at 300 oC for 1 h.
Catalysis, 2016, 28, 5185 | 77
View Online
a
Table 3 Semihydrogenation of terminal alkynes using Au25(SR)18/TiO2 catalyst. Adapted

with permission from ref. 69. G. Li and R. Jin, J. Am. Chem. Soc., 2014, 136, 11347.
Copyright (2014) American Chemical Society.
17:23:38.
R'
Entry
Au25(SR)18/TiO2
Pyridine, EtOH/H2O
100C, 20 bar H2, 20h
R'
Alkyne
Alkene
Conv. [%]
99.2
B100
2
3
99.5
99.6
99.4
a
Reaction conditions: 100 mg Au25(SR)18(1wt%)/TiO2 catalyst, 0.1 mmol alkynes, 0.2 mmol
pyridine, 1 mL EtOH/H2O (10:1, V/V), 100 1C, 20 bar H2, 20 h.
Table 4 Comparison of conversion in semihydrogenation catalyzed by ligand-on and

ligand-o catalysts.a,b Adapted with permission from ref. 69. G. Li and R. Jin, J. Am.
Chem. Soc., 2014, 136, 11347. Copyright (2014) American Chemical Society.
R1
R2
Au25/TiO2
Pyridine, EtOH/H2O
100C, 20 bar H2, 20h
R1
R2
Entry
Catalyst
R1
R2
Conv.
Sel.
1
2
3
4
5
6
7
8
Ligand-on Au25
sphere
PhC2H4
Ph
Ph
n-C6H13
PhC2H4
Ph
Ph
n-C6H13
H
CH3
Ph
CO2CH3
H
CH3
Ph
CO2CH3
499
o1
o1
o1
95.6
52.8
59.7
52.6
97
499
99
Reaction conditions: as noted in Table 3.

Stereoselectivity (Sel.) for Z-alkene.
Ligand-o Au25
The bare Au25/TiO2 catalyst indeed catalyzed the semihydrogenation

reactions of internal alkynes including 1-phenyl-1-propyne, diphenylacetylene, and methyl 2-nonynoate (Table 4, entries 68), with 52.6 to
59.7% conversions of internal alkynes and 497% stereoselectivity for
Z-alkene. Thus, the presence of ligands on Au25 inhibit the activation of
internal alkynes but not to the terminal alkynes.
With respect to the support eect, blank experiments using gold-free
TiO2 gave no conversion, thus the catalytic eect is associated with the Au
nanocluster. To further investigate the potential eect of the oxide supports, four dierent oxides (i.e., CeO2, SiO2, TiO2, and Al2O3) were used as
supports for the Au25(SR)18 nanocluster.69 No apparent dierence in the
four types of Au25(SR)18/oxide catalysts was observed in the catalytic
78 | Catalysis, 2016, 28, 5185
17:23:38.
View Online
Fig. 18 (A) The atomic structure of Au25(SR)18 nanoclusters. (B) The terminal alkyne
molecule adsorption on the surface of the Au25(SC2H4Ph)18 nanocluster (shown in spacefill mode). Adapted with permission from ref. 69. G. Li and R. Jin, J. Am. Chem. Soc., 2014,
136, 11347. Copyright (2014) American Chemical Society.
semihydrogenation reaction of 4-phenyl-1-butyne (i.e., all gave high

conversion and selectivity around 99%).
The above results imply that the activation process of the terminal
alkynes on Au25(SR)18 nanocluster catalyst is dierent from the case of
bare Au25 nanocluster and conventional gold nanoparticle catalysts. The
structure of the Au25(SR)18 nanocluster shows a 13-atom icosahedral Au13
core protected by six dimeric staple motifs (Fig. 18A). Two open facets of
triangular Au3 are seen in the structure (Fig. 18B), with the exposed gold
atoms from the three adjacent staples. These sites may constitute the
catalytic sites. Previous theoretical simulation indicated the adsorption
of phenylacetylene on such Au3 sites on the surface of Au25(SR)18
nanocluster,41 and the an adsorption energy was computed to be
0.40 eV. We proposed that the semihydrogenation reaction involves the
following initial adsorption configuration: the phenyl ring of phenylacetylene faces one of the external Au atoms of the triangular Au3 site, and
the CCH group of phenylacetylene points to a second Au atom of the
Au3 site (Fig. 18B). In such a configuration, the third Au atom of the Au3
site is unoccupied and thus available for possible hydrogen adsorption
and activation facilitated by pyridine.
3.2.5 Catalytic conversion of propargylic acetates to a,b-unsaturated
ketones or aldehydes. Li et al.70 reported the utility of Au25(PET)18 in
catalytic conversion of propargylic acetates to a,b-unsaturated ketones
or aldehydes (Scheme 9 and Table 5). For the size eect, Au25(SR)18 was
found to be superior than Au38(SR)24, Au144(SR)60, B3 nm and B23 nm
gold nanoparticles. The extraordinary activity of Au25(SR)18 is probably
associated with the anionic nature. The presence of ligands seems not
to aect the activity. The catalytic process has been proposed to be initialized by an SN2 0 addition of OH. A dramatic solvent eect was observed. The cluster was confirmed to be stable in the catalytic reaction
Catalysis, 2016, 28, 5185 | 79
View Online
17:23:38.
Scheme 9 Catalytic conversion of 1,3-diphenylprop-2-ynyl acetate to (E)-chalcone over

Au25(SR)18 in DMSO/H2O and the presence of K2CO3 at 80 1C.
Table 5 Conversion of propargylic acetates to a,b-unsaturated

ketones or aldehydes catalyzed by Au25(SR)18.a
Entry
R1
R2
Yield (isolated)
1
2
3
4
Ph
Ph
H
H
Ph
Me
Ph
Me
76%
78%
75%
70%
Reaction conditions: propargylic acetate (0.20 mmol), Au25(SR)18

(3.0 mg), K2CO3 (0.20 mmol), DMSO (2.0 mL), H2O (0.2 mL), 80 1C, 2 h.
a
by the observed of identical signals as those of the fresh catalyst in

UV-vis and mass spectrometry analyses.
3.3 Nanoclusters as electron transfer catalysts
In a recent report, Chong et al.71 Au25(SC2H4Ph)18 nanoslusters were
employed as electron transfer catalysts. The catalytic reduction was
conducted in THF at 20 1C for 2 h, with NaBH4 as the reductant. The
nanocluster catalyzed an intramolecular cascade reaction of 2-nitrobenzonitrile under ambient conditions with high conversion (87%) and
selectivity (96%). Based on the structure and redox states of Au25(PET)18,
mechanistic understanding of the catalytic reaction was achieved. Specifically, electron spin-resonance spectra reveled the consecutive electron
transfer from the anionic Au25(SR)18 to the reactant and the formation of
N radical (Scheme 10). The radical could be readily transformed into a
five-member heterocyclic intermediate and finally to 2-(hydroxyamino)
benzonitrile as the product. A range of reactants were tested and The
Au25(SR)18 nanoclusters were confirmed to be intact after the catalytic
circles, evidenced by the unchanged UV-vis absorption spectrum and
mass spectrometry analysis.
3.4 Catalytic eect of [Au25(SR)18]q charge state
The majority of Aun(SR)m nanoclusters are charge neutral, but exceptions
are [Au25(SR)18] and [Au23(SR)16],72 both of which carries a 1 charge
for their native state. Since [Au25(SR)18] was obtained earlier, much work
has been carried out on [Au25(SR)18] but not on the [Au23(SR)16]. As
discussed in Section 2.3, [Au25(SR)18] readily loses one electron and
converts to [Au25(SR)18]0. A controlled oxidation can further lead to
[Au25(SR)18]1.43 The neutral state is paramagnetic due to the presence of
80 | Catalysis, 2016, 28, 5185
17:23:38.
View Online
Scheme 10 Au25(SR)18 as an electron transfer catalyst for intramolecular cascade reaction in 2-nitrobenzonitrile with the mechanistic steps shown in (1) and (2). Ligands are
omitted in the cluster formula.
Table 6 Reaction turnover frequencies (TOF: molecules/Au25q/second) determined

from polarization curves at the indicated potentials; standard deviations are from three
runs with freshly deposited Au25q/CB samples.
TOF
(molec./Au25q/s)
CO2 Reduction
(1.0 V)
CO Oxidation
(0.89 V)
O2 Reduction
(0.5 V)
Au25
Au250
Au251
95.2 5.7
61.1 2.5
40.1 1.9
57.4 3.9
93.2 4.7
133.6 5.9
25.3 1.5
20.0 1.7
16.0 0.5
one unpaired electron, while both the anionic and cationic states are
diamagnetic.
In terms of the catalytic eect of the [Au25(SR)18]q, Kauman et al.
found distinct charge state-dependent electrocatalytic activity for CO2
reduction, CO oxidation and O2 reduction reactions in aqueous media.73
The results show that the anionic cluster favors the CO2 electroreduction
as well as the O2 electroreduction, while the electrooxidation of CO
requires the cationic Au25 (Table 6). A combined study of experiment and
density functional theory simulation identified a relationship between
the catalytic reactivity of Au25(SR)18q charge state and the adsorption of
reactants or products. Specifically, the anionic [Au25(SR)18] promoted
CO2 electroreduction by stabilizing co-adsorbed CO2 and H1 reactants.
In contrast, the cationic Au251 cluster promoted CO electroxidation by
stabilizing co-adsorbed CO and OH reactants. In the ORR, stronger
adsorption of OH product was found on Au251, which inhibited O2 reduction rates.
The charge state-mediated reactivity may be extended to a wide range
of applications, including fuel cells, water splitting, batteries, and sensors, because of the participation of H1 and OH in aqueous electrocatalytic reactions (Scheme 11). The charge state eect of Au25q and Au23
nanoclusters are also worth being investigated in other reaction systems,
such as photocatalysis and thermocatalysis. The well-defined charge state
of nanoclusters may create new opportunities for tuning reactant,
intermediate and product interactions with the catalysts.
Catalysis, 2016, 28, 5185 | 81
View Online
+2e
CO2 + 2H+
CO + H2O
-2e- CO + H O
CO + 2OH-
2
2
-
+4e 4OHO2 + 2H2O
17:23:38.
+2e
2OOH- + OHO2 + H2O
Scheme 11 Eects of the cluster charge state in electrocatalytic processes.
Lu et al.74 reported a correlation between the Au25(SC12H25)18 charge

state (q 1, 0 and 1) and the electrocatalytic activity for the production of H2O2 from O2 in alkaline media. The Au25 nanoclusters
passivated by short-chain phenylethylthiol ligands instead favored a
4-electron reduction of O2 to H2O (as opposed to H2O2). For the charge
states of Au25(SC12H25)18, the negative charge state was found to give a
higher H2O2 production (B90%) than the neutral and cationic ones,
which was attributed to the more ecient electron transfer from the
anionic Au25 cluster into the LUMO (p*) of O2, leading to the activation
of O2 molecules and generation of peroxo-like species. Accelerated
durability tests (ADTs) that were performed in O2-saturated 0.1 M KOH
solution by applying a cyclic potential sweep between 0.60 and 0.00 V (vs.
SCE) demonstrated high stability of the Au25(SC12H25)18, as the characteristic absorption features of the clusters were retained after ADTs.
4 Summary
In this chapter, we have discussed the catalytic application of atomically
precise Aun(SR)m nanoclusters, including (i) catalytic oxidation (such as
selective oxidation of styrene to benzaldehyde or styrene epoxide, selective oxidation of sulfides to sulfoxides, selective oxidation of benzyl
alcohol to benzaldehyde, and catalytic CO oxidation to CO2), (ii) catalytic
hydrogenation (such as nitrophenol to aminophenol, chemoslective
hydrogenation, semihydrogenation), (iii) electron transfer catalysis, and
(iv) electrocatalysis.
Overall, the Aun(SR)m nanoclusters possess several distinct features
that are of particular interest to catalysis compared to conventional
metallic Au nanoparticle catalysts. First of all, metallic gold nanoparticles
(2 to 100 nm) adopt a face-centered cubic (FCC) structure, but Aun(SR)m
nanoclusters often adopt dierent types of atom-packing structures;9
such structures oer an opportunity to probe the structural sensitivity in
catalysis. Second, the ultrasmall size not only increases the surface-tovolume ratio, but also induces strong electron-energy quantization in
nanoclusters (as opposed to the continuous conduction band in metallic
gold nanoparticles);8 the quantization eect and the high surface-tovolume eect may exhibit dierent behavior. Third, some unique
structures have been discovered in the crystal structures of nanoclusters,
such as the staple-like motifs and volcano-like atomic sites on cluster
surfaceswhich can act as catalytically active centers in chemical
reactions.2
82 | Catalysis, 2016, 28, 5185
17:23:38.
View Online
Future work should investigate the following aspects, (i) the size dependence, the available sizes would allow for detailed studies on the
nanocluster size eect; (ii) the tunable characteristic of bimetallic
nanoclusters75 where are expected to further improve the catalytic reaction conditions toward mild temperatures, and to enhance catalytic
activity, selectivity, and stability for specific reaction processes; (iii) It is
also worth further investigation on the catalytic application of such
nanoclusters in electrochemical catalysis,76 and photocatalysis;7780
(iv) mechanistic studies, the well-defined nanoclusters provide exciting
opportunities for gaining mechanistic insight by combining experimental and theoretical eorts. The detailed characterization of the
supported nanocluster catalysts should be strengthened and issues such
as charge transfer should be investigated. As a new class of catalysts,
Aun(SR)m nanoclusters will bridge the gap between organometallic
complex-based homogeneous catalysis and nanoparticle-based heterogeneous catalysis, and also bridge the gap between the bulk crystal
structure model catalysts and real-world catalysts.
Acknowledgements
The work was supported by U.S. Department of Energy-Oce of Basic
Energy Sciences (Grant DE-FG02-12ER16354).
References
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
M. Haruta, S. Tsubota, T. Kobayashi, H. Kageyama, M. J. Genet and

B. Delmon, J. Catal., 1993, 144, 175.
G. Li and R. Jin, Acc. Chem. Res., 2013, 46, 1749.
S. Yamazoe, K. Koyasu and T. Tsukuda, Acc. Chem. Res., 2014, 47, 816.
R. Jin, Nanotechnol. Rev., 2012, 1, 31.
G. J. Hutchings and C. J. Kiely, Acc. Chem. Res., 2013, 46, 1759.
D. Astruc, F. Lu and J. R. Aranzaes, Angew. Chem., Int. Ed., 2005, 44, 7852.
Y. Negishi, K. Nobusada and T. Tsukuda, J. Am. Chem. Soc., 2005, 127, 5261.
R. Jin, Nanoscale, 2010, 2, 343.
H. Qian, M. Zhu, Z. Wu and R. Jin, Acc. Chem. Res., 2012, 45, 1470.
R. Jin, H. Qian, Z. Wu, Y. Zhu, M. Zhu, A. Mohanty and N. Garg, J. Phys. Chem.
Lett., 2010, 1, 2903.
Z. Wu, M. A. MacDonald, J. Chen, P. Zhang and R. Jin, J. Am. Chem. Soc.,
2011, 133, 9670.
C. Zeng, C. Liu, Y. Pei and R. Jin, ACS Nano, 2013, 7, 6138.
Y. Chen, C. Zeng, D. R. Kauman and R. Jin, Nano Lett., 2015, 15, 3603.
M. Zhu, C. M. Aikens, F. J. Hollander, G. C. Schatz and R. Jin, J. Am. Chem.
Soc., 2008, 130, 5883.
C. Zeng, H. Qian, T. Li, G. Li, N. L. Rosi, B. Yoon, R. N. Barnett, R. L. Whetten,
U. Landman and R. Jin, Angew. Chem., Int. Ed., 2012, 51, 13114.
H. Qian, W. T. Eckenho, Y. Zhu, T. Pintauer and R. Jin, J. Am. Chem. Soc.,
2010, 132, 8280.
N. K. Chaki, Y. Negishi, H. Tsunoyama, Y. Shichibu and T. Tsukuda, J. Am.
Chem. Soc., 2008, 130, 8608.
C. Zeng, Y. Chen, G. Li and R. Jin, Chem. Mater., 2014, 26, 2635.
Catalysis, 2016, 28, 5185 | 83
View Online
19
20
17:23:38.
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
Y. Levi-Kalisman, P. D. Jadzinsky, N. Kalisman, H. Tsunoyama, T. Tsukuda,

D. A. Bushnell and R. D. Kornberg, J. Am. Chem. Soc., 2011, 133, 2976.
Y. Negishi, C. Sakamoto, T. Ohyama and T. Tsukuda, J. Phys. Chem. Lett.,
2012, 3, 1624.
C. Zeng, Y. Chen, K. Kirschbaum, K. Appavoo, M. Y. Sfeir and R. Jin, Sci. Adv.,
2015, 1, e1500045.
H. Qian and R. Jin, Nano Lett., 2009, 9, 4083.
H. Qian, Y. Zhu and R. Jin, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 696.
R. Jin, Nanoscale, 2015, 7, 1549.
M. Zhu, E. Lanni, N. Garg, M. E. Bier and R. Jin, J. Am. Chem. Soc., 2008,
130, 1138.
Z. Wu, J. Suhan and R. Jin, J. Mater. Chem., 2009, 19, 622.
X. Yuan, B. Zhang, Z. Luo, Q. Yao, D. T. Leong, N. Yan and J. Xie, Angew.
Chem., Int. Ed., 2014, 53, 4623.
H. Qian, Y. Zhu and R. Jin, ACS Nano, 2009, 3, 3795.
M. W. Heaven, A. Dass, P. S. White, K. M. Holt and R. W. Murray, J. Am. Chem.
Soc., 2008, 130, 3754.
kkinen and H. Gro
nbeck, J. Am.
J. Akola, M. Walter, R. L. Whetten, H. Ha
Chem. Soc., 2008, 130, 3756.
C. M. Aikens, J. Phys. Chem. Lett., 2010, 1, 2594.
M. A. MacDonald, D. M. Chevrier, P. Zhang, H. Qian and R. Jin, J. Phys.
Chem. C, 2011, 115, 15282.
H. Qian, M. Zhu, U. N. Andersen and R. Jin, J. Phys. Chem. A, 2009, 113, 4281.
Y. B. Losovyj, S.-C. Li, N. Lozova, K. Katsiev, D. Stellwagen, U. Diebold,
L. Kong and C. S. S. R. Kumar, J. Phys. Chem. C, 2012, 116, 5857.
G. Li, C. Zeng and R. Jin, J. Am. Chem. Soc., 2014, 136, 3673.
C. Zeng, C. Liu, Y. Chen, N. L. Rosi and R. Jin, J. Am. Chem. Soc., 2014,
136, 11922.
C. Zeng, T. Li, A. Das, N. L. Rosi and R. Jin, J. Am. Chem. Soc., 2013,
135, 10011.
C. Zeng, Y. Chen, G. Li and R. Jin, Chem. Commun., 2014, 50, 55.
Y. Zhu, H. Qian and R. Jin, Chem. Eur. J., 2010, 16, 11455.
X. Nie, H. Qian, Q. Ge, H. Xu and R. Jin, ACS Nano, 2012, 6, 6014.
G. Li, D.-e. Jiang, C. Liu, C. Yu and R. Jin, J. Catal., 2013, 306, 177.
Z. Wu and R. Jin, ACS Nano, 2009, 3, 2036.
M. Zhu, W. T. Eckenho, T. Pintauer and R. Jin, J. Phys. Chem. C, 2008,
112, 14221.
M. Zhu, C. M. Aikens, M. P. Hendrich, R. Gupta, H. Qian, G. C. Schatz and
R. Jin, J. Am. Chem. Soc., 2009, 131, 2490.
H. Kawasaki, S. Kumar, G. Li, C. Zeng, D. R. Kauman, J. Yoshimoto,
Y. Iwasaki and R. Jin, Chem. Mater., 2014, 26, 2777.
Y. Zhu, H. Qian, M. Zhu and R. Jin, Adv. Mater., 2010, 22, 1915.
Y. Zhu, H. Qian and R. Jin, J. Mater. Chem., 2011, 21, 6793.
Z. Wu, D.-e. Jiang, A. K. P. Mann, D. R. Mullins, Z.-A. Qiao, L. F. Allard,
C. Zeng, R. Jin and S. H. Overbury, J. Am. Chem. Soc., 2014, 136, 6111.
kkinen, Nat. Chem.,
O. Lopez-Acevedo, K. A. Kacprzak, J. Akola and H. Ha
2010, 2, 329.
X. Nie, C. Zeng, X. Ma, H. Qian, Q. Ge, H. Xu and R. Jin, Nanoscale, 2013,
5, 5912.
G. Ma, A. Binder, M. Chi, C. Liu, R. Jin, D.-e. Jiang, J. Fan and S. Dai, Chem.
Commun., 2012, 48, 11413.
S. Gaur, J. T. Miller, D. Stellwagen, A. Sanampudi, C. S. S. R. Kumar and
J. J. Spivey, Phys. Chem. Chem. Phys., 2012, 14, 1627.
84 | Catalysis, 2016, 28, 5185
View Online
53
54
17:23:38.
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
P. Huang, G. Chen, Z. Jiang, R. Jin, Y. Zhu and Y. Sun, Nanoscale, 2013,

5, 3668.
J. Liu, K. S. Krishna, Y. B. Losovyj, S. Chattopadhyay, N. Lozova, J. T. Miller,
J. J. Spivey and C. S. S. R. Kumar, Chem. Eur. J., 2013, 19, 10201.
Y. Liu, H. Tsunoyama, T. Akita and T. Tsukuda, Chem. Commun., 2010,
46, 550.
C. Liu, S. Lin, Y. Pei and X. C. Zeng, J. Am. Chem. Soc., 2013, 135, 18067.
Y. Zhu, H. Qian, A. Das and R. Jin, Chin. J. Catal., 2011, 32, 1149.
T. Yoskamtorn, S. Yamazoe, R. Takahata, J.-i. Nishigaki, A. Thivasasith,
J. Limtrakul and T. Tsukuda, ACS Catal., 2014, 4, 3696.
C. Lavenn, A. Demessence and A. Tuel, J. Catal., 2015, 322, 130.
G. Li, H. Qian and R. Jin, Nanoscale, 2012, 4, 6714.
H. Yamamoto, H. Yano, H. Kouchi, Y. Obora, R. Arakawa and H. Kawasaki,
Nanoscale, 2012, 4, 4148.
A. Shivhare, S. J. Ambrose, H. Zhang, R. W. Purves and R. W. J. Scott, Chem.
Commun., 2013, 49, 276.
Y. Song, J. Zhong, S. Yang, S. Wang, T. Cao, J. Zhang, P. Li, D. Hu, Y. Pei and
M. Zhu, Nanoscale, 2014, 6, 13977.
M.-B. Li, S.-K. Tian, Z. Wu and R. Jin, Chem. Commun., 2015, 51, 4433.
J. Fang, J. Li, B. Zhang, X. Yuan, H. Asakura, T. Tanaka, K. Teramura, J. Xie
and N. Yan, Nanoscale, 2015, 7, 6325.
Y. Zhu, H. Qian, B. A. Drake and R. Jin, Angew. Chem., Int. Ed., 2010, 49, 1295.
ckner, C. Mohr and H. Hofmeister, J. Am. Chem. Soc., 2000,
P. Claus, A. Bru
122, 11430.
G. Li, D.-e. Jiang, S. Kumar, Y. Chen and R. Jin, ACS Catal., 2014, 4, 2463.
G. Li and R. Jin, J. Am. Chem. Soc., 2014, 136, 11347.
M.-B. Li, S.-K. Tian and Z. Wu, Nanoscale, 2014, 6, 5714.
H. Chong, P. Li, S. Wang, F. Fu, J. Xiang, M. Zhu and Y. Li, Sci. Rep., 2013,
3, 3214.
A. Das, T. Li, K. Nobusada, C. Zeng, N. L. Rosi and R. Jin, J. Am. Chem. Soc.,
2013, 135, 18264.
D. R. Kauman, D. Alfonso, C. Matranga, P. Ohodnicki, X. Deng, R. C. Siva,
C. Zeng and R. Jin, Chem. Sci., 2014, 5, 3151.
Y. Lu, Y. Jiang, X. Gao and W. Chen, Chem. Commun., 2014, 50, 8464.
R. Jin and K. Nobusada, Nano Res., 2014, 7, 285.
D. R. Kauman, D. Alfonso, C. Matranga, H. Qian and R. Jin, J. Am. Chem.
Soc., 2012, 134, 10237.
C. Yu, G. Li, S. Kumar, H. Kawasaki and R. Jin, J. Phys. Chem. Lett., 2013,
4, 2847.
M. Lee, P. Amaratunga, J. Kim and D. Lee, J. Phys. Chem. C, 2010, 114, 18366.
J. Lee, H. S. Shim, M. Lee, J. K. Song and D. Lee, J. Phys. Chem. Lett., 2011,
2, 2840.
A. Kogo, N. Sakai and T. Tatsuma, Electrochem. Commun., 2010, 12, 996.
Catalysis, 2016, 28, 5185 | 85
Research and development of

hydrocracking catalysts and technology
Chong Peng,a,b Xiangchen Fang*a,b and Ronghui Zenga
17:23:48.
DOI: 10.1039/9781782626855-00086
Hydrocracking (HCK), one of the main approaches to deep process heavy oil, is a catalytic
conversion process where feedstock undergoes hydrogenation, S/N removal, molecular
restructuring, cracking, and other reactions. It can process straight-run gasoline/diesel,
vacuum gas oil, and other secondary processing fractions such as fluid catalytic cracking
(FCC) diesel, FCC clarified oil, coker diesel, coker gas oil, and deasphalted oil and produce
various quality clean fuels such as liquefied gas, gasoline, kerosene, jet fuel, diesel, and
various quality petrochemical materials such as light/heavy naphtha and tail oil. In this
chapter, the research progress on commercial HCK technology and its relative catalysts
are discussed. The typical technical characteristics and the representative processes from
dierent corporations, such as Universal Oil Products, Albemarle, Criterion, Haldor Topsoe, and SINOPEC, are also presented. The development trend of HCK technology in the
future is outlined.
Introduction
Hydrocracking (HCK) is one of the most versatile petroleum-refining

processes. It usually converts heavy low-quality feedstock into lighter and
valuable transportation fuels or chemical raw materials through carbon
carbon bond breaking, accompanied by simultaneous or sequential
hydrogenation, which contribute significantly to the overall profitability
of the refinery. HCK is a process of considerable flexibility because it
allows the conversion of a wide range of feedstock to a variety of products.1 The HCK process has become the primary diesel production
method in many refineries. As environmental regulations on transportation fuels continue to tighten, HCK will be one of the tools used by
refineries to meet new product specifications. Unlike fluid catalytic
cracking unit (FCCU) processes, HCK can eectively yield ultra-low sulfur
diesel (ULSD) streams, whereas middle-distillate range FCC products
(i.e., light cycle oil, LCO) regularly require additional treatment to meet
product blending specifications.
Hydrocrackers are designed for and run at a variety of conditions
depending on many factors, such as type of feed, desired cycle length,
and expected product slate. However, in general, they will operate at the
following range of conditions: liquid hourly space velocity (LHSV)
(0.52.0 h1), H2 circulation (8501700 Nm3 m3), H2PP (10.015.7 MPa),
and SOR (start of operation) temperatures ranging between 357 1C and
405 1C.
a
Fushun Research Institute of Petroleum and Petrochemicals, SINOPEC, Liaoning

Fushun 113001, China
b
The State Key Laboratory of Chemical Engineering, East China University of
Science and Technology, Shanghai 200237, China. E-mail: fxc@ecust.edu.cn
86 | Catalysis, 2016, 28, 86118

c
17:23:48.
View Online
A typical vacuum gas oil (VGO) feedstock consists of paranic, naphthenic, aromatic, and naphtheno aromatic species, along with heteroatom impurities such as sulfur, nitrogen, and oxygen compounds.
Significant amount of metals may also be present. VGO can be classified
as light VGO (LVGO) and heavy VGO (HVGO). LVGO has a boiling range
of about 345430 1C. Typically, saturates in this range account for 40.0
60.0 wt% of the fraction, which consists mainly of parans and alkylnaphthenes with onefour rings. They have carbon numbers between 15
and 30. The aromatic content of LVGO may be as high as 50 wt%. Most
types of compounds in this fraction are alkylbenzenes, naphthalenes,
and phenanthrenes, with or without naphthenic rings. Sulfur compounds such as benzo-, dibenzo-, naphthobenzo-, thiophenes, and other
derivatives are also present. The elemental sulfur, oxygen, and nitrogen
content are 3.0 wt%, 0.2 wt%, and 1500 ppm, respectively.2 The polar
compounds in LVGO may range in content from 1.0 wt% to 10.0 wt%,
which consists of compounds containing nitrogen, oxygen, or both.
The atmospheric equivalent boiling range of HVGO is about 430540 1C
with a carbon number ranging from 20 to 50. Compared with LVGO, a
dramatic decrease in the amount of saturates and a corresponding increase in aromatics and polar compounds can be observed in HVGO.
Most aromatic and naphthenic molecules in this range primarily contain
onefour rings. Compounds with more than four rings are present at
much lower concentrations. Elemental sulfur, oxygen, and nitrogen
content can be about 4.0 wt%, 0.3 wt%, and 2000 ppm, respectively.
Typically, based on their boiling point ranges, the products obtained
from the fractionation of the hydrocracker euent are defined as follows:
light ends, light naphtha, heavy naphtha, jet fuel/kerosene, diesel fuel,
and unconverted fractionator bottoms. The products obtained from HCK
are generally of high quality. The light naphtha with octane number
between 78 and 85 can be used for the blending stock in the gasoline
pool. Heavy naphtha is a good-quality reformer feedstock for the production of high-octane gasoline. The jet fuel obtained is low in aromatics
and has a high smoke point. The diesel fraction has a relatively high
cetane number and very low sulfur content. High-quality lube oil, FCC,
and thermal cracking feedstock are also obtained from the HCK process.
In this chapter, the research progress on commercial HCK technology
and its relative catalysts is discussed. The typical technical characteristics
and the representative processes from dierent corporations, such as
Universal Oil Products (UOP), Albemarle, Criterion, Haldor Topsoe and
SINOPEC, are also discussed. The development trend of HCK technology
in the future was outlined.
History
Thermal HCK is one of the oldest hydrocarbon conversion processes

used for coal conversion, which was developed in Germany between 1915
and 1945. Its typical reaction conditions were quite severe: 21.069.0 MPa
and 375525 1C, producing liquid fuels that were highly aromatic because
of thermodynamic equilibrium. The large-scale industrial development
Catalysis, 2016, 28, 86118 | 87
17:23:48.
View Online
of hydrogenation in Europe, particularly in Germany, was entirely because of military considerations. Germany used hydrogenation extensively during World War II to produce gasoline: 3.5 million tons were
produced in 1944.
After World War II, HCK became less important because the availability
of Middle Eastern crude oil removed the incentive to convert coal to liquid
fuels, and the newly developed catalytic cracking processes proved to be
more economical for converting heavy petroleum fractions to gasoline.
The use of HCK changed in the early 1950s with the advent of catalytic
reforming, which made by product hydrogen available. In 1950, the
development of a new generation of catalysts, which were capable of
operating at much lower pressures between 10.0 and 22.0 MPa, made the
process economically aordable, which brought back the process to
the refineries. The first plant of this new generation started operations in
the United States to meet the demand for conversion of surplus fuel oil in
gasoline oriented refineries. By mid-1970s, HCK had become a mature
technology and today it is a well-established process.
Before 1970, HCK catalysts mostly used amorphous support. The
activity of this kind of catalyst was low, which led to high operating
temperature at SOR and short catalyst life. To increase the activity of HCK
catalysts, researchers started to develop zeolite HCK catalysts in mid1950s and successfully developed ultra-stable Y zeolite (USY), rare earth
dealuminium Y zeolite, silica substituted zeolite, nitrogen tolerant Y
zeolite, dealuminated Y zeolite, b-zeolite, ZSM-5 zeolite, macroporous
alumina, microporous alumina, and amorphous alumina silicate support
with higher activity and mid-barrel selectivity. At present, HCK catalysts
that meet various purposes can be produced by flexibly adjusting the
important properties of support, such as ratio of silica to alumina,
specific surface area, pore size, pore distribution, and acid strength.
Hydrocracking catalyst
3.1 Characterization of hydrocracking catalyst

Most of HCK catalysts of commercial interest are dual functional in
nature, which consist of both a hydrogenationdehydrogenation component and an acidic support. The reactions catalyzed by the individual
components are quite dierent. The cracking function is provided by an
acidic support, whereas the hydrogenationdehydrogenation function is
provided by metals. Figure 1 depicts the composition of HCK catalysts in
detail.
The dierent kinds of acidic supports can be (i) amorphous oxides,
e.g., silica-alumina, (ii) crystalline zeolites, mostly USYs and a binder,
e.g., alumina, or (iii) hybrid supports having a mixture of zeolites and
amorphous oxides. Cracking and isomerization steps take place on the
acidic support. The metal can be noble metals (palladium and platinum)
or non-noble metal sulfides from groups VIA (molybdenum and tungsten)
and VIIIA (cobalt and nickel). These metals produce the reactive olefins for
the cracking and isomerization steps and catalyze the hydrogenation of the
aromatics in the feedstock and heteroatom removal. The ratio of catalyst
88 | Catalysis, 2016, 28, 86118
View Online
Hydrocracking Catalyst
(Dual Function)
17:23:48.
Cracking Function
(Acid Support)
Amorphous
(SiO2Al2O3.X-Al2O3
X=halogen
Low-Zeolite
/Amorphous
(Modify.Y/SiO2
-Al2O3)
Hydrogenation Function
(Metals)
High-Zeolite
+Binder
(Modify.Y+
Al2O3)
Noble metals
(Pt, Pd)
Non-noble metals
[MxSy from gr VIA (Mo, W)
+gr VIIIA (Co, Ni)
Fig. 1 Composition of HCK catalysts.1,3

Table 1 Strength of hydrogenation and cracking function in bifunctional catalysts.
Relative strength
Hydrogenation function
Cracking function
Co/MooNi/MooNi/WoPt (Pd)
Increasing hydrogenation activity (in low-S environment)
Al2O3oAl2O3-halogenoSi2-Al2O3ozeolite
Increasing cracking activity (acidity)
acidic and metal activities is adjusted to optimize activity and selectivity.

The relative strength of dierent hydrogenation components and cracking
(acid) components in HCK catalysts are shown in Table 1.4
For a HCK catalyst to be eective, a rapid molecular transfer between
the acid sites and hydrogenation sites is important to avoid undesirable
secondary reactions. Rapid molecular transfer can be achieved by having
the hydrogenation sites located in the proximity of the cracking (acid)
sites. The HCK catalyst performance is evaluated by four key criteria:
initial activity, which is measured by the temperature required to obtain
desired product at the start of the run; stability, which is measured by the
rate of increase of temperature required to maintain conversion; product
selectivity, which is the ability of a catalyst to produce the desired product
slate; and product quality, which is the ability of the process to produce
products with the desired use specifications, such as pour point, smoke
point, or cetane number.
3.1.1 Hydrocracking catalysts for maximally producing petrochemical naphtha. This series of catalysts has been developed to solve the
problem of refineries producing less light distillates even though a high
demand for petrochemical naphtha (feedstock for catalytic reformer)
exists. The series of catalysts has high HCK activities and can usually
fully convert 4350 1C distillates to o180 1C distillates under higher LHSV
of 1.52.0 h1 and lower temperature of 365380 1C. The yield of petrochemical naphtha (65177 1C) can be as high as 65% (to feedstock).
3.1.2 Hydrocracking catalysts for flexible production of medium
distillates and chemical naphtha. The characteristics of this series of
Catalysis, 2016, 28, 86118 | 89
17:23:48.
View Online
catalysts is that they have a moderate activity and can be used to produce chemical naphtha and middle distillates (jet fuel and diesel).
When the target product is chemical naphtha, its mass yield can be
30%40% (the yield of medium distillates is 45%55%); when the target product is medium distillates, its yield can be 60%65% (the yield
of chemical naphtha is 15%20%). Because of their flexibility, refineries can meet the demands of the market by adjusting operating
conditions.
3.1.3 Hydrocracking catalysts for maximally producing middle and
heavy distillates. This series of catalysts has been developed to meet
the increasing demands for middle and heavy distillates. These catalysts have medium activities, with the mass yield of middle distillates
equal to 70%80%. The catalysts can be used for the direct production
of 3# jet fuel and sulfur-free clean diesel with a high cetane number
and a low aromatic content.
3.1.4 Single-stage hydrocracking catalysts. This series of catalysts
has two categories: one is amorphous catalyst and the other is zeolite
catalyst. The former is better at producing more middle distillates
(75%80%). Its yield of diesel fraction can be over 40% (the diesel yield
of the series of catalysts for maximally producing middle/heavy distillates as mentioned above is only 25%30%). For those catalysts that do
not have higher cracking activities, under typical operating conditions,
the temperature at SOR is over 410 1C. Although adding some zeolites
to the support can significantly increase the cracking activity of the
HCK catalysts (the later ones), it also decreases the mid-barrel selectivity (the average yield is 60%65%). To integrate the strength of both
categories, some new single stage catalysts have been developed; for
example, SINOPEC developed the single-stage HCK catalyst FC-14,
which not only had the equivalent mid-barrel activity as amorphous
catalysts, but also had higher activity (operating temperature at SOR is
less than 10 1C).
3.2 Recent progress on hydrocracking catalyst
The recent progress on commercial HCK catalysts mainly focus on increasing the selectivity for middle distillates (diesel, kerosene, and jet
fuel) or providing flexibility for shifting the balance of production between distillates and naphtha. Catalyst selectivity depends on the concentration of Brnsted acid centers. If the concentration of Brnsted acid
centers is high, cracking activity can favor the formation of lighter
products, i.e., gases and naphtha. Therefore, many distillate-selective
catalysts are based on zeolite Y, which has been dealuminated to reduce
the acidity. Others are supported on amorphous silica-alumina. Selectivity for distillates versus naphtha can also be controlled by process
conditions. All of the major suppliers of HCK catalysts oer distillate
selective products. A smaller number of them have catalysts with flexible
naphtha/distillate selectivity including Criterion Catalysts & Technologies, SINOPEC, and UOP.5
90 | Catalysis, 2016, 28, 86118
17:23:48.
View Online
Axens provides catalysts for the conventional HCK of coker and visbreaker VGO streams, deasphalted oil (DAO), LCO, and heavy cycle oil
(HCO). The latest Ni-Mo zeolite catalysts for high-pressure HCK processes
include HYK 732, HYK 742, HYK 752, and HYK 762. These catalysts are
said to have optimized dispersion of zeolite crystals in the alumina
matrix, which improves metals impregnation and thus lowers the distance between the acid and metal sites, consequently resulting in higher
activity, improved hydrogenation eciency, better middle distillate selectivity, extended cycle lengths, and long-term stability. The alumina
matrix utilized in HYK 700 series catalysts also allows higher metal
loadings and improves porosity, which results in better diusion of both
reactants and products. HYK 742 is Ni-Mo zeolite catalyst with higher
activity than its predecessors. Compared with amorphous catalyst, it has
the same advantages of lower temperature operation, high stability, and
regenerability. HYK 742 is used in the high-pressure HCK processes for
converting a variety of feeds to high-quality diesel and kerosene, along
with high-viscosity index (VI) lube oil bases and steam cracker feed. HYK
742 is more active for conversion and has a better balance between hydrogenation and cracking function because of closer proximity between
acid and metal sites, higher metal loading, and controlled acidity. HYK
742 is more active than the previous HYC 642 catalyst by about 8 1C for
the same middle distillate selectivity.
Clariant Company developed a metal impregnated zeolite-based
catalyst named HYDEX-G, which could be regenerated for extended service life and are used to improve the cold flow properties of middles
distillates. HYDEX-G can be utilized in a stand alone operation or within
a middle distillate hydrotreater. It can be operated at 3.015.0 MPa and
280580 1C.6
Recent catalyst development from Chevron Lummus Global (CLG) has
focused on maximizing the yield of high-quality products from poorquality feeds. In addition to selectivity improvement, enhanced activity
and H2 consumption reduction have emerged as major spurs of catalyst
developments. The HCK catalyst portfolio of CLG presently comprises
three generations of catalyst oerings. The first generation largely
includes catalysts developed prior to 2000, but many are still utilized
in HCK units. The third generation of HCK catalysts with improved activities and selectivities have become commercially available, although
the large-scale industrial application of these newest catalysts has been
limited so far. ICR 250 and ICR 255 are the first members of thirdgeneration catalysts that have been commercialized with the first applications noted in 2010. ICR 250 is claimed to be an improvement from ICR
240 in terms of middle distillate yield, with specific gains attributed
to better hydrogenation functionality. ICR 250 is particularly suitable for
HCK schemes configured for recycle. ICR 255 is a jet fuel selective
second-stage middle distillate HCK catalyst that allows a significant drop
in weighted average bed temperature compared with ICR 240 while also
improving overall middle distillates yield.7,8 ICR 214 is also included in
the third-generation catalysts group designed to maximize naphtha
production. ICR 214 incorporates base metals as the active components
Catalysis, 2016, 28, 86118 | 91
17:23:48.
View Online
and is said to lower gas and premature catalyst deactivation via improved
nitrogen resistance while also improving the yield of naphtha.9
Criterion Catalysts & Technologies announced that Rive Technology
would work with Molecular Highwayt Zeolite Technology on HCK catalysts at the 2014 AFPM conference in March.10 Criterion and Zeolyst
oered state-of-the-art catalysts with the goal of improving naphtha and
distillate flexibility. Z-2513, Z-2623, Z-2723, Z-3723, Z-3733, Z-853, and
Z-863 are oered for a wide range of distillates selectivity and cracking
activity, ranging from high diesel of Z-2513 and high naphtha of Z-863.
These catalysts, when compared with earlier technologies, provide the
same activity but, at the same time, do not occupy as much space. This
quality is credited to the advanced trilobe extra (ATX) shape of the catalysts, which allows better diusion of reactants and larger void fraction to
pick up particulates. The ATX shape maintains the advantages of reduced
delta pressure and enhanced diusion, which was provided by the TX
shape, but with higher strength. Unlike the previous TX shape, the ATX
shape can be applied to the majority of the catalyst portfolio. Compared
with the conventional trilobe shape, the ATX shape has a shorter diffusion path, which results in enhanced catalyst utilization by reducing
overcracking, and better liquid yields. The larger void fraction of ATX also
increases particulate uptake and delays the onset to rapid pressure differential buildup. The ATX shape produces a lower start-of-run pressure.
The company is currently developing the Z-FX10, Z-FX20, Z-NP10, and
Z-NP20 catalysts for expanded flexibility and selectivity. In pilot plant
testing, where feeds of both HVGO and Canadian synthetic VGO were
tested, the Z-FX10 catalyst improved middle distillate selectivity by 3% for
HVGO and 4% for Canadian synthetic, relative to the Z-3723 catalyst.
Diesel cetane did not change the HVGO feed, whereas it improved
Canadian synthetic crude by 1%.11
Haldor Topsoe also has emphasized on achieving both higher activity
and greater selectivity for middle distillates. The HCK catalysts oerings
portfolio of Topsoe has been generally separated in to three series: the
red series, which was tailored to maximize hydrogenation, the blue series, which provided the maximum yield of middle distillates that were
characterized by good cold flow properties, and the new D-selt series,
which aimed to maximize diesel yield in the hydrocracker. Compared
with the red series, the D-sel series increases diesel yields by 6% while
maintaining the same conversion level. TK-939 is a Ni-W-cylinder shape
catalyst that has the same HCK ability as TK-931, but only with higher
selectivity for middle distillates. TK-949 is a Ni-W catalyst that has similar
activity as TK-941, but with higher selectivity towards middle distillates.12
SINOPEC Fushun Research Institute of Petroleum and Petrochemicals
(FRIPP) has focused on developing catalysts for use in HCK. The catalysts
in the FC series of FRIPP range from moderate activity with high selectivity for middle distillates to high activity and high selectivity for
naphtha. FC-32 catalyst provides flexibility in the ratio of naphtha to
middle distillates depending on the process conditions. The catalyst has
good resistance to nitrogen and shows excellent stability because it accommodates variability in the feedstock. The FC-32 catalyst can provide
92 | Catalysis, 2016, 28, 86118
17:23:48.
View Online
once-through conversion of about 80% because it operates on VGO feed

with a pressure of about 15 MPa and temperature of about 380 1C. The
FC-14 catalyst displays moderate activity but very high selectivity for
middle distillates because it uses modified b-zeolite. The FC-60 catalyst
has less activity but higher selectivity for middle distillates. The FC-20
catalyst is noted for high activities for hydrogenation and ring-opening.
Among the FC catalysts, FC-46 has the highest activity and lowest selectivity for middle distillates. This catalyst is oered as a means of
meeting increasing the demands for naphtha in China.
The UOP researchers aim to understand the molecular transformations enabled by the catalyst so that they can couple that knowledge with
catalyst design and process conditions.13 The catalyst portfolio of UOP
oers the options of maximum production of distillates and diesel,
maximum yield of naphtha, or the flexibility to make high yields of either
distillates or naphtha. Some of the catalysts are also designed to use
hydrogen more eectively. UOP has developed a model for designing
catalysts, which is called the Catalyst Design Engine. The model relates
the catalyst physical properties with the desired features, such as higher
stability or hydrogenation activity. This allows H2 to be used selectively
and avoids adding excess H2 in to boiling ranges that do not add value.
The Engine relies on specific physical characteristics of the catalyst, such
as the acidity or pore geometry, as well as an extensive amount of empirical data accumulated on dierent catalyst formulations. The Catalyst
Design Engine enables first-iteration catalyst development, which comes
very close to meeting desired performance goals while greatly shortening
the amount of time needed for catalyst development. In addition, UOP
has enlarged its view on how to characterize catalyst performance.
Traditionally, activity and selectivity were the key parameters, but these
parameters did not help refiners who have concerns with hydrogen
management. As a result, UOP now adds hydrogenation to its catalyst
modeling and considers unicracking catalysts in the three dimensions of
activity, selectivity, and hydrogenation. HC-140LT has been formulated to
produce a blend of distillate and naphtha, or adjust between maximum
distillate and maximum naphtha modes depending on market conditions. The diesel produced with the new catalyst exhibits a high cetane
number, whereas any jet fuel produced has a high smoke point. HC140LT can also maximize yields of heavy naphtha while minimizing light
naphtha and LPG make. HC-140LT displays a higher activity than the
previous catalysts and gives higher distillate yields. HC-140LT can be
used in single-stage, single-stage with recycle, and two-stage hydrocracker configurations. HC-150LT is the best diesel catalyst that is said to
reduce H2 consumption by 10%15% while increasing activity at 6 1C
compared with previous catalysts. HC-185LT is a flexible HCK catalyst
that is able to alternate between maximum distillate and naphtha production within a single catalyst cycle based on market conditions. The
flexible catalyst is claimed to give diesel yields of 3050 vol% at nominal
conversion with improved cold flow properties while also reducing H2
consumption by 10%20% compared with other flexible catalysts. HC205LT is a new second-stage HCK catalyst designed to maximize yields of
Catalysis, 2016, 28, 86118 | 93
17:23:48.
View Online
diesel, ethylene cracker feed, high-quality lube oil base stocks, and
jet fuel/kerosene. HC-205LT has been proven to increase distillate yield
by 23 wt% and diesel yield by 35 wt% compared with other alternative
second-stage HC catalysts. HC-205LT has also been utilized in applications that process heavy and dicult feeds and produce high-quality
products.
Commercially available HCK catalysts are listed in Table 2, which were
mainly referred to in the Oil and Gas Journal and Worldwide Refinery
Processing Review.5,14
4 Hydrocracking processes
Major companies in the HCK technology licensing include Axens, CLG,
Haldor Topsoe, Shell Global Solutions, SINOPEC, IFP, and UOP. Each of
them provides a range of solutions for treating a variety of feeds that can
be operated at varying severities. HCK technologies generally include
mild, medium-pressure, and conventional HCK processes (high pressure). Mild HCK (pressure of 410 MPa and conversion of 15%45%) is
generally used to convert VGO into middle distillates, particularly ULSD,
and an upgraded residual stream for FCCU feed. Medium-pressure HCK
bridges the gap between mild and high-pressure HCK in terms of cost,
conversion, and product quality. The high-pressure HCK process converts a variety of feeds into high-quality diesel and kerosene, along with
high-VI lube oil and steam cracker feed.15 Feeds include VGO from many
sources (DAO, furfural extracts, parans, LCO, and HCO). The pressure
ranges from 10 MPa to 14 MPa or more.
The HCK unit and process can be configured in a number of ways. The
unit can consist of one or two reactors loaded with one type of catalyst,
one reactor with stacked beds of two types of catalysts, or multiple reactors with dierent catalysts. The process can utilize one or two stages
and be operated in once-through or recycle mode, and the recycle gas
system can be shared by both stages or used separately.
4.1 Basic flow schemes
Three basic processes configurations exist: single-stage, single-stage in
series, and two-stage processes. In all three configurations, the feed is
first sent through a pretreatment section (hydrotreating reactions) and
then through a HCK section (HCK reactions).
The single-stage once-through or recycle process is shown in Fig. 2.
This configuration also provides high yields of naphtha, middle distillates, and unconverted oil (UCO). With a single-stage HCK process, the
reactions of hydrodenitrogenation (HDN), hydrodesulfurization (HDS),
and HCK can take place in the same reactor. Therefore, this process has
advantages such as simple flow scheme, high LHSV, and low construction and operation costs. The light naphtha, with an average octane
number of 80, is sent to the gasoline pool for blending, whereas the
heavy naphtha (naphthene content usually 450%) is suitable for reformer feed. The kerosene product has a smoke point of 25 mm, and
the diesel has a cetane number of more than 60. The UCO from the
94 | Catalysis, 2016, 28, 86118
17:23:48.
Table 2 Commercially available HCK catalysts.

Conventional hydrocracking
Company
Maximum
naphtha, jet
Maximum diesel, jet, lube
Advanced Refining Technologies
Albemarle Catalysts
KC
KC
KC
KC
2301
2601
2710
2715
Axens
Chevron Lummus Global
Catalysis, 2016, 28, 86118 | 95
Exxon Mobil Research &

Engineering
ICR
ICR
ICR
ICR
ICR
ICR
ICR
ICR
ICR
ICR
ICR
ICR
ICR
ICR
ICR
136
139
141
147
160
209
210
211
214
230
160
180
183
185
255
KC 3210
KC 3211
KC 2301
KC 2601
KC 2610
KC 2611
HYC 642
HYC 652
HYK 732
HYK 742
HYK 752
HYK 762
ICR 106
ICR 120
ICR 126
ICR 142
ICR 150
ICR 155
ICR 162
ICR 177
ICR 180
ICR 220
ICR 240
ICR 245
ICR 250
Mild hydrocracking
Flexible produce
distillate or naphtha
Maximum
naphtha, jet
Maximum
diesel, jet, lube
KC-2601
KC-2301
KC-2601
KC-2602
HTH 548
ICR 141
ICR 147
ICR 160
RT-3
TN-8
ICR
ICR
ICR
ICR
ICR
106
126
142
150
162
Conventional hydrocracking
96 | Catalysis, 2016, 28, 86118
17:23:48.
Table 2 (Continued)
Company
Maximum
naphtha, jet
Haldor Topsoe
TK-965
JGC Catalysts & Chemicals Ltd.
V-HYC
NHC-97-13
Kataleuna GmbH Catalysts

SINOPEC
UOP
Zeolyst Int. (marketed by Criterion)
FC-46
FC-24
3825
3905
3955
DHC-41LT
HC-24
HC-26
HC-28
HC-34
HC-38
HC-80
HC-29
HC-53
HC-150LT
HC-175LT
HC-190LT
Z-723 Z-733
Z-743 Z-753
Z-773 Z-803
Z-853 Z-863
Z-3723
Z-3733
Z-NP10
Maximum diesel, jet, lube

TK-925 TK-926
TK-931 TK 933
TK-941 TK-943
TK-947 TK-951
TK-961
V-HYC
NHC-97-13
KL-8380
KL-9514
FC-14FC-16 FC-20FC-26
FC-50 FC-60 RIC-1RT-1 RT-5 38823901 3974
DHC-2
DHC-8
DHC-32LT
DHC-39LT
DHC-41LT
HC-35
HC-43LT
HC-53LT
HC-115LT
HC-120LT
HC-190LT
HC-205LT
HC-215LT
Z-503 Z-513Z-603 Z-623Z-673 Z-723Z-733
Z-2513Z-2623 Z-2723Z-3723 Z-3733Z-5723
Z-FX10 Z-HD10 Z-HD11Z-MD10
Mild hydrocracking
Flexible produce
distillate or naphtha
Maximum
naphtha, jet
Maximum
diesel, jet, lube
TK-961 TK-962
TK-965
NHC03-35
3824 3903 3971

3976 FC-12
FC-32 FC-36
HC-43LT
HC-140LT
HC-150LT
HC-185LT
HC-35
DHC-2
Z-723
MHC-210
View Online
R-1
Product Gas
Recycle H2
Makeup H2
Fresh Feed
Light Naphtha
Heavy Naphtha
17:23:48.
Jet Fuel/Kerosene
Diesel
Single-Stage
Product
Fuel
FCC Feed
Ethylene Feed
Lube Oil Base
Recycled Boltoms
Fig. 2 Simplified flow diagram of the single-stage with and without recycle scheme.
R-1(HT)
R-2(HC)
Separators
Product Gas
Recycle Gas
Compressor
Preheater
Light
Naphtha
Heavy
Naphtha
Wash Water
HP Cold
Flash Gas
Jet Fuel
Etesel
Fuel
Heal
Fresh Feed
Makeup Hydrogen
Fig. 3
Exch H P Hot
LP Cold
Sour Water
Simplified flow diagram of the single-stage in series configuration.
single-stage once-through process has a Bureau of Mines Correlation

Index (BMCI) of less than 10, which indicates a highly hydrogenated
product that is suitable as a feedstock for the catalytic reforming process.
After dewaxing, the UCO from this process is also suitable as a Group III
lube oil base stock.
The single stage in series HCK process (shown in Fig. 3) needs at least
two reactors: one is loaded with a pretreating catalyst that has superior
HDS, HDN, and aromatics saturation activities for the feedstock to
undergo the relative reactions in this reactor; the other is loaded with a
HCK catalyst, in which the feedstock with N content is lowered to certain
range (typically 1020 ppm) to undergo HCK reactions. This process has a
good adaptability to feedstock, a flexible operation selection, good
product quality, and a long run length. This process can also take various
recycle operations to meet the specific demands of refineries, which include full recycling mode for maximally producing chemical naphtha, jet
fuel and diesel, middle distillates recycling mode for maximally producing chemical naphtha and ethylene feedstock (hydrocracker tail oil),
and once-through/part recycling modes for producing chemical naphtha,
middle distillates, and ethylene feedstock. Despite its advantages, the
Catalysis, 2016, 28, 86118 | 97
17:23:48.
View Online
single-stage in series HCK process also has some weak points. Its LHSV is
low, and the construction cost is high. The operation is limited by the
activity of catalysts (zeolite HCK catalysts). The average middle distillates
yield is only 65%75%. Moreover, the product selectivity changes obviously at the end of operation.
The first stage of the two-stage HCK process can be either a single-stage
serial connection mode or a single-stage mode. The HCK reactors of both
stages can be loaded with zeolite catalysts or amorphous catalysts, and
with same or dierent catalysts. The process has advantages such as good
adaptability to feedstock, high yields of target products, good product
quality, and high LHSV. In the one-stage process, the existence of ammonia in the recycling gas inhibits the activity of HCK catalysts to some
extent. In the two-stage process, the recycling hydrogen in the second
stage is nearly free of ammonia and can ensure that the catalysts function
at high LHSV and low temperature. Furthermore, this process also features flexible operation. By changing the HCK catalysts of the second
stage, shifting from max-middle distillate mode to max-chemical naphtha mode is easy. Moreover, this process does not have such engineering
problems as mal adjustment of operating temperature/activity among
dierent catalyst beds. This process is superior to the single-stage processes, in which the LHSV and middle distillate yields are much higher
and consumes less hydrogen, but its flow scheme is more complex and
it needs more construction investment. With the HCK unit becoming
increasingly huge, the two-stage HCK process will have its position in
refining industries.
Both the single-stage with recycle scheme and the two-stage configuration achieve a similar yield of C5 , with the two-stage configuration
giving a better diesel/kerosene ratio, which is shown in Fig. 4. The singlestage with recycle scheme gives a higher distillate yield with similar
quality and a slightly higher H2 consumption than the single-stage oncethrough scheme. The single-stage with recycle configuration uses a small
purge to prevent buildup of polycyclic aromatic hydrocarbons (PNAs)
in the recycle loop. This purge steam can be used as part of the FCC
Product Gas
Makeup
Hydrogen
Fresh Feed
Recycle
Gas
Recycle
Gas
Light
Naphtha
Heavy
Naphtha
Kerosene
Diesel
First-Stage
Product
Fig. 4
Second-Stage
Product
Simplified flow diagram of the two-stage configuration.
98 | Catalysis, 2016, 28, 86118
View Online
Hydrocracking Processes
17:23:48.
Single-stage
(no intermediate separation
of products)
HC only (1reactor)
single catalyst
low N and S feed
HT+HC (1reactor or
2reactors)
dual catalyst
high N and S feed
without recycle
once-through process
Two-stage
(Separation of products
between stage 1 and stage 2)
HT+HC (2reactors)
dual catalyst
fractionation of HT
effluent
recycle to HC reactor
HT+HC (2reactor or
3reactors)
dual or triple catalyst
fractionation of effluent
from 1st HC reactor
recycle to 2nd HC reactor
with recycle
to HC reactor
HC-hydrocracking; HT-hydrotreating
Fig. 5 Classification of HCK processes.1
feedstock or as feed for a steam cracker. In the two-stage process, middle

distillates yield can exceed 90% with a maximum amount of diesel, which
meets the ULSD specifications. Similar with the single-stage with recycle
configuration, a limited purge is needed. H2 consumption is similar to
the other two configurations.
The key characteristics of dierent HCK processes are summarized in
the Fig. 5. The two-stage configuration oers more flexibility than the
single-stage configuration. The former is better suited for processing
heavier feedstock that is rich in aromatic and nitrogen. Separation of
ammonia and hydrogen sulfide between the first and second stages
oers more flexibility in the selection of a HCK catalyst for the second
stage and results in a deeper hydroconversion of the feedstock.
4.2 Recent progresses on hydrocracking processes
4.2.1 HyCycle Unicracking. HyCycle Unicracking is a low-pressure
process that can be used to achieve 99.5% conversion at 20%40% conversion per pass.16 The mode of operation can be changed to maximize
the yield of higher boiling products. The HyCycle flow scheme employs
the engineering innovations of an enhanced hot separator (EHS).17
In the HyCycle process, after the feed passes through the hydrotreating
catalyst zone, the HyCycle separator/finisher separates UCO from
cracked products and selectively hydrogenates the aromatics of the
product stream. Most of the UCO is routed to the HCK catalyst zone to
feed the hydrotreating zone. The remaining portion of the UCO is removed as a purge stream using a novel split-wall fractionator design.
Hydrogen is recovered from the product stream before the stream is
flashed, stripped, and fractionated.
According to UOP, the combination of HyCycle of patented design
features oers several advantages. One advantage is the HCK of cleaner
feedstock provided by the back-staged reactor configuration, which
raises the activity of the catalyst and there by lowers the amount of
Catalysis, 2016, 28, 86118 | 99
View Online
Makeup
Hydrogen
Offgas
HC Rx
Amine
Scrubber
To LPG
Recovery
HP Cold
Fresh
Feed
17:23:48.
Distillate
Product
Feed Gas
HT Rx
Separator
/finisher
Hot
Flash
Drum
Product
Fractionator
Stripper
Cold
Flash
Drum
0.5% UCO
Fig. 6 UOPs HyCycle Unicracking process.
catalyst needed and enables the unit to be operated at a lower pressure of

25% compared with commonly used in hydrocrackers.18 Additionally, the
HyCycle separator/finisher produces diesel that meets Euro-V specifications and allows unconverted materials to be recycled, which eectively
lowers the per pass conversion requirements. This processing scheme
increases diesel yields by 10%15% over conventional units and minimizes the secondary reactions that produce undesirable coke and gas.
The 20% reduction of H2 consumption and the more ecient use of
process heat lower the operating costs by up to 15%.19 Figure 6 gives a
diagram of the unit.
4.2.2 Advanced partial conversion unicracking. Advanced partial
conversion unicracking (APCU) process is based on HyCycle Unicracking. In this process, high-sulfur feedstock, such as VGO or heavy coker
gas oil (HCGO), is mixed with a heated stream of recycle hydrogen and
sent downward through a reactor containing a top bed of high-activity
hydrotreating catalyst and a bottom bed of distillate-selective unicracking catalyst. The reactor euent is sent to an EHS. The cracked products that are overhead streamed from this unit can be mixed with
distillates such as kerosene, light CGO, and mildly hydrotreated diesel
before hydrogenated in a distillate unionfining step and then fractionated. Desulfurized FCC feed flows from the bottom of the EHS directly
to the fractionation column. The unit can assist refiners who operate
an FCCU to rebalance dieselgasoline made in a cost-eective way.
Using the APCU design can also oer a significant reduction in hydrogen usage of 5%10% on the unit when compared with a conventional
mild HCK process.20 A diagram of the APCU process is shown in Fig. 7.
The use of the EHS enables the enhancement of distillate product quality
while preventing over treatment of the FCC feed, which consequently
minimizes hydrogen consumption. The process configuration provides
the flexibility to vary the amounts of high-cetane ULSD blending stock
and low-sulfur FCC feed, which are produced according to seasonal demand. This process also yields high-naphthene-content naphtha with
100 | Catalysis, 2016, 28, 86118
View Online
High-sulfur
FCC Feed
Makeup
Hydrogen
HC Rx
Amine
Scrubber
HT Rx
HP Cold
HC Rx
17:23:48.
Offgas
ULSD
Co-feed
Hot
Flash
Drum
Cold
Flash
Drum
Enhanced Hot
Separator
Low-sulfur
FCC Feed
Fig. 7
UOPs APCU process.
ultra-low sulfur of o0.5 ppm that can be sent to reformers to boost the
octane.
This technology can be implemented as a grassroots unit or as a
revamp of a FCC feed hydrotreater and a hydrocracker. It reduces its
investment and operating expense relative to those of other pretreatment
technologies. It can also produce higher quality diesel using lower
conversions of 20%50% and lower operating pressure of 79 MPa.
4.2.3 Optimized partial conversion and selective staging. The optimized partial conversion (OPC) technology has low initial investment
costs and allows for less expensive revamping to meet higher product
quality standards or to operate at higher conversion. The OPC technology uses reverse staging instead of adding more reactor volume downstream of the existing reactor and/or a saturation reactor for product
treatment, which is described in Fig. 8. Nitrogen and other catalyst poisons are removed from the feed in the existing reactor, which contains
layers of HDN and HCK catalysts. The liquid product from this reactor
flows into a fractionator. A portion of the bottom stream from this
column is recycled to the new reactor, which is fed makeup hydrogen
and operates at a higher pressure and lower temperature than conventional HCK. The HCK reactions, including paran conversion, occur
more rapidly in the upstream reactor in the absence of the impurities.
The OPC process has the ability to optimize the yield and quality of
FCC feed and distillate products by limiting conversion in the first reactor, modifying the recycle stream, and varying the operating conditions
and catalysts of the second reactor,21 through which results that are
comparable to the use of a downstream reactor with two or three times
the catalyst volume are achieved.22
The selective staging process is a modification of the OPC process,
which is designed to process HVGO feeds with very high boiling points
and refractory feeds, such as HCGO, into high-quality jet and diesel
products. The selective staging process is shown in Fig. 9. The fresh feed
is first hydrotreated, and the UCO is sent to the vacuum column.
Catalysis, 2016, 28, 86118 | 101
View Online
Makeup
Hydrogen
Product Gas
Recycle gas
Light
Naphtha
Fractionator
17:23:48.
Fresh Feed
Heavy
Naphtha
Kerosene
Diesel
Separator
FCC Feed
Recycle Oil
Fig. 8 Chevrons OPC process.
2-Stage Reactor
Product Gas
1-Stage Reactor
Makeup
Hydrogen
Recycle
Gas
Fresh Feed
Fractionator
Kerosene
Diesel
To Vacuum System
Vacuum
column
FCC Feed
Recycle oil
Fig. 9
Chevrons selective staging HCK process.
Chevron Lummus Globle also oers a new process that combines

selective staging with reverse staging for the coprocessing of highboiling refractory feeds with middle distillates to maximize high-quality
middle distillates and FCC feed and optimize H2 consumption.
Figure 10 depicts the selective/reverse staging HCK scheme. Similar
with the selective staging process, the selective/reverse staging scheme
produces a jet product with sulfur of o10 ppm and a smoke point of
425 mm, a diesel product with a sulfur content of o10 ppm, and cetane
number of 450.23
4.2.4 Single-stage reaction sequencing. The single-stage reaction sequencing (SSRS) scheme has important benefits: It can reduce quench
requirements, as in the split-feed injection arrangement, and reduce circulating gas requirements by sending unused H2 from the second stage
to the first stage. It can also realize high yields of high-quality middle
distillates, high conversion, and low cost. The first commercial implementation of SSRS was installed in 2006 at the 1.5 Mt a1 hydrocracker
102 | Catalysis, 2016, 28, 86118
View Online
Distillates
Reactor
LCGO/CND
Recycle Gas
LN
Stripper
HVGO/HCGO
HN
17:23:48.
Kerosene
ULSD
HHPS
HCR
Reactor
HDT
Reactor
Vacuum
Column
HHPS
FCC
Feed
Stripper
Reactor
VGO
Fig. 10 Chevrons selective/reverse staging HCK process.
R1102
TC
FC
C1202
H1101
V1104
H1201
R1101
E1105
E1104
P1102
P1103
Fig. 11 Single-stage reaction sequencing scheme.
of West Pacific Petrochemical in Dalian, China. The feed of the unit was
a full-range Middle Eastern VGO, and it was operated in max middle distillate mode with CLG ICR 162 catalyst in the first stage and ICR 180 in
the second stage. SSRS was operated at a temperature range of 350390 1C
and a pressure of 14.6 MPa. The performance test results showed that
with 100% conversion, the yield of the middle distillates reached
83.1%, the smoke point of aviation fuel was 30 mm, the sulfur content
of heavy diesel fraction was less than 10 mg g1, and its cetane index was
61. Furthermore, the energy consumption of the unit was 61.3 MJ t1
lower than the design data. Figure 11 shows the SSRS scheme.24
4.2.5 Hydrotreating and hydrocracking integrated technology. To
provide the most economical solutions for customers, UOP developed
Catalysis, 2016, 28, 86118 | 103
View Online
DAO RCD Section
Distillate
VGO/DAO
Unionfining Unicracking
Section
Section
VGO Feed
17:23:48.
AGO Feed
DAO Feed
Amine
Scrubber
HP Cold
HP Hot
HP Hot
Woth Post
Treat Bed
To sour Gas
Recovery
Stipper
Hot
Flash
Drum
Cold
Flash
Drum
Stabilized
Naphtha
Kerosene
Fractionator
ULSD
VGO
Fig. 12 Integrated flow scheme for heavy oil upgrading.
a hydrotreating and hydrocracking integrated technology, including

RCD Unionfiningt, Distillate Unionfining, and VGO/DAO Unicracking
processes. Integrating these three processes into one combined unit
could save capital and operating costs.25 The integrated unit design
shown in Fig. 12 reduces the equipment count and utility requirement
for compression, pumping, and process heating.
This integrated unit utilizes the RCD Unionfining process to upgrade
DAO and integrates with the VGO and distillate hydroprocessing units by
using a hot high-pressure separator operating at fairly high temperature.
Most of the naphtha and about half the distillate generated in the RCD
Unionfining Process reactors are still in vapor phase during operating
conditions in the hot separator. Sucient hydrogen is available from the
hot separator to hydrotreat all Atmospheric Gas Oil (AGO) available from
the crude unit atmospheric column. The overhead vapor stream from the
hot separator is therefore routed directly to the inlet of a Distillate
Unionfiningt Process reactor where it is co-currently processed with the
AGO fraction of the feed. The operating conditions and reactor volume
are set to produce distillate stream that meets all ULSD specifications.
The only additional hydrogen required is that needed for reactor interbed quenches.
Hydrotreated DAO from the bottom of the first hot separator is
pumped to the inlet of the Unicracking process reactor train. The
addition of a high-pressure low-head pump at the bottom of the hot
104 | Catalysis, 2016, 28, 86118
17:23:48.
View Online
separator greatly reduces the power required to pump hydrotreated

DAO to the reactor inlet pressure, and reduces the design pressure.
Therefore, the process lessens capital cost of several pieces of
equipment.
VGO from the crude vacuum column is sent to the inlet of the
hydrocracking section. The VGO/DAO blend is converted to produce a
desired quantity of VGO-range material for synthetic crude blend. A highactivity pre-treat catalyst available to UOP through its alliance with
Albemarle, such as KF-848 STARS, would be loaded in the first reactor.
Meanwhile, a stacked loading of Unicracking catalysts is used for
hydrocracking. This step allows organic nitrogen to slip from the first
reactor to be monitored and controlled for optimum unit performance
and cycle length. The reactor euents from both hydrotreating and
hydrocracking sections are combined in a common separator and fractionation system.26
4.2.6 ISOFLEX. Chevron Lummus Global has devised a highconversion ISOFLEX scheme that accomplishes deep HDN in a firststage hydrocracker, followed by hydrotreatment and aromatic saturation
in a hot hydrogen stripper-reactor, and finally, further hydrocracking in
a second-stage reactor. Compared with SSOT isocracking, this scheme
operates at lower pressure and with lower catalyst temperature and
hydrogen consumption, while providing the same yield of distillates.
This technology has the following features. The first feature is multifeed capability for VGO and diesel-range streams, with the best conditions for treating each kind of feedstock being processed. The second
feature is maximized conversion and high product selectivity, while
minimizing equipment requirements. The third feature is minimal
hydrogen consumption as a result of reduced production of light ends
and by introduction of hydrogen when and where it is needed to process
dierent fractions. Finally, this technology uses second stage processing
to provide high treatment rates.
Figure 13 shows the mild hydrocracking scheme that incorporates the
ISOFLEX concept. The modified process provides economic advantages
of mild hydrocracking along with the capability for converting VGO,
CGO, AGO, and LCO into high-quality FCC feed and middle distillates.
Gas oil, which is in excess due to reduction in fuel oil production, can be
converted along with external distillates to valuable middle distillates.
Euent from the MHC reactor is flashed in a hot, high-pressure separator to produce bottoms oil that can be used as FCC feedstock and vapor
overhead. The latter is stripped with hydrogen and fed to a hydrotreating
reactor along with AGO, LCO, and/or recycle diesel, kerosene to be upgraded to high-quality middle distillates by desulfurization and aromatic
saturation reactions.
With this process scheme, the hydrocracking reactor needs to be
run only to meet the requirements of the FCC feed, specifically sulfur
content. External feeds go into the same high-pressure loop as the
hydrocracked vapor, which eliminates the need for additional gas compression equipment to process these external feeds.
Catalysis, 2016, 28, 86118 | 105
View Online
17:23:48.
Oil
Feed
H 2S
Absorber
First-stage
Reactor
Hot H2
Stripeer Water
Reactor Injection
External
Feed
Makeup
Hydrogen
HP Cold
HP Hot
Recycle
Diesel
Kerosene
To Fractionation
Sour water
To Fractionation
Middle distillate
Upgrader Reactor
Fig. 13 Chevron Lummus globals mild hydrocracking with ISOFLEX.
4.2.7 Single-stage, double-catalyst hydrocracking. SINOPEC has designed an improved process based on single-stage process. The singlestage in series process is known as Single-stage, Double-catalyst and
produces more middle distillates or FDC. Substituting a portion of
the hydrocracking catalyst in single-stage process with hydrotreating
catalyst lowers reaction temperature, increases middle distillate selectively, and decreases hydrogen consumption.27 An example of hydrocracking is given by an Iranian VGO in a single-stage with recycle of
4385 1C fraction. FDC process reduced mean temperature by 4 1C,
raised middle distillate by 3.0%, and lowered hydrogen consumption
by 0.16% when compared with conventional hydrocracking process.
A schematic diagram of the FDC process is presented in Fig. 14.
The FDC process has the advantages of simplicity, easy operation,
and reduced capital costs compared with the single-stage in series
processes that require separate reactors for hydrotreating and hydrocracking. The FDC process can be operated in once-through, partial
recycle, or full recycle configuration, depending on the desired distribution and outlet for UCO.
4.2.8 LCO hydrocracking. Light cycle oil (LCO) contains 60% to
85% aromatics, which leads to its high density and low cetane number. Hence, producing clean diesel is dicult through routine hydroupgrading technology from LCO. UOP and SINOPEC developed LCO
hydrocracking technologies, namely, LCO Unicrcaking28 and FD2G, respectively. The technology is highly selective in controlling hydrocracking reactions in partial conversion operating mode to achieve
good hydrogenation of diesel fraction, while preserving aromatics in
the gasoline range product. Depending on a refiners product needs
and product quality targets, aromatic ring manipulation is the key to
producing high-value gasoline and diesel from LCO. To accomplish
these parallel objectives, UOP and SINOPEC have combined process
innovations with HC 190 catalyst and FC-24B catalyst, respectively,
106 | Catalysis, 2016, 28, 86118
View Online
Makeup
Hydrogen
Recycle Hydrogen
17:23:48.
Fresh
Feed
Light Ends
Separator
Liquid Products
Unconverted Oil
Recycle Oil
Fig. 14 SINOPECs FDC hydrocracking process.
which were new, high-activity catalysts for eectively upgrading

LCO in partial conversion hydrocracking process configuration. LCO
Unicracking process uses partial conversion hydrocracking to produce
high quality gasoline and diesel stocks in a simple once-through flow
scheme, while SINOPEC uses single-stage process with partial recycle.
The pressure of this technology usually falls above that of highseverity hydrotreating, but is significantly low than that of a fullconversion unit.
LCO hydrocracking technology (FD2G) was applied in hydrocracking
unit I at Jinling Company, SINOPEC in 2013. The aim was to produce
gasoline with high research octane number and diesel with good quality
successfully, only by changing the catalyst system and optimizing process
conditions without any big unit modification. The results of industrial
application showed that LCO hydrocracking technology can produce 35%
to 50% gasoline blend stocks with high octane number of 91 to 94, sulfur
content less than 10 mg g1, and clean diesel blend stocks with sulfur
content less than 10 mg g1. In addition, the cetane number improved
10 to 14 units compared with feed stock.
The property of feed stock, major operating conditions, and main
products at the time of unit running 256 days are listed in Tables 3, 4,
and 5, respectively. The density of LCO feed is 0.948 g cm3, which has
aromatic contents of 65.1 wt% and cetane number 17. The feed rate of
unit is 75.1 t h1, average temperature of hydrocracking reactor is
393.9 1C, hydrogen consumption is 4.31%, and chemical hydrogen consumption is 3.80%. The research octane number of gasoline (65210 1C)
is 94.6, and sulfur content is 2.4 mg g1. The research octane number of
gasoline (150210 1C) is 97.1, and sulfur content is 0.7 mg g1. The cetane
number of diesel (4210 1C) is 31, which has improved by 14 units compared with feed stock.
4.2.9 SHEER hydrocracking technology. To reduce energy consumption, SINOPEC developed a high eciency hydrocracking technology
Catalysis, 2016, 28, 86118 | 107
View Online
Table 3 Property of industry feed stock.
17:23:48.
Jinling
Density (ASTM D 4052)
Distillation (ASTM D 86)
IBP-FBP
T50-T95
Cetane number (ASTM D 4737)
S
N
Aromatic content
g cm3
0.948
1C
1C
181357
269341
17
0.4201
0.0952
65.1
wt%
wt%
wt%
Table 4 Operating conditions of industry unit.

Item
t h1
t h1
MPa
h1
1C
%
%
Fresh feed
Recycle oil
Pressure
LHSV (hydrocarcking reactor)
Average temperature of hydrocracking reactor
The one-way conversion
Hydrogen consumption
75.1
7.0
10.7
1.33
393.9
50.95
4.31
Table 5 Properties of main products.
Fraction range
Density (ASTM D 4052)
Distillation (ASTM D 86)
IBP-FBP
T50-T90
Cetane number (ASTM D 4737)
RON
S
N
Aromatic content
Naphthene content
Paran content
gasoline
gasoline
diesel
1C
g cm3
150210
0.8554
65210
0.8248
4210
0.8721
1C
1C
121210
172194
97.1
0.7
o0.5
73.28
15.7
11.02
79204
144186
94.6
2.4
o0.5
62.01
25.81
12.18
229365
253329
31.0
16.3
48.5
13.4
38.1
mg g1
mg g1
wt%
wt%
wt%
named SHEER. The technology fulfilled the integration of high temperature and high pressure counter-current heat transfer, state-of-theart reactor internals, micro-swirl separation, and non-direct-contact
fouling prevention. The new start-up scheme for a set of partial autothermal hydrogenation technology was developed, and the new hydrocracking technology was invented for start-up working furnace. The
scheme lowered project investment and operation costs. The calculation results of SHEER hydrocracking process are based on a hydrocracking unit showing that high eciency SHEER hydrocracking
technology can reduce unit energy consumption by 44.85% and save
4.56% project investment. This scheme has good economic and social
benefits, as well as good application prospects. The schematic diagram
of the SHEER process is displayed in Fig. 15, and the inlet and outlet
108 | Catalysis, 2016, 28, 86118
View Online
Amine
Scrubber
Rich
amine
Makeup
Hydrogen
Fresh
Feed
Product
Oil
Fig. 15 Process flow diagram of 2.0 Mt a1 SHEER hydroprocessing unit in one company.
350
4500
300
3600
250
2700
200
1800
150
Concentration of H2 S/%
Inlet and outlet temperature of the reactor /C
17:23:48.
HP Cold
900
100
50
0
10
15
20
25
30
35
40
45
50
0
55
Time/h
Inlet temperature of the reactor;
Outlet temperature of the reactor;
Concentration of H2S
Fig. 16 Reactor inlet and outlet temperature during catalyst pre-sulfide.
temperatures of the reactor during catalyst pre-sulfide are shown in

Fig. 16.29
This technology, which developed a new wet start-up method for
hydrogenation unit, comprises the following steps:
(a) Utilize low-temperature heat source to heat up start-up activating
oil. This step leads to heated start-up activating oil and circulating
hydrogen into a hydrogenation unit that contains a sulfurized type
hydrogenation catalyst.
Catalysis, 2016, 28, 86118 | 109
17:23:48.
View Online
(b) Stop using the low-temperature heat source, and utilize the reaction heat generated from activating the reaction of the sulfurized type
hydrogenation catalyst to heat up the materials in the hydrogenation unit
to 180 10 1C. Introduce olefin-rich gas and/or olefin-rich light distillate
oil into the hydrogenation unit to carry out olefin hydrogenation
reaction.
(c) Utilize reaction heat generated from the olefin hydrogenation reaction to further heat up materials in the hydrogenation
unit to 230 10 1C. Keep materials at constant temperature for at
least 4 h.
(d) Further heating up to a temperature higher than 240 1C, gradually
reduce the introduction amount of olefin-rich gas and/or olefin-rich
light distillate oil. Replace start-up activating oil with feedstock gradually,
and utilize reaction heat generated from the feedstock hydrogenation
reaction to further heat up the materials in the hydrogenation unit until
the target temperature. Compared with prior technology, the present
invention cancels the heating furnace during start-up method of the
hydrogenation unit, and the start-up process operates smoothly and
steadily.
Kinetic models of hydrocracking
The development of an accurate kinetic model describing the hydrocracking process can provide a wealth of valuable information concerning the yields of industrial hydrocrackers. With an accurate kinetic
model, countless expensive experiment in pilot plants can be avoided by
the use of computer algorithm. Full potential of the fundamental model
is realized when operating conditions of the process are optimized to
provide product distributions, which are high in demand in the refining
industry. To predict product distributions of an industrial hydrocracker
and to obtain kinetic parameters, which are independent of feed stock
composition, the process must be described through its fundamental
chemical steps.
Kinetic modelling of hydrocracking with dierent approaches has
been reported in the literature. Kinetic studies consider each compound
and all possible reactions are complex due to the huge number of
hydrocarbons involved. However, these studies allow a mechanistic
description of hydrocracking based on detailed knowledge of the
mechanism of dierent reactions. In most cases, applying this method
to hydrocracking of real feeds is dicult because of analytical complexity and computational limitations. The situation is clear, the more
compounds a model includes, intrinsically is the need to estimate more
kinetic parameters. Consequently, more experimental information is
required.30
Kinetic models of hydrocracking processing are classified into (1)
models based on lumping technique,3138 (2) models based on continuous mixtures,3941 (3) structure oriented lumping, and (4) single event
models.
110 | Catalysis, 2016, 28, 86118
17:23:48.
View Online
5.1 Traditional lumping

The kinetics of hydrocracking of gas oil was studied by Qader and Hill42
in a continuous fixed-bed tubular flow reactor. These authors found that
the rate of hydrocracking was of first order with respect to feed concentration (Fig. 17), with an activation energy of 21.1 kcal mol1. Kinetic
data were obtained at 10.34 MPa pressure, 400500 1C temperature,
0.53.0 h1 space velocity, and a constant H2/oil ratio of 500 Nm3 m3.
The liquid product was distilled into gasoline (IBP-200 1C), middle distillate (200300 1C), and diesel (300 1C ). This research seems to be the
first experimental study in which kinetics of hydrocracking of real feed is
reported.
Orochko studied the kinetics of hydrocracking of vacuum distillates
from Romashkin and Arlan crude oils in a fixed-bed reactor over an
aluminacobalt molybdenum catalyst. The study used a first-order
kinetic scheme involving four lumps, according to the scheme showed
in Fig. 18. The rate of a first-order heterogeneous catalytic reaction is
expressed by the following equation:
at ln
1
by
1 y
where a is the rate constant, t is the nominal reaction time, y is the

total conversion, and b is the inhibition factor of the process by
the reaction products formed and absorbed on the active surface of the
catalyst and by their eects on the mass transfer in the heterogeneous
process.
Diesel fraction yield:
0
GO
1 yk 1 y
1 k0
k1
Products
Fig. 17 Reaction schemes for hydrocracking lump-kinetic models.
Diesel
VGO
Gasoline
Gases
Fig. 18 Reaction schemes for hydrocracking lump-kinetic models.
Catalysis, 2016, 28, 86118 | 111
View Online
Gasoline yield:
00
x k0
k
1 yk 1 yk
0 1 y 1 y
k
1 k0 k0 k00
1 k0 1 k00
00
17:23:48.
Gases yield:
g y (z x)
where k 0 and k00 are kinetic factors with similar meanings to rate
constants. Rate constants are determined from the experimental data
and are dependent on the equivalent kinetic temperature of the process and the catalyst activity. For the Romashkin petroleum vacuum
distillate at 1013 MPa, the values of k 0 and k00 are 1.3 and 2.0,
respectively.
Based on the principle of carbon balance in chemical reaction, a
method for calculating stoichiometric coecients of every reactant and
product in each individual hydrocracking reaction has been developed
by Guoqing Li and Yong Xia,43 using the distribution function in
Stangelands model. Consequently, hydrogen consumption of each
individual lump reaction can be calculated accurately, which results in
correct estimation of the heat of reaction, material balance closure, and
energy balance closure in each lump reaction process. An application in
an industrial VGO hydrocracker showed that the improved model,
which is based on the strict description on each individual lump
hydrocracking reaction, could simulate and predict the behavior of the
reactor precisely. In addition, the proposed model could predict temperature distribution along with the reactor and yield of products. The
model parameters were estimated by using the Genetic Algorithm based
on industrial data. The average error of the flow temperatures in the
outlets of the beds between calculated values and measured ones in the
plant was less than 0.99%.
SINOPECs Fushun Research Institute of Petroleum and Petrochemicals (FRIPP) has established a six-lumped kinetic model for
hydrocracking catalysts grading system. The principle of lump division is
as follows. First, crude oil (feedstock and hydrogenated oil) is divided
into six lumps according to fixed distillation interval, namely feedstockhydrogenated residue oil, diesel fraction, kerosene fraction, heavy
naphtha fraction, light naphtha fraction, and refinery gas lump. The
reaction scheme is depicted in Fig. 19. To predict hydrocracking product
distribution, the six-lumped kinetic model was established as follows.
The kinetic model parameters were regressed by hydrocracking experiments at dierent conversion degrees for two dierent hydrocracking
catalysts. The kinetic model parameters mainly included exponential
factors and activation energy. For the prediction of hydrocracking product properties, the function was established by product properties and
cracking conversions. Relative coecient (R2) was high enough to predict
the properties.
112 | Catalysis, 2016, 28, 86118
17:23:48.
View Online
Fig. 19 Reaction scheme of six-lumped hydrocracking kinetic model.
The equations are as follows:

dy1
k12 k13 k14 k15 k16 y1
dt
dy2
k12 y1 k23 k24 k25 k26 y2
dt
dy3
k13 y1 k23 y2 k34 k35 k36 y3
dt
dy4
k14 y1 k24 y2 k34 y3 k45 k46 y4
dt
dy5
k15 y1 k25 y2 k35 y3 k45 y4 k56 y5
dt
dy6
k16 y1 k26 y2 k36 y3 k46 y4 k56 y5
dt
0
kij kij0 *eEij =RT *PH2a kij0 *eEij =RT i; j 1 6

The lumped kinetic model parameters were regressed from the
hydrocracking experiment data of two dierent catalysts by fmincon
nonlinear least squares algorithm in Matlab 2011b (Fig. 20). Optimized
model parameters were used as the initial value to predict the product
distribution of hydrocracking catalysts grading experiment through
modulating partial parameters. A similar method was used to predict
product properties. The results showed that the calculation was in good
agreement with the experimental product yields and product properties,
and thus, had strong guidance for industrial process. Table 6 presents
the comparison of calculation and experiment results.
Catalysis, 2016, 28, 86118 | 113
View Online
Basic model of catalyst
Experiment data of A
Experiment data of B
Output as input
Catalysts combination
17:23:48.
Kinetic model of A
Kinetic model of B
Catalyst bed (A) above the hydrocracker
Catalyst bed (B) below the hydrocracker
Experimental results
Model calculation results

Comparison
Parameters optimization of model
No
No Parameters optimization of model

OK
Yes
Determine kinetic model parameters
End
Fig. 20 Calculation flow of kinetic model of hydrocracking catalyst combination.

Table 6 Comparison of calculation and experiment results.
Items
Feed
Yield of A bed-a
Yield of B bed-b
Experiment results
Residue-y1
Diesel-y2
Kerosene-y3
Heavy naphtha-y4
Light naphtha-y5
Gas (C4-)
0.9429
0.0571
0
0
0
0
0.5625
0.1085
0.1772
0.1359
0.0110
0.0050
0.3123
0.1322
0.2740
0.2398
0.0264
0.0154
0.3278
0.1369
0.2572
0.2180
0.0319
0.0151
5.2 Models based on continuous mixtures

Laxminarasimhan et al.44 developed a kinetic model for hydrocracking
a petroleum mixture based on the continuous theory of lumping.
The model considers the properties of the reaction mixture, underlying
pathways, and associated selectivity of the reactions. The parameter
of characterization is the true boiling point temperature (TBP). During
the reaction of a particular feed, the mixtures distillation curve
changes continuously inside the reactor. As residence time increases,
most of the heavier components are converted in to lighter components.
114 | Catalysis, 2016, 28, 86118
View Online
A normalized TBP as a function of an index (y) is used instead of the TBP.

Normalized TBP is defined by the following equation:
17:23:48.
TBP TBPl
TBPh TBPl
Reactivity is considered to be monotonic and can be represented by a

simple power law type function:
k
kmax
y1=a
where k is the reaction rate of a particular compound, kmax is the reaction

rate of the compound of higher TBP, and a is a model parameter. The
model equations are formulated as a function of reactivity following the
procedure proposed by Chou and Ho.45 To express the equation with k as
independent reactivity, a transformation operator is required, which is
approximated by equation:
Dk
Na
ka1
a
kmax
D(k) can be considered as a species-type distribution function, where

N represents the number of compounds in the mixture and tends toward
infinitum in a heavy fraction of oil.
Given the material balance of species of reactivity k, the core of the
kinetic model can be expressed with an integral dierential equation:
kmax
dCk; t
kCk; t
pk; K K CK; t DKdK
dt
k
5.3 Structure oriented lumping

Structure oriented lumping kinetic models employ most of the information obtained with modern analytical techniques for model reaction
modelling at a molecular level. These models have been proposed for some
catalytic processes. The lumps are defined according to the structure of the
compounds in there acting mixture. Liguras and Allen46 utilized contribution group concepts, which provide a mechanism for making use of
pure compound data in modelling complex reactions. They described the
conversion of vacuum gas oil in terms of a relatively large number of
pseudo components, most of which were lumps in their own right. Quann
and Jae47,48 developed a procedure to describe molecules and reactions
with a notation of vectors, which allowed a computer program to represent
reaction networks. These authors expressed chemical transformations in
terms of a typical structure of the molecules without completely eliminating lumps and rate parameters that depended on feedstock composition.
5.4 Single event models
Martens and Marin49 reported a model for the hydrocracking of hydrogenated vacuum gas oil based on theoretical and mechanistic
Catalysis, 2016, 28, 86118 | 115
17:23:48.
View Online
considerations. The reaction mechanism was described by a set of single

events, each of which could be ascribed a rate equation or a term in a
single rate equation. The model considered reaction rules for carbenium
ion of the secondary and tertiary types. A computer algorithm was used
for generating reaction networks. Froment50 has recently reviewed the
single event approach, which retained the full detail of the reaction
pathways of the individual feed components and reaction intermediates.
This approach was illustrated by means of methanol-to-olefins and
catalytic cracking of oil fraction reactions. Other important processes
with complex feedstock, such as catalytic reforming, hydrocracking,
alkylation, and isomerization, can be modelled with the single event
concept.
6 Consideration for hydrocracking technology

development
After the eorts of several generations of researchers, great improvements in hydrocracking technology has been achieved. As diesel demand
continues to grow and environmental regulations tighten, refiners will
begin to rely more heavily on hydroprocessing units to produce highquality and high-value products. Hydrocracking may emerge as the
favored primary conversion unit in refineries around the world, as
hydrocrackers are now more capable of handling low-quality feeds and
can provide increased supplies of diesel that more closely match final
product specifications. Conventional hydrocracking technology focuses
on upgrading feeds with very high boiling points and refractory feeds,
such as HCGO, LCO, into high-quality middle distillate products. Catalyst
developers continue to improve hydrocracking catalyst activity, increase
middle distillate selectivity, and reduce energy and H2 use via novel
formulations for active metal components and acidic catalyst support
materials. Research on the probing subjects must be combined with
fundamental study.
Implemented environmental regulations will aect hydrocracking
production, as the contaminant contents of transportation fuels become
more stringent. Sulfur content continues to be one of the main focuses of
transportation fuel regulations, and many countries currently are or will
be reducing permissible content of sulfur within these refined products.
Before long, many countries will require sulfur compositions of 5 ppm or
lower, and steps will be made to hydrocrackers towards improving
hydrodesulfurization capabilities of these units.
In the future, a major area of focus for technology developers and researchers will be hydrocracking unconventional feeds, such as residue
feeds, tight oil, and renewable sources. Another area that will continue to
be a main focus of technology developers is integration of the hydrocracker with other conversion units, such as delayed coker, solvent
deasphalter, FCC, etc., within the refinery to increase the yield and
quality of transportation fuels.
All in all, as one of the main approaches to process heavy oil, hydrocracking has the characteristics of good feedstock adaptability, flexible
116 | Catalysis, 2016, 28, 86118
View Online
flow scheme, high liquid yield, and superior product quality. Hydrocracking technology still has room to grow in terms of catalyst selectivity,
middle distillate maximization, and H2 use/energy eciency improvement.
References
17:23:48.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
J. Scherzer and A. J. Gruia, Hydrocracking Science and Technology, CRC Press,

1996.
K. H. Algelt and M. M. Boduszynski, Composition and Analysis of Heavy Petroleum Fractions, Marcel Dekker, Inc., New York, 1994.
X. Fang, Hydrocracking, China Petrochemical Press, Beijing, 2008.
J. W. Ward, Applied Industrial Catalysis, Academic Press, New York, 1984.
H. P. Company, Worldwide Refinery Processing Review, Hydrocarbon
Publishing Company, 2014.
R. A. Rakoczy and P. M. Morse, Hydrocarbon Processing, 2013.
2012 AFPM Q&A and Technology Forum: Answer Book, in 2012 AFPM Annual
Question and Answer Session, Salt Lake City, UT, 2012 [CD-ROM]; American
Fuel & Petrochemical Manufacturers: Washington, D.C., Salt Lake City, UT,
2012, p. 139.
A. Jiminez, Hydrocarbon Engineering, 2009, pp. 4849.
K. Bihler, D. Dominiak, B. Keith and J. Johnson, Hydrocarbon Process., 2012,
91(100), 9196.
R. D. Deepak, M. Goodman, A. Moreland and W. Whitecotton, in AFPM
Annual Meeting, Orlando, FLA, 2014.
W. Koester and A. C. Estrada, in LARTC 3rd Annual Meeting, Mexico, 2014.
D. Vannauker and H. Rasmussen, in 112th American Fuel and Petrochemical
Manufacturers Annual Meeting Conference, Hydrocarbon Processing, Daily, 2014.
S. F. Abdo, in AIChE Spring National Meeting, Orlando, FL, 2006.
OGJ International Refining Catalyst Compilation-2009, Oil Gas J., 2009, 9,
218.
F. Morel, in Proceedings of the 5th Bottom of the Barrel Technology Conference
and Exhibition, Athens, Greece, 2007.
T. E. A. Kalnes, in NPRA Annual Meeting, New Orleans, LA, 2001.
D. Hairston, Chem. Eng., 2003, 110(9), 2731.
K. P. Mc Cormick, in Proceeding of the 3rd European Catalyst Technology
Conference, London, 2002.
T. N. E. A. Kalnes, in European Refining Technology Conference, Paris, France,
2002.
S. Simpson, V. Thakkar, D. Ackelson and L. Wang, in 19th World Petroleum
Congress, Madrid, Spain, 2008.
A. S. E. A. Krishna, in European Refining Technology Conference Annual Meeting, Rome, Italy, 2000.
D. V. Law, A. Krishna, D. Farshid and A. Dahlberg, in Proceedings of the 1st
Bottom of the Barrel Technology Conference & Exhibition, Hamburg, Germany,
2000.
H. Virdi, G. Sieli and D. Torchia, in NPRA Annual Meeting, Phoenix, AZ, 2010.
L. G. Y. C. Y. X. W. Dehui, Pet. Process. Petrochem., 2010, 41, 2124.
V. T. Donald, B. Ackelson and B. Dziabala, NPRA Meeting, 2007, AM-07-47.
G. Brierley, in 4th NCUT Conference on Upgrading and Refining of Heavy Oil,
Bitumen, and Synthetic and Crude Oil, Alberta, Canada, 2006.
Z. R. Peng Chong and X. Huang, Pet. Process. Petrochem., 2014, 45, 5255.
S. F. Abdo, V. P. Thakkar, V. A. Gembicki and J. F. Mc Gehee, in NPRA Annual
Meeting, San Francisco, 2005.
Catalysis, 2016, 28, 86118 | 117
View Online
29
30
31
17:23:48.
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
F. X. Peng Chong and R. Zeng, Acta Pet. Sin., Pet. Process. Sect., 2015, 31,
706711.
nchez and M. A. Rodrguez, Catal. Today, 2005, 109, 7692.
J. Ancheyta, S. Sa
V. I. Topilnikov and M. K. Sosna, Chem. Technol. Fuels Oils, 2012, 48,
135142.
H. Zhou, J. Lu, Z. Cao, J. Shi, M. Pan, W. Li and Q. Jiang, Fuel, 2011, 90,
35213530.
V. Kumar and P. Balasubramanian, Fuel, 2009, 88, 21712180.
P. C. Krishna and P. Balasubramanian, Ind. Eng. Chem. Res., 2009, 48,
66086617.
nchez and J. Ancheyta, Energy Fuels, 2007, 21, 653661.
S. Sa
H. Kumar and G. F. Froment, Ind. Eng. Chem. Res., 2007, 46, 58815897.
nchez, M. A. Rodrguez and J. Ancheyta, Ind. Eng. Chem. Res., 2005, 44,
S. Sa
94099413.
P. Balasubramanian, S. J. Bettina, S. Pushpavanam and K. S. Balaraman, Ind.
Eng. Chem. Res., 2003, 42, 47234731.
M. Adam, V. Calemma, F. Galimberti, C. Gambaro, J. Heiszwolf and
R. Ocone, Chem. Eng. Sci., 2012, 76, 154164.
I. Elizalde and J. Ancheyta, Fuel, 2011, 90, 35423550.
H. Kumar and G. F. Froment, Ind. Eng. Chem. Res., 2006, 46, 40754090.
S. A. Qader and G. R. Hill, Ind. Eng. Chem. Process Des. Dev., 1969, 8, 98105.
G. Li, Y. Xia and W. Zeng, Fuel, 2013, 103, 285291.
C. S. Laxminarasimhan, R. P. Verma and P. A. Ramachandran, AIChE J., 1996,
42, 26452653.
M. Y. Chou and T. C. Ho, AIChE J., 1989, 35, 533538.
D. K. Liguras and D. T. Allen, Ind. Eng. Chem. Res., 1989, 28, 665673.
R. J. Quann and S. B. Jae, Ind. Eng. Chem. Res., 1992, 31, 24832497.
R. J. Quann and S. B. Jae, Chem. Eng. Sci., 1996, 51, 16151635.
G. G. Martens and G. B. Marin, AIChE J., 2001, 47, 16071622.
G. F. Froment, Catal. Rev., 2005, 47, 83124.
118 | Catalysis, 2016, 28, 86118
Titano-silicates: highlights on
development, evolution and application
in oxidative catalysis
17:23:54.
Ayomi Sheamilka Perera* and Marc-Olivier Coppens*

DOI: 10.1039/9781782626855-00119
Titano-silicates are a class of highly useful zeolite materials, predominantly used as

heterogeneous catalysts in selective oxidation. This chapter reviews the history and
significant advances in synthesis, characterisation and application of various types of
titano-silicates. In particular, selective oxidation catalysis of alkenes to epoxides is
discussed in detail as a key application of the material, while highlighting its potential
in green chemical processes. Finally, a brief overview of recent advances and future
prospects is given.
Introduction
1.1 Titano-silicates: a brief history

Titano-silicates can be generally characterised as zeolite derivatives,
containing isolated Ti centres embedded within the silica matrix. The
first titano-silicate, called titano-silicalite-1 (TS-1) was synthesised at
ENICHEM laboratories by Taramasso et al., in 1983.1 TS-1 is constituted
of an MFI structure in which, isolated Ti41 sites are located within the
silica framework, via isomorphous substitution of Si41. This material
has the general molar composition xTiO2(l x)SiO2, where x varies from
0 to 0.025.2
The primary application of the material at the time was selective and
ecient oxidation of various organic compounds. These include alkenes
to epoxides,3,4 hydroxylation of cyclic aromatic compounds,5,6 formation
of cyclooximes4 and oxidation of alkanes to ketones and alcohols.7,8 It
was established early on that the high catalytic activity of the material is
brought about by the tetrahedrally coordinated Ti41 centres within the
silicate matrix. Soon afterwards, another titano-silicalite (TS-2) with an
MEL structure was developed by Bellussi et al., in 1989.9 Since then, many
variations of TS-1 have evolved, targeting higher eciency as well as
dierent applications.1012 Examples of dierent variations of the
TS category include microporous types such as Ti-b, Ti,Al-b13 and
mesoporous types such as Ti-MCM-41,14,15 Ti-SBA-15,16 Ti-HMS17 and
Ti-MSU-1,18 among others. Extensive research is continuously conducted
to understand and analyse the active centre and overall crystal structure
of this highly versatile catalyst. Theoretical models combining QM/MM
with DFT calculations and shell model potentials were used recently by
Catlow et al.19 and To et al.20 to elucidate the structure of TS-1 (Fig. 1).
Department of Chemical Engineering, University College London, Torrington Place,

London, WC1E 7JE, UK. E-mail: ayomi.perera@ucl.ac.uk; m.coppens@ucl.ac.uk
Catalysis, 2016, 28, 119143 | 119

c
17:23:54.
View Online
Fig. 1 Structure of the catalytic active centre of TS-1 according to (A) Catlow et al.,
Adapted from ref. 19 and (B) To et al., Reproduced from ref. 20 with permission from
The Royal Society of Chemistry.
The most significant modification to the original TS-type material was

introduced in 1989, by Kuznicki, with the introduction of Engelhard titanosilicates (ETS-type).21 Developed at the Engelhard Corporation, USA and
named correspondingly, these materials dier predominantly from the TStype titano-silicates, in that the coordination of the active Ti centre is either
mixed octahedral/tetrahedral (Ex. ETS-4) or octahedral (Ex. ETS-1022).
The ETS-type titano-silicates were developed specifically to address
some of the drawbacks of the TS-type materials. Certain variations of
the latter were utilised as agents for cation relocation, ion exchange, dehydrating agents,23,24 adsorbents25 and membranes.2630 However, they
suered from framework deformations which largely limited these
applications. In contrast, it was demonstrated that the first member of the
ETS family, namely ETS-4, could withstand the strain of such applications
while retaining its structural integrity.31 The reason behind this enhanced
performance is the adjustability of pore sizes of ETS-4, accomplished via
controlled contraction by dehydration, which, in turn, renders tuneable
properties to the material. Hence, the structure of ETS-type materials is
inherently defective. This, however, is its most advantageous feature.
ETS-10 is a highly versatile adsorbent utilised to separate gas mixtures.3234 Other prominent members of the ETS family include ETS-1 and
ETS-2,35 whose remarkable ability to absorb heavy metal ions from
aqueous environments make them applicable in water purification.36
Other variations of Engelhard titano-silicates are also utilised for removal of dissolved heavy metals, such as Pd, Hg, Cd, Sr and Cs.3740
The third and most recent class of titano-silicates, namely, layered
titano-silicates were introduced by Roberts et al. and further characterised by Du et al. in 1996.41,42 By general definition, layered titano-silicates
consist of several layers of titano-silicates, interconnected with 4-ring
SiO4 tetrahedra, with isolated sites of TiO5 pyramids or TiO6 octahedra.
The interlayer space of the material is filled with water molecules and
cations such as Na1, Mg21 and Ca21.43 Since these ions are easily
interchangeable, layered titano-silicates can be applied as excellent ion
exchange materials. Examples of this class of titano-silicates include
120 | Catalysis, 2016, 28, 119143
17:23:54.
View Online
Fig. 2 SEM image of JDF-L1. Reproduced from ref. 45 with permission from The Royal
Society of Chemistry.
paranatisite, bafertisite, natisite, and most prominently, JDF-L1 (Jilin

Davy Faraday-Layered solid 1) (Fig. 2) and AM-4 (Aveiro-Manchester
material number 4).44
The first of such materials, JDF-L1 (Na4Ti2Si8O22 4H2O)45 is one of the
rare forms of synthetic titano-silicate which contains 5-coordinated Ti.41
Isolated TiO5 pyramids exist in the SiO4 layer, which have unshared vertexes
that point towards the interlayer space. The space between interlayers of
JDF-L1 (also referred to as AM-146 and NTS47) consists of water and Na1
ions. A similar variety is AM-4 (Na3(Na,H)Ti2O2[Si2O6]2 2H2O), which is
analogous to the mineral lintisite.48 In contrast to JDF-L1, AM-4 contains
TiO6 octahedra that interconnect with the SiO4 layer.49
Apart from ion exchange and adsorption, layered titano-silicates are
also potentially useful as pillaring templates for introducing inorganic
and organic metal oxides between the interlayer space.43,50 Pillared
materials have advantageous features such as enhanced catalytic
properties and structural stability. JDF-L1 has been synthesised using
glass fibres as a mount, as means of engineering nano-sized hydrogen
adsorbents.51 JDF-L1 and AM-4 have proven to be eective anti-bacterial
agents, when ion-exchanged with biocidal ions such as Ag1, Zn21 and
Cu21.52 The above two materials have also been utilised as growth
templates for pattern and size controlled overgrowth of MFI-type zeolite.44 It was found that the layered titano-silicates can act as nucleation
sites for guiding the growth of the surface zeolite, while imparting some
of their inherent characteristics onto it. This is a remarkable feature,
which can be utilised for potential generation of hierarchical catalysts
with mixed capabilities. Development of new types of layered titanosilicates is presently a highly active field of research.53,54
One of the most talked about layered titano-silicate materials in recent
years is the Ti-MWW type. The most important breakthrough in Ti-MWW
synthesis was achieved in 1995, with B-containing MWW zeolite ERB-1
via use of boric acid as supporting agent.55 This material had superior
Catalysis, 2016, 28, 119143 | 121
View Online
17:23:54.
Table 1 Summary of the history of the development of prominent titano-silicates.

Titano-silicate
type
Developed by
Year
Major application/s
Key features
Ref.
TS-1
TS-2
Ti,Al-b
Ti-MCM-41
Ti-SBA-15
ETS-4
ETS-10
ETS-1
ETS-2
JDF-L1
AM-4
Ti-MWW
Tramasso et al.
Bellussi et al.
Sato et al.
Corma et al.
Morey et al.
Kuznicki et al.
Kuznicki et al.
Kuznicki et al.
Kuznicki et al.
Roberts et al.
Lin et al.
Corma et al.
1983
1989
1994
1994
2000
1989
1990
1991
1991
1996
1997
1999
Oxidative catalysis
Oxidative catalysis
Oxidative catalysis
Oxidative catalysis
Oxidative catalysis
Ion exchange, adsorbent
Oxidative catalysis
Microporous
Microporous
Microporous
Mesoporous
Mesoporous
Microporous
Microporous
Microporous
Microporous
Layered
Layered
Layered
1
9
13
14
16
21
22
35
35
41
48
56
properties in terms of activity and crystallinity. The first successful use of

Ti-MWW type material (Ti/ITQ-2) in alkene epoxidation was reported by
Corma et al. in 1999.56 Although highly ecient and selective for the above
reaction, the material could only be used with TBHP as oxidant and was
inactive with H2O2, owing to its increased hydrophobicity. Since then many
variations of the Ti-MWW type have been investigated for catalysing
epoxidation reactions, including B-free type.57,58 It has also gained prominence in ammoximation of methyl ethyl ketone, surpassing TS-1 in terms
of eciency and selectivity.59 At present Ti-MWW is, arguably, the only
other titano-silicate with the potential to surpass the industrial scale
application of TS-1, in terms of activity, stability and selectivity.
Table 1 summarises a few chemically and industrially relevant titanosilicates along with their approximate date of discovery, principal
applications and pore type.
Synthesis and characterisation of titano-silicates
2.1 Evolution of dierent synthetic techniques

Typically, synthesis of titano-silicates is carried out by a hydrothermal
process that employs temperatures ranging from 150230 1C and reaction
times ranging from several hours to ca. 30 hours.1 Early syntheses were
carried out in autoclaves lined with Teflon. The initial synthesis of TS-1
was described by Taramasso et al. and consists of tetraethyl-ortho-silicate
(TEOS), tetra-ethyl-ortho-titanate (TEOT) and tetrapropyl-ammoniumhydroxide (TPAOH).1 This method is often referred to as the classical
method of titano-silicate synthesis.6 Another hydrothermal synthesis
method of a TS-1 derivative, utilising (n-C3H7)NOH, was introduced
by Peregot et al.60 This material had enhanced selectivity in alkene
epoxidation, together with higher yields. Significant variations of this
synthesis followed, utilising dierent starting materials and techniques.
This has been an area of growing research interest since. The early
variations included the so called TS-2 with a MEL structure, which later
developed to incorporate elements other than Ti.9,61 Procedures that
utilise much higher reaction temperatures such as 750 1C were reported by
122 | Catalysis, 2016, 28, 119143
17:23:54.
View Online
Harrison et al. to achieve single crystals of a derivative of the mineral

pharmacosiderite, analogous to titano-silicate (Cs3HTi4O4(SiO4)3 4 H2O).62
This synthesis required 40 hours and was carried out in a gold tube.
Apart from direct hydrothermal synthesis, two other well-known
methods of titano-silicate synthesis are the drygel conversion (DGC)
method63 and the post-synthesis method.64 DGC is an ecient technique
for the synthesis of TS-2 with a lower Al31 content and smaller crystal
size.65 The solgel method is a useful synthesis technique, which allows
greater control over reaction parameters and allows the use of hydrothermal or DGC routes.6668 The disadvantages of this technique,
however, are that it produces non-crystalline material with heterogeneities in pore volume and framework chemistry, together with low
surface area. Kosuge et al. attempted to address some of these issues by
employing a modified solgel method with TEOS and TEOT as framework
sources and aqueous octylamine as a template, to produce novel
mesoporous titano-silicates, MTP.69 These catalysts have the advantages
of an increased specific surface area and a more homogeneous distribution of Ti. Li and Coppens extended this technique further by using a
surfactant/oil mixture for templating in order to achieve mesoporous
hollow Ti-silica spheres with increased surface area and robustness.70
It was established early on that particle size of TS-1 directly aect
its catalytic activity, with smaller particles showing higher activity due
to decreased diusion limitations.71 Later on, post-synthetic treatments evolved, one of the most prominent being the use of tetrapropylammonium hydroxide (TPAOH) as post-treatment agent. The latter
creates hollow TS-1 particles after calcination, with greater surface area,
leading to enhancement of catalytic activity and lifetime.7274 This process
has gained industrial relevance in recent years. Similar treatments have
been conducted using Ti-MOR, where implantation of fluoride species via
post-treatment leads to a large increase in activity for ammoximation of
cyclohexanone.75
2.2 Characterisation techniques: past to present
Early characterisation was aimed towards determining Ti : Si ratios and
confirming the tetrahedral geometry of the Ti41 centres. The original
characterisation of TS-1 by Notari involved a study of the change in unit
cell volume of pure silicalite crystal, after introduction of Ti41.6 This
phenomenon is explained by:

3
dTi
Vx VSi VSi 1 3 x
(1)
dSi
In here, VSi is the unit cell volume of pure silicalite and dTi and dSi are
the tetrahedral TiO and SiO bond distances, respectively. The factor x is
known as the degree of isomorphous substitution, and is calculated by the
atomic ratio Ti/(Ti Si). It was found that the upper limit of x is equal to
0.025.76 When the Ti content was increased to values higher than 0.025,
the correlation between Ti and the unit cell volume was lost and formation
of an additional phase of TiO2 was observed.6 This is a highly undesirable
Catalysis, 2016, 28, 119143 | 123
17:23:54.
View Online
Fig. 3 Original experimental values for unit cell volume change vs. x. Reproduced from
B. Notari, Titanium silicalites, Catal. Today, 18, 163172. Copyright (1993) with permission
from Elsevier.6
feature, as TiO2 does not contribute to catalysis of alkene epoxidation and

causes decomposition of the oxidant hydrogen peroxide.77
Millini et al. performed XRD on titano-silicates and used Rietveld
analysis to confirm the above.2 Deo et al. conducted a comprehensive
analysis of the titano-silicates as well as the titania (TiOx) species present
in titano-silicate materials, using Raman spectroscopy, methanol oxidation, X-ray absorption spectroscopy and pyridine adsorption.78 These
pioneering studies shed light into the fundamental relationship between
the structure of titano-silicates and that of titania species present within
them, and how it aects the reactivity of the former. It was found that
stability of the TiOSi bond in the presence of liquid water plays a
crucial role in the catalytic eciency of the material (Fig. 3).
The structure of ETS-10 was elucidated by Anderson et al.79,80
The disordered structure, together with the fact that it could only be
synthesised in powder form, proved to make accurate structural
determination challenging (Fig. 4). A combination of various analytical
techniques, such as 29Si-NMR, HRTEM (High Resolution Transition
Electron Microscopy), chemical analysis, XRD (powder X-Ray Diraction)
and DLS (Distance-Least-Squares) had to be employed for a comprehensive structural analysis.
At present, a combination of key characterisation techniques is used to
accurately determine the structure of titano-silicates. These include FTIR
and Raman spectroscopy, solid state UV-vis spectroscopy, 29Si MAS-NMR,
nitrogen adsorption, SEM, TEM, TGA and XRD58,60,47,81,82 (Table 2).
Application of titano-silicates as oxidative catalysts
3.1 Oxidation of alkenes to epoxides

Titano-silicates are selective and eective catalysts for the epoxidation of
alkenes16 (Scheme 1). Epoxides from alkenes such as propene and
124 | Catalysis, 2016, 28, 119143
17:23:54.
View Online
Fig. 4 The structure of ETS-10, showing defects and dierent chemical and crystallographic environments. Reproduced with permission from ref. 82 with permission from
John Wiley and Sons r 2000 WILEY-VCH Verlag GmbH, Weinheim, Fed. Rep. of
Germany.
cyclohexene are industrially relevant precursors in pharmaceutical, food

and agro-chemical industries.83,84
One of the major challenges in epoxidation of straight chain and cyclic
alkenes is the formation of undesired side products. Before the advent of
titano-silicates, these reactions were catalysed via group IVVI metal
oxides.12 For example, a typical reaction using methanol as solvent and
hydrogen peroxide (H2O2) as oxidant rendered mixtures of alcohols, dialcohols, ketones and ether products. Nowak et al. describes the possible
side products that can be formed via dierent reaction pathways in the
epoxidation of cyclohexene, using various zeolites85 (Scheme 2).
The selectivity of TS-type titano-silicates towards the formation of
the desired product epoxide is truly remarkable. One of the most
advantageous features of TS-1 is its ability to selectively form epoxide
from double bonds, regardless of any other functional group present, in
the reactant molecule.12 Straight-chain alkenes, for example, are converted in nearly quantitative yields and formation of undesired side
products is negligible.86,87 The rate of epoxidation is higher for smaller,
unrestricted alkenes versus bulky compounds, since diusion into the
Catalysis, 2016, 28, 119143 | 125
View Online
17:23:54.
Table 2 Characterisation techniques and features of titano-silicates.

Technique
Characteristic Signal
Key characterisation features
FTIR
960 cm1
Raman
955 cm1
1100 cm1
Absence of 140, 395, 515, 640 cm1,
210 nm
TiOSi asymmetric stretching

mode
TiOTi symmetric stretch
TiOTi antisymmetric stretch
Absence of anatase TiO2
Ligand-to-metal-chargetransfer (LMCT) band for
4-coordinated Ti
LMCT band for 6-coordinated
Ti
Q1 [Si(OSi)1(OH)3],
Q2 [Si(OSi)2(OH)2],
Q3 [Si(OSi)3(OH)], and
Q4 [Si(OSi)4]
Crystalline/amorphous nature
Solid-state UV-vis
(Kubelka-Munk)
225 nm
29
Si CP MAS-NMR
XRD
N2-isotherms
TGA
89
96
104 and
111.5 ppm
Characteristic spectrum or peaks
for crystalline materials
Nature and shape of hysteresis,
formed by adsorption and
desorption isotherms
SEM
Sharp weight loss bands

Constant weight bands at high
temperature
Surface image
TEM
Image of particles in transmission
Alkene
(in excess)
Ex.
Propene
Oxidant
(limiting reagent)
Ex.
H2O2
O O
TS catalyst
Conditions
Specific surface area

Pore volume
Average pore diameter
Pore size distribution
Adsorbed species
Thermal stability
Physical morphology of
particles
Physical morphology and
porosity of particles
Epoxide + alcohol by
product + other side
products
1-Hexene
O
H O
Cyclopentene
Cyclohexene
Scheme 1 General reaction scheme for epoxide formation from alkenes, catalysed by
titano-silicates.
TS-1 catalyst is hindered. This would explain why the epoxidation rate of
cyclohexene is much slower than that of 1-hexene.87
Other TS-type titano-silicates, such as TS-2, Ti-b, Ti,Al-b, and the
layered type Ti-MWW are also eective as selective epoxidation catalysts.
Ti-b and Ti-MWW type catalysts were developed to facilitate the
126 | Catalysis, 2016, 28, 119143
View Online
OH
17:23:54.
OH
OH
OH
OR
OH
Scheme 2 Possible pathways and side products formed during cyclohexene oxidation,
using zeolite catalysts. Adapted from I. Nowak et al., Epoxidation of cyclohexene on Nbcontaining meso- and macroporous materials, Catal. Today, 78, 487498, Copyright
(2003) with permission from Elsevier.85
epoxidation of hindered alkenes (e.g. norbornene, cyclododecene,

methylcyclohexenes, camphene, methyl oleate, etc.), which were unable to
diuse into the pores of TS-1.58,8890 Ti-MWW catalyst is unique in terms of
isomer-selectivity. According to Wu et al., from a mixture of 50 : 50 linear
cis/trans-alkenes, Ti-MWW is able to selectively catalyse epoxidation of
the trans-isomer for ca. 80%.91 This unprecedented trans-selectivity is not
aected by the type of alkene or the reaction conditions. This is said to be
due to the sinusoidal 10-MR channels present in Ti-MWW.
For hindered and bulky alkenes, b-type and Ti-MWW titano-silicates
are found to be more eective. Wu et al. have investigated relative
catalytic eciencies for various epoxidations with Ti-MWW, Ti-b and
TS-1. Ti-MWW in acetonitrile and water has shown superior activity for
synthesis of glycidol.92 Epoxidation of diallyl ether was also found to be
more ecient with Ti-MWW than TS-1, when the reaction is conducted in
acetonitrile and acetone.93 Indeed, such solvent eects have proven to be
a key factor in catalytic eciency in various titano-silicate materials.94
The increased pore volume and bi-functionality of certain b-type titanosilicate catalysts, predominantly Ti,Al-b, is advantageous for epoxidation
of bulky organic alkenes. Cyclohexene and cyclopentene have been
converted to trans-2-alkoxycycloalkanols, which are important pharmaceutical intermediates, using Ti,Al-b as catalyst.95 Other industrially
relevant examples include epoxidation of linalool and a-terpineol.96,97
3.2 Use of H2O2 as oxidant in alkene epoxidation
H2O2 is a highly eective oxidant for epoxidation of alkenes. Epoxidation of propene, for example, converts 490% of H2O2 with a product
selectivity of 496%.86,98 The use of H2O2 is also favourable in that only
dilute solutions are required at or near room temperature.98 However, the
reaction is not without limitations. Water is formed as a side product
during decomposition and can act as a catalyst poison by adsorbing onto
the catalyst surface. In addition, when water is present, a TS catalyst can
act as a weak acid to cleave the epoxide ring, resulting in formation of
open chain alcohols. One solution is to add small quantities of bases
Catalysis, 2016, 28, 119143 | 127
17:23:54.
View Online
in order to neutralise the acid sites of the catalyst, but at the price of
reducing catalyst activity.86
Propene epoxidation is an industrially significant reaction using
titano-silicates as catalysts and as H2O2 oxidant. Much eort has been
directed towards optimising the process of using H2O2 for propene
epoxidation over the years. In 1998 Hayashi et al. introduced a procedure
where a catalyst composed of gold nanoparticles incorporated on titania
(AuTiO2) was used together with an Oxygen/hydrogen mixture, which
generated H2O2 in situ for propene epoxidation.99 Later on, Nijhuis et al.
demonstrated that if the catalyst was changed to a titano-silicate type
material, the product yields increased significantly.100 Since then, other
titano-silicate materials, such as Ti-SBA-15,101 TS-1102,103 and Ti-SiO2104
have been investigated for their use in the above application. The process
achieved considerable industrial interest when several companies, including Dow, BASF, Evonik and SKC launched plants which couple
the epoxidation process with synthesis of H2O2, the latter via the
anthraquinone process.105,106
3.3 Use of TBHP as oxidant in alkene epoxidation
In order to counteract the disadvantages of using H2O2, organic hydroperoxides, of which the most significant is TBHP (tert-butyl hydroperoxide), were introduced as an alternative.107109 During early stage
research, studies were conducted to compare the rates of epoxide
formation between H2O2 and TBHP, with various types of alkenes and
dierent titano-silicate catalysts.110 In general, TBHP showed lower rates of
epoxidation than H2O2. This was attributed to the diculty of formation of
the sterically hindered TBHP-titano-silicate intermediate inside the zeolite
pores and the low electrophilicity of the TiOOR, compared to TiOOH.
However, TBHP had other superior qualities: it showed almost 100%
epoxide selectivity for almost all alkenes under mild conditions and it is
also a better oxidant for catalysis with large-pore titano-silicates, such as
Ti-b. It also did not show the epoxide decomposition that is associated
with H2O2 when used together with methanol.
The ability of TBHP to selectively oxidise bulky alkenes with other substituent groups was further investigated by Bhaumik et al.110 Building on
previously published work that showed enhanced catalytic performance of
silylated titano-silicates,111 they synthesised Ti-MCM-41, functionalised
with various straight-chain and aromatic organic groups, such as methyl,
allyl, vinyl, pentyl, 3-chloropropyl, and phenyl. The 3-chloropropylTi-MCM-41catalyst showed superior epoxidation performance for sterically
hindered unsaturated alcohols with TBHP, compared to unmodified
Ti-MCM-41. The epoxidation and consequent formation of cyclic ethers,
was drastically better with TBHP, compared to H2O2, with 4 to 5-fold increased rates. Wu et al. demonstrated that epoxidation of cyclohexene
using Ti-MWW catalyst is more ecient with TBHP than with H2O2.112
Overall, it is evident that TBHP is a better oxidant to be used with
mesoporous titano-silicate catalysts than with their microporous
analogues, for the epoxidation of alkenes. This is largely due to the fact
that the relatively larger pores of the mesoporous material facilitate
128 | Catalysis, 2016, 28, 119143
View Online
17:23:54.
Table 3 Preferred epoxidation substrates for various titano-silicates against H2O2 and
TBHP.
Titano-silicate
Preferred oxidant
Preferred substrate
Ref.
TS-1
Ti-b, Ti,Al-b
Ti-MCM-41
Ti-MWW
H2O2
TBHP
TBHP
TBHP
Straight-chain alkenes
Bulky and functionalised alkenes
Bulky alkenes
Bulky and cyclic alkenes
86, 87
8890, 9597
89
58, 92, 93
diusion of the hydrophobic TBHP molecules. It is also a better oxidant

with microporous catalysts, such as Ti-b, when utilised for epoxidation of
bulky substrates.
Table 3 summarises the preferred oxidation substrates for dierent
types of titano-silicate catalysts.
Novel types of titano-silicate catalysts are constantly being added to the
sea of already existing literature on the subject. It is undeniable that the
new categories are designed around the use of TBHP as a preferred oxidant. Kosuge et al. employed the novel titano-silicate MTP for epoxidation of 1-hexene, 1-octene, 1-decene and cyclohexene with TBHP.69 The
shorter-length alkenes showed higher activity and the overall selectivity was
close to 100% for all substrates. Baca et al. utilised a previously synthesised70 Ti-silica catalyst, produced as hollow, mesoporous particles via a
surfactant-templating technique, as an ecient catalyst for cyclohexene
epoxidation with TBHP.113 The latter was also successfully used as an oxidant under mild conditions by Modak et al., for epoxidation of several cyclic
alkenes, via a novel organic-hybrid-titano-silicate catalyst, Ti-LHMS-3.114
There are however, some limitations to the use of TBHP. The most
industrially relevant titano-silicate to date, TS-1, cannot be used with
TBHP as oxidant for alkene epoxidation.12 This is due to the fact that
TBHP molecules are too bulky to diuse into the micropores of TS-1.
Also, epoxidation reactions generally take longer times with TBHP
compared to H2O2. Most importantly, strictly anhydrous conditions need
to be employed when utilising TBHP, whereas such drastic measures are
not required with H2O2, even though water can poison the catalyst when
using H2O2 as well.
3.4 Reaction mechanisms
It is widely accepted that the isolated Ti centre of titano-silicates needs to
be coordinated in a specific geometry in order to be catalytically active.
The catalytic mechanism, however, has been a topic of controversy and
much debate. Notari postulated that there are three major pathways
for oxygen transfer from the oxidant to the Ti41 centre, during catalysis:
(1) electrophilic oxidation, (2) metallacyclo intermediates, and (3) radical
intermediates.6 At the time, two intermediate structures between H2O2
and the titano-silicate active centre were proposed for the electrophilic
oxidation mechanistic pathways (Fig. 5).
Later on, other hypotheses based on theoretical studies suggested
dierent intermediates involved and dierent methods of oxygen transfer. Vayssilov and van Santen conducted rigorous DFT calculations
Catalysis, 2016, 28, 119143 | 129
View Online
OH2
Ti O
H 2O O
OH2
Ti O OH
OH2
(1)
(2)
17:23:54.
Fig. 5 Intermediate structures proposed by Notari for adsorption of H2O2 on titanosilicate active site. Adapted from B. Notari, Titanium Silicates, Catal. Today, 18, 163172,
Copyright (1993) with permission from Elsevier.6
142
H
133
C H
H C
H
O H
H O
149
C H
C
H
151
H
146
H
C
145
211
O
H O H
Scheme 3 Mechanism of alkene epoxidation according to Vayssilov and van

Santen (bond distances indicated are in pm). Adapted from Catalytic Activity of Titanium
Silicalites a DFT study, J. Catal., 175, 170174, Copyright (1998) with permission from
Elsevier.115
involving energy barriers of catalytic steps of alkene epoxidation by

titano-silicates.115 Their study showed that cleavage of the OH bond in
H2O2 is unnecessary and is unlikely during catalysis. This conclusion was
based on two observations: (1) the energy barrier for non-dissociative
catalysis of alkene epoxidation by H2O2 is much lower than the energy
required for adsorbed H2O2 deprotonation, and (2) the calculated energy
barrier for the first step in Scheme 3 is much lower that its gas phase
analogue without the catalyst.
Sankar et al. conducted extensive theoretical studies using DFT
calculations combined with X-ray absorption fine structure (XAFS)
spectroscopy to investigate the geometry of the 6-coordinated Ti41 centre
of calcined and reacted TS-1 and Ti-MCM-41 and model compounds
Ti-(OSiPh3)4, Ti-(OGePh3)4, fresnoite (Ba2TiSi2O8), JDF-L1 and ETS-10116
(Fig. 6). Based on the intensity and position of the pre-edge-peak
observed for these compounds, they concluded that Ti41 is in a tetrahedral geometry for all the compounds.
It is now generally accepted that heterogeneous catalysis by hetero
atom-substituted zeolites is facilitated via an open active site in the
catalytic active centre.117 Open active sites are also accepted to be
responsible for a variety of other reactions, such as CO adsorption onto Pt
surfaces,118,119 coking,120 watergas shift catalysis,121 oxidation122 and
synthesis of ammonia.123 The open active site of a titano-silicate catalyst,
according to Nandi et al.,124 is depicted in Fig. 7.
The mechanism of alkene epoxidation by titano-silicates was further
investigated by Notestein et al. via theoretical and experimental studies
(Scheme 4).125 They describe a general catalytic mechanism, applicable
for titano-silicates of the TS-type, and other catalysts with isolated Ti
centres with 4-fold coordination. According to this catalytic mechanism,
130 | Catalysis, 2016, 28, 119143
17:23:54.
View Online
Fig. 6 Plot of pre-edge intensity and position for model compounds and calcined and
reacted (with TBHP) Ti-MCM-41 compounds. Reproduced with permission from
G. Sankar, J. M. Thomas, C. R. A. Catlow, C. M. Barker, D. Gleeson and N. Kaltsoyannis,
J. Phys. Chem. B, 2001, 105, 9028. Copyright (2001) American Chemical Society.116
O
Si
HO
Ti
O
H O
Si
O
Si
Si
Fig. 7 Open active site of titano-silicates.124
described with TBHP as oxidant, there are three basic steps: (1) reversible
coordination of hydroperoxide to the Ti active site, (2) electrophilic attack
of alkene to O1 oxygen in bound peroxide, and (3) dissociation of alcohol
by-products. Epoxidation is said to occur either as rate determining step
or prior to it. The overall rate constant k is given as k k3K2.
Similar to the reaction mechanism, the kinetics of titano-silicate
catalysed reactions have also been debated. The LangmuirHinshelwood
dual-site mechanism is described as a possible option, leading to:126
rPO k1 TS 1
K1 K2 H2 O2 C3 H6
1 K1 H2 O2 K3 C3 H6 1 K2 C3 H6 K4 CH3 OH
(2)
However, the latter was found to be only applicable for lower reactant
concentrations, which are not representative of industrial scale
Catalysis, 2016, 28, 119143 | 131
View Online
O
O
Ti
K2
Ti
HO
17:23:54.
H
O
k3
O
O
Ti
Ti
Ti
H
K4
O
O
O
Ti
HO
Scheme 4 Mechanism of alkene epoxidation by titano-silicates according to Notestein

et al. Adapted with permission from J. M. Notestein, A. Solovyov, L. R. Andrini, F. G.
Requejo, A. Katz and E. Iglesia, J. Am. Chem. Soc., 2007, 129, 15585. Copyright (2007)
American Chemical Society.125
processes. The most accepted kinetic mechanism is the Eley-Rideal type

mechanism, as most recently confirmed by Russo et al.127 and supported
by previous studies; this leads to the following expression for the reaction
rate:128,129
rPO k1 TS 1
K1 H2 O2 C3 H6
1 K1 H2 O2 K2 C3 H6 K3 H2 O K4 PO
(3)
3.5 Developments towards green catalytic processes

Use of titano-silicates as catalysts for oxidation of alkenes via H2O2 has
been called a green process, due to the remarkable eciency of the
catalyst and water being the only byproduct of the oxidant.130 However,
careful scrutiny of the synthesis procedure of H2O2, together with its
production cost, challenges the very definition of it being characterised
as a green chemical (Table 4).
Industrial scale synthesis of H2O2 is predominantly achieved via the
anthraquinone auto-oxidation process (Scheme 5).131,132 Typically, 2-alkyl
substituted anthraquinone (A) is reduced to its di-alcohol counterpart,
anthrahydroquinone (B), catalysed by supported Pd or Pt. The anthrahydroquinone is subject to oxidation in air to regenerate the anthraquinone with equimolar amounts of H2O2. The di-alcohol compound can
also convert to tetrahydro-anthrohydroquinone (C), which can also be
oxygenated to form H2O2. The produced H2O2 needs to be separated from
the reaction mixture via water based counter current extraction.
Not only is the above process extremely expensive and complex, it
also involves costly and time consuming separation and purification
procedures for the product. Since use of commercially available H2O2
would not generate sucient profit in industrialised propene oxide
132 | Catalysis, 2016, 28, 119143
View Online
17:23:54.
Table 4 Summary of industrial and green applicability of some major titano-silicates.
Titano-silicate
Oxidant
TS-1
H2O2 (from
anthraquinone)
H2O2 (from O2 H2)
H2O2 (from O2 H2)
H 2 O2
TBHP
TS-1
TS-1/Pd/Pt
Ti-MWW
Ti-MCM-41
Application at
industrial scale
Application in
green chemical
engineering
High
Yes
No
Low
Low
High
Low
No
No
Yes
No
Yes
No
Yes
Yes
H2
H2
Pd or Pt
Pd or Pt
OH
OH
R
Overall
epoxidation
yields
B
O2
OH
OH
O2
O
R
H 2O 2
H 2O 2
O
Scheme 5 Production of H2O2 via anthraquinone auto-oxidation process. Adapted from

C. Samanta, Direct synthesis of hydrogen peroxide from hydrogen and oxygen: An
overview of recent developments in the process, Appl. Catal., A, 133149. Copyright
(2008) with permission from Elsevier.132
production,133 on-site production techniques had to be developed.134 In

addition, the large amounts of H2O2 needed for industrial scale propene
oxide production carries significant risks in transportation and handling.
If the above facts do not expose the non-green nature of H2O2 production,
the toxicity and carcinogenicity of the major starting material anthraquinone should certainly suce.135
Fortunately, alternative pathways that lead to more sustainable processes have been introduced over the years. Major alternative pathways of
H2O2 production include direct synthesis from elemental O2 and H2,
process integration and in situ production.12 One such approach is to
integrate the propylene oxide epoxidation with H2O2 synthesis, but with
replacement of anthraquinone as starting material to a mixture of O2 and
H2, in a plasma reactor.136 This method is not only greener, but also does
not require a solvent and the product obtained is in vapour form, hence
is pure and in no need of further separation.137 An industrial adaptation
of the process was carried out by Klemm et al.138 Previous studies have
also indicated that the product selectivity can be increased by utilisation
of low temperatures, and addition of mineral acids and additives.131
Unfortunately, the product yields and H2O2 conversion rates for these gas
Catalysis, 2016, 28, 119143 | 133
17:23:54.
View Online
phase reactions were very much inferior to their liquid-phase counterpart. Hence, industrial scale application is still a challenge.
In situ generation of H2O2 is another viable alternative for sustainable
synthesis. One modified process utilised TS-1 incorporated with Pd to
create a bi-functional catalyst.139 Pd catalyses the formation of H2O2 from
O2 and H2, which subsequently diuses into the TS-1, where propene
epoxidation takes place. A major disadvantage of this method is that Pd
acts as a strong reducing agent and causes the reduction of propene to
propane. Another drawback is that when methanol is used as a solvent,
methyl formate is formed from its oxidation, together with oxidation
products from propene, such as acrolein and acetone.140 These unwanted
products exceed the concentration of epoxide formed after several hours.
A solution to reduce the formation of such side products is to use
supercritical CO2 instead of methanol or water as solvent.141
Process integration is perhaps the most ecient alternative to counteract the problems associated with H2O2. However, these processes do
not necessarily lead to green technologies. Since this technique uses
H2O2 without further processing directly into the epoxidation reactor,
additional steps of purification, concentration and stabilisation are no
longer required, which results in significant cost reduction. Starting
materials for H2O2 generation include anthraquinone, isopropanol and
1-phenylethanol,128,142 which do not qualify as green materials.
4 Outlook: is the future bright for titano-silicate

catalysts?
The preceding pages provide ample evidence for the importance of
titano-silicates in selective heterogeneous catalysis. Three decades into
their advent, progress of novel research in terms of synthesis and applications is as alive and dynamic as ever. This final section attempts to
review a few of the most recent trends observed in titano-silicate research
and provide insight into their future role in the field of catalysis.
4.1 Hybrid materials with enhanced applications
The incorporation of organic components into titano-silicates gives
rise to new types of organic-inorganic hybrid catalysts with new and
enhanced features.
Modak et al. synthesised a new titano-silicate Ti-LHMS-3, functionalised with a phloroglucinol moiety.114 The organic component was
introduced via a silica precursor containing bis(propyliminomethyltriethoxysilyl)-phloroglucinol. Synthesis of the latter was achieved
through formylation of phloroglucinol via a VilsmeierHaack reaction
and subsequent Schi base formation with 3-aminopropyltriethoxysilane
(APTES). This hybrid catalyst has demonstrated reasonable eciency and
selectivity in epoxidation of several cyclic alkenes, with cyclohexene being
the best in terms of conversion. The catalyst is found to remain active up
to five cycles with marginal loss of activity. However, it is not superior to
well-established titano-silicates such as TS-1 and Ti-MWW, either in
134 | Catalysis, 2016, 28, 119143
17:23:54.
View Online
eciency or selectivity. In fact, selectivity towards epoxide is decreased

after a few hours into the reaction, when formation of diol species takes
precedence.
Burmann et al. reported on the development of polysulfone membranes
embedded with fillers composed of zeolites (ETS-10 and MCM-41) and
metal-organic-frameworks (MOF) (ZIF-8 and NH2-MIL-53).143 The material
integration was done via spin coating to produce thin, homogeneous
substrates. The additional materials were intended to increase the gas
permeability of the polysulfone membrane. The MOFs displayed significant increases in permeability of H2/CH4 and O2/N2 gas mixtures, while the
zeolites proved to be less eective. However, the MCM-41 incorporated
membrane proved to be the most robust among all others.
A recent publication by Xu et al. reports synthesis of a layered Ti-MWW
catalyst incorporated with piperidine ligands, to form a new organic
inorganic hybrid, Re-Ti-MWW-PI.144 The new catalyst has been found
to have a 6-coordinated orientation around Ti, according to UV-Raman,
UV-Vis and XPS studies. The presence of the Piperidine ligand is said to
supress acidity of hydroxyl groups and subsequently enhances peroxide
selectivity. The catalyst has shown remarkably high epoxide yields and
recyclability for straight chain and cyclic alkenes. However, conversion of
cyclohexene, one of the most prominent epoxidation substrates, seemed
to be relatively low.
4.2 Enhanced adsorbents/ion exchange materials
The property of adsorbing and exchanging various ions is a major
application of Engelhard type titano-silicates, due to their defective
microporous structure. One of their best features is the tuneable pore
size and structure based on dehydration.23 Cheung et al. studied the effects of partially ion-exchanged Na-ETS-4 and Na-CTS-1 (CTS contracted
titanium silicate) on CO2 adsorption.145 The ions exchanged were group
cations Mg21, Ca21, Sr21 and Ba21. Out of these materials, NaMg-CTS-1
displayed superior CO2 adsorption properties with high selectivity of CO2
from a CO2/N2 mixture. This is thought to be due to the contractible
nature of CTS network. An interesting observation was the realisation
that the CTS-1 network was in fact a dehydrated version of ETS-4. Preparation temperature was found to be a crucial factor in determining
whether a CTS or ETS framework was formed, with higher temperatures
favouring formation of a CTS framework. Such knowledge is crucial for
future research in terms of pre-treatment for existing materials, as well as
development of new materials.
Liu et al. have utilised ETS-1 and ETS-2 as agents for water treatment,
for the first time.36 A comparative study with zeolites ETS-4, Na-ZSM-5
and Na-Y revealed that ETS-1 showed the best adsorption of Cs1 ions and
ETS-2 showed significant adsorption for Sr21 ions. However, the selectivity for Cs1 needed further improvement.
4.3 New trends in photo-activity
One of the most significant dierences and advantages of titanosilicates to traditional zeolites is that they are photo-active.146149
Catalysis, 2016, 28, 119143 | 135
17:23:54.
View Online
The photo-activity and associated photo-catalysis using these materials is

a vast and dynamic area of research, and does not fall within the scope of
this chapter. However, a discussion on the outlook of titano-silicates
would not be complete without mentioning the most recent advances in
the field.
Galioglu et al. have achieved partial photochromic behaviour of silver
(Ag0) nanoparticle-incorporated ETS-10 films, for the first time.150
Photochromism can be defined as the light induced reversible change
of colour.151 First, Ag1 has been introduced into ETS-10 by ion exchange
with Na1 and K1, by mixing ETS-10 crystals with an AgNO3 solution
(denoted Ag1-ETS-10). Next, the material was converted to Ag nanoparticle form, I-Ag0-ETS-10 via thermal reduction. The reduction was
associated with a colour change from white to dark brownish yellow,
corresponding to a colourless bleached state and a colourful activated
state, respectively. When the I-Ag0-ETS-10 system was exposed to visible
laser light, re-oxidation of the Ag nanoparticles, led to partial diminishing of colour, thus generating a photochromic system. After a second
reduction step, followed by laser treatment, the darker colour partially
reappeared, confirming that the process is reversible, to an extent. It was
also observed that the extent of bleaching the colour of the system increased with increased power of the laser. Such photochromic systems
have the potential to be applied in a variety of fields such as optical
information storage, variable transmission optical materials, change in
material refractive index etc. The most prominent and commercial
applications of this phenomenon are smart windows and sunglasses.
4.4 New theoretical studies targeted at understanding oxidative
catalysis
Computational studies are of paramount importance in understanding
complex structureactivity relationships of titano-silicate materials.
These need to go hand-in-hand with experimental results in order to
bring about significant advances in their applications. Ko et al. recently
conducted DFT studies within the gradient-corrected approximation
(GCA) of two ETS-10 models. The goal was to better understand the eect
of ion exchange on the material structure and electronic properties.152
The TiO bond angles and distances were considered to measure the
extent of distortion of TiOTi quantum wires. It was found that two
dierent TiO bonds exist in the Na1 form and that K1, Ag1, and Ca21
cause small distortions to the wire, when fully cation-exchanged. Heavy
metal ions Zn21, Au31, and Ru31 seemed to cause significant structural
changes compared to the Na-form. This has also been confirmed by
Raman spectroscopic investigations. XRD studies have shown a significant loss in crystallinity of samples that were ion exchanged with Ru31
and Au31. Such observations are critical in fine-tuning specific applications related to ion exchange in titano-silicates.
Another recent study conducted by Kwon et al. was aimed at investigating the kinetics of gas-phase epoxidation of cyclohexene via H2O2,
with mesoporous TS-1 as catalyst.153 It was found that the epoxidation
follows an EleyRideal type mechanism, similar to the liquid-phase
136 | Catalysis, 2016, 28, 119143
View Online
Si
Si
O
Ti
O
Si
+ H 2O 2
O
Si
k1
-k1
Si
O
17:23:54.
Si
H
O
O
Ti
O
O O H +C H
6 10
Si
O
Si
Si
k2
-k2
k3
-k3
Si
O
+ H2O
Ti
O
O
Si
Si
O
Si
O
Ti
O
O
Si
Si
O
Si
+ H 2O
Si
Si
O O
Ti
O
O
Si
Si
O
Si
Si
H
O
O
Ti
O
O O H
Si
O
Si
Si
k4
-k4
O
+
Ti
O
O
Si
Si
O
Si
Si
H
O
O
H
Ti
O
O
Si
Si
O
Si
Scheme 6 Proposed EleyRideal type reaction mechanism for alkene epoxidation

with titano-silicate catalysts. Adapted from S. Kwon et al., A kinetic study of vapourphase cyclohexene epoxidation by H2O2 over mesporous TS-1, J. Catal., 326, 107115.
Copyright (2015) with permission from Elsevier.153
analogue (Scheme 6). The reaction between the TiOOH and adsorbed
alkene was determined to be the rate-determining step. In varying the
partial pressure of H2O2, a compensation eect was observed, due to the
surface coverage of intermediate species. The competition between water
molecules and H2O2 to be adsorbed onto Ti active sites was found to be
crucial for the above eect. Kinetics of the gas-phase epoxidation of alkenes had not been reported previously. Such studies play a key role in
the understanding of reaction mechanisms and deriving kinetics for
these industrially relevant processes.
These recent studies indicate that titano-silicate materials are actively
being investigated in multi-disciplinary research areas, which could lead
to their utilisation beyond the traditional scope. Although clearly not yet
optimised, such advances are significant to drive the evolution of synthesis methods and application of titano-silicates.
Summary
The development, evolution and applications of titano-silicate materials

have been discussed. Progress in synthesis methods and characterisation
of dierent titano-silicates with varying pore structure and morphology
has been reviewed. Variations in synthesis target specific applications
that expand their catalytic ability to suit dierent substrates and
Catalysis, 2016, 28, 119143 | 137
17:23:54.
View Online
oxidants. Selective catalytic epoxidation of alkenes, in the presence of

various oxidants has been discussed in detail, as it is a major application
with industrial relevance and academic interest. The use of two major
oxidants, hydrogen peroxide and tert-butylhydroperoxide, was compared.
The validity of some of the so-called green processes in the field was
critiqued. A brief overview of reaction mechanisms and kinetics has been
given; this is critical in understanding catalysis and further improving
these materials. However, much more work needs to be done in this area
to provide more insightful understanding of the complex behaviour of
titano-silicate materials. Finally, recent research and future prospects of
the field were discussed. It was made evident that the synthesis and
application of titano-silicates is rapidly evolving and has enormous potential to expand well beyond their current scope.
Acknowledgements
The EPSRC is gratefully acknowledged for support via a Frontier
Engineering award, EP/K038656/1.
References
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
M. Taramasso, G. Perego and B. Notari, U.S. Pat. 4410501, 1983.

R. Millini, E. P. Massara, G. Perego and G. Bellussi, J. Catal., 1992, 137, 497.
C. Neri, B. Anfossi, A. Esposito and F. Buonomo, Eur. Pat. 0 100 119, 1984.
P. Raa, G. Leofanti, A. Cesana, M. Mantegazza, M. Padovan, G. Petrini,
S. Tonti, V. Gervasutti and R. Varagnolo, Chim. Ind., 1990, 72, 598.
A. Esposito, C. Neri, F. Buonomo and M. Taramasso, U.K. Pat. 2 116 97,
1983.
B. Notari, Catal. Today, 1993, 18, 163.
T. Tatsumi, M. Nakamura, S. Negishi and H. Tominaga, J. Chem. Soc., Chem.
Commun., 1990, 476.
D. R. C. Huybrechts, L. De Bruycker and P. A. Jacobs, Nature, 1990, 345, 240.
G. Bellussi, A. Carati, M. G. Clerici, A. Esposito, R. Milliniand, and
F. Buonomo, Bel. Pat. No. 1001038, 1989.
G. Perego, G. Bellussi, C. Corno, M. Taramasso, F. Buonomo and A.
Esposito, New Developments in Zeolite Science Technology, ed. Y. Murakami,
A. Iijima and J. W. Ward, Kodansha, Tokyo, 1986, p. 129.
Y. S. S. Wan, J. L. H. Chau, A. Gavriilidis and K. L. Yeung, Microporous
Mesoporous Mater., 2001, 42, 157.
M. G. Clerici, Titanium Silicalite 1, Metal Oxide Catalysis. ed. S. David
Jackson and J. S. J. Hargreaves, WILEY-VCH Verlag GmbH & Co. KGaA,
Weinheim, Copyright r 2009, ISBN: 978-3-527-31815-5.
T. Sato, J. Dakka and R. A. Sheldon, J. Chem. Soc., Chem. Commun., 1994,
1887.
rez-Pariente, J. Chem. Soc., Chem. ComA. Corma, M. T. Navarro and J. Pe
mun., 1994, 147.
K. M. Reddy, I. L. Moudrakovski and A. Sayari, Stud. Surf. Sci. Catal., 1995,
98, 19.
M. S. Morey, S. OBrien, S. Schwarz and G. D. Stucky, Chem. Mater., 2000,
12, 898.
P. T. Tanev, M. Chibwe and T. J. Pinnavaia, Nature, 1994, 368, 321.
S. A. Bagshaw, E. Pouzet and T. J. Pinnavaia, Science, 1995, 269, 1242.
138 | Catalysis, 2016, 28, 119143
View Online
19
20
17:23:54.
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
C. R. A. Catlow, S. A. French, A. A. Sokol and J. M. Thomas, Philos. Trans. R.

Soc., A, 2005, 363, 913.
J. To, P. Sherwood, A. A. Sokol, I. J. Bush, C. R. A. Catlow, H. J. J. van Dam,
S. A. French and M. F. Guest, J. Mater. Chem., 2006, 16, 1919.
S. M. Kuznicki, US Pat. 4853202, 1989.
S. M. Kuznicki, US Pat. 4938939, 1990.
T. M. Neno, J. B. Parise, G. A. Jones, L. G. Galya, D. R. Corbin and
G. D. Stucky, J. Phys. Chem., 1996, 100, 14256.
G. M. Johnson, B. A. Reisner, A. Tripathi, D. R. Corbin, B. H. Toby and
J. B. Parise, Chem. Mater., 1999, 11, 2780.
B. F. Mentzen and P. Gelin, Mater. Res. Bull., 1995, 30, 373.
G. Guan, K. Kusakabe and S. Morooka, Microporous Mesoporous Mater.,
2001, 50, 109.
H.-K. Jeong, J. Krohn, K. Sujaoti and M. Tsapatsis, J. Am. Chem. Soc., 2002,
124, 12966.
llez, J. Coronas and J. Santamara,
Z. Lin, J. Rocha, A. Navajas, C. Te
Microporous Mesoporous Mater., 2004, 67, 79.
n, C. Te
llez, J. Coronas and J. Santamara, Eur. J. Inorg. Chem.,
V. Sebastia
2008, 2448.
X. Li, C. Zhou, Z. Lin, J. Rocha, P. F. Lito, A. S. Santiago and C. M. Silva,
Microporous Mesoporous Mater., 2011, 137, 43.
S. M. Kuznicki, V. A. Bell, S. Nair, H. W. Hillhouse, R. M. Jacubinas,
C. M. Braunbarth, B. H. Toby and M. Tsapatsis, Nature, 2001, 412, 720.
llez, J. Coronas and J. Santamara, J. Membr. Sci.,
I. Tiscornia, S. Irusta, C. Te
2008, 311, 326.
n, G. T. Palomino, F. Geobaldo, C. Lamberti,
A. Zecchina, C. O. Area
G. Spotoa and S. Bordiga, Phys. Chem. Chem. Phys., 1999, 1, 1649.
A. Anson, C. C. H. Lin, S. M. Kuznicki and J. A. Sawada, Chem. Eng. Sci.,
2009, 64, 3683.
S. M. Kuznicki, US Pat., 5011591, 1991.
H. Liu, A. Yonezawa, K. Kumagai, M. Sano and T. Miyake, J. Mater. Chem. A,
2015, 3, 1562.
G. X. S. Zhao, J. L. Lee and P. A. Chia, Langmuir, 2003, 19, 1977.
Y. D. Noh and S. Komarneni, Environ. Sci. Technol., 2011, 45, 6954.
L. D. Barreira, P. F. Lito, B. M. Antunes, M. Otero, Z. Lin, J. Rocha, E. Pereira,
A. C. Duarte and C. M. Silva, Chem. Eng. J., 2009, 155, 728735.
C. B. Lopes, J. Coimbra, M. Otero, E. Pereira, A. C. Duarte, Z. Lin and
J. Rocha, Quim. Nova, 2008, 31, 321.
M. A. Roberts, G. Sankar, J. M. Thomas, R. H. Jones, H. Du, J. Chen, W. Pang
and R. Xu, Nature, 1996, 381, 401.
H. Du, M. Fang, J. Chen and W. Pang, J. Mater. Chem., 1996, 6, 1827.
lu, ISBN 978-953-51K.-W. Park, Ion Exchange Technologies, ed. A. Kilisliog
0836-8, Published: November 7, 2012 under CC BY 3.0 license. r The
Author(s).
S. Ferdov, CrystEngComm, 2014, 16, 4467.
S. Ferdov, V. Kostov-Kytin and O. Petrov, Chem. Commun., 2002, 1786.
J. Rocha and M. W. Anderson, Eur. J. Inorg. Chem., 2000, 801.
M. Veltri, D. Vuono, P. De Luca, J. B. Nagy and A. Nastro, J. Therm. Anal.
Calorim., 2006, 84, 247.
o, A. Ferreira, A. P. Esculcas and J. D. Pedrosa de
Z. Lin, J. Rocha, P. Branda
Jesus, J. Phys. Chem. B, 1997, 101, 7114.
M. S. Dadachov, J. Rocha, A. Ferreira, Z. Lin and M. W. Anderson, Chem.
Commun., 1997, 2371.
Catalysis, 2016, 28, 119143 | 139
View Online
50
51
52
17:23:54.
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
G. Ferraris, Solid State Phenom., 2006, 111, 47.

rez-Carvajal, P. Aranda, A. Berenguer-Murcia, D. Cazorla-Amoro
s,
J. Pe
J. Coronas and E. Ruiz-Hitzky, Langmuir, 2013, 29, 7449.
rez-Carvajal, P. Lalueza, C. Casado, C. Te
llez and J. Coronas, Appl. Clay
J. Pe
Sci., 2012, 56, 30.
`meth, G. Ferraris, G. Radno
`czi and O. A. Ageeva, Can. Mineral., 2005,
P. Ne
45, 973.
G. Ferraris, A. Bloise and M. Cadoni, Microporous Mesoporous Mater., 2008,
107, 108.
R. Millini, G. Perego, W. O. Parker Jr., G. Bellussi and L. Carluccio, Microporous Mater., 1995, 4, 221.
s, J. L. Jorda
, M. Domine and F. Rey, Chem.
A. Corma, U. Daz, V. Forne
Commun., 1999, 9, 779.
P. Wu and T. Tatsumi, Chem. Commun., 2002, 1026.
P. Wu, D. Nuntasri, J. Ruan, Y. Liu, M. He, W. Fang, O. Terasaki and
T. Tatsumi, J. Phys. Chem. B, 2004, 108, 19126.
F. Song, Y. Liu, L. Wang, H. Zhang, M. He and P. Wu, Appl. Catal., A, 2007,
327, 22.
G. Peregot, G. Bellussi, C. Corno, M. Taramasso, F. Buonomot and
A. Esposito, Stud. Surf. Sci. Catal., 1986, 28, 129.
J. S. Reddy, R. Kumar and P. Ratnasamy, Appl. Catal., 1990, 58, L1L4.
W. T. A. Harrison, T. E. Gier and G. D. Stucky, Zeolites, 1995, 15, 408.
T. Tatsumi and N. Jappar, J. Phys. Chem. B, 1998, 102, 7126.
Y. Kubota, Y. Koyama, T. Yamada, S. Inagaki and T. Tatsumi, Chem. Commun., 2008, 46, 6224.
N. Jappar, Q. Xia and T. Tatsumi, J. Catal., 1998, 180, 132.
Z. Liu and R. J. Davis, J. Phys. Chem., 1994, 98, 1253.
M. D. C. Dutoit, M. Schneider and A. Baiker, J. Catal., 1995, 153, 165.
S. Klein, S. Thorimbert and W. F. Maier, J. Catal., 1996, 163, 476.
K. Kosuge and P. S. Singh, J. Phys. Chem. B, 1999, 103, 3563.
W. Li and M.-O. Coppens, Chem. Mater., 2005, 17, 2241.
A. J. H. P. Van der Pol, A. J. Verduyn and J. H. C. van Hoo, Appl. Catal., A,
1992, 92, 113.
Y. Wang, M. Lin and A. Tuel, Microporous Mesoporous Mater., 2007, 102, 80.
S.-T. Tsai, P.-Y. Chao, T.-C. Tsai, I. Wang, X. Liu and X.-W. Guo, Catal.
Today, 2009, 148, 174.
Y. Xue, Y. Wen, H. Wei, M. Liu, X. Huang, X. Ye, X. Wang and B. Li, RSC
Adv., 2015, 5, 51563.
Y. Yang, J. Ding, B. Wang, J. Wu, C. Zhao, G. Gao and P. Wu, J. Catal., 2014,
320, 160.
R. Millini and G. Perego, Gazz. Chim. Ital., 1996, 126, 133.
D. R. C. Huyberchts, P. L. Buskens and P. A. Jacobs, J. Mol. Catal., 1992,
71, 129.
G. Deo, A. M. Turek, I. E. Wachs, D. R. C. Huybrechts and P. A. Jacobs,
Zeolites, 1993, 13, 365.
M. W. Anderson, O. Terasaki, T. Ohsuna, A. Philippou, S. P. MacKay,
A. Ferreira, J. Rocha and S. Lidin, Nature, 1994, 367, 347.
M. W. Anderson, O. Terasaki, T. Oshuna, P. J. OMalley, A. Philippou,
S. P. Mackay, A. Ferreira, J. Rocha and S. Lidin, Philos. Mag. B, 1995, 71, 813.
L. Marchese, E. Gianotti, V. Dellarocca, T. Maschmeyer, F. Rey, S. Coluccia
and J. M. Thomas, Phys. Chem. Chem. Phys., 1999, 1, 585.
J. Rocha and M. W. Anderson, Eur. J. Inorg. Chem., 2000, 801.
M. Dusi, T. Mallat and A. Baiker, Catal. Rev.: Sci. Eng., 2000, 42, 213.
140 | Catalysis, 2016, 28, 119143
View Online
84
85
17:23:54.
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
J. Gao, Y. Chen, B. Han, Z. Feng, C. Li, N. Zhou, S. Gao and Z. Xi, J. Mol.
Catal. A: Chem., 2004, 210, 197.
I. Nowak, B. Kilos, M. Ziolek and A. Lewandowska, Catal. Today, 2003,
78, 487.
M. G. Clerici, G. Bellussi and U. Romano, J. Catal., 1991, 129, 159.
M. G. Clerici and P. Ingallina, J. Catal., 1993, 140, 71.
J. C. van der Waal, M. S. Rigutto and H. van Bekkum, Appl. Catal., A, 1998,
167, 331.
M. A. Camblor, A. Corma, P. Esteve, A. Martnez and S. Valencia, Chem.
Commun., 1997, 795.
A. Corma, P. Esteve, A. Martnez and S. Valencia, J. Catal., 1995, 152, 18.
P. Wu and T. Tatsumi, J. Phys. Chem. B, 2002, 106, 748.
P. Wu and T. Tatsumi, J. Catal., 2003, 214, 317.
P. Wu, Y. Liu, M. He and T. Tatsumi, J. Catal., 2004, 228, 183.
W. Fan, P. Wu and T. Tatsumi, J. Catal., 2008, 256, 62.
E. G. Derouane, G. J. Hutchings, W. F. Mbafor and S. M. Roberts,
New J. Chem., 1998, 797.
L. J. Schofield, O. J. Kerton, P. McMora, D. Bethell, S. Ellwood and
G. J. Hutchings, J. Chem. Soc., Perkin Trans. 2, 2002, 1475.
A. Bhaumik and T. Tatsumi, J. Catal., 1999, 182, 349.
X. Liu, X. Wang, X. Guo and G. Li, Catal. Today, 2004, 9395, 505.
T. Hayashi, K. Tanaka and M. Haruta, J. Catal., 1998, 178, 566.
T. A. Nijhuis, B. J. Huizinga, M. Makkee and J. A. Moulijn, Ind. Eng. Chem.
Res., 1999, 38, 884.
T. A. Nihjuis, E. Sacaliuc and B. M. Weckhuysen, The epoxidation of propene over gold nanoparticle catalysts, in Mechanisms in Homogeneous and
Heterogeneous Epoxidation Catalysis, ed. S. T. Oyama, Elsevier, Amsterdam,
2008, pp. 339386.
J. J. Bravo-Suarez, K. K. Bando, T. Fujitani and S. T. Oyama, J. Catal., 2008,
257, 32.
J. H. Huang, E. Lima, T. Akita, A. Guzman, C. X. Qi, T. Takei and M. Haruta,
J. Catal., 2011, 278, 8.
T. A. Nijhuis, T. Q. Gardner and B. M. Weckhuysen, J. Catal., 2005, 236, 153.
A. H. Tullo and P. L. Short, Chem. Eng. News, 2006, 84, 22.
P. Bassler, M. Weidenbach and H. Goebbel, The new HPPO Process for
Propylene Oxide: From Joint Development to Worldscale Production, Chem.
Eng. Trans., 2010, 21, 571576, DOI: 10.3303/CET1021096.
A. Sayari, Chem. Mater., 1996, 8, 1840.
R. Murugavel and H. W. Roesky, Angew. Chem., Int. Ed. Engl., 1997, 36, 477.
A. Corma, Chem. Rev., 1997, 97, 2373.
A. Bhaumik and T. Tatsumi, J. Catal., 2000, 189, 31.
T. Tatsumi, K. A. Koyano and N. Igarashi, Chem. Commun., 1998, 325.
P. Wu, T. Tatsumi, T. Komatsu and T. Yashima, J. Catal., 2001, 202, 245.
M. Baca, W. J. Li, P. Du, G. Mul, J. A. Moulijn and M.-O. Coppens, Catal.
Lett., 2006, 109, 207.
A. Modak, M. Nandi and A. Bhaumik, Catal. Today, 2012, 198, 45.
G. N. Vayssilov and R. A. van Santen, J. Catal., 1998, 175, 170.
G. Sankar, J. M. Thomas, C. R. A. Catlow, C. M. Barker, D. Gleeson and
N. Kaltsoyannis, J. Phys. Chem. B, 2001, 105, 9028.
A. Zecchina, M. Rivallan, G. Berlier, C. Lamberti and G. Ricchiardi, Phys.
Chem. Chem. Phys., 2007, 9, 3483.
Y. Iwasawa, R. Mason, M. Textor and G. A. Somorjai, Chem. Phys. Lett., 1976,
44, 468.
Catalysis, 2016, 28, 119143 | 141
View Online
119
120
121
17:23:54.
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
K. M. Bratlie, H. Lee, K. Komvopoulos, P. Yang and G. A. Somorjai, Nano

Lett., 2007, 7, 3097.
J. Lu et al., Science, 2012, 335, 1205.
M. Shekhar, J. Wang, W.-S. Lee, W. D. Williams, S. M. Kim, E. A. Stach,
J. T. Miller, W. N. Delgass and F. H. Ribeiro, J. Am. Chem. Soc., 2012,
134, 4700.
X. Deng, B. K. Min, A. Guloy and C. M. Friend, J. Am. Chem. Soc., 2005,
127, 9267.
rnqvist, L. Jensen, H. Topse,
C. J. H. Jacobsen, S. Dahl, P. L. Hansen, E. To
D. V. Prip, P. B. Menshaug and I. Chorkendor, J. Mol. Catal. A: Chem.,
2000, 163, 19.
P. Nandi, W. Tang, A. Okrut, X. Kong, S.-J. Hwang, M. Neurock and A. Katz,
Proc. Natl. Acad. Sci. U. S. A., 2013, 110, 2484.
J. M. Notestein, A. Solovyov, L. R. Andrini, F. G. Requejo, A. Katz and
E. Iglesia, J. Am. Chem. Soc., 2007, 129, 15585.
S. B. Shin and D. Chadwick, Ind. Eng. Chem. Res., 2010, 49, 8125.
V. Russo, R. Tesser, E. Santacesaria and M. Di Serio, Ind. Eng. Chem. Res.,
2014, 53, 6274.
X. H. Liang, Z. T. Mi, Y. L. Wu, L. Wang and E. H. Xing, React. Kinet. Catal.
Lett., 2003, 80, 207.
G. Wu, Y. Wang, L. Wang, W. Feng, H. Shi, Y. Lin, T. Zhang, X. Jin, S. Wang,
X. Wu and P. Yao, Chem. Eng. J., 2013, 215216, 306.
V. Russo, R. Tesser, E. Santacesaria and M. Di Serio, Ind. Eng. Chem. Res.,
2013, 1168.
J. M. Campos-Martin, G. Blanco-Brieva and J. L. G. Fierro, Angew. Chem. Int.
Ed., 2006, 45, 6962.
C. Samanta, Appl. Catal., A, 2008, 350, 133.
http://www.icis.com/chemicals/channel-info-chemicals-a-z/, costs given for
2007 (accessed 24/05/2015).
http://www.solvay.com/en/media/press_releases/20111005hydrogenperoxideplant.html, (accessed 24/05/2015).
L. E. Sendelbach, Toxicology, 1989, 57, 227.
J. L. Zhao, J. C. Zhou, J. Su, H. C. Guo, X. S. Wang and W. M. Gong, AIChE J.,
2007, 53, 3204.
J. Su, J. Zhou, C. Liu, X. Wang and H. Guo, Chin. J. Catal., 2010, 31, 1195.
E. Klemm, E. Dietzsch, T. Schwarz, T. Kruppa, A. L. de Oliveira, F. Becker,
G. Markowz, S. Schirrmeister, R. Schutte, K. J. Caspary, F. Schuth and
D. Honicke, Ind. Eng. Chem. Res., 2008, 47, 2086.
M. G. Clerici and P. Ingallina, Catal. Today, 1998, 41, 351.
G. Jenzer, T. Mallat, M. Maciejewski, F. Eigenmann and A. Baiker, Appl.
Catal., A, 2001, 208, 125.
T. Danciu, E. J. Beckman, D. Hancu, R. N. Cochran, R. Grey, D. M. Hajnik
and J. Jewson, Angew. Chem., Int. Ed., 2003, 42, 1140.
R. J. Saxton, Top. Catal., 1999, 9, 43.
llez and J. Coronas, Chem. Eng. Sci., 2014,
P. Burmann, B. Zornoza, C. Te
107, 66.
L. Xu, D.-D. Huang, C.-G. Li, X. Ji, S. Jin, Z. Feng, F. Xia, X. Li, F. Fan, C. Lib
and P. Wu, Chem. Commun., 2015, 51, 9010.
O. Cheung, J. Su, Z. Bacsik, J. Li, L. Samain, X. Zou and N. Hedin, Microporous Mesoporous Mater., 2014, 198, 63.
J. Zhang, D. S. Zhao, L. Y. Yang and Y. B. Li, Chem. Eng. J., 2010, 156, 528.
G. D. Lee, S. K. Jung, Y. J. Jeong, J. H. Park, K. T. Lim, B. H. Ahna and
S. S. Hong, Appl. Catal., A, 2003, 239, 197.
142 | Catalysis, 2016, 28, 119143
View Online
148
149
150
17:23:54.
151
152
153
K. Ikeue, H. Yamashita, M. Anpo and T. Takewaki, J. Phys. Chem., 2001,

105, 8350.
R. F. Howe and Y. K. Krisnandi, Chem. Commun., 2001, 1588.
S. Galioglu, M. Isler, Z. Demircioglu, M. Koc, F. Vocanson, N. Destouches,
Rasit Turan and B. Akata, Microporous Mesoporous Mater., 2014, 196, 136.
H. Durr and H. Bouas-Laurent, Photochromism: Molecules and Systems,
Elsevier BV Press, The Netherlands, 2003.
M. Ko, S. Galioglu, D. Tooli, H. Ustunel and B. Akata, J. Phys. Chem. C,
2014, 118, 27281.
S. Kwon, N. M. Schweitzer, S. Park, P. C. Stair and R. Q. Snurr, J. Catal.,
2015, 326, 107.
Catalysis, 2016, 28, 119143 | 143
Nanofiber-supported metal-based
catalysts
17:23:59.
Adeniyi S. Ogunlaja,a Phumelele E. Kleyi,a

Ryan S. Walmsleyb and Zenixole R. Tshentu*a
DOI: 10.1039/9781782626855-00144
Catalysis utilizing heterogeneous catalysts remains favoured in the chemical industry due
to their ease of separation and recyclability compared to homogeneous catalysts. Electrospun nanofibers as catalyst support materials can enhance catalyst performance due to
increased surface area-to-volume ratio. Recently, metal oxides and metallic nanoparticles
immobilized onto electrospun nanofibers have displayed enhanced catalytic activities
towards various reactions. Metal ion complexes supported on electrospun nanofibers, via
coordination to the desired functional groups of polymer chains, have also been applied as
heterogeneous catalysts in some organic syntheses. The nanofiber-based catalytic materials exhibited good catalytic activities for various reactions, as well as good recyclability
and reusability. Concerns over the mechanical and chemical stability of electrospun
nanofibers as well as the metal ion leaching sometimes occurring when employed under
extreme conditions are also emphasized. This review covers progress in the fabrication
and catalytic applications of various metal-based catalysts immobilized onto nanofibers.
It will also highlight the challenges associated with the use of electrospun nanofibers in
catalysis.
Introduction
Catalysis is arguably one of the areas of modern science that has driven
the size and shape of the chemical industry. It finds use in the
pharmaceutical, petrochemical and fine chemicals industries. There is
continuous research and development in this area of science.14 Catalysts
can be subdivided into two main types namely; homogeneous and heterogeneous catalysts. The value of homogeneous catalysts cannot be
understated, especially under laboratory scale conditions. However, the
industrial applicability of these catalysts is quite limited compared to the
heterogeneous counterparts. The main reason for this can be linked
primarily to the diculties associated with separation of homogeneous
catalysts from the reaction solution. Not only does this lead to increased
costs, but also environmental concerns. The very nature of these catalysts
also often limits their use to batch rather than continuous flow reactions,
another negative for industrial application. Therefore catalysis is now
increasingly being done using solid phase supports as the catalyst host.
Metals hosted on microspherical support systems show excellent
promise as heterogeneous catalysts.5,6 However, there is great eort
directed towards miniaturization of supports in order to achieve
a
Department of Chemistry, Nelson Mandela Metropolitan University,

P.O. Box 77000, Port Elizabeth 6031, South Africa.
E-mail: zenixole.tshentu@nmmu.ac.za
b
Department of Chemistry, Rhodes University, P.O. Box 94, Grahamstown 6139,
South Africa
144 | Catalysis, 2016, 28, 144174

c
17:23:59.
View Online
Fig. 1 Characteristics of nanofibers and nanofibrous membranes.
improved catalytic activities as a result of greater catalytic surface area.7,8

In addition to nanotubular supports and nanospheres,7,8 small diameter
fibrous materials in the form of electrospun nanofibers also present an
alternative (Fig. 1). The diameter of each fiber can be made very small in
order to achieve a high surface area, and the three-dimensional network
structure allows these materials to be separated very easily from the reaction solution.9 The fiber flexibility and porosity (inter-fiber) of the mat
makes these materials ideal for continuous flow applications. While the
microspherical supports are often highly porous and hence may have
comparatively larger surface areas than the fibers, it is questionable how
accessible these pores are to substrates. Thus, it is likely that a large
degree of catalysis takes place on the surface of the solid microspheres
and since the fiber diameters are much smaller, a greater degree of
catalytic surface area is directly exposed to the reaction solution.
Fibers may be produced by a variety of methods including drawing,10
template synthesis,11 deposition on a substrate, thermally induced phase
separation and spinning. Of the spinning methods, electrospinning
allows for fabrication of extremely fine (low nanometer diameter)
fibers.12 During the process of electrospinning, the viscous polymer solution or a melt forms a droplet at the spinneret (typically a blunt end
needle). The electric field draws this droplet into a structure known as the
Taylor cone, as a result of the surface tension on the droplet.13 The
surface tension is overcome when there is sucient electrostatic repulsion resulting in the formation of a polymer jet. As the jet accelerates
towards a counter electrode (typically a grounded collector) it undergoes
Catalysis, 2016, 28, 144174 | 145
17:23:59.
View Online
bending and elongation and at the same time the solvent evaporates until
a fine solid fiber forms.14 Although electrospinning seems simple from
an instrumentation point of view, there are several parameters that influence this process. These parameters can be roughly divided into three
groups: polymer solution (such as concentration, viscosity, surface tension, and conductivity), process (applied voltage, polymer solution feed
rate, and the tip-to-collector distance), and ambient conditions (temperature, humidity, and atmospheric conditions).15
Polymer nanofibers produced by electrospinning have been studied for
a wide variety of applications including filtration, wound dressings, tissue engineering scaolds, artificial blood vessels, energy devices, drug
delivery, colorimetric probes and solid phase extraction sorbents,1623 to
mention but a few. There has been, however, moderate interest in using
nanofibers as catalyst supports. The simplest examples of electrospun
fibers acting as catalyst supports are the non-metallated variety.2426 With
these, an organic molecule acts as the catalyst and may be introduced
either by copolymerization or by post-electrospinning functionalization.
Polymer nanofibers are also increasingly being used as supports for metal
nanoparticles. Polymer nanofibers are also increasingly being used as
supports for metal nanoparticles and to a lesser extent as metal oxide
fibers produced by calcination, and applied in catalysis. There is, however,
scant literature reporting on the utilization of the coordination chemistry
approach for catalyst immobilization on nanofibers. This review will,
therefore, attempt to cover advances in the area of catalytic nanofibers.
Nanofiber-supported metal/metal oxide catalysts
2.1 Metallic nanoparticles

2.1.1 Cross-coupling reactions. The discovery of CC cross-coupling
reactions, a few decades ago, such as Heck,27 Suzuki,28 Negishi29 and
Stille30 reaction revolutionized the field of organic synthesis since the
syntheses of compounds by the CC formation became possible. The
reactions usually occur in the presence of a metallic catalyst, most
commonly Pd(0). However, the major challenges associated with these
catalysts include, but are not limited to, designing suitable ligands for
the stabilization of the metal in the zero oxidation state, the catalyst efficiency, and its recyclability.31,32 These catalyst challenges have been
addressed by immobilization of metal nanoparticles on polymer-based
solid support materials.3336 Metal nanoparticles have high surface
area-to-volume ratios, good electrical conductivities and optical properties, and thus they possess higher catalytic activities towards various reactions compared to bulk catalysts.33
Electrospun nanofibers have possible application as membranes in
membrane catalysis, and more commonly as solid supports for metal
nanoparticles and metal oxides produced by calcination. Most research
in catalytic application of electrospun nanofibers involves the deposition
of metal nanoparticles on the surface of electrospun nanofibers,3740 or
preparation of inorganic/metal oxide based nanofibers.41,42 Polymer
nanofibers fabricated by the electrospinning process have been employed
146 | Catalysis, 2016, 28, 144174
17:23:59.
View Online
as solid support materials for metal nanoparticle-based catalysts due to

their complementary properties (high surface area-to-volume ratio and
porous nature). The use of carbon nanofibers as a favoured metal catalyst
support material is well documented.36 This is because carbonized materials are swelling resistant and insoluble in most solvents. Moreover,
the catalytic sites are located at a large conjugated electron structure of
graphite which may enhance the catalyst activity/selectivity by the formation of charge-transfer complex.40 Carbon nanofibers containing
palladium nanoparticles have been used as a catalyst for the Sonogashira
coupling reaction.40 The catalyst was prepared by electrospinning a solution of polyacrylonitrile (PAN, 10 wt%) and Pd(OAc)2 (3.3 wt%) in DMF.
Calcination of the PAN/Pd(OAc)2 composite nanofibers resulted to the
carbonation of PAN as well as the reduction of Pd21 to Pd0 and thus the
formation of Pd nanoparticles (PdNPs). Figure 2 illustrates the TEM
micrographs of the electrospun carbon nanofiber composites (PdNP/
CNFs).
The PdNP/CENFs composite was evaluated for catalytic activity towards
the reaction of terminal acetylenes and aryl halides the Sonogashira
reaction (Scheme 1).
The catalyst showed excellent product yields (8488%) and was
recovered without loss of material. It was also reported that the catalyst
was recovered without leaching of the PdNPs. The reason was that the
nanoparticles were formed by surface diusion and aggregation of
smaller nanoparticles within the carbon nanofiber matrix. The catalyst
was also reusable for up to the 10 cycles without any reduction of
Fig. 2 TEM micrographs of PdNP/CENFs. Reproduced from L. Chen et al., Novel

Pd-carrying composite carbon nanofibers based on polyacrylonitrile as a catalyst for
Sonogashira coupling reaction, Catal. Commun., 9, 22212225. Copyright (2008), with
permission from Elsevier.40
I
+
PdNP/CENFs
Na3PO4.12H2O, reflux
Scheme 1 The reaction phenylacetylene and iodobenzene catalyzed by PdNP/CENFs.

Catalysis, 2016, 28, 144174 | 147
View Online
17:23:59.
43
catalytic activity. In a similar study, Guo et al. reported on Pd nanoparticles supported onto electrospun carbon nanofibers as a catalyst
for the Heck reaction. The catalyst was fabricated by electrospinning a
solution of PAN (8 wt%) and PdCl2 in DMF such that the ratio PAN
and PdCl2 was 50. The formed PAN-PdCl2 nanofibers were then subjected
to a gas phase hydrogen reduction reaction using a hydrogen-filled kettle
at 100 1C. The nanofibers were subsequently calcined, at dierent
temperatures, to induce carbonization of PAN nanofibers. Figure 3
illustrates the process for the fabrication of the carbon nanofibersupported Pd nanoparticles catalyst (PdNPs/CNFs).
The catalytic activity of the catalyst was evaluated in the reaction of
alkyl acrylates and iodobenzene (Scheme 2).
It was observed that the catalyst exhibited excellent conversion eciencies (100%) and product selectivities (9899%). The catalyst was recovered with a slight loss of material and was reusable up to 5 cycles with
product yields of over 95%. Unfortunately, no information was furnished
by the authors regarding the amount of leaching of the PdNPs. Obuya
et al.39 also evaluated the catalytic activity of the electrospun TiO2
nanofiber-Pd nanoparticles composite on the Heck reaction (reaction
of styrene and iodobenzene). The nanofiber catalyst was fabricated by
electrospinning a solution of polymethylmethacrylate (PMMA) and titanium isopropoxide [Ti(OPri)4] in CHCl3 and DMF. These nanofiber
composites were calcined in the presence of air at 500 1C. The Pd nanoparticles were subsequently deposited on the surface of the nanofibers
under UV light with ethyleneglycol (EG) used as a reducing agent.
Reduction
Electrospinning
Carb
PdCl2/PAN/DMF
solution
oniza
tion
Fig. 3 Process for fabrication of PdNPs/CNFs catalyst.43
I +
catalyst, Et3N
125 C
R = C4H6O2, C5H8O2, C6H12O2

Scheme 2 The reaction of alkyl acrylates and iodobenzene.
148 | Catalysis, 2016, 28, 144174
17:23:59.
View Online
An interesting observation was reported in that the catalyst was quasihomogeneous, that is, the Pd nanoparticles on the surface of nanofibers
would be oxidized to Pd21 in the reaction solution and later re-deposit on
the surface towards the end of the reaction.
Fabrication of electrospun ceramic nanofibers containing noble metal
nanoparticles have also been reported by Formo et al.44 Specifically, Pt,
Pd and Rh nanoparticles were deposited on the surface of TiO2 (anatase
and rutile) and ZrO2 nanofibers. The nanofibers were fabricated by
electrospinning a solution of polyvinylpyrrolidone (PVP) and TiO2 or ZrO2
in ethanol. The ZrO2 nanofibers were obtained by calcination at 550 1C
while calcination was carried out at 510 and 800 1C for anatase and rutile
nanofibers, respectively. It was observed that the rough surface texture of
the ceramic nanofibers presented ideal sites for the nucleation of metal
atoms, hence allowed for direct deposition of noble metals. The catalytic
activity of the nanofibers was evaluated for the Suzuki reaction
(Scheme 3) in a continuous flow reactor.
It was observed that the conversion eciency decreased as reaction
proceeded which was attributed to poisoning of the catalyst. The authors
suggested that the solution to the challenge was to reactivate the catalyst
by intermittently passing 1 M HNO3 through the reactor. This poses
another challenge since it requires that the reaction be stopped and
started again, which could lead to erroneous results. It was also observed
that the concentration of HNO3 resulted to leaching of the nanoparticles,
which was solved by using a diluted HNO3.
2.1.2 Methanol oxidation reactions: direct methanol fuel cells. Direct methanol fuel cells (DMFCs) have received much research interest
because methanol is an inexpensive, readily available and easy to store
and to transport liquid fuel.45 The important process in a DMFC is the
catalytic oxidation of methanol by a metal catalyst (e.g., Pt, Ru), which
produces energy (electricity) with CO2 and H2O as by-products. The fuel
cell technology is being applied in various technical areas including
stationary power sources, electro-tractions, computers and mobile devices.46,47 Methanol cross-over and poor oxidation at the anode remain
the major challenges for applications of DMFCs.46,48 Moreover, the Pt
catalyst, which is used as the anode, is highly susceptible to poisoning
by CO adsorbed on the surface and this drawback can be averted by
using solid supports such as TiO2 and SnO2.47 Alternatively, Pt-based
binary, tertiary or quaternary compounds can be used.49 Formo et al.44
fabricated electrospun anatase nanofibers as supports for Pt nanostructures, the combination being used as a catalyst for DMFCs. Electrospinning a solution of PVP and Ti(OPri)4, followed by calcination
at 510 1C, gave anatase nanofibers. Three types of anatase nanofibers
were then prepared by surface-coating (at dierent times: 3, 7 and 19 h)
HO
O 2N
Br +
B
HO
Pd/TiO 2
K2CO 3
O2N
Scheme 3 The reaction of 4-bromo-nitrobenzene and phenylboronic acid.

Catalysis, 2016, 28, 144174 | 149
17:23:59.
View Online
with Pt nanoparticles using EG as a reducing agent. A second type of

catalyst was prepared by growing Pt nanowires on the surface of the
nanoparticle-coated anatase nanofibers. The methanol oxidation reaction
carried out using cyclic voltammetry showed that the Pt/TiO2 nanofibers
(coated for 3 h) were the best catalyst due to exposure of TiO2 surfaces.
The reason was that TiO2 reduces catalyst poisoning by CO since the
hydroxides adsorbed on the surface of anatase react with the CO intermediates. Another observation was that the Pt nanowire-coated nanofibers exhibited better electrocatalytic activity than both the commercial
Pt/C and Pt nanoparticle-coated nanofibers. A similar investigation
was carried out where Pt nanoparticles were supported on the surface
of electrospun CNFs.49 It was observed that these catalysts displayed
excellent electrocatalytic activity and stability.
In an attempt to investigate the influence of the nanofibrous morphology on the catalytic activity of NiO nanostructures towards methanol
electrooxidation, Barakat et al.50 fabricated NiO nanoparticles and electrospun polyvinylpyrrolidine nanofibers. It was observed that the nanofibers exhibited superior electrocatalytic activity compared to just the
nanoparticles. This eect was attributed to the 1-D structure of nanofibers
which facilitates electron movement. It was also observed that, for
nanofibers, the required optimum methanol concentration was 10-fold
that required for nanoparticles. A year later Barakat et al.51 also reported
a novel CoNPs/CuNPs-incorporated carbon nanofiber catalyst, fabricated
by electrospinning of a Co (25 wt%)/Cu (25 wt%)/PVA (80 wt%) solgel in
water, followed by calcination at 750 1C. The electrocatalytic and stability
properties of the electrospun nanofibers were very promising for DMFCs.
2.1.3 Hydrogenation/reduction reactions. The hydrogenation reaction is very important in various sectors of industry, inter alia, food,
pharmaceutical, petrochemical and fine chemicals. In the food industry, for example, solid spread margarine is produced by hydrogenation
of a liquid vegetable oil.52 Most hydrogenation reactions are performed
at elevated temperature or pressure, in the presence of homo- or heterogeneous catalysts. Immobilizing these catalysts on solid support materials renders their separation and recyclability simple. Demir et al.53
evaluated the catalytic hydrogenation activity of electrospun poly(acrylonitrile-co-acrylic acid)-PdNPs (PAN-AA/PdNPs) nanofibers in the hydrogenation of unsaturated alcohols. Four nanofiber samples were
fabricated by electrospinning solutions of PAN-AA (8 wt%) and PdCl2
with varying concentrations of AA and PdCl2 in DMF. Subsequently,
placing the nanofibers in a solution of aqueous hydrazine as reducing
agent resulted in the formation of Pd0 nanoparticles. It was observed
that increasing the AA (5.48.1%) resulted in a corresponding decrease
in the reaction rate. This eect was attributed to the increased
size of the Pd nanoparticles in the copolymer containing 8.1% AA.
Increasing the concentration of Pd nanoparticles was observed to
decrease the catalytic activity as well. However, compared to Pd/Al2O3,
the catalytic activity of the PAN-AA/PdNPs nanofibers was 4.5 times
better. Ebert et al.54 investigated the catalytic activity of electrospun
150 | Catalysis, 2016, 28, 144174
17:23:59.
View Online
poly(amideimide) (PAI) nanofibers coated with PdNPs towards the hydrogenation of methyl oleate to methyl stearate. The PdNP-coated
nanofibers displayed better catalytic activity towards hydrogenation of
methyl oleate, compared to the commercial Pd/Al2O3, although the
reaction pathways of the two catalysts were dierent. The observed
experimental rates of hydrogenation for Pd/Al2O3 and PAI/PdNPs were
0.3 and 2.0 mmol/(h mg Pd), respectively. The electrospun nanofiber
catalyst could be used up to four cycles while maintaining the same
rate of hydrogenation. This stability was attributed to lack of leaching
of PdNPs.
Porous polymer-silica electrospun nanofibers containing AgNPs were
fabricated using a solgel method and evaluated for the catalytic reduction of methylene blue dye in the presence of NaBH4.55 The progress
of the catalytic reduction of methylene blue was monitored using UV
spectroscopy which showed the gradual disappearance of the methylene
blue band (665 nm) over time. It was reported that no deactivation or
poisoning of the catalyst was observed but unfortunately no information
was provided on the reusability of the catalyst. In a similar study, Pt
nanoparticles incorporated into electrospun TiO2 nanofibers were shown
to have a potential for the catalytic reduction of methylene red in the
presence of NaBH4.56,57 Graeser et al.58 reported on electrospun core
shell polyethylene oxide (PEO) nanofibers incorporated with Pd, Rh and
Pt nanoparticles as catalysts for hydrogenation of alkenes and nitrobenzenes. The nanaofibers were composed of a hollow core incorporating nanoparticles or a polymer core in which nanoparticles were
dispersed. The shell acted as a support, regulating access to the nanoparticles while also facilitating their recovery. The core acted as a reaction
chamber allowing easy access to the catalyst sites (Fig. 4).
The coreshell PEO nanofibers with bimetallic nanoparticles (Pd/Rh)
were evaluated as catalysts for the hydrogenation of cyclopentadiene and
nitrobenzenes using an autoclave. Direct reduction of cyclopentadiene to
cyclopentane was complete although the reduction of the second double
Fig. 4 SEM micrographs of electrospun PEO nanofibers with metal nanoparticles:

(a) after hydrogen reduction, (b) after solvent extraction of PEO core.58 Reproduced
with permission from M. Graeser, E. Pippel, A. Greiner and J. H. Wendor, Macromolecules, 2007, 40, 60326039. Copyright (2008) American Chemical Society.
Catalysis, 2016, 28, 144174 | 151
17:23:59.
View Online
bond occurred slower. The nitrobenzene was also completely reduced to

the aminobenzene, with observed side products or starting material. For
both model reactions, the catalysts were readily recovered and could be
used several times without any loss of activity. In separate studies Fang
et al.59 and Lu et al.60 also investigated the catalytic activity of PdNPs
encapsulated into electrospun polyethyleneimine/polyvinyl alcohol and
polypyrrole nanofiber composites, respectively, towards hydrogenation of
p-nitrophenol, respectively.
2.1.4 Photocatalysis. Heterogeneous photocatalysis has, over the
past decades, seen an upsurge of research activity for energy and environmental applications. Photocatalysis is a process whereby a photochemical reaction is accelerated by the presence of a catalyst and
recent interest has been on semiconductor materials.61 The mechanism
of a photocatalytic reaction (Fig. 5) involves absorption of UV radiation
from a source, and the energy is used for the decomposition of the
substrate (organic and inorganic compounds, dyes, pesticides and
crude oils). Once irradiated, an electron is excited from the valence
band to the conduction band, creating a hole in the valence band.
Superoxides are generated by the interaction of oxygen with excited
electrons and holes, and they subsequently react with a substrate.
Some of the applications of photocatalyis include the removal of pollutants from air and water as well as solar water splitting.62,63
For easy handling and separation, the photocatalysts are usually immobilized on solid support materials (e.g., polymers). Electrospun polymer nanofibers have become the solid support materials of choice due
to their high surface-area-to-volume ratio, porous nature and ease of
tunability. Electrospun polystyrene (PS) nanofibers doped with Zn(II)
phthalocyanine-gold nanoparticle conjugates have been shown to be
potential photocatalysts for the degradation of 4-chlorophenol (4-CP) and
Orange G (OG).63 The photodegradation of both 4-CP and OG was
observed to obey first-order kinetics, as the computed rate constants
decreased with decreasing concentration. In a related study, electrospun
Fig. 5
Mechanism of a photocatalytic reaction of nanoparticles.61
152 | Catalysis, 2016, 28, 144174
17:23:59.
View Online
Fig. 6 Possible structure of Zn(II) phthalocyanine-magnetite nanoparticles on electrospun nylon 6 nanofibers.64
nylon 6 nanofibers functionalized with Zn(II) phthalocyanine-magnetite

nanoparticles displayed photocatalytic activity towards the degradation
of OG.64 Figure 6 illustrates the structure of electrospun nylon 6 nanofibers functionalized with Zn(II) phthalocyanine-magnetite nanoparticles
conjugates. It was mentioned that there was no leaching observed despite
the fact that the conjugates were embedded and not covalently linked to
the electrospun nanofibers.
2.2 Metal oxide catalysts
The last decade has seen an increased research activity focused towards
the use of metal oxide electrospun nanofibers and electrospun nanofibersupported metal oxide-based photocatalysts.65 Titanium oxide (TiO2)
nanofibers have attracted the most research interest since TiO2 is biologically and chemically inert, and it is relatively inexpensive.66 TiO2 is
understood to be photocatalytically unselective, and thus selectivity can
be induced by modifying surface functionality, or altering the surface
charge, or by incorporation of transition or noble metals.66 From the
incorporation of SiO2 into electrospun TiO2 nanofibers by combining a
solgel and two-capillary co-electrospinning method, a TiO2/SiO2 nanofiber composite selectively degraded methylene blue (MB) in water in the
presence of disperse red (DR).66 The disappearance of the absorption
band for MB (664 nm) and the persistence of the band for DR (228 nm)
indicated the selective photodegradation of MB. The catalytic selectivity
was attributed to the presence of internal channels of mesoporous SiO2
that allowed small MB molecules to enter these pores but prevented
entrance of larger DR molecules.
Several research groups have reported electrospun TiO2/CoFe2O4
nanofibers as photocatalysts for the degradation of MB.6769 The
electrospun TiO2/CoFe2O4 displayed excellent activity towards the
Catalysis, 2016, 28, 144174 | 153
17:23:59.
View Online
photodegradation of MB, however, the photodegradation reactions were

performed at dierent wavelengths. CoFe2O4 was used to impart magnetic properties to the nanofiber composite to enhance recovery and
reuse. In one study, the presence of CoFe2O4 resulted in enhanced photocatalytic activity of TiO2, since Co and Fe act as electron trapping sites
which reduce the rate of electron-hole recombination.68 In another study,
the improved photocatalytic activity was attributed to the possible substitution of Ti41 sites by Co21 and/or Fe31 in the TiO2 lattice, and the
presence of CoFe2O4 broaden the response region of UV light as well as
improve the absorbance of UV light.69 Vanadium-doped electrospun TiO2
nanofibers have also been reported to have enhanced photocatalytic activity towards photodegradation of MB.70 The enhanced photocatalytic
activity was also attributed to the ability of V41 to act as an electron and
hole trap, reducing the rate of electron-hole recombination, as shown in
the mechanism in Fig. 7. The photocatalytic activity of electrospun TiO2
nanofibers has also been improved by the incorporation of Ag nanoparticles into the nanofibers matrix.71,72 Barakat et al.71 reported that an
increase in silver-doping on the nanofibrous TiO2 material enhanced
photocatalytic degradation of methylene blue dye possibly due to the
reduction of the energy gap.71 Other research groups have reported the
photodegradation of MB using TiO2 nanofibers without the incorporation of another metal or metal oxide to enhance the photocatalytic activity of TiO2.7378 Finally, electrospun graphene-ZnO79 nanofibers have
also been investigated for their photocatalytic activity towards the degradation of MB. Graphene was added because it has the ability to improve
the electron transport and thus increase the photocatalytic activity.
A combined solgel process and electrospinning technique was used to
fabricate electrospun ZnO/SnO2 nanofiber for the photodegradation of
Rhodamine B (RhB).80 The ZnOSnO2 nanofibers formed a heterojunction which enhanced the separation of the generated electron-hole pairs,
Fig. 7 The energy diagram and the mechanism of V-doped TiO2 nanofibers. Reproduced
from Z. Zhang et al., Electrospun nanofibres of V-doped TiO2 with high photocatalytic
activity, J. Colloid Interface, 351, 5762. Copyright (2010), with permission form Elsevier.70
154 | Catalysis, 2016, 28, 144174
17:23:59.
View Online
resulting to improved photocatalytic activity. It was also observed that

upon UV excitation of the ZnO/SiO2 heterojunction with energy equal to
or higher than the band gaps of ZnO and SnO2 the electrons in the valence band could be excited to the conduction band with concomitant
generation of an equal number of holes in the valence band. It was further observed that the photogenerated electrons were transferred from
the conduction band of ZnO to that of SnO2, and conversely, the photogenerated holes were transferred from the valence band of SnO2 to that
of ZnO. In a similar study, Shi et al.81 reported the photocatalytic activity
of porous SnO2/TiO2 nanofibers fabricated by calcination of electrospun
polyvinylpyrrolidone nanofibers doped with SnO2 and TiO2. The SnO2/
TiO2 nanofibers displayed enhanced photocatalytic activity towards the
degradation of RhB, with the enhancement of photocatalytic activity attributed to the presence of SnO2, since it aected the separation and
transfer of photogenerated electron-hole pairs. Horzum also reported the
photocatalytic activity of electrospun CeO2/TiO2 nanofibers, fabricated by
colloidal electrospinning, towards the photodegradation of RhB.82 Du
et al.83 performed a comparative study of the photocatalytic activities of
electrospun ZnO and TiO2 nanofibers towards the degradation of RhB. It
was observed that the electrospun ZnO nanofibers displayed superior
initial photocatalytic activity compared to TiO2 nanofibers. However, the
electrospun TiO2 nanofibers exhibited superior overall photocatalytic
activity. Other electrospun nanofibers composited with TiO2,84 and
PtNPs/TiO2,85 BiFeO386 and ZnO/Bi2O387 have been fabricated and found
to exhibit enhanced photocatalytic activity for the degradation of
RhB. Various other metal oxide-doped electrospun nanofibers have also
exhibited enhanced photocatalytic activity towards the degradation of
other organic dyes such as methyl orange88 and Congo red.89
Gholamvand90 reported the fabrication and photocatalytic activity investigation of an electrospun carbon nanofiber/TiO2 composite toward
photodegradation of ethanol. He observed that the photocatalytic degradation decreased as the quantity of adsorbed ethanol (onto the nanofiber composite) increased at 1000 1C. The observation was attributed to
the partial transformation of anatase to rutile at 1000 1C.90 Basheer
employed electrospun TiO2/PVC nanofibers packed in a capillary microreactor as a photocatalyst for the oxidation of benzene to phenol using
hydrogen peroxide as an oxidant.91 He observed that performing the reaction for more than 10 min resulted in the formation of side products
(hydroquinone and quinine). Furthermore, application of a potential
dierence across the capillary increased the conversion of benzene due to
the increase of electrooxidation force (EOF) in the capillary channel.
Electrospun TiO2/SiO2 nanofiber composites have also been employed
as a photocatalyst for the reduction and oxidation of Ag1 ions and
acetaldehyde, respectively.92 Javid et al.93 reported that degradation of
tetracycline, using TiO2 nanofibers as a photocatalyst could be enhanced
by raising the pH of the solution from 410. An interesting study was
performed by Zhang et al.,94 where the photocatalytic activity of electrospun polyoxometalates (Ti0.75PW12O40)/TiO2 nanofiber composites were
evaluated for the reduction of Cr(VI) in aqueous solution. The observation
Catalysis, 2016, 28, 144174 | 155
View Online
17:23:59.
Table 1 Summary of some metal nanoparticles hosted onto polymer nanofibers for
catalysis.
Polymer nanofibers
Metal nanoparticles
Reactions catalysed
Ref.
Polyacrylonitrile
PdNPs
(1) Sonogashira
4043
coupling reaction
(2) Heck coupling
reaction
Polyvinyl pyrrolidine (PVP)
TiO2, ZrO2
Suzuki reaction
44
Ti(OPri)4, TiO2, PtO2,

SnO2
Oxidation of
methanol
4649
NiO
Methanol
electro-oxidation
50
(1) Polyacrylonitrileco-acrylic acid

(2) Polyamideimide (PAI)
(3) Poly ethylene oxide
(4) Polyethyleneimine/
poly vinyl alcohol (PVA)
PdNPs, RhNPs, PtNPs,

(Pd0), TiO2
Hydrogenation
reaction
5360
Polystyrene;
Poly(ethyleneterephthalate);
Nylon;
Polyvinyl alcohol (PVA).
AuNPs, AgNPs, MgNPs, Photocatalysis

TiO2/CoFe2O4, TiO2,
Graphene-ZnO,
ZnO-SnO2, SnO2/TiO2,
ZnO/TiO2, Pt TiO2,
ZnO/Bi2O3
63, 64, 66,

67, 6997
was that the reduction of Cr(VI) increased as the sintering temperature

was increased to 600 1C, at which point the reduction activity decreased
due to poor performance of the high temperature phase WO3. Other
metal oxide-doped electrospun nanofibers have been fabricated and
employed as catalysts for the photocatalytic evolution of hydrogen,95
oxidation of CO,96 and photocatalytic reduction of CO2.97 Table 1 below
lists the summary of some metal nanoparticles immobilized onto polymer nanofibers for catalysis.
Nanofiber-supported metal complexes
Polymers with nitrogen, oxygen and sulfur donor atoms are commonly
employed as ligands for coordination with metal ions, while simultaneously acting as support materials. Most polymers do not usually
come with the desired functional groups; hence they are functionalized
to introduce the desired chemistry. This can be done in two ways, namely
via a pre- or post-electrospinning modification strategy. The two approaches employed in preparing functional polymers are: (i) the polymerisation or copolymerisation of monomers containing the desired
functionality;98100 and (ii) the chemical modification of the available
polymer.100 The latter approach is potentially simpler due to the direct
functionalization of polymer, in which the incorporation of ligand onto
156 | Catalysis, 2016, 28, 144174
17:23:59.
View Online
polymer can be observed by monitoring the leaving group from the

polymer upon functionalization. Polymers are preferably electrospun to
nanofibers prior to the introduction of various chemical modifications
(post-electrospinning functionalization), as some polymers tend to
crosslink upon functionalization, making the polymer insoluble in the
most suitable electrospinning organic solvents.
Metal complex formation can be achieved by soaking functional
polymer nanofibers with a metal salt solution under suitable reaction
conditions. Coordination patterns (geometric structures) of metal ions
with polymers are somewhat dicult to characterize fully. Nonetheless, a
monodentate, bidentate or other complex of higher denticity can be
formed between functional polymer and metal ion (Fig. 8). The intermolecular bridged complexes can also be expected between polymer
strands where a metal ion acts as the bridging atom (Fig. 9). The above
mentioned phenomena have been reported as probable coordination
patterns within functional polymers.101110
Card and Neckers106 functionalised polystyrene by attaching 2,2 0 bipyridine to the benzene rings of polystyrene-2% divinylbenzene
copolymer through the reaction sequence shown in Scheme 4. Firstly,
the styrene ring bromination was carried out, followed by lithiation
and lastly functionalization with bipyridine ligand. Various metal ions
+
L
L
M
M
L
M
Polymer nanofiber
L = Ligand
M = Metal ions
Fig. 8 A schematic showing the pendant complex formation between the functional
polymer and metal ions.
L
L
M = Metal ions; L = Ligand;
M
L
= Polymer strands
Fig. 9 Intermolecular bridged complex formation between functional polymer strands

and metal ions.
Catalysis, 2016, 28, 144174 | 157
17:23:59.
158 | Catalysis, 2016, 28, 144174
nBuLi / THF
Br2
0 - RT
FeBr3
n
Br
Li
Air/
Reflux
N
m
MLx
THF
MLx
N
Scheme 4 Functionalisation of polystyrene with 2,2 0 -bipyridine.106
17:23:59.
View Online
were bound to the 2,2 0 -bipyridine-functionalized polymer using salts

such as Cr(NO3)3, MnI2, FeCl3, CoCl2, Ni(NO3)2, Pd(O2CCH3)2 and CuBr2.
Welleman et al.107 introduced imidazoles on chloromethylated styreneco-DVB in the presence of sodium or lithium salts using dimethyl formamide or tetrahydrofuran as solvent.
Metal complexes supported on polymers are widely used as heterogeneous catalysts in organic synthesis to overcome some of the serious
limitations of the reactions in the homogeneous phase.108134 Herein, we
review vanadium, palladium and zinc catalysts hosted on functional
nanofibers categorized by the dierent types of reactions they can
catalyse.
3.1 Vanadium catalysts
3.1.1 Oxidation reactions. Walmsley et al.108 reported the catalytic
oxidation of thioanisole using electrospun nanofibers loaded with
oxidovanadium(IV) complexes. The oxidation reaction was carried
out using hydrogen peroxide as the oxidant and oxidovanadium(IV)
complexes as pre-catalysts. The oxidovanadium(IV) complexes of the
series; 2-(4,5-diphenyl-1H-imidazol-2-yl)-4R-phenol ligands (R hydrogen,
bromo, methoxy and nitro) were hosted on polystyrene naofibers by the
incorporation method rather than as covalent linkages. In fact the
diphenyl rings were introduced to facilitate p-interactions with polystyrene (Scheme 5). The complex catalyst of the ligand with the nitro substituent gave the highest oxidation yield of methylsulfonylbenzene
(99.8%) within 10 min (Scheme 5). The order of substituted ligand
reactivity was 4-nitro-2-(4,5-diphenyl-1H-imidazol-2-yl)phenol44-bromo2-(4,5-diphenyl-1H-imidazol-2-yl)phenol42-(4,5-diphenyl-1H-imidazol2-yl)phenol44-methoxy-2-(4,5-diphenyl-1H-imidazol-2-yl)phenol, which
O
Catalyst + H2O2
Catalyst + H2O2
CH3CN, 25 C
CH3CN, 25 C
Thioanisole/
methylsulfinylbenzene
methylsulfonylbenzene
methyl(phenyl)sulfane
HN
Catalyst
O O
N
NH
Scheme 5 Application of electrospun polystyrene nanofiber-supported catalyst in the

oxidation of thioanisole.
Catalysis, 2016, 28, 144174 | 159
17:23:59.
View Online
followed the order of greater electron withdrawing eect of the

substituents.
Walmsley et al.109 also reported the co-polymerization of 1-vinylimidazole with styrene, p(ST-co-IM), (Scheme 6), after which the resulting
functional polymer was electrospun to obtain nanofibers followed by the
immobilization of oxidovanadium(IV). It was, however, reported that a
direct reaction between dissolved copolymer and VOSO4 in dimethylformamide (DMF) produced a green solid which was insoluble in several
solvents and therefore could not be electrospun. This was attributed to
the possibility of vanadium bonding to imidazole groups from neighbouring polymer chains, thereby forming crosslinks and rendering the
polymer insoluble.
Complex formation between the functional polymer and vanadium(IV)
ions is believed to occur on the nitrogen site (Fig. 10). Oxidovanadium(IV)
complexes supported on nanofibers were also employed for the oxidation
of thioanisole under continuous flow system. An excellent yield of
methylsulfonylbenzene (99.9%) was obtained for flow rates in the range of
14 mL h1. Oxidovanadium(IV) was also incorporated into poly(styreneco-4 0 -ethenyl-2-hydroxyphenylimidazole) for the oxidation of thioanisole
using hydrogen peroxide as oxidant.110 The electrospun nanofibers supported catalyst gave a high oxidation yield. The poly(styrene-co-4 0 -ethenyl2-hydroxyphenylimidazole) copolymer was prepared by polymerization of
vinylated monomer, 2-(2 0 -hydroxy-5 0 -ethenylphenyl)imidazole and styrene
(Scheme 7) in dimethyformamide (DMF) since 2-(2 0 -hydroxy-5 0 -ethenylphenyl)imidazole was not suciently soluble in styrene alone.
Oxidovanadium(IV) was also immobilized onto electrospun polybenzimidazole nanofibers, and was employed as a heterogeneous catalyst
AIBN
+
Polymerization
Scheme 6 Synthesis of p(ST-co-IM) copolymer using 2,2 0 -azobisisobutyronitrile (AIBN)

as initiator.
N
N
N
N
Available site for vanadium ion

(V=O) co-ordination
Fig. 10 p(ST-co-IM) nanofibers showing available nitrogen sites for oxidovanadium(IV) to
coordinate.
160 | Catalysis, 2016, 28, 144174
View Online
H
N
AIBN, DMF
Polymerization
N
HO
OH
HN
17:23:59.
N
Scheme 7 Synthesis of poly(styrene-co-4 0 -ethenyl-2-hydroxyphenylimidazole).
V
N
H
N
N
H
Available site for vanadium ion (V=O) to

co-ordinate
Fig. 11 Oxovanadium(IV) hosted on polybenzimidazole nanofibers (A), showing available

nitrogen sites for vanadium to coordinate (B).
in the oxidation of thioanisole.111 An oxidation yield of 99.8% methylsulfonylbenzene was recorded after 30 min under batch oxidation conditions. Complex formation between polybenzimidazole and vanadyl
oxide is also believed to take place on the available benzimidazole nitrogen site (Fig. 11). However, the geometry of the complexes formed
within the polymer could not be reconciled.
An oxovanadium(IV) Schi base vanadyl complex was attached to
electrospun polyvinyl benzylchloride nanofibers (Scheme 8) and was
evaluated for the oxidation of dibenzothiophene to dibenzothiophene
sulfone under batch conditions at 40 1C (Scheme 9).112 The catalytic
oxidation reaction of dibenzothiophene was carried out by employing
hydrogen peroxide as an oxidant. A maximum oxidation yield of dibenzothiophene sulfone (76%) was observed when a mole ratio of 4 : 1
(H2O2 : dibenzothiophene) was employed for the oxidation reaction.
A high amount of vanadium, B40% of the initial vanadium content,
leached into the solution during the first oxidation cycle raising concerns
about the stability of this catalytic material. A reduction in the overall
yield of dibenzothiophene sulfone to B52% in the second oxidation cycle
was ascribed to the low amount of vanadium available on the recycled
nanofiber mat.
3.2 Palladium catalysts
Palladium(II) bound on polymeric materials has been applied for various
organic reactions and, most importantly, it has been applied in catalysing
carboncarbon cross-coupling reactions such as the Suzuki, Heck113,114
and Sonogashira reactions.115 However, most studies have been based on
Catalysis, 2016, 28, 144174 | 161
17:23:59.
View Online
Scheme 8 Oxovanadium(IV) Schi base vanadyl complex attached to electrospun polyvinyl benzylchloride nanofibers.112
H2 O 2
H2 O
H 2 O2
H2 O
S
Catalyst, MeOH,
40C
O O
S
Catalyst, MeOH,
40C
Scheme 9 Application of electrospun nanofiber-supported catalyst in the oxidation of

dibenzothiophene.
polymer microspheres, with very few reports on the use of polymer

nanofibers.113118
3.2.1 Aryl amination (CN bond formation/coupling). Christensen
et al.116 evaluated the catalytic activity of palladium-based catalysts
attached to phosphine ligands of a polypropylene backbone (Johnson
Matthey FibreCat), for the BuchwaldHartwig amination reaction
between p-bromotoluene and piperazine in the presence of a base
NaO-t-Bu (Scheme 10). The overall reaction with the observed products
1-(4-methylphenyl)piperazine (1), 1,4-bis(4-methylphenyl)piperazine (2),
toluene (3), and 1,1 0 -dimethyl-4,4 0 -biphenyl (4) is shown in Scheme 10.
A similar product distribution was also observed when the homogeneous catalyst was employed.116 The amination products obtained
162 | Catalysis, 2016, 28, 144174
View Online
H
N
H
N
+
Catalyst,
NaO-t-Bu
+ NaBr + HO-t-Bu
17:23:59.
Solvent
N
N
H
Br
(1)
(2)
(3)
(4)
Scheme 10 Buchwald-Hartwig amination reaction using a polymer-incarcerated palladium catalyst.
O
O
O
O
butyl acrylate
O
(0.001 mol%) Pd
+
Br
Bu3N, DMF, 150C,

4h
1-(4-bromophenyl)ethanone
butyl 3-(4-acetylphenyl)acrylate
Scheme 11 Application of poly-M1-co-DMNH6-supported catalyst in the Heck coupling

of butyl acrylate with 4-bromoacetophenone.
when recycled supported catalysts were employed indicate that the

activity of the catalyst decreases significantly, going from nearly full
conversion in the first reaction cycle to about 50% conversion in
second and third reaction cycles. Further studies on the catalyst
confirmed that the deposition of NaBr on the surface of the catalyst
support as well as the leaching of palladium from the FibreCat support
are the major factors contributing to loss in activity.
3.2.2 Heck coupling reaction. Bradshaw and his co-worker117
developed Pd(II)-chitosan composite nanofibers catalyst for Heck crosscoupling reaction between iodobenzene and n-butylacrylate. Iodobenzene (1 eq.) was reacted with n-butylacrylate (1.2 eq.) in a solution
containing 2.5 eq. of triethylene amine. High cross-coupling yields
were reported and the catalyst showed excellent reusability properties.
Pawar et al.118 also reported the use of Pd(II) based complex supported
on a copolymer of 1,4,4a,5,8,8a-hexahydro-1,4,5,8-exo,endo-dimethanonaphthalene (DMNH6) and 3-(bicyclo[2.2.1]hept-5-en-2-ylmethyl)-1-(2propyl)-3,4,5,6-tetrahydropyrimidin-1-ium-2-carboxylate (M1) {poly-M1co-DMNH6} for the Heck-type coupling reaction between arylbromides
and butylacrylate. The coupling of butylacrylate and 1-(4-bromophenyl)ethanone took place at a temperature of 150 1C in the presence of
tributylamine as base (Scheme 11). A 100% yield was recorded for butyl-3(4-acetylphenyl)acrylate with turnover numbers as high as 100 000
(TOF 25 000).
Catalysis, 2016, 28, 144174 | 163
17:23:59.
View Online
3.3 Zinc catalysts

3.3.1 Photocatalytic reactions. Metallated phthalocyanine has been
reported to absorb ultraviolet light which can be applied for the degradation of organic pollutants to less toxic compounds.119 The anchoring of phthalocyanines onto polymer nanofibers can be achieved
through chemical modification of linkers (L) attached to phthalocyanines (Fig. 12). However, most studies report the use of polymer
embedded catalysts rather than covalently linked catalysts.
Zugle et al.119 reported the use of zinc(II) phthalocyanines embedded
onto electrospun polystyrene and polysulfone nanofibers for photodegradation of 4-chlorophenol, 4-nitrophenol and methyl orange. The
experiments were carried out at various concentrations of substrate and
pH, namely pH 11 for 4-chlorophenol, pH 8.2 for 4-nitrophenol, and pH
9.2 for methyl orange in aqueous buer solutions. Zinc(II) phthalocyaninesfunctionalized nanofibers (20 mg) were suspended in each sample
solution, after which a measured intensity of the irradiated light source
was applied to the reaction mixture. Benzoquinone, hydroquinone and
4,4 0 -oxydiphenol were obtained as degradation products. In a similar
study, quaternized zinc(II) 2,9,16,23-tetrakis[4-(N-methylpyridyloxy)]phthalocyanine was anchored on electrospun polysulfone nanofibers for
the photo-degradation of methyl orange.120 The degradation products of
methyl orange were identified as 2-amino-5-(3-hydroxy-4-oxo-cyclohexa2,5-dienylideneamino)-benzene sulfonic acid and poly(catechol), an
oligopolymeric product. Lutetium phthalocyanine substituted at a nonperipheral position with phenoxy groups was also embedded on electrospun polystyrene nanofibers.121 The polystyrene-supported catalyst
was employed for the photodegradation of 4-nitrophenol. The photodegradation products consisted of 4-nitrocatechol, benzoquinone and
hydroquinone. Zugle and co-worker119121 reported that the use of
nanofibers as supports for photodegradation catalysts, such as metallated phthalocyanines, demonstrated high photostability properties
N
+
M
N
N
L
Functional Polymer Nanofiber
L = Linker
M = Metal ions
Fig. 12 Proposed structure of phthalocyanine metal complexes hosted on polymer

nanofibers.
164 | Catalysis, 2016, 28, 144174
View Online
17:23:59.
115
with support material reusability. Likewise, Kann

established that
attaching an organometallic catalyst to a solid support oers numerous
advantages such as the reduction of metallic residue contamination
of final reaction products compared to running the reaction with
unsupported organometallic catalyst in solution.
Nanofiber-supported metal complexes oer excellent properties for use
as catalyst support amongst which is a high surface area-to-volume ratio,
giving reactants easy access to a large number of active sites. Concerns
have been raised over nanofiber solubility, nanofiber strength and metal
catalyst leaching, especially when supports are employed under extreme
conditions such as high temperature, but nanofiber supports are mostly
employed under conditions mild enough to reduce catalyst support
breakdown and catalyst leaching.111,116 Some catalyzed reactions do require high temperature in order to achieve the desired products, and
strategies for crosslinking of nanofibers are currently being developed so
as to improve the strength of the support.122,123 Some of these interventions result in the support materials being brittle and less robust.
Some of the various concerns encountered with the use of nanofibers as
catalyst support are discussed in Section 4. On the other hand, crosslinked polymer microspheres have been applied extensively as a support
for organometallic catalysts despite the low surface area-to-volume
ratio they oer as compared to electrospun nanofibers. Various other
applications of polymer supported catalysts as reviewed by Kann115 and
Tsuchida et al.124 are presented in Table 2.
4 Challenges associated with electrospun nanofibers

4.1 Solubility problems and morphological changes
Walmsley136 demonstrated that electrospun nanofibers of styrene copolymerized with vinyl imidazole p(ST-co-IM) dissolve or change
morphology when they are allowed to come into contact with some
selected organic solvents. Electrospun p(ST-co-IM) fibers when placed in
acetonitrile or even in an acetonitrile/water mixture, lose their structural integrity as shown in the SEM image in Fig. 13. In a related
situation, styrene copolymerized with 4-vinylbenzyl chloride p(STco-PVC) showed a change in fiber morphology when it came into
contact with toluene (Fig. 13). Both nanofibers {p(ST-co-IM) and p(ST-coPVC)} were also reported to be soluble in organic solvents such as
chloroform, dichloromethane, and dimethylformamide (DMF),
amongst others. However, methanol did not dissolve or change the
nanofiber morphology, and for this reason, methanol was chosen as
the suitable solvent for vanadium-incorporated nanofiber-catalyzed
reactions.136
Most electrospun nanofibers are reported to have low tensile strength
with the exception of a few such as polybenzimidazole and nylon. The
tensile strength of nanofibers can be improved through crosslinking, but
most crosslinking agents compete with the reactive donor atoms on
polymers in coordinating metal ions. This phenomenon may result in
the reduction of polymer active sites available for metal ion coordination.
Catalysis, 2016, 28, 144174 | 165
View Online
17:23:59.
Table 2 Polymer support used in catalysis.
Functional polymer support
Metal/metal
complex
involved
Merrifield resin
Cu(II)
Polymer-bound proline
Cu(II)
Poly(4-vinyl pyridine)
Cu(II)
2-Vinylpyridine crosslinked
with ethylene glycol acrylate
Polystyrene-supported
bidentate aminophosphanephosphite and -phosphinite
ligands
Polystyrene supported
diphosphite ligand named
JanaPhos
Salen-attached to Merrifield
resin
Rh(II)
Polystyrene bound
diphenylphosphinooxazoline
(PHOX) ligands
Merrifield resin bound with
bis(2-picolyl)amine
Polymer bound pyridine
Merrifield resin
Reactions catalysed
Ref.
Cycloaddition reaction of
various imines with
N-phenylmaleimide and
methyl fumarate
Cross-coupling of oximes with
arylboronic acids
1. Oxidation of thioderivatives
2. Oxidative polymerisation of
phenols
Carbonylation of methanol
125
126
127, 128
129
Rh(II)
Asymmetric hydrogenation of
conjugated methyl esters
130
Rh(I)
Conjugate addition of boronic

acids to enones
131
Pd(II)
Suzuki reaction of aryl

bromides with
phenylboronic acid
Enantioselective amination
of allylic acetates
132
Ru(III)
Ring Opening of Epoxides
134
Ruthenium
carbine
V(IV)
Metathesis
135
Oxidation reactions
6, 136139
Pd(II)
133
As mentioned earlier, polymer nanofibers sometimes become brittle

upon crosslinking and the surface area can also be compromised.136,140
The nanofiber morphology can also be altered upon crosslinking, as was
observed in an attempt to crosslink chitosan nanofibers with glutaraldehyde (Scheme 12 and Fig. 14).140
4.1.1 Leaching. It is an undisputed fact that the use of electrospun
nanofibers as solid supports for inorganic catalysts has heralded a new
era in heterogeneous catalysis on the laboratory scale an era of catalysts that are easy to separate from reaction mixtures that are easy to
recover (recyclable) and are also reusable. However, these successes
have come with two challenges, the leaching of immobilized or embedded catalyst, and the solubility of the nanofibers. It seems that the
metal oxide-based catalysts do not suer from leaching, which can be
attributed to their lack of solubility in most reaction media. Metal
nanoparticles, on the other hand, have been found to undergo various
degrees of leaching; from not observable40 or insignificant43 to a significant amount of leached nanoparticles.44 The use of highly oxidizing
reaction media results in significant amounts of leaching due to the
166 | Catalysis, 2016, 28, 144174
17:23:59.
View Online
Fig. 13 SEM images of (A) p(ST-co-IM) nanofibers, (B) p(ST-co-IM) nanofibers in acetonitrile, (C) p(ST-co-PVC), and (D) p(ST-co-PVC) nanofibers in toluene.
oxidation of Mo to Mn1. In most cases, leaching of the nanoparticles is

manifested by the loss of catalyst eciency between subsequent reaction cycles.
Leaching of coordinated metal ions in functional nanofibers is another
major concern, as the activity of the reused supported catalysts is reduced
in subsequent cycles.111,116,141 Christensen et al.116 reported that the reused catalyst (FibreCat) for the amination reaction gave more side
products with a reduction in the overall yields, which is most likely due
to leaching of palladium(II) from the nanofiber support. Walmsley
et al.108111 also reported the leaching trend of oxidovanadium(IV) coordinated to polymer nanofibers. The catalysts incorporated through
embedding within the polymer via weak pp interactions were reported
to result in significant leaching of the metal complexes.108 Of interest was
their study of the oxidation of thioanisole using oxidovanadium(IV) coordinated to polybenzimidazole.109 An initial leaching of 7.29% vanadium after the first reaction was observed, however, a lower degree of
leaching between 0.52 and 0.69% in the subsequent cycles under batch
conditions was observed. The drop in activity observed during the recycling experiments was attributed to the leaching of vanadium, and a
similar trend was also observed by Miller and Sherrington in microspherical catalyst hosts.142 In order to reduce leaching of vanadium, a
continuous flow oxidation process was developed by Walmsley et al.108
They also reported that 0.37%0.81% vanadium leached out of the
nanofiber after several oxidation cycles. This small degree of leaching did
Catalysis, 2016, 28, 144174 | 167
17:23:59.
168 | Catalysis, 2016, 28, 144174
HO
NH2
NH2
OH
O*
OH
OH
HO
O
O
OH
OH
NH2
OH
*O
n
HO
HO
HO
HO
OH
OH
OH
NH2
O
crosslinker
HO
NH2
N
NH2
OH
*O
OH
OH
OH
OH O
OH
OH
HO
OH
NH2
OH
HO
OH
NH2
OH
NH2
OH
*O
NH2
Scheme 12 Proposed reaction scheme for the crosslinking of chitosan nanofibers using glutaraldehyde as crosslinker.
OH
OH
17:23:59.
View Online
Fig. 14 SEM images of (A) electrospun chitosan nanofibers and (B) crosslinked chitosan
nanofibers (cross linker is glutaraldehyde).
not aect the oxidation conversions in each cycle. Leaching of vanadium

from the electrospun nanofibers support was reduced by incorporating
a bidentate ligand (Scheme 7) instead of the monodentate imidazole
ligand (Scheme 7), as the bidentate coordination resulted in greater
stabilization of oxidovanadium(IV) relative to the monodentate
counterpart.136
Conclusions
The utilization of functional nanoparticles and metal oxides immobilized

on electrospun nanofibers continues to advance, and it seems that the
use of electrospun nanofibers as solid support materials will also progress well into the future. The flexibility of this support system as well as
the additional advantages such as high surface area and porosity make it
stand out as a support material of choice for application in certain
catalysed systems, be they membrane catalysts or particulate catalysts in
batch reactors. Challenges such as catalyst leaching and solubility of the
support system continue to plague progress in the application of these
materials. However, strategies are being sought to ameliorate these
technical problems. There are also a few studies reporting on the coordination chemistry approach for immobilization of catalysts onto
electrospun nanofibers. The relative paucity of such studies may be attributed to the infancy of the field or to some of the challenges associated
with working with these non-crosslinked materials.
References
1
2
3
4
5
L. Li, S. Fan, X. Mu, Z. Mi and C. J. Li, J. Am. Chem. Soc., 2014, 136, 7793.
J. S. Cannon, L. Zou, P. Liu, Y. Lan, D. J. Oleary, K. N. Houk and
R. H. Grubbs, J. Am. Chem. Soc., 2014, 136, 6733.
A. Budhai, B. Omondi, S. O. Ojwach, C. Obuah, E. Y. Osei-Twum,
E. Y. Darkwa and J. Tandem, Catal. Sci. Technol., 2013, 3, 3130.
L. M. Ombaka, P. Ndungu and V. O. Nyamori, Catal. Today, 2013, 217, 65.
Z. R. Tshentu, C. Togo and R. S. Walmsley, J. Mol. Catal. A: Chem., 2010,
318, 30.
Catalysis, 2016, 28, 144174 | 169
View Online
6
7
8
17:23:59.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
A. S. Ogunlaja, S. M. Khene, E. Antunes, T. Nyokong, N. Torto and

Z. R. Tshentu, Appl. Catal., A, 2013, 462463, 157.
L. M. Sikhwivhilu, N. J. Coville, D. Naresh, K. V. R. Chary and
V. Vishwanathan, Appl. Catal., A, 2007, 324, 52.
L. M. Sikhwivhilu, S. S. Ray and N. J. Coville, Solid State Phenom., 2008,
140, 61.
Z. M. Huang, Y. Z. Zhang, M. Kotaki and S. A. Ramakrishna, Compos. Sci.
Technol., 2003, 63, 2223.
T. Ondaruhu and C. Joachim, Europhys. Lett., 1998, 42, 215.
C. R. Martin, Chem. Mater., 1996, 8, 1739.
J. H. He, Y. Liu, L. F. Mo, Y. Q. Wan and L. Xu, Electrospun Nanofibers and
their Applications, iSmithers, Shopshire, Shawbury, UK, 2008.
Y. Dzenis, Science, 2004, 304, 1917.
A. Greiner and J. H. Wendor, Angew. Chem., Int. Ed., 2007, 46, 5670.
A. Kumar, Nanofibers, Intech, Vukovar, Croatia, 2010.
D. Bjorge, N. Daels, S. De Vrieze, P. Dejans, T. van Camp, W. Audenaert,
P. Westbroek, K. De Clerck, C. Boeckaert and S. W. H. van Hulle, Water SA.,
2010, 36, 151.
Z. Dong, S. J. Kennedy and Y. Wu, J. Power Sources, 2011, 196, 4886.
lgen, S. Egri and I. A. Isoglu, Nanomedicine, 2007, 2, 441.
E. Piskin, N. Bo
U. Boudriot, R. Dersch, A. Greiner and J. H. Wendor, Artif. Organs, 2006,
30, 785.
D. S. Katti, K. W. Robinson and C. T. Laurencin, J. Biomed. Mater. Res.,
Part B, 2004, 70, 286.
J. Venugopal and S. Ramakrishna, Appl. Biochem. Biotechnol., 2005, 125, 147.
D. A. Ondigo, Z. R. Tshentu and N. Torto, Anal. Chim. Acta, 2013, 804, 228.
S. Chigome, G. Darko and N. Torto, Analyst, 2011, 136, 2879.
n and R. Leino, Org. Lett., 2001, 3, 2551.
S. Degni, C. E. Wile
C. Xuefei, H. Yang and X. Zhen-Liang, Mater. Lett., 2011, 65, 1719.
ben, N. Rosenberger, F. Schleth, A. Studer, A. Greiner and
M. Stasiak, C. Ro
J. H. Wendor, Polymer, 2007, 48, 5208.
R. F. Heck and J. P. Nolley Jr., J. Org. Chem., 1972, 37, 2320.
N. Miyaura and A. Suzuki, J. Chem. Soc., Chem. Commun., 1979, 866.
S. Baba and E. Negishi, J. Am. Chem. Soc., 1976, 98, 6729.
D. Milstein and J. K. Stille, J. Am. Chem. Soc., 1979, 101, 4992.
N. Miyaura and S. L. Buchwald, Cross-coupling reactions: A practical guide,
Springer, New York, 2002, vol. 219.
Palladium-catalyzed coupling reactions: Practical aspects and future develop. Molna
r, Wiley-VCH, Germany, 2013, vol. 1.
ments, ed. A
R. Narayanan, Molecules, 2010, 15, 2124.
B. Karimi, H. Behzadnia, E. Farhangi, E. Jafari and A. Zamani, Curr. Org.
Synth., 2010, 7, 543.
J. Guerra and M. A. Herrero, Nanoscale, 2010, 2, 1390.
rez, P. Lara, A. B. Garca and K. Philippot, Recent Pat.
E. J. Garca-Sua
Nanotechnol., 2013, 7, 247.
H. Dong, D. Wang, G. Sun and J. P. Hinestroza, Chem. Mater., 2008,
20, 6627.
S. N. Sidorov, L. M. Bronstein, Y. A. Kabachii, P. M. Valetsky, P. L. Soo,
D. Maysinger and A. Eisenberg, Langmuir, 2004, 20, 3543.
E. A. Obuya, W. Harrigan, D. M. Andala, J. Lippens, T. C. Keane and
W. E. Jones Jr., J Mol. Catal. A: Chem., 2011, 340, 89.
L. Chen, S. Hong, X. Zhou, Z. Zhou and H. Hou, Catal. Commun., 2008,
9, 2221.
170 | Catalysis, 2016, 28, 144174
View Online
41
42
43
17:23:59.
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
K. Sambasivudu, Y. B. Reddy, J. S. Yadav, G. Sabitha and D. Shailaja, Int. J.

Polym. Mater., 2008, 57, 891.
S. Chuangchote, J. Jitputti, T. Sagawa and S. Yoshikawa, ACS Appl. Mater.
Interfaces, 2009, 1, 1140.
L. Guo, J. Bai, C. Li, H. Liang, W. Sun, Q. Meng and T. Xu, New J. Chem.,
2013, 37, 4037.
E. Formo, M. S. Yavuz, E. P. Lee, L. Lane and Y. Xia, J. Mater. Chem., 2009,
19, 3878.
J. Prakash, D. A. Tryk and E. B. Yeager, J. Electrochem. Soc., 1999, 146, 4145.
ver, J. Electroanal. Chem., 1999, 461, 14.
S. Wasmus and A. J. Ku
E. Formo, Z. Peng, E. Lee, X. Lu, H. Yang and Y. Xia, J. Phys. Chem. C, 2008,
112, 9970.
Z. Chen, L. Xu, W. Li, H. Waje and Y. Yan, Nanotechnology, 2006, 17,
5254.
G. H. An and H. J. Ahn, ECS Solid State Lett., 2014, 3, M29M32.
N. A. M. Barakat, M. A. Abdelkareem, M. El-Newehy and H. Y. Kim,
Nanoscale Res. Lett., 2013, 8, 402.
N. A. M. Barakat, M. El-Newehy, S. S. Al-Deyab and H. Y. Kim, Nanoscale Res.
Lett., 2014, 9, 2.
F. Sonmez, H. Ercan, H. Genc, M. Arslan, M. Zengin and
M. Kucukislamoglu, J. Chem., 2013, 169109.
M. M. Demir, M. A. Gulgun and Y. Z. Menceloglu, Macromolecules, 2004,
37, 1787.
K. Ebert, G. Bengtson, R. Just, M. Oehring and D. Fritsch, Appl. Catal., A,
2008, 346, 72.
A. C. Patel, S. Li, C. Wang, W. Zhang and Y. Wei, Chem. Mater., 2007,
19, 1231.
E. Formo, E. Lee, D. Campbell and Y. Xia, Nano Lett., 2008, 8, 668.
Y. Dai, B. Lim, Y. Yang, C. M. Cobley, W. Li, E. C. Cho, B. Grayson,
P. T. Fanson, C. T. Campbell, Y. Sun and Y. Xi, Angew. Chem., Int. Ed., 2010,
49, 8165.
M. Graeser, E. Pippel, A. Greiner and J. H. Wendor, Macromolecules, 2007,
40, 6032.
X. Fang, H. Ma, S. Xiao, M. Shen, R. Guo, X. Cao and X. Shi, J. Mater. Chem.,
2011, 21, 4493.
X. Lu, X. Bian, G. Nie, C. Zhang, C. Wang and Y. Wei, J. Mater. Chem., 2012,
22, 12723.
A. O. Ibhadon and P. Fitzpatrick, Catalysts, 2013, 3, 189.
J. M. Herrmann, Catal. Today, 1999, 53, 115.
S. Tombe, E. Antunes and T. Nyokong, J. Mol. Catal. A: Chem., 2013,
371, 125.
P. Modisha and T. Nyokong, J. Mol. Catal. A: Chem., 2014, 381, 132.
J. Sundaramurthy, N. Li, S. Kumar and S. Ramakrishna, Biofuel Res. J., 2014,
2, 44.
S. Zhan, D. Chen, X. Xiao and Y. Song, Chem. Commun., 2007, 2043.
and M. Ritala, Nanotechnology, 2009,
E. Santala, M. Kemell, M. Leskela
20, 035602.
V. Pongsorrarith, C. Srissithiratkul, K. Laohhasurayotin and N. Intasanta,
Mater. Lett., 2010, 67, 1.
C. J. Li, J. N. Wang, B. Wang, J. R. Gong and Z. Lin, Mater. Res. Bull., 2012,
47, 333.
Z. Zhang, C. Shao, L. Zhang, X. Li and X. Liu, J. Colloid Interface Sci., 2010,
351, 57.
Catalysis, 2016, 28, 144174 | 171
View Online
17:23:59.
71
N. A. M. Barakat, M. A. Kanjwal, S. S. Al-Deyab and I. S. Chronakis, Mater.

Sci. Appl., 2011, 2, 1188.
72 S. Mishra and S. P. Ahrenkiel, J. Nanomater., 2012, 902491.
73 J. P. Jeun, D. W. Park, D. K. Seo, H. B. Kim, Y. C. Nho and P. H. Kang, Rev.
Adv. Mater. Sci., 2011, 28, 26.
74 S. Liu, B. Liu, K. Nakata, T. Ochiai, T. Murakami and A. Fujishima,
J. Nanomater., 2012, 491927.
75 H. Y. Kim, B. H. Kim and K. S. Yang, Carbon, 2012, 50, 2472.
76 M. S. Amin and A. Z. Moshfegh, Fabrication and characterization of electrospun composite nanofibers: A photocatalytic study-Proceedings of the 4th
International Conference on Nanostructures (ICNS4), 1214 March, 2012, Kish
Island, I.R., Iran.
77 C. G. Song, S. K. Koppala and J. W. Yoon, J. Ceram. Process. Res., 2013,
14, 653.
78 M. Hu, M. Fang, C. Tang, T. Yang, Z. Huang, Y. Liu, X. Wu and X. Min,
Nanoscale Res. Lett., 2013, 8, 548.
79 S. An, B. N. Joshi, M. W. Lee, N. Y. Kim and S. S. Yoon, Appl. Surf. Sci., 2014,
294, 24.
80 Z. Zhang, C. Shao, X. Li, L. Zhang, H. Xue, C. Wang and Y. Liu, J. Phys. Chem.
C, 2010, 114, 7920.
81 M. Shi, M. Zhou, D. Song, X. Pan, J. Fu, J. Zhou, S. Ma and T. Wang, Ceram.
Int., 2014, 40, 10383.
oz-Esp, G. Glasser, M. M. Demir, K. Landfester and
82 N. Horzum, R. Mun
D. Crespy, ACS Appl. Mater. Interfaces, 2012, 4, 6338.
83 P. Du, L. Song, J. Xiong and H. Cao, J. Mater. Sci., 2013, 48, 8386.
84 J. Li, H. D. Y. Qiao, C. Chen, X. Li, J. Cui, D. Kumar and Q. Wei, Sci. World J.,
2012, 154939.
85 E. A. Obuya, P. C. Joshi, T. A. Gray, T. C. Keane and W. E. Jones Jr., Int. J.
Chem., 2014, 6, 1.
86 W. Wang, N. Li, Y. Chi, Y. Li, W. Yan, X. Li and C. Shao, Ceram. Int., 2013,
39, 3511.
87 Y. Yang, L. Xu, C. Su, J. Che, W. Sun and H. Gao, J. Nanomater., 2014,
130539.
88 X. Meng, X. Feng, Z. Li and Y. Zhand, Curr. Nanosci., 2012, 8, 3.
89 Y. N. Feng, H. C. Wang, Y. D. Luo, Y. Shen and Y. H. Lin, J. Appl. Phys., 2013,
113, 146101.
90 Z. Gholamvand, Int. J. Mod. Phys.: Conf. Ser., 2012, 5, 622.
91 C. Basheer, J. Chem., 2013, 562305.
92 M. Jin, X. Zhang, A. V. Emeline, Z. Liu, D. A. Tryk, T. Murakami and
A. Fujishima, Chem. Commun., 2006, 4483.
93 A. Javid, S. Nasseri, A. Mesdaghinia, A. H. Mahvi, M. Alimohammadi,
R. M. Aghdam and N. Rastkari, J. Environ. Health Sci. Eng., 2013, 11, 24.
94 D. Zhang, X. Li, H. Tan, G. Zhang, Z. Zhao, H. Shi, L. Zhang, W. Yu and
Z. Sun, RSC Adv., 2014, 4, 44322.
95 S. Chuangchote, J. Jitputti, T. Sagawa and S. Yoshikawa, ACS Appl. Mater.
Interfaces, 2009, 1, 1140.
96 S. Xu, D. Sun, H. Liu, X. Wang and X. Yan, Catal. Commun., 2011, 12, 514.
97 J. Fu, S. Cao, J. Ju, J. Low and Y. Lei, Dalton Trans., 2014, 3, 9158.
98 R. L. Gustafson and J. A. Lirco, J. Phys. Chem., 1965, 69, 2849.
99 F. Yamashita, T. Komatsu and T. Nagagawa, Bull. Chem. Soc. Jpn., 1976,
49, 2073.
100 A. Nagai, A. Takahashi, T. Komatsu and T. Nakagawa, Polym. J., 1988,
20, 609.
172 | Catalysis, 2016, 28, 144174
View Online
101
102
17:23:59.
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
M. Kaneko, E. Tsuchida, K. Yamada and Y. Kurimura, J. Polym. Sci., 1971,

A-19, 3521.
B. B. De, B. B. Lohray, S. Sivaram and P. K. Dhal, J. Polym. Sci., Polym. Chem.
Ed., 1997, 35, 1809.
J. H. Barnis and G. F. Esstemont, Makromol. Chem., 1976, 177, 307.
Y. Morishima, K. Frejisawa and S. Nazakura, J. Polym. Sci. Polym. Lett. Ed.,
1978, 6, 141.
C. U. Pittman, R. L. Voges and J. Elder, Macromolecules, 1971, 4, 291, 298,
302.
R. J. Card and D. C. Neckers, J. Am. Chem. Soc., 1977, 99, 7733.
J. A. Welleman, F. B. Hulsbergen and J. Reedijk, Makromol. Chem., 1981,
182, 785.
R. S. Walmsley, S. Chigome, N. Torto and Z. R. Tshentu, Catal. Lett., 2012,
142(2), 243.
R. S. Walmsley, A. S. Ogunlaja, M. J. Coombes, W. Chidawanyika,
C. Litwinski, N. Torto, T. Nyokong and Z. R. Tshentu, J. Mater. Chem., 2012,
22, 5792.
R. S. Walmsley, C. Litwinski, E. Antunes, P. Hlangothi, E. Hosten,
C. McCleland, T. Nyokong, N. Torto and Z. R. Tshentu, J. Mol. Cat. A: Chem.,
2013, 379, 94.
R. S. Walmsley, P. Hlangothi, C. Litwinski, T. Nyokong, N. Torto and
Z. R. Tshentu, J. Appl. Polym. Sci., 2013, 127(6), 4719.
A. S. Ogunlaja, E. C. Hosten and Z. R. Tshentu, S. Afr. J. Chem., 2015,
68, 172.
R. F. Heck and J. P. Nolley Jr., J. Org. Chem., 1972, 37(14), 2320.
T. Mizoroki, K. Mori and A. Ozaki, Bull. Chem. Soc. Jpn., 1971, 44(2), 581.
N. Kann, Molecules, 2010, 15, 6306.
H. Christensen, S. Kiil and K. Dam-Johansen, Org. Process Res. Dev., 2007,
11, 956.
M. Bradshaw, J. Zou, L. Byrne, K. SwaminathanIyer, S. G. Stewart and
C. L. Raston, Chem. Commun., 2011, 47, 12292.
G. M. Pawar and M. R. Buchmeiser, Adv. Synth. Catal., 2010, 352, 917.
R. Zugle and T. Nyokong, J. Appl. Polym. Sci., 2013, 128, 1131.
R. Zugle and T. Nyokong, J. Mol. Catal. A: Chem., 2013, 366, 247.
R. Zugle and T. Nyokong, J. Mol. Catal. A: Chem., 2012, 358, 49.
A. C. Oyrton, Monteiro Jr. and A. Claudio, Int. J. Biol. Macromol., 1999,
26, 119.
H. Y. Li and Y. L. Liu, J. Mater. Chem. A, 2013, 1, 1171.
E. Tsuchida and H. Nishide, Adv. Polym. Sci., 1977, 24, 1.
B. Martin-Matute, S. I. Pereira, E. Pena-Cabrera, J. Adrio, A. M. S. Silva and
J. C. Carretero, Adv. Synth. Catal., 2007, 349, 1714.
L. Wang, C. Y. Huang and C. Cai, Catal. Commun., 2010, 11, 532.
M. K. F. Chanda, OPriscoll and G. C. Ranpel, J. Catal., 1980, 61, 533.
E. Tsuchida, M. Kaneko and H. Nishide, Makromol. Chem., 1973, 164, 203.
S. Zhang, C. Guo, Q. Qian and G. Yuan, Catal. Commun., 2008, 9, 853.
R. den Heeten, B. H. G. Swennenhuis, P. van Leeuwen, J. G. de Vries and
P. C. J. Kamer, Angew. Chem., Int. Ed., 2008, 47, 6602.
R. Jana and J. A. A. Tunge, Org. Lett., 2009, 11, 971.
N. T. S. Phan and P. Styring, Green Chem., 2008, 10, 1055.
D. Popa, R. Marcos, S. Sayalero, A. Vidal-Ferran and M. A. Pericas, Adv.
Synth. Catal., 2009, 351, 1539.
S. H. Lee, E. Y. Lee, D. W. Yoo, S. J. Hong, J. H. Lee, H. Kwak, Y. M. Lee,
J. Kim, K. A. Cheal and J. K. Lee, New J. Chem., 2007, 31, 1579.
Catalysis, 2016, 28, 144174 | 173
View Online
135
136
17:23:59.
137
138
139
140
141
142
A. Kirschning, K. Harmrolfs, K. Mennecke, J. Messinger, U. Schon and

K. Grela, Tetrahedron Lett., 2008, 49, 3019.
R. S. Walmsley, An investigation into the antidiabetic and catalytic properties
of oxovanadium(IV) Complexes, PhD thesis, Rhodes University, 2011.
A. S. Ogunlaja, W. Chidawanyika, E. Antunes, M. A. Fernandes, T. Nyokong,
N. Torto and Z. R. Tshentu, Dalton Trans., 2012, 4145, 13908.
M. R. Maurya, A. Arya, A. Kumar and J. C. Pessoa, Dalton Trans., 2009, 2185.
M. R. Maurya, M. Kumar, A. Kumar and J. C. Pessoa, Dalton Trans., 2008,
4220.
A. S. Ogunlaja, M. J. Coombes, N. Torto and Z. R. Tshentu, React. Funct.
Polym., 2014, 81, 61.
ben, N. Rosenberger, F. Schleth, A. Studer, A. Greiner and
M. Stasiak, C. Ro
J. H. Wendor, Polymer, 2007, 48, 5208.
M. M. Miller and D. C. Sherrington, J. Catal., 1995, 152, 377.
174 | Catalysis, 2016, 28, 144174
17:24:10.
Elucidation of mechanistic and kinetic

aspects of watergas shift reaction on
supported Pt and Au catalysts via
transient isotopic techniques
Angelos M. Efstathiou
DOI: 10.1039/9781782626855-00175
With growing interest in hydrogen fuel cell applications, novel and robust watergas shift
(WGS) industrial catalysts are required with high activity, preferably at low-temperatures
(To250 1C). Supported Pt and Au on reducible metal oxides appear very promising
candidates for such applications. Towards this goal, the need for in-depth investigations of
the actual mechanism(s) and micro-kinetics of potential catalytic systems and the need to
obtain intrinsic correlations between site reactivity and chemical composition (quality and
quantity) of the active adsorbed phase are imperative. The use of in situ kinetics coupled
with spectroscopic measurements (operando methodology) fulfils these requirements.
Steady state isotopic transient kinetic analysis (SSITKA) used in the operando mode
coupled with other transient isotopic experiments can sensitively probe the chemical
nature and surface composition of the active reaction intermediates. Rival mechanisms
(redox versus associative with OH group regeneration) can also be screened and
kinetic parameters (e.g. rate constants) can be obtained through advanced micro-kinetic
modelling of the transient isotopic responses of both gaseous and adsorbed species. The
mechanisms of the WGS reaction on supported Pt (or Au) metal catalysts on the basis
of recent DFT computations and experimental investigations, and the application of
SSITKA-operando (FTIR-Mass Spectroscopy) methodology with a focus on its advantages
and limitations are presented through several case studies.
Introduction
The watergas shift (WGS) reaction (eqn (1)) discovered by the Italian
physicist F. Fontana in 1780 is a reversible and slightly exothermic
chemical reaction known as an old industrial process to obtain hydrogen
and carbon dioxide.
CO H2O2CO2 H2
(DH1 40.6 kJ mol1)
(1)
WGS is nowadays a very important step in the industrial catalytic

processes of hydrogen production via the conversion of natural gas and
some other hydrocarbons feedstock through steam reforming and partial
oxidation,18 and the production of other important feedstock chemicals,
such as ammonia and methanol.911 The current use of the WGS reaction
followed the increased industrial demand for hydrogen. The state-of-theart steam reforming of methane (SRM) industrial process carries out
WGS in two stages in order to overcome the thermodynamic limitations
of the reaction at high temperatures but to also achieve complete CO
conversion (unconverted CO below 0.1 vol%).1,12,13 The modern two-stage
Chemistry Department, Heterogeneous Catalysis Laboratory, University of Cyprus,
1 University Ave., University Campus, 1678 Nicosia, Cyprus. E-mail: efstath@ucy.ac.cy
Catalysis, 2016, 28, 175236 | 175

c
17:24:10.
View Online
WGS system reduces the CO concentration to about 0.1 vol% from

the high level of 1020 vol% obtained in the outlet of the reformer
(reformate gas product) in a high-temperature (350500 1C) shift (HTS)
over Fe2O3/Cr2O3 catalysts, followed by a low-temperature (200250 1C)
shift (LTS) over Cu/ZnO/Al2O3 catalysts.1,2,12,13
With growing interest in small scale hydrogen fuel cell mobile applications,14,15 the two-stage currently practiced WGS is not an appropriate
choice because of technical complexities (e.g. starting up and shutting
down the engine) and safety concerns due to the pyrophoric nature of the
catalysts used.13 In the last decade, WGS reactors in fuel cell systems were
using industrial FeCr- and Cu-based catalysts.16 The former catalytic
systems in spite of the fact of being cheap and stable, they operate at
temperatures higher than 350 1C for acceptable activity, whereas Cubased catalysts in spite of their good activity at low-temperatures are
susceptible to sulphur poisoning. In fact, pre-removal of sulphur to below
100 ppb is required, whereas deactivation of these catalysts limit their
life-time to 24 years.2 In addition, they exhibit a rather narrow
temperature window due to their sintering problems (T4250 1C).2,12,13
Several other drawbacks of Cu-based catalysts include: (i) slow and controlled activation procedures, (ii) need to be purged with inert gas upon
shutdown to avoid water condensation and re-oxidation, and (iii) catalyst
deactivation due to water condensation that could result upon repeated
start-up and shutdown operation cycles.12,17 All the above mentioned
drawbacks of the existing WGS industrial catalysts make imperative the
development of new and robust catalysts towards the production of clean
hydrogen for fuel cell applications.
In recent years, reducible metal oxides used as carriers of nano-size
Au and Pt particles appeared to be very promising catalytic systems
for fuel cell reformers because of their exceptional high activity at
low-temperatures (To300 1C) and favourable kinetics compared to
Cu-based catalysts.2,5,1830 The Andreeva group in 199618,19 and the
Flytzani-Stephanopoulos group in 200120 were the first to report the
excellent activity at low-temperatures of nanostructured Au deposited on
Fe2O3 and CeO2, respectively. In these early research works the important
role of support was addressed and in a recently published review paper29
recent advances in the development of supported gold catalysts for
the WGS reaction (e.g. chemical composition of support, preparation
methods of supported Au, morphology and particle size of the support,
chemical promotion) and fundamental insights that can be gained (e.g.
nature of active sites, reaction and deactivation mechanisms) are
described. In another recent review paper,30 it is shown that the eect of
support is indirect, where partially oxidized gold atoms, AuO(OH)x-s
are the active catalytic sites of low-temperature WGS, with s being a
support atom in the vicinity of the Au atom that activates water and
supplies OH groups. On the other hand, the key question of whether the
support chemical composition and morphological characteristics aect
the mechanism of the WGS reaction is still a debated issue.
Supported platinum on reducible metal oxide carriers has received
much interest for low-temperature WGS given the prior knowledge of the
176 | Catalysis, 2016, 28, 175236
View Online
17:24:10.
1
dierences in turnover frequency (TOF, s ) among the extended surfaces

of Cu and Pt.3133 In fact, it has been reported31,32 that the TOF of Pt(111)
is initially five times larger than that obtained on Cu(111). However, due
to carbon formation by the Boudouard reaction (2 CO-s2C-s CO2),
platinum surfaces are prone to deactivation. It has been reported34,35 that
Pt-based catalysts were used in a fuel processor for stationary applications (210 kW electric load) in a single stage WGS reactor, where a
significant reduction in reactor volume was achieved compared to a
Cu-based catalyst. Among the various metal oxide supports investigated
as reviewed elsewhere,2,28,36 those based on CeO2 or TiO2 appear to be the
most promising ones for low-temperature WGS. In recent works,3740
CeO2 was doped with La31, Zr41 and Ti41 at the level of 2030 at.% using
modified solgel methods, ultimately forming a solid solution. After the
deposition of 0.5 wt% Pt (by wet impregnation), the highest kinetic rate
of WGS (3 vol% CO/10 vol% H2O/He) at 300 1C was obtained in the case of
Ti41-doped ceria supported Pt. The latter exhibited 7.5 and 5.8 times
higher activity compared to Pt/CeO2 and Pt/TiO2 catalysts, respectively.39
Using the SSITKA technique the authors were able to show that the
kinetic rate of WGS (mmol CO g1 s1) was correlated with the concentration (mmol g1) of the active carbon-containing intermediates (C-pool)
formed on both the Pt and a region of supports surface around each Pt
nanoparticle (dPt 1.41.7 nm) that extended to less than 10 (equivalent
to about two support lattice constants) from the Pt-support interface. The
formation of sites of dierent reactivity at the Pt-support interface was
proposed to explain these results.3740 This explanation finds support
from recent density functional theory (DFT) studies over a model Pt/CeO2
(111) catalytic surface, where large electronic perturbations for small Pt
particles enhances their ability to dissociate the OH bonds of water and
to better accommodate H2O, OH and H adsorbed species.41 Turnover
frequencies of 12.4 s1 (yPt 0.15 ML) and 0.53 s1 were estimated for
Pt/CeO2(111) and Pt(111), respectively. A comparison of the performance
of Pt/CeO2(111) with those of Cu/CeO2(111) and Au/CeO2(111) illustrated
that Pt is the best admetal in terms of metal site reactivity and low
loading.41,42
The enhanced WGS catalytic activity of 0.5 wt% Pt/Ce0.8Ti0.2O2d
compared to 0.5 wt% Pt/CeO2 and 0.5 wt% Pt/TiO239 is in harmony
with the findings of Bruix et al.41 who have shown that PtCeOx/TiO2(110)
were about 2.5 times more active than the corresponding Pt/CeO2(111)
catalyst. The key conclusion from the latter studies41 was that the
Pt2ceria and ceria2titania interactions significantly enhance the
ability of Pt to adsorb water, dissociate the OH bonds in the adsorbed
water molecule and better accommodate the OH and H adspecies. This is
the result of the active phase of PtCeOx/TiO2(111) exhibiting a density of
Pt 5d states near the Fermi level of Pt that is much smaller than that
expected for bulk metallic Pt and is a consequence of preparing small
clusters of Pt and nanoparticles of reduced ceria (Ce2O3). The synergy of
these two results in the enhancement of water dissociation, thus of WGS
activity; dissociation of water is considered the rate-limiting step on
unsupported Pt surfaces.24,32
Catalysis, 2016, 28, 175236 | 177
View Online
17:24:10.
The eect of alkali ions (Na or K ) added in small amounts on an

irreducible metal oxide support (e.g. SiO2 or Al2O3) on the enhancement
of the low-temperature WGS activity of the corresponding supported Pt
catalyst has been demonstrated by Zhai et al.43 A Pt/SiO2 catalyst showed
practically no conversion of CO at 250 1C (2% CO/10% H2O/He,
W/F 0.09 g s cm3), whereas Pt/Na-SiO2 (Na/Pt 3/1 at./at.) resulted in a
remarkable increase of catalytic activity (close to 100% in CO conversion).
This result was explained to be due to the formation of alkali ionassociated OH groups which are activated by CO at low temperatures
(even at 100 1C) in the presence of atomically dispersed Pt, based on
HAADF/STEM, XPS and CO-TPR experimental studies coupled with
density functional theory calculations. The authors43 commented that
their findings are very useful for the design of highly active and stable
WGS catalysts that contain traces of precious metal without the need for a
reducible metal oxide support (e.g. CeO2). The same group44 recently
reported on the same subject but using oxygen-free multi-walled carbon
nanotubes as a support of Pt. It was shown via HAADF-STEM that
atomically dispersed Pt species are stabilized by the presence of Na.
In situ XPS studies have demonstrated that a PtNaxOy(OH)z active-site
structure is formed, as previously reported on Pt/Na-SiO2,43 responsible
for this high LT-WGS catalytic activity. Other investigations in the
LT-WGS activity over alkali-promoted supported Pt catalysts (g-Al2O3,
TiO2 and ZrO2) have been recently reported,45,46 which also show
significant improvements in the activity compared to that obtained with
the non-promoted catalyst. The authors concluded that alkali modifies
the support properties and creates new active sites which accommodate
active reaction intermediates. However, according to in situ XANES
studies45 the relative amount of oxidised Pt did not track the rate promotion; the WGS reaction rate per mol of Pt increased with the fraction
of Pto (not oxidized platinum) on the samples that showed the presence
of PtO. The latter conclusion contradicts that reported by the group of
Flytzani-Stephanopoulos.43,44
The formation of new active Pt-&s-Ti31 sites by the addition of alkali
(Li, Na, K or Cs) and alkaline earth metals (Mg, Ca, Sr or Ba) onto the TiO2
surface was proposed as the main reason for the significantly enhanced
LT-WGS catalytic activity (3% CO/10% H2O/He) observed over the
corresponding supported Pt catalysts.47,48 The authors suggested that the
eects of alkali promotion on the chemisorptive (towards H2 and CO)
and catalytic properties of Pt/TiO2 might be described as a permanent
SMSI eect.47,48 The Davis group49,50 investigated the eect of Li, Na and
K promoter on ZrO2-supported Pt, where the enhanced LT-WGS activity
was correlated with the significant weakening of the formate CH bond
(in situ DRIFTS studies). Based on the view that formate decomposition
formed on support sites (by the aid of Pt) is the rate-limiting step in the
WGS mechanism, the authors stated that a new direction in watergas
shift catalyst design based on the weakening of formate CH bond could
be considered. However, the actual role of formate in WGS (kinetically
important intermediate or spectator species) is still a very controversial
issue as outlined in Sections 2 and 4.
178 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
The above-mentioned investigations on Pt and Au nanoparticles supported on reducible metal oxides (e.g. CeO2 and TiO2 or Ce1xTixO2d)
and alkali-promoted carriers have clearly indicated the important and
decisive role a reducible support or a chemical promoter can play in
enhancing the WGS activity for low-temperature applications. The need
for in-depth investigations of the actual mechanism(s) and detailed microkinetics (on the basis of elementary steps) of WGS over these catalytic
systems for activity optimisation naturally arises. This requirement points
to the use of in situ techniques and operando methodologies for obtaining
true correlations between the site reactivity and chemical composition and
structure of both the active catalytic sites and the accommodated adsorbed
reaction intermediates (active adsorbed phase).28,5167 Furthermore, the
application of appropriate in situ dynamic spectroscopic techniques and
kinetic rate measurements coupled with the use of stable isotopes under
steady state reaction conditions (e.g. Steady State Isotopic Transient
Kinetic Analysis, SSITKA) becomes necessary. The operando-SSITKA
technique allows discrimination between rival mechanisms and kinetic
parameter estimations (surface concentrations of active species and rate
constants of individual elementary reaction steps), as first proposed and
illustrated by Happel,68,69 Bennett,70,71 Biloen72 and Chuang,59,73 who were
inspired by the earlier seminal work of Tamaru and his coworkers.74
In the SSITKA technique, once the steady state of the overall reaction
rate is reached, an abrupt switch (step change) of the isotopic composition of the reaction mixture is made, when one of the reactants is
suddenly replaced by the same chemical species but with one of its atoms
replaced by one of its stable isotopes, while at the same time monitoring
(use of gas detectors, usually a mass spectrometer) the relaxation
and evolution of both non-labeled and labeled reactants and reaction
products, respectively, experimental details of which are given in
Section 3. In this SSITKA gas switch, the on-going surface catalytic reaction
is not disturbed, since the partial pressures of reactant components,
temperature and space velocity are kept constant.52,71 Therefore, the
overall reaction remains at a steady state, but a transfer of the isotope
element from the reactant molecular species through the various active
reaction intermediates to the various reaction products occurs, populating them with the isotope element. This process is properly termed
tracing of the reaction path from which the dynamics of gas and
adsorbed phases are developed. Based on the gas-phase transient isotopic
response curves recorded, usually by on line mass spectrometer with a fast
response (ca. 3050 ms) inlet system (SSITKA-Mass Spectrometry), various
intrinsic kinetic parameters can be determined as described in Sections 3
and 4, while at the same time the use of suitable spectroscopic techniques
(e.g. transmission FTIR, DRIFTS or Raman) would allow identifying the
chemical composition and structure of both the truly active reaction
intermediate and inactive (spectator) species (operando methodology, e.g.
SSITKA-DRIFTS-Mass Spectrometry).52,53 It has been illustrated3740,7578
that besides SSITKA, the combination of the use of 13CO and 18O2 in a
series of step-gas concentration switches is also a powerful methodology to
estimate important kinetic parameters (e.g. surface concentration) of the
Catalysis, 2016, 28, 175236 | 179
17:24:10.
View Online
WGS reaction under working conditions and to probe for both the active
and less kinetically important reaction intermediates.
The scope of this chapter is to show how the correct use of stable
isotopes in combination with in situ transient techniques and operando
methodologies provide important mechanistic and kinetic information
for the watergas shift reaction on supported Pt and Au working catalytic
surfaces. Through various examples the role of support in influencing
(i) the nature and concentration of active reaction intermediates found in
the carbon and hydrogen reaction paths of WGS, (ii) the nature of the
active catalytic sites (metalsupport interface vs. metal), (iii) the nature
of the inactive (spectator) species formed which can lead to catalyst
deactivation and (iv) the likely dierent mechanisms, is illustrated.
Watergas shift reaction mechanisms
Likely mechanisms of the Water Gas Shift reaction (1) on supported

metal catalysts have been studied for some decades mainly on the
basis of steady-state rate measurements and the application of LH
kinetics. Recently, more advanced micro-kinetic modelling combined
with DFT calculations1,2,7995 were considered in the investigation of
WGS reaction mechanisms. In recent years, various in situ spectroscopic
and operando isotopic techniques (e.g. XAFS, NAP-XPS, IRRAS, DRIFTS)
were used to obtain important information about the nature and
composition of surface catalytic sites and adsorbed active reaction
intermediates.2628,3740,7578,96109 This information has advanced significantly our basic understanding of possible reaction routes and intermediates of the WGS reaction. The importance of catalytic sites (structure
and intrinsic reactivity) at the metal support interface (supported Pt and
Au catalysts) has also been illustrated in recent years, based on both
experimental and theoretical evidence.25,41,9194,110113 In the following
Sections 3 and 4, the application of transient isotopic techniques and
results obtained in several cases for supported Pt and Au catalysts will be
presented and discussed in relation to these mechanisms.
2.1 Redox mechanisms
Pt and Au metal nanoparticles supported on reducible metal oxides
(e.g. CeO2 and doped-CeO2, La2O3 or TiO2) are considered bifunctional
WGS catalytic systems, where interfacial metal atoms and reducible
oxide support atoms at the three-phase boundary (TPB) of the metal
nanoparticle, oxide support and gas-phase, are the active catalytic
sites.25,87,94,111,114118 On the basis of this information, Heyden and his
group94 recently postulated reaction paths of possible CO- and H-assisted
WGS reaction steps at the TPB of Pt/CeO2(111), extending beyond the
classical redox pathway.
2.1.1 Classical redox pathway. The classical redox pathway
proposed earlier on several reducible metal oxides used as carriers of
Pt2,77,78,119,120 is presented in Scheme 1. The first step according to this
mechanism is the adsorption of CO on an empty interfacial Pt site
180 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
Scheme 1 Representation of the classical redox pathway at an interfacial site of a metal

nanoparticle supported on a reducible metal oxide (Adopted from Aranifard et al., On the
importance of metal-oxide interface sites for the water-gas shift reaction over Pt/CeO2
catalysts, J. Catal., 309. Copyright (2014) with permission from Elsevier).94
(PtOint), followed by a spill-over to the interfacial oxygen of support to

create an adsorbed CO2 structure (oxidation of CO). In a subsequent
step, CO2 desorbs from the surface creating an oxygen vacancy at the
interface (Vint). Adsorbed water on Pt (free site)-Vint dissociates and
transfers one H atom to Pt forming at the same time an interfacial OH
group. A second H-transfer from the OH to the Pt cluster takes
place, followed by the formation of di-hydrogen (with a concomitant
re-oxidation of the support surface shell) to close the catalytic cycle.94
The above-described model was not successful in reproducing all
experimental results, because the predicted apparent activation
energies were too high and the TOFs too low.
2.1.2 CO- and H-assisted redox pathways. Scheme 2 presents an
alternative redox mechanism (Redox (I)) that accounts for more
appropriate molecular configurations of the water dissociation step
(lower activation barriers) compared to that of the classical redox pathway (Scheme 1) and for the adsorption of a second CO molecule on a
Pt free site next to the interfacial OHint group. The reaction elementary
steps (R1)(R7) are similar to the classical redox pathway shown in
Scheme 1 but with some modification. After CO chemisorption (R1),
the H atom from the interfacial OH group has to be transferred to the
Pt cluster (R2) before the adsorbed CO can spill over to the interfacial
oxygen (R3). Desorption of CO2 leads to the formation of an interfacial
oxygen vacancy (R4), where H2O can then adsorb and dissociate (R5) by
transferring one H atom to the metal Pt cluster (R6). The final step
(R7) is the formation and desorption of H2(g).
Scheme 3 is a representation of an alternative redox mechanism
(Redox (II)) that accounts for the possibility of a second CO adsorption
on the vacant Pt site in the intermediate structure shown in step (R4),
before H2O adsorbs and dissociates.94 For both Redox (I) and (II) reaction
pathways, it was found94 that in the case of Pt/CeO2(111) the highest activation barriers are for the transfer of the H atom from an interface OH
group to the Pt metal cluster (step (R2)) and the CO2 desorption step (R4).
Catalysis, 2016, 28, 175236 | 181
17:24:10.
View Online
Scheme 2 Network of possible CO- and H-assisted WGS reaction elementary steps at the
three-phase boundary (TPB) of Pt/CeO2 (111) (Adopted from Aranifard et al.94 Copyright
2014 Elsevier).
Scheme 3 Network of possible CO- and H-assisted WGS reaction elementary steps at
the three-phase boundary (TPB) of Pt/CeO2 (111) (Adopted from Aranifard et al.94 Copyright
2014 Elsevier).
In the presence of additional surface H atoms, the energy barrier for

the interfacial H-transfer process is reduced compared to the pathway
depicted in Scheme 2.
182 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
2.1.3 Associative carboxyl with redox regeneration pathway. Computational methods applied to both Pt79,80,88,89 and Au85,121 single crystals,
as well as on model Au/CeO2(111),95 Pt/TiO2(110),91,93 Pt/CeO2(111)92 and
Au/CeOxTiOx87 catalytic surfaces, suggested the importance of the associative carboxyl (OCOH) and redox regeneration pathways. Scheme 4 is
a representation of the associative carboxyl with Redox (I) regeneration
pathway.94 In step (R12), adsorbed CO spills over to the interface OH
group forming a carboxyl (OCOH) intermediate. The next step (R13) is
the dissociation of the OCOH to the nearby surface OsH to form
OsHH (adsorbed water at the surface oxygen vacancy) and an adsorbed
CO2 species at the interface. Subsequently, the H in OsHH is transferred back to the free oxygen of the CO2 to form a stable HOCO(Pt-int)
species (step (R14)). Reaction steps (R13)(R14) describe the H-transfer
between the two oxygen atoms within the OCOH species, which leads to
the formation of a more stable intermediate and at the same time permits
the dissociation of OCOH at the metal cluster.94 The trans-OCOH
intermediate in step (R15) then dissociates leading to the formation of
HPtCOPtCO2(Pt-int)OsH, where the transition state (TS) of this step has
the highest energy.94 However, it is more stable than the H-transfer TS in
the redox pathway, thus the associative carboxyl with redox regeneration
pathway could be considered as the dominant WGS reaction pathway.
2.1.4 Associative formate (HCOO) with redox regeneration pathway.
The IR spectroscopic identification of the formate (HCOO) adsorbed
Scheme 4 Associative carboxyl with redox regeneration WGS reaction pathway at the
three-phase boundary (TPB) of Pt/CeO2 (111) (Adopted from Aranifard et al.94 Copyright
2014 Elsevier).
Catalysis, 2016, 28, 175236 | 183
17:24:10.
View Online
species in a large number of in situ IR (transmission or diuse reflectance mode) investigations in WGS over supported Pt catalysts has been
investigated theoretically to determine whether it could be considered
an active and kinetically important intermediate.79,80,88,89,94 It is, therefore, appropriate to consider the investigations performed on Pt
clusters supported on single crystal faces of metal oxides as adopted
also in the previous sections for other WGS reaction pathways.
Aranifard et al.94 have investigated the associative formate with redox
regeneration pathway on a Pt10 cluster supported on CeO2(111) surface.
The possibility of a formate (HCOO) intermediate at the interface
(steps (2)(3)) was considered:94
2COPtOHintOsH-COPtCHOPtOintOsH
(2)
COPtCHOPtOintOsH-COPt*PtHCOOintOsH
(3)
Steps (2) and (3) show that the formation of HCOO from adsorbed CO on
Pt and an interface OH group is a two-step process. In the first step (2),
CO abstracts H from the interface OH group forming a formyl (CHO)
intermediate, which then spills over to the interface oxygen leading to the
formation of formate species (step (3)). The formation of CHO requires
an activation barrier of about 2 eV and when this barrier is compared to
those of the redox pathway, it can be concluded that reaction step (2)
might possibly proceed by transferring the H atom directly to the metal
rather than to the adsorbed CO.94
2.1.5 Insights from microkinetic modelling. The calculated TOFs
for the CO- and H-assisted reaction pathways presented in Schemes 2
and 3 were found to be significantly higher than those for the classical
redox and CO-assisted only redox pathways. The computed rates
suggested that the associative carboxyl pathway with redox regeneration
is the dominant pathway when tested against experimental conditions.94 However, it was found that the rate of the Redox (II) pathway
is only one order of magnitude lower than the rate of the associative
carboxyl pathway. DFT computations clearly suggested that very high
TOFs are possible at interfacial Pt/CeO2(111) sites, which are 23 orders
of magnitude higher than the computed TOFs on Pt (111) and on
stepped Pt (211) and Pt (322) surfaces.94 For the first time, the latter result reports that the WGS in the 200300 1C range is indeed significantly faster at the metaloxide interface. In fact, the authors have pointed
out that virtually all catalytic activity of Pt/CeO2(111) originated from
metaloxide interface sites. This result finds support from an earlier
operando-SSITKA experimental investigation of WGS in the 250300 1C
range on a series of Pt/CeO2 catalysts of varying mean platinum particle
size,78 where a strong linear correlation of the specific WGS kinetic rate
per length of the Ptceria interface (mmol CO cm1 s1) with the Pt
particle size was reported.
The improved activity at the metaloxide interface sites was explained
by the significantly enhanced water activation and dissociation at the
interface oxygen vacancies.94 At the same time, the stability of these
184 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
oxygen vacancy structures can increase the activation barrier and reduce
the low-T WGS activity on the basis of DFT computations on
Pt/CeO2(111)94 and Pt/TiO2(110)91 catalytic surfaces. Aranifard et al.94
have also suggested some design criteria for improved WGS catalysts at
low-temperatures. First, reducing the CO adsorption strength on the
interfacial Pt atoms could increase the number of active sites; this could
be achieved by bimetallic clusters (e.g. AuPt). Second, on the support
side, destabilizing the vacancy structure in ceria could decrease the
activation barrier and increase the low-T WGS rate; using a mixed oxide
support or an appropriate doping strategy could achieve this goal.
2.2 Classical associative mechanisms (formate, carboxyl and
carbonate as important intermediates)
R. Burch24 in his review paper on Au catalysts for the WGS reaction has
introduced the concept of a universal mechanism, where apparent discrepancies appearing in the open literature between carboxyl, formate
and carbonate as kinetically important intermediates can be rationalised.
The concept of this universal mechanism is that the importance of the
kinetically relevant carbon-containing reaction intermediate will be a
function of the experimental conditions,24 including the choice of support and the active metal. Alternatively, the nature of the active catalyst
surface and that of active catalytic site(s) will be a function of reaction
conditions. In fact, as pointed out by Ratnasamy and Wagner,2 these
features regarding the active catalytic sites were part of the landmark
postulate of Hugh Taylor in 1926,122 in which (i) particular atoms or
group of atoms on the surface of solids are the active sites responsible for
the catalytic activity and selectivity, (ii) the nature and concentration of
the active sites on a catalyst are dependent not only on the procedures
adopted during the preparation, but also on the particular reaction
conditions.2 As pointed out by R. Burch,24 at high temperatures the
decomposition of formate and carbonate species is expected to be fast in
the presence of a high concentration of water, as proved also experimentally.40,78,101,123125 At these conditions, the support metal oxide
surface will be in an oxidised state, and one would anticipate that the
redox mechanism would be dominant. At lower temperatures, where
formate, carbonate or carboxyl species have a significant lifetime on the
surface, then the rate-determining step will involve the formation or
decomposition of one or other of these species.24 The importance of
reaction conditions (e.g. H2O/CO2 feed ratio) encountered in practical
WGS (composition of a reformate gas) in how they would aect the
nature of the kinetically important reaction intermediate has also been
illustrated.24 The concentration of OH groups will be highest when there
is a high concentration of water in the feed gas stream and low concentration of carbon dioxide. This is because it would be expected that
water would dissociatively adsorb onto the support surface to give OH
but that carbon dioxide would compete for surface sites and could block
water adsorption sites by forming relatively stable surface carbonate
species. Since water facilitates the decomposition of carbonates, a high
H2O to CO2 ratio would favour surface OH over surface CO3 groups,
Catalysis, 2016, 28, 175236 | 185
17:24:10.
View Online
thus favouring formate formation. On the other hand, under conditions

where the H2O/CO2 gas ratio is low, the reverse would be expected and
the formation of formates would be inhibited in competition with the
formation of carbonates.24 Furthermore, if the reaction steps which give
CO2 from the decomposition of formate or carbonate have dierent
activation energy barriers, then a changeover in the dominant mechanism (formate vs carbonate associative mechanism) could occur as the
WGS temperature varies.24
Scheme 5a and b present the carbonate/carboxylate and universal
mechanisms, respectively, proposed by R. Burch,24 where formate
(HCOO) was suggested to be favourably produced through the addition
of an H atom to an adsorbed CO on a metal particle. The shaded box
identifies a carbonate-type species through which the associative formate
mechanism passes. The same species was identified in the associative
carboxyl mechanism.24 Thus, if the carbonate-type decomposition step to
release CO2 becomes slow, then this reaction path is named an associative
carbonate mechanism. On the other hand, if the rate of formate decomposition by water and the involvement of metal particle becomes
Scheme 5 Proposed carbonate/carboxylate (a) and universal (b) mechanisms for the
watergas shift reaction (Adopted from Burch24 Copyright 2006 Royal Society of
Chemistry).
186 | Catalysis, 2016, 28, 175236
View Online
rate-limiting, then this reaction path is named an associative formate

mechanism.
17:24:10.
SSITKAoperando methodology
In this section some basic and important theoretical and experimental

background on the application of the SSITKA-operando methodology
using both mass spectrometry and infrared spectroscopy (transmission
or diuse reflectance mode) to estimate important kinetic parameters
and to investigate mechanistic issues of the WGS reaction as outlined in
the Introduction section is provided.
3.1 Steady state isotopic transient kinetic analysis (SSITKA)
Figure 1 presents an illustrative scheme of the concept of SSITKA
experiment applied in studying a reversible gassolid heterogeneous
catalytic reaction, such as the watergas shift:
R (g) other reactants (g)2A intermediates (ads)2P (g) other products (g)
(4)
This scheme was adopted from a review on SSITKA by Shannon and
Goodwin57 and depicts the compartmental model that best describes
the SSITKA experiment.51,52 In this model the catalyst surface is
considered to be composed of a system of interconnected pools or
compartments,68,69 where each pool represents a homogeneous subsystem within the reaction pathway, which reflects a unique adsorbed
active reaction intermediate only, having the same site reactivity, kn. It
is also considered that there is no mixing time associated with each
pool except for the mean residence time of the given reaction intermediate, tn (s).
After steady state is achieved (Fig. 1a), it is expected that the rate
of reaction, rR (e.g. mol R gcat1 s1) is the result of the net rate of
pool transfer (sum of rates of input sum of rates of output) of the
Fig. 1 Catalyst surface compartmental model showing the isotopic distribution between
n pools in series during SSITKA R - *R (n is the number of dierent active reaction
intermediates) (Adopted from S. L. Shannon, J. G. Goodwin Jr., Chem. Rev., 1995, 95(3),
677. Copyright (1995) American Chemical Society).57
Catalysis, 2016, 28, 175236 | 187
17:24:10.
View Online
non-isotopic species of interest from reactant R. Furthermore, a constant

abundance, Ni (mol) (i 1,2,. . .n), surface concentration (mol m2 or
mol gcat1) or surface coverage (y) of intermediate species A in the i-th
pool (chemical nature of A is dierent in each i-th pool) is established. At
time t 0, an instantaneous fast switch of the reaction feed gas stream
containing the reactant R to an equivalent one having the same chemical
composition and flow-rate and which contains the isotopic labeled
element in R*, is made via a 4-way chromatographic valve of zero dead
volume.51,52
At t Z0, a non-ideal step change51 in the isotopic composition of R* is
made at the inlet of the catalyst bed and as a result, the new isotope
element is distributed only between the various active adsorbed reaction
intermediates which are found in the reaction path to form eventually
product P. Therefore, as the reactant R* passes through the catalyst bed
volume and reacts on the catalyst surface, the new isotope element in R*
is transferred via all the reaction intermediates A found in the reaction
path from R to P to form eventually the new isotopic labelled P* (Fig. 1b).
Ultimately, the entire old isotope element is displaced. The distribution
of the isotope element in each reaction intermediate (pool i-th) is strictly
dependent on the intrinsic kinetic rates ri between the pools and
the intrinsic reactivity, ki, associated with each adsorbed reaction
intermediate A. The transient isotopic rate of products P and P* is decreasing and increasing, respectively, during the SSITKA experiment,
where after some time a new steady-state is achieved (rp rp*). In the
absence of any mass isotopic eects during SSITKA, the following
relationship holds:
rps.s. rp(t) rp*(t)
(5)
where rps.s. is the overall steady-state reaction rate (mol s1) determined
under the non-isotopic feed gas composition reaction conditions, rp(t) is
the transient rate of product P at a given time, t, and rp*(t) is the transient
rate of product P* (containing the isotope element) at the same time.
Based on a material balance applied at steady-state reaction conditions
for an open flow-reactor, the following relationship applies:
rps.s. FT yps.s.
(6)
where, FT (mol s1) and yps.s. are the total molar flow rate of reaction gas
stream and the mole fraction of product P, respectively, measured at the
exit of the reactor at steady-state reaction conditions. A normalized
transient response, Zp(t) can be defined:
Zp(t) yp(t)/yps.s.
(7)
where, at t 0 (isotopic switch), Zp(t) 1.0, while at t-t*s.s. (new steady

state under the isotopic feed gas stream), Zp(t-t*s.s.) 0.0.
In the absence of any perturbation in the reaction system (e.g. flow rate,
pressure or T) and of any mass isotope eects, the functions Zp(t) and
Zp*(t) are mirror images of each other and they should cross each other
188 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
Fig. 2 Typical SSITKA switch R-R* for measuring the total number of adsorbed
reaction intermediates found in the reaction path: R (reactant)-A (adsorbed reaction
intermediates)-P (product). I: inert tracer (e.g. Ar, Kr or He), R*: isotope labeled reactant
R, P*: isotope labeled product P.
at Z 0.5 as illustrated in Fig. 2. The ZI(t) function reflects the gasphase hold-up from the switching valve through the reactor to the
detector (e.g. mass spectrometer) when the SSITKA gas switch R/other
reactants/Ar/He-R*/other reactants/He is made; more details on the
ZI(t) function were previously reported.51,57
Based on the experimental response curves of ZI(t) and Zp(t) and a
material balance, the concentration (mol gcat1) of all the active adsorbed
reaction intermediates, Np N1 N2 . . . Nn (Fig. 1) leading to the
formation of product P can be estimated according to the following
relationship:
t*S:S:

1
P
Np mol gcat FT =W yS:S:
Zp t ZAr t dt
(8)
0
where, y s.s. is the steady-state concentration (mole fraction) of product P

measured experimentally under the non-isotopic reaction feed stream
and W is the amount of catalyst used. Eqn (8) is applicable without the
need of knowing the true micro-kinetics and mechanism for the catalytic
reaction system under investigation. This is the power of SSITKA technique, which permits the in situ determination of the concentration of
the active reaction intermediates only, excluding inactive (spectator)
species. However, readsorption eects of the gaseous product(s) on nonactive catalytic sites do not allow the use of eqn (8), and this limitation
applies in the present case of the WGS reaction (e.g. CO2 readsorption to
form carbonate-type surface species). Furthermore, various H/D exchange
reactions that might occur during WGS do not allow the correct
Catalysis, 2016, 28, 175236 | 189
View Online
17:24:10.
estimation of the concentration of the active intermediates found in

the H-path of reaction via eqn (8). These two limitations, along with
alternative experimental methodologies proposed to overcome these
problems are described in detail in Section 3.3.
3.2 Steady state isotopic transient kinetic analysis (SSITKA) DRIFTS
methodology
Infrared (IR) absorption spectroscopy was one of the first techniques
used for the characterization of catalysts, and it is still one of the most
popular.126 The popularity of IR as a powerful tool for the characterization of solid catalysts and the adsorbed phase under reaction conditions (in situ IR) is that a number of setups are available to adopt the
experiments to the nature of the sample to be investigated.126 The
diuse-reflectance (DRIFTS) mode has become a very popular method for
in situ IR studies in spite of some limitations,126 especially when the solid
is opaque to IR radiation and not amenable to manipulation to produce
self-standing samples as required for the transmission IR mode. Band
intensities can also be several times more intense in DRIFTS than in
transmission mode, and enhanced sensitivity is typical for signals from
surface adsorbed species compared to gas-phase.126 The quantification of
signal intensities in DRIFTS to extract reliable kinetic parameters,
including surface coverages of spectroscopically observed adsorbates,
and some problems faced in the analysis of DRIFTS spectra have been
recently reviewed.126130
The combined transmission FTIR and SSITKA techniques first
introduced by Chuang and co-workers73,131 for the investigation of
methanation, ethylene hydroformylation and dry reforming of methane
reactions should be appreciated. The same methodology but extended to
SSITKA-DRIFTS and the design of kinetically-appropriate DRIFTS cells
has been discussed.61,102,108,126,132,133 Several cases of SSITKA-DRIFTS
and SSITKA-Mass spectrometry applications towards WGS will be presented and discussed in Section 4. At this point, some important aspects
of the analysis and interpretation of DRIFTS spectra recorded under the
SSITKA gas switch (Section 3.1) related to the investigation of WGS
reaction mechanisms are presented.
3.2.1 Interpretation of the red isotopic shift in IR bands at the
CO/H2O-13CO/H2O SSITKA switch. In a typical SSITKA-DRIFTS
experiment, where the diuse reflectance cell is the micro-reactor, the
chemical structure of the true active reaction intermediates found
in the carbon-path of WGS will be traced correctly by DRIFTS only if
(i) the surface concentrations are within the detection limit, (ii) appropriate procedures for background subtraction and baseline correction
are followed,129 (iii) the theoretically expected red isotopic shift of the
13
C-labelled carbon-containing intermediates134 is larger than the resolution of the FTIR, (iv) the red isotopic shift is not due to any other
surface chemical reaction step(s) but only to those associated with the
carbon-path of WGS, and (v) appropriate deconvolution procedures are
applied134 for overlapping IR bands which might arise from active and
12
190 | Catalysis, 2016, 28, 175236
View Online
17:24:10.
inactive (spectator) intermediate species. The same criteria should

apply in the case of SSITKA gas switch CO/H2O-CO/D2O made for
tracing the H-path of the WGS reaction.
3.2.1.1 The case of adsorbed CO-s species. Figure 3 presents SSITKADRIFTS spectra in the 22501850 cm1 range recorded after 30 min in
WGS (steady-state) at 250 1C in 12CO/H2O (upper spectrum, solid line),
and after 30 min of the gas switch to 13CO/H2O (lower spectrum,
dashed line) over the Pt/Ce0.8La0.2O2d catalytic system.38 The DRIFTS
spectrum recorded under 12CO/H2O is typical of n(CO) stretching vibrational modes of carbon monoxide in both the gaseous and adsorbed
states (various carbonyl species) on Pt surfaces.135137 Deconvolution and curve-fitting (Gaussian band shape) procedures138 are then
necessary to establish the position of individual bands due to various
adsorbed states of CO. The IR bands labelled as 1 and 2, respectively,
centred at 2182 and 2116 cm1, are due to gas phase 12CO.135 The
low-intensity IR band centred at 2069 cm1 (band 3) is related to a
high-frequency (HF) linearly bound CO on Pt, whereas the infrared
band at 2046 cm1 (band 4) is assigned to a low-frequency (LF) linearly
bound CO on Pt atoms present in small platinum particles of low
coordination (dPto15 ).136,137 The IR band 5 recorded at 1995 cm1 is
attributed to CO adsorbed on Pt sites of low coordination in very small
Fig. 3 SSITKA-DRIFTS spectra recorded in the 22501850 cm1 range over 0.5 wt%
Pt/Ce0.8La0.2O2d at 250 1C. Solid-line spectra were recorded under 3 vol% 12CO/10 vol%
H2O/Ar/He and dashed-line spectra under 3 vol% 13CO/10 vol% H2O/Ar feed gas
mixtures. Deconvolution and curve fitting procedures are also shown (Adopted from
Kalamaras et al., The eect of La31-doping of CeO2 support on the water-gas shift
reaction mechanism and kinetics over Pt/Ce1xLaxO2d, Appl. Catal., B, 136137, 225
238. Copyright (2013) with permission from Elsevier).38
Catalysis, 2016, 28, 175236 | 191
View Online
17:24:10.
136,137
Pt particles.
Further discussion on the positions of these
adsorbed CO IR bands on the present supported Pt surface can be
found elsewhere.38
As depicted in Fig. 3 (bottom spectrum), all IR bands described
above that are due to molecularly adsorbed CO show a red isotopic shift
upon the switch 12CO/H2O-13CO/H2O (bands labelled with * after
spectrum deconvolution and curve-fitting) consistent with theory
(v(12C16O)/v(13C16O) 1.023).134 During SSITKA the surface coverage of
a given active 13C-containing reaction intermediate remains the same as
the initial corresponding 12C-containing one; due to the short time
treatment of the catalyst surface with 13CO/H2O, it is likely that the same
is true for the inactive species. This is important since no surface
coverage eects need to be considered in the deconvolution of the
13
C-labelled adsorbed CO DRIFT spectra.
The observed red isotopic shift for all three populated states of linear
CO cannot be safely used to argue that these three kinds of adsorbed CO
take part in the carbon-path of the WGS towards the formation of CO2.
The red isotopic shift observed could be simply due to the process of
exchange of a reversibly adsorbed 12CO-s with 13CO(g). On the other
hand, since the only C-containing reactant species in WGS is the CO,
then at least one kind of adsorbed CO-s must be considered as an active
intermediate. Therefore, SSITKA-DRIFTS alone cannot convincingly
provide an answer which of these three adsorbed CO-s species (Fig. 3)
should be regarded as active reaction intermediates. In the following
Section 3.3.1, this point is addressed, where an alternative experimental
methodology has been recently proposed to determine the kind(s) of
active CO-s present during WGS.38,78
3.2.1.2 The case of formate (HCOO-s) species. Figure 4 presents
characteristic nCH stretching mode IR bands recorded at 2940, 2890
and 2843 cm1 (after deconvolution and curve fitting) which are due to
adsorbed formate species (HCOO)139142 formed during WGS at 250 1C
on Pt/Ce0.8La0.2O2d (Fig. 3). A recent work139 has combined DFT and
experimental IR investigations targeting the reassignment of the complex IR spectra of surface carbonates, hydrogen-carbonates and formate
species in various coordination modes to the ceria surface. Regarding
the formate (HCOO) species, it has been reconfirmed that an IR band
in the 28972945 cm1 range is associated with the nCH stretching
mode. The observation of an infrared band at lower wavenumbers (ca.
2850 cm1) is assigned convincingly to combination modes (e.g. nOCO
and dCH), in agreement with earlier reports.140142 The combination
mode could have a high absorption cross section and due to a Fermi
resonance could lead to similar or higher intensities than the fundamental excitations.143
Based on the previous paragraph and the expected theoretical isotopic
shift for nCH (1.003 isotopic shift factor),134 the observed IR band at
2843 cm1 (Fig. 4, upper graph) should reflect the combination band
dCH nOCOsym of the same formate species responsible for the nCH
band observed at 2940 cm1. The small IR band recorded at 2890 cm1
192 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
H2O/Ar/He and dashed-line spectra under 3 vol% 13CO/10 vol% H2O/Ar feed gas mixtures.
Deconvolution and curve fitting procedures are also shown (Adopted from Kalamaras
et al., The eect of La31-doping of CeO2 support on the water-gas shift reaction
mechanism and kinetics over Pt/Ce1xLaxO2d, Appl. Catal., B, 136137, 225238. Copyright (2013) with permission from Elsevier).38
Scheme 6 Types of formate (HCOO) species (I, II and III) chemisorbed on metal oxides
(MOx) and carboxyl species (IV) on metals (e.g. Pt, Au).
(Fig. 4, band 2) might be assigned to a second type of formate species

according to Scheme 6. However, the range of the nCH, dCH nOCOasym
and dCH nOCOsym vibrational modes overlap,140 and a clear-cut discrimination among them is dicult from experimental data alone. In the
present case of Pt/Ce0.8La0.2O2 (Fig. 4), Kalamaras et al.38 assigned the
2890 cm1 IR band to the dCH nOCOasym mode, therefore only one type
of formate species was claimed to have been formed during WGS at
250 1C.
The site location of the HCOO species, whether on the Ce0.8La0.2O2d
support, the Pt or at the metalsupport interface, is very dicult to
resolve based only on the recorded experimental IR band positions.118,123,135,144146 A strong indication that active HCOO resides on
the support can only be provided after estimating its surface concentration. For example, if yHCOO (based on Pts) is estimated to be greater
Catalysis, 2016, 28, 175236 | 193
17:24:10.
View Online
than unity, then part of this active species is definitely formed on the
support surface. In the following Section 3.3, it is demonstrated how
other than SSITKA transient isotopic experiments combined with DRIFTS
allow probing whether active HCOO is formed on support or Pt-support
interface sites.
It is shown in Fig. 4 (lower graph) that all three IR bands due to HCOO
gave the red isotopic shift. This is an indication that the latter species
could be an active intermediate in the carbon-path of WGS. However, as
discussed in the case of adsorbed CO-s (Section 3.2.1.1), it is likely that
gaseous CO may interact reversibly with OH groups to form the formate
(HCOO) species. Calatayud et al.140 on the basis of DFT computations
and experimental infrared spectroscopic measurements have found that
insertion of CO with surface OH groups leads to the formation of formate
species over metal oxide surfaces. In the case of Pt/Ce0.8La0.2O2 (Figs. 3
and 4), after the following sequence of step-gas switches 3% 12CO/10%
H2O/Ar (30 min)-3% 12CO/Ar (10 min)-3% 13CO/Ar (10 min) at 250 1C
was performed,38 the HCOO infrared bands shown in Fig. 4 (upper
graph) gave red isotopic shifts after 10 min treatment in 13CO/Ar.
This result strongly supports the view that for the given catalytic system
and reaction conditions, formation of HCOO must pass through the
following reversible steps (9)(11):
CO(g) OH-s2HCOO-s
CO(g) s2CO-s
CO-s OH-s2HCOO-s s
(s: surface support site)

(s: surface Pt site)
(9)
(10)
(s: surface support and Pt sites) (11)
Therefore, unambiguous confirmation that HCOO is a true active

reaction intermediate of the WGS at 250 1C on Pt/Ce0.8La0.2O2 cannot be
provided by the SSITKA-DRIFTS technique alone. This issue is further
discussed in Section 3.3.1.
3.2.1.3 The case of carboxyl (s-COOH) and carbonate-type species.
Figure 5 shows DRIFTS spectra in the 17001150 cm1 region recorded
under WGS-SSITKA at 250 1C (upper spectrum: 12CO/H2O; bottom
spectrum: 13CO/H2O) over the Pt/Ce0.8La0.2O2d catalyst (Figs. 3 and 4).38
A detailed band assignment for this specific infrared region regarding
potential WGS reaction intermediates (active or inactive) on fully oxidised and reduced ceria surfaces, on the basis of a model Ce21O42 nanoparticle accompanied by DFT computations and experimental data, has
been reported and discussed with respect to previous assignments.139
Scheme 7 depicts the structures of the various surface carbonate-related
species formed on ceria,135,141143,147 structural characterisation details
of which were critically discussed by Vayssilov et al.139
The IR band at 1580 (band 1) which is shifted to 1548 cm1 (band 1*)
after the SSITKA switch (Fig. 5) is due to the nOCOas mode and it cannot
be clearly assigned to a single species among those described in
Schemes 6 and 7. Formates (Scheme 6) and carbonates (Scheme 7)
194 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
H2O/Ar/He and dashed-line spectra under 3 vol% 13CO/10 vol% H2O/Ar feed gas mixtures.
Deconvolution and curve fitting procedures are also shown (Adopted from Kalamaras
et al., The eect of La31-doping of CeO2 support on the water-gas shift reaction
mechanism and kinetics over Pt/Ce1xLaxO2d, Appl. Catal., B, 136137, 225238. Copyright (2013) with permission from Elsevier).38
Scheme 7 Structures of surface carbonates and related species formed on ceria

surfaces.
formed on ceria show this type of vibrational mode in the 15701530

and 17001490 cm1 regions, respectively.139 Furthermore, hydrogen
carbonates of certain structural characteristics show a nOCOas mode in
the 15881632 cm1 region.139
The 13C isotopic shift (D13C) for this vibrational mode was reported
to be in the 3845 cm1 range for formate-, carbonate- and hydrogen
carbonate-type species, in good agreement with the results reported in
Catalysis, 2016, 28, 175236 | 195
17:24:10.
View Online
Fig. 5 for the Ce0.8La0.2O2 surface. Since at least one type of formate
species was identified on the same surface as reported in Fig. 4, then
if part of the band 1 integral intensity is also due to a carbonate typespecies, it is rather clear (on the basis of the deconvolution/curve fitting
applied) that the carbonate-type species providing the same red isotopic
shift might also be considered an active intermediate in WGS (Section
2.2). However, as will be illustrated in Section 3.3.1, reversible chemisorption of CO2 (eqn (12)) is a likely step to occur on the surface of metal
oxides of basic character at WGS reaction temperatures:
CO2(g) O-s2CO3-s
(s: surface oxygen support site)
(12)
13
It becomes, therefore, clear that upon the CO/H2O SSITKA switch the
CO2 will start replacing 12CO3-s on inactive catalytic sites, thus
providing the red 13C isotopic shift. The 1353 cm1 band 4 formed under
the 12CO/H2O gas mixture (Fig. 5, upper spectrum) can be assigned to the
nOCOs vibrational mode of formates (Scheme 6), carbonates and hydrogen carbonates (Scheme 7), while the IR bands centered at 1515 (band 2)
and 1468 cm1 (band 3) to carbonates (Scheme 7, types IIII) and polydentate carbonates.135,139 All these infrared bands show a red isotopic
shift (Fig. 5, lower spectrum).
Several DFT computations related to the mechanism of WGS on Pt
(111) and Pt/CeO2 (111) surfaces41,79,80,85,88,89,91,94 conclude that the WGS
mechanism passes through a carboxyl (COOH) rather than a formate
(HCOO) active reaction intermediate. It was reported146 that COOH on Pt
(Scheme 6, IV) exhibits three infrared bands, namely: 1732 (vas(OCO),
1264 (d(CH)) and 1104 cm1 (vs(OCO)). Some stable carbonate structures
on ceria resulting in a nOCOas vibrational mode in the 17001755 cm1
range were also reported,139 where the lower frequency infrared bands of
PtCOOH are also characteristic of carbonate-type species (Scheme 7).
The SSITKA-DRIFTS cannot unambiguously identify the carboxyl species
as an active intermediate because: (i) its concentration can be too low
(a very active intermediate), or (ii) its main vibrational modes overlap
with those of other carbonate-type adsorbed species having similar 13C
isotopic shifts.139 The experimental results in Figs. 35 support these
conclusions.
13
3.3 Application of SSITKA-mass spectrometry in WGS: inherent

problems
The correct estimation of the total concentration of active reaction
intermediates, Np (eqn (8)) through the SSITKA experiment requires
a priori that the final reaction product P is not able to readsorb on catalytic
sites that are not involved in the catalytic cycle under steady-state
reaction conditions. When this important criterion is not satisfied, then
the Np value is overestimated. This phenomenon was first reported by
Goodwin and his co-workers57 who suggested experimental methodologies to correct for interparticle and intraparticle product readsorption.
In the former case, decreasing the bed length or increasing the weight
hourly space velocity (WHSV) tends to minimize interparticle
196 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
readsorption in the catalyst bed. A series of experiments at various flow

rates (WHSVs) or catalyst loadings (bed lengths, Lb) can provide Np.
A plot of calculated Np against Lb or 1/WHSV, with extrapolation to zero Lb
provides the amount of active intermediates leading to a given product,
free of external readsorption eects.57 The latter methodology does not,
however, account for intraparticle readsorption. It was suggested57 that
the latter can be accounted for using unlabelled product in the labelled
feed stream in order to allow competition for readsorption sites with the
labelled product formed during the SSITKA switch. If readsorption of
product occurs at nonreactive sites, increasing the concentration of unlabelled product (Cpin) at the reactor inlet will decrease the measured Np
obtained from the labelled product response curve, ultimately reaching a
limiting value, which is equivalent to the true surface concentration of
intermediates (Np) leading to the given product.57 However, concerns
remain when product readsorption occurs at active catalytic sites. In this
case, the external co-feed of unlabelled product during the SSITKA switch
might change the rates of elementary steps, thus the surface coverages of
reaction intermediates according to the pool series depicted in Fig. 1.
In the present SSITKA-WGS, the correct estimation of the total concentration of active C-containing (C-pool) reaction intermediates (NCO2,
mmol g1) is limited by CO2(g) readsorption on inactive and active catalytic sites. This eect becomes large when WGS is operated using metal
oxides of basic character as supports of active metals. When SSITKA is
used to estimate the total concentration of active H-containing (H-pool)
reaction intermediates found in the reaction path(s) from H2O to the H2
gaseous product, the accurate estimation of it is limited by H/D side
exchange reactions between reactant H2O and H2 product and the various
adsorbed species. These reactions are very likely to occur, limiting
application of SSITKA in accurately estimating the concentrations of the
C-pool and H-pool; the issue has been recently addressed by Efstathiou
and his group,78,97 as described in the next two sections.
3.3.1 Overestimation of SSITKA C-pool: CO2 readsorption eect.
Efstathiou and his group3740,77,78,97 were the first to report on the development of an operando methodology to estimate the active C-pool
for a Pt/CeO2 catalyst free of CO2 readsorption eects. This methodology can in principle be applied to any other catalytic system. The
experiment was as follows. After a steady-state WGS kinetic rate is
reached with the 12CO/H2O/Ar feed gas stream, a step gas switch to a
12
CO/Ar gas mixture (same 12CO concentration as in 12CO/H2O/Ar) is
performed for 3 min, quenching the reaction rate but keeping the surface concentration of CO similar to that established under steady-state
reaction conditions (based on the CO IR band recorded under
12
CO/H2O/Ar and 12CO/Ar). The feed is then switched to an equivalent
13
CO/Ar gas mixture, replacing any reversibly adsorbed 12CO-s with
13
CO-s and labeling any other adsorbed reaction intermediates able to
interact directly with gaseous 13CO and/or adsorbed 13CO-s with 13C.
For example, considering that HCOO-s (formate) is formed during WGS
on a given catalyst surface, this species will then appear as H13COO-s
Catalysis, 2016, 28, 175236 | 197
View Online
17:24:10.
13
(under the 3 min treatment in CO/Ar) if there is a chemical step

associated with the interaction of this species with gaseous or adsorbed
13
CO-s. At the end of the 13CO/Ar gas treatment, the feed is switched to
a H2O/Ar gas mixture (same composition in water as in WGS) and the
transient evolution of 12CO2, 13CO2 and H2 gases is followed by mass
spectrometry. At the same time, the transient depletion or formation of
adsorbed species under the H2O/Ar gas is followed by DRIFTS. Details
of the experimental set-up used to perform this operando isotopic
experiment are given elsewhere.51,52,75,78,148
The delivery sequence of step-gas switches for the above described
experiment and for the gas composition used was as follows:
3 vol% 12CO/10 vol% H2O/Ar (T, steady-state)-3 vol% 12CO/Ar (T, 3 min)
-3 vol% 13CO/Ar (T, 3 min)-10 vol% H2O/Ar (T, t)
The importance of introducing the 12CO/Ar switch before the 13CO/Ar one
is because during the 12CO/H2O/Ar (T, steady-state)-13CO/Ar gas switch
it will take time to remove water from the gas-phase of reactor, where
otherwise it would react with 13CO to form labeled 13C-containing reaction intermediates.
Figure 6 shows the transient isotopic response curves of 12CO2, 13CO2
and H2 obtained during the 10 vol% H2O/Ar gas switch in the operando
DRIFTS-MS experiment described above, implemented on a 0.5 wt%
Pt/Ce0.8La0.2O2d catalyst.38 Most CO2 formed is 13CO2 (54.7 mmol g1, or
Fig. 6 Transient isothermal response curves of H2, 12CO2 and 13CO2 obtained on the
0.5 wt% Pt/Ce0.8La0.2O2d catalyst according to the following gas delivery sequence
at 250 1C: 3 vol% 12CO/10 vol% H2O/Ar (30 min)-3 vol% 12CO/Ar (10 min)-3 vol%
13
CO/Ar (10 min)-10 vol% H2O/Ar (t) (Adopted from Kalamaras et al.38 Copyright 2013
Elsevier).
198 | Catalysis, 2016, 28, 175236
View Online
17:24:10.
1
y 2.1, based on the value of 26.0 mmol Pts g ), with only a small
amount of 12CO2 (5.4 mmol g1, or y 0.2). In the case of hydrogen
response, this corresponds to 50 mmol g1. As shown next, this amount
reflects (i) the contribution of HCOO (formate) on the support and any
PtCOOH (carboxyl) species found in the C-pool which reacted during
H2O/Ar (2 mols of COOH or HCOO corresponds to 1 mol H2), and (ii)
the contribution of adsorbed CO (1 mol CO corresponds to 1 mol H2).
Therefore, the transient 13CO2 response curve depicted in Fig. 6 is due to
the reaction with water of these adsorbed species. The fact that a single
symmetrical 13CO2 peak is observed (Fig. 6) may reasonably suggest that
the kinetic rate of reaction of adsorbed CO and formate/carboxyl with
water are similar. The small 12CO2 response obtained with its peak
maximum shifting to higher reaction times compared to 13CO2 is due to
the presence of a small concentration of formates/carbonate-type species
that react slowly with water, as will be shown next. The latter species are
formed by the irreversible interaction of CO(g) with OH groups (no
exchange with 13CO). This small concentration cannot be seen in the
DRIFT spectrum reported in Fig. 4.
Figure 7 reports DRIFT spectra recorded in the 30001150 cm1 range
during the 10 vol% H2O/Ar gas-flow (corresponding MS responses in
Fig. 6). The characteristic region of the vCH stretching mode due to
formate (HCOO) is shown in Fig. 7a,38 where the spectrum marked with
t 0 s corresponds to that recorded just after the 10 min 13CO/Ar gas
treatment and for which the red isotopic shift of the formate IR bands
was noted. The latter result strongly indicates that the elementary step(s)
of formate that involve CO(g) is/are reversible, as indicated by eqn (9) and
(11). Practically all adsorbed formate was depleted within the first 5 min
in H2O/Ar gas stream (Fig. 7a). It is important to note that in another
experiment, following the gas switch CO/H2O (30 min, 250 1C)-Ar
(250 1C, t), the authors38 observed only a small decrease in the intensity of
formates IR bands after 10 min in Ar gas-flow. This result demonstrates
that practically the rate of decomposition of formate in Ar is very small at
250 1C. Therefore, the results shown in Fig. 7a strongly suggest that
depletion of formate surface concentration is largely due to the presence
of water. This depletion of formate concentration leads to CO2 and H2
(Fig. 6). The authors38 suggested that under H2O/Ar gas treatment the
formation of OH/H species on Pt and/or Pt-support interface, after the
removal by reaction with water of adsorbed CO on Pt, enhances the rate
of formate decomposition into CO2 and H2, in agreement also with
previous works.78,101,124,146,149 Figure 7b presents corresponding DRIFT
spectra recorded in the 11501700 cm1 range, where the assignment of
infrared bands has been previously discussed (Fig. 5). It was found that
after 10 min in H2O/Ar gas-flow the intensity of the IR band centered at
B1540 cm1 clearly changed, whereas those corresponding to lower
wavenumbers change only slightly. This result suggests that part of these
species react only slowly under the water reactive atmosphere. Given the
fact that all the surface concentration of formate (HCOO) species was
consumed after 10 min in the H2O/Ar gas stream (Fig. 7a), and that
carboxyl intermediates (COOH) that potentially could be formed on Pt
Catalysis, 2016, 28, 175236 | 199
17:24:10.
View Online
Fig. 7 DRIFT spectra recorded over the 0.5 wt% Pt/Ce0.8La0.2O2d catalyst at 250 1C under
the 10 vol% H2O/Ar (t) switch of the following gas delivery sequence: 3 vol% 12CO/10 vol%
H2O/Ar (30 min)-3 vol% 12CO/Ar (10 min)-3 vol% 13CO/Ar (10 min)-10 vol% H2O/Ar (t)
(Adopted from Kalamaras et al.38 Copyright 2013 Elsevier).
200 | Catalysis, 2016, 28, 175236
View Online
17:24:10.
88
38
are even more reactive than formate species, the authors have
concluded that carbonate-type species should not be considered as part
of the active C-pool of WGS (e.g., they are spectator species). The inability
of carbonate-type adsorbed species to react with water was proved also by
a separate DRIFTS experiment, where after a feed stream of 985 ppm
CO2/He was passed over the catalyst at 250 1C for 30 min, the feed was
then changed to Ar for 5 min and subsequently to 10 vol% H2O/Ar
(250 1C, t). No change in the infrared bands intensities was seen under
the latter gas switch.38
Figure 7c presents corresponding DRIFT spectra in the region of
adsorbed CO (after deconvolution and curve fitting).38 In the inset of
Fig. 7c, DRIFT spectra before and after deconvolution corresponding to
5 min in H2O/Ar are provided. It is observed that adsorbed CO (all three
kinds present) is relatively unreactive with water. The most refractory
adsorbates were the high frequency linear adsorbed CO on Pt and
adsorbed CO on Pt sites of low coordination (Section 3.2.1.1). The least
refractory was the low-frequency linear CO associated with small Pt
clusters (dPto1.5 nm); about half of its surface coverage reacted towards
CO2 and H2 in the H2O/Ar gas treatment (Fig. 7c, 1994 cm1). It should be
noted that all three kinds of adsorbed CO gave the red isotopic shift
under 13CO/Ar. These results clearly indicate that not all reversibly
adsorbed CO formed on Pt under WGS is able to react with water. The
extent of CO2 readsorption estimated in this manner over a series of
supported Pt catalysts37,38,78 is reported in Table 1. The overestimation of
the C-pool was found to be in the 2580% range, where Ce0.8La0.2O2d
support exhibits the smallest eect and Ce0.8Ti0.2O2d the largest one.
3.3.2 Overestimation of SSITKA H-pool: H/D exchange. Efstathiou
and his co-workers97 have developed an experimental method to
account for the overestimation of the H-pool due to possible H/D
exchange reactions that could occur during the low-temperature WGS
on metal oxide-supported catalysts. It should be noted that this is the
first experimental work published in the open literature that addressed
this issue. The main concept leading to development of this method is
that H/D exchange contributions to the measured H-pool during the
Table 1 Concentration (mmol g1) and surface coverage (y, based on Pts) of active
C-pool in WGS (3 vol% CO/10 vol% H2O/He, 300 1C), corrected for CO2 readsorption,
over various supported Pt catalysts.
0.5 wt% Pt/support
C-pool (SSITKA)
(mmol g1)
CeO2
Ce0.8Ti0.2O2d
Ce0.8La0.2O2d
TiO2
Ce0.5Zr0.5O2d
61.0
542.0
98.2
11.4
109.0
(2.6)a
(28.9)
(3.8)
(0.8)
(10.3)
Titration by H2O
following WGS and
13
CO exchange (mmol g1)
41.3
114.2
75.3
5.3
44.5
(1.7)a
(6.5)
(2.9)
(0.4)
(4.2)
CO2 readsorption
eect (mmol g1)
19.7
427.8
22.9
6.1
64.5
(0.9)a
(22.4)
(0.9)
(0.4)
(6.1)
a
Values in parentheses represent the equivalent surface coverage based on Pts, where
C-containing/Pts is considered to be 1 : 1.
Catalysis, 2016, 28, 175236 | 201
17:24:10.
View Online
SSITKA gas switch CO/H2O-CO/D2O becomes important only over the

time-scale of the HD and H2 transient isotopic responses, which are
used to estimate the amount of the active H-pool.40
The experiment consists of the following four step-gas switches as
applied to a 0.5 wt% Pt/Ce0.8La0.2O2d WGS catalyst.38,97 The ordinary
WGS feed gas composition consisted of 3 vol% CO/10 vol% H2O/He. The
concentration of 0.25 vol% D2 that appears in the following gas switches
was chosen based on the D2 concentration measured at the time where
maximum concentration of HD was formed following the SSITKA switch.
Step gas switches applied:97
(i) 3 vol% CO/10 vol% H2O/Ar/Kr (2 h, 300 1C)-3 vol% CO/10 vol%
D2O/Ar (t) (SSITKA switch)
(ii) 3 vol% CO/10 vol% H2O/Ar/Kr (2 h, 300 1C)-0.25 vol% D2/Ar (t)
(iii) 3 vol% CO/10 vol% H2O/Ar/Kr (2 h, 300 1C)-3 vol% CO/
0.25 vol% D2/Ar (t)
(iv) 3 vol% CO/10 vol% H2O/Ar/Kr (2 h, 300 1C)-10 vol% D2O/Ar (t).
The above four step-gas switches (i)(iv) were performed based on the
following reasoning:
(a) Gas switch (i) is the normal SSITKA switch designed to follow the
H-path of WGS, where in the absence of H/D exchange, the actual concentration of the active H-pool can be estimated.40 In the presence of H/D
exchange, the estimated amount of H-pool represents an upper bound on
the size of the H-pool.
(b) Gas switch (ii) allows recording the transient rates of HD and H2
formation that might likely be obtained during H/D exchange between
the non active H-containing species formed during steady-state WGS, the
H2O(g) and the H2(g) gas product which remain in the reactor for Dt (s)
after the SSITKA switch, and the D2(g) produced after the SSITKA switch.
(c) Gas switch (iii) allows recording the transient rates of HD and H2
formation that might be obtained after H/D exchange between D2(g)
and the species mentioned in (b), but in the presence of 3 vol% CO
(simulating SSITKA switch conditions); this allows one to study the eect
of CO(g) on the H/D exchange processes after comparing this result with
that of step-gas switch (ii).
(d) Gas switch (iv) allows probing both the H/D transient exchange
kinetics between D2O(g) and the species mentioned in (b), and the
reaction of D2O(g) with all active C-containing species formed under
steady-state WGS.
Figure 8a illustrates that the HD(g) transient formation rate via H/D
exchange processes observed during the D2/Ar and CO/D2/Ar gas treatments is largely dierent from that observed during the ordinary SSITKA
switch (i) and the D2O/Ar gas treatment (step-gas switch (iv)). In
particular, there is a significant delay and a large shift towards higher
times in the appearance of the peak maximum in the rate of HD
formation during the D2/Ar and CO/D2/Ar gas treatments, compared to
those observed during the CO/D2O/Ar and D2O/Ar gas treatments.
Furthermore, at t 100 s after the SSITKA switch, the formation rate of
202 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
Fig. 8 (a) Transient response curves of HD formation obtained during the following gas
switches: 3 vol% CO/10 vol% H2O/Ar/Kr (2 h, 300 1C)-3 vol% CO/10 vol% D2O/Ar (t)
(SSITKA switch) or-0.25 vol% D2/Ar (t) or-3 vol% CO/0.25 vol% D2/Ar (t) or-10 vol%
D2O/Ar (t); (b) Transient response curves of Kr, H2, HD and D2 obtained during the
gas switch: 3 vol% CO/10 vol% H2O/Ar/Kr (2 h, 300 1C)-3 vol% CO/10 vol% D2O/Ar (t)
(SSITKA switch to follow the H-path of WGS); Wcat 0.05 g (dilution with SiC, 9 : 1);
F T 100 NmL min1 (Adopted from Efstathiou et al. 97 Copyright 2014 Elsevier).
HD(g) is significantly lower than the maximum HD(g) formation rate

obtained during gas switches (ii) or (iii). Also, at t 162 s, where the HD
response observed after switch (i) is complete (Fig. 8b),97 a significant
amount of HD(g) was measured under D2/Ar and CO/D2/Ar gas treatments. All these observations strongly suggest that the rate of HD
Catalysis, 2016, 28, 175236 | 203
17:24:10.
View Online
formation obtained during SSITKA, which is controlled by the ratelimiting step present in the H-path of WGS for H2 formation, must be
considered significantly larger than the total rate of H/D exchange between H2(g), H2O(g) and D2(g) and the non active H-containing adsorbed
species formed during WGS. As a result of this analysis, a lower limit for
the active H-pool was suggested.97 The latter is obtained after subtracting
the equivalent amount of H obtained after integrating the HD response
recorded during the CO/D2/Ar switch until the end of the SSITKA-HD
transient response at t 162 s (Fig. 8b), from that of the HD response of
the normal SSITKA gas switch (i). The former quantity was estimated to
be 124 mmol g1 and the latter 282 mmol g1. Thus, the net amount of
HD(g) corresponds to 158 mmol g1, or to an equivalent amount of
yH 6.07 (expressed in surface monolayers of Pts). It should be clarified
that in the three gas switches (ii)(iv), the maximum possible concentrations of D2 and D2O were used, whereas after the SSITKA switch and
until the end of the HD transient response, lower concentrations of these
gases appear (Fig. 8b). This leads to maximum H/D exchange rates and a
maximum equivalent amount of H which must be subtracted from the
H-pool measured during SSITKA switch (i).
The observed HD response from the 10 vol% D2O/Ar gas treatment (gas
switch (iv)) shows the same peak maximum position in time and a very
similar shape to that observed during the SSITKA switch (i) (Fig. 8a). The
dierence in the absolute concentration values of the HD(g) obtained in
the two gas switches is the result of a larger transient of D2(g) than HD(g)
observed in the D2O/Ar gas switch. It is also important to mention that
after the gas switch 10 vol% H2O/Ar (2 h, 300 1C)-10 vol% D2O/Ar (t)
neither HD nor D2 gas was observed.38 This result along with the HD
responses shown in Fig. 8a (gas switches (i)(iv)) suggest that surface OH
groups cannot engage H/D exchange with D2O(g) under the gas environment characteristic of the D2O-SSITKA switch. The results of Fig. 8
also suggest that OH groups are able to exchange with D2(g) in the
presence or absence of CO(g) but with lower rates when compared to the
rate of HD formation via the WGS (SSITKA switch (i)).
Figure 9 presents the normalized concentration of Kr (tracer) and H2(g)
formed during the gas switches (ii)(iv).38 The ZH2 response observed for
gas switch (iii) was the same as for gas switch (ii). It is seen that within
the first 45 s of the transient, no H2(g) is formed in the case of D2/Ar gas
treatment, in harmony with the HD response (Fig. 8a); the opposite was
found for the D2O/Ar gas treatment. The dierent ZH2 responses observed
between 50 and 250 s during the D2/Ar and D2O/Ar gas treatments should
be noted. The equivalent amount of H in terms of yH which is estimated
up to the end of the transient (Fig. 9, 250 s) was found to be 3.5, 3.15 and
1.95 for the D2/Ar, CO/D2/Ar and D2O/Ar gas treatments, respectively. In
the case of the SSITKA switch (i) (Fig. 8b), this was found to be yH 4.8.
After considering the largest value of H/D exchange observed in Fig. 8a
(HD response), and subtracting this from the corresponding SSITKA
value (Fig. 8b), a net amount of yH 1.3 is obtained.97 A lower limit for the
H-pool can then be estimated, yH 7.4 (based on the lower values of HD
and H2 formed via H/D exchange).
204 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
Fig. 9 Transient response curves of the dimensionless H2 gas concentration (ZH2) formed
after the switch 3 vol% CO/10 vol% H2O/Ar/Kr (2 h, 300 1C)-0.25 vol% D2/Ar (t) or 3 vol%
CO/0.25 vol% D2 /Ar (t) or 10 vol% D2O/Ar (t); Wcat 0.05 g (dilution with SiC, 9 : 1);
F T 100 NmL min1 (Adopted from Efstathiou et al. 97 Copyright 2014 Elsevier).
Table 2 Concentration (mmol g1) and surface coverage (y, based on Pts) of the active
H-pool in WGS (3 vol% CO/10 vol% H2O/He, 300 1C), corrected for H/D exchange
reactions, over various supported Pt catalysts.
0.5 wt% Pt/support
H-pool (SSITKA)
(mmol g1)
H/D exchange
(mmol g1)
H-pool (lower value)

(mmol g1)
CeO2
Ce0.8Ti0.2O2d
Ce0.8La0.2O2d
TiO2
La2O3
285 (12.1)a
735 (39.3)
408 (15.9)
88 (6.2)
353 (13.8)
87
98
218
18
273
198 (8.4)a
637 (34.0)a
190 (7.4)a
70 (4.9)a
80 (3.1)a
a
Values in parentheses represent the equivalent surface coverage based on Pts, where
H-containing/Pts is considered to be 1 : 1.
The above described methodology for estimating a lower limit for the
size of the active H-pool in WGS can be further extended to investigate the
role of carbon dioxide in the transient rates of possible H/D exchange
processes, in a manner similar to the role of CO (gas switch (iii)). Table 2
illustrates how the support chemical composition influences the overestimation of the active H-pool in WGS over various supported Pt catalysts.38,39,97 In the case of La2O3-supported Pt, the overestimation of the
active H-pool due to H/D exchange was the largest one (77%), whereas
Ce0.8Ti0.2O2d-supported Pt exhibited the smallest contribution to H/D
exchange (13%).
As stated in Section 3.1, one of the unique features of the SSITKA experiment is to allow the estimation of the concentration of active surface
Catalysis, 2016, 28, 175236 | 205
17:24:10.
View Online
reaction intermediates. In the case of supported metal catalysts, the value

of this kinetic parameter provides a rather definite answer as to whether
the catalytic reaction should be considered bi-functional. For example,
in the above-described case study (Figs. 8 and 9), a surface coverage
of yH 7.4 (Table 2) dictates that Pt metal alone cannot accommodate
this coverage (maximum yH for Pts 1). In fact, for this particular
Pt/Ce0.8La0.2O2d catalytic system, the size of the active C-pool was found
to be yC 2.1 (Table 1).38 It is reasonable to argue that yCO is less than
one monolayer (CO-s on Pt). Irrespective of the chemical structure of
active species comprising the C-pool (e.g. formate (HCOO), carboxyl
(COOH) or both), the equivalent H-pool must be larger than one monolayer. It is also unreasonable to assume that the surface coverage of active
H and OH on Pt are zero. Therefore, it is clear that more than 5.0
equivalent surface monolayers of Pts from the H-pool reside on the
Ce0.8La0.2O2d support.
For an H-pool size of yH 7.4, a reactive zone around each Pt
nanoparticle (B1.0 nm mean particle size) of a depth of B22 was
estimated,38 the latter being equivalent to about six lattice constants of the
Ce0.8La0.2O2 support (solid solution). This means that OH groups are
able to diuse within this zone towards the Pt-support interface, and
react with CO adsorbed on Pt sites at the interface to form HCOO active
intermediates under WGS reaction conditions; it was suggested38 that
surface diusion of CO towards the Pt-support interface and within this
reactive zone (to reach the OH sites) might also be considered.
4 Application of SSITKA and other transient isotopic

techniques towards elucidation of WGS reaction
mechanisms
4.1 Supported Pt catalysts
4.1.1 Non reducible metal oxide-supported Pt
4.1.1.1 Pt/g-Al2O3. The SSITKA-MS and SSITKA-operando methodologies for the WGS reaction at 350 1C over Pt/g-Al2O3 (0.5 wt% Pt) were
first applied by Olympiou et al.96 and Kalamaras et al.75 It was reported
that the carbon-path of the WGS reaction passes through an active
formate (HCOO) species, whereas another kind of formate species
interacting reversibly with CO-s (Pt) and OH (support or Pt-support
interface) was considered as inactive (spectator). This conclusion was
based on two important observations. First, during SSITKA-DRIFTS two
kinds of formate species were identified (Scheme 6, IIII), where both
species showed a red isotopic shift. Second, the SSITKA-operando
methodology revealed that the H12COO-s transient decay followed that
of 12CO2(g) as illustrated in Fig. 10,75 a result inconsistent with the
observation that the CO-s precursor is feeding the HCOO-s pool from
which CO2(g) is formed (Fig. 1). As will be discussed in detail in
Section 4.2, unambiguous evidence that formate is a kinetically important reaction intermediate for the Pt/g-Al2O3 catalyst requires measuring
the rate of formate decomposition to CO2 (mmol HCOO g1 s1) under
206 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
Fig. 10 Transient response curves (Z normalized concentration) of 12C-labelled CO and

CO2, and adsorbed 12COL-s and H12COO-s species found in the carbon-path of the
WGS reaction at 350 1C over the 0.5 wt% Pt/g-Al2O3 catalyst (Adopted from Kalamaras
et al.75 Copyright 2008 Elsevier).
SSITKA-WGS and comparing this rate with that of the steady-state WGS
determined independently.
The surface concentration of the active carbon-pool on Pt/g-Al2O375,76
was found to increase with reaction temperature, namely 1.3 mmol g1
(350 1C) vs. 31.7 mmol g1 (500 1C). The equivalent surface coverage (in
terms of monolayers of exposed Pt surface) of the carbon-pool was
yC 0.1 and 2.6, respectively, for T 350 and 500 1C. Because the surface
coverage of CO-s on Pt at these temperatures is expected to be less than
unity, part of the active formate species at 500 1C must reside on the
support. The same group has later estimated that for a Pt mean particle
size of B2 nm (the present one) and for the same Pt loading (0.5 wt%), a
surface coverage of yC 2.6 is accommodated within 13 lattice constants
of support (410 ) around each Pt nanoparticle.38,39 This result lends
strong support to previous works of the Davis group98,100,150 who
presented evidence that formate intermediate decomposition to CO2(g)
and H-s is facilitated by the presence of Pt. Similar experimental
observations were also reported recently by Petallidou et al.40 in the case
of Pt supported on CeO2, La2O3 and La31-doped CeO2 carriers.
For the hydrogen-path of the reaction, the concentration of the active
H-pool (e.g. H/OH) was found to be significantly larger than that of the
C-pool (350 and 1664 mmol g1 or yH 28.5 and 135.6, respectively, at
T 350 and 500 1C). In spite of the fact that in these works75,96 the
authors have not corrected the estimated C-pool and H-pool based on
their later experimental methodology38,39,7678,97 (Section 3.3.2), the main
conclusion regarding the site location of the active formate and H/OH
species remains valid considering the acidic character of alumina surface
Catalysis, 2016, 28, 175236 | 207
17:24:10.
View Online
and the size of the H-pool. In fact, the back-spillover of OH/H

species from the alumina support was evidenced via transient isotopic
experiments (precovered surface by OD followed by WGS).75 A normal
deuterium KIE was observed (RH2/RD2 B2.1 and 1.8) after WGS (3%
CO/10% H2O/He) at 350 and 500 1C, respectively.75,96 Surface diusion
of H species (present as OH) towards the Ptalumina interface, and
decomposition of formate (HCOO-s2 Pt-s1-CO2(g) H-s1 s2) at the
metalsupport interface (Pt-s1) were suggested to be likely slow steps in
the H-path of WGS over Pt/g-Al2O3.75,96 The increase in the concentration
of active C-pool and H-pool with reaction temperature could be explained
on the basis of a surface formate/OH group diusion mechanism
proposed by Duprez and his co-workers,62 and later adopted by Jacobs
and Davis101 in order to explain their SSITKA-DRIFTS results on Pt/CeO2.
This model incorporates the concept that the rate of surface diusion of
an adsorbed species increases with reaction temperature, thus enabling
it to approach catalytic sites and become active.
Density functional theory (DFT) combined with micro-kinetic modelling and Monte Carlo simulations have been widely used to assess WGS
reaction mechanisms on unsupported Pt and Au metal clusters and crystal
faces,80,85,8790,121,151 but only recently have attempts been made to
investigate the eect of the metalsupport interface.41,87,9195,110 Clay
et al.80 in a recent elegant DFT work concluded that several important
discrepancies between several DFT models on metal-only surfaces and
experimentally observed WGS reaction rates on supported metals do
exist. While water dissociation appears as the rate-limiting step on the
metal-only surface (significantly larger activation barrier than other
steps), the rate-limiting step can shift due to the diminished water dissociation barrier at the metalsupport interface.25,9194,152 Grabow et al.88
have presented a microkinetic model for low-temperature WGS
(250300 1C) catalysed by Pt(111) combined with DFT-GGA calculations,
where predictions for reaction rates, orders and apparent activation
energies agree with experimental results for 3 wt% Pt/g-Al2O3 (dPtB3.5 nm).
They have found that the carboxyl (COOH) intermediate on Pt is the
kinetically significant one, where water acts as a promoter of the
COOH OH decomposition path, in line with results of the early work
of Jacobs et al.99,101 and Iwasawa et al.123,124 Also, the direct formate
decomposition step (HCOO-s-CO2 H-s) accounted for more than 75%
of the CO2 production, whereas after increasing the water concentration
in the feed the OH concentration on Pt increases, making the low-energy
HCOO OH decomposition path kinetically accessible. It was claimed
that formate was formed from CO and OH not by a single step, but via a
two-step reaction path, and it was considered as a spectator species.
These results may not contradict those discussed above for the 0.5 wt%
Pt/g-Al2O3 catalyst for the following reasons: (i) the mean Pt particle
size of the latter catalyst was almost half of that used in the work of
Grabow et al.,88 thus the number and site reactivity of interfacial
sites of Pt at the metalsupport interface might have been increased
significantly; (ii) no DFT calculations were attempted with a model
alumina-supported Pt cluster of less than 2.0 nm, where other than
208 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
Pt(111) sites may be important; (iii) the elementary steps of formate

formation and their associated energetics were not considered (sites
either on the alumina alone or at the metalsupport interface). On the
other hand, the diculty in identifying the red isotopic shift of the
carboxyl-IR band during SSITKA-DRIFTS it should be pointed out
(Section 3). Therefore, one cannot exclude the presence of carboxyl
species on the catalyst surface of Pt/g-Al2O3.
4.1.1.2 Pt/ZrO2. Another non-reducible single metal oxide carrier of
Pt of interest in investigating the associative reaction route (e.g. formate or carboxyl) is ZrO2. Tibiletti et al.153 have investigated the nature
and reactivity of the surface species formed during WGS (2% CO/7%
H2O/12% H2/He) at 200 1C over 4 wt% Pt/ZrO2 (dPt 1.5 nm) by operando SSITKA. On the basis of the SSITKA-MS results (Fig. 11a) and the
H13COO SSITKA-DRIFTS response curve (Fig. 11b), which was estimated from the growth of the integral IR band intensity (log (1/R))
recorded in the 31002700 cm1 region (due solely to the n(13CH) and
d(13CH) ns(O13CO) vibrational modes of formate), it was concluded
that the formate route must be ruled out as the main WGS reaction
path at these conditions. Even though the authors did not consider the
fact that at least one of the formate IR bands did give a red isotopic
shift (so potentially an active intermediate), they stated that a small
IR-invisible formate species formed at the PtZrO2 interface cannot be
excluded as an active reaction intermediate. This explanation is in line
with that provided earlier for Pt/g-Al2O3,75 where the 12CO2 and
H12COO decay response curves appeared on a significantly dierent
time scale. In the case of Pt/ZrO2 the concentration of C-containing
active intermediates (C-pool) was lower than one monolayer (based on
Pts), in contrast to Pt/g-Al2O3.75 Tibiletti et al.153 have not considered
that in the integrated IR region more than one kind of formate species
was seen (more than two overlapping IR bands appeared in the
31002700 cm1 region), where the strongest IR band did clearly show
the red isotopic shift. This is again the result of forming at least one
kind of HCOO by the reversible interaction of CO and OH, as discussed
in Section 3.2.1.2 and reported also for WGS over ZrO2 and Pt/ZrO2 at
240 1C.154 Tibiletti et al.153 have not considered that a redox mechanism
may operate on Pt/ZrO2 at 200 1C with 12% H2 in the feed gas stream.
Leerts and his group154 through various transient experiments using
in situ transmission FTIR (not SSITKA) have investigated the role and
reactivity of dierent coordinated hydroxyl groups with CO in the
240400 1C range. On the basis of the observed results, and earlier
CO-pulsing experimental studies118 and DFT estimations,155,156 it was
concluded that during WGS partial reduction of zirconia does occur,
limited to the structural defect sites in the vicinity of Pt nanoparticles. In
particular, the associative formate with redox regeneration hybrid WGS
reaction mechanism was proposed. In this mechanism, CO activation on
Pt is not required to produce formate species on zirconia, where identical
formate structures were detected over both ZrO2 and Pt/ZrO2 surfaces. As
the temperature of WGS increases, the surface coverage on the Pt surface
Catalysis, 2016, 28, 175236 | 209
17:24:10.
View Online
Fig. 11 (a) Mass spectrometry traces of Kr, 13CO and 13CO2 obtained for the SSITKA
switch 2% 12CO/7% H2O/12% H2/Ar-2% 13CO/7% H2O/12% H2/Kr/Ar at 200 1C over the
Pt/ZrO2 catalyst. (b) Normalised MS trace of 13CO2 and intensity of IR band of the
13
C-containing formate following the isotopic switch (Adopted from Tibiletti et al.153
Copyright 2006 Elsevier).
decreases, and as a result the rate of formate decomposition at the metal

support interface also increases. As depicted in Scheme 8, monocoordinated hydroxyls are involved in formate formation, whereas
multi-coordinated hydroxyls and Pt are needed to enable formate decomposition into CO2 and H2.154 Reduction of the support was proposed
to be limited to defective zirconia sites at edges and kinks.118,154 These
sites can be regenerated with H2O, implying that the role of water in
the WGS mechanism is the regeneration of hydroxyl groups and simultaneous re-oxidation of ZrO2 surface (Scheme 8).
Davis and his group49 have also investigated Pt/ZrO2 and Na-promoted
Pt/ZrO2 catalytic systems for the WGS reaction mechanism using various
210 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
Scheme 8 Watergas shift reaction mechanism for Pt/ZrO2. 1: Formate formation

reaction between CO and mono-coordinated hydroxyls. 2: Formate decomposition
involving Pt and multi-coordinated hydroxyls. 3: Hydroxyl regeneration with water
(Adopted from Graf et al.154 Copyright 2009 Elsevier).
in situ DRIFTS (not SSITKA) and H/D exchange experiments.

The authors have also presented experimental evidence for the lowtemperature surface reduction of zirconia shell in the presence of Pt. The
formation of an oxygen vacancy was linked to its participation in the
dissociation of water to form bridging OH groups (type-II, multicoordinated hydroxyls) or in the direct dissociation of hydrogen on Pto,
and spillover of active H on the zirconia support. As claimed by
Graf et al.154 and discussed above, type-II OH groups were the active
hydroxyl species in forming the formate intermediate (associative formate
mechanism), the decomposition of which via the participation of surface
Pt sites is considered as rate-limiting step (CH bond cleavage).124,125,157
The addition of Na1 (2.5 wt%) on the 2 wt% Pt/ZrO2 surface led to an
important weakening of the CH bond, thus to a significant increase in
the rate of formate decomposition under WGS.49
As shown by Tibiletti et al.153 and previously discussed, the importance
of the associative formate mechanism for the Pt/ZrO2 catalytic system can
be demonstrated via operando-SSITKA methodology. For the formate
species to be a true active reaction intermediate, the transient isotopic
response of H13COO in the 12CO/H2O-13CO/H2O SSITKA switch must
not lag behind that of 13CO2(g). A quantitative evaluation of the pure
redox and associative formate mechanisms (Sections 2.1 and 2.2) can best
be performed via micro-kinetic modelling of the SSITKA-DRIFTS and
SSITKA-MS transients. However, such modelling is yet to be reported for
WGS over these alumina- and zirconia-supported Pt catalysts.
4.1.2 Reducible metal oxide-supported Pt
4.1.2.1 Pt/CeO2. The research group of B. H. Davis has extensively
investigated over the last 15 years the role of formate species in WGS
over mainly ceria-based supported Pt catalysts using SSITKA-DRIFTS
(using both 13CO and D2O), D-kinetic isotopic eect, transient formate
decomposition and other types of transient kinetics and chemisorption
Catalysis, 2016, 28, 175236 | 211
17:24:10.
View Online
Scheme 9 Surface formate mechanism in WGS on CeO2-supported Pt catalysts

(Adopted from Jacobs et al.98 Copyright 2015 Springer).
experiments.98101,125,144,145,150,158161 The main conclusions from these

studies can be summarized in Scheme 9, which encompasses the
associative formate with OH group regeneration mechanism mainly
operative in low-temperature WGS, in which the role of water and
metal in the Pt/CeO2 have been extensively investigated. OperandoSSITKA-DRIFTS was used,101 with the DRIFTS cell also used to quantify
CO and CO2, thus allowing an estimate of the WGS reaction rate.
Figure 12 presents transient DRIFT spectra obtained (i) after steadystate WGS at 225 1C with 12CO (2.8% 12CO/46.7% H2O/50.5% H2) (dash
line), (ii) after 11.9 min, at which 50% of 12CO2(g) had been exchanged
by 13C (13CO-SSITKA switch), and (iii) after 27.2 min at which the new
steady-state under the isotopic 13CO-containing feed gas was obtained
over the 0.5 wt% Pt/CeO2 catalytic system.101 Similar DRIFT spectra are
also presented over two other ceria-supported Pt catalysts with higher
loadings (1.0 and 2.0 wt% Pt) in an eort to investigate the role of Pt
metal loading on the assumed associative formate mechanism in WGS.
The authors noted that as one moves from lower to higher Pt loading:
(a) The CO2(g) production increases, (b) the surface formate steadystate coverage is decreasing, suggesting that the formates are reacting
faster, (c) the times to achieve 50% fractional isotopic exchange in
the formate (nCH) and gas-phase CO2 (nCO2) infrared bands nearly
coincide, (d) the band for adsorbed CO on Pt had already completely
exchanged when the time to achieve only 50% fractional isotopic exchange in nCO2 is reached, and (e) the times to achieve 50% fractional
212 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
Fig. 12 Impact of Pt metal loading on the formate coverage, the CO2 production rate,
and the half-lives of reactive exchange for surface formate and CO2 during 13CO-SSITKADRIFTS. Note that increasing the metal loading decreases the formate coverage and
decreases the half-lives of formate and CO2 during the reactive exchange process
(Adopted from Jacobs and Davis.101 Copyright 2007 Elsevier).
Catalysis, 2016, 28, 175236 | 213
17:24:10.
View Online
isotopic exchange in the formate n(CH) and nCO2 bands decrease with
increasing Pt metal loading. The latter was interpreted by the
authors101 as an indication that formate dehydrogenation is accelerated at the Ptceria interface.
The above results (Fig. 12) were also discussed by Jacobs and Davis101
in relation to a surface formate diusion mechanism, the latter being
elaborated by 16O/18O and H/D isotopic exchange studies by Duprez and
co-workers.62,162 This diusion mechanism model considers three formate types: (i) those reacting rapidly at the metalsupport interface
(rapid reaction zone), (ii) those at path lengths sucient to eventually
diuse to the metal and contribute to the overall activity (intermediate
surface diusion zone), and (iii) those that are essentially locked onto the
surface due to excessive diusional path lengths to the metaloxide
interface (stranded intermediate zone). This last would apply at low
temperatures and for low metal loadings. Jacobs and Davis101 argued
that ceria loaded with 0.5 and 1.0 wt% Pt (Fig. 12) belong to the case
(ii), whereas for the largest Pt loading of 2.5 wt% the catalyst approaches
case (iii). For case (i) the fractional exchange of total formate would be
slower than that of CO2 product exchange.
The discussion oered by Jacobs and Davis101 on the experimental
results of Fig. 12 should be re-examined with respect to the following
points. First, the variation in absorbance intensity shown in Fig. 12 for
the three solid catalysts has not been shown to follow a linear relationship with surface concentration of formate species, a critical issue for the
accurate quantification of the DRIFTS integral band intensity of an
adsorbed species.126130 Secondly, the fast isotopic exchange of 12CO-s by
13
CO-s (red isotopic shift) compared to those of H12COO-s and 12CO2(g)
cannot be considered as an indication of the absence of a redox
mechanism, as claimed by the authors. Referring to Fig. 1, both WGS
mechanisms require that an active pool of CO-s (1st pool) formed on
Pt feeds a pool of OH species to form adsorbed HCOO-s (associative
formate mechanism, 2nd pool) followed by a pool of adsorbed CO2-s (3rd
pool), or the CO-s (1st pool) feeds a pool of O-s (lattice oxygen at the
metalsupport interface, redox mechanism) to form CO2-s (2nd pool).
Decomposition/desorption of CO2-s (carbonate-type species) leads then
to the formation of CO2 gas product. In the SSITKA switch, the exchange
time (e.g. 50% of the surface concentration) depends on surface coverage
and site reactivity. A fast exchange of CO-s due simply to its reversible
adsorption on Pt may result in similar isotopic exchange response irrespective of whether CO-s participates in the redox or the associative
formate mechanism. Third, the decreasing concentration of formate and
the concomitant increase of WGS rate (CO2 formation) with increasing Pt
loading can also be the result of the increase of Pt-support interface
length, Io (cm g1), which also favours the redox pathway.78 Finally, as
will be further discussed below and in Section 4.2, convincing evidence
for formate being an important active intermediate in WGS is when its
rate of decomposition (e.g. mols HCOO g1 s1) to CO2 and H (or H2), not
to CO and OH or CO and H2O, is comparable to the kinetic rate of WGS
(CO2 or H2 formation rate). The latter determination was not performed
214 | Catalysis, 2016, 28, 175236
View Online
101
17:24:10.
in the work of Jacobs and Davis,

or in earlier SSITKA-DRIFTS
investigations.125
The mechanism shown in Scheme 9 merits the following further
discussion.
(i) The forward rate of formate decomposition to CO2 and H-s requires
CH bond breaking, the rate of which was found to be at least an order of
magnitude larger when Pt (or Au) is loaded on ceria, compared to the
ceria alone.40,78,98,163 As shown in Scheme 9, decomposition of the formate intermediate associated with a CeO pair site (see also Scheme 6) is
largely facilitated by the presence of adjacent Pt sites (Ptceria interface)
which enhance CH bond cleavage and stabilisation of abstracted
hydrogen via PtH bonding. The latter was first proposed by Shido and
Iwasawa124 and later by Jacobs et al.,158,163 and recently by Efstathiou and
his group40,78 on the basis of transient DRIFTS experiments on various
ceria-based supported Pt (0.5 wt%) catalysts. The increase in metal
loading (Pt and Au) increased the formate decomposition rate.158 The
authors suggested that by increasing the metal loading, the number of
metal nanoparticles per gram of catalyst increases, thus increasing
the surface concentration of activated formate species on the ceria support.158 This concept has recently been strongly supported by Efstathiou
and co-workers,3840,77,78 who have proposed (after quantification of the
active C-pool) that a reactive zone of a small extent (515 ) around each Pt
nanoparticle accommodates the active C-pool (including formates) on
ceria-based supported Pt catalysts.
(ii) The formate decomposition rate is largely enhanced by the presence of water as the result of a favourable tilting rotation of formate
(HCOO), as indicated in Scheme 9 (activated complex is shown in
brackets).
(iii) Bridging hydroxyl groups associated with surface Ce31 appear to
be active intermediates,99 produced by water dissociation on Ce31Vo
sites (Vo: oxygen vacant site). The bridging hydroxyl group reacts with CO
to generate surface HCOO. This type of hydroxyl group (IR band at
3650 cm1) has previously been identified as thermally stable even at
500 1C.164 The bridging OH active site density increases with the extent of
ceria surface shell reduction.
(iv) The metal (Pt or Au) plays a role in the formation of bridging OH
groups by facilitating either the removal of O (reduction of Ce41 to Ce31),
thus allowing H2O dissociation onto Vo, or the H2 rupture, allowing for H
spillover to the ceria surface to form OH groups.98
(v) The normal kinetic isotopic eect (NKIE) observed in the SSITKA
switch CO/H2O-CO/D2O may not be associated with water dissociation
on oxygen vacant sites as proposed earlier,42,77,165 since recent
investigations by a combination of in situ and operando spectroscopic
techniques and DFT computations suggested that this step is fast over Pt
supported on ceria-based reducible metal oxides.110 The CH bond
breaking of formate (HCOO-s) was suggested to be a more likely possibility in the case of an associative formate with OH group regeneration mechanism.
Catalysis, 2016, 28, 175236 | 215
17:24:10.
View Online
Fig. 13 Comparison of the relative exchange of the gas-phase CO2 and surface formate
(HCOO-s) species during SSITKA (2% 12CO/7% H2O/Ar - 2% 13CO/7% H2O/Kr/Ar) over a
2 wt% Pt/CeO2 at 160 and 220 1C (Adopted from Meunier108 Copyright 2010 Elsevier).
The similar %-exchange of 12C to 13C labelled adsorbed and gaseous

species (e.g. CO2(g) and HCOO-s) observed in the SSITKA work of
Jacobs and Davis,101,125 which was used to decide whether the adsorbed
intermediate is active or inactive (spectator), has been questioned by
Meunier et al.106,108 These authors observed in a SSITKA operando WGS
experiment (DRIFTS and MS) on a 2 wt% Pt/CeO2 catalyst at 160 1C that
exchange of formate was significantly slower than that of CO2 (Fig. 13),
suggesting that formates were unimportant surface intermediates.
However, the exchange of these two species was essentially identical at
220 1C (Fig. 13).
The SSITKA-DRIFTS-MS results shown in Fig. 13 were analysed on the
basis of a simple first order kinetic model describing the rate of depletion
of the adsorbed H12COO-s species during the SSITKA switch.132
Rate of formate exchange d [H12COO-s]/dt k [H12COO-s] (13)
where [H12COO-s] is the surface formate concentration. The value of k
was estimated based on the transient formate exchange curve shown
in Fig. 13 (SSITKA-DRIFTS). The concentration of formate under
steady-state WGS was estimated via calibration curves as detailed elsewhere.105,132 Assuming that the rate of CO2 formation under steady-state
WGS (RCO2 (s.s.)) is controlled by the formate decomposition step, then
the rate estimated via eqn (13) should equal RCO2 (s.s.). But it was
found108 that at both temperatures (160 and 220 1C) the rate of formate
decomposition was significantly lower, leading the authors to conclude
that formate was only a minor reaction intermediate in WGS.
Efstathiou and co-workers97 have expressed concerns about the validity
of eqn (13) that may lead to wrong conclusions about the comparison of
216 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
WGS and formate decomposition rates. They argued that decomposition

of HCOO-s involves metal sites (e.g. M Pt, Au) at the interface, as
illustrated in Scheme 9, thus the rate should also be proportional to the
concentration of empty metal sites. Estimation of an apparent rate constant (kapp, eqn (13)) associated with the rate of formate decomposition,
to be compared with an intrinsic rate constant (k) estimated on the basis
of a description of the mechanism of formate decomposition (Scheme 9)
via a sequence of elementary steps could lead to a dierent result. This
subject is further elaborated in Section 4.2.
The use of a micro-kinetic analysis to simulate the SSITKA transient
response curves of Fig. 13 in order to investigate rival mechanisms
(redox and associative formate) is another powerful tool to justify the
importance of formate species in the WGS reaction path. This approach
has been followed successfully recently in the case of the more complex
FischerTropsch catalytic reaction mechanism.166
Efstathiou and co-workers3740,77,78 have investigated the influence of
Pt mean particle size and the synthesis method of the ceria support on
the mechanism of WGS in the 250300 1C range using the operandoSSITKA technique and various other transient isotopic experiments.
A strong linear relationship between the specific kinetic rate of WGS estimated on the basis of the length of Pt-ceria interface (mmol CO cm1 s1)
and the mean Pt particle size, dPt (nm) in the 1.38.0 nm range was
found (Fig. 14).78 After measuring the active C-pool via the operando
methodology (Section 3.3.1), results of which are presented in Fig. 15a, it
was shown that the concentration of these species along with the active
H-pool (via SSITKA-MS) can be justified as being formed on Pt and along
the Ptceria interface in the case of 1.3, 3.0 and 8.0 nm in size Pt particles.
Fig. 14 Eect of mean Pt particle size (dPt, nm) on the specific kinetic rate of WGS reaction
based on the length of perimeter of Ptceria interface (cm g1), RCO (mmol cm1 s1) at 250,
275, and 300 1C (Adopted from Kalamaras et al.78 Copyright 2011 Elsevier).
Catalysis, 2016, 28, 175236 | 217
17:24:10.
View Online
These results were interpreted considering that active Pt and support

sites at the interface can possess dierent site reactivities due to the
dierent local electronic environment as Pt particle size varies. The
results of this work78 are in harmony with the recent DFT computations
of WGS activity and the prevailing mechanism on Pt/CeO2(111) compared
to Pt single crystals94 presented in Section 2.
The 12CO2 response curve shown in Fig. 15a was linked to active formate species, which after 13CO/Ar gas treatment did not show the red
isotopic shift (Section 3.3.1) but did show the shift after the SSITKADRIFTS switch.78 The latter results probed for an active formate species
formed by the irreversible interaction of CO(g) with OH groups of the
ceria support at the metalsupport interface. The 13CO2 response shown
in Fig. 15a was linked to adsorbed CO-s on Pt formed during WGS and
which was all exchanged under the 13CO/Ar gas treatment. To exclude the
possibility that the 12CO2 transient response curve (Fig. 15a) resulted
from a reaction of adsorbed carbonate-type species with water during the
10% H2O/Ar gas treatment, the authors performed the following experiment, results of which are presented in Fig. 15b. The catalyst was first
treated with 900 ppm 12CO2/Ar gas at 300 1C for 20 min, purged with He
for 3 min and the gas flow was then switched to 3 vol% 13CO/Ar for 3 min.
The feed was then switched to 10 vol% H2O/Ar and the transient
responses of H2, 12CO2 and 13CO2 were recorded with time on stream. It is
seen (Fig. 15b) that no formation of 12CO2 was obtained, whereas the
13
CO2 response was interpreted as the result of the reaction of adsorbed
13
CO with water. It is also noted that after the 13CO/Ar gas treatment no
12
CO was formed, therefore excluding exchange of carbonyl in 12CO32
with gas phase 13CO.
The dierent transient reactivity of HCOO-s and CO-s towards water as
depicted in Fig. 15a was considered as evidence for the parallel operation
of the redox and associative formate with OH group regeneration
mechanisms on this particular Pt/CeO2 catalytic system.78 The former
mechanism was also supported by experimental evidence obtained by the
following 16O/18O isotopic exchange followed by WGS. After the 0.6 wt%
Pt/CeO2 catalyst was pretreated in 5 vol% 18O2/He at 600 1C for 30 min, it
was purged in He flow and cooled to 80 1C. The feed was then switched to
H2 (1 bar) at 80 1C for 20 min to reduce only Pt18Ox and then the catalyst
was heated from 80 to 300 1C in Ar. Then the catalyst was exposed to the
WGS reaction mixture (3 vol% C16O/10 vol% H2O/Ar) at 300 1C and the
evolution of isotopic forms of CO2 monitored (Fig. 16).78 Integration of
the C18O16O and C18O2 response curves provided the amount of
18
O (mmol O g1) present in the CeO2 support which participated in the
carbon-path of WGS. This quantity was 41.9 mmol18O gcat1 or 1.4
equivalent monolayers of surface Pt (yO 1.4).78 This result illustrates
clearly the participation of lattice oxygen of the ceria support in
the carbon-path of WGS. The estimated concentration of 18O which resulted in CO2 formation is larger than the estimated surface concentrations of the active OH and COOH intermediates. Under WGS at
300 1C, diusion of surface 18O towards the Ptceria interface is
possible.62
218 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
Fig. 15 (a) Transient response curves of H2, 12CO2 and 13CO2 obtained on 0.6 wt%
Pt/CeO2 catalyst at 300 1C with the gas delivery sequence: 3 vol% 12CO/10 vol% H2O/Ar
(300 1C, 30 min)-3 vol% 13CO/Ar (300 1C, 2 min)-10 vol% H2O/Ar (300 1C, t);
(b) Transient response curves of H2, 12CO2 and 13CO2 obtained on 0.6 wt% Pt/CeO2
catalyst at 300 1C with the gas delivery sequence: 900 ppm 12CO2/Ar (300 1C,
20 min)-3 vol% 13CO/Ar (300 1C, 2 min)-10 vol% H2O/Ar (300 1C, t) (Adopted from
Kalamaras et al.78 Copyright 2011 Elsevier).
The operando-SSITKA results (WGS at 300 1C, same feed gas composition) reported by Kalamaras et al.38,78 on Pt (1.2 and 3.0 nm in size)
supported on two dierent ceria supports (Aldrich (5.6 m2 g1) and
Catalysis, 2016, 28, 175236 | 219
17:24:10.
View Online
Fig. 16 Transient response curves of C16O2, C16O18O and C18O2 obtained during 18O/16O
isotope exchange at 600 1C followed by WGS (C16O/H216O/He) at 300 1C over the 0.6 wt%
Pt/CeO2 catalyst (Adopted from Kalamaras et al.78 Copyright 2011 Elsevier).
lab-synthesized by the citrate solgel method (14.5 m2 g1)) revealed that

the amounts of the active H-pool and C-pool were significantly larger for
the smaller size Pt particles. This result could be due to the larger
number density (per gram basis) of small Pt nanoparticles activating a
larger concentration of OH groups in their vicinity, in agreement with
the previously discussed view of Jacobs and Davis101 for formate species.
4.1.2.2 Pt/TiO2. Mechanistic aspects of low-temperature WGS on a
0.5 wt% Pt/TiO2 (Degussa P25, SSA 41 m2 g1) were reported by
Kalamaras et al.76 using the operando-SSITKA-DRIFTS methodology. On
the basis of the SSITKA-DRIFTS results obtained at 200 1C (Fig. 17),
where the red isotopic shift was not observed in the 13001600 cm1
and 28003000 cm1 regions, the authors concluded that two types of
formates (Fig. 17B) and carbonate adsorbed species were clearly
observed (Fig. 17A), but these species cannot be considered as active
reaction intermediates. On the other hand, three types of linear
adsorbed CO were identified (Fig. 17C, IR bands at 2076, 2046 and
1995 cm1), which all showed the red isotopic shift. The estimated concentration of the active C-pool was found to be very small, 1.6 mmol g1
(yC 0.007), whereas a larger concentration of reversibly adsorbed CO
formed during WGS was estimated (yCO 0.95). These results indicated
that there was a significant amount of adsorbed CO on Pt which did
not participate in the formation of CO2. The alternative redox mechanism was also investigated76 after performing a similar experiment to
that described in Fig. 16 for the case of Pt/CeO2. The amount of 18O of
titania support that exchanged with 16O and which participated in the
220 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
Fig. 17 SSITKA-DRIFT spectra recorded in the 16501300 cm1 (A), 30002800 cm1 (B)
and 21001850 cm1 (C) ranges during WGS using 3% 12CO/10% H2O/Ar/He (solid line
spectra) and 3% 13CO/10% H2O/Ar (dashed line spectra) feed gas compositions. In Figs. (B)
and (C) deconvoluted DRIFT spectra are also shown (Adopted from Kalamaras et al.76
Copyright 2009 Elsevier).
WGS at 200 1C was found to be 12.6 mmol g1 or yO 0.55 (based on

surface Pt). On the basis of these results and others reported,76 the
authors concluded that redox is the dominant mechanism of WGS
Catalysis, 2016, 28, 175236 | 221
17:24:10.
View Online
(3% CO/10% H2O/Ar) at 200 1C and that formates and carbonates are
spectator species. In a subsequent publication Kalamaras et al.77
investigated a 0.5 wt% Pt supported on another commercial titania
carrier (Alfa-Aesar) of larger SSA (150 m2 g1) but of similar mean Pt
particle size, and using a H2O/CO feed ratio of B7.0 (4.2% CO/29.3%
H2O/He), which was larger than that used in the previous work (H2O/
CO 3.3).76 It was again found that neither formates nor carbonates
were active intermediates at 200 or 300 1C (based on the absence of red
isotopic shift in the SSITKA-DRIFTS and on CO2 isothermal adsorption/
desorption studies). In the same work,77 deposition of 6.6 wt% CeO2
on 0.5 wt% Pt/TiO2 (similar mean Pt particle size) gave greater activity
at T4275 1C but lesser at To275 1C. At 200 1C, only the redox mechanism was operative, whereas at 300 1C both the redox and associative
formate with OH group regeneration mechanisms were suggested to
operate. The surface coverages of the active C-pool and H-pool were
found to depend on support composition and reaction temperature
(200300 1C).77 A similar NKIE (1.141.18) was reported for Pt/TiO2, Pt/
CeO2 and Pt/6.6 wt% CeO2TiO2.
The redox mechanism strongly suggested in the works of Kalamaras
et al.76,77 over the 0.5 wt% Pt/TiO2 catalysts finds strong support in the
recent theoretical and experimental work of Ammal et al.93 (DFT computations and other considerations to those reported in Section 2 for
Pt/CeO2(111)). The authors concluded that in the 230300 1C range, the
rates of the associative carboxyl pathway were six orders of magnitude
smaller than those of the redox pathway, suggesting that the oxygen
vacancy structure (Section 2) plays an essential role in promoting water
adsorption and dissociation. Furthermore, the authors were able to show
that the WGS reaction rates increase dramatically at the corner Pt atoms
present at the Pt/TiO2(110) interface. The edge Pt atoms at the same
interface are active at low temperatures, whereas the corner atoms
contribute more to the overall rate at higher temperatures. For Pt
particles in the 1.52.0 nm range (cub-octahedral), it is expected that
there are more edge than corner Pt atoms.167
Iida and Igarashi,168 in early work on low-temperature WGS with
Pt/TiO2, have used the SSITKA-FTIR methodology (transmission mode,
with D2O) to examine the superior catalytic activity of Pt supported on
rutile titania (R) compared to anatase (A). Four types of adsorbed CO were
identified in the 21001950 cm1 range (after deconvolution), where their
reactivity towards water was investigated by time resolved FTIR, applying
a first order kinetics model for the rate of depletion of CO-s.168 Two linear
weak (LW, IR band o2050 cm1) and two linear strong (LS, IR band
42050 cm1) CO adsorbed states were resolved. The LW states were
found to be more abundant and reactive (initially) than the LS states in
both rutile and anatase titania-supported Pt. FTIR spectra collected in the
OD region (29002600 cm1) revealed that the linear-type OD group had
higher oxidizing ability (to formate, leading to CO2 and H2) than the
bridging-type OD group. The main conclusion was that weakly adsorbed
CO on Pt in conjuction with surface OH groups with oxidizing ability
accelerate the synthesis and decomposition of formates on the Pt/TiO2
222 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
(R), resulting in high LT-WGS catalytic activity. The latter conclusion

contradicts the results of the operando-SSITKA work of Kalamaras
et al.76,77 and those of the DFT computations93 mentioned in the previous
paragraphs. However, it should be noted that (i) the authors could not
observe formate bands in the nCH (nCD) region, (ii) no experimental
studies were conducted to probe for the redox mechanism,76,77 and (iii)
formate decomposition rates were not compared to those obtained under
steady-state WGS.
4.1.2.3 Pt/Ce1xMxO2d (M La31, Ti41, Zr41). Operando-SSITKA investigations of WGS have been reported by Efstathiou et al.3740,77,78
over Pt supported on ceria-doped (solid solution) metal oxides
(dopants: La31, Ti41 or Zr41). It was found (Fig. 18) that the specific
kinetic rate of WGS (mmol CO g1 s1) at 300 1C when ceria is doped
with the optimum atom-% concentration (2050 atom-%) correlates
with the concentration (mmol g1) of the active C-pool. Ti41-doped
Pt/ceria (20 atom-% Ti41) exhibited 7.2 times larger specific rate and
2.9 times larger C-pool (mmol g1) than Pt/CeO2. In these investigations
the Pt mean particle sizes were similar (1.21.8 nm), therefore, the kinetic parameters (e.g. rate, surface coverage of active intermediates) were
largely influenced by the support chemical composition. No clear correlation was found between the concentration of the active H-pool and
the WGS kinetic rate, and the authors suggested that the likely ratelimiting step of WGS lies within the carbon-path. The works of Heyden
et al.91,94 on Pt/CeO2(111) and Pt/TiO2(110) are very informative for providing fundamental reasons for the higher activity of Pt/Ce0.8Ti0.2O2d
compared to Pt/CeO2 (Fig. 18).39 The authors reported that nearly all
interfacial Pt atoms on the Pt/TiO2(110) could be active because CO
Fig. 18 Correlation of active C-pool and specific reaction rate, rCO (mmol g1 s1) over
0.5 wt% Pt/CeO2-doped catalysts as a function of dopant (Zr41, La31 and Ti41) (Adopted from
Petallidou et al.39 Copyright 2014 Elsevier).
Catalysis, 2016, 28, 175236 | 223
17:24:10.
View Online
adsorption on these Pt atoms is not very strong, whereas very strong

adsorption of CO was found on the interfacial Pt atoms of
Pt/CeO2(111), precluding WGS at low temperatures.91,94 It was also
concluded that the lower TOF on Pt/CeO2(111) could result from fewer
active interface sites and the higher activation energy barrier reflects
the stability of the oxygen vacancy structure which was identified as the
dominant intermediate for both catalytic systems. In fact, the authors
have suggested that destabilizing the oxygen vacancy structure in ceria
by an appropriate doping could decrease the activation barrier and thus
increase the rate of WGS. It is important to point out that on Pt/
TiO2(110) the redox mechanism prevails, whereas on Pt/CeO2(111) both
the associative carboxyl with OH group regeneration and the redox
can take place as shown in Sections 4.1.2.1 and 4.1.2.2. However, no information was provided which mechanism was the prevailing one.91,94
The eect of La31, Ti41 and Zr41 dopants on the concentration of active
H-pool (OH/H, HCOO and COOH species) and C-pool and the extent of a
reactive zone, Dx (nm) established around the Pt nanoparticles during
WGS reaction conditions is shown in Fig. 19.39 In all cases the reactive
zone is less than B5.0 nm (less than B15 support lattice constants) but
depends on the metal dopant used. The size of the pool is determined by
the individual rates of important steps in the reaction path of WGS, thus
on the site reactivity (k, s1) within this zone including the Ptsupport
interface.
4.2 Supported Au catalysts
Meunier et al.105 were the first to apply the operando SSITKA-DRIFTS-MS
methodology in WGS on a Au/La-CeO2 catalytic system at low temperatures (155220 1C), with the goal being to investigate the extent of
participation of formate in the reaction-path of CO2 formation.
A quantitative analysis was carried out to determine both the rate constant (k) of formate decomposition (leading to CO2 and H-s) and the
surface coverage of formate under steady-state WGS reaction conditions.
The authors claimed that the analysis performed allows one to compare
the specific rate of CO2 formation in WGS (mol CO2 g1 s1) with that of
formate decomposition leading to CO2 (mol HCOO or CO2 g1 s1), thus
to conclude whether formate must be considered as an important
intermediate or a minor one (rate of HCOO decomposition{rate of
WGS). The relative IR integral band intensity or equivalently the relative
dimensionless concentration of H12COO (Z Abs(t)/Abs(t 0)) during
the SSITKA-DRIFTS switch 2% 12CO/7% H2O/Ar-2% 13CO/7% H2O/Ar
was modelled as the sum of two exponential terms given by the following
eqn (14), where a satisfactory fit was obtained.
Z(t) Af exp(kf t) As exp(ks t)
(14)
The physical meaning of each term in eqn (14) is that two dierent kinds
of formate species were considered with a dierent distribution (A factor)
and reactivity (k). The first kind of formate species was considered
very reactive (fast exchange, kf), while the second one less reactive (slow
exchange, ks), with the sum of Af and As equal to unity.
224 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
Fig. 19 Eect of Zr41, La31 and Ti41-doping of CeO2 on the active C-pool (a) and H-pool
(b) of WGS on 0.5 wt% Pt/CeO2-doped catalysts. The estimated extent of the reactive
zone, Dx (nm) accommodating the active C-pool and H-pool is also shown in (a) and (b),
respectively (Adopted from Petallidou et al.39 Copyright 2014 Elsevier).
The formate (H12COO) transient exchange curve could not be fitted by

a simple exponential term as in eqn (13). The rate of formate decomposition was considered as the sum of the fast and slow formate
decomposition via the following eqn (15):105
Rate of HCOO-s decomposition kf [HCOO-s]f ks [HCOO-s]s
(15)
Figure 20 presents the temperature dependence of the rates of formate

decomposition and CO2 production in WGS. It is seen that the rate of
formate decomposition is more than one order of magnitude smaller
Catalysis, 2016, 28, 175236 | 225
17:24:10.
View Online
Fig. 20 Rate of CO2 formation in WGS and rate of formate decomposition estimated via
eqn (15) following the SSITKA-DRIFTS switch 2% 12CO/7% H2O/Ar-2% 13CO/7% H2O/Ar
over the 0.6 at% Au /7.3 at% La-CeO2 catalyst in the 155220 1C range. (Adopted from
Meunier et al.105 Copyright 2007 Elsevier).
than the rate of WGS. In particular, at 220 1C a 60-fold lower rate of

formate decomposition is noted. The analysis of the 13CO2 SSITKA-MS
transient response curves recorded was approximated by a single
exponential function (first order kinetics) from which an apparent
activation energy of 23.7 kJ mol1 was obtained. This value was
dierent from that of CO2 formation obtained from steady-state WGS
(40 kJ mol1).
The SSITKA-DRIFTS-MS analysis employed by the authors105 attempted
to estimate the rate of an individual elementary reaction step, that of
formate decomposition, to be compared to the overall WGS kinetic rate.
If formate is an important intermediate formed in the carbon-path of
WGS, and assuming that only the associative formate with OH regeneration mechanism applies for the present catalytic system, then the
estimated rate of formate decomposition towards CO2 formation ought to
be similar to that measured under WGS.
The simple kinetics used by the authors105 in describing the dynamics
of H12COO-s depletion during SSITKA is not consistent with Jacobs
et al.98 mechanism (Scheme 9), where formate decomposition is enhanced by the presence of H2O and requires also the presence of adjacent
metal sites; the description of the transient H12COO-s decomposition
rate should consider all these facts. The following eqn (16)(23) provide
the basic elementary steps based on which the rate of H12COO-s
depletion must be described according to Scheme 9:
k1
COg s1 ! CO-s1
k2
CO-s1 ! COg s1
226 | Catalysis, 2016, 28, 175236
(16)
(17)
View Online
k3
CO-s1 HO-s2 ! HCOO-s2 s1

k4
HCOO-s2 s1 ! CO-s1 HO-s2

k5
17:24:10.
HCOO-s2 H2 Og s2 ! A*
k6
A* ! CO2 g H2 g OH-s2

k7
CO2 g s2 ! CO2 -s2

k8
CO2 -s2 ! CO2 g s2
(18)
(19)
(20)
(21)
(22)
(23)
DRIFTS normalised integral absorbance intensities of formate

(H12COO-s2) and 12CO-s1 exchange, and normalized exchange response
curves of gaseous 12CO and 12CO2 should then be used to estimate all ks
for the above associative formate with OH group regeneration
mechanism by curve-fitting and optimisation procedures. Eqn (16)(17)
and (22)(23) describe the reversible chemisorption of CO on metal free
sites (s1: Pt or Au), and of CO2 on support s2 sites, respectively. These
steps have been shown to occur by the authors105 and others.26,27 The
activated complex [A]* shown in eqn (20)(21) is that indicated in
Scheme 9 in brackets. The authors105 also compared the apparent activation energies of decomposition of the two kinds of formates with the
apparent activation energy of the WGS measured independently from
kinetic studies. The latter quantity incorporates activation energies of
individual steps in addition to that of formate decomposition, and is
usually measured through steady-state rate measurements using a power
rate law relationship.
Leppelt et al.169 investigated the kinetics and mechanism of low-T WGS
on Au/CeO2 catalyst using also transient DRIFTS experiments with the
goal being to estimate the initial rate of formate decomposition during
the gas switch 1% CO/2% H2O/N2 (s.s. WGS)-2% H2O/N2 (t) at 180 1C. By
estimating the initial surface coverage of formate before the H2O/N2
switch (based on a TPD run after WGS for 22 h), the authors were able to
estimate from the initial slope of the curve Aabs (2830 cm1) versus time
the initial rate of formate decomposition in the presence of water. The
latter was found to be 5.6 mmol g1 s1 to be compared with the WGS
reaction rate of B10.0 mmol g1 s1. This result along with other findings
prompted the authors to propose that the associative formate with OH
regeneration mechanism applies on Au/CeO2 with formate decomposition to be the rate-limiting step. One could argue that the
evaluation of the rate of formate decomposition was not performed
under SSITKA conditions (undisturbed surface composition of adsorbed
phase) as attempted in the work of Meunier et al.105 However, it is not
unlikely that even though the surface composition of the adsorbed phase
Catalysis, 2016, 28, 175236 | 227
17:24:10.
View Online
may change upon the H2O/N2 switch, the reactivity of sites responsible
for formate decomposition according to Scheme 9 would remain close to
those under working WGS reaction conditions.
In spite of concerns over the appropriateness of the analysis of the
transient exchange rates of formate and CO2 reported by Meunier
et al.,105 of interest is the result of their analysis for non-uniform reactivity of formates. The authors suggested that this could be related to
the presence of Ce and La cations or even to surface patches of the
segregated oxides on the support. The fact that the distribution ratio of
fast to slow formate species varied widely over the temperature range
used (Af and As values) was explained by (i) a variable extent of spillover of
formate species with temperature, (ii) a temperature-dependent promoting eect of Au due to an enhanced metalsupport electronic eect,
e.g., an increased extent of active reduced ceria around the Au nanoparticles as proposed for Pt/CeO2,144,170 or to an increased capture zone
of slowly diusing surface species169 and (iii) an increased mobility of
surface Au with temperature. The notion of the existence of a reactive zone
around Au nanoparticles supported on CeLa2O3 carrier which accommodates formate species has been also discussed for CeO2-based supported Pt catalysts in the previous Sections 3.3.2 and 4.1.2.
The main conclusion from the operando-SSITKA work of Meunier
et al.105 over the Au/La-CeO2 catalyst is that the associative formate with
OH group regeneration mechanism cannot explain the high rate of
WGS in the 155220 1C range. As stated by the authors, the main WGS
reaction path remains unclear. It was suggested that formates could act
as a kind of buer species storing CO at the surface, but this does not
mean that formate is the main reactive intermediate leading to the formation of CO2.105
Ribeiro and his co-workers26 in an elegant work have used the operando
SSITKA-FTIR-MS methodology to investigate the deactivation mechanism
and the nature of Au active species of a Au/TiO2 catalyst for WGS. In spite
of the very high activities of supported Au catalysts for low-temperature
WGS,24 the catalysts deactivate with time on stream, and mechanisms
for this deactivation have been proposed, namely: sintering of Au for
Au/CeZrO4,171 or increased blocking of the ceria surface by the formation
of surface carbonates, formates or hydroxycarbonates for Au/CeO2.169,172
Figure 21A presents SSITKA-FTIR (transmission mode) spectra of WGS
(6.8% CO/11% H2O/37.5% H2/Ar/He) at 120 1C before (a) and after 2 s (b)
and 120 s (c) of switching to the equivalent feed containing the
isotopically labelled 13CO. After switching to 13CO, IR bands due to
adsorbed CO on Auo and Aud shifted from 2100 to 2052 cm1 and 2040
to 1992 cm1, respectively,26 indicating the exchange of non-labeled CO
with 13CO. The bands from gas-phase CO and CO2 also shifted upon
switching to 13CO. The response curves of these shifts were fast, whereas
the IR bands from formates and carbonates remained essentially unchanged for more than 2 min. The authors suggested that the overwhelming majority of the surface carbonates and formates are spectators
for WGS at 120 1C. Figure 21B shows the normalized SSITKA-MS signals
(F) for Ne (tracer gas), 13CO(g) and 13CO2(g), as well as the normalized IR
228 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
Fig. 21 (A) IR spectra of 2.3 wt% Au/TiO2 at steady-state WGS at 120 1C (a) for the nonlabelled feed: 6.8% CO/11% H2O/37.5% H2/Ar/He (bold red solid line) and for the corresponding 13CO labelled feed, (b) 2 s after the SSITKA switch (bold blue dashed line), and (c)
120 s after the SSITKA switch (pink solid line). (B) Comparison of the relative exchange of
the gas phase (normalized MS signal, F) for Ne, 13CO(g), 13CO2(g) and surface 13CO species.
The 13CO adsorption on Auo (dotted line, 2052 cm1) and 13CO adsorption on Aud
(dashed line, 1992 cm1) are also shown (Adopted from Wang et al.26 Copyright 2012
Elsevier).
band area due to adsorbed CO on Auo and Aud as a function of time. It is

noted that both kinds of adsorbed CO exchanged faster than gas-phase
13
CO2 but slower than gas-phase 13CO. This indicates that at least one of
these adsorbed CO species should be considered as an active precursor of
CO2 formation.
On the basis of eqn (8) and the transient responses of Ne and 13CO2,
the amount (mols) of active reaction intermediates per mol of Au atoms
was estimated. This was found to increase with increasing temperature
from 120 to 140 1C (0.74 vs. 0.99 mols COx/mols Au (total)). Also, the
amount of surface CO during WGS was estimated based on the area
Catalysis, 2016, 28, 175236 | 229
View Online
17:24:10.
13
between Ne and CO(g) (Fig. 21B) after accounting for the conversion of
CO by WGS. The result was that B5% of total Au was covered by CO,
which was found to be very close to the amount of Au corner atoms (B3%
of total Au), similar to that estimated for cubo-octahedral particles; the
average Au particle size measured by TEM was 3.0 0.8 nm.26 On the
basis of the above results, the number of active sites for WGS on
the 2.3 wt% Au/TiO2 catalyst investigated was most closely related to the
amount of corner Au atoms. As shown in Fig. 21B, CO adsorbed on Auo
exchanged with 13CO over a similar time scale as CO adsorbed on Aud.
The authors have concluded that although both kinds of adsorbed CO
might be considered active intermediates in the C-path of WGS (SSITKA
results), steady-state kinetics revealed that the surface active species are
metallic and not charged Au atoms. Furthermore, less than 1% of the
total metal Au atoms are responsible for most of the catalytic activity,
where corner Auo atoms are the dominant active sites for WGS over Au/
TiO2 catalysts, while perimeter or surface Au atoms are unlikely to
function as active sites. The change in the surface coverage of CO adsorbed on Auo was found to best correlate with the change in WGS kinetic
rate (deactivation), whereas CO adsorbed on Aud and carbonate and
formate species did not, strongly suggesting that these species are
spectators.26 It was concluded that deactivation of Au/TiO2 occurred
mainly due to Au particles sintering.
In isotopic work similar to that described above, Ribeiro and coworkers,27 using poisoning by Br of Au atoms on the same Au/TiO2
catalyst, have concluded that low coordination metallic corner Au atoms
are the active WGS sites. They suggested that in order to further improve
the catalytic rate for Au catalysts, it is important first to increase the
percentage of low coordination Au atoms either by synthesizing smaller
size or dierently shaped nanoparticles, or by developing supported Au
species with AuAu coordination less than four.27
Conclusions
It has been demonstrated by several case studies that SSITKA-operando

methodology and several kinds of transient isotopic tracing experiments
(use of 13CO, D2 and D2O), and titration with H2O, are powerful tools to
investigate rival mechanisms of the watergas shift reaction and estimate
important kinetic parameters such as the chemical composition and
surface concentration of active reaction intermediates found in the
carbon- and hydrogen-paths. The bifunctional nature of WGS over Pt
supported on reducible (e.g. CeO2, TiO2, La2O3, Ce1xMxO2d) and non
reducible (e.g. g-Al2O3, ZrO2) metal oxides is proved unequivocally after
the correct in situ estimation of the concentration (mmol g1) of active
reaction intermediates. When the latter kinetic parameter takes values
larger than the equivalent of one surface monolayer of Pt (or Au), this
means that the catalyst support provides active sites within a region
(reactive zone) around the Pt (or Au) nanoparticles, including the metal
support interface. Recent DFT computations along with experimental
evidence strongly support the presence of such active sites at the metal
230 | Catalysis, 2016, 28, 175236
17:24:10.
View Online
(Pt or Au)support interface. Quantification of the rate of depletion of

adsorbed formate (H12COO-s) towards CO2 formation (in terms of mol
g1 s1 after calibration of the integral absorbance IR signal) during
SSITKA-operando (12CO/H2O-13CO/H2O) and comparison with the
steady-state WGS kinetic rate, is a powerful method to determine the
kinetic importance of formate in the associative WGS reaction mechanism. Advanced micro-kinetic modelling of the various experimental
transient isotopic response curves obtained in SSITKA-operando (Mass
Spectrometer and IR) remains the best approach for discriminating rival
mechanisms (redox, or associative formate and/or carboxyl) in WGS for
practical catalytic systems. The latter approach uses a large number of
experimental data from the dynamic rate measurements through both
the gas and adsorbed phases without disturbing the on-going overall
steady-state catalytic rate. This approach had not yet been implemented
in WGS despite its great success in other heterogeneous catalytic reaction
systems.
Acknowledgements
The author would like to thank his former PhD students: Dr Kyriaki
Polychronopoulou, Dr George G. Olympiou, Dr Christos M. Kalamaras
and Dr Klito C. Petallidou for their contribution and excellent collaboration towards the advancement of fundamental knowledge in WGS over
supported metal catalysts. The financial support of the Research
Committee of the University of Cyprus and Cyprus Research Promotion Foundation through the grants PENEK/ENISX/0308/50 and
TEXNOLOGIA/YLIKA/0308(BE) is greatly appreciated.
References
1
2
3
D. S. Newsome, Catal. Rev.: Sci. Eng., 1980, 21, 275.

C. Ratnasamy and J. P. Wagner, Catal. Rev., 2009, 51, 325.
S. Chernik, R. French, C. Feik and E. Chornet, Ind. Eng. Chem. Res., 2002,
41, 4209.
4 R. D. Cortright, R. R. Davda and J. A. Dumesic, Nature, 2002, 418, 964.
5 Q. Fu, H. Saltsburg and M. Flytzani-Stephanopoulos, Science, 2003, 301, 935.
6 A. C. Basagiannis and X. E. Verykios, Appl. Catal., A, 2006, 308, 182.
7 T. L. LeValley, A. R. Richard and M. Fan, Int. J. Hydrogen Energy, 2014,
39, 16983.
8 K. Polychronopoulou, C. M. Kalamaras and A. M. Efstathiou, Recent Pat.
Mater. Sci., 2011, 4, 122.
9 C. Bosch and W. Wild, Hydrogen production, Canadian patent 153379,
1914.
10 S. A. Topham, Catalysis, Science and Technology, ed. J. R. Anderson and
M. Boudart, Springer Verlag, Berlin, 1985, vol. 7, pp. 150.
11 A. P. G. Kieboom, J. A. Moulijn, R. A. Sheldon and P. W. N. M. van Leeuwen,
Studies in Surface Science and Catalysis, ed. R. A. van Santen, P. W. N. M. van
Leeuwen, J. A. Moulijn and B. A. Averill, Elsevier Science B.V., Amsterdam,
1999, vol. 123, ch. 2, pp. 5761.
12 C. N. Satterfield, Heterogeneous Catalysis in Industrial Practice, McGraw-Hill,
New York, 2nd edn, 1991.
Catalysis, 2016, 28, 175236 | 231
View Online
13
17:24:10.
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
C. H. Bartolomew and R. J. Farrauto, Fundamentals of Industrial Catalytic

Processes, ed. C. H. Bartolomew and R. J. Farrauto, John Wiley & Sons,
Hoboken, New Jersey, 2006, p. 370.
M. T. Gencoglu and Z. Ural, Int. J. Hydrogen Energy, 2009, 34, 5242.
A. Qi, B. Peppley and K. Karan, Fuel Process. Technol., 2007, 88, 3.
V. A. Deshpande and C. Krause, Fuel Processor for Producing a Hydrogen Rich
Gas, Texaco, USA, 2007.
R. Farrauto, S. Hwang, L. Shore, W. Ruettinger, J. Lampert, T. Giroux, Y. Liu
and O. Ilinich, Annu. Rev. Mater. Res., 2003, 33, 1.
D. Andreeva, V. Idakiev, T. Tabakova, A. Andreev and R. Giovanoli, Appl.
Catal., A, 1996, 134, 275.
D. Andreeva, V. Idakiev, T. Tabakova and A. Andreev, J. Catal., 1996,
158, 354.
Q. Fu, A. Weber and M. Flytzani-Stephanopoulos, Catal. Lett., 2001, 77, 87.
X. Wang, J. A. Rodriguez, J. C. Hanson, M. Perez and J. Evans, J. Chem. Phys.,
2005, 123, 221101.
G. Jacobs, P. M. Patterson, L. Williams, E. Chenu, D. Sparks, G. Thomas and
B. H. Davis, Appl. Catal., A, 2004, 262, 177.
D. Tibiletti, A. Amieiro-Fonseca, R. Burch, Y. Chen, J. M. Fisher, A. Goguet,
C. Hardacre, P. Hu and D. Thompsett, J. Phys. Chem. B, 2005, 109, 22553.
R. Burch, Phys. Chem. Chem. Phys., 2006, 8, 5483.
rez, Science, 2007,
J. A. Rodriguez, S. Ma, P. Liu, J. Hrbek, J. Evans and M. Pe
318, 1757.
J. Wang, V. F. Kispersky, W. N. Delgass and F. H. Ribeiro, J. Catal., 2012,
289, 171.
M. Shekhar, J. Wang, W.-S. Lee, M. C. Akatay, E. A. Stach, W. N. Delgass and
F. H. Ribeiro, J. Catal., 2012, 293, 94.
J. A. Rodriguez, J. C. Hanson, D. Stacchiola and S. D. Senanayake, Phys.
Chem. Chem. Phys., 2013, 15, 12004, and references therein.
F. Tao and Z. Ma, Phys. Chem. Chem. Phys., 2013, 15, 15260.
M. Flytzani-Stephanopoulos, Acc. Chem. Res., 2014, 47, 783.
J. Nakamura, J. M. Campbell and C. T. Campbell, J. Chem. Soc., Faraday
Trans., 1990, 86, 2725.
A. A. Gokhale, J. A. Dumesic and M. Mavrikakis, J. Am. Chem. Soc., 2008,
130, 1402.
D. W. Flaherty, W.-Y. Yu, Z. D. Pozun, G. Henkelman and C. B. Mullins,
J. Catal., 2011, 282, 278.
A. F. Ghenciu, Curr. Opin. Solid State Mater. Sci., 2002, 6, 389.
A. F. Ghenciu, Fuel Cell Magazine, 2003, April/May.
G. Jacobs and B. H. Davis, Catalysis, 2007, 20, 122.
C. M. Kalamaras, D. D. Dionysiou and A. M. Efstathiou, ACS Catal., 2012,
2, 2729.
C. M. Kalamaras, K. C. Petallidou and A. M. Efstathiou, Appl. Catal., B, 2013,
136137, 225.
K. C. Petallidou, C. M. Kalamaras and A. M. Efstathiou, Catal. Today, 2014,
228, 183.
K. C. Petallidou, S. Boghosian and A. M. Efstathiou, Catal. Today, 2015,
242, 153.
A. Bruix, J. A. Rodriguez, P. J. Ramirez, S. D. Senanayake, J. Evans, J. B. Park,
D. Stacchiola, P. Liu, J. Hrbek and F. Illas, J. Am. Chem. Soc., 2012, 134,
8968.
J. A. Rodriguez, P. Liu, J. Hrbek, J. Evans and M. Perez, Angew. Chem., Int.
Ed., 2007, 46, 1329.
232 | Catalysis, 2016, 28, 175236
View Online
43
44
17:24:10.
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
Y. Zhai, D. Pierre, R. Si, W. Deng, P. Ferrin, A. U. Nilekar, G. Peng,

J. A. Herron, D. C. Bell, H. Saltsburg, M. Mavrikakis and M. FlytzaniStephanopoulos, Science, 2010, 329, 1633.
B. Zugic, S. Zhang, D. C. Bell, F. Tao and M. Flytzani-Stephanopoulos, J. Am.
Chem. Soc., 2014, 136, 3238.
J. H. Pazmino, M. Shekhar, W. D. Williams, M. C. Akatay, J. T. Miller,
W. N. Delgass and F. H. Ribeiro, J. Catal., 2012, 286, 279.
H. Xie, J. Lu, M. Shekhar, J. W. Elam, W. N. Delgass, F. H. Ribeiro, E. Weitz
and K. R. Poeppelmeir, ACS Catal., 2013, 3, 61.
P. Panagiotopoulou and D. I. Kondarides, J. Catal., 2009, 267, 57.
P. Panagiotopoulou and D. I. Kondarides, Appl. Catal., B, 2011, 101, 738.
J. M. Pigos, C. J. Brooks, G. Jacobs and B. H. Davis, Appl. Catal., A, 2007,
319, 47.
J. M. Pigos, C. J. Brooks, G. Jacobs and B. H. Davis, Appl. Catal., A, 2007, 328, 14.
A. M. Efstathiou and X. E. Verykios, Appl. Catal., A, 1997, 151, 109.
A. M. Efstathiou, J. T. Gleaves and G. S. Yablonsky, Characterization of Solid
Materials: From Structure to Surface Reactivity, ed. M. Che and J. C. Vedrine,
Wiley-VCH, Weinheim, Germany, 2012, vol. 2, ch. 22, pp. 10131073.
ares, Catal. Today, 2005, 100, 71.
M. A. Ban
G. A. Somorjai, Chem. Rev., 1996, 96, 1223.
Catalysis Today, ed. R. Burch, 1991, vol. 9 (12).
J. W. Niemansverdriet, Spectroscopy in Catalysis, VCH, New York, 1993.
S. L. Shannon and J. G. Goodwin Jr., Chem. Rev., 1995, 95(3), 677.
zinger, J. Raman Spectrosc., 2002, 33, 325.
S. Kuba and H. Kno
G. Srinivas, S. S. C. Chuang and M. W. Balakos, AIChE J., 1993, 39(3), 530.
S. S. C. Chuang, M. A. Brundage, M. W. Balakos and G. Srinivas, Appl.
Spectrosc., 1995, 49(8), 1151.
F. C. Meunier, Chem. Soc. Rev., 2010, 39, 4602.
D. Duprez, Isotopes in Heterogeneous Catalysis, ed. J. S. J. Hargreaves,
S. D. Jackson and G. Webb, Catalytic Science Series, Imperial College Press,
London, 2006, vol. 4, pp. 133182.
C. Ledesma, J. Yang, D. Chen and A. Holmen, ACS Catal., 2014, 4, 4527.
C. Mirodatos, Catal. Today, 1991, 9, 83.
B. S. Balzhinimaev, E. M. Sadovskaya and A. P. Suknev, Chem. Eng. J., 2009,
154, 2.
drine, Appl. Catal., A, 2014, 474, 40.
J. C. Ve
T. Gott and S. Ted Oyama, J. Catal., 2009, 263, 359.
J. Happel, Chem. Eng. Sci., 1978, 33, 1567.
J. Happel, Isotopic Assessment of Heterogeneous Catalysis, Academic Press,
New York, 1986.
C. O. Bennett, ACS Symp. Ser., 1982, 178, 1.
C. O. Bennett, Adv. Catal., 1999, 44, 329.
P. Biloen, J. Mol. Catal., 1983, 21, 17.
M. W. Balakos, S. S. C. Chuang and G. Srinivas, J. Catal., 1993, 140, 281.
K. Tamaru, Adv. Catal., 1964, 15, 65.
C. M. Kalamaras, G. G. Olympiou and A. M. Efstathiou, Catal. Today, 2008,
138, 228.
C. M. Kalamaras, P. Panagiotopoulou, D. I. Kondarides and
A. M. Efstathiou, J. Catal., 2009, 264, 117.
C. M. Kalamaras, I. D. Gonzalez, R. M. Navarro, J. L. G. Fierro and
A. M. Efstathiou, J. Phys. Chem. C, 2011, 115, 11595.
C. M. Kalamaras, S. Americanou and A. M. Efstathiou, J. Catal., 2011,
279, 287.
Catalysis, 2016, 28, 175236 | 233
View Online
79
80
81
17:24:10.
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
M. Stamatakis, Y. Chen and D. G. Vlachos, J. Phys. Chem. C, 2011,

115, 24750.
J. P. Clay, J. P. Greeley, F. H. Ribeiro, W. N. Delgass and W. F. Schneider,
J. Catal., 2014, 320, 106.
B. Smith, M. Loganathany and S. S. Shanthaz, Int. J. Chem. React. Eng., 2010,
8, 1.
P. Liu, New and Future Developments in Catalysis: Hybrid Materials, Composites and Organocatalysts, ed. S. L. Suib, Elsevier, 2013, ch. 9, pp. 213238.
D. G. Vlachos, A. B. Mhadeshwar and N. S. Kaisare, Comput. Chem. Eng.,
2006, 30, 1712.
M. Maestri, D. Livio, A. Beretta and G. Groppi, Ind. Eng. Chem. Res., 2014,
53, 10914.
C.-H. Lin, C.-L. Chen and J.-H. Wang, J. Phys. Chem. C, 2011, 115, 18582.
G. Bond, Gold Bull., 2009, 42, 337.
J. B. Park, J. Graciani, J. Evans, D. Stacchiola, S. D. Senanayake, L. Barrio,
P. Liu, J. F. Sanz, J. Hrbek and J. A. Rodriguez, J. Am. Chem. Soc., 2010,
132, 356.
L. C. Grabow, A. A. Gokhale, S. T. Evans, J. A. Dumesic and M. Mavrikakis,
J. Phys. Chem. C, 2008, 112, 4608.
D. W. Flaherty, W.-Y. Yu, Z. D. Pozun, G. Henkelman and C. B. Mullins,
J. Catal., 2011, 282, 278.
S.-K. Wu, R.-J. Lin, S. Jang, H.-L. Chen, S.-M. Wang and F.-Y. Li, J. Phys.
Chem. C, 2014, 118, 298.
S. C. Ammal and A. Heyden, J. Catal., 2013, 306, 78.
S. Aranifard, S. C. Ammal and A. Heyden, J. Phys. Chem. C, 2014, 118, 6314.
S. C. Ammal and A. Heyden, ACS Catal., 2014, 4, 3654.
S. Aranifard, S. C. Ammal and A. Heyden, J. Catal., 2014, 309, 314.
W. Song and E. J. M. Hensen, ACS Catal., 2014, 4, 1885.
G. G. Olympiou, C. M. Kalamaras, C. D. Zeinalipour-Yazdi and
A. M. Efstathiou, Catal. Today, 2007, 127, 304.
A. M. Efstathiou and K. C. Petallidou, Appl. Catal., B, 2014, 152153, 439.
G. Jacobs, S. Ricote, U. M. Graham and B. H. Davis, Catal. Lett., 2015,
145, 533.
G. Jacobs and B. H. Davis, Int. J. Hydrogen Energy, 2010, 35, 3522.
G. Jacobs, S. Khalid, P. M. Patterson, D. E. Sparks and B. H. Davis, Appl.
Catal., A, 2004, 268, 255.
G. Jacobs and B. H. Davis, Appl. Catal., A, 2007, 333, 192.
A. Goguet, F. C. Meunier, D. Tibiletti, J. P. Breen and R. Burch, J. Phys. Chem.
B, 2004, 108, 20240.
F. C. Meunier, D. Tibiletti, A. Goguet, D. Reid and R. Burch, Appl. Catal., A,
2005, 289, 104.
D. Tibiletti, A. Goguet, D. Reid, F. C. Meunier and R. Burch, Catal. Today,
2006, 113, 94.
F. C. Meunier, D. Reid, A. Goguet, S. Shekhtman, C. Hardacre, R. Burch,
W. Deng and M. Flytzani-Stephanopoulos, J. Catal., 2007, 247, 277.
F. C. Meunier, D. Tibiletti, A. Goguet, S. Shekhtman, C. Hardacre and
R. Burch, Catal. Today, 2007, 126, 143.
F. C. Meunier, A. Goguet, S. Shekhtman, D. Rooney and H. Daly, Appl.
Catal., A, 2008, 340, 196.
F. C. Meunier, Catal. Today, 2010, 155, 164.
R. Burch, A. Goguet and F. C. Meunier, Appl. Catal., A, 2011, 409410, 3.
J. Vecchietti, A. Bonivardi, W. Xu, D. Stacchiola, J. J. Delgado, M. Calatayud
and S. E. Collins, ACS Catal., 2014, 4, 2088.
234 | Catalysis, 2016, 28, 175236
View Online
111
112
17:24:10.
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
K. Mudiyanselage, S. D. Senanayake, L. Feria, S. Kundu, A. E. Baber,

J. Graciani, A. B. Vidal, S. Agnoli, J. Evans, R. Chang, S. Axnanda, Z. Liu,
J. F. Sanz, P. Liu, J. A. Rodriguez and D. J. Stacchiola, Angew. Chem., Int. Ed.,
2013, 52, 5101.
B. Zugic, S. Zhang, D. C. Bell, F. Tao and M. Flytzani-Stephanopoulos,
J. Am. Chem. Soc., 2014, 136, 3238.
M. Yang, J. Liu, S. Lee, B. Zugic, J. Huang, L. F. Allard and M. FlytzaniStephanopoulos, J. Am. Chem. Soc., 2015, 137, 3470.
T. Ishida, N. Kinoshita, H. Okatsu, T. Akita, T. Takei and M. Haruta, Angew.
Chem., Int. Ed., 2008, 47, 9265.
Z. Zhou, S. Kooi, M. Flytzani-Stephanopoulos and H. Saltsburg, Adv. Funct.
Mater., 2008, 18, 2801.
M. Haruta, Gold Bull., 2004, 37, 27.
M. Boaro, M. Vicario, J. Llorca, C. de Leitenburg, G. Dolcetti and
A. Trovarelli, Appl. Catal., B, 2009, 88, 272.
K. G. Azzam, I. V. Babich, K. Seshan and L. Leerts, J. Catal., 2007, 251, 153.
R. J. Gorte and S. Zhao, Catal. Today, 2005, 104, 18.
K. Li, Q. Fu and M. Flytzani-Stephanopoulos, Appl. Catal., B, 2000, 27, 179.
Z.-P. Liu, S. J. Jenkins and D. A. King, Phys. Rev. Lett., 2005, 94, 196102.
H. S. Taylor, J. Phys. Chem., 1926, 30, 145.
T. Shido and Y. Iwasawa, J. Catal., 1992, 136, 493.
T. Shido and Y. Iwasawa, J. Catal., 1993, 141, 71.
G. Jacobs, A. C. Crawford and B. H. Davis, Catal. Lett., 2005, 100, 147.
F. Zaera, Chem. Soc. Rev., 2014, 43, 7624.
J. Sirita, S. Phanichphant and F. C. Meunier, Anal. Chem., 2007, 79, 3912.
H. Li, M. Rivallan, F. Thibault-Starzyk, A. Travert and F. C. Meunier, Phys.
Chem. Chem. Phys., 2013, 15, 7321.
A. Paredes-Nunez, I. Jbir, D. Bianchi and F. C. Meunier, Appl. Catal., A, 2015,
495, 17.
J. Couble and D. Bianchi, Appl. Catal., A, 2011, 409, 28.
R. W. Stevens and S. S. C. Chuang, J. Phys. Chem. B, 2004, 108, 696.
F. C. Meunier, A. Goguet, C. Hardacre, R. Burch and D. Thompsett, J. Catal.,
2007, 252, 18.
J. Ryczkowski, Catal. Today, 2001, 68, 263.
K. Hadjiivanov, M. Mihaylov, D. Panayotov, E. Ivanova and K. Chakarova,
Spectrosc. Prop. Inorg. Organomet. Compd., 2014, 45, 43.
C. Li, Y. Sakata, T. Arai, K. Domen, K.-I. Maruya and T. Onishi, J. Chem. Soc.,
Faraday Trans. 1, 1989, 85, 929.
P. Bazin, O. Saur, J. C. Lavalley, M. Daturi and G. Blanchard, Phys. Chem.
Chem. Phys., 2005, 7, 187.
M. Primet, J. Catal., 1984, 88, 273.
B. C. Smith, Fundamentals of Fourier Transform Infrared Spectroscopy, CRC
Press, 1996.
G. N. Vayssilov, M. Mihaylov, P. St. Petkov, K. I. Hadjiivanov and
K. M. Neyman, J. Phys. Chem. C, 2011, 115, 23435.
s and A. L. Bonivardi, Phys. Chem.
M. Calatayud, S. E. Collins, M. A. Baltana
Chem. Phys., 2009, 11, 1397.
C. Binet, A. Badri, M. Boutonnet-Kizlingt and J.-C. Lavalley, J. Chem. Soc.,
Faraday Trans., 1994, 90, 1023.
F. Romero-Sarria, L. M. Martinez, T. Miguel, A. Centeno and J. A. Odriozola,
J. Phys. Chem. C, 2007, 111, 14469.
B. E. Hayden, K. Prince, D. P. Woodru and A. M. Bradshaw, Phys. Rev. Lett.,
1983, 51, 475.
Catalysis, 2016, 28, 175236 | 235
View Online
144
17:24:10.
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
G. Jacobs, U. M. Graham, E. Chenu, P. M. Patterson, A. Dozier and

B. H. Davis, J. Catal., 2005, 229, 499.
G. Jacobs, S. Ricote and B. H. Davis, Appl. Catal., A, 2006, 302, 14.
C. D. Zeinalipour-Yazdi and A. M. Efstathiou, J. Phys. Chem. C, 2008,
112, 19030.
G. Busca and V. Lorenzelli, Mater. Chem., 1982, 7, 89.
C. N. Costa, T. Anastasiadou and A. M. Efstathiou, J. Catal., 2000, 194, 250.
spar, P. Fornasiero, M. Graziani and M. S. Islam, J. Phys.
G. Balducci, J. Ka
Chem. B, 1998, 102, 557.
M. C. Ribeiro, G. Jacobs, L. Linganiso, K. G. Azzam, U. M. Graham and
B. H. Davis, ACS Catal., 2011, 1, 1375.
X. Q. Gong, P. Hu and R. Raval, J. Chem. Phys., 2003, 119, 6324.
J. A. Rodriguez, J. Evans, J. Graciani, J.-B. Park, P. Liu, J. Hrbek and
J. F. Sanz, J. Phys. Chem. C, 2009, 2, 7364.
D. Tibiletti, F. C. Meunier, A. Goguet, D. Reid, R. Burch, M. Boaro,
M. Vicario and A. Trovarelli, J. Catal., 2006, 244, 183.
P. O. Graf, D. J. M. de Vlieger, B. L. Mojet and L. Leerts, J. Catal., 2009,
262, 181.
S. T. Korhonen, M. Calatayud and A. O. I. Krause, J. Phys. Chem. C, 2008,
112, 6469.
A. Ignatchenko, D. G. Nealon, R. Dushane and K. Humphries, J. Mol. Catal.
A: Chem., 2006, 256, 57.
G. Jacobs, U. M. Graham, E. Chenu, P. M. Patterson, A. Dozier and
B. H. Davis, J. Catal., 2005, 229, 499.
G. Jacobs, S. Ricote, U. M. Graham, P. M. Patterson and B. H. Davis, Catal.
Today, 2005, 106, 259.
G. Jacobs, P. M. Patterson, U. M. Graham, A. C. Crawford and B. H. Davis,
Int. J. Hydrogen Energy, 2005, 30, 1265.
G. Jacobs, P. M. Patterson, U. M. Graham, A. C. Crawford, A. Dozier and
B. H. Davis, J. Catal., 2005, 235, 79.
G. Jacobs, P. M. Patterson, U. M. Graham, D. E. Sparks and B. H. Davis, Appl.
Catal., A, 2004, 269, 63.
D. Duprez, Catal. Today, 2006, 112, 17.
G. Jacobs, S. Ricote, P. M. Patterson, U. M. Graham, A. Dozier, S. Khalid,
E. Rhodus and B. H. Davis, Appl. Catal., A, 2005, 292, 229.
C. Binet, M. Daturi and J. C. Lavalley, Catal. Today, 1999, 50, 207.
Y. Chen, H. Wang, R. Burch, C. Hardacre and P. Hu, Faraday Discuss., 2011,
152, 121.
N. S. Govender, F. G. Botes, M. H. J. M. de Croon and J. C. Schouten,
J. Catal., 2008, 260, 254.
M. Che and C. O. Bennett, Adv. Catal., 1989, 36, 55.
H. Iida and A. Igarashi, Appl. Catal., A, 2006, 298, 152.
R. Leppelt, B. Schumacher, V. Plzak, M. Kinne and R. J. Behm, J. Catal.,
2006, 244, 137.
A. Goguet, F. C. Meunier, J. P. Breen, R. Burch, M. I. Petch and
A. F. Ghenciu, J. Catal., 2004, 226, 382.
H. Daly, A. Goguet, C. Hardacre, F. C. Meunier, R. Pilasombat and
D. Thompsett, J. Catal., 2010, 273, 257.
C. H. Kim and L. T. Thompson, J. Catal., 2005, 230, 66.
236 | Catalysis, 2016, 28, 175236
Recent progresses on the use of

supported bimetallic catalysts for the
preferential oxidation of CO (PROX)
17:24:19.
Alina Moscu,* Yves Schuurman and Frederic C. Meunier*

DOI: 10.1039/9781782626855-00237
The preferential oxidation of CO in the presence of a large excess of H2 can be used as a

purification technique to obtain PEM fuel cell-grade H2. Improvements of the catalyst
performances at low temperature are yet required in terms of specific activity, selectivity
and stability. Several bimetallic compounds were shown to exhibit much improved PROX
performances as compared to the monometallic counterparts. The improved activity of
alloys and bimetallic can be due to geometric or electronic eects. The present review
discusses these matters giving a special attention to in situ IR-based studies realized over
Pt-based formulations, since CO is both a reactant and a molecular probe enabling the
determination of the state of metals under reaction conditions. The need to study the
catalysts under relevant operating conditions is emphasized. In situ and operando conditions enable determining the nature of the true active phases, because alloy segregation
can readily occur even upon some minor modification of the experimental conditions.
1 Benefits and challenges associated with the use of

multi-metallic materials as catalysts and electrodes
Supported metallic particles catalyse many reactions of industrial and
environmental importance.1 Bi- and multi-metallic catalysts have raised a
lot of interest because of potential synergies between the elements
forming the multi-component catalysts, as those observed for instance in
the case of the preferential oxidation of CO in the presence of excess
hydrogen (PROX).25 The possibility to dilute expensive noble metals into
base metals while maintaining similar activities or improving durability
is also highly valuable.68
The catalytic activity in terms of reactant conversion rates and selectivity to various reaction pathways will directly depend on the surface
composition and local geometric arrangement of atoms.9 Traditional
methods based on volumetric titration of metallic sites (e.g. by CO, H2,
NO, N2O) and high resolution transmission electron microscopy are
routine ex situ methods used to characterise metal dispersion and particle size distribution. The use of in situ and operando methods have yet
stressed that the structure of the metallic particles can be dramatically
modified under reaction conditions,10,11 both in terms of geometry and
chemical composition. The likelihood of particle restructuring is even
higher when multi-metallic particles are used.
Marked evolutions of the structure and composition of catalytic materials were shown to occur under or after reaction, as for instance in the
Universite Lyon 1, CNRS, UMR 5256, IRCELYON, Institut de Recherches sur la
Catalyse et lEnvironnement de Lyon, 2 Avenue Albert Einstein, 69626 Villeurbanne,
France. E-mail: alina.moscu@gmail.com; fcm@ircelyon.univ-lyon1.fr
Catalysis, 2016, 28, 237267 | 237

c
17:24:19.
View Online
case of PtSn materials. PtSn-based catalysts exhibit improved activity,

selectivity and/or resistance to deactivation over the analogous Pt catalysts in hydrogenation12 and dehydrogenation13 reactions, and oxidation
of ethanol, methanol14 and carbon monoxide.15,16 The preferential oxidation of CO by O2 in the presence of excess H2 has been envisaged as a
purification route for ultra pure H2 to be used in PEM fuel cells and alloys
based on PtSn have demonstrated a high activity for this reaction.16
Somorjai and co-workers15 have shown that Sn was more readily oxidised
than Pt, leading to partly irreversible alloy segregation. A reversible segregation of PtSn alloys was also reported by Moscu et al. merely upon
lowering the temperature below ca. 175 1C under a CO/H2 stream.11
Bimetallics are also of interest for the development of new electrodes,
for instance to be used in low-temperature Proton Exchange Membrane
(PEM) fuel cells. The activation of O2 (oxygen reduction reaction ORR)
is considered as rate-limiting and electrodes made upon alloying Pt with
a 3d-transition metal (Co, Ni, Fe, Cu) appear to provide improved ORR
rates.17,18 Noble metals such as Pt have been alloyed with many base
metals with a view at decreasing the poisoning eect of CO, which is a
potent poison of H2 oxidation electrodes. Coreshell PtSn@Pt-type
structures were found to be considerably more resistant against CO
poisoning than plain Pt.19 The structure and stability of such bimetallics
is complex. In the case of PtCo alloys a CO-induced segregation was
observed, in which the surface became richer in Pt, supposedly due to the
stronger bond formed between CO and Pt.20 The gradual formation of
hollow platinum and platinum-cobalt particles was observed by ex situ
TEM analyses following the electrochemical reduction of O2. The presence of hollow nanoparticles was explained by the nanoscale Kirkendall
eect, a vacancy-mediated diusion mechanism in binary alloys in which
one metal diuses faster than the other.18
The examples reported above stress the need to monitor the structure
of metals under realistic operating conditions if a relevant description
of the active working surface is to be determined. Therefore, the use of
in situ methods (i.e. monitoring of the structure under controlled conditions) or, better, operando techniques (i.e. with simultaneous monitoring of the structure and the activity of the same sample) appear crucial
to better understand the structure of the active surface and ageing
processes.1722
This short review starts with briefing recalling the current interest in
the PROX reaction for clean H2 production, then highlights some of the
main features of supported bimetallic catalysts used for this reaction,
focussing on PtSn systems and finishing o with some recent reports
using in situ or operando technique enabling a direct characterisation of
the PROX catalyst under reaction conditions.
Hydrogen as energy carrier in fuel cell applications
The human demand of carbon-based energy resources is ever increasing

and the rate of fossil fuel consumption is superior to that of natural
fuel replenishment. The environmental impact of conventional energy
238 | Catalysis, 2016, 28, 237267
17:24:19.
View Online
systems is also significant through the emission of pollutants such as

CO, SOx, NOX, heavy metals and green-house gases such as N2O and
CO2. The use of truly renewable sources of energy would mitigate
high consumption of conventional fuels and decrease CO2 emissions.
During the last decades molecular hydrogen has gained worldwide interest as a substitute of fossil fuels.2325 The hydrogen to be used as an energy
carrier can be derived from various sources, which are intrinsically
cleaner, more renewable and less geopolitically sensitive than fossil fuels.
Instantaneous energy in the form of electricity and heat can be generated by reacting hydrogen and air in fuel cells. Fuel cells, operating
with hydrogen or hydrogen-rich fuels, are the major factors in catalysing
the transition to a future sustainable energy system with low CO2 emissions.26 Hydrogen as an energy carrier, requires optimised technologies
for production, transport, storage and usage.25,27 Hydrogen can be produced from fossil fuels and renewable fuels, enabling a smooth transition between the two types of feedstocks. Production from fossil fuels and
hydrocarbons refers to steam reforming of natural gas or oil, steam reforming of methanol, coal gasification and partial oxidation of hydrocarbons. The hydrogen production from renewable sources becomes
more and more interesting and includes the thermochemical process
from biomass, the thermochemical decomposition of water, photoelectrochemical conversion and water electrolysis.
The storage of hydrogen is an important issue in automotive applications, for which weight and volume should be kept as low as possible.
Dierent hydrogen storage options include: compressed hydrogen at
high pressure (up to 700 bars) and kept in glass micro balls, chemical
storage in metal hydrides or storage by means of carbon nanostructures.
Hydrogen utilisation as a fuel demands an ecient transportation
and distribution network, which must meet the usual requirements of
security and quick and easy supply stations. Currently, hydrogen is
delivered through pipeline networks or transported by road via trucks.
All these options increase the final cost of hydrogen for the end users.
Due to the high costs of hydrogen distribution and due to security
concerns (leakage and explosions), the development of a compact reformer with on-site hydrogen production will probably be more preferable than centralized distribution through pipelines. Nowadays, vehicles
based on hydrogen fuel cells are gaining interest and utilisation with a
view at atmospheric pollution reduction.
Fuel cells convert the chemical energy liberated during the electrochemical reaction of H2 and O2 into electrical energy with zero emission
of pollutants. Hydrogen (or a hydrogen containing source) and air are
introduced at the anode and cathode respectively, of the fuel cells, and
then the electrochemical reaction between H2 and O2 takes place. Lowtemperature proton exchange membrane fuel cells (PEMFC) oer higher
power density than other fuel cell systems.2830 Their main advantage is
the use of a solid polymer as an electrolyte and porous carbon electrodes
as a replacement of corrosive liquid electrodes. Proton exchange membrane fuel cells are constituted from a thin polymer membrane which
plays the role of the electrolyte and a separator between the anode
Catalysis, 2016, 28, 237267 | 239
17:24:19.
View Online
and cathode. They operate at low temperature (6090 1C) to lessen the
membrane humidification. Due to the low operating temperature,
hydrogen is the only fuel to be successfully converted electrochemically at
acceptable rates.
Hydrogen production from C-containing sources, such as natural gas,
coal, alcohols or biomass requires a reforming unit independent of the
PEMFC.31 The reformate obtained is constituted from hydrogen, CO2 and
CO. Due to the strong adsorption of CO on the electro catalysts, blocking
the catalyst sites, the electro oxidation of H2 which has to take place over
the free sites is impeded. CO concentrations in the range of tens of ppm
are sucient to generate significant voltage losses.29,32,33 The resistance
to CO poisoning could be alleviated by using fuel cells that operate at
higher temperatures.34
The existing hydrogen purification technologies are divided into
physical purification processes and chemical processes. Examples of
physical purification processes are membrane separation, pressure swing
adsorption and solvent adsorption. Chemical purification processes are
for instance water gas shift reaction (CO H2O2CO2 H2) and prefer1
ential oxidation of CO (PROX: CO O2-CO2), of which PROX could
2
also be used either internally at the fuel cell anode or in a separate external reactor. Internal preferential oxidation is an alternative method
where a small amount of oxygen is added to the fuel cell anode to oxidize
the CO.35,36 The advantage is reduction of PROX reactor volume and
possibly even the elimination of the PROX reactor. Short-term observations show negligible eect on the fuel cell stack performance. However, the exothermicity of CO oxidation on the anode appears to cause
sintering of the electrode catalyst and may result in accelerated aging of
the fuel cell membrane.
In the view of increased interest for transportation applications of the
PEM fuel cell, a hydrogen purification process to assure very low CO
concentration becomes paramount. The water-gas shift reaction followed
by the preferential oxidation of CO are currently the preferred methods
of hydrogen purification from on-board reformers. The flow diagram of
hydrogen production unit is represented in Fig. 1. The present review
focuses on the PROX and attempts to better understand this reaction
from the mechanistic and the kinetics point of views.
Main features of the preferential CO oxidation (PROX)
Low temperature water-gas shift (LT-WGS) reaction precedes the PROX

reaction and it is commonly used in the hydrogen production industry to
adjust the CO/H2 ratio in synthesis gas. The WGS process is represented
by the following reaction:
CO H2O2CO2 H2
DH1298K 41 kJ mol1
The preferential oxidation of CO (PROX) is used as a complement to

the WGS reaction to reach a CO concentration below 10 ppm. It is best
operated between 80 1C, the operating temperature of PEMFC, and 200 1C
240 | Catalysis, 2016, 28, 237267
17:24:19.
View Online
Fig. 1 Hydrogen production unit flow diagram.
which is the operating temperature of the water-gas shift unit. The CO

PROX process involves two competitive reactions:
1
CO O2-CO2 DH1298K 283 kJ mol1
2
1
H2 O2-H2O DH1298K 242 kJ mol1
2
The parallel hydrogen oxidation is an undesired reaction leading to the
consumption of hydrogen as 100% selectivity towards CO2 has not yet
been achieved. Minimum loss of hydrogen is controlled by adjusting the
O2 content at the inlet. In addition, the formed water decreases the
catalyst performance. The reactions involved are highly exothermic requiring good temperature control and heat removal.
3.1 Monometallic catalysts for PROX
PROX catalyst design is about enhancing the CO oxidation and inhibiting
the parallel oxidation of hydrogen. The main requirements for an ideal
PROX catalyst in order to achieve very low CO concentration (i.e. o10 ppm)
are to be highly active, stable, resistant towards deactivation by carbon
dioxide and water present in the feed and selective in the lower temperature range. The PROX catalysts should exhibit a high activity in a
wide operation temperature window (70200 1C). Moreover, the catalytic
performance at higher temperatures range (4250 1C) should not be
1
altered by side reactions such as H2 oxidation (H2 O2-H2O), reverse
2
Catalysis, 2016, 28, 237267 | 241
17:24:19.
View Online
Fig. 2 Scheme of PROX catalysts classification. (Reprinted with permission from ref. 62:
K. Liu, A. Wang and T. Zhang, ACS Catal., 2012, 2, 11651178. Copyright r 2012 American
Chemical Society.)
water-gas shift reaction (CO2 H2-CO H2O) and CO methanation

(CO 3H2-CH4 H2O). These reactions consume large amounts of
hydrogen and favor the steam generation, which can lead to catalyst
deactivation. PROX catalysts can be classified as illustrated in Fig. 2 into
the following groups:
Promoted or unpromoted supported noble catalysts (Platinum
Group Metal - PGM): Pt, Pd, Ir, Ru or Rh.3745
Nano-gold catalysts.4651
Cu-based catalysts supported on oxides with dierent acidic, basic
and redox properties (MgO, La2O3, CeO2, SiO2-Al2O3, etc.).5259
Gold-based catalysts have triggered a marked interest due to their high
activity in low temperature CO oxidation.60,61 However, the main unresolved issues of the gold catalysts are the low CO2 selectivity at high
temperature and the poor resistance towards CO2 and H2O.62 Moreover,
the catalytic performance is aected by the fast deactivation due to sintering during reaction.
Among copper-based catalysts, the CeO2-supported materials display
high activity above 100 1C. Nevertheless, these catalysts are not widely
used as PROX catalysts due to their low specific activity as compared to
that of noble metals. In addition, the hydrothermal instability and the
poor resistance towards CO2 of these oxides limit their application in the
PROX reaction, although those are cost-eective.63
Monometallic PGM catalysts on inert supports revealed very low activities in the low temperature range (o100 1C).40,41,64,65 Ru/Al2O3 and
242 | Catalysis, 2016, 28, 237267
View Online
17:24:19.
66
Rh/Al2O3 were reported to be very active at 100 1C. In spite of that, it has
been shown that on Ru0 the CO removal using CO oxidation is accompanied by CO methanation at higher temperatures. Methanation of CO2
could also occur via CO2 4H2-CH4 2H2O.42 Hence, the loss of
hydrogen appears when these catalysts are used and decreases their
interest.
Supported Pt catalysts have been thoroughly studied due to their CO2
and H2O-resistance.64,67 On the other hand, the disadvantages of these
catalysts are their low activity at low temperature (i.e. o120 1C) due to the
strong adsorption of CO on platinum and the unselective oxidation of
hydrogen at higher temperatures. Moreover, their high costs and limited
availability (the yearly world production of platinum could not cover the
potential demand if a massive production of fuel cells were considered)
restrict their application as commercial catalysts. However, a high dispersion of the metal onto the support can overcome this issue. It has
been shown that Pt is best dispersed on acidic oxides (SiO2, Al2O3,
SiO2-Al2O3)40,68 and additionally it reveals enhanced activity and CO2
selectivity at lower temperature by maximizing the number of active sites
involved in the CO removal.69
The key for improving Pt-based catalysts for CO oxidation at temperatures between 80 and 120 1C is either to weaken the CO adsorption
on the platinum sites in order to free up sites for oxygen adsorption or to
allow the possibility of O2 adsorption on a dierent site than Pt atoms,
leading to a non-competitive dual-site mechanism of CO and O2. In this
respect new catalysts are being developed by providing Pt catalysts with
an additional site for oxygen adsorption via addition of a second metal or
by using a support that can activate oxygen. For the last group Pt/CeO2 is
a good example showing much higher activity than Pt/Al2O3 or Pt/SiO2.
A particular attention will be paid to the bimetallic catalysts based on Pt
in the following section.
3.2 Bimetallic Pt-containing catalysts for PROX

Bimetallic catalysts, which display dierent electronic and chemical
properties as compared to their monometallic counterpart, oer the
opportunity to obtain new catalysts with enhanced activity, selectivity and
stability. Bimetallic catalysts, can exhibit a large variety in structure and
dierent phases can coexist. Thus, the apprehension of the structureactivity relationship is an ongoing research activity in numerous research
groups. The understanding of the correlation between the structure/
composition and the catalytic activity requires the synthesis of wellcontrolled nanoparticles that display narrow size distribution, homogeneous shape and composition. Achieving well-defined bimetallic
nanoparticles (NPs) through conventional synthesis routes (i.e., incipient
wetness impregnation, co-impregnation, etc.) is dicult. However, over
the last decade, advanced synthesis approaches that allow the control of
these NPs characteristics have been elaborated. For example, recently
developed colloidal synthesis allows control of the formation of welldefined bimetallic catalysts.
Catalysis, 2016, 28, 237267 | 243
17:24:19.
View Online
Fig. 3 Possible structures of bimetallic NPs induced by pretreatment or reaction conditions. (Reprinted with permission from ref. 62: K. Liu, A. Wang and T. Zhang, ACS Catal.,
2012, 2, 11651178. Copyright r 2012 American Chemical Society.)
The use of this synthesis technique allows formation of bimetallic

catalysts that display a wide variety of dierent structures such as alloys
or coreshell nanoparticles, which yet may be prone to structural
modifications under reaction conditions. As illustrated in Fig. 3,
structural changes can appear as a response to changes to the temperature or the reacting atmosphere15,70 leading to modifications
of the catalytic properties. Hence, bimetallic catalysts can display
dierent surface composition with respect to the bulk and furthermore
those can undergo complete segregation of the metals under reaction
conditions.
Studies on bimetallic Pt-containing catalysts for PROX reaction are
numerous. Bimetallic catalysts such as PtCo,71,72 AuPt,73 PtCu,7476
PtPd,77,78 PtFe,79 PtNi,80 PtSn,15,48 RuPt,81 RhPt,82 PtMn83 are
reported in the literature and grouped in several reviews.62,84 In all cases
superior performance of the bimetallic catalyst as compared to that of the
pure Pt were reported. Table 1 gives a non-exhaustive list of the bimetallic
Pt catalysts supported on inert supports and their catalytic performance
in PROX. However, a multitude of dierent experimental conditions
makes the comparison between catalysts dicult.
The superior activity of bimetallic Pt-based catalysts in PROX obviously
derives from the new properties of the multi-metallic material, the second metal possibly altering the properties of Pt through two main eects:
(i) a geometric eect, where the number of Pt atoms is dispersed in
244 | Catalysis, 2016, 28, 237267
17:24:19.
Table 1 Dierent bimetallic Pt catalysts and their activities in PROX.

Catalyst
Composition
PtCo/CNT
4 wt% Pt
0.7 wt% Co
4 wt% Pt
0.7 wt% Co
2.54 wt% Pt
0.85 wt% Cu
2.8 wt% Pt
0.014 wt% Co
4 wt% Pt
0.20 wt% Fe
1 wt% Pt
0.11 wt% Fe
1 wt% Pt
1.5 wt% Ni
1 wt% Pt
1 wt% Ru
2 wt% Pt
1 wt% Re
0.7 wt% Pt
0.22 wt% Sn
0.38 wt% Pt
10.5 wt% Mn
PtCo/g-Al2O3
PtCu/g-Al2O3
PtCo/g-Al2O3
PtFe/C
PtFe/SiO2
PtNi/g-Al2O3
Catalysis, 2016, 28, 237267 | 245
PtRu/SiO2
PtReOx/SiO2
Pt3Sn/SBA-15
PtMnOx/Al2O3
CO (1C)
Max. CO conversion (%)
Feed
Ref.
40
100
1% CO, 1% O2, 50% H2, N2 balance
71
100
100
1% CO, 1% O2, 50% H2, N2 balance
71
120
50
1% CO, 0.5% O2, 67% H2, N2 balance
75
80
60
1% CO, 0.5% O2, 67% H2, N2 balance
75
40
100
1% CO, 0.5% O2, 50% H2, N2 balance
77
110
95
0.5% CO, 0.5% O2, 45% H2, N2 balance
85
77
90
1% CO, 1% O2, 80% H2, 2% H2O, He balance
80
120
100
0.5% CO, 0.5% O2, 45% H2, N2 balance
81
110
100
1% CO, 1.25% O2, 60% H2, He balance
86
170
100
2% CO, 2% O2, 48% H2, He balance
87
140
90
1% CO, 1.5% O2, 60% H2, He balance
83
at max
17:24:19.
View Online
between the second metal atoms and (ii) an electronic eect, where the
electronic structure of Pt is modified.
The enhanced activity in PROX due to the geometric eect is ascribed
to an increased concentration of catalytic sites. The bimetallic catalyst
could also possibly works as dual active sites where CO adsorbed on Pt
reacts with oxygen activated by the second metal, which may therefore act
as a catalytic centre. Hence detailed kinetic investigations would be required to determine the full role of the added metal.
In this context, Zhang and coworkers88 investigated a PtFe/C system
in PROX and concluded by means of ICP-AES, XPS, XRD, TEM, HS-LEIS
that the sample adopts a Fe-rich structure with a Pt core. The extremely
high activity at low temperature is the eect of O2 activation by Fe
species located on the catalyst surface, as was indicated by CO and O2
pulse experiments. The PtFe/C catalyst presented a better catalytic performance in the O2 pulse experiments as compared with the standard
Pt/C even if CO was pre-adsorbed. At the first pulse, the conversion of O2
was close to 100% (versus 30% for Pt/C), implying that O2 could be effectively activated and reacted with the absorbed CO. It was clearly
shown that the Fe species, highly dispersed on the surface of Pt nanoparticles, worked as O2 activation sites and greatly enhanced catalytic
activity.
The enhancement of the activity of Pt-based catalysts in the low temperature PROX driven by the electronic eect is rationalized by a weakened CO bond with the Pt sites, which are otherwise poisoned.
The modifications of the electronic structure of Pt-based catalysts aect
the availability of the valence orbitals to form bonds with the reactants
during the catalytic process.89 When an adsorbate is chemically adsorbed
on a metal surface, new electronic states are formed. These electronic
states consist of bonding and antibonding states. When the antibonding
states are pushed above the Fermi level, the adsorbate bonds to the
metal. Hammer and Nrskov90 have proposed the adsorbate-metal
bonding model based on d-band theory. According to this theory, the
adsorption energy of the adsorbates, and in parallel the reactivity of the
metal surface, depends on the energy position of the d band relative to
the Fermi level. The d band shape of a metal and, thus, the d band centre
are prone to change after alloying with a dierent metal.
Contrary to the geometric eect, the identification of electronic eect
is less straightforward and the majority of work elucidating the electronic
structure of supported bimetallic Pt-based catalysts relies on using IR
spectroscopy of chemisorbed CO.16,91 High energy X-ray absorption
spectroscopy (XAS) gives useful insight into the electronic structure of
alumina-supported PtSn catalysts.92 The d band of Pt sample after
alloying with Sn is narrowed and presents a downward shift with respect
to the Fermi level. The electronic modification inhibited the H2 and CO
adsorption over the bimetallic catalyst. However, the electronic structure
of Pt in the monometallic catalyst after CO adsorption is similar to the Pt
after CO exposure of the PtSn sample.92 These observations could possibly actually suggest that the PtSn phase was destroyed in the presence
of CO. As a matter of fact, Moscu et al.11 at showed that at temperatures
246 | Catalysis, 2016, 28, 237267
17:24:19.
View Online
below 175 1C the adsorption of CO modified PtSn particles through the

segregation of the two metals.
Proof of the new properties of a bimetallic catalyst which is modified
electronically is reported by Eichhorn and co-workers.82,93 These authors
reported a series of nano-catalysts made of a transition metal core
(M Ru, Rh, Ir, Pd, or Au) covered with a 12 monolayer-thick shell of Pt
atoms. The synthesis procedure produced a PGM@Pt coreshell structure. Among the studied systems, Ru@Pt coreshell NPs exhibited significantly higher activity for the PROX reaction when compared with
monometallic mixtures and bulk non-segregated bimetallic nanoalloys.
The best Ru@Pt catalyst showed a 100% CO conversion below 20 1C
(0.1% CO and 0.5% O2 in H2) and fair stability. Since Ru is found in the
core of NPs, the dual site mechanism is not operative. This means that
the electronic structure of the Pt surface should be modified by the Ru
core. The interaction between Pt-monolayer and Ru core downshifts the
d-band center. As a result, the interaction strength between Pt surface
and adsorbates is weakened, and CO is significantly destabilized over the
Pt atoms. This leads to lower CO saturation coverage and provides much
more CO-free sites for activation of other reactants. DFT studies indicate
that the dierences in the catalytic activities for the various coreshell
NPs originate from a combination of (i) the relative availability of CO-free
Pt surface sites on the M@Pt NPs, which are necessary for O2 activation,
and (ii) a hydrogen-mediated low-temperature CO oxidation process
that is clearly distinct from the traditional bifunctional CO oxidation
mechanism.
Dierent techniques (XPS, TPR, CO TPD, XRD, TEM, HS-LEIS, IR) are
often used to study the bimetallic catalyst, but the characterization is
usually performed in a manner that only a static description of the local
structure is provided. More in situ or operando studies are required to
provide insight and understanding of the actual catalytically active site
formed under reaction conditions in the presence of the reactants,
products, and at the relevant temperature.
3.3 Mechanistic and kinetic studies over monometallic Pt-containing

catalysts
PROX over monometallic Pt supported catalysts follow a Langmuir
Hinshelwood reaction mechanism.94,95 This mechanism described in
Fig. 4 implies a competitive adsorption on the same Pt active sites between CO, O2 and H2. At low temperature, on a surface covered predominantly with CO, the adsorption of O2 and H2 is inhibited. Increase of
temperature will lead to the depletion of CO. Thus, the O2 adsorption and
activation over the newly formed Pt free sites becomes possible. The
LangmuirHinshelwood mechanism predicts a low activity of Pt-based
catalysts at low temperature due to the inhibited O2 dissociation over the
CO-covered Pt sites.
The low temperature regime or the low rate branch in which the
surface is largely covered by CO is associated with a reaction order close
to 1 for pCO and to 1 for pO2. The high rate branch or the high
Catalysis, 2016, 28, 237267 | 247
17:24:19.
View Online
Fig. 4 Competitive LangmuirHinshelwood mechanism over monometallic Pt catalysts.

(Adapted with permission from ref. 62: K. Liu, A. Wang and T. Zhang, ACS Catal., 2012, 2,
11651178. Copyright r 2012 American Chemical Society.)
Table 2 Comparison of reaction orders and apparent activation energy in PROX over
monometallic Pt-based catalysts.
Catalyst
aCO
aO2
Ea (kJ mol1)
Ref.
1 wt% Pt/Mordenite
1 wt% Pt/Al2O3
0.04 wt% Pt/Al2O3
0.69
0.47
0.51
0.68
0.66
0.76
40
63
78
97
97
98
temperature regime exhibits a reaction order of 1 for pCO and zero for
pO2.96 These assumptions are supported by Behm and coworkers96 who
investigated over a wide range of CO concentrations (0.021.5%) and low
stoichiometric O2 excess (pO2/pCO 0.51.5) the kinetics over 0.5 wt%
Pt/gAl2O3. The reaction orders with respect to pCO (0.4) and pO2 (0.8)
and the apparent activation energy of 71 kJ mol1 are in agreement
with the values for the low rate branch. Hence, oxygen adsorption is the
rate-determining step. The loss of CO2 selectivity at high temperature
(i.e. 250 1C) and at low CO concentration (SB15% at 0.02% CO) is attributed to the increase of hydrogen oxidation rate. Consequently, the
oxidation rate of co-adsorbed hydrogen is also limited by the dissociative
O2 adsorption step. Other literature values of the reaction orders and
apparent activation energy over Pt supported catalysts in PROX are reported in Table 2.
Schubert and co-workers48 performed DRIFTS measurements to
analyze the CO coverage over 0.5 wt% Pt/Al2O3 during selective CO oxidation in idealized reformate with 10 kPa CO and l 2 (where l 2 pO2/
pCO) at dierent temperatures (150, 175, 200 1C). For comparison, DRIFTS
spectra were recorded under CO adsorptiondesorption equilibrium
(1 kPa CO, balance N2). The spectra obtained in both cases are identical
and the steady state CO coverage is not significantly altered by the PROX
reaction. The CO coverage decreases by less than 20% as the partial
pressure is reduced from 1 kPa to 0.05 kPa. However, the adsorption of
oxygen and hydrogen were still inhibited by COads during the reaction.
Thereby, the selectivity towards CO2 in the PROX reaction on the Pt/Al2O3
is controlled by the high CO coverage.
In addition, DFT calculations have been used to study the selective CO
oxidation with respect to H2 on Pt (111) surfaces.99 The step involving the
248 | Catalysis, 2016, 28, 237267
17:24:19.
View Online
OH formation (O H-OH) has a higher activation energy barrier than

the step yielding H2O (OH OH-H2O O), suggesting that the PROX
selectivity is mainly determined by the CO2 and OH formation rates for
the mechanism proposed by Kandoi et al.99 On Pt (111) the barrier for OH
formation (0.83 eV) is lower than for CO oxidation (0.96 eV) and thus it
can be explained the low CO2 selectivity at low temperature (i.e. 150 1C).
The presence of hydrogen in the feed mixture was found in many
studies to have a positive eect enhancing the reaction rate of CO oxidation over monometallic Pt catalysts.100,101 This eect could be attributed to the interaction between COads and Hads or to the competitive
adsorption between H2 and CO that weakens the CO bonding over the
metal.96,102,103
3.4 Mechanistic and kinetic studies over bimetallic Pt-containing

catalysts
In spite of the vast amount of work concerning the enhanced performance in PROX over bimetallic Pt-based catalysts, very few detailed kinetic
studies have been reported. The incorporation of a second metal (or/and
MOx) or a reducible oxide in order to enhance the performance of Pt
catalysts in low temperature PROX leads to a change in the reaction
mechanism by creating a dual-site reaction pathway.16,104 The new route
implies the formation of new free sites that are able to adsorb and dissociate O2.
Moreover, the Mars van Krevelen mechanism (the so-called redox
mechanism) is proposed, as the MOx is easily reduced at the reaction
temperature by H2 dissociation over Pt. This mechanism involves the
participation of the surface lattice oxygen directly in the CO oxidation
reaction, providing an oxygen vacancy. Further, the gas-phase oxygen is
activated by the electron rich vacancy. This change in reaction mechanism is often reflected in a change of the reaction order of oxygen.15 CO
PROX over Pt usually has an O2 reaction order close to 1, while in the case
of a Mars van Krevelen mechanism the O2 reaction order is closer
to zero.104
Catalysts like MnOxPt/Al2O3,83 SnOxPt/Al2O3105 or PtFe/Al2O379 are
known to improve the oxygen supply and the CO oxidation reaction rate
and to follow a non-competitive LangmuirHinshelwood mechanism.
The O2 dissociation process is conducted by MOx and the activated
oxygen reacts with adsorbed CO through spill-over or at the metalsupport interface. In this mechanism, the O2 adsorption is no longer the
rate-determining step.
Similarly to the case of monometallic Pt catalysts, it has been reported
that hydrogen has a strong influence on the CO oxidation reaction over
bimetallic catalysts. At low temperatures an enhancement of CO oxidation is observed. In an attempt to better understand the improved
activity at low temperature induced by hydrogen, Tanaka et al.106 proposed for FeOx/Pt/TiO2 catalysts that hydroxyl groups (reaction intermediates assignable as OH species obtained via ionic dissociation of H2O
over FeOx) participate in the low temperature CO oxidation via reaction
Catalysis, 2016, 28, 237267 | 249
View Online
with formates: HCOO OH-CO2 H2O which would be the rate determining step of PROX reaction.
In the absence of hydroxyl groups onto the support, it is proposed that
the O2 dissociation is assisted by hydrogen93,107 through the following
steps:
17:24:19.
O2 H*-HO2*
HO2* -O* OH*
Some literature data report that carboxyl groups (COOH) play an important role in the PROX mechanism. Microkinetic studies for CO oxidation, WGS reaction and PROX over Pt and Rh catalysts reveal that the
reaction between CO and O is slow and the proposed mechanism involves the COH2 coupling, which would be a crucial step for the CO
oxidation. It is shown that the oxidation of CO by OH via the carboxyl
intermediate, and not by O, is the dominant path in PROX. Formation of
CO2 is attained through the COH2 coupling via CO* OH* reaction or
indirectly via the formate or carboxyl intermediates.102
CO* OH*2CO2* H* (direct pathway)
CO* OH*2COOH* * (indirect pathway)
COOH* *2CO2* H*
However, Liu et al.108 reported by DRIFT measurements over a Ir-Fe
catalyst that no formates or carbonates are formed at the catalyst surface
in the temperature range of 20150 1C in the H2 rich atmosphere. The
micro kinetic model proposed involves the following equations:
(1)
(2)
(3)
(4)
(5)
(6)
(7)
CO Ir2IrCO
O2 2*-2O*
IrCO O*-CO2 Ir *
H2 2Ir22HIr
HIr O*2OH* Ir
OH* HIr-H2O * Ir
OH* COIr-CO2 HIr *
where the * represents an iron site (Fe21). The model suggested that no
competitive adsorption between CO and O2 was observed for CO oxidation (steps 13) and in this case, the surface reaction between CO and
O was the rate-determining step. On the other hand, in the presence of
hydrogen, CO could be oxidized by both atomic O and by the surface OH
(step 7). The rate-limiting step is the reaction between adsorbed H and O
for the formation of OH (step 5). The main pathway for CO oxidation is
via oxidation by surface OH (step 7). Moreover, the presence of H2 favors
the increase of OH concentration and increases the PROX reaction rate.
The important role of OH groups played in the CO oxidation reaction is
observed when water is present in the reaction feed.106,109 However, few
works reported that water addition has increased the activity for low
temperatures PROX reaction.71
250 | Catalysis, 2016, 28, 237267
View Online
17:24:19.
4 Platinum-tin-based catalysts
Among the promising bimetallic Pt catalysts for PROX, PtSn based
catalysts showed enhanced activity in the low temperature PROX.16,110
Knowing the fact that neither Pt nor Sn can catalyze CO oxidation at low
temperature, a synergy between these two metals has been proposed. It
must be stressed that promoting eect of Sn in the PROX reaction has
also been evidenced in the case of Pd.2
4.1 Structural properties of PtSn catalysts
Pt and Sn form highly exothermic alloys. The phase diagram of PtSn
described in Fig. 5 reveals five intermediate stoichiometric phases (PtSn4
orthorhombic; PtSn2 cubic; Pt2Sn3 hexagonal; PtSn hexagonal and Pt3Sn
face centered cubic).
Of these five compounds, the Pt3Sn and PtSn phases have been
studied the most thoroughly, not only for their surface structural properties,111 but also for their reactivity and catalytic properties.112,113
To explain the high bonding energy in the alloys, the Engel-Brewer
model115 was used to qualitatively predict the strength of the intermetallic bond. This model predicted the charge transfer from dierent
metallic species because of electronegativity dierences. The basic EngelBrewer model has been confirmed when the intermetallic bond has been
correlated to a shift in the overlayer local d-electron band and a simultaneous dip in the noble metal (e.g. Pt) d-electron local density of states
(LDOS) at the Fermi level.116 The new electronic structure of Pt leads to
Fig. 5 Phase diagram of PtSn. (Reprinted from P. Anres, M. Gaune-Escard, J. Bros and
E. Hayer, Enthalpy of formation of the (PtSn) system, J. Alloys Compd., 280, 158167,
Copyright r 1998, with permission from Elsevier.114)
Catalysis, 2016, 28, 237267 | 251
17:24:19.
View Online
new interesting properties (i.e., change in the reactivity towards adsorbates). XPS studies89,92 indicate a binding energy shift of 1 eV in the Pt 4f
towards higher binding energy after alloying with Sn. This shift is associated with the electronic transfer from the Pt 4f level to the PtSn bond.
Singh et al.92 studied the electronic structure of supported Pt and PtSn
NPs and revealed that the dierence in the electronic structure of Pt after
alloying consists of a downward shift and narrowing of the d band of Pt
relative to the Fermi level. According to Hammer and Norskov,90 the
closer the d band centre is to the Fermi level, the stronger the bond
between the metal and the adsorbates will be. This assumption is supported by Singh et al.,92 who concluded that PtSn does not favor the
adsorption of hydrogen and carbon monoxide.
The changes in the electronic and geometric properties of the PtSn
alloys are known to play an important role in the enhancement of the
catalytic activity during PROX or other reactions such as alkane dehydrogenation117,118 or hydrocarbon hydrogenation.119
4.2 PtSn-based catalysts for PROX

Over the past years, PtSn containing catalysts have been studied in
PROX and showed improved activity in the low temperature regime (i.e.,
o120 1C). Studies over Pt3Sn single crystals,112,120 unsupported PtSn110
or supported PtSn16,37,103,105,121125 catalysts carried out under PROX or
CO oxidation15,89,113,126,127 conditions aimed at identifying the active
sites during the reaction. The high activity of these catalysts has been
ascribed to a synergistic bifunctional mechanism in which Pt provides
the adsorption sites for CO, while oxygen is activated over tin or tin oxide.
However, the simple addition of tin is not enough to achieve high
activities. In some cases, the poor activity in PROX103 at low temperature
could be assigned to formation of ill-defined catalysts, containing various
PtSnx phases. These types of catalysts are usually synthesized through
conventional techniques (e.g., impregnation, co-impregnation) leading to
ill-defined PtSn phases or separated Pt and Sn sites deposited onto the
support. This aects the performance of the Pt catalysts during PROX,
but the identification of the active catalytic sites, the optimum Pt/Sn ratio
or the discrimination between surface and bulk eects remain dicult.
Hence, dierent synthesis techniques have been developed (i.e., colloidal
techniques,87 controlled surface reaction between tin compounds and Pt
NPs118) in order to achieve well-defined PtSn NPs.
Even so, bimetallic PtSn catalysts under PROX or CO oxidation reaction conditions are prone to structural and surface modifications.
Somorjai and coworkers15 prepared well-defined monodispersed 2 nm
PtSn NPs via colloidal techniques and tested those in the CO oxidation
reaction. While intermixed metallic Pt and Sn were found under reducing
conditions, these metals were found to reversibly separate into isolated Pt
and oxidic tin domains when exposed to reaction mixture. However, the
modifications in the NPs structure are not necessarily a negative eect as
the newly formed phase provided new sites and prevented competition
between CO and O2 for the same active sites. Somorjai suggested that the
252 | Catalysis, 2016, 28, 237267
17:24:19.
View Online
addition of tin changes the reaction mechanism in the CO oxidation

reaction with tin oxide being the active Sn species under oxidative
atmosphere. For a PtSn/SiO2 catalysts, Margitfalvi et al.123,124 reported
that under oxidative conditions Sn41Pt ensemble sites are formed. The
Sn41 formed in situ is involved in the activation of CO adsorbed on the
neighboring Pt sites.
Despite the extensive theoretical and in situ experimental studies that
have been carried out to elucidate the reaction mechanism and the active
phases, many questions remain and more operando work would be required to address these questions.
4.3 Mechanistic and kinetic studies over PtSn-based catalysts

In general, the high activity of PtSn-containing catalysts for PROX has
been ascribed to a bifunctional mechanism in which activated oxygen
provided by SnOx enhances the oxidation of chemisorbed CO on Pt sites.
For the bifunctional mechanism, some groups15,128 proposed that the
SnOx surface dissociates gas phase oxygen, and the O-atoms migrate at
the Pt interface with SnOx phases where those react with adsorbed CO.
Gangal et al.129 however, proposed that it is the lattice oxygen of SnOx
that participates in CO oxidation and the reduced tin oxides are subsequently re-oxidized by gas-phase oxygen.
Kinetic studies support the non-competitive or Mars-van Krevelen
mechanism. Behm and coworkers,16 reported at 80 1C a 0.15 s1 and
2103 s1 TOF over PtSn/C and Pt/Al2O3, respectively. This enhancement is accompanied by a decrease of the apparent activation energy;
49 5 kJ mol1 over PtSn versus 77 7 kJ mol1 for pure Pt. The Ea for
PtSn exhibits a value much lower than the CO desorption barrier on PtSn
alloys (i.e., 90 kJ mol1). These results agree with the non-competitive
LangmuirHinshelwood mechanism, where CO oxidation takes place
between CO adsorbed on Pt sites and oxygen adsorbed on SnOx islands/
Sn sites or by spill-over of one of these species and subsequent reaction to
CO2. As oxygen activation is not inhibited by CO desorption, the PtSn
catalyst shows a higher CO conversion than that of Pt catalyst, at least in
the regime of high CO coverage.
However, more work is still needed to obtain a detailed micro-kinetic
model for PtSn catalysts and to unravel the nature of the active phase
(surface Sn oxide versus chemisorbed atomic oxygen). A recent combined kinetic and in situ PROX study by Moscu et al.104 on well-defined
nanoparticles indicated again markedly dierent reaction orders and
apparent activation energies between the case of plain Pt and PtSn
catalysts. The PROX kinetics over Pt/Al2O3 could be described by oxygen
adsorption as being rate-determining step, while this was not the case
for PtSn. For the PROX reaction over PtSn/Al2O3 a bifunctional mechanism was proposed with CO and H being adsorbed on the platinum
sites and oxygen being channeled from neighboring SnOx sites. The
authors proposed a Mars-Van Krevelen-type of mechanism for the SnOx
site, in agreement with the low value of 0.2 for the oxygen reaction order
over PtSn.104
Catalysis, 2016, 28, 237267 | 253
View Online
17:24:19.
Surface analyses by in situ infrared spectroscopy
5.1 CO adsorption analyses by IR over monometallic Pt-based

catalysts
Carbon monoxide adsorption on metals has been widely studied by
vibrational spectroscopy. The strong electric dipole associated with the
CO bond confers to this molecule a high molar extinction coecient
in IR spectroscopy, making detection possible even at minute concentrations. In addition, the CO stretching frequency is an indicator of the
way CO is bound to the substrate.
The COPt interaction is often described using the so-called Blyholder
model,130 based on p and r bonds formed between CO and the nearest
metal atom d orbitals. As a consequence, the CO molecule will compete
with the neighboring metal atoms for the d electrons of the atoms involved in the MCO bond. The coordination number of the surface metal
atom involved in the MCO bond will aect the frequency of the CO
bond vibration. When the metal coordination number increases, the
competition between CO and the other surface neighboring sites decreases the e retrodonation from the metal site involved in the adsorption towards the 2p* of CO, leading to an increased nCO frequency.
Indeed, it has been reported that dispersion of Pt, and hence the coordination number, is aecting the wavenumber of adsorbed CO: for
smaller Pt particles131,132 a higher metal electron density is available for
back bonding into the 2p* orbital of adsorbed CO molecules. This reduces the CO bond strength and hence the CO stretching frequency.131
The Blyholders model is represented in Fig. 6 and it is recalled to explain
the COads band wavenumbers in Table 3.
It is also proposed that CO molecules tend to adsorb on defect Pt sites
with lower coordination number, because of the larger bonding energy
associated with defect sites. Greenler et al.134 reported the presence of
three bands for CO adsorption over a Pt/SiO2 catalyst. The intensity of the
bands located at 2081, 2070 and 2063 cm1 vary in the same manner as
the relative numbers of face, corner and edge atoms in the sample vary
with particle size.
Moreover, a coverage-dependent shift is rationalized through an increase in the number of competing CO 2p* orbitals for the available
Fig. 6 (a) Scheme of Blyholder model and (b) typical wavenumber of CO(ads) on metals.
(Adapted from ref. 133 with permission from The Royal Society of Chemistry.)
254 | Catalysis, 2016, 28, 237267
View Online
137
Table 3 Factors that determine the COads IR frequency over Pt based samples.
High ratio of PtPt bonds:
PtCO bonds
2100 cm1
17:24:19.
Large particles/low dispersion

High surface coverage (higher
dipoledipole interaction)
Pt sites of higher coordination
number (terrace, step)
Blyholders Model49
COads band
frequency !
Low ratio of PtPt bonds:

PtCO bonds
2000 cm1
Small particles/high
dispersion
Low surface coverage (lower
dipoledipole interaction)
Pt sites of lower coordination
number (edge atoms,
kinked atoms)
d-electrons of the metal. Nevertheless, part of the shift observed by

changing the CO coverage appears to be due to the dipoledipole coupling between the species. This assumption is supported by Shigeishi
and King135 who reported that the coverage related frequency shift is
consistent with dipoledipole coupling between chemisorbed species
and the degree of coupling. According to Stoop et al.,136 the adsorption of
e donor molecules (i.e. ethylene) has an electronic eect over the surface
site structure. The electronic eect describes an increase in the electron
density of the metal and thus it results in an enhanced e back donation
towards the 2p* orbital of CO, that lowers the nCO frequency. In addition, the geometric eect consists of dispersion of the CO molecules in
between the ethylene molecules leading to a decrease in the dipole
dipole coupling between adjacent CO molecules. In return, a red-shift in
the CO frequency occurs.
In the IR spectrum obtained upon the CO adsorption on Pt/Al2O3
samples two nCO bands appear at (i) 20902040 cm1 assigned to linearly adsorbed CO on Pt atoms and (ii) 18601780 cm1 attributed to CO
bridging two or more Pt atoms.131
For a CO/metal system it can be useful to determine the CO heat of
adsorption as a function of CO coverage (y) at the surface under
equilibrium conditions. Bianchi and coworkers138 showed by means
of in situ IR spectroscopy under adsorption equilibrium (AEIR) that
changes in the CO coverage on Pt sites are related to the evolution of
the IR band intensities of adsorbed CO with the adsorption temperature
(Ta) at a given CO partial pressure. The dependence y f (Ta) allowed
the determination of the CO heat of adsorption of linear, bridged
and three folded CO species. The CO heat of adsorption calculations are
based on an adsorption model that assumes the presence of immobile
adsorbed species and that the heats of adsorption linearly decrease
with increase in CO coverage. With this study, Bianchi concluded that
in the presence of a CO pressure o103 Pa and for an adsorption temperature 4300 K, the Pt/alumina sample exhibits sites that are not fully
covered even at low temperatures. These sites are related to bridged CO
species. The heat of adsorption at zero-coverage (E0) and that at full
coverage (E1) for linear, bridged and three fold CO species are reported
in Table 4.
Catalysis, 2016, 28, 237267 | 255
View Online
17:24:19.
Table 4 Wavenumber and CO heat of adsorption observed over dierent CO species.138

CO species
Wavenumber
(cm1)
E0 (kJ mol1)
y0
E1 (kJ mol1)
y1
Linear (L)
Bridged (B)
Three fold (3F)
2075
18781824
1800
206
94
135
115
45
104
Fig. 7 Schematic representation of an ordered PtSn surface, in which the increased

distance between CO(ads) dipole leads to a decreased coupling, thereby lowering CO
vibration frequencies. The more electropositive Sn also donates electronic density to Pt,
resulting in increased back donation to CO anti-bonding orbital, thereby lowering CO
vibration frequencies.
The high values for the heats of adsorption at zero coverage evidence
the strong adsorbate-substrate bond strength. The heat of adsorption are
coverage-dependent due to the lateral adsorbateadsorbate interactions,
and chemical shift or due to the surface heterogeneity.139,140 Overall, the
heat of adsorption is much lower at y 1. The significant advantage of
Bianchis method is that the heats of adsorption are determined specifically for a given band (i.e. a given type of site), even when several are
present, contrary to the case of calorimetric methods that provide an
average value. This method is therefore well-suited for the analysis of
bimetallic samples in which several phases can be simultaneously
present.
5.2 CO adsorption analyses by IR over Pt-based alloys
CO adsorption has been widely studied on well-defined surface alloys, to
investigate site-blocking eects and electronic phenomena within the
bimetallic alloys. In bimetallic catalysts, the alloying process with an
element more electropositive than Pt makes that the second metal increases the electron density of Pt, resulting in an additional donation of
electrons from Pt to the antibonding orbitals of adsorbed CO (Fig. 7).
In this way, the CO bond is weakened and as a result its vibration
frequency is decreased. This is the case of the PtSn system, for which the
Pauling electronegativity of Pt and Sn are 2.28 and 1.96, respectively.
The wavenumber of CO(ads) is indeed significantly lower in the case of
256 | Catalysis, 2016, 28, 237267
17:24:19.
View Online
175C
2054
200C
225C
250C
275C
Log (1/R)
300C
0.01
325C
2045
2120
2080
2040
Wavenumber
2000
1960
(cm-1)
Fig. 8 (Top) In situ DRIFTS spectra of 1% Pt/Al2O3 after reduction at 400 1C. The
temperature was decreased from 400 to 50 1C by 50 1C steps in the presence of 2%
CO in H2. (Bottom) In situ DRIFTS spectra recorded under 2% CO/H2 over PtSn/Al2O3
pre-reduced at 400 1C. The temperature was decreased by 25 1C steps. (Adapted from
ref. 11 with permission from The Royal Society of Chemistry.)
the PtSn alloy (Fig. 8, bottom) as compared to that of the plain Pt sample
(Fig. 8, top).
A second eect induced by the alloy formation is the geometric eect.
In ordered bimetallic alloys, Pt atoms are diluted in between the second
metal atoms (Fig. 7) and thus the amount of Pt atoms exposed at the
surface decreases. The adjacent CO molecules are (i) separated further
away as compared to the case of the pure Pt and (ii) the dipole density
decreases, both eects leading to a much decreased dipoledipole
Catalysis, 2016, 28, 237267 | 257
17:24:19.
View Online
coupling which limits the usual dipole shift (towards higher frequencies
with increasing coverages). This eect can be clearly seen in the case of
the pure Pt,sample (Fig. 8, top) for which a large band shift (416 cm1) is
obtained at higher coverages (i.e. lower temperatures), while the band
shift observed in the case of the PtSn alloy (Fig. 8, bottom) is markedly
lower (ca. 9 cm1).
Moreover, the disappearance of bridged-bonded CO is reported due to
the lack of neighbouring Pt atoms141 or due the influence of the support
used. On silica-supported Pt samples the formation of bridged CO species is less likely as compared to the case of alumina-supported Pt.142
The trends induced in the CO adsorption frequency by either the
electronic or geometric eect is supported by studies on alloys such as
PtSn,16,141 PtGe,132 PtAu,142,143 PtPd,144 PtCu,145 PtRu.146 Some of
the mentioned alloys (i.e. PtAu or PtCu) do not aect the strength of CO
bond to Pt and the amount of adsorbed CO is similar to that obtained on
the monometallic Pt samples. This assumption leads to the conclusion
that the eect of Au and Cu are mainly geometric in nature. On the other
hand, studies on the PtSn/alumina or PtPb/alumina147 reported a decrease in the adsorbed CO with increasing the second metal content and
this was attributed to the weakening of the s PtCO bond due to an
electronic eect rather than a geometric one. Moscu et al.11 have recently
reported a direct measure of the heat of adsorption of CO on Pt and PtSn
using the AEIR method (Fig. 9). The heat of adsorption of CO on Pt surface
atoms at zero-coverage was almost divided by two on the alloy (i.e. 95 kJ
mol1) as compared to the case of the plain Pt sample (i.e. 180 kJ mol1).
This observation can explain the greater resistance against CO poisoning
of PtSn as compared to Pt, at least in the case of reducing conditions.
FTIR spectroscopy is an important tool to probe small variations of the
surface compositions of bimetallics and to discriminate adsorbates on
Fig. 9 (Left) Evolution with temperature of the DRIFTS band intensity of carbonyl species
over (O) Pt and (n) PtSn materials pre-reduced at 400 1C. Feed: 2% CO in H2. The
temperature was ramped down by 25 1C steps. The dotted lines correspond to a modified
Temkin adsorption model using the E0 and E1 values reported in the table on the right.
(Right) Table giving the heats of CO adsorption at zero (E0) and full (E1) coverage
determined through a modified Temkin model for the PtSn/Al2O3 and Pt/Al2O3 catalysts.
(Adapted from ref. 11 with permission from The Royal Society of Chemistry.)
258 | Catalysis, 2016, 28, 237267
View Online
17:24:19.
148
dierent adsorption sites. Maillard et al.

reported for a PtRu/C electrocatalyst CO stretching bands at n 2030 cm1 which are attributed to
a-top COads on Pt atoms electronically modified by the presence of Ru. As
the Ru domains become more present at the surface of agglomerated Pt
Ru particles, the position of the band shifted by ca. 5 cm1 and developed
a shoulder specific for CO adsorbed on Ru atoms. Hence, two vibrational
features similar to COad/Ru and COad/Pt are observed.
5.3 In situ IR of CO oxidation reaction over Pt-based samples
It is of upmost importance to get a detailed insight into the elementary
steps that take place at the surface of a catalyst during a reaction and IR
spectroscopy is one of the most adapted techniques to perform in situ
studies under reaction conditions.
The elementary steps of the CO oxidation reaction have been summarized by Engel and Ertl149 in 1979 into the three-step Langmuir
Hinshelwood (LH) reaction scheme:
COgas2COads
(O2)gas-2Oads
COads Oads-CO2 gas
where indices gas and ads refer to gas-phase and adsorbed species.
However, this broad picture of the reaction mechanism cannot be applied for supported Pt nanoparticles working under real conditions
(elevated temperatures and pressures) and additional information is
required.
By using in situ IR, Bianchi and coworkers150,151 have elucidated parameters playing significant roles in the elementary of the CO oxidation
over Pt/Al2O3. By performing kinetic studies they observed that the linear
adsorbed CO species are converted into CO2 at temperatures lower than
350 K at dierent O2/He concentrations. The oxidation of linear (L)
species requires the dissociation of chemisorbed oxygen but without any
competition for the adsorption sites; each species is adsorbed on different sites. Two kinetic models were proposed deriving from LH
mechanism. One of the models involves the reaction between the linear
CO species and a weakly adsorbed oxygen species (Owads). It is operative
at (i) whatever temperature in excess of CO and (ii) only at low temperature values in excess of O2. In contrast, the other model assumes the
reaction between linear CO and strongly adsorbed oxygen species (Osads)
and it is eective at high temperature in excess of oxygen. Moreover,
during the ignition process a surface phase transformation occurs from
a Pt surface covered with L CO to a Pt surface covered by Osads. This is
a result of an oxidative removal of the adsorbed CO into CO2 and not due
to a competitive chemisorption. Fridell et al.152 supported the observations made by Bianchi, by means of high-resolution step-scan FT-IR
spectroscopy.
Furthermore, the comparison of the transient CO oxidation reaction
between Pt samples with dierent metal loadings and dispersion152
Catalysis, 2016, 28, 237267 | 259
17:24:19.
View Online
(i.e. 1% Pt/Al2O3 versus 4% Pt/Al2O3) indicated the higher reactivity for the
catalyst containing larger particles. Bianchi et al.153 stated that whatever
the dispersion of Pt catalysts, the IR band of the linear CO species
dominates the IR spectra. At 300 K, the IR frequencies depend on the
dispersion: 2073, 2066 and 2054 cm1 for D 44%, 65% and 75%, respectively. The band shift is explained on the basis of the Blyholder
model. The position of the IR band of the bridged CO species also depends on the dispersion. For oxidation of the linear species by O2, it has
been shown that there is an induction period (the duration of oxidation
before the CO2 peak appears) for DZ0.6 but which is not observed at
lower dispersions. This is rationalized by considering the formation rate
of Pt sites that adsorb oxygen during oxidation depending on the dispersion. These sites are attributed to the bridged CO adsorption. The CO
heat of adsorption for the bridged species increases with dispersion and
thus, the removal of B CO (by desorption and oxidation) controls the
induction period.
Anderson154 recorded FTIR spectra of Pt/Al2O3 under in situ conditions over a wide range of O2 : CO ratios (0.2535 : 1) at temperatures
between 323 and 573 K in order to investigate the oxygen influence
over the CO surface coverage. When the sample is heated only in the
presence of CO, the COads peak on Pt displayed a shift from 2080 cm1
to 2055 cm1. In contrast, at O2 : CO ratios at or below stoichiometry,
the minimum value of the frequency is 20582059 cm1. Above the
stoichiometric ratio, the lowest frequency reached 2064 cm1.
Additionally, the oxidizing conditions did not lead to variations in
n(CO), suggesting either high CO coverage was maintained in the
presence of oxygen or that oxygen electronically influenced the carbonyl
frequency. To explain this behavior under oxidative conditions, Cant
and Donaldson155 suggest that invariance of n(CO) is rationalized by
high regional CO coverage over step or edge atoms which confine the
CO molecules forming CO islands within particular areas of the particle
surface. Likewise, the CO migration in the presence of co-adsorbed
oxygen is less likely.156
By studying CO reaction on Pt(335) Xu and Yates157 showed that both
dissociative O2 adsorption and CO adsorption occur preferentially on
step sites. It was found that chemisorbed CO on (111) terrace sites is
more reactive than chemisorbed CO on (100) step sites. In contrast,
chemisorbed O on step sites is more reactive at high CO coverages than
chemisorbed O on terrace sites. At high CO coverages the most reactive
geometry involves step site chemisorbed O interacting with terrace site
chemisorbed CO. This information provides a conceptual basis for
understanding the interplay between geometrical and energetic factors
influencing the CO oxidation reaction.
The above-mentioned in situ studies provided experimental understanding of structure-activity relationship and information on surface
species present during the catalytic process. However, if IR studies on CO
oxidation over Pt/alumina could be found in the literature, studies on
PROX under reductive environment or real reformate mixtures for fuel
cell application are scarce.
260 | Catalysis, 2016, 28, 237267
17:24:19.
View Online
5.4 In situ IR of PROX reaction over Pt-based samples

Behm et al.48 related the large dierence in activity during PROX over a
Pt/alumina and Au/Fe2O3 with the dierence in steady-state CO coverage.
Whereas on the Pt catalyst this is always near saturation under reaction
conditions, the amount of COads on gold particles (indicated by an IR
band at B2110 cm1) strongly depends on the CO partial pressure. In
addition, DRIFTS spectra recorded over the Pt sample under CO
adsorptiondesorption equilibrium (1 kPa CO, balance N2) and during
the PROX reaction (1 kPa CO, l 2, 75% H2) are practically identical,
suggesting that the steady-state CO coverage is essentially unperturbed
by the selective CO oxidation reaction.
Besides CO adspecies, assignment of other surface species is essential
to probe the possible reaction pathways. Schubert and co-workers158 reported that formate species are accumulating during PROX reaction with
75% H2, with the formation of three peaks at 2900, 1594 and 1394 cm1.
TPR experiments showed that formate species band intensity reached a
maximum rate at 200 1C and decreased at higher temperature. Reversible
decomposition of formate species is also possible and COads is formed
as a product. However, the authors did not suggest the evolution of
other surface species (i.e. carbonate/bicarbonate). Similarly, on a PtSn/C
sample, Caglayan et al.125 reported by performing DRIFTS/adsorption
desorption measurements that Pt3Sn alloy formation on surface oxygen
group-rich nitric acid-treated activated carbon has led to augmented CO
adsorption and faster surface reaction involving intermediate hydroxyl
groups which bring along increase in CO conversion and selectivity as the
temperature decreases within the range of 110135 1C.
In contrast, Zhang et al.159 proposed by dynamic in situ DRIFTS (adding
or removing CO and/or H2) over a FeOx/Pt/TiO2 catalyst that the CO
oxidation in the presence of hydrogen proceeds via HCOO intermediate
and the subsequent oxidation of HCOO by the reaction with OH, which is
the rate-determining step.
As briefly afore-mentioned, Moscu et al.104 studied the PROX over welldefined alumina-supported Pt and PtSn particles. The reaction rate was
much faster over PtSn/Al2O3 than over Pt/Al2O3. Drastically dierent
apparent activation energies and oxygen reaction orders were found for
both samples. Whereas the PROX kinetics over Pt/Al2O3 could be described by oxygen adsorption as being rate-determining step, this was not
the case for PtSn. In situ DRIFTS showed that the alloyed phase PtSn,
present initially under a mixture of 1% CO/80% H2, readily segregated
upon the introduction of 2% O2. A typical PtCO(ads) band at 2050 cm1
was observed under reducing conditions before O2 introduction, while
the IR spectra were immediately modified upon O2 introduction, indicating that the PtSn alloy had been converted into segregated phases of
Pt and SnOx (Fig. 10).
For the PROX reaction over PtSn/Al2O3 a bifunctional mechanism is
proposed with CO and H being adsorbed on the platinum sites and
oxygen being channeled from neighboring SnOx sites. This work suggests
a Mars-Van Krevelen type of mechanism for the SnOx site, in agreement
with the low value of 0.2 for the oxygen reaction order over PtSn.
Catalysis, 2016, 28, 237267 | 261
17:24:19.
View Online
Fig. 10 In situ DRIFTS spectra recorded at 225 1C at various times under 1% CO 2%

O2 80% H2 in Ar over PtSn/Al2O3 pre-reduced at 400 1C. The dashed spectrum was
collected just before the addition of O2. (Adapted from A. Moscu, L. Veyre, C. Thieuleux,
F. Meunier and Y. Schuurman, Catal. Today, in press, doi: 10.1016/j.cattod.2014.12.036.
Copyright (2015) with permission from Elsvier.104)
Conclusions
Many bimetallic compounds exhibit much improved PROX performances

as compared to the monometallic counterparts. The improved activity
can be due to geometric or electronic eects. The signal of adsorbed CO,
the molecule of interest in the PROX reaction, can be monitored by in situ
or operando IR spectroscopy, a truly surface-sensitive technique. Such IRbased studies enable a detailed description of the metallic surface at
work and can help describe the nature of the active phases involved. The
reports discussed here thus reemphasized the need to study the catalysts
under relevant operating conditions, since alloy segregation can readily
occur upon the modification of the experimental conditions. The case of
the PtSn system was described in detail, showing that any PtSn alloy
phase present in the absence of molecular oxygen was readily converted
to Pt SnOx as soon as O2 was present.
References
1
2
S. Chansai, R. Burch, C. Hardacre, J. Breen and F. Meunier, J. Catal., 2011,

281, 98105.
R. Lanza, M. Bersani, L. Conte, A. Martucci, P. Canu, M. Guglielmi,
G. Mattei, V. Bello, M. Centazzo and R. Rosei, J. Phys. Chem. C, 2014, 118,
2539225402.
D. I. Potemkin, E. Y. Semitut, Y. V. Shubin, P. E. Plyusnin, P. V. Snytnikov,
E. V. Makotchenko, D. Y. Osadchii, D. A. Svintsitskiy, S. A. Venyaminov,
S. V. Korenev and V. A. Sobyanin, Catal. Today, 2014, 235, 103111.
ronzier, F. Morfin, M. Lomello and J.-L. Rousset, J. Catal., 2014, 311,
T. De
221229.
262 | Catalysis, 2016, 28, 237267
View Online
17:24:19.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
C. Zlotea, F. Morfin, T. S. Nguyen, N. T. Nguyen, J. Nelayah, C. Ricolleau,

M. Latroche and L. Piccolo, Nanoscale, 2014, 6, 99559959.
R. Ferrando, J. Jellinek and R. L. Johnston, Chem. Rev., 2008, 108, 845910.
R. Ghosh Chaudhuri and S. Paria, Chem. Rev., 2011, 112, 23732433.
J. H. Sinfelt, in Bimetallic Catalysts: Discoveries, Concepts and Applications,
Wiley-Interscience, New York, 1983.
B. Roldan Cuenya and F. Behafarid, Surf. Sci. Rep., 2015, 70, 135187.
P. L. Hansen, J. B. Wagner, S. Helveg, J. R. Rostrup-Nielsen, B. S. Clausen
and H. Topse, Science, 2002, 295, 20532055.
A. Moscu, Y. Schuurman, L. Veyre, C. Thieuleux and F. Meunier, Chem.
Commun., 2014, 50, 85908592.
A. B. Merlo, V. Vetere, J. F. Ruggera and M. L. Casella, Catal. Commun., 2009,
10, 16651669.
R. Burch and L. C. Garla, J. Catal., 1981, 71, 360372.
, J. Power
A. O. Neto, R. R. Dias, M. M. Tusi, M. Linardi and E. V. Spinace
Sources, 2007, 166, 8791.
W. D. Michalak, J. M. Krier, S. Alayoglu, J.-Y. Shin, K. An, K. Komvopoulos,
Z. Liu and G. A. Somorjai, J. Catal., 2014, 312, 1725.
M. M. Schubert, M. J. Kahlich, G. Feldmeyer, M. Huttner, S. Hackenberg,
H. A. Gasteiger and R. J. Behm, Phys. Chem. Chem. Phys., 2001, 3, 11231131.
J. Snyder, K. Livi and J. Erlebacher, Adv. Funct. Mater., 2013, 23, 54945501.
taz, P. Bayle-Guillemaud, M. Chatenet,
M. Lopez-Haro, L. Dubau, L. Gue
, N. Caque
, E. Rossinot and F. Maillard, Appl. Catal., B, 2014,
J. Andre
152153, 300308.
Z. Liu, G. S. Jackson and B. W. Eichhorn, Angew. Chem., Int. Ed., 2010, 49,
31733176.
K. J. J. Mayrhofer, V. Juhart, K. Hartl, M. Hanzlik and M. Arenz, Angew.
Chem., Int. Ed., 2009, 48, 35293531.
B. M. Weckhuysen, Chem. Commun., 2002, 97110.
F. C. Meunier, Catal. Today, 2010, 155, 164171.
M. Momirlan and T. Veziroglu, Renewable Sustainable Energy Rev., 2002, 6,
141179.
M. Momirlan and T. N. Veziroglu, Int. J. Hydrogen Energy, 2005, 30, 795802.
G. Cipriani, V. Di Dio, F. Genduso, D. La Cascia, R. Liga, R. Miceli and
G. Ricco Galluzzo, Int. J. Hydrogen Energy, 2014, 39, 84828494.
P. P. Edwards, V. L. Kuznetsov, W. I. F. David and N. P. Brandon, Energy
Policy, 2008, 36, 43564362.
M. A. DeLuchi, Int. J. Hydrogen Energy, 1989, 14, 81130.
V. Mehta and J. S. Cooper, J. Power Sources, 2003, 114, 3253.
X. Cheng, Z. Shi, N. Glass, L. Zhang, J. Zhang, D. Song, Z.-S. Liu, H. Wang
and J. Shen, J. Power Sources, 2007, 165, 739756.
S. J. Peighambardoust, S. Rowshanzamir and M. Amjadi, Int. J. Hydrogen
Energy, 2010, 35, 93499384.
J.-M. Bae, S. Ahmed, R. Kumar and E. Doss, J. Power Sources, 2005, 139,
9195.
A. Rodrigues, J. C. Amphlett, R. F. Mann, B. Peppley and P. R. Roberge, in
Energy Conversion Engineering Conference, 1997. IECEC-97, Proceedings of the
32nd Intersociety, IEEE, 1997, pp. 768773.
N. Zamel and X. Li, Prog. Energy Combust. Sci., 2011, 37, 292329.
L. N. Cleemann, F. Buazar, Q. Li, J. O. Jensen, C. Pan, T. Steenberg, S. Dai
and N. J. Bjerrum, Fuel Cells, 2013, 13, 822831.
S. Gottesfeld and J. Paord, J. Electrochem. Soc., 1988, 135, 26512652.
J. J. Baschuk and X. Li, Int. J. Energy Res., 2003, 27, 10951116.
Catalysis, 2016, 28, 237267 | 263
View Online
37
38
17:24:19.
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
S. K. Jain, E. M. Crabb, L. E. Smart, D. Thompsett and A. M. Steele, Appl.

Catal., B, 2009, 89, 349355.
J. L. Ayastuy, A. Gil-Rodriguez, M. P. Gonzalez-Marcos and M. A. GutierrezOrtiz, Int. J. Hydrogen Energy, 2006, 31, 22312242.
C.-Y. Huang, Y.-Y. Chen, C.-C. Su and C.-F. Hsu, J. Power Sources, 2007, 174,
294301.
S. Y. Chin, O. S. Alexeev and M. D. Amiridis, J. Catal., 2006, 243, 329339.
S. L. Zhou, Z. S. Yuan and S. D. Wang, Int. J. Hydrogen Energy, 2006, 31, 924
933.
M. Echigo and T. Tabata, Catal. Lett., 2004, 98, 3742.
Y. H. Kim, E. D. Park, H. C. Lee, D. Lee and K. H. Lee, Catal. Today, 2009,
146, 253259.
Y. F. Han, M. J. Kahlich, M. Kinne and R. J. Behm, Phys. Chem. Chem. Phys.,
2002, 4, 389397.
Y. Huang, A. Wang, X. Wang and T. Zhang, Int. J. Hydrogen Energy, 2007, 32,
38803886.
S. Zhou, G. S. Jackson and B. Eichhorn, Adv. Funct. Mater., 2007, 17, 3099
3104.
S. Monyanon, S. Pongstabodee and A. Luengnaruemitchai, J. Power Sources,
2006, 163, 547554.
M. M. Schubert, M. J. Kahlich, H. A. Gasteiger and R. J. Behm, J. Power
Sources, 1999, 84, 175182.
S. Monyanon, S. Pongstabodee and A. Luengnaruemitchai, J. Chin. Inst.
Chem. Eng., 2007, 38, 435441.
P. Landon, J. Ferguson, B. E. Solsona, T. Garcia, S. Al-Sayari, A. F. Carley,
A. A. Herzing, C. J. Kiely, M. Makkee, J. A. Moulijn, A. Overweg,
S. E. Golunski and G. J. Hutchings, J. Mater. Chem., 2006, 16, 199208.
R. M. T. Sanchez, A. Ueda, K. Tanaka and M. Haruta, J. Catal., 1997, 168,
125127.
S. Zeng, Y. Wang, S. Ding, J. J. H. B. Sattler, E. Borodina, L. Zhang,
B. M. Weckhuysen and H. Su, J. Power Sources, 2014, 256, 301311.
C. Wang, B. Li, H. Lin and Y. Yuan, J. Power Sources, 2012, 202, 200208.
J. Kugai, T. Moriya, S. Seino, T. Nakagawa, Y. Ohkubo, H. Nitani and
T. A. Yamamoto, Int. J. Hydrogen Energy, 2013, 38, 44564465.
X. Zhao, Y. Huang, W. Zhang, X. Liu, A. Wang, J. Wang and T. Zhang,
J. Nanosci. Nanotechnol., 2014, 14, 69706976.
A. Lopez Camara, V. Cortes Corberan, L. Barrio, G. Zhou, R. Si, J. C. Hanson,
M. Monte, J. C. Conesa, J. A. Rodriguez and A. Martinez-Arias, J. Phys.
Chem. C, 2014, 118, 90309041.
Z. Zhao, J. Lin, G. Wang and T. Muhammad, AIChE J., 2015, 61, 239252.
N. Ren, L. Guo, X. Dong and C. Wen, Transition Met. Chem., 2015, 40,
207216.
S. Yao, K. Mudiyanselage, W. Xu, A. C. Johnston-Peck, J. C. Hanson, T. Wu,
D. Stacchiola, J. A. Rodriguez, H. Zhao, K. A. Beyer, K. W. Chapman,
P. J. Chupas, A. Martinez-Arias, R. Si, T. B. Bolin, W. Liu and
S. D. Senanayake, ACS Catal., 2014, 4, 16501661.
S. Dulnee, A. Luengnaruemitchai and R. Wanchanthuek, Int. J. Hydrogen
Energy, 2014, 39, 64436453.
A. Luengnaruemitchai, S. Chawla and R. Wanchanthuek, Int. J. Hydrogen
Energy, 2014, 39, 1695316963.
K. Liu, A. Wang and T. Zhang, ACS Catal., 2012, 2, 11651178.
E.-Y. Ko, E. D. Park, K. W. Seo, H. C. Lee, D. Lee and S. Kim, Catal. Today,
2006, 116, 377383.
264 | Catalysis, 2016, 28, 237267
View Online
64
65
17:24:19.
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
A. Manasilp and E. Gulari, Appl. Catal., B, 2002, 37, 1725.

O. S. Alexeev, S. Y. Chin, M. H. Engelhard, L. Ortiz-Soto and M. D. Amiridis,
J. Phys. Chem. B, 2005, 109, 2343023443.
S. H. Oh and R. M. Sinkevitch, J. Catal., 1993, 142, 254262.
G. Avgouropoulos, T. Ioannides, C. Papadopoulou, J. Batista, S. Hocevar and
H. Matralis, Catal. Today, 2002, 75, 157167.
o, C. Descorme and D. Duprez, Appl. Catal., B, 2004, 54, 5966.
F. Marin
K. Y. Koo, U. H. Jung and W. L. Yoon, Int. J. Hydrogen Energy, 2014, 39, 5696
5703.
Y. Uemura, Y. Inada, K. K. Bando, T. Sasaki, N. Kamiuchi, K. Eguchi,
A. Yagishita, M. Nomura, M. Tada and Y. Iwasawa, J. Phys. Chem. C, 2011,
115, 58235833.
C. Wang, B. Li, H. Lin and Y. Yuan, J. Power Sources, 2012, 202, 200208.
H. Xu, Q. Fu, X. Guo and X. Bao, ChemCatChem, 2012, 4, 16451652.
S. Zhou, G. S. Jackson and B. Eichhorn, Adv. Funct. Mater., 2007, 17, 3099
3104.
J. Kugai, T. Moriya, S. Seino, T. Nakagawa, Y. Ohkubo, H. Nitani,
Y. Mizukoshi and T. A. Yamamoto, Appl. Catal., B, 2012, 126, 306314.
J. Kugai, T. Moriya, S. Seino, T. Nakagawa, Y. Ohkubo, H. Nitani and
T. A. Yamamoto, Int. J. Hydrogen Energy, 2013, 38, 44564465.
T. Komatsu, M. Takasaki, K. Ozawa, S. Furukawa and A. Muramatsu, J. Phys.
Chem. C, 2013, 117, 1048310491.
H. Zhang, M. Jin, H. Liu, J. Wang, M. J. Kim, D. Yang, Z. Xie, J. Liu and
Y. Xia, Acs Nano, 2011, 5, 82128222.
O. Rosseler, C. Ulhaq-Bouillet, A. Bonnefont, S. Pronkin, E. Savinova,
A. Louvet, V. Keller and N. Keller, Appl. Catal., B, 2015, 166, 381392.
J. Yin, J. Wang, T. Zhang and X. Wang, Catal. Lett., 2008, 125, 7682.
E.-Y. Ko, E. D. Park, K. W. Seo, H. C. Lee, D. Lee and S. Kim, Catal. Lett.,
2006, 110, 275279.
S. Y. Chin, O. S. Alexeev and M. D. Amiridis, J. Catal., 2006, 243, 329339.
S. Alayoglu and B. Eichhorn, J. Am. Chem. Soc., 2008, 130, 1747917486.
lez-Marcos, J. R. Gonza
lez-Velasco and
J. L. Ayastuy, M. P. Gonza
rrez-Ortiz, Appl. Catal., B, 2007, 70, 532541.
M. A. Gutie
W. Yu, M. D. Poroso and J. G. Chen, Chem. Rev., 2012, 112, 57805817.
A. Siani, B. Captain, O. S. Alexeev, E. Stafyla, A. B. Hungria, P. A. Midgley,
J. M. Thomas, R. D. Adams and M. D. Amiridis, Langmuir, 2006, 22, 5160
5167.
T. Ebashi, Y. Ishida, Y. Nakagawa, S. Ito, T. Kubota and K. Tomishige,
J. Phys. Chem. C, 2010, 114, 65186526.
M. Boualleg, J.-M. Basset, J.-P. Candy, V. Caps, J.-C. Jumas, S. Norsic,
E. A. Quadrelli, L. Veyre and C. Thieuleux, ChemCatChem, 2012, 4, 1729
1732.
H. Zhang, D. Lin, G. Xu, J. Zheng, N. Zhang, Y. Li and B. H. Chen, Int. J.
Hydrogen Energy, 2015, 40, 17421751.
B. C. Vicente, R. C. Nelson, J. Singh, S. L. Scott and J. A. van Bokhoven,
Catal. Today, 2011, 160, 137143.
B. Hammer and J. K. Nrskov, Surf. Sci., 1995, 343, 211220.
M. Arenz, V. Stamenkovic, B. B. Blizanac, K. J. Mayrhofer, N. M. Markovic
and P. N. Ross, J. Catal., 2005, 232, 402410.
J. Singh, R. C. Nelson, B. C. Vicente, S. L. Scott and J. A. van Bokhoven, Phys.
Chem. Chem. Phys., 2010, 12, 56685677.
A. U. Nilekar, S. Alayoglu, B. Eichhorn and M. Mavrikakis, J. Am. Chem. Soc.,
2010, 132, 74187428.
Catalysis, 2016, 28, 237267 | 265
View Online
94
17:24:19.
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
N. P. Lebedeva, M. T. M. Koper, J. M. Feliu and R. A. van Santen, J. Electroanal. Chem., 2002, 524525, 242251.
H. P. Bonzel and R. Ku, Surf. Sci., 1972, 33, 91106.
M. J. Kahlich, H. A. Gasteiger and R. J. Behm, J. Catal., 1997, 171, 93105.
S. Ren and X. Hong, Fuel Process. Technol., 2007, 88, 383386.
D. H. Kim and M. S. Lim, Appl. Catal., A, 2002, 224, 2738.
S. Kandoi, A. A. Gokhale, L. C. Grabow, J. A. Dumesic and M. Mavrikakis,
Catal. Lett., 2004, 93, 93100.
H. Muraki, S.-I. Matunaga, H. Shinjoh, M. S. Wainwright and D. L. Trimm,
J. Chem. Technol. Biotechnol., 1991, 52, 415424.
K. Tanaka, H. He, M. Shou and X. Shi, Catal. Today, 2011, 175, 467470.
A. B. Mhadeshwar and D. G. Vlachos, J. Phys. Chem. B, 2004, 108, 15246
15258.
J. L. Ayastuy, M. P. Gonzalez-Marcos and M. A. Gutierrez-Ortiz, Catal.
Commun., 2011, 12, 895900.
A. Moscu, L. Veyre, C. Thieuleux, F. Meunier and Y. Schuurman, Catalysis
Today, 2015, 258, 241246.
nsan and R. Yldrm, Catal. Lett., 2006, 111, 173176.
G. Uysal, A. N. Akn, Z. I_ . O
K. Tanaka, M. Shou, H. He, X. Shi and X. Zhang, J. Phys. Chem. C, 2009, 113,
1242712433.
S. Salomons, R. E. Hayes and M. Votsmeier, Appl. Catal., A, 2009, 352, 2734.
K. Liu, A. Wang, W. Zhang, J. Wang, Y. Huang, X. Wang, J. Shen and
T. Zhang, Ind. Eng. Chem. Res., 2011, 50, 758766.
H.-F. Wang, R. Kavanagh, Y.-L. Guo, Y. Guo, G.-Z. Lu and P. Hu, Angew.
Chem., Int. Ed., 2012, 51, 66576661.
D. Teschner, A. Wootsch and Z. Paal, Appl. Catal., A, 2012, 411, 3134.
K. Balakrishnan and J. Schwank, J. Catal., 1991, 127, 287306.
C. Dupont, F. Delbecq, D. Loreda and Y. Jugnet, J. Catal., 2011, 278,
239245.
Y. Jugnet, D. Loreda, C. Dupont, F. Delbecq, E. Ehret, F. J. C. S. Aires,
B. S. Mun, F. A. Akgul and Z. Liu, J. Phys. Chem. Lett., 2012, 3, 37073714.
P. Anres, M. Gaune-Escard, J. Bros and E. Hayer, J. Alloys Compd., 1998, 280,
158167.
L. M. Pecora and P. Ficalora, J. Solid State Chem., 1979, 27, 239253.
S. Speller and U. Bardi, in The Chemical Physics of Solid Surfaces, ed.
D. P. Woodru, Elsevier, 2002, vol. 10, pp. 184224.
F. Humblot, J. P. Candy, F. Le Peltier, B. Didillon and J. M. Basset, J. Catal.,
1998, 179, 459468.
H. Zhu, D. H. Anjum, Q. Wang, E. Abou-Hamad, L. Emsley, H. Dong,
P. Laveille, L. Li, A. K. Samal and J.-M. Basset, J. Catal., 2014, 320, 5262.
M. Englisch, V. S. Ranade and J. A. Lercher, J. Mol. Catal. A: Chem., 1997,
121, 6980.
V. Stamenkovic, M. Arenz, B. B. Blizanac, P. N. Ross and N. M. Markovic,
J. New Mater. Electrochem. Syst., 2004, 7, 125132.
zkara and A. E. Aksoylu, Appl. Catal., A, 2003, 251, 7583.
-S. O
P. Marques, N. F. P. Ribeiro, M. Schmal, D. A. G. Aranda and
M. M. V. M. Souza, J. Power Sources, 2006, 158, 504508.
J. L. Margitfalvi, I. Borbath, M. Hegedus, A. Szegedi, K. Lazar, S. Gobolos
and S. Kristyan, Catal. Today, 2002, 73, 343353.
J. L. Margitfalvi, I. Borbath, M. Hegedus, E. Tfirst, S. Gobolos and K. Lazar,
J. Catal., 2000, 196, 200204.
B. S. Caglayan, I. I. Soykal and A. E. Aksoylu, Appl. Catal., B, 2011, 106,
540549.
266 | Catalysis, 2016, 28, 237267
View Online
126
17:24:19.
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
S. Stevanovic, D. Tripkovic, V. Tripkovic, D. Minic, A. Gavrilovic, A. Tripkovic

and V. M. Jovanovic, J. Phys. Chem. C, 2014, 118, 278289.
S. Axnanda, W.-P. Zhou and M. G. White, Phys. Chem. Chem. Phys., 2012, 14,
1020710214.
K. Grass and H.-G. Lintz, J. Catal., 1997, 172, 446452.
N. D. Gangal, N. M. Gupta and R. M. Iyer, J. Catal., 1990, 126, 1325.
G. Blyholder, J. Phys. Chem., 1964, 68, 27722777.
A. Bourane and D. Bianchi, J. Catal., 2003, 218, 447452.
L. C. deMenorval, A. Chaqroune, B. Coq and F. Figueras, J. Chem. Soc.,
Faraday Trans., 1997, 93, 37153720.
B. L. Mojet, S. D. Ebbesen and L. Leerts, Chem. Soc. Rev., 2010, 39, 4643
4655.
R. G. Greenler, K. D. Burch, K. Kretzschmar, R. Klauser, A. M. Bradshaw and
B. E. Hayden, Surf. Sci., 1985, 152153(Part 1), 338345.
R. A. Shigeishi and D. A. King, Surf. Sci., 1976, 58, 379396.
F. Stoop, F. J. C. M. Toolenaar and V. Ponec, J. Catal., 1982, 73, 5056.
S. P. TEH, Thesis, Universite Claude Bernard-Lyon1, 2005.
A. Bourane, O. Dulaurent and D. Bianchi, J. Catal., 2000, 196, 115125.
V. M. Browne, S. G. Fox and P. Hollins, Catal. Today, 1991, 9, 114.
D. Tibiletti, A. Goguet, D. Reid, F. C. Meunier and R. Burch, Catal. Today,
2006, 113, 94101.
K. Balakrishnan and J. Schwank, J. Catal., 1992, 138, 491499.
K. Balakrishnan, A. Sachdev and J. Schwank, J. Catal., 1990, 121, 441455.
D. Mott, J. Luo, P. N. Njoki, Y. Lin, L. Wang and C.-J. Zhong, Catal. Today,
2007, 122, 378385.
J. Xu, L. Ouyang, G.-J. Da, Q.-Q. Song, X.-J. Yang and Y.-F. Han, J. Catal.,
2012, 285, 7482.
L. Li, Z. Xu, F. Liu, Y. Shao, J. Wang, H. Wan and S. Zheng, J. Photochem.
Photobiol., A, 2010, 212, 113121.
H. Hartmann, T. Diemant, J. Bansmann and R. J. Behm, Phys. Chem. Chem.
Phys., 2012, 14, 1091910934.
A. G. T. Bastein, F. J. C. Toolenaar and V. Ponec, J. Catal., 1984, 90, 8895.
F. Maillard, A. Bonnefont, M. Chatenet, L. Guetaz, B. Doisneati-Cottignies,
H. Roussel and U. Stimming, Electrochim. Acta, 2007, 53, 811822.
T. Engel and G. Ertl, in Advances in Catalysis, ed. D. D. Eley, H. Pines and
P. B. Weisz, Academic Press, 1979, vol. 28, pp. 178.
A. Wille and E. Fridell, Appl. Catal., B, 2007, 70, 294304.
A. Bourane, S. Derrouiche and D. Bianchi, J. Catal., 2004, 228, 288297.
J. Anderson, J. Chem. Soc., Faraday Trans., 1992, 88, 11971201.
N. W. Cant and R. A. Donaldson, J. Catal., 1981, 71, 320330.
D. M. Haaland and F. L. Williams, J. Catal., 1982, 76, 450465.
J. Xu and J. T. Yates Jr., J. Chem. Phys., 1993, 725.
rgen Behm, J. Catal., 1997, 172,
M. M. Schubert, H. A. Gasteiger and R. Ju
256258.
X. Shi, K. Tanaka, H. He, M. Shou, W. Xu and X. Zhang, Catal. Lett., 2008,
120, 210214.
Catalysis, 2016, 28, 237267 | 267
3D MoS2/Graphene hybrid layer materials

as counter electrodes for dye-sensitized
solar cells
17:24:22.
Wei Wei and Yun Hang Hu*

DOI: 10.1039/9781782626855-00268
3D MoS2/Graphene hybrid layer materials were prepared by in situ synthesizing molybdenum sulfide on 3D honeycomb-like structured graphene sheets via a hydrothermal
approach. Furthermore, the composite materials were explored as counter electrodes for
dye-sensitized solar cells (DSSCs). It was found that the DSSC with 3D MoS2/Graphene
hybrid counter electrode exhibited a high energy conversion eciency of 7.31%, which is
larger than those with either MoS2 (2.81%) or graphene counter electrode (6.72%). The
excellent electrocatalytic activity of the hybrid counter electrode was further demonstrated by cyclic voltammograms, namely, the MoS2/Graphene hybrid electrode showed
higher peak current density and lower peak-to-peak separation positions than either MoS2
or graphene counter electrode.
Introduction
The conversion of solar energy to electrical energy is one of the major

strategies to solve global energy issues.1,2 Since the pioneering research
tzel in 1991,3 dye-sensitized solar cells (DSSCs) have
by O Regan and Gra
attracted considerable attention as an alternative to conventional silicon
based solar cells because of their low cost, low energy consumption,
simple fabrication process, and high power conversion eciency.4,5
Typically, DSSCs are composed of a photoelectrode (a transparent
conducting subtract with a dye coated TiO2 film), an electrolyte, and a
counter electrode (CE). As shown in Fig. 1, electrons transfer from an
excited dye sensitizer into the semiconductor conduction band, then
pass through the transparent conductive electrode to the external circuit,
and finally to the counter electrode (CE). The CE catalyzes the regeneration of redox couples and collects electrons from external circuit. For an
optimized cell, the CE material should possess low sheet resistance, high
catalytic activity for reduction of the redox electrolyte, excellent chemical
stability, and low cost.
For most cases, the iodide-triiodide couple has been exploited as the
redox mediator and the overall redox reaction in DSSCs can be
expressed as
I3 2e23I
(1)
The tri-iodides are produced near the dye sensitized TiO2 electrode
and reduced at the counter electrode. Platinum (Pt) is widely used for
the counter electrodes of DSSCs due to its high catalytic activity
Department of Materials Science and Engineering, Michigan Technological
University, 1400 Townsend Drive, Houghton MI 49931-1295, USA.
E-mail: yunhangh@mtu.edu
268 | Catalysis, 2016, 28, 268280

c
17:24:22.
View Online
Fig. 1 DSSC scheme.
and excellent conductivity. However, Pt is not only expensive, but

also unstable in I/I3 redox-based electrolyte because it can form a
PtI4 compound.6,7 Those issues with Pt-based CE stimulated a
great interest to develop ecient Pt-free CEs for DSSCs, including
carbon materials (graphite, carbon black, carbon nanotubes, and
graphene),824 conducting polymers,2527 metal sulfides,2831 metal
carbides,32 metal phosphides,33 metal nitrides,3448 and metal oxides.49,50 Among them, graphene is the most promising CE material for
DSSCs because of its high stability, large surface area, and excellent
electrical conductivity.5153
Graphene, a two-dimensional (2D) sheet of sp2-hybridized carbon
atoms, has attracted worldwide attention and research interest. Numerous synthesis approaches, such as mechanical cleavage,51 chemical
vapor deposition,54 and epitaxial growth,55 have been developed for the
first generation graphene (flat graphene sheets on a substrate). In the
mechanical cleavage method,51 graphene is detached from a graphite
crystal using an adhesive tape. However, it is dicult to obtain large
amounts of graphene by this method. Chemical vapor deposition (CVD)
is a well-known process, in which carbon-based gas compounds decompose on a surface to grow a graphene film.54 The CVD growth techniques has its the high compatibility with the current complementary
metal-oxide-semiconductor technology. However, its need of expensive
substrate materials would limit its applications for large-scale production. Epitaxial growth is also a substrate-based method,55 where
isolated monolayer of graphene is grown on a single-crystal silicon carbide (SiC) by vacuum graphitization. Its major advantage is that epitaxialgrown graphene can be patterned using standard lithography methods.
However, it is very dicult to control the thickness of graphene in
the routine production. All of those techniques are used to prepare
flat graphene sheets on a substrate. Unlike the first generation, the second generation of graphene emphasizes controlling shape and tuning
property of graphene, such as 3D graphene sheets. Very recently,
Catalysis, 2016, 28, 268280 | 269
17:24:22.
View Online
Fig. 2
MoS2 structure.
we synthesized 3D graphene directly from CO and CO2 via invented

reactions.15,21,22 The 3D graphene exhibited excellent performance for
DSSCs with a high energy conversion eciency of 6B8%.15,21,22
Molybdenum sulfide (MoS2), which is a typical transition metal sulfide,
possesses a layer-structure similar to graphite. Its basic unit consists of
three atomic layers, one Mo layer sandwiched between two S layers. This
is dierent from graphene, which is a single atomic layer. The triple
layers in MoS2 are stacked and held together via weak van der Waals
interaction (Fig. 2).56,57 MoS2 is widely applied as solid lubricants and
hydrodesulfurization catalysts. However, its application for DSSCs is
limited.29,5863 MoS2, produced by a hydrothermal method, has been
employed as CE materials for DSSCs.28 Furthermore, carbon materials
(such as 2D graphene sheets and 1D carbon nanotubes) have been added
into MoS2 to improve MoS2 performance for DSSCs.5963
In this chapter, we report the synthesis of 3D MoS2/Graphene hybrid
layer-structure materials using 3D honeycomb-structured graphene sheets.
These hybrid layer-structure materials can be used as counter electrodes
for DSSCs, and have shown excellent photovoltaic conversion eciency.
Preparation of DSSCs
Briefly, the synthesis of materials and fabrication steps for DSSCs are
described in the following sections. It is the main novelty to fabricate 3D
MoS2/Graphene counter electrode compared to those reported above
(Section 1).
270 | Catalysis, 2016, 28, 268280
17:24:22.
View Online
2.1 Synthesis of honeycomb-structured graphene (HSG)

Honeycomb-structured graphene (HSG) was synthesized as the previous
report15 and briefly described as follow: Lithium oxide (Li2O) power
(Aldrich) was loaded into a ceramic tube reactor and exposed to CO at
pressure of 35 psi. The reactor temperature increased from room temperature to 550 1C at a rate of 10 1C min1 and then remained constant
for 24 hours, followed by cooling down to room temperature. This solid
product was treated by 36.5 wt% hydrochloric acid (HCl) and washed
with de-ionized (DI) water for more than 10 times. The remained solid
was separated from water by centrifugation (3600 rpm) and then dried
overnight at 80 1C to get graphene powder.
2.2 Synthesis of MoS2/Graphene hybrid layer materials
MoS2/Graphene hybrid layer materials were prepared with synthesizing
MoS2 on 3D honeycomb-structured graphene with a typical hydrothermal method as follows: 1 mmol of Na2MoO4 2H2O and 5 mmol of
thiourea were dissolved in 60 ml of distilled water, followed by adding
0.108 g of the synthesized 3D honeycomb-structured graphene (HSG).
Then, the obtained mixture was transferred into a 100 ml Teflon-lined
autoclave and kept for reactions at 210 1C for 24 h. Finally, the black
precipitate was collected by centrifugation, washed three times with
distilled water and ethanol, and then dried in an oven at 80 1C for 12 h.
The BET surface area of the MoS2/Graphene sample was determined by
nitrogen adsorption at 77 K with a Micromeritics ASAP 2000 adsorption
instrument. Before the nitrogen adsorption, the sample was degassed
at 100 1C.
2.3 Preparation of photoelectrodes
FTO glass plates were ultrasonically washed with water and the conducting surface of the FTO was treated with a 0.4 mM TiCl4 aqueous
solution at 70 1C for 30 min. A commercial TiO2 sol (Solaronix,
Ti-nanooxide T/SP) was used to form a TiO2 film on a FTO by the doctorblade method. The TiO2 film was thermally treated in air at 325 1C for
5 min, 375 1C for 5 min, 450 1C for 15 min, and 500 1C for 15 min. The
obtained TiO2/FTO glass plate was post-treated with the 0.4 mM TiCl4
solvent at 70 1C, followed by heating in air at 500 1C for 30 min. Finally,
the obtained film was immersed in a cis-bis(isothiocyanato)bis(2,2 0 bipyridyl-4,40 -dicarboxylato)-ruthenium(II)-bis-tetrabutylammonium (N719,
0.3 mM in ethanol) dye solution for 24 hours to achieve dye sensitization.
2.4 Preparation of counter electrodes
First, the fluorine-doped tin oxide (FTO) transparent glass substrates
(Hartford Glas Co., Inc, 13 O sq1) were ultrasonically cleaned sequentially in water, acetone, and ethanol for 10 min, respectively. The obtained MoS2/Graphene powder was mixed with ethanol with stirring for
5 min to prepare a homogeneous paste. The paste was deposited on
fluorine-doped tin oxide (FTO) coated glass to form a composite film by
doctor blading method. The MoS2/Graphene counter electrode was
Catalysis, 2016, 28, 268280 | 271
View Online
17:24:22.
obtained by heating the MoS2/Graphene coated FTO glass at 80 1C for 4 h.

For comparison, pure MoS2 and 3D honeycomb-like structured graphene
counter electrodes were also prepared using the same approach. The
X-ray diraction spectrometer (Scintag XDS2000 Powder Diractometer)
was employed to characterize the crystalline structure of MoS2, graphene,
and MoS2/Graphene films.
2.5 Fabrication and characterization of DSSCs
The sandwich DSSCs were assembled using the N719-sensitized TiO2
photoelectrode and MoS2/Graphene composite counter electrode with
liquid electrolyte (I/I3 redox couple) between them. The electrolyte
consists of 0.025 M LiI, 0.04 M I2, 0.28 M tert-butyl pyridine (TBP), 0.6 M
1-butyl-3-methylimidazolium iodide (BMII), and 0.05 M guanidinium
thiocyanate in acetonitrile/valeronitrile solvent with 85/15 volume ratio.
The photovoltaic performance of the obtained MoS2/Graphene DSSCs
was evaluated by Keithley 2400 under illumination of AM 1.5 simulate
sunlight (1 sun, 100 mW cm2) using a Newport solar simulator.
Electrochemical impedance spectroscopy (EIS) data were collected at a
potential equal to the open-circuit voltage of a DSSC and 10 mV amplitude over the frequency range of 0.1100 kHz by CHI 660A electrochemical workstation in dark condition. Cyclic voltammetry (CV)
measurements were carried out using an electro chemical workstation
(EG&G Princeton Applied Research) with a three-electrode system
(MoS2/Graphene composite as a working electrode, Pt wire as a counter
electrode, and Ag/AgCl as a reference electrode) containing acetonitrile
solution of 10 mM LiI, 1 mM I2, and 0.1 M LiClO4.
Characterization of DSSCs
3.1 The photovoltaic performances of DSSCs with MoS2/Graphene

counter electrodes
The photovoltaic performances of DSSCs with counter electrodes of
MoS2/Graphene composites were examined by photocurrent densityvoltage (J-V) measurements. The obtained J-V curves are illustrated in
Fig. 3 and photovoltaic parameters listed in Table 1. As the graphene
content increased, the power conversion eciency increased up to the
maximum value of 7.31% at MoS2/Graphene molar ratio of 1/9 and then
decreased. This indicates that the optimum composition is 1/9 MoS2/
Graphene. CE with dierent MoS2/Graphene ratios exhibited similar Voc
values. Since Voc is dependent on the TiO2 film and the electrolyte, no
eect of MoS2/Graphene ratio on Voc indicates that the MoS2/Graphene
did not cause any changes for the TiO2 film and the I3/I electrolyte.
However, the short circuit current density (Jsc) of the DSSC always increased with increasing graphene content, whereas its fill factor (FF)
decreased with increasing graphene content after the MoS2/Graphene
ratio is smaller than 1/9. This indicates that the power conversion enhancement with increasing graphene content is due to the increase of
current density, whereas further increasing graphene content caused the
272 | Catalysis, 2016, 28, 268280
17:24:22.
View Online
Fig. 3
J-V curves of DSSCs with MoS2/Graphene hybrid counter electrodes.
Table 1 The photovoltaic performance of the DSSCs with dierent materials as CEs.
Samples (MoS2 : Graphene molar ratio)
Jsc (mA cm2)
Voc (V)
FF
Z (%)
12 : 1
9:1
3:1
1:1
1:3
1:9
1 : 12
1 : 20
1 : 30
7.72
8.20
9.76
13.59
15.58
16.64
16.97
17.30
17.75
0.64
0.72
0.72
0.73
0.71
0.76
0.76
0.76
0.77
0.50
0.59
0.55
0.58
0.56
0.58
0.55
0.52
0.49
2.49
3.47
3.85
5.73
6.17
7.31
7.12
6.78
6.76
reduction of FF and thus resulted in the decrease of the power conversion

eciency.
For the comparison of DSSCs fabricated with MoS2, graphene, and
optimized MoS2/Graphene counter electrodes, their J-V curves are illustrated in Fig. 4 and photovoltaic parameters listed in Table 2. One can see
that the DSSC with the optimized MoS2/Graphene counter electrode exhibited the highest power conversion eciency because it has a larger
current density than that with a MoS2 counter electrode and higher FF
factor than that with a graphene counter electrode.
3.2 Structure characterization of MoS2/Graphene hybrid layer
materials
The crystal structures of MoS2, graphene, and MoS2/Graphene hybrid
layer material were evaluated by X-ray diraction measurements. No
Catalysis, 2016, 28, 268280 | 273
17:24:22.
View Online
Fig. 4 J-V curves of DSSCs with MoS2, graphene, and MoS2/Graphene hybrid counter
electrodes.
Table 2 The photovoltaic performance of the DSSCs with dierent materials as CEs.
Samples
Jsc (mA cm2)
Voc (V)
FF
Z (%)
Rs
Rct
Zn
MoS2
Graphene
MoS2/Graphene
6.76
18.92
16.64
0.64
0.77
0.76
0.50
0.46
0.58
2.18
6.72
7.31
20
25
20
90
35
80
11 000
265
6500
obvious diraction peak was detected for 3D graphene (Fig. 5). In contrast, both MoS2 and MoS2/Graphene samples showed the strong diffraction peaks at 14.421, 33.251, and 58.581, corresponding to the crystal
planes of (002), (100), and (110) of hexagonal MoS2 (PDF no. 771716).
Furthermore, the strong peak at 14.421 reveals a well-stacked layered
structure.56 This confirms that MoS2 was synthesized on graphene
sheets.
The formation of MoS2 on graphene sheets was further supported by
SEM images. Figure 6a shows that the hydrothermal synthesized MoS2
without graphene is composed of interlaced nanosheets, each of which is
about 4 nm thick. The graphene sheet, which has a thickness of about
2 nm, possesses a 3D honeycomb-like structure shape with a cell size in
the range of 50500 nm (Fig. 6b). In the MoS2/Graphene hybrid layer
material, one can see MoS2 in the cells of the graphene honeycomb
(Fig. 6c). BET surface areas for MoS2, graphene, and MoS2/Graphene, are
101, 153, and 165 m2 g1, respectively. This indicates that MoS2 and
graphene slightly increased the surface area.
274 | Catalysis, 2016, 28, 268280
17:24:22.
View Online
Fig. 5 XRD patterns of honeycomb-like structured graphene (HSG), MoS2, and MoS2/
Graphene (1/9 molar ratio) hybrid material.
Fig. 6 SEM image of (a) MoS2, (b) 3D honeycomb-like structured graphene, and (c) MoS2/
Graphene (1/9 molar ratio) hybrid material.
3.3 Electrochemical characterization of DSSCs with MoS2/Graphene

counter electrodes
Two functions are necessary for a DSSC counter electrode (CE): (1) collecting electrons from external circuit and (2) catalyzing the reduction
Catalysis, 2016, 28, 268280 | 275
View Online
17:24:22.
I3
8,15,21,22,26,63
of oxidized electrolyte (namely, the reduction of

to I ).
Furthermore, the electro-catalytic reduction of I3 consists of two steps: (a)
I3 is adsorbed on the counter electrode and (b) the adsorbed I3 obtains
electrons to convert into I. Because MoS2 is an ionic compound, it would
be excellent for the adsorption of I3 ions. However, as a semiconductor,
MoS2 is a poor collector of electrons, which are needed to be transferred to
the adsorbed I3. Indeed, the DSSC with a MoS2 counter electrode showed
a low power conversion eciency (2.18%) (Table 2). In contrast, graphene
possesses an excellent conductivity for electrons. When MoS2 was replaced
by 3D honeycomb-like structured graphene for a counter electrode, the
DSSC reached a power conversion eciency of 6.72% (Table 2). When
MoS2 and graphene are combined, their best electrocatalytic sites should
be the MoS2/Graphene interface, which possesses both the excellent
I3 adsorption of MoS2 and the high electrical conductivity of graphene.
Therefore, there are three types of electrocatalytic sites with an activity
order as: Sinterface4Sgraphene4SMoS2. When MoS2 is dominant, the increase
of graphene content boosts the number of the active sites of both graphene
and enlarge the MoS2/Graphene interface, leading to the enhancement of
power conversion eciency. When the composition of the MoS2/Graphene
creates the largest interface, the power conversion eciency can reach the
maximum value. The further increase of graphene content would decrease
the interface and thus reduce the eciency. This can explain the relationship between composition of MoS2/Graphene and its power conversion eciency (Table 1).
Electrochemical impedance spectra (EIS) were obtained for MoS2,
graphene, and MoS2/Graphene hybrid CEs. Two semicircles were observed in the frequency range of 0.1100 KHz for all three samples
(Fig. 7). The intercept of the first semicircle (left one) represents the
ohmic serial resistance (Rs) related to the intrinsic resistance of
assembled cells.64,65 The semicircle in lower Z 0 (corresponding to highfrequency) is associated with the charge transfer resistance (Rct) at the
CE/electrolyte interface, which changes inversely with catalytic ability of
CEs for the reduction of I3 to I, while the semicircle in the higher Z 0
(Low-frequency range) arises from the Nernst diusion impedance (Zn) of
the triiodide/iodide couple.6669 All EIS curves were fitted by equivalent
circuit model with Z-view software and summarized in Table 2. As shown
in Table 2, one can see the variation of Rs was negligible (2025 O) for all
three CEs, whereas Rct and Zn values are strongly dependent on their
composition. The graphene has the lowest Rct and Zn, MoS2 the highest,
and MoS2/Graphene in between. This indicates the electrical conductivity
order: graphene4MoS2/Graphene4MoS2. This confirms that the power
conversion eciency is not only dependent on electrical conductivity. In
other words, although MoS2/Graphene hybrid electrode has a poorer
conductivity than graphene, the hybrid one possesses an optimum
combination of electrical conductivity and I3 adsorption, leading to
a higher energy conversion eciency of 7.31%. This is further supported
by cyclic voltammograms (CVs) (Fig. 8). For a typical CV curve, two pairs
of oxidation and reduction peaks should be observed for the I/I3 and
I3/I2 redox pairs, respectively.44,45,70
276 | Catalysis, 2016, 28, 268280
17:24:22.
View Online
Fig. 7 EIS with the corresponding circuit model inset for dierent counter electrodes.
Fig. 8 CVs for MoS2, graphene, and MoS2/Graphene hybrid counter electrodes.
3I 2 I3 2e

2I3 2 3I2 2e
Catalysis, 2016, 28, 268280 | 277
17:24:22.
View Online
Furthermore, the peak current density and peak-to-peak separation

(Epp) are two important parameters to examine catalytic activities of
dierent electrodes.70 The higher peak current density and lower Epp, the
better catalytic activity that counter electrode has. Figure 8 shows
that both 3D graphene and MoS2/Graphene exhibited typical reduction
and oxidation characteristics for I/I3 redox couple. However, the
MoS2/Graphene hybrid electrode showed higher peak current density
(1.38 mA cm2) and lower peak-to-peak separation positions (0.2 V). This
confirms that the MoS2/Graphene hybrid electrode possesses better
electrocatalytic activity than the graphene electrode.
4 Conclusions
3D MoS2/Graphene (1/9 molar ratio) hybrid layer materials were prepared
by an in situ hydrothermal method. These hybrid materials have been
characterized as counter electrodes for dye-sensitized solar cells (DSSCs).
The 3D MoS2/Graphene hybrid counter electrode exhibited improved
properties compared to either MoS2 or graphene counter electrode. Cyclic
voltammograms and EIS spectra further revealed the excellent electrocatalytic activity of the 3D MoS2/Graphene hybrid counter electrode.
A DSSC with the 3D MoS2/Graphene hybrid counter electrode achieved a
high power conversion eciency (7.31%).
Acknowledgements
This work was supported by the U.S. National Science Foundation
(NSF-CBET-0931587) and the ACS Petroleum Research Fund (PRF-51799ND10). Hu also thanks Charles and Carroll McArthur for their great
support.
References
1
A. Paracchino, V. Laporte, K. Sivula, M. Gratzel and E. Thimsen, Nat. Mater.,

2011, 10, 456.
2 Y. D. Hou, B. L. Abrams, P. C. K. Vesborg, M. E. Bjorkeun, K. Herbest, L. Bech,
A. M. Setti, C. D. Damsgaard, T. Pedersen, O. Hansen, J. Tossmeisl, S. Dahl,
J. K. Norskov and I. Chorkendor, Nat. Mater., 2011, 10, 434.
tzel, Nature, 1991, 353, 737.
3 B. ORegan and M. Gra
tzel, Nature, 2001, 414, 338.
4 M. Gra
tzel, Acc. Chem. Res., 2009, 42, 1788.
5 M. Gra
6 E. Olsen, G. Hagen and S. E. Lindquist, Sol. Energy Mater. Sol. Cells, 2000,
63, 267.
tzel, Sol. Energy Mater. Sol. Cells, 1996, 44, 99.
7 A. Kay and M. Gra
8 H. Wang and Y. H. Hu, Energy Environ. Sci., 2012, 5, 8182.
9 H. Wang, S. L. Leonard and Y. H. Hu, Ind. Eng. Chem. Res., 2012, 51,
10613.
10 H. Choi, H. Kim, S. Hwang, Y. Han and M. Jeon, J. Mater. Chem., 2011,
21, 7548.
11 H. Choi, H. Kim, S. Hwang, W. Choi and M. Jeon, Sol. Energy Mater. Sol. Cells,
2011, 95, 323.
12 F. Gong, H. Wang and Z.-S. Wang, Phys. Chem. Chem. Phys., 2011, 13, 17676.
278 | Catalysis, 2016, 28, 268280
View Online
13
14
17:24:22.
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
tzel, Nano Lett., 2011, 11, 5501.

L. Kavan, J.-H. Yum and M. Gra
T. N. Murakami, S. Ito, Q. Wang, M. K. Nazeeruddin, T. Vesso, I. Cesar,
tzel, J. ElectroP. Liska, R. Humphry-Baker, P. Comte, P. Pechy and M. Gra
chem. Soc., 2006, 153, A2255.
H. Wang, K. Sun, F. Tao, D. Stacchiola and Y. H. Hu, Angew. Chem., Int. Ed.,
2013, 52, 9210.
H. Wang and Y. H. Hu, J. Colloid Interface Sci., 2013, 391, 21.
K. Suzuki, M. Yamaguchi, M. Kumagai and S. Yanagida, Chem. Lett., 2003,
32, 28.
Z. Huang, X. Z. Liu, K. X. Li, D. M. Li, Y. H. Luo, H. Li, W. B. Song, L. Q. Chen
and Q. B. Meng, Electrochem. Commun., 2007, 9, 596.
F. Ghamouss, R. Pitson, F. Odobel, M. Boujtita, S. Caramori and
C. A. Bignozzi, Electrochim. Acta, 2010, 55, 6517.
M. X. Wu, X. Lin, T. H. Wang, J. S. Qiu and T. L. Ma, Energy Environ. Sci., 2011,
4, 2308.
W. Wei and Y. H. Hu, Int. J. Energy Res., 2015, 39, 842.
W. Wei, K. Sun and Y. H. Hu, J. Mater. Chem. A, 2014, 2, 16842.
J. B. Xia, N. Masaki, K. J. Jiang and S. Yanagida, J. Mater. Chem., 2007,
17, 2845.
M. K. Wang, A. M. Anghel, B. Marsan, N. L. C. Ha, N. Pootrakulchote,
tzel, J. Am. Chem. Soc., 2009, 131, 15976.
S. M. Zakeeruddin and M. Gra
J. H. Wu, Q. H. Li, L. Q. Fan, Z. Lan, P. J. Li, J. M. Lin and S. C. Hao, J. Power
Sources, 2008, 181, 172.
W. Wei, H. Wang and Y. H. Hu, Int. J. Energy Res., 2014, 38, 1099.
J. M. Pringle, V. Armel and D. R. Macfarlane, Chem. Commun., 2010, 46, 5367.
H. C. Sun, D. Qin, S. Q. Huang, X. Z. Guo, D. M. Li, Y. H. Luo and Q. B. Meng,
Energy Environ. Sci., 2011, 4, 2630.
X. K. Xin, M. He, W. Han, J. H. Jung and Z. Q. Lin, Angew. Chem., Int. Ed.,
2011, 50, 11739.
M. X. Wu, Y. D. Wang, X. Lin, N. S. Yu, L. Wang, L. L. Wang, A. Hagfeldt and
T. L. Ma, Phys. Chem. Chem. Phys., 2011, 13, 19298.
J. Y. Lin, J. H. Liao and S. W. Chou, Electrochim. Acta, 2011, 56, 8818.
M. X. Wu, X. Lin, A. Hagfeldt and T. L. Ma, Angew. Chem., Int. Ed., 2011,
50, 3520.
J. S. Jang, D. J. Ham, E. Ramasamy, J. Lee and J. S. Lee, Chem. Commun., 2010,
46, 8600.
A. R. Ko, J. K. Oh, Y. W. Lee, S. B. Han and K. W. Park, Mater. Lett., 2011,
65, 2220.
Y. Y. Dou, G. R. Li, J. Song and X. P. Gao, Phys. Chem. Chem. Phys., 2012,
14, 1339.
W. Wei, H. Wang and Y. H. Hu, J. Mater. Chem. A, 2013, 1, 14350.
G. R. Li, F. Wang, Q. W. Jiang, X. P. Gao and P. W. Shen, Angew. Chem., Int.
Ed., 2010, 49, 3653.
M. X. Wu, J. Bai, Y. D. Wang, A. J. Wang, X. Lin, L. Wang, Y. H. Shen,
Z. Q. Wang, A. Hagfeldt and T. L. Ma, J. Mater. Chem., 2012, 22, 11121.
Q. W. Jiang, G. R. Li and X. P. Gao, Chem. Commun., 2009, 6720.
G. R. Li, J. Song, G. L. Pan and X. P. Gao, Energy Environ. Sci., 2011, 4, 1680.
M. X. Wu, Q. Y. Zhang, J. Q. Xiao, C. Y. Ma, X. Lin, C. Y. Miao, Y. J. He,
Y. R. Gao, A. Hagfeldt and T. L. Ma, J. Mater. Chem., 2011, 21, 10761.
X. Y. Zhang, X. Chen, S. M. Dong, Z. H. Liu, X. H. Zhou, J. H. Yao, S. P. Pang,
H. X. Xu, Z. Y. Zhang, L. F. Li and G. L. Cui, J. Mater. Chem., 2012, 22, 6067.
C. P. Lee, L. Y. Lin, K. W. Tsai, R. Vittal and K. C. Ho, J. Power Sources, 2011,
196, 1632.
Catalysis, 2016, 28, 268280 | 279
View Online
44
45
17:24:22.
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
M. H. Yeh, L. Y. Lin, C. P. Lee, H. Y. Wei, C. Y. Chen, C. G. Wu, R. Vittala and

K. C. Ho, J. Mater. Chem., 2011, 21, 19021.
G. R. Li, F. Wang, J. Song, F. Y. Xiong and X. P. Gao, Electrochim. Acta., 2012,
65, 21620.
J. Song, G. R. Li, F. Y. Xiong and X. P. Gao, J. Mater. Chem., 2012, 22, 20580.
Q. W. Jiang, G. R. Li, S. Liu and X. P. Gao, J. Phys. Chem. C, 2010, 114, 13397.
B. Yoo, K. J. Kim, Y. H. Kim, K. Kim, J. Ko, W. M. Kim and N. G. Park, J. Mater.
Chem., 2011, 21, 3077.
H. Wang, W. Wei and Y. H. Hu, J. Mater. Chem. A, 2013, 1, 6622.
H. Wang, W. Wei and Y. H. Hu, Top. Catal., 2014, 57, 607.
K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos,
I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666.
R. R. Nair, P. Blake, A. N. Grigorenko, K. S. Novoselov, T. J. Booth, T. Stauber,
N. M. R. Peres and A. K. Geim, Science, 2008, 320, 1308.
C. Lee, X. Wei, J. W. Kysar and J. Home, Science, 2008, 321, 385.
K. S. Kim, Y. Zhao, H. Jang, S. Y. Lee, J. M. Kin, K. S. Kim, J.-H. Ahn, P. Kim,
J.-Y. Choi and B. H. Hong, Naure, 2009, 457, 706.
P. Sutter, Nat. Mater., 2009, 8, 171.
R. Tenne, L. Margulis, M. Genut and G. Hodes, Nature, 1992, 360, 444.
Y. F. Li, Z. Zhou, S. B. Zhang and Z. F. Chen, J. Am. Chem. Soc., 2008,
130, 16739.
M. Wu, Y. Wang, X. Lin, N. Yu, L. Wang, A. Hagfeldt and T. Ma, Phys. Chem.
Chem. Phys., 2011, 13, 19298.
C. J. Liu, S. Y. Tai, S. W. Chou, Y. C. Yu, K. D. Chang, S. Wang, F. S. S. Chien,
J. Y. Lin and T. W. Lin, J. Mater. Chem., 2012, 22, 21057.
G. Yue, J. Y. Lin, S. Y. Tai, Y. Xiao and J. Wua, Electrochim. Acta, 2012, 85, 162.
S. Y. Tai, C. J. Liu, S. W. Chou, F. S. S. Chien, J. Y. Lin and T. W. Lin, J. Mater.
Chem., 2012, 22, 24753.
J. Y. Lin, C. Y. Chan and S. W. Chou, Chem. Commun., 2013, 49, 1440.
G. Yue, J. Wu, Y. Xiao, M. Huang, J. Lin and J. Y. Lin, J. Mater. Chem. A, 2013,
1, 1495.
M. Gratzel, J. Photochem. Photobiol., C, 2003, 4, 145.
F. Gong, H. Wang, X. Xu, G. Zhou and Z. S. Wang, J. Am. Chem. Soc., 2012,
134, 10953.
S. Sakurai, H. Q. Jiang, M. Takahashi and K. Kobayashi, Electrochim. Acta,
2009, 54, 5463.
L. Y. Han, N. Koide, Y. Chiba, A. Islam, R. Komiya, N. Fuke, A. Fukui and
R. Yamanaka, Appl. Phys. Lett., 2005, 86, 213501.
P. Joshi, L. F. Zhang, Q. L. Chen, D. Galipeau, H. Fong and Q. Qiao, ACS Appl.
Mater. Interfaces, 2010, 2, 3572.
J. K. Koh, J. Kim, B. Kim, J. H. Kim and E. Kim, Adv. Mater., 2011, 23, 1641.
Z. Z. Wu, D. Z. Wang, Y. Wang and A. Sun, Adv. Eng. Mater., 2010, 12, 534.
280 | Catalysis, 2016, 28, 268280

Catalysis - Volume 28

Uploaded by

Copyright:

Available Formats

Catalysis - Volume 28

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Catalysis - Volume 28

Uploaded by

Copyright:

Available Formats

17:23:09.

Published on 19 February 2016 on http://pubs.rsc.org | doi:10.1039/9781782626855-FP001

A Specialist Periodical Report

Wei Wei, Michigan Technological University, MI, USA

Chapter 1: Tungsten containing materials as heterogeneous catalysts for

Catalysis, 2016, 28, viix | vii

The Royal Society of Chemistry 2016

Chapter 3: Recent advances in the synthesis and catalytic applications of

this application is reviewed here. But several potential green chemical

x | Catalysis, 2016, 28, viix

Tungsten containing materials as heterogeneous catalysts for

Wei-Lin Dai, Jing Ding, Quanjing Zhu, Ruihua Gao and

Alumina ceramic foams as catalyst supports

Erfan Behravesh, Leena Hupa, Tapio Salmi and

The Royal Society of Chemistry 2016

Recent advances in the synthesis and catalytic applications

Research and development of hydrocracking catalysts and

Chong Peng, Xiangchen Fang and Ronghui Zeng

Titano-silicates: highlights on development, evolution and

Ayomi Sheamilka Perera and Marc-Olivier Coppens

Nanofiber-supported metal-based catalysts

Elucidation of mechanistic and kinetic aspects of watergas

Recent progresses on the use of supported bimetallic catalysts

Alina Moscu, Yves Schuurman and Frederic C. Meunier

3D MoS2/Graphene hybrid layer materials as counter electrodes

Wei Wei and Yun Hang Hu

Catalysis, 2016, 28, xixiii | xiii

Tungsten containing materials as

Green Chemistry is now a central issue in both academia and industry

The Royal Society of Chemistry 2016

Pristine W-based catalyst for green catalytic oxidation

Pristine tungsten-based catalyst mainly includes tungsten trioxide,

photocatalytic oxidation of organic pollutants under solar irradiation.

H2O2 H1 e-HO H2O

Fig. 1 The possible route of degradation/mineralization of phenolic substrates. Reprinted

can be taken for large-scale application in pharmaceutical industries for

elaboration of carbide synthesis have led to significant improvement in

and reuse. Hence, it is urgent to design new kinds of tungsten-based

3 W-based catalyst supported on dierent carriers for

Scheme 1 The selective oxidation of 1,2-benzenedimethanol by H2O2. Reprinted from

heterogeneous catalyst with tungsten and zinc oxides supported on SnO2

spheres, the strong interaction, and the medium-strong acidity of the

uniform composite structure containing amorphous WO3 and anatase

Fig. 6 Schematic illustration of the consecutive photocatalytic process. Reprinted from

Fig. 7 Schematic illustration of the selective oxidation of 1,2-benzenedimethanol.

applied to the oxidative lactonization of 1,2-benzenedimethanol to

theoretical analysis revealed that the combined presence of Lewis and

Scheme 2 Proposed mechanism for DPPHWO3/Al2O3 catalyzed aerobic oxidation of

3.2 W-based catalyst supported on siliceous mesoporous molecular

Scheme 3 Plausible mechanism route to cyclohexane liquid-phase oxidation products

support, adsorbent, sensor and an excellent inorganic template for the

Scheme 4 The oxidation process of 1,3-butanediol by H2O2. Reprinted from Y. Su et al.,

Scheme 5 Incorporating catalytic oxotungsten tetrahedra centers into the framework of

These features extend its applications in the field of catalysis, molecular

In this section, several heterogeneous silica-supported WO3 catalysts

3.3 W-based catalyst supported on sustainable carbon materials

the photocatalytic degradation of CH3CHO using the composite was

Scheme 7 Possible degradation mechanism of CH3CHO over a g-C3N4/WO3 composite

Scheme 8 Proposed photodegradation mechanism of RhB over WO3 nanorods/