Vous êtes sur la page 1sur 217

Catalysis

Volume 21
A Specialist Periodical Report

Catalysis
Volume 21

A Review of Recent Literature

Editors
James J. Spivey, Louisiana State University, USA
Kerry M. Dooley, Louisiana State University, USA
Contributors
J. A. Anderson, University of Aberdeen, UK
Marco J. Castaldi, Columbia University, USA
Gabriele Centi, University of Messina, Italy
Claus Hviid Christensen, Technical University of Denmark, Denmark
Kresten Egeblad, Technical University of Denmark, Denmark
M. Fernandez-Garca, Instituto de Catalisis y Petoleoquimica (CSIC), Spain
Amit C. Gujar, Mississippi State University, USA
Charlotte C. Marsden, Technical University of Denmark, Denmark
Nora M. McLaughlin, Columbia University, USA
Siglinda Perathoner, University of Messina, Italy
Jeppe Rass-Hansen, Technical University of Denmark, Denmark
Esben Taarning, Technical University of Denmark, Denmark
Mark G. White, Mississippi State University, USA
Ye Xu, Oak Ridge National Laboratory, USA
If you buy this title on standing order, you will be given FREE access
to the chapters online. Please contact sales@rsc.org with proof of
purchase to arrange access to be set up.
Thank you

ISBN 978-0-85404-249-4
ISSN 0140-0568

A catalogue record for this book is available from the British Library

r The Royal Society of Chemistry 2009

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

Typeset by Macmillan India Ltd, Bangalore, India


Printed and bound by Lings, Dorchester, Dorset, UK
Preface
James J. Spivey* and Kerry M. Dooley*
DOI: 10.1039/b820546a

The eld of catalysis enjoys signicant scientic prominence due to its


importance in areas that aect the general publicclean energy, environ-
mental protection, and conversion of sustainable feedstocks, for example.
These driving forces, among others, will guide research eorts in our eld
for the foreseeable future. This volume of the Royal Society of Chemistrys
Specialist Periodical Reports: Catalysis book series addresses these issues
directly, providing up-to-date reviews on subjects of current interest.
First, Claus Christensen and colleagues Kresten Egeblad, Jeppe Rass-
Hansen, Charlotte Marsden, and Esben Taarning (Technical University of
Denmark, Lyngby) review the production of high-value chemicals and
intermediates from biomass. This is important because a wide range of
biomass feedstocks have the potential to replace fossil-based raw materials
to produce these end products. Among other challenges, heterogeneous
catalysts with extremely high activities and selectivities must be developed
to compete with current processes.
James Anderson and M. F. Garcia (Univ. Aberdeen, UK) show that the
signicant challenges in developing processes for water purication can be
addressed using photocatalytic reactions to remove both organic and
inorganic pollutants. They point out the diculties in studying the funda-
mentals of catalytic reactions in an aqueous medium, and the need to
improve the typically low quantum yield in the processese.g., by the
addition of noble metals to titania.
Gabriele Centi and Siglinda Perathoner (Univ. Messina, Italy), report on
approaches to the synthesis of titania catalysts, particularly ways to control
the structure at the nanometer scale. They show approaches to develop
specic active sites, and to direct the synthesis in a way that also produces a
local 3-D environment around the active site with desired properties.
Computational catalysis has enjoyed rapid progress as computer speed
and available codes have allowed more realistic catalytic cycles to be
studied. Ye Xu (Oak Ridge National Lab, USA) shows that the transition
in heterogeneous catalysis from a primarily empirical science to one that is
based on rst principles will provide new materials for experimental
research. Coupled with new imaging methods with greatly improved spatial
resolution, and atomically precise synthesis methods, computational
approaches hold great promise for the development of catalysts with
unprecedented levels of activity and selectivity.
In addition to their use as solvents, surfactants, and biocides, ionic liquids
are attractive for use in catalytic reactions due to their ability to activate
reactant molecules, the ease of separation from nal products, thermal
stability, solubility of gaseous reactants, among other properties. Amit
Gujar and Mark White (Mississippi State Univ., USA) show, for example,

Gordon A. and Mary Cain Dept. Chemical Engineering, Louisiana State University,
Baton Rouge LA 70803, USA. E-mail: jjspivey@lsu.edu. E-mail: dooley@lsu.edu

Catalysis, 2009, 21, 78 | 7


This journal is
c The Royal Society of Chemistry 2009
how these liquids can be used in a number of dierent catalyst-liquid
systems, e.g., monophasic systems in which the catalyst and substrate are
dissolved in the ionic liquid, or monophasic systems in which the ionic
liquid acts as both the solvent and the catalyst.
Finally, Nora McLaughlin and Marco Castaldi (Columbia University,
USA) provide a review of in situ techniques to study catalytic reaction
mechanisms. Because the catalyst is not static but can change during a
reaction, it is important to be able to characterize the surface at reaction
conditions. In addition, identication of reaction intermediates can help us
understand the reaction mechanism. The authors review surface measure-
ment techniques and recent developments in spectroscopy that can help us
examine these catalytic properties.
We greatly appreciate the eorts of the authors who have contributed to
this volume. We thank the Royal Society of Chemistry for their support of
this series. Comments are welcome.

8 | Catalysis, 2009, 21, 78


This journal is
c The Royal Society of Chemistry 2009
CONTENTS

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 Studios.

Preface 7
James J. Spivey* and Kerry M. Dooley*

Heterogeneous catalysis for production of value-added chemicals 13


from biomass
Kresten Egeblad, Jeppe Rass-Hansen, Charlotte C. Marsden,
Esben Taarning and Claus Hviid Christensen*
Introduction 13
Setting a new scene 14
Catalytic CC bond breaking 17
Catalytic hydrolysis 23
Catalytic dehydrations 25
Catalytic oxidations 32
Catalytic hydrogenations 39
Summary and outlook 43

Catalytic and photocatalytic removal of pollutants from aqueous sources 51


J. A. Anderson* and M. Fernandez-Garca
General introduction 51
Catalytic elimination of inorganics 53
Photocatalytic removal of inorganics 60
Catalytic and photocatalytic removal of organometallics 65
Catalytic and photocatalytic removal of organics 65
Removal of microorganisms 73

Catalysis, 2009, 21 | 9
This journal is
c The Royal Society of Chemistry 2009
Nano-architecture and reactivity of titania catalytic materials. 82
Part 2. Bidimensional nanostructured films
Gabriele Centi and Siglinda Perathoner
Introduction 82
Outlooks for the development of catalysts based 89
on the concept of nanostructured films
Synthesis of titania nanostructured films 100
Uses, with focus on catalysis 106
Conclusions 119

Recent advances in heterogeneous catalysis enabled by first-principles 131


methods
Ye Xu
Introduction 131
Theory-aided catalyst design 133
Molecular-level effects of reaction environment 137
Outlook 146

Ionic liquids as catalysts, solvents and conversion agents 154


Amit C. Gujar and Mark G. White
Introduction 154
Solubility of substrates in ionic liquids 155
Physical and chemical properties of ionic liquids 156
Demonstration of utility of RTILs as reaction solvents 157
Review articles 157
Synthesis of aluminum-containing ILs 158
Alkoxy carbonylation 159
Arene carbonylation 159
Catalytic oxidations 162
DielsAlder reactions in ILs 162
Dimerization of olefins in ILs 164
Enzyme-catalyzed reactions 165
Fischer esterifications in ILs 166
FriedelCrafts reactions in ILs 166
Heck reaction 169
Henry reactions in ILs 170
Hydrogenation in ILs 171
Hydroformylation 173
Isomerization 176
Metathesis of olefins 177
Michael reaction 177
Pechmann condensation in ILs 179
Sonogashira reaction 179
Sulfonation of arenes 181

10 | Catalysis, 2009, 21
This journal is
c The Royal Society of Chemistry 2009
Supported analogs of ionic liquid catalysts 181
Task specic ionic liquids (TSIL) 182
Telomerization in ILs 184

Measurement techniques in catalysis for mechanism development: 191


kinetic, transient and in situ methods
Nora M. McLaughlin and Marco J. Castaldi
Introduction 191
Structure to kinetics 192
Surface measurement techniques 193
Current in situ measurement techniques 197
Future directions 215

Catalysis, 2009, 21 | 11
This journal is
c The Royal Society of Chemistry 2009
Heterogeneous catalysis for production
of value-added chemicals from biomass
Kresten Egeblad, Jeppe Rass-Hansen, Charlotte C. Marsden,
Esben Taarning and Claus Hviid Christensen*
DOI: 10.1039/b712664f

1. Introduction
Almost everything around us is in some way a product of controlled
chemical processes. That is either chemical processes conducted in Nature
or chemical processes conducted in the chemical industry. In the most
developed parts of the World, it is in fact products from the chemical
industry that completely dominate our everyday lives. These products range
from fuels and fertilizers to plastics and pharmaceuticals.1 To make these
products widely available, a huge amount of resources have been invested
during the last century to develop the chemical industry to its current level
where it is the largest industry worldwide, a cornerstone of contemporary
society, and also a platform for further global economic growth.2,3 It can be
argued that the enormous success of the chemical industry can be attributed
to the almost unlimited availability of inexpensive fossil resources, and to a
continuously increasing number of catalysts and catalytic processes that
make it possible to eciently transform the fossil resources into all the
required compounds and materials. Accordingly, more than 95% of the
fuels and chemicals produced worldwide are derived from fossil resources,
and more than 60% of the processes and 90% of the products in chemical
industry somehow rely on catalysis. It has been estimated that 2030% of
the production in the industrialized world is directly dependent on catalytic
technology.4 Therefore, it is not surprising that we are continuously
expanding our already vast empirical knowledge about catalysis to further
improve the eciency of existing catalysts and processes, to discover
entirely new ways of valorizing available resources, and to lower the
environmental impact of human activities.5 Due to the overwhelming
importance of fossil resources during the 20th century, most catalysis
research eorts have, so far, concerned the conversion of these resources
into value-added fuels and chemicals. There are, however, indications that
the era of easy access to inexpensive fossil resources, especially crude oil, is
coming to an end. The resources are certainly limited and the demand from
everywhere in the world is growing rapidly. At the same time, it is becoming
increasingly clear that the emission of CO2 that follows the use of fossil
resources is threatening the climate of the Earth. Together this makes the
development of a chemical industry based on renewable resources one of the
most important challenges of the 21th century.
This challenge has two dierent facets. One is the discovery and develop-
ment of methods to use renewable resources to supply suitable energy
carriers, in sucient quantities at acceptable costs, and with minimal impact

Center for Sustainable and Green Chemistry, Department of Chemistry, Technical University
of Denmark, Building 206, Lyngby DK-2800, Denmark. E-mail: chc@kemi.dtu.dk

Catalysis, 2009, 21, 1350 | 13


This journal is
c The Royal Society of Chemistry 2009
on the environment. The other is the discovery and development of new ways
to provide all the chemicals needed to sustain a modern society. Whereas
there are several possible energy scenarios that do not involve carbon-
containing energy currencies, it is in fact impossible to envisage how it should
be possible to provide the required chemicals and materials without relying
extensively on carbon-containing compounds. Thus, to develop a chemical
industry that does not depend on fossil resources, there are only two
alternative carbon sources and that is CO2 and biomass. Since transformation
of CO2 into useful chemicals always requires a signicant energy input and
since it is usually only available in minute concentrations, it appears attractive
to instead utilize biomass as the dominant feedstock for chemical industry. In
this way, it is possible to harvest the energy input from the Sun that is stored
by photosynthesis in the CC, CH, CO, and OH bonds of the biomass.
Clearly, a shift from fossil resources to renewable resources as the preferred
feedstock in chemical industry is a formidable challenge. However, it is worth
pointing out that during the early part of the 20th century, before fossil
resources became widely available, biomass was the preferred feedstock for
the emerging chemical industry, and today, biomass still nds use as a
feedstock for a range of very important chemicals.6 Interestingly, these
processes often rely mostly on the availability of biological catalysts whereas
the processes for conversion of hydrocarbons use mostly heterogeneous
catalysts. However, to explore the full potential of biomass as a feedstock
in chemical industry, it appears necessary to integrate processes that rely on
biological catalysts with processes that use heterogeneous or homogeneous
catalysts to develop new, cost-competitive and environmentally friendly
technologies.7 Here, we will survey the possibilities for producing value-
added chemicals from biomass using heterogeneous catalytic processes.

2. Setting a new scene


2.1 Biomass for production of fuels and chemicals
Currently, there exists a strong focus on the manufacture of transportation
fuels from biomass.8,9 Clearly, this can be attributed to a desire to relinquish
our dependence on fossil fuels, in particular crude oil, and also to signicantly
lower the emission of greenhouse gasses to minimize global warming. In some
regions of the world, it seems that production of bio-ethanol is indeed already
cost-competitive with gasoline8 and this demonstrates the potential of biomass
as a renewable raw material. However, it is also clear that widespread use of
biomass as a raw material for biofuel production remains controversial from
both an economical and an ecological perspective. These issues must, of course,
be resolved soon in a fully transparent way to identify sustainable paths
forward. However, it is undisputable that we will eventually need alternatives
to the fossil resources for producing chemicals and materials.911 It can be
argued that if the amount of biomass available is too limited to substitute fossil
resources in all its applications and if suciently ecient methods for
transforming biomass into value-added chemical can be developed, this will
represent the optimal use of biomass.7 There are two reasons for this. First of
all, most chemicals, even most of the simple petrochemical building blocks, are
signicantly more valuable than transportation fuels. This can be illustrated in

14 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
a semi-quantitative way by comparing the value chains in a chemical industry
based on fossil and renewable resources, respectively.7 In this context, it is
instructive to compare the cost of renewable resources to fossil resources over
time. It is noteworthy that today, the cost of glucose is comparable to the cost
of crude oil (on a mass-to-mass basis). Secondly, it is clear that by use of
renewable resources as a feedstock for the chemical industry, signicantly
higher reductions in the emissions of green-house gases can be achieved than
what is possible by production of biofuels. This can be attributed to the fact
that production of many large-scale commodity chemicals from fossil resources
is associated with a substantial co-production of CO2 as expressed e.g., by the
C-factor (kg CO2 produced by kg of desirable product).7 This can often be
attributed to the high temperature required to transform hydrocarbons. To
illustrate this, the C-factor for industrial production of hydrogen from natural
gas is about 9 and for ethylene from naphtha it is 0.65. If hydrogen or ethylene
was produced eciently from biomass, the C-factor would approximately
express the amount of CO2 emission that would be saved compared to what
would be possible by production of biofuels instead. Since ethylene alone is
currently produced in an annual amount close to 100 mill. tonnes, it is obvious
that this would have a substantial impact on the total emission of green-house
gases.

2.2 Biomass in chemical industry


There are many ways in which biomass can be envisaged to become an
increasingly important feedstock for the chemical industry, and this has
already been the topic of numerous studies.1022 The most comprehensive
study was published recently by Corma et al.10 and it contains a very
detailed review of possible routes to produce chemicals from biomass.
In Fig. 1, we illustrate schematically how selected commodity chemicals
could be produced using abundant bio-resources, i.e., carbohydrates
(starch, cellulose, hemi-cellulose, sucrose), lipids and oils (rapeseed oil,
soy oil, etc.), and lignin as the sole raw materials. From these bio-resources,
it is possible to directly obtain all the compounds classied in Fig. 1 as
primary renewable building blocks (of which only selected examples are
given) with only one purication step. For example, ethanol can be obtained
by fermentation of sucrose, glucose by hydrolysis of starch, glycerol by
transesterication of triglycerides (or by fermentation of glucose), xylose
by hydrolysis of hemi-cellulose, fructose by hydrolysis of sucrose (and by
isomerization of glucose), and nally synthesis gas can be obtained directly
by gasication of most bio-resources or by steam-reforming of the other
primary renewable building blocks. From the primary renewable building
blocks a wide range of possible commodity chemicals can be produced in a
single step, and again examples of selected transformations are shown in
Fig. 1. For instance, acetic acid can be produced by fermentation of glucose
or by selective oxidation of ethanol. Lactic acid is available by fermentation
of glucose, and 5-hydroxymethyl furfural can be obtained by dehydration of
fructose. These compounds can again be starting materials for other
desirable products and so forth. Some of the commodity chemicals shown
are already produced on a large scale from fossil resources, e.g., ethylene,

Catalysis, 2009, 21, 1350 | 15


This journal is
c The Royal Society of Chemistry 2009
Fig. 1 Overview of how selected commodity chemicals could be produced from primary
renewable building blocks.

acetic acid, acrolein and butadiene. Others are envisaged to become


important large-scale commodity chemicals in the future when biomass
gradually becomes a more important feedstock.14 The dierent commodity
chemicals are labeled to categorize them according to their number of
carbon atoms. It is seen that a wide range of C1 to C6 compounds can be
made available by quite simple means. Moreover, the chemical transforma-
tions in Fig. 1 are labeled with dierent arrows to illustrate specic ways to
convert one building block into another. As it is apparent, the reactions all
require a suitable catalyst, and this can be either a biological catalyst or a
heterogeneous/homogeneous catalyst. Most of the primary renewable
building blocks are produced today from bio-resources using mainly
biocatalytic processes, and similarly several of the proposed commodity
chemicals can also be produced from the primary renewable building blocks
using biological catalysts. On the other hand, it is also clear that a very
substantial number of the desirable transformations rely on the availability
of suitable heterogeneous or homogeneous catalysts. Thus, it appears likely
that a chemical industry based on renewable resources as the dominant
feedstock will feature biological and chemical processes intimately integrated
to eciently produce all the desired chemicals and materials.

2.3 Heterogeneous catalysis and biomass


Often, it appears that the possible role of heterogeneous catalysis in this
scenario is not receiving sucient attention in comparison with that of the
biocatalytic methods. Therefore, in the present chapter we will highlight
some of the existing possibilities for converting bio-resources, primary

16 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
renewable building blocks, and commodity chemicals derived from these
into value-added chemicals. We will focus on production of chemicals that
can prove useful on a larger scale since they will contribute most to the
valorization of signicant quantities of biomass, and thereby contribute
most to relinquishing the dependence on fossil fuels and to lowering the
emission of green-house gases. Hopefully, this will be useful as a starting
point for others to discover and develop new reactions and catalysts that
can become useful in the eorts to make biomass a more useful resource for
chemical industry. Our emphasis here is the catalytic reactions and the
corresponding catalysts. Therefore, we have organized the literature
covered in separate chapters according to ve important reaction types,
specically, CC bond breaking, hydrolysis, dehydration, oxidation, and
hydrogenation. We envisage that these reaction types will be the most
important for producing value-added chemicals from biomass since they
can be conducted on large scale and they do not involve expensive reagents
that will make them prohibitively expensive for industrial applications.
Clearly, other reactions will also be important but several of those will be
analogues to current methods in chemical industry. In each chapter,
the presentation is organized hierarchically to rst discuss the catalytic
conversion of compounds that are most closely related to the bio-resources
(carbohydrates, lipids and oils, and lignin) and then successively those
derived from these renewable raw materials.

3. Catalytic CC bond breaking

3.1 Introduction
This section concerns catalytic processes that transform chemicals from
renewables by CC bond breaking. Among these are thermochemical
processes, such as pyrolysis and also gasication, catalytic reactions, such
as catalytic cracking and dierent reforming reactions, and decarbonylation
and decarboxylation reactions. Many of these reactions occur simulta-
neously, particularly in the thermochemical processes. Another technically
important class of CC bond breaking reactions is the fermentation pro-
cesses, however, they will not be considered in this section since they do not
involve heterogeneous catalysis.

3.2 CC Bond breaking reactions involving bio-resources


3.2.1 Crude biomass. Next to combustion, gasication is probably the
easiest and most primitive method for degradation of biomass. In the
simplest form, gasication involves heating of biomass (or any other
carbonaceous material) to temperatures around 800900 1C, in an atmos-
phere with only little oxygen, until it thermally decomposes into smaller
fragments. This partial oxidation process obviously requires a signicant
energy input and is not particularly selective; on the other hand, it is
reasonably exible since essentially all types of biomass can be gasied.
Gasication, in particular of coal, has been known for long and was
previously used to produce town gas. However, the gas resulting from
gasication has a relatively low heating value of only 1050% of that of

Catalysis, 2009, 21, 1350 | 17


This journal is
c The Royal Society of Chemistry 2009
natural gas,2325 and this was a major reason for replacing town gas with
natural gas. During World War II, biomass gasication advanced in
Europe, but it was not until the oil crisis in the 1970s that new developments
in the area truly took place.24 Today, the main purpose of biomass
gasication is to produce synthesis gas, with a H2:CO ratio close to two,
which is suitable for methanol synthesis or Fischer-Tropsch fuels.
There exist many dierent types of gasication furnaces but they generally
work by having several dierent cracking and reforming zones. These zones
are typically a pyrolysis zone, an oxidation zone and a reduction zone.
Biomass is broken down either by pyrolysis (without oxygen) or by partial
oxidation (with oxygen or air as oxidant) to a mixture of CO, CO2, H2O, H2,
CH4, other light hydrocarbons, some tar, char and ash, as well as some
nitrogen and sulfur containing gasses such as HCN, NH3, HCl, H2S etc.25
The hydrocarbons and the char are further partially oxidized to mainly CO
and H2O (14) and steam reformed (56) or dry reformed (79) to CO and
H2. The heat from the exothermic oxidation reactions is used to supply the
heat for the endothermic cracking reactions. Finally, the H2:CO ratio can be
adjusted by the water gas shift reaction (10).2326
CH4 + 1/2O2 = CO + 2H2 (1)
H2 + 1/2O2 = H2O (2)
CnHm + (n/2 + m/4)O2 = nCO + (m/2)H2O (3)
C + 1/2O2 = CO (4)
CnHm + nH2O = nCO + (n + m/2)H2 (5)
C + H2O = CO + H2 (6)
CnHm + nCO2 = 2nCO + (m/2)H2 (7)
C + CO2 = 2CO (8)
CH4 + CO2 = 2CO + 2H2 (9)
CO + H2O = CO2 + H2 (10)

The major challenge in gasication is to avoid the formation of tars, which


have a tendency to clog lters and condense in end-pipelines. Tars are
considered as the condensable fraction of the organic gasication products,
and consist mainly of dierent aromatic hydrocarbons with benzene as the
main species. For removal of tars three types of catalysts have been widely
investigated; alkali metal salts, alkaline earth metal oxides and supported
metallic oxides.2426
Alkali metal salts can be mixed directly with the biomass before entering
the gasication furnace. They enhance the gasication reactions and lower
the tar content, but recovery of the catalyst is dicult and costly making the
alkali metals unattractive as catalysts for industrial use.25,26 Another family
of catalysts, which can be used eectively for gasication, is the alkaline
earth metal oxides and carbonates. Of these, mainly the naturally occurring
mineral dolomite (MgCO3  CaCO3) has been used.25 It enhances the
degradation of especially the tars and hydrocarbons into light gasses,
though it is not active for methane reforming. When dolomite is calcined

18 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
at 800 1C, CO2 is eliminated, yielding a far more active catalyst. These
catalysts are deactivated by carbon formation and attrition but they are
inexpensive and disposable, and therefore easily replaceable. The third type
of catalysts used are metals on a support, typically nickel on various oxide
supports. Nickel catalysts are highly eective in tar destruction, the
reforming of hydrocarbons and in adjusting the composition of the syn-
thesis gas by the water gas shift reaction (10). They are operated as
secondary catalysts in a downstream reactor, which can be operated at
conditions dierent from those in the gasier. Nickel catalysts primarily
deactivate due to carbon formation and nickel particle sintering. Therefore,
dolomite is often used in guard beds upstream of the nickel catalyst bed to
remove most of the higher hydrocarbons.2426
Instead of gasifying biomass, it can be subjected to liquefaction in a
pyrolysis process. Pyrolysis is actually one of the main processes occurring
during gasication, however, in a dedicated pyrolysis plant, the desired
products are liquid hydrocarbons rather than synthesis gas. In the current
development of pyrolysis reactors, this is achieved by a fast pyrolysis
process. Here, the biomass is heated rapidly to temperatures of around
500600 1C, which leads to formation of a dark brown liquid known as
bio-oil along with some gasses and chars. Other types of liquefaction
processes are high pressure pyrolysis (350 1C, 20 MPa) and non-pyrolytic
liquefaction (aqueous/non-aqueous) (250425 1C, 1035 MPa).27 The liquid
products from these processes are of relatively pure quality with a heating
value of around half that of conventional oil. Alternatively to being used as
heating oil they can be upgraded to transportation fuels or chemical
feedstocks by hydrotreatment and catalytic cracking.
A possibly more sophisticated method for utilizing biomass to produce
synthesis gas is by aqueous phase reforming (APR), a processing method
that was developed for carbohydrates and other more readily accessible
biomass oxygenates by Dumesic et al.2832 Valenzuela et al.,33 however,
were the rst to report APR of real woody biomass. They used sawdust from
pine, which was milled to an average diameter of 375 mm. The biomass was
mixed with water, sulfuric acid (5%) and a catalyst (Pt/Al2O3) in a batch
reactor. The acid catalyzed the hydrolysis of the biomass to decompose
it nto smaller soluble molecules, which were reformed over the platinum
catalyst to yield mostly hydrogen and carbon dioxide. The process was
operated at 225 1C, with hydrogen accounting for 33% of the non-condensable
product gasses.

3.2.2 Bio-oils. In the 1970s, it was shown that bio-oils from plant extracts
such as rubber latex, corn oil, and peanut oil can be converted into a mixture
of mainly gasoline and liquid petroleum gas over a ZSM-5 catalyst, at
temperatures between 400500 1C.34 These bio-oils were investigated as
feedstocks for the reaction because they have high hydrogen to carbon
ratios and low oxygen contents and therefore a hydrocarbon-like structure.
It was suggested that such renewable plant resources, due to their siginifant
content of highly reduced photosynthetic products, would be suitable for
producing fuels or chemical raw materials.35 The high hydrogen-to-carbon

Catalysis, 2009, 21, 1350 | 19


This journal is
c The Royal Society of Chemistry 2009
ratios in the biomass feed is desirable because oxygen usually must be
removed and/or hydrogen must be added to achieve useful products.
Recently, several groups have investigated the catalytic conversion of
bio-oils or model bio-oils over HZSM-5 catalysts,3638 and recently a review
was published describing how biomass could be converted into fuels or
chemicals in a conventional petrochemical renery in FCC or hydrotreating
renery units.39

3.2.3 Carbohydrate resources. Carbohydrate resources, such as hydro-


lyzed starch and sucrose as well as xylose and glucose, can be processed into
hydrocarbons in a process similar to the one performed with bio-oils as
described above (section 3.2.2), i.e. by using a HZSM-5 catalyst operated at
around 510 1C and ambient pressure.40 This process is perhaps a little
surprising since carbohydrates do not resemble the desired hydrocarbon
product as much as the bio-oils do. However, formation of hydrocarbon
compounds was found to occur as a result of oxygen removal from the
carbohydrate by decarbonylation and decarboxylation reactions.40 This
process is probably one of the rst attempts to conduct catalytic cracking
of biomass.
Carbohydrate resources have also been processed under hydrotreating
conditions, i.e. high hydrogen pressures (35300 bar) and high temperatures
(300600 1C) in the presence of CoMo or NiMo-based catalysts; although
other precious metals like Ru and Pt can also be used.39 The main reaction
involved under these conditions is hydrodeoxygenation (HDO), as, for
example, described by Elliot et al.41 The important advantage of this
technology is that excellent fuels and useful chemicals can be produced in
good yields, but the process is expensive and requires high hydrogen
pressures.

3.3 CC Bond breaking reactions involving primary renewable building


blocks
3.3.1 Aqueous-phase reforming (APR). Aqueous phase reforming of
glucose, glycerol and other biomass oxygenates, such as methanol, ethylene
glycol and sorbitol, was carefully investigated by the group of Dumesic.2832
They showed how various biomass oxygenates can be converted into H2,
CO2 and some light alkanes with good conversions and high selectivities
over a Pt/Al2O3 catalyst operated at 225265 1C and 2956 bar,28 as well as
over a specially designed non-precious metal catalysts (Raney NiSn).29 It
was shown that this reaction could be used to supply hydrogen that could
simultaneously be used for reduction of sorbitol to hexane.30 This was
achieved using a bifunctional catalyst that caused sorbitol to be partly
cleaved over a metal catalyst (Pt, Pd) to form H2 and CO2 and at the same
time sorbitol was also dehydrated over a solid acid catalyst. By carefully
balancing these reaction steps, the hydrogen produced could be used
directly for hydrogenation of the dehydrated sorbitol to eventually yield
alkanes.30 Alternatively, hydrogen could be co-fed, whereby the production
of CO2 was avoided and the conversion to alkanes (especially hexane) is
improved.30

20 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
3.3.2 Steam reforming of ethanol. Steam reforming (SR) is probably the
most investigated process for breaking CC bonds in chemicals available
from biomass. Particularly, ethanol SR for production of hydrogen has been
extensively examined,4244 but also other primary renewable building blocks
have received attention, such as SR of glycerol45,46 or SR of bio-oils.47,48
SR of methane/natural gas is one of the largest catalytic processes in the
world and is by far the most important method for producing industrial
hydrogen today. The process is well described in literature and it is typically
carried out at 800950 1C over nickel-based catalysts.49 The main reactions
are methane SR (11) and water-gas-shift (WGS) (12).
CH4 + H2O = CO + 3H2 (11)
CO + H2O = CO2 + H2 (12)

SR of ethanol has mainly been conducted under similar conditions as


methane SR, which means relatively high temperatures, ambient pressure,
and primarily with Ni- or Rh-based catalysts.4244 Ideally, one mole of
ethanol is converted into 6 moles of hydrogen (13). During SR, ethanol
decomposes mainly through two dierent routes; either by dehydrogenation
to acetaldehyde (14) or dehydration to ethylene (15). These two intermediates
can be further catalytically reformed to a thermodynamically equilibrated
reaction mixture of H2, CO, CO2, CH4 and H2O (12, 1618).50
CH3CH2OH + 3H2O = 2CO2 + 6H2 (13)
CH3CH2OH = CH3CHO + H2 (14)
CH3CH2OH = CH2CH2 + H2O (15)
CH3CHO = CH4 + CO (16)
CH3CHO + H2O = 3H2 + 2CO (17)
CH2CH2 + 2H2O = 4H2 + 2CO (18)

A substantial diculty in ethanol SR is a too rapid catalyst deactivation due


to coking. This can occur by several reactions, such as methane decomposi-
tion (19) or the Boudouard reaction (20), but primarily the polymerization of
ethylene is thought to cause the problems (21). Unlike the situation for
methane SR, it appears that for ethanol SR the deactivation by coke
formation is lower at high temperatures.
CH4 = 2H2 + C (19)
2CO = CO2 + C (20)
CH2CH2 = polymeric deposits (coke) (21)

SR of ethanol is an endothermic reaction and relatively high temperatures


are required to convert ethanol into hydrogen and carbon monoxide and
eventually carbon dioxide after equilibration by the WGS reaction (12).
Thus, the drawback of this process is the energy requirements, which
perhaps are not so disadvantageous. If the hydrogen is used in a high
eciency fuel cell, compared to combusting the ethanol in a motor engine
with a relatively low eciency, the overall energy output could be signi-
cantly improved.50 Alternatively, the steam reforming reaction can be

Catalysis, 2009, 21, 1350 | 21


This journal is
c The Royal Society of Chemistry 2009
performed as a partial oxidation (22).51 Less hydrogen is formed in this way,
but instead the reaction is slightly exothermic, thus making hydrogen from
renewable resources without the need of adding extra energy in terms of heat.
CH3CH2OH + 2H2O + 1/2O2 = 2CO2 + 5H2 (22)

3.3.3 Decarbonylation. Furfural is easily obtained from biomass waste


such as oat and rice hulls that are rich in pentosans. Further valorisation of
furfural can be done by decarbonylation to produce furan, which can be
further converted into tetrahydrofuran by catalytic hydrogenation.
Pure decarbonylation typically employs noble metal catalysts. Carbon
supported palladium, in particular, is highly eective for furan and CO
formation.52 Typically, alkali carbonates are added as promoters for the
palladium catalyst.52,53 The decarbonylation reaction can be carried out at
reux conditions in pure furfural (165 1C), which achieves continuous
removal of CO and furan from the reactor. However, a continuous ow
system at 159162 1C gave the highest activity of 36 kg furan per gram of
palladium with potassium carbonate added as promoter.54 In oxidative
decarbonylation, gaseous furfural and steam is passed over a catalyst at
high temperatures (300400 1C). Typical catalysts are zinc-iron chromite or
zincmanganese chromite catalyst and furfural can be obtained in yields of
around 90% at full conversion.53 Again, addition of alkali metal carbonates
promotes the reaction.

3.3.4 Deformylation. Levulinic acid is used as a starting material for


the preparation of organic chemicals, dyes, polymers, pharmaceutically active
compounds and avoring agents. Acidic catalysts are required to procure
levulinic acid from sugars, and/or 5-HMF. Acidic ion-exchange resins
have been tested for dehydration of sucrose in pure water at 100 1C.55 And
levulinic acid could be achieved with up to 83% selectivity using all four tested
ion-exchange resins (Dowex MSC-1H, Amberlyst 15, Amberlyst XN-1010 and
Amberlyst XN-1005) although the overall yields were quite low (924%) even
after 24 h reaction times.55 Better results were achieved using zeolites as
catalysts. Zeolite LZY was tested for fructose dehydration in pure water at
various temperatures with the main product being levulinic acid formed in ca.
66% yield after 15 h at 140 1C.56 Levulinic acid was also observed as one of the
main products from aqueous phase dehydration of glucose using zeolite H-Y
(with a SiO2/Al2O3 ratio of 6.5) as well as with acidic montmorillonite clays as
catalysts,57,58 but signicantly lower yields were reported. With the possibilities
of levulinic acid as a renewable chemical building block, it seems interesting to
develop the zeolite-catalyzed process from cellulosic feedstocks.

3.3.5 Hydrogenolysis. CC and CO bond breaking by hydrogenolysis of


dierent polyols (glycerol, xylitol, erythritol and sorbitol) has been investi-
gated by Montassier et al.59,60 Predominantly ruthenium and copper-based
charcoal catalysts were studied at 210260 1C and 16 MPa hydrogen
pressures. The main products from the aqueous glycerol conversions were
propylene glycol using copper catalysts and ethylene glycol along with
methane using ruthenium catalysts. The hydrogenolysis of glycerol to

22 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
ethylene glycol and propylene glycol using ruthenium on a range of dierent
supports at 180 1C and 5 MPa hydrogen pressure showed the highest
conversion on a TiO2 support.61 Blanc et al. reported the treatment of
aqueous sorbitol solutions on CuOZnO catalysts at 180 1C 130 bar
hydrogen pressure.62 The purpose of the analysis was to achieve a high
C4+ selectivity suitable in the synthesis of alkyd polymers, and the CuOZnO
catalyst was superior in achieving a high C4+ selectivity (73% yield)
compared to Ru and Ni catalysts which mainly yielded C1C3 products.
A commercial example of a hydrotreating technology is examined below. The
IPCI (International Polyol Chemicals, Inc.) hydrogenolysis process is carried out
at 100300 1C and at hydrogen pressures of 70300 bar.63 The hydrogenolysis
process is used to cleave carbohydrates to smaller polyol fragments. Specically,
sorbitol and mannitol are reformed to propylene glycol and ethylene glycol as
the main products, and to dierent butanediols in smaller quantities.64 The
primary product, propylene glycol, is formed by hydrocracking either of sorbitol
directly (23) or, more likely, through glycerol (24,25).
C6H14O6 + 3H2 = 2C3H8O2 + 2H2O (23)
C6H14O6 + H2 = 2C3H8O3 (24)
C3H8O3 + H2 = C3H8O2 + H2O (25)

The composition of the hydrogenolysis products is very dependent on


the actual process conditions and on the catalysts used in the reaction. So
far, mostly supported nickel catalysts are being applied. IPCI has
constructed a 10 000 MT/y pilot plant in China in 2005, and in 2007, a
commercial 200 000 MT/y plant was commissioned, also in China.63

4. Catalytic hydrolysis
4.1 Introduction
Hydrolysis is the process by which a compound is broken down by reaction
with water, thus it can be thought of as the opposite reaction of dehydration,
where water is of course removed. Hydrolysis is a key reaction type in
biomass chemistry, for it is central in the depolymerisation of polysaccharides
to simpler monosaccharide building blocks, such as fructose, glucose, and
xylose.

4.2 Hydrolysis reactions involving renewable resources


4.2.1 Sucrose, maltose and cellubiose. Sucrose can be hydrolyzed to give
inverted sugars, i.e. a mixture of fructose and glucose (Scheme 1). For the
transformation of biomass into value-added chemicals, this is a key reaction
since it provides major building blocks for further chemical synthesis,
fructose and glucose, from widely occurring sucrose. In the past, and on
an industrial level, this reaction has been performed with the use of enzymes
as the catalyst. However, due to the production of waste, low thermal
stability, problems with separation of products and enzymes, and recovery,
and low rate due to glucose and fructose inhibiting the reaction, a dierent
path has been sought for.

Catalysis, 2009, 21, 1350 | 23


This journal is
c The Royal Society of Chemistry 2009
Scheme 1 Sucrose is hydrolyzed into a mixture of glucose and fructose when exposed to an
acid catalyst. Sulfuric acid has been used for this, but also heterogeneous solid acid catalysts
have found use, especially acidic ion exchange resins.

It has been established early that acids catalyze this hydrolysis reaction,
thus liquid sulfuric acid has been used. Heterogeneous catalysis can
potentially provide simpler and environmentally more benign processes,
however, via ease of separation and recovery. Thus, solid acids, such as
acidic ion-exchange resins,65,66 zeolites and heteropolyacids, can replace the
homogeneous acids. Hydrolysis of sucrose is in fact already established on
an industrial scale using acid ion-exchangers,67 but the main route is still
via enzyme catalysis. Transfer to the heterogeneous system shows problems
regarding the microenvironment of the swollen polymer, i.e. limitation
of diusion and restricted accessibility, as well as the production of many
by-products.68
In an eort to make this switch to heterogeneous catalysts viable, various
acidic exchange resins have been tested, including those prepared by
radiation-induced grafting to produce graft copolymers capable of hosting
sulfonic groups.6971 A common problem with solid acids, including ion-
exchange resins, is that they are subject to poisoning by water. Thus,
sulfonated mesoporous silicas were investigated as a new class of solid
acids, giving glucose and fructose in 90% yield after four hours at 80 1C.72
Zeolites are also acid ion-exchangers. A conversion of sucrose of up to 90%
with close to 90% selectivity and very few by-products formed was achieved
using highly dealuminated zeolite Y at 70 1C.73 Similarly, the activity of
various dealuminated zeolites was compared, again showing high selecti-
vities and few by-products, regardless of the conversion.74 Similarly, the
hydrolysis of maltose was studied by comparing the performance of acid
zeolites, ion-exchange resins, amorphous silica-aluminas and also the
ordered mesoporous material, MCM-41.75 The best results were achieved
with zeolite beta (Si/Al = 50) at 130 1C and 10 bar where a conversion of
85% and a selectivity of 94% was reported. Most recently, the use of
organicinorganic hybrid mesoporous silica catalysts was reported for the
hydrolysis of cellubiose to yield glucose. Cellubiose was used as a model for
oligosaccharides, and it was possible to achieve 100% conversion at 175 1C
but signicant glucose degradation was observed simultaneously.76
It can be seen then that heterogeneous catalysis may nd an opportunity
for replacing the enzymatic catalysis of disaccharides to its monosaccharides,
and thereby provide industry with a more ecient and benign route.
However, it is also clear that more selective catalysts are required.

4.2.2 Triglycerides. Triglycerides can be hydrolyzed to give fatty acids and


glycerol (Scheme 2). The fatty acids obtained have many industrial uses, mostly
for the manufacture of soap. Glycerol is currently viewed as a by-product from

24 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
this reaction, but maybe in the future it will be considered a commodity due to
the current drive to develop it as a feedstock. Technologies in this area have
often featured high temperatures and pressures because of low reaction rates.
In an attempt to develop low temperature and pressure processes, as well as
methods that are easy to implement, heterogeneous catalysis has been pursued
as an alternative.

Scheme 2 Triglycerides can be hydrolyzed to give fatty acids and glycerol. This can be
catalyzed by solid acid catalysts like zeolites or acid exchanged resins.

Similarly to the hydrolysis of sucrose, acid exchanged resins can be


utilized, in one case to give 75% hydrolysis of triglycerides after six hours
at 155 1C. It was shown that the Brndsted acid sites catalyze the hydrolysis
reaction, which was performed in the liquid phase with continuous steam
injection.77 The same authors reported that polystyrene sulfonic cation-
exchange resin, loaded with 13% of the superacid H3Mo, gave 74.5%
hydrolysis of palm oil at 155 1C in a batch reactor also operated with steam
injection.78

4.2.3 Polysaccharides. Before the introduction of enzymes (a-amylase


and glycoamylase) to facilitate the hydrolysis of polysaccharides, this
transformation was typically achieved using strong mineral acids. There
have also been studies of the use of ion-exchange resins and of the zeolite
mordenite to catalyze the hydrolysis of amylose and starch at 130 1C and
10 atm. With the ion-exchange resin, it was seen that the selectivity towards
glucose was not lowered by lengthening the reaction time. However, this
was not the case for mordenite where substantial degradation of the glucose
was observed.75 With the ion-exchange resin, it was possible to obtain 35%
glucose after 24 hours reaction time. Similarly, the performance of an ion-
exchange resin (Amberlyst 15), naon-silica and sulfonated mesoporous
silicas were compared for starch hydrolysis. The best yields reported were
39% glucose and 18% maltose obtained at 130 1C.79

5. Catalytic dehydrations

5.1 Introduction
There are several examples of dehydrations of chemicals derived by renew-
able resources by use of heteregeneous catalytic approaches in the literature.
These can be categorized into three types of reactions: (a) reactions in which
one (or more) molecule(s) of water is eliminated from a single substrate
molecule, (b) reactions in which one (or more) molecule(s) of water is
generated as the result of an esterication reaction between an alcohol and a
carboxylic acid or carboxylic acid derivative and (c) reactions in which one
(or more) molecule(s) of water is generated due to an etherication reaction
between two alcohol functionalities.

Catalysis, 2009, 21, 1350 | 25


This journal is
c The Royal Society of Chemistry 2009
5.2 Dehydration reactions involving bio-resources and primary renewable
building blocks
5.2.1 Bio-oils. Transesterication of vegetable oils to produce fatty acid
methyl esters (FAME) which can be used as biodiesel has been studied
intensely in recent years. Mainly solid bases such as MgO and hydrotalcites
are used as catalysts,80,81 however, solid acids are also studied for these
reactions.81 In a study with soybean oil as the source of fatty acids, MgO
catalysts prepared in dierent ways, as well as a hydrotalcite catalyst, were
all reported to be eective catalysts for the transesterication reactions
yielding between 75 and 95% FAME after 1 h at 180 1C, whereas
application of alumina resulted in less than 5% FAME.80 At 200 1C, all
the tested base catalysts (except one of the four MgO ones) resulted in a
yield of FAME of 95100%. In a similar study, experiments carried out at
180 1C showed a dierence in the yields obtained using hydrotalcite and
MgO catalysts.81 Using hydrotalcite, 92% yield was obtained, whereas the
yield using MgO was 75%. However, the yield obtained using these
catalysts were similar (7580%) when the reaction mixture also contained
some free fatty acid. In this study, pure and metal-substituted vanadyl
phosphates (MeVPO) as well as titanated silica (tetraisopropoxide titanium
grafted onto silica) were also tested for the reaction.81 The best results were
obtained with GaVPO with which a yield of 82% FAME yield was
obtained. The transesterication reactions have also been studied using
alumina-supported solid base catalysts at methanol reux temperatures, e.g.
using catalysts made by calcining KNO3 adsorbed on Al2O3.82 The study
showed that 35 wt% KNO3/Al2O3 calcined at 500 1C was the optimum
catalyst for the reaction, and this catalyst gave 87% yield after a reaction
time of 7 h. Recently, KF/Al2O3 has also been reported as catalyst for
transesterication at about 65 1C in a study using cottonseed oil as the fatty
acid source,83 and even poultry fat has been transesteried recently using
hydrotalcite as the catalyst.84

5.2.2 Syngas and methanol. Methanol is one of the top industrial


chemicals today. It is produced on a very large scale from fossil-derived
syngas by use of a CuZnAl-oxide catalyst, however, it can of course also
be produced in a similar manner from bio-derived syngas. Methanol (and
also syngas) can be used as a feedstock to produce dimethyl ether via
catalytic dehydration. However, the chemistry involved in these processes is
well-known, and will not be considered here, since it has been extensively
dealt with in detail elsewhere.8587

5.2.3 Ethanol. Ethanol is the most important chemical produced by


fermentation, and it has the potential to become a major feedstock for the
chemical industry since many other large-scale chemicals can be produced
from ethanol. In fact, ethanol can in many respects be considered a
renewable alternative to ethylene, which is the largest volume carbon-
containing chemical produced from fossil resources today. Via catalytic
dehydration, ethanol can easily be converted into ethylene and diethyl ether,
both of which are well-known acid catalyzed processes. Almost all available

26 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
solid dehydration catalysts have been tested for these reactions, and a
comprehensive review of this eld is beyond the scope of this review. The
reader is referred elsewhere for reviews on these topics.8891

5.2.4 Glycerol. It has long been known that glycerol can be dehydrated
to produce acrolein by heating aqueous glycerol with a mixture of nely
powdered KHSO4 and K2SO4 (Scheme 3).92 Recently, the reaction has
received attention again, as several acidic solid oxide catalysts were tested as
catalysts for the reaction.9395 The best results were obtained with silica-
supported heteropolyacids such as silicotungstic acid with which 86%
selectivity towards acrolein was obtained at 98% conversion of glycerol at
275 1C.93 Also tungstated zirconia has been reported to be a selective
catalyst for acrolein formation in a comparative study of many dierent
catalysts; using 15 wt% WO3/ZrO2, 65% selectivity towards acrolein was
achieved at 100% conversion at 325 1C.95

Scheme 3 Acrolein can be obtained by dehydration of glycerol. The reaction was reported
many years ago using powdered KHSO4/K2SO4 as catalyst. Recently, the use of silica-
supported heteropolyacids has also been described, notably with silicotungstic acid as catalyst.

Another dehydration product from glycerol is hydroxyacetone, or acetol


(Scheme 4). In one study, several catalysts were tested for this reaction.96 Of
the tested catalysts, however, only copper-chromite appeared to be eective
for this transformation. Using this catalyst, 80% selectivity towards
hydroxyacetone was achieved at 86% conversion in a reactive distillation
experiment carried out under a slight vacuum (98 kPa) at 240 1C.96

Scheme 4 Hydroxyacetone can be obtained from glycerol by dehydration. The reaction has
been reported using copperchromite as catalyst.

5.2.5 Xylose. Catalytic dehydration of xylose, which is the most abun-


dantly available pentose monomer in hemicellulose, has been known for a
long time (Scheme 5). In fact, as early as 1922, an industrial process
involving sulfuric acid catalyzed dehydration of xylose to produce furfural
was developed by the Quaker Oats Co.

Scheme 5 Xylose can be dehydrated to produce furfural. The reaction has been reported using
several dierent catalysts including zeolites, sulfonic acid functionalized MCM-41 and
immobilized heteropolyacids. The best selectivity towards furfural was achieved using zeolite
H-mordenite, although at low conversion of xylose.88 Overall, the best yield of furfural was
obtained using sulfonic acid functionalized MCM-41.

Catalysis, 2009, 21, 1350 | 27


This journal is
c The Royal Society of Chemistry 2009
Recently, several reports concerning this reaction have appeared in literature
describing the use of zeolites,97 ion-exchange resins,98 sulfonic acid functiona-
lized MCM-41,98 immobilized heteropolyacids,99101 niobium silicates102 and
exfoliated titanate and niobate nanosheet structures103 as solid acid catalysts.
In 1998, Moreau et al. compared zeolites HY and H-Mordenite in batch
experiments at 170 1C with 1:3 water/toluene or water/methyl isobutylketone
mixtures as the reaction media. It was found that HY was generally the most
active catalyst whereas H-mordenite was the most selective towards furfural
formation; up to 96% selectivity at 27% conversion after 30 min. using
water/toluene as the reaction media.97 Using sulfonic acid-functionalized
MCM-41, 82% selectivity at 91% conversion was achieved after 24 h at
140 1C using either DMSO or toluene/water as the extraction phase.98 Under
the same conditions, application of Amberlyst-15 resulted in only 70%
selectivity towards furfural at 90% conversion. Catalysts made by functiona-
lizing MCM-41 with heteropolyacids have also been tested for the reaction. In
general, however, the performances of these catalysts are not particularly good,
as the highest selectivity achieved in a study of several heteropolyacids using
dierent extraction phases were 67% at 94% conversion after 4 h at 140 1C.99
Moreover, microporous AM-11 niobium silicate and ordered mesoporous
niobium silicates have been reported as catalysts for dehydration of xylose,
however, at 160 1C no more than 56% selectivity at 89% conversion was
obtained using microporous AM-11.102 However, the study also showed that
the reaction temperature could be raised to 180 1C whereby the furfural yield
increased from ca. 20% after 1 h at 160 1C to ca. 45%. Very recently, exfoliated
and acidied layered titanates, niobates and titanoniobates prepared by heating
mixtures of TiO2 or Nb2O5 with alkali carbonates followed by immersion in
aqueous HCl or HNO3 and nally exfoliating the sheets with tetrabutyl-
ammonium hydroxide.103 Using these catalysts, furfural yields up to 55%
could be obtained at 92% conversion after 4 h at 160 1C.

5.2.6 Glucose and fructose. Sucrose is one of the largest chemicals readily
available from biomass. It is produced from sugar cane or sugar beats and
can be easily hydrolyzed into its constituent monomers, glucose and fructose.
In general, dehydration of these carbohydrates will lead to the formation of
many dierent products, however, some control of the dehydration products
obtained can be achieved using dierent acid catalysts. The target chemical
in most reports concerning solid acid catalyzed dehydration of hexoses is
5-hydroxymethyl furfural (Scheme 6), and to a lesser extent levulinic acid,
which is formed along with formic acid from HMF by a rehydration
decomposition reaction. HMF is sometimes referred to as a sleeping giant
due to its enormous potential importance as a key chemical intermediate,104
and several reviews are available concerning the production as well as
chemistry of HMF.105,106 Very recently, signicant achievements were made
in the production of HMF by homogeneous catalytic approaches.107 Typi-
cally, the starting material for HMF synthesis is fructose, however, there
exists, of course, great interest in establishing a commercially viable process
directly from glucose, since glucose is less expensive than fructose. Several
types of catalysts have been applied for dehydration of fructose to produce

28 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
HMF. These were categorized into ve classes of catalysts (organic acids,
inorganic acids, salts, Lewis acids and other) by Cottier et al. who also
categorized the dierent methods by which the dehydration reaction was
carried out into ve dierent types (aqueous media below 100 1C, aqueous
media above 100 1C, non-aqueous media, mixed-solvent systems and
solvent-free/microwave processes).105 Most of the solid acids applied in the
synthesis of HMF from fructose, zeolites, ion-exchange resins, solid inorganic
phosphates, belong to the group of other catalysts according to the
categorization of Cottier et al. Although levulinic acid can be an interesting
target in itself, it is nonetheless an undesirable byproduct in processes
targeting HMF, particularly, when water is used as the reaction media.
Therefore, the most successful approaches to circumvent levulinic acid
formation by HMF rehydrationdecomposition is to carry out the reaction
in a mixed water-organic solvent system, so that HMF is removed from the
aqueous phase as it forms. Literature covering the synthesis of levulinic acid
can be found in section 5.3.

Scheme 6 Fructose can be transformed into 5-hydroxymethyl furfural (HMF) via acid-
catalyzed dehydration. Solid acid catalysts applied to facilitate the reaction are zeolites, ion-
exchange resins and solid inorganic phosphates. With sporadic success, notably with inorganic
phosphates, other carbohydrate sources such as inulin can also be transformed into HMF.

With zeolites as the solid acid catalyst, the best results for HMF synthesis
were obtained by Moreau et al. who tested acidic mordenites with dierent
Si/Al ratios in batch experiments and reported that dealuminated H-Mordenite
with Si/Al ratio of 11 exhibited the highest selectivity and even so at reasonably
high fructose conversion (91% selectivity at 76% conversion after 60 min. at
165 1C using water/methyl isobutyl ketone as the reaction media).108 Other
zeolites, H-Y, H-Beta and H-ZSM-5 were also tested for the reaction, however,
none of these catalysts were as selective as H-mordenite.109
Also acidic ion-exchange resins were tested as catalysts for fructose dehydra-
tion. Using PK-216, a solution of water-DMSO-polyvinylpyrrolidone contain-
ing 10 wt% fructose was dehydrated to HMF after 816 h at 90 1C with 71%
selectivity at 80% conversion using MIBK as the extraction phase.107 Using a
more concentrated fructose solution (30 wt%), 65% selectivity was achieved at
83% conversion. Working in more dilute solution (0.5 M in DMSO), also with
PK-216 as the catalyst, an HMF yield of 90% was obtained after 5 h at
80 1C.110 Even more remarkable perhaps was the observation that the reaction
could be carried out in a continuous process with no signs of deactivation even
after 900 h, in this case using Amberlite IR-118 as the catalyst. Recently,
Amberlyst 15 was also reported as catalyst for fructose dehydration to produce
HMF at 80 1C using a solvent system comprising DMSO and either a
hydrophilic (BMIM-BF4) or hydrophobic (BMIM-PF6) ionic liquid.111 In
both cases, HMF yields of ca. 80% were achieved after 24 h, however, when
the reaction was carried out without DMSO as co-solvent, a maximum yield of

Catalysis, 2009, 21, 1350 | 29


This journal is
c The Royal Society of Chemistry 2009
only ca. 50% yield could be achieved (after 3 h) using BMIM-BF4. Very
recently, ionic liquids immobilized on silica (ILIS) were used as catalysts for
dehydration of fructose to HMF.112 In the study, it was shown that both Lewis
and Brndsted acidic ILIS were eective for the transformation. 70% yield was
obtained at 100% conversion after 4 min. of 200 W microwave irradiation in
DMSO using Brndsted acidic 3-allyl-1-(4-sulfobutyl)imidazolium triuor-
methanesulfonate [ASBI][Tf], whereas a yield of 67% was achieved using
Lewis acidic 3-allyl-1-(4-sulfurylchloride butyl)imidazolium triuormethane-
sulfonate [ASCBI][Tf].
A third class of solids, which were tested as catalysts for carbohydrate
dehydrations are inorganic phosphates. In a comparative study including
vanadyl phosphate (VOPO4  2H2O) and partially metal-substituted vanadyl
phosphates (MxVO1xPO4  2H2O, M being Cr, Mn, Fe, Al and Ga), it was
shown that vanadyl phosphate, which contains both Brndsted and Lewis
acid sites, is very selective towards HMF formation (80% selectivity at 50%
conversion after 1 h) under mild conditions (80 1C) and in pure water.113
Moreover, it was shown that partial Fe-doping increased the performance
quite signicantly, so that even at very high fructose concentrations, a
reasonable yield of HMF could be achieved (84% selectivity at 71%
conversion using 30 wt% fructose as the reaction media). The study also
showed that the catalyst performances are very similar when inulin is used
as the carbohydrate source in stead of fructose, perhaps opening up the
possibility of producing HMF from an even more inexpensive source than
fructose. Overall, the performance of the vanadyl phosphate catalysts at
80 1C was quite similar to the performance of zirconium and titanium
phosphate and pyrophosphate catalysts operated at 100 1C. Of the latter
two types of catalysts, the best performance was achieved with cubic
zirconium pyrophosphate and g-titanium phosphate, that gave selectivities
of up to 99.8% and 98.3% after 30 min., respectively.114 Also selectivities up
to ca. 100% were reported for niobium phosphate systems at 100 1C,
although at low conversions of fructose.115,116 Very recently, niobic acid and
niobium phosphate catalysts were also studied under continuous ow
conditions in aqueous medium. It was shown that the niobium phosphate
catalyst was more active than the niobic acid catalyst due to it having a
more acidic surface.117 The use of zirconium phosphates under subcritical
water conditions was also reported recently. In the study, it was shown that
HMF yield of 61% at 80% fructose conversion could be achieved after only
2 min. at 240 1C.118 Using glucose as the reactant under otherwise similar
conditions, only 39% selectivity was achieved.
Dehydration of glucose and fructose was also reported using solid oxide
catalysts. With anatase-TiO2 (a-TiO2) and a mixture of monoclinic and
tetragonal ZrO2 (m/c-ZrO2) it was reported that fructose could be relatively
selectively dehydrated to HMF after only 5 min. at 200 1C, although in quite
low yields (ca. 25%).119,120 With the same catalysts but with glucose as the
starting material, a-TiO2 was much more selective towards HMF than
m/c-ZrO2, which gave more or less the equilibrium mixture of 1,6-anhydro-
glucose and HMF, which was also obtained in the absence of a catalyst.
However, it should be noted that m/c-ZrO2 presumably catalyzes the
isomerization of glucose into fructose since more than 60 mol% fructose

30 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
was present in the reaction mixture after experiments starting from aqueous
glucose.
In 2000, Kroger et al. reported that 2,5-furan-dicarboxylic acid (FDCA)
could be produced in a combined one pot dehydrationoxidation reaction
starting from fructose, however, only in 25% yield.121 Very recently, this was
improved signicantly, when Ribeiro and Schuchardt reported that also
SiO2-gel containing Co(acac)3 could be used as catalyst for the combined
dehydrationoxidation reaction of fructose to yield FDCA.122 With pure
SiO2-gel, ca. 100% HMF selectivity could be achieved at ca. 50% conversion
after 65 min. at 160 1C. However, with Co(acac)3 encapsulated in SiO2-gel, a
70% one-pot yield of 2,5-furan-dicarboxylic acid was achieved. This proves
an important point, namely that an alternative strategy to improving the
HMF yield by adding an extraction phase is simply to react it further in situ
to desired end-products.

5.3 Dehydration reactions involving commodity chemicals


5.3.1 1,2- and 1,3-Propanediols. Dehydration of 1,2- and 1,3-propanediols
to produce allyl alcohol was studied using CeO2 as the catalyst.123,124 With
1,3-propanediol, the reaction is very specic towards allyl alcohol which is
formed with 99% selectivity at 51% conversion at 325 1C.123 At the same
temperature and with 1,2-propanediol as the substrate, only 44% selectivity
was achieved and at very low conversion. At elevated temperatures, higher
conversion of 1,3-propanediol can be achieved, however, at a substantial drop
in selectivity. At 425 1C, the selectivity towards allyl alcohol was 54% at 78%
conversion of 1,3-propanediol.

5.3.2 Succinic acid. Succinic acid is also available via fermentation of


glucose, and has the potential to become a large-scale industrial chemical in
the future. However, there are only a few reports on dehydration reactions
involving succinic acids in the literature, and most of these are concerned with
esterication to produce dialkyl esters. The synthesis of various dialkyl esters
was reported using metal exchanged montmorillonite clays (Na+, Mn2+,
Zn2+, Ni2+, Cr3+, Fe3+ and Al3+) as the catalysts.125127 For dimethyl
succinate, 70% isolated yield was achieved using Fe3+-exchanged mont-
morillonite after 4.5 h at methanol reux temperature.127 For dibutyl
succinate, the best results were obtained with Al-montmorillonite (94% yield
after 8 h, also at reux temperature), which also proved to be a good catalyst
for other esterication reactions, e.g. for diesterication with isobutyl alcohol
to produce di-(isobutyl) succinate in 98% yield.126 Very recently, a new family
of materials known as Starbons was also applied for esterication of succinic
acid with ethanol in aqueous ethanol solution.128,129 Using sulfonated
Starbon-400-SO3H as the catalyst, diethyl succinate was obtained in
499% after ca. 9 h at 80 1C.128

5.3.3 Levulinic acid and itaconic acid. Levulinic acid could also become an
important intermediate chemical in the future since it can be produced by acid
catalyzed dehydrationdecomposition of fructose. The synthesis of diethyl
levulinate was recently reported using sulfonated Starbon-400-SO3H. The

Catalysis, 2009, 21, 1350 | 31


This journal is
c The Royal Society of Chemistry 2009
selectivity towards the diester was 499% at 85% conversion after 6 h at
80 1C.128 In the paper esterication of itaconic acid was also reported.
However, this reaction is much slower and less selective, as ethyl itaconate
was achieved with 75% selectivity at 81% conversion after 24 h.128

5.3.4 Sorbitol. Sorbitol is the sugar alcohol obtained by reduction of


glucose and it can be dehydrated to either isosorbide or to 1,4- and 2,5-sorbitan
in acid or base catalyzed processes, respectively. Using sulfonic acid functiona-
lized MCM-41 type materials lauric acid esters of isosorbide can be achieved
quite selectively starting from sorbitol (495% selectivity towards isosorbide
dilaurate at 33% lauric acid conversion) in a dehydrationesterication
reaction.130

6. Catalytic oxidations

6.1 Introduction
Oxidation as a process to transform biomass into value-added chemicals is a
key one. Here, we focus on oxidations using molecular oxygen as the
oxidant, with the aim of illustrating selected interesting reactions that could
be important in the eorts to develop sustainable chemistry since they only
require abundant bio-resources as reactants and have water as the only, or
at least the main, byproduct.

6.2 Oxidation reactions involving primary renewable building blocks


6.2.1 Ethanol. Acetic acid, an important chemical reagent and industrial
chemical with a global demand of around 10 million tonnes per year, can be
produced from the oxidation of bioethanol. It has a very large number of
industrial applications e.g., in the production of cellulose acetate for
photographic lms and in polyvinyl acetate for wood glue. Moreover, it
also nds use in the food industry as an acidity regulator, as the additive
E260.
The oxidation of ethanol to acetic acid was among the rst heterogeneous
catalyzed reactions to be reported, but it has not attracted continued interest.
During the 1990ies, however, 100% conversion of ethanol coupled with
100% selectivity to acetic acid was reported in a gas-phase reaction using
molybdenum oxide catalytic systems on various supports, at temperatures
below 250 1C.131 Similarly, a tin oxide and molybdenum oxide catalyst was
used with a feed consisting of 80% aqueous ethanol to produce acetic acid at
320 1C.132 Recently, it was reported that a mixed Mo-V-Nb oxide also
catalyzes the selective oxidation of ethanol to acetic acid with oxygen in a gas
phase reaction at about 235 1C with 95% selectivity at full conversion at
about 235 1C.133
At even milder conditions, gold nanoparticles were found to be eective
heterogeneous catalysts for this reaction in aqueous phase. A yield of acetic
acid of 92% after eight hours and at the considerably lower temperature of
180 1C and 3.5 MPa air pressure was achieved, employing Au/MgAl2O4 as
the catalyst, and starting from ethanol concentrations comparable to those

32 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
obtained by fermentation (ca. 5 wt% aqueous ethanol). Platinum and
palladium, as traditional catalysts in this area, were compared to gold to
show that a superior selectivity and conversion can be achieved using gold
catalysis. At 180 1C, 3 MPa air pressure and after four hours, gold yielded
83% acetic acid, platinum 16% and palladium 60%. Moreover, gold
catalysts exhibited a selectivity of 97%, whereas platinum and palladium
gave only 82% and 93%, respectively.134 At higher ethanol concentrations
(above 60%), the same catalytic reaction leads primarily to formation of
ethyl acetate135 and the reaction was shown to proceed via acetalaldehyde as
intermediate.

6.2.2 Glycerol. Recently, the possibilities for oxidizing glycerol into


valuable chemicals have received signicant attention, and there are several
recent reviews with entire sections devoted to this particular topic.136138
The market for glycerol oxidation products is not yet developed due to the
current catalytic routes providing too low selectivities and yields. However,
since many expect that glycerol will be widely (and inexpensively) available
as a result of glycerol being a major waste product in the production of
biodiesel by transesterication,136 there is a signicant drive to develop new
catalytic technologies. Glycerol possesses two (identical) primary alcohol
functionalities and one secondary. Consequently, there exist a multitude of
possible products from the oxidation reaction. Catalytic selectivity is there-
fore a central factor to consider in developing this chemistry. In fact, the
nature of the metal, as well as the pH, largely controls the selectivity for
converting either type of alcohol. The main products possible from glycerol
oxidation are glyceric acid or glycerate depending on pH, dihydroxyacetone
(DHA), and glyceraldehyde. Under acidic conditions, oxidation of glycerol
usually leads to the formation of DHA (Scheme 7). In initial experiments
performed at pH of 24, platinum on charcoal showed low catalytic activity
for oxidation of the secondary hydroxy group of glycerol with a dihydroxy-
acetone yield of only 4% at a glycerol conversion of 37%. Several
promoters were tried, including bismuth, tellurium, lead, tin, and selenium.
On addition of bismuth (mass ratio of bismuth/platinum = 0.2), a drastic
increase was seen in the DHA selectivity, which increased from 10% to
80%.139 The bismuth addition decreases the conversion slightly, but the
yield still increased from 4% to 20%.

Scheme 7 Oxidation of glycerol in acid media leads to dihydroxyacetone, using PtBi/C as


catalyst.

The selective oxidation of a 50% aqueous solution of glycerol was


performed at 50 1C with an oxygen/glycerol ratio of 2, in a continuous
xed bed process using a PtBi catalyst supported on charcoal. Here, a
DHA selectivity of 80% at a conversion of 80% was obtained.140

Catalysis, 2009, 21, 1350 | 33


This journal is
c The Royal Society of Chemistry 2009
Under basic conditions, oxidation of glycerol mainly leads to the formation
of glycerate (Scheme 8). By employing a 5 wt% Pd/C catalyst, the selectivity
to glycerate can be as high as 70% at 100% conversion at pH = 11.141,142

Scheme 8 Oxidation of glycerol in basic media leads to glyceric acid, using Pd/C or Au/C as
catalyst.

During the last few years, signicant attention has been devoted to the
aerobic oxidation of glycerol using heterogeneous gold catalysts. Glycerol was
oxidized to glycerate in the presence of NaOH with 100% selectivity at 60 1C
using water as the solvent, and at an oxygen pressure of 0.3 MPa. The
catalysts used were either 1% Au/charcoal or 1% Au/graphite both giving
around 55% conversion.143 A range of dierent supports for the gold
nanoparticles catalysts were investigated, TiO2, MgO, and Al2O3, but all
showed low activity compared to Au/C. Depending on the base concentration
and the reaction time, the selectivity of the Au/C catalyzed liquid phase
glycerol oxidation could be controlled.144 Most recently, it was shown
that hydrogen peroxide is formed during oxidation of glycerol using gold
catalysts145 and that this leads to CC bond breakage and a resulting loss of
selectivity. This was independently supported by the fact the AuPd catalysts
showed higher selectivity to glycerate than the monometallic Au catalyst,
which was shown to be related to the higher eciency of Pd to catalytically
decompose the produced hydrogen peroxide in situ.146 By employing a
reaction temperature of 100 1C and an air pressure of 21 bar with methanol
as the solvent, it is possible to obtain dimethyl mesoxalate in yields as high as
89%147 and this clearly illustrates the eect of temperature on the degree of
oxidation of the glycerol feedstock.

6.2.3 Glucose. Glucose can be selectively oxidized to a number of


products. Currently, gluconic acid, glucuronic acid, glucaric acid and
2-keto-gluconic acid can be formed from such catalytic transformations.
All these oxidations can be performed with air or oxygen, in an aqueous
medium, under mild conditions and using a supported noble metal catalyst.
Owing to the multifunctionality of glucose, the possibility of controlling the
catalytic selectivity is again highly important. By use of Pt/C as the catalyst,
the anomeric carbon atom of glucose is oxidized most readily, followed by
the terminal primary alcohol moiety. The least oxidizable group is found to
be the equatorial alcohol moieties.148150 A plausible mechanism of the
reaction is that a dehydrogenation of glucose takes place to form gluconic
acid and adsorbed hydrogen on the platinum surface. The hydrogen is
then subsequently oxidized by oxygen to form water. This mechanism is
supported by the observation that gluconic acid is formed even in the
absence of oxygen in strongly basic solution and in the presence of the Pt
catalyst.148,149 The main product from the reaction of glucose is gluconic

34 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
acid (Scheme 9). This represents one of the rst examples where oxidation
occurs solely by dehydrogenation, and thus it is clearly an attractive option
with the only reagents required being simply water and hydroxide ions.151,152

Scheme 9 Oxidation of glucose to gluconic acid has been reported using oxidation catalysts
such as Pt/C, Pd/C, PdBi/C and Au/C.

Gluconic acid is found naturally in many fruits, in tea and in wine.


Gluconic acid is also used as an additive, to regulate pH in food (E574).
Currently, gluconic acid is produced industrially from glucose, using
glucose oxidase enzymes. The market of gluconic acid is in the range of
60.000 t/year.153 However, heterogeneous catalytic oxidation is also a viable
method, and one under currently intense investigation. It has been debated
whether the currently best heterogeneously catalyzed process is possibly
better than the industrial biocatalytic process.154
Oxidation is usually selective for the anomeric position, as mentioned
previously. Thus when glucose is oxidized in the presence of supported metal
catalysts, specically Pd and Pt, gluconic acid is achieved in high yields, for
example with 5% Pt/C, a 70% yield is obtained,155 or, also reported, complete
conversion is seen with Pd/C after six hours.156 However, when using
palladium catalysts deactivation is observed at high conversions. This problem
can be alleviated using a modied palladium catalyst. Thus, a PdBi/C
catalyst was found to be capable of oxidizing glucose to gluconic acid with
excellent selectivity (95100%) at rates up to 20 times greater than that of
the Pd/C catalyst.157,158 An oxidative dehydrogenation mechanism is also
proposed here.159
The use of gold as a catalyst in the oxidation of glucose to gluconic acid has
also been reported, on supports such as activated carbon, CeO2, TiO2, Fe2O3.
Gold exhibits somewhat lower selectivity than the Pd-Bi/C catalyst, though.
However, the activity of gold is strongly dependent on particle size, and it is
less sensitive to low pH, being active even under acidic conditions.160164 In
the aerobic oxidation of glucose over gold catalysts, hydrogen peroxide has
been observed to be a reaction product, just as it was the case in the oxidation
of glycerol.165 Most recently, it has proved possible to achieve long-term
stability (recycled 17 times without noticeable loss of activity) of an Au/TiO2
catalyst in the oxidation of glucose at pH = 9, and at temperature between
40 and 60 1C.166

6.3 Oxidation reactions involving commodity chemicals

6.3.1 1,2-Propanediol and 1,3-Propanediol. 1,2-propanediol and


1,3-propanediol can be obtained as hydrogenation products of glycerol.
The rst example of gold catalysts being able conduct aerobic oxidations of

Catalysis, 2009, 21, 1350 | 35


This journal is
c The Royal Society of Chemistry 2009
alcohols was provided by Prati and Rossi, who showed that 1,2-propanediol
can be oxidized at about 80 1C in alkaline aqueous solution to yield lactate
with a selectivity of 90100% at 8094% conversion using an Au/C
catalyst.167 Later, it was shown that similar results can be achieved using
Au nanosols stabilized with poly(vinylalcohol).168 If the reaction is instead
performed in methanol, it is only necessary to add a catalytic amount of
base, and under these conditions 1,2-propanediol and 1,3-propanediol
are oxidized to yield methyl lactate and methyl-3-hydroxypropionate,
respectively. Methyl lactate can be formed with 71% selectivity at 99%
conversion in the presence of 2 wt% Au/TiO2. The reaction requires about
20 hours at 100 1C and with 2.5 MPa air pressure.147 Similarly, methyl
3-hydroxypropionate can be produced with a somewhat higher selectivity of
85% at 99% conversion.147

6.3.2 Acrolein. Acrylic acid and acrylate esters constitute an important


group of chemicals in todays chemical indstry, for example, they are used
to make water-based paints, solvent-based coatings and acrylic coatings.
Typically, acrolein is obtained by catalytic oxidation of propene. However,
acrolein can also be obtained from glycerol, and in this way it can be
thought considered a renewable feedstock. Industrially, acrolein is oxidized
to acrylic acid in a gas-phase process operated at temperatures above 350 1C
employing mixed oxide catalysts.169 Therefore, it is noteworthy that methyl
acrylate can be produced from acrolein with 87% selectivity at 97%
conversion in the presence of catalytic amounts of Au/ZnO, suspended in
methanol, at room temperature and ambient pressure.170

6.3.3 Glyceric acid. The selectivity for the catalytic oxidation of glyceric
acid, and the calcium salt, can be controlled by the nature of the catalyst, and
the pH, in a similar way to that of glycerol as discussed above. In this way, it
is possible to obtain products corresponding to the oxidation of the primary
and secondary alcohol moieties, i.e., tartronic acid and hydroxypyruvic acid,
respectively. As reported by Fordham et al., oxidation of glyceric acid under
basic conditions leads to the formation of tartronic acid whereas hydro-
xypuric is aorded under acidic conditions.171 The catalyst was suspended in
glyceric acid, a calcium salt added, and oxygen gas bubbled through. NaOH
was added to keep the pH constant. Two catalytic systems were tested; 5 wt%
Pt/C at pH 10-11 yields tartronic acid with a selectivity of 60% at a
conversion of 94%; when 2% of bismuth is added, the same product is
obtained but with a selectivity of 83% at 90% conversion.171 Hydroxypyruvic
acid was obtained by aerobic oxidation of glyceric acid using a bismuth-
promoted platinum catalyst under acidic conditions (pH 34) to give a 64%
yield at 75% conversion.171 After prolonged contact with the catalyst,
tartronic acid was oxidized to oxalate whereas hydroxypyruvic acid was
oxidised even more rapidly to glycolic acid.

6.3.4 Lactic acid. Pyruvic acid and its derivatives are in increasing
demand due to their use as precursors in the synthesis of drugs and
agrochemicals.10 It has proved dicult to obtain pyruvic acid directly from

36 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
lactic acid via heterogeneous catalysis, because the major part of lactic acid
is converted to form acetaldehyde and CO2 by the oxidative C-C bond
cleavage over most catalysts, e.g. V2O5- or MoO3-based mixed oxide
catalysts.172
The vapor-phase oxidation of lactic acid with air was executed using an
iron phosphate catalyst with a P/Fe atomic ratio of 1.2. It was found that
lactic acid is selectively converted to form pyruvic acid by oxida-
tive dehydrogenation. The one-pass yield reached 50 mol%; however,
acetaldehyde, acetic acid, and CO2 was still formed, and the pyruvic acid
produced decomposes over time to give acetic acid and CO2.173
Oxidation has also been tried over iron phosphates with a P/Fe atomic
ratios of 1.2, including FePO4, Fe2P2O7 and Fe3(P2O7)2, at 230 1C. The
catalysts containing both Fe2+ and Fe3+ performed better than those with
just one oxidation state present. The best results were 62% selectivity at
60% conversion.174

6.3.5 Furfural. Furfural is a commodity chemical that is readily avail-


able from dehydration of pentoses, and it can be produced in a very large
scale if necessary. Furfural nds limited use today, though, which is
reected in its pricing being similar to the cheapest fossil bulk chemicals
such as benzene and toluene.10 In this way then, it can be seen that furfural
could become a key feedstock in the future, especially given the new turn
towards using bio-resources as a feedstock for chemicals when oil supplies
become more and more insecure and/or expensive. Furoic acid is used as a
feedstock in organic syntheses, and as an intermediate in the synthesis of
perfumes and medicines. The oxidation of furfural to furoic acid is mainly
described in patents, which discloses the use of various dierent catalysts
including Ag2O and Ag2O/CuO mixtures. However, during the 1990s, the
use of PbPt/C catalysts was also investigated.175 Very recently, methyl
furoate was synthesized using gold catalysis. Au/TiO2 was suspended in a
solution of furfural in methanol, a catalytic amount of sodium methoxide
was added, and methyl furoate could be produced with 90% selectivity at
more than 90% conversion.176

6.3.6 5-Hydroxymethyl furfural. 2,5-diformylfuran (DFF) is a furan


derivative that has many uses, including use as a polymer building
block.177,178 By utilizing a platinum catalyst supported on carbon, and
running the reaction in water at high temperatures, DFF is produced as the
major product in neutral solution. If low temperatures and high pH are
employed, 2,5-furandicarboxylic acid results.179
2,5-furandicarboxylic acid (FDCA) is another furan derivative available
from oxidation of HMF (Scheme 10). It holds a great promise in the
polymers industry because it can potentially replace terephthalic acid, which
is produced in a massive scale for making PET plastics.179,180 HMF is
converted to FDCA under strongly alkaline conditions, pH 12 or above,
with oxygen gas bubbled through the alkaline system. Platinum, or platinum
and a mixture of silver and copper oxides, preferably supported on carbon,
were employed as catalysts.

Catalysis, 2009, 21, 1350 | 37


This journal is
c The Royal Society of Chemistry 2009
Scheme 10 Oxidation of both the aldehyde and alcohol moieties of HMF leads to formation
of 2,5-furandicarboxylic acid. The oxidation reaction is catalyzed by Pt.

Oxidation of HMF was also attempted in situ directly from fructose, using
a membrane reactor or encapsulating PtBi/C into a polymeric silicone
matrix, and again, with air as the oxidant. However, the yield was
never more than 25%.181 A further attempt to obtain FDCA directly from
fructose involved a one pot reaction in the presence of cobalt acetyl-
acetonate encapsulated in solgel silica, at 155 1C and with 2 MPa of air
pressure giving FDCA with 99% selectivity directly from fructose at a
conversion of 72%.122
HMMF can be obtained from HMF by oxidation with Au/TiO2 (Scheme 11),
under very mild conditions25 1C, 1 bar O2 and 8% NaOMein near
quantitative yields. This is an intermediate on the route to FDMC, but the
reaction can be stopped at this stage.176

Scheme 11 Under mild conditions, oxidation of HMF in methanol can be tuned to yield
methyl (5-hydroxymethyl)-furoate via oxidation of the aldehyde moiety.

Similarly to above, but at a higher temperature and pressure130 1C and


4 bar O22,5-furan dimethyl furoate (FDMC), a direct analogue of
FDMA, can be obtained in near quantitative yields from HMF.176 It is
possible that the formation of a diester directly, rather than a diacid, could
save a further synthesis step on the route to polymerization.

6.3.7 Gluconic acid. Glucaric acid can be furnished by the selective


oxidation of the primary alcohol of gluconic acid with Pt-based catalysts.182
Platinum is preferred over palladium due its greater selectivity for the
oxidation of primary alcohols.183 Low rates of oxidation of primary the
alcohols is usually a complication, since products and byproducts bind
more strongly to the platinum surface than the primary alcohol moiety and
in eect poison the catalyst. Oxidation of secondary alcohols can also occur,
leading to the formation of highly oxidized species such as oxalic acid,
resulting in a poor selectivity towards gluraric acid. Gluconic acid was
oxidized to glucaric acid with 55% selecticity at 97.2% conversion, using a
Pt/C catalyst.182,184
When modifying the Pt catalyst by addition of bismuth or lead, a
signicant change in selectivity occurs. Oxidation of the primary terminal
alcohol moiety in gluconic acid is no longer the dominating reaction.
Instead oxidation of the a-hydroxy group, next to the carboxylic acid takes

38 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
place, resulting in the formation of 2-keto gluconic acid with a 98%
selectivity under slightly acidic conditions.184,185 It has been proposed that
the carboxylic acid coordinates to the promoter, as does the a-alcohol
group, thus promoting the oxidation of the a-alcohol group.184 This is
supported by the reaction proceeding in a weakly acidic medium; when run
in a basic medium, other products form and this is thought to be due to
further coordination with the remaining alcohol groups.186

7. Catalytic hydrogenations

7.1 Introduction
Catalytic hydrogenation represents a set of reactions that will be extremely
important in the production of value-added chemicals from biomass.
Already now, they play a signicant role in todays industry, and holds
great promise for further developments. Here, selected examples of hetero-
geneously catalyzed hydrogenations of chemicals available from renewables
resources are presented.

7.2 Hydrogenation reactions involving bio-resources and primary renewable


building blocks

7.2.1 Cellulose. The hydrolysis of polysaccharides, e.g. starch, inulin, can


also be combined with hydrogenation processes to yield polyols directly, in a
one step process. A one-pot process was previously reported using a homo-
genous catalyst based on Ru(TPPTS)3,187 however, heterogeneous catalysis
would be preferential in terms of the ease of recovery and re-use of the
catalyst. In this way, a heterogeneous system was developed whereby
ruthenium is supported on carbon, which is made acidic by treatment with
dierent oxidizing agents, thereby catalyzing the hydrolysis part of the
reaction. Selectivities to mannitol of 3740% were achieved, which is in line
with the yields from the non-coupled hydrolysis reaction, i.e. simply the
hydrogenation reaction, from fructose to mannitol.188 Cellulose, making up
around 4050% of biomass by weight makes it the largest component of
biomass. Cellulose is a linear polysaccharide consisting of many thousands of
glucose subunits. The glucose monomers are adjoined by 1-4-b glycosidic
bonds, which can be hydrolysed by strong acids at high temperature. Direct
hydrogenation of cellulose to sorbitol would be a highly desirable way to
valorize biomass. The major complication in this strategy is the insolubility of
cellulose in water. However, hydrogenation of cellulose to sorbitol has been
achieved in superheated water (190 1C) using platinum and ruthenium on
acidic supports.189 The highest activity can be achieved using a 2.5%
Pt/g-Al2O3 catalyst which aords sorbitol in 25% yield and mannitol,
resulting from epimerisation, is formed in 6% yield. The catalyst is reported
to remain active after several runs. However, the maximum overall yield is
limited to 31%, which is attributed to the complex structure of cellulose that
does not allow it to undergo complete hydrolysis under these conditions.
In a dierent study, the 1-4-b glucose dimer cellubiose was used as a
model substrate for cellulose. Using a polymer supported ruthenium

Catalysis, 2009, 21, 1350 | 39


This journal is
c The Royal Society of Chemistry 2009
nanocluster catalyst; cellubiose was converted to sorbitol in a one-pot
hydrolysis-hydrogenation reaction in 100% yield under acidic conditions
using an ionic liquid solvent.190 This procedure was found to result in a
complicated separation of the sorbitol from the ionic liquid and catalyst. A
modication to the procedure was reported recently, in which cellulose is
hydrolyzed in superheated water (200250 1C) and reduced using a carbon-
supported ruthenium catalyst.191 Cellulose conversion of 85% and 39%
selectivity towards hexitols (sorbitol and mannitol) was achieved with a
hydrogen pressure of 60 bar using this procedure.

7.2.2 Glycerol. Hydrogenation of glycerol to 1,2-propanediol or


1,3-propanediol has been reported using dierent metal catalysts including
nickel, copper, copper-chromite, ruthenium, rhodium palladium and
platinum.192195 For these reactions, the diculties lie in achieving either diol
with high selectivity. One method that has proven useful for producing
1,2-propanediol selectively is to carry out the hydrogenation reaction in the
presence of an ion-exchange resin in addition to a hydrogenation catalyst
containing Ru.193196 In this approach, the ion-exchange resins functions as a
dehydration catalyst which presumably facilitates the dehydration of glycerol
to hydroxyacetone, that is subsequently hydrogenated into 1,2-propanediol.

7.2.3 Glucose. Catalytic hydrogenation of glucose leads to the formation


of sorbitol (Scheme 12). Typically, Raney nickel is used to catalyze the
reaction,197 however, several other catalysts including platinum and ruthenium
have been reported to be active for the reaction, and in many cases these
catalysts are more eective than standard Raney nickel, which can be proble-
matic due to leaching of nickel.197200 Very recently, an impressive yield of
more than 99.5% sorbitol was obtained using Pt supported on microporous
activated carbon cloth.198 The experiments were conducted at 100 1C using a
40 wt% aqueous solution of glucose in a 300 ml stirred autoclave pressurized to
a hydrogen pressure of 80 bar. The high selectivity towards sorbitol exhibited
by the catalyst was attributed to fast desorption of sorbitol from the catalyst
surface, which eectively lowers sorbitol epimerization and thus suppresses the
formation of mannitol.

Scheme 12 Various catalysts have been applied to facilitate the catalytic hydrogenation of
glucose to sorbitol, notably Pt supported on activated carbon cloth using which 99.5% yield of
sorbitol can be obtained.

7.2.4 Xylose and fructose. Xylose can be hydrogenated into xylitol


(Scheme 13). This reaction was reported using hydrogenation catalysts such
as Raney nickel as well as platinum group metal catalysts.201,202

40 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
Scheme 13 Hydrogenation of xylose to produce xylitol can be achieved with 95% selectivity
using Raney nickel or Ru/C as catalyst.

Raney nickel, copper and platinum group metal catalysts have also been
used as catalysts for transforming fructose into mannitol via catalytic
hydrogenation (Scheme 14).197199,203,204 Ruthenium supported on carbon
is among the most studied catalysts for this reaction,205 and it is, in fact, also
eective for the combined hydrolysis-hydrogenation of inulin to mannitol
when the carbon support has been made acidic prior to the catalytic
experiments.188 The bifunctional catalyst applied in the study was made
by pre-oxidizing activated carbon (SX1G) with various oxidants such as
nitric acid and ammonium persulfate and then introducing Ru onto this
support by incipient wetness impregnation followed by reduction with
NaBH4. The oxidized carbon catalyzes the hydrolysis of inulin to a mixture
of glucose and fructose which is subsequently hydrogenated to a mixture of
glucitol and mannitol. Increasing the hydrogen pressure (up to 100 bar)
apparently also increases the rate of hydrolysis dramatically.

Scheme 14 Hydrogenation of fructose to mannitol is catalyzed by hydrogenation catalysts


such as Raney nickel and Ru/C.

7.3 Hydrogenation reactions involving commodity chemicals


7.3.1 3-Hydroxypropanal. 3-Hydroxypropanal can be formed by fermen-
tation of glucose and is thus an attractive starting material for production of
1,3-propanediol, which can be polymerized with tere-phthalic acid to produce
polytrimethylene terephthalate (PTT). PTT is used in the bers industry in
the production of stain resistant carpets etc.
Aqueous solutions of 3-hydroxypropanal were reduced using TiO2
supported ruthenium catalysts at 4060 1C using 40 bar of hydrogen.206
The most stable catalysts were found to be ruthenium catalysts supported
on low surface area macroporous rutile.

7.3.2 Lactic acid. Propylene glycol (1,2-propanediol) can be employed


as a de-icing agent replacing ethylene glycol, which is currently produced
from fossil resources. Furthermore, propylene glycol is a safe alternative to
ethylene glycol, which is toxic to humans due to its metabolism to oxalic
acid.

Catalysis, 2009, 21, 1350 | 41


This journal is
c The Royal Society of Chemistry 2009
Hydrogenation of lactic acid represents a simple route from a biomass
chemical to propylene glycol. Lactic acid has been hydrogenated in the
vapor phase using a Cu/SiO2 catalyst at 140220 1C with a hydrogen
pressure of 0.10.72 MPa. The selectivity of 1,2-propanediol was 88% at
full conversion, with 2-hydroxypropionaldehyde and propionic acid formed
as the major by-products.207 In a dierent study, aqueous phase hydro-
genation of lactic acid was achieved using a carbon supported ruthenium
catalyst. The hydrogenation is operated at temperatures from 100170 1C
with a hydrogen pressure of 714 MPa resulting in the formation of
1,2-propanediol in 90% selectivity at 95% conversion.208 Disappointingly
though, hydrogenation of salts of lactic acid did not result in the formation
of 1,2-propandiol. Hydrogenation has also been carried out using a
magnesia supported poly-g-aminopropysiloxane-ruthenium complex in
aqueous solution at 240 1C and 5 MPa hydrogen pressure for 18 hours,
giving 100% yield of 1,2-propandiol, with no apparent deactivation of the
catalyst.209

7.3.3 Furfural. Furfural is readily obtainable from dehydration of pen-


toses. Reduction of furfural can lead to a variety of products that are more
volatile, more stable and possibly also more useful than furfural itself. Selective
reduction of the aldehyde moiety leads to furfuryl alcohol (Scheme 15), whereas
further reduction of the furan core will lead to tetrahydrofurfuryl alcohol.
Reductive deoxygenation can result in the formation of either 2-methylfuran or
2-methyltetrahydrofuran, which can be used as liquid fuels or solvents.

Scheme 15 Hydrogenation of furfural to furfuryl alcohol is catalyzed by Cu-containing


catalysts.

Furfuryl alcohol has traditionally been obtained from furfural by hydro-


genation with copper containing catalysts, e.g. copperbariumchromium
oxide, copper oxide supported on silica or alumina, copperchromium
oxide and coppercobalt oxide on silica yields furfuryl alcohol as the major
product.210 Due to its toxicity, attempts have been made to eliminate
chromium from such catalytic systems, especially due to new restrictions
that prevent used copper chromite catalysts from being deposited in landll
sites.211 In recent years, several new catalytic systems have been demon-
strated to successfully catalyze the hydrogenation of furfural to furfuryl
alcohol. Copper has been investigated as a catalyst for this reaction on its
own. Furfural hydrogenation over copper dispersed on three forms of
carbonactivated carbon, diamond and graphitized berswas studied.
Similar to other copper-containing catalysts, only products corresponding
to hydrogenation of the carbonyl bond were detected, and the selectivity to
furfuryl alcohol was comparable to that obtained with commercial copper
chromite catalysts.212 Copper supported on magnesium oxide has also been
prepared, via the coprecipitation method, giving a 98% selectivity of
furfuryl alcohol at 98% conversion of furfural. This is attributed to the

42 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
higher number of surface copper sites and the defective sites at the copper
and magnesium oxide interfacial region. Similarly, CuCa/SiO2 catalysts
revealed a selectivity of 98% in the gas phase hydrogenation of furfural to
furfuryl alcohol at a conversion of 98% and a temperature of 130 1C.213
Various other catalytic systems have recently been used to promote hydro-
genation, including a molybdenum doped cobalt-boron amorphous catalyst
exhibiting excellent activity and nearly 100% selectivity to furfuryl alcohol
during liquid phase hydrogenation of furfural, at 100 1C and 1 MPa
hydrogen pressure.214
Reduction of the alcohol group to produce 2-methyl furan can be
achieved using a commercial Cu/Zn/Al/Ca/Na catalyst with the atomic
ratio 59:33:6:1:1. This catalyst was found to achieve 99.7% conversion with
87.0% selectivity to 2-methyl fural at 250 1C. Hydrogenation from furfuryl
alcohol yields a slightly higher selectivity of 92.7% at 98.1% conversion
under similar conditions.215

7.3.5 Levulinic acid. Hydrogenation of levulinic acid resulting in the


reduction of the ketone moiety leads to 4-hydroxy pentanoic acid. This acid
can cyclize to form g-valerolactone (GVL) which is a useful industrial
solvent. A 94% yield of GVL was obtained with a Raney nickel catalyst,
and a hydrogen pressure of 5 MPa at a temperature of 100150 1C.216
1,4-Pentanediol (PDO) holds promise for being used in the synthesis of
polyesters. It has been synthesized from GVL in the presence of a copper
chromite catalyst. At 150 1C and 20.330.4 MPa hydrogen pressure, 78.5%
PDO was produced together with 8.1% 1-pentanol.217

7.3.6 5-Hydroxymethylfurfural. 2,5-Di(hydroxymethyl)furan can be


synthesized from 5-HMF via hydrogenation (Scheme 16). Under a hydro-
gen pressure of 7 MPa at 140 1C in the presence of platinum or copper
catalysts, practically quantitative yields of 2,5-di(hydroxymethyl)furan
can be obtained.218 However, under similar conditions but with palladium
or nickel as catalyst, hydrogenation of the ring system occurs so that
2,5-di(hydroxymethyl)-tetrahydrofuran is obtained as the predominant
product.

Scheme 16 Hydrogenation of HMF to 2,5-di(hydroxymethyl)furan is catalyzed by Pt and Cu.

8. Summary and outlook


From a chemical perspective, renewable feedstocks being highly functiona-
lized molecules are very dierent from fossil feedstocks that are generally
unfunctionalized. Therefore, the challenge in converting fossil resources, in
particular crude oil, into useful products has been to develop methods that
allow controlled addition of desirable chemical functionality to the hydro-
carbon feedstock. Due to the quite low reactivity of the hydrocarbons; it has

Catalysis, 2009, 21, 1350 | 43


This journal is
c The Royal Society of Chemistry 2009
been possible to develop ecient catalytic processes that operate satis-
factorily at relatively high temperatures and pressures. Here, heterogeneous
catalysis has proven most successful and therefore played a dominant role in
chemical industry. The challenge of converting renewable feedstocks into
useful chemicals is very dierent. Still the desirable transformations are
entirely dependent on catalysis, but now it is often a question of controlled
removal of superuous chemical functionality under suciently mild con-
ditions to prevent uncontrolled degradation of the renewable feedstock. So
far, most emphasis has been on using biocatalytic processes to facilitate
these transformations but it appears likely that heterogeneous catalysis
could also play a signicant role in the future valorization of renewables.
Since the conversion of bio-resources into the primary renewable building
blocks is typically achieved using biocatalytic processes operating in water
as the natural solvent, it seems likely that there will be a signicant drive to
develop heterogeneous catalysts that also operate in water, and preferably
at low temperatures. In this way, it will be possible to achieve maximum
process integration between the biocatalytic processes and the hetero-
geneously catalyzed processes. This integration will lead to lower costs of
the resulting products since the need for expensive unit operations,
especially separations, will be minimized. Clearly, it represents a signicant
challenge to discover and develop heterogeneous catalysts that exhibit
sucient activity and selectivity under these conditions but it seems likely
that this will be one of the new directions that heterogeneous catalysis will
take during the next decade. The progress made in this endeavor will
obviously determine how large a role heterogeneous catalysis will eventually
play in the production of value-added chemicals from biomass. Here, we
have shown that there are several reaction types where heterogeneous
catalysis already oers some very promising opportunities but that there
clearly exists a great need for further discoveries and developments in this
emerging eld.

References
1 C. H. Christensen and J. K. Nrskov, J. Chem. Phys., 2008, 128, 182503.
2 S. Shaik, Angew. Chem., Int. Ed., 2003, 42, 3208.
3 S. Senkan, Angew. Chem., Int. Ed., 2001, 40, 312.
4 I. Maxwell, Stud. Surf. Sci. Catal., 1996, 101, 1.
5 Handbook of Heterogeneous Catalysis, eds. G. Ertl, H. Knozinger, F. Schuth
and J. Weitkamp, Wiley-VCH, Weinheim, 2008.
6 B. A. Tokay, Biomass Chemicals, in Ullmanns Encyclopedia of Industrial
Chemistry, Wiley-VCH, Weinheim, 2005.
7 C. H. Christensen, J. Rass-Hansen, C. C. Marsden, E. Taarning and
K. Egeblad, ChemSusChem., 2008, 1, 283.
8 J. Goldemberg, Science, 2007, 315, 808.
9 G. W. Huber, S. Iborra and A. Corma, Chem. Rev., 2006, 106, 4044.
10 A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107, 2411.
11 A. J. Ragauskas, C. K. Williams, B. H. Davison, G. Britovsek, J. Cairney,
C. A. Eckert, W. J. Frederick, Jr, J. P. Hallett, D. J. Leak, C. L. Liotta,
J. R. Mielenz, R. Murphy, R. Templer and T. Tschaplinski, Science, 2006, 311, 484.
12 D. R. Dodds and R. A. Gross, Science, 2007, 318, 1250.

44 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
13 P. Maki-Arbela, B. Holmbom, T. Salmi and D. Y. Murzin, Catal. Rev. Sci.
Eng., 2007, 49, 197.
14 Top Value Added Chemicals From BiomassResults of Screening for Potential
Candidates from Sugars and Synthesis Gas, eds. T. Werpy and G. Petersen, US
Department of Energy, Oak Ridge, TN, August 2004, vol. 1; available at
www.eere.energy.gov/biomass/pdfs/35523.pdf.
15 BioreneriesIndustrial Processes and Products, eds. B. Kamm, P. R. Gruber
and M. Kamm, Wiley-VCH, Weinheim, 2005.
16 B. O. Palsson, S. Fathi-Afshar, F. F. Rudd and E. N. Lightfoot, Science, 1980,
213, 513.
17 E. S. Lipinsky, Science, 1981, 212, 1465.
18 B. Kamm, Angew. Chem., Int. Ed., 2007, 46, 5056.
19 J. Rass-Hansen, H. Falsig, B. Jrgensen and C. H. Christensen, J. Chem.
Technol. Biotechnol., 2007, 82, 329.
20 F. W. Lichtenhaler and S. Peters, C. R. Chimia, 2004, 7, 65.
21 P. Gallezot, Catal. Today, 2007, 121, 76.
22 P. Gallezot, Green Chem., 2007, 9, 295.
23 M. A. Paisley, Biomass Energy, in Kirk-Othmer Encyclopedia of Chemical
Technology, Wiley VCH, Weinheim, 2003.
24 Milne et al. 1998_NREL/TP-570-25357.
25 Dayton 2002_Nrel/TP-510-32815.
26 D. Sutton, B. Kelleher and J. R. H. Ross, Fuel Process. Technol., 2001, 73, 155.
27 D. L. Klass, Biomass for Renewable Energy, Fuels, and Chemicals, Academic
Press, San Diego, 1998.
28 R. D. Cortright, R. R. Davda and J. A. Dumesic, Nature, 2002, 418, 964.
29 G. W. Huber, J. W. Shabaker and J. A. Dumesic, Science, 2003, 300, 2075.
30 G. W. Huber, R. D. Cortright and J. A. Dumesic, Angew. Chem., Int. Ed.,
2004, 43, 1549.
31 R. R. Davda, J. W. Shabaker, G. W. Huber, R. D. Cortright and J. A. Dumesic,
Appl. Catal. B, 2005, 56, 171.
32 G. W. Huber, J. N. Chheda, C. J. Barrett and J. A. Dumesic, Science, 2005,
308, 1446.
33 M. B. Valenzuela, C. W. Jones and P. K. Agrawal, Energy Fuels, 2006, 20, 1744.
34 P. B. Weisz, W. O. Haag and P. G. Rodewald, Science, 1979, 206, 57.
35 P. E. Nielsen, H. Nishimura, J. W. Otvos and M. Calvin, Science, 1977, 198, 942.
36 R. K. Sharma and N. N. Bakhshi, Biomass Bioenergy, 1993, 5(6), 445.
37 A. G. Gayubo, A. T. Aguayo, A. Atutxa, R. Aguado and J. Bilbao, Ind. Eng.
Chem. Res., 2004, 43, 2610.
38 A. Corma, G. W. Huber, L. Sauvanaud and P. OConnor, J. Catal., 2007, 247, 307.
39 G. W. Huber and A. Corma, Angew. Chem., Int. Ed., 2007, 46, 7184.
40 N. Y. Chen, T. F. Degnan, Jr and L. R. Koenig, Chemtech, 1986, August, 506.
41 D. C. Elliott, D. Beckman, A. V. Bridgwater, J. P. Diebold, S. B. Gevert and
Y. Solantausta, Energy Fuels, 1991, 5, 399.
42 A. Haryanto, S. Fernando, N. Murali and S. Adhikari, Energy Fuels, 2005, 19,
2098.
43 P. D. Vaidya and A. E. Rodrigues, Chem. Eng. J., 2006, 117, 39.
44 M. Ni, D. Y. C. Leung and M. K. H. Leung, Int. J. Hydrogen Energy, 2007, 32,
3238.
45 A. Behr, J. Eilting, K. Irawadi, J. Leschinski and F. Lindner, Green Chem.,
2008, 10, 13.
46 S. Adhikari, S. Fernando and A. Haryanto, Energy Fuels, 2007, 21, 2306.
47 D. Wang, S. Czernik, D. Montane, M. Mann and E. Chornet, Ind. Eng. Chem.
Res., 1997, 36, 1507.

Catalysis, 2009, 21, 1350 | 45


This journal is
c The Royal Society of Chemistry 2009
48 S. Czernik, R. Evans and R. French, Catal. Today, 2007, 129, 265.
49 J. R. Rostrup-Nielsen, J. Sehested and J. K. Nrskov, Adv. Catal., 2002, 47, 65.
50 J. Rass-Hansen, C. H. Christensen, J. Sehested, S. Helveg, J. R. Rostrup-
Nielsen and S. Dahl, Green Chem., 2007, 9, 1016.
51 G. A. Deluga, J. R. Salge, L. D. Schmidt and X. E. Verykios, Science, 2004,
303, 993.
52 P. Lejemble, A. Gaset and P. Kalck, Biomass, 1984, 4, 263.
53 H. E. Hoydonckx, W. M. Van Rhijn, W. Van Rihjn, D. E. De Vos and
P. A. Jacobs, Furfural and Derivatives, in Ullmanns Encyclopedia of Industrial
Chemistry, Wiley-VCH, Weinheim, 2007.
54 K. J. Jung and A. Gaset, Biomass, 1988, 16, 89.
55 R. A. Schraufnagel and H. F. Rase, Ind. Eng. Chem. Prod. Res. Dev., 1975, 14,
40.
56 J. Jow, G. L. Rorrer, M. C. Hawley and D. T. A. Lamport, Biomass, 1987, 14,
185.
57 K. Lourvanij and G. L. Rorrer, Ind. Eng. Chem. Res., 1993, 32, 11.
58 K. Lourvanij and G. L. Rorrer, Appl. Catal. A, 1994, 109, 147.
59 C. Montassier, J. C. Menezo, L. C. Hoang, C. Renaud and J. Barbier, J. Mol.
Catal., 1991, 70, 99.
60 C. Montassier, J. M. Dumas, P. Granger and J. Barbier, Appl. Catal. A, 1995,
121, 231.
61 J. Feng, H. Fu, J. Wang, R. Li, H. Chen and X. Li, Catal. Comm., 2008, 9,
1458.
62 B. Blanc, A. Bourrel, P. Gallezot, T. Haas and P. Taylor, Green Chem., 2000,
2, 89.
63 www.polyolchem.com.
64 T. Brix, BBI Conference, Portland, 911 October, 2007.
65 B. Satyanarayana and Y. B. G. Varma, Indian J. Technol., 1970, 8, 58.
66 A. Masroua, A. Revillon, J. C. Martin, A. Guyot and G. Descotes, Bull. Soc.
Chim., 1988, 561.
67 G. Siegers and F. Martinola, Int. Sugar. J., 1985, 87, 23.
68 C. Buttersack, React. Polymers, 1989, 10, 143.
69 T. Mizota, S. Tsuneda, K. Saito and T. Sugo, Ind. Eng. Chem. Res., 1994, 33,
2215.
70 T. Yoshioka and M. Shimamura, Bull. Chem. Soc. Jpn., 1984, 57, 334.
71 M. M. Nasef, H. Saidi and M. M. Senna, Chem. Eng. J., 2005, 108, 13.
72 P. L. Dhepe, M. Ohashi, S. Inagaki, M. Ichikawa and A. Fukuoka, Catal.
Lett., 2005, 102, 163.
73 C. Buttersack and D. Laketic, J. Mol. Catal., 1994, L283.
74 C. Moreau, R. Durand, F. Alie`s, M. Cotillon, T. Frutz and M. Theoleyre, Ind.
Crops Prod., 2000, 11, 237.
75 A. Abbadi, K. F. Gotlieb and H. van Bekkum, Starch/Starke, 1998, 50, 23.
76 J. A. Bootsma and B. H. Shanks, Appl. Catal. A, 2007, 327, 44.
77 C. J. Yow and K. Y. Liew, J. Am. Oil Chem. Soc., 1999, 76, 529.
78 C. J. Yow and K. Y. Liew, J. Am. Oil Chem. Soc., 2002, 79, 357.
79 P. L. Dhepe, M. Ohashi, S. Inagaki, M. Ichikawa and A. Fukuoka, Catal.
Lett., 2005, 102, 163.
80 M. D. Serio, M. Ledda, M. Cozzolino, G. Minutillo, R. Tesser and
E. Santacesaria, Ind. Eng. Chem. Res., 2006, 45, 3009.
81 M. D. Serio, M. Cozzolino, M. Giodano, R. Tesser, P. Patrono and
E. Santacesaria, Ind. Eng. Chem. Res., 2007, 46, 6379.
82 W. Xe, H. Pong and L. Chen, Appl. Catal. A, 2006, 300, 67.

46 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
83 C. Lingfeng, X. Guomin, X. Bo and T. Guangyuan, Energy Fuels, 2007, 21,
3740.
84 Y. Liu, E. Lotero, J. G. Goodwin Jr and X. Mo, Appl. Catal. A, 2007, 331, 138.
85 S. Lee and A. Sardesai, Top. Catal., 2005, 32, 197.
86 J. Rostrup-Nielsen, Proceedings of the World Petroleum Congress, 1998, 2, 767.
87 J. J. Spivey, Chem. Eng. Commun., 1991, 110, 123.
88 Y. C. Hu, Hydrocarbon Processing, 1983, april issue, 113.
89 Y. C. Hu, Hydrocarbon Processing, 1983, may issue, 88.
90 J. Schulz and F. Bandermann, Chem. Eng. Technol., 1994, 17, 179.
91 W. J. Toussaint, J. T. Dunn and D. R. Jackson, Ind. Eng. Chem., 1947,
february issue, 121.
92 Organic Syntheses, 1941, Coll. vol. 1, p. 15.
93 E. Tsukuda, S. Sato, R. Takahashi and T. Sodesawa, Catal. Comm., 2007, 8,
1349.
94 S.-H. Chai, H.-P. Wang, Y. Liang and B.-Q. Xu, J. Catal., 2007, 250, 342.
95 S.-H. Chai, H.-P. Wang, Y. Liang and B.-Q. Xu, Green Chem., 2007, 9, 1130.
96 C.-W. Chiu, M. A. Dasari and G. J. Suppes, AIChE Journal, 2006, 52, 3543.
97 C. Moreau, R. Durand, D. Peyron, J. Duhamet and P. Rivalier, Ind. Crops
Prod., 1998, 7, 95.
98 A. S. Dias, M. Pillinger and A. A. Valente, J. Catal., 2005, 229, 414.
99 A. S. Dias, M. Pillinger and A. A. Valente, Appl. Catal. A, 2005, 285, 126.
100 A. S. Dias, S. Lima, M. Pillinger and A. A. Valente, Carbohydrate Res., 2006,
341, 2946.
101 A. S. Dias, M. Pillinger and A. A. Valente, Micropor. Mesopor. Mater., 2006,
94, 214.
102 A. S. Dias, S. Lima, P. Brandao, M. Pillinger, J. Rocha and A. A. Valente,
Catal. Lett., 2006, 108, 179.
103 A. S. Dias, S. Lima, D. Carriazo, V. Rives and M. Pillinger, J. Catal., 2006,
244, 230.
104 M. Bicker, J. Hirth and H. Vogel, Green Chem., 2003, 5, 280.
105 L. Cottier and G. Descotes, Trends Heterocycl. Chem, 1991, 2, 233.
106 J. Lewkowski, Arkivoc, 2001, i, 17.
107 Y. Roman-Leshkov, J. N. Chheda and J. A. Dumesic, Science, 2006, 312,
1933.
108 C. Moreau, R. Durand, S. Razigade, J. Duhamet, P. Faugeras, P. Rivalier,
P. Ros and G. Avignon, Appl. Catal. AGeneral, 1996, 145, 211.
109 P. Rivalier, J. Duhamet, C. Moreau and R. Durand, Catal. Today, 1995, 24,
165.
110 Y. Nakamura and S. Morikawa, Bull. Chem. Soc. Jpn., 1980, 53, 3705.
111 C. Lansalot-Matras and C. Moreau, Catal. Commun., 2003, 4, 517.
112 Q. Bao, K. Qiao, D. Tomida and C. Yokoyama, Catal. Comm., 2008, 9, 1383.
113 C. Carlini, P. Patrono, P. Raspolli, A. M. Galletti and G. Sbrana, Appl. Catal.
A, 2004, 275, 111.
114 F. Benvenuti, C. Carlini, P. Patrono, P. Raspolli, A. M. Galletti, G. Sbrana,
M. A. Massucci and P. Galli, Appl. Catal. A, 2000, 193, 147.
115 C. Carlini, M. Giuttari, A. M. R. Galletti, G. Sbrana, T. Armaroli and
G. Busca, Appl. Catal. AGeneral, 1999, 183, 295.
116 T. Armaroli, G. Busca, C. Carlini, M. Giuttari, A. M. R. Galletti and
G. Sbrana, J. Mol. Catal. A, 2000, 151, 233.
117 P. Carniti, A. Gervasini, S. Biella and A. Auroux, Catal. Today, 2006, 118,
373.
118 F. S. Ashgari and H. Yoshida, Carbohydrate Res., 2006, 341, 2379.

Catalysis, 2009, 21, 1350 | 47


This journal is
c The Royal Society of Chemistry 2009
119 M. Watanabe, Y. Aizawa, T. Iida, R. Nishimura and H. Inomata, Appl. Catal.
A, 2005, 295, 150.
120 M. Watanabe, Y. Aizawa, T. Iida, T. M. Aida, C. Levy, K. Sue and
H. Inomata, Carbohydrate Research, 2005, 340, 1925.
121 M. Kroger, U. Prusse and K.-D. Vorlop, Top. Catal., 2000, 13, 237.
122 M. L. Ribeiro and U. Schuchardt, Catal. Commun., 2003, 4, 83.
123 S. Sato, R. Takahashi, T. Sodesawa, N. Honda and H. Shimizu, Catal.
Comm., 2003, 4, 77.
124 S. Sato, R. Takahashi, T. Sodesawa and N. Honda, J. Mol. Catal. A, 2004,
221, 177.
125 C. R. Reddy, P. Iyengar, G. Nagendrappa and B. S. J. Prakasha, J. Mol.
Catal. A, 2005, 229, 31.
126 C. R. Reddy, P. Iyengar, G. Nagendrappa and B. S. J. Prakash, Catal. Lett.,
2005, 101, 87.
127 M. L. Kantam, V. Bhaskar and B. M. Choudary, Catal. Lett., 2002, 78, 185.
128 V. L. Budarin, J. H. Clark, R. Luque and D. J. Macquarrie, Chem. Commun.,
2007, 634.
129 V. Budarin, R. Luque, D. J. Macquarrie and J. H. Clark, Chem. Eur. J., 2007,
13, 6914.
130 W. M. Van Rhijn, D. E. De Vos, B. F. Sels, W. D. Bossaert and P. A. Jacobs,
Chem. Commun., 1998, 317.
131 E. C. Alyea, K. F. Brown, L. Durham and I. Svazic, Stud. Surf. Sci. Catal.,
1992, 73, 309.
132 P. A. Awasarkar, A. Y. Sonsale and A. K. Chatterjee, React. Kinet. Catal.
Lett., 1988, 36, 301.
133 X. Li and E. Iglesia, Chem. Eur. J., 2007, 13, 9324.
134 C. H. Christensen, B. Jrgensen, J. Rass-Hansen, K. Egeblad, Kresten,
R. Madsen, S. K. Klitgaard, S. M. Hansen, M. R. Hansen, H. C. Andersen
and A. Riisager, Angew. Chem., Int. Ed., 2006, 45, 4648.
135 B. Jrgensen, S. E. Christiansen, M. L. D. Thomsen and C. H. Christensen,
J. Catal., 2007, 251, 332.
136 C.-H. Zhou, J. N. Beltramini, Y.-X. Fan and G. Q. Lu, Chem. Soc. Rev., 2008,
37, 527.
137 D. T. Johnson and K. A. Taconi, Environmental Progress, 2007, 26, 338.
138 M. Pagliaro, R. Ciriminna, H. Kimura, M. Rossi and C. D. Pina, Angew.
Chem., Int. Ed., 2007, 46, 4434.
139 H. Kimura, K. Tsuto, T. Wakisaka, Y. Kazumi and Y. Inaya, Appl. Catal. A,
1993, 96, 217.
140 H. Kimura, Appl. Catal. A, 1993, 105, 147.
141 R. Garcia, M. Besson and P. Gallezot, Appl. Catal. A, 1995, 127, 165.
142 N. Dimitratos, J. A. Lopez-Sanchez, D. Lennon, F. Porta, L. Prati and
A. Villa, Catal. Lett., 2006, 108, 147.
143 S. Carrettin, P. McMorn, P. Johnston, K. Grin and G. J. Hutchings, Chem.
Commun., 2002, 696.
144 S. Demirel-Guelen, M. Lucas and P. Claus, Catal. Today, 2005, 102, 166.
145 W. C. Ketchie, M. Murayama and R. J. Davis, Top. Catal., 2007, 44, 307.
146 W. C. Ketchie, M. Murayama and R. J. Davis, J. Catal., 2007, 250, 264.
147 E. Taarning, A. T. Madsen, J. M. Marchetti, K. Egeblad and C. H. Christensen,
Green Chem., 2008, 10, 408.
148 K. Heyns and H. Paulsen, Angew. Chem., 1957, 69, 600.
149 K. Heyns and H. Paulsen, Adv. Carbohydr. Chem., 1962, 17, 169.
150 K. Heyns, W. D. Soldat and P. Koll, Chem. Ber., 1975, 108, 3619.

48 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
151 G. Dewit, J. J. Devlieger, A. C. Kockvandalen, R. Heus, R. Laroy,
A. J. Vanhengstum, A. P. G. Kieboom and H. Van Bekkum, Carbohydr.
Res., 1981, 91, 125.
152 G. D. Wit, J. J. D. Vlieger, A. C. Kockvandalen, A. P. G. Kieboom and
H. Van Bekkum, Tetrahedron Lett., 1978, 1327.
153 F. W. Lichtenthaler, Acc. Chem. Res., 2002, 35, 728.
154 M. Comotti, C. D. Pina, E. Falletta and M. Rossi, J. Catal., 2006, 244, 122.
155 P. Gallezot, Catal. Today, 1997, 37, 405.
156 M. Besson, P. Gallezot, F. Lahmer, G. Fleche and P. Fuertes, in Catalysis of
Organic Reactions, eds. J. R. Kosak and T. A. Johnson, Marcel Dekker,
New York, 1993, vol. 53, pp. 169180.
157 A. Abbadi and H. Van Bekkum, J. Mol. Catal. A, 1995, 97, 111.
158 S. Karski, I. Witonska and J. Goluchowska, J. Mol. Catal. A, 2005, 245, 225.
159 M. Besson, F. Lahmer, P. Gallezot, P. Fuertes and G. Fleche, J. Catal., 1995,
152, 116.
160 A. Abbadi, M. Makkee, W. Visscher, J. A. R. Vanveen and H. Van Bekkum,
J. Carbohydr. Chem., 1993, 12, 573.
161 A. Abad, P. Concepcion, A. Corma and H. Garcia, Angew. Chem., Int. Ed.,
2005, 44, 4066.
162 A. Corma and P. Serna, Science, 2006, 313, 332.
163 A. Mirescu and U. Prusse, Catal. Commun., 2006, 7, 11.
164 S. Biella, L. Prati and M. Rossi, J. Catal., 2002, 206, 242.
165 M. Comotti, C. D. Pina, E. Falletta and M. Rossi, Adv. Synth. Catal., 2006,
348, 313.
166 A. Mirescu, H. Berndt, A. Martin and U. Prusse, Appl. Catal. A, 2007, 317,
204.
167 L. Prati and M. Rossi, J. Catal., 1998, 176, 552.
168 P. G. N. Mertens, M. Bulut, L. E. M. Gevers, I. F. J. Vankelecom,
P. A. Jacobs and D. E. De Vos, Catal. Lett., 2005, 102, 57.
169 T. V. Andrushkevich, Catal. Rev. Sci. Eng., 1993, 35, 213.
170 C. Marsden, E. Taarning, D. Hansen, L. Johansen, S. K. Klitgaard, K. Egeblad
and C. H. Christensen, Green Chem., 2008, 10, 168.
171 P. Fordham, M. Besson and P. Gallezot, Appl. Catal. A, 1995, 133, L179.
172 S. Sugiyama, N. Shigemoto, N. Masaoka, S. Suetoh, H. Kawami, K. Miyaura
and H. Hayashi, Bull. Chem. Soc. Jpn., 1993, 66, 1542.
173 M. Ai and K. Ohdan, Appl. Catal. A, 1997, 150, 13.
174 M. Ai and K. Ohdan, Appl. Catal., 1997, 165, 461.
175 P. Verdegeur, N. Merat and A. Gaset, Appl. Catal. A, 1994, 112, 1.
176 E. Taarning, I. Nielsen, K. Egeblad, R. Madsen and C. H. Christensen,
ChemSusChem, 2008, 1, 75.
177 A. Gandini and M. N. Belgacem, Prog. Polym. Sci., 1997, 22, 1203.
178 F. W. Lichtenthaler, S. Nishiyama and T. Weiner, Liebigs Ann., 1989, 1163.
179 F. W. Lichtenthaler and U. Kraska, Carbohydr. Res., 1977, 58, 363.
180 C. Moreau, M. N. Belgacem and A. Gandini, Top. Catal., 2004, 27, 11.
181 M. Kroger, U. Prusse and K. D. Vorlop, Top. Catal., 2000, 13, 237.
182 M. Besson, G. Fleche, P. Fuertes, P. Gallezot and F. Lahmer, Recl. Trav.
Chim. Pays-Bas, 1996, 115, 217.
183 A. W. Heinen, J. A. Peters and H. Van Bekkum, Carbohydr. Res., 1997, 304, 155.
184 P. C. C. Smits, B. F. M. Kuster, K. Van der Wiele and H. S. Van der Baan,
Appl. Catal., 1987, 33, 83.
185 P. C. C. Smits, B. F. M. Kuster, K. Van der Wiele and H. S. Van der Baan,
Carbohydr. Res., 1986, 153, 227.
186 A. Abbadi and H. Van Bekkum, Appl. Catal. A, 1995, 124, 409.

Catalysis, 2009, 21, 1350 | 49


This journal is
c The Royal Society of Chemistry 2009
187 A. W. Heinen, G. Papadogianakis, R. A. Sheldon, J. A. Peters and
H. van Bekkum, J. Mol. Catal. A, 1999, 142, 17.
188 A. W. Heinen, J. A. Peters and H. van Bekkum, Carbohydrate Res., 2001, 330,
381.
189 A. Fukuoka and L. Dhepe, Angew. Chem., Int. Ed., 2006, 45, 5161.
190 N. Yan, C. Zhao, C. Luo, P. J. Dyson, H. Liu and Y. Kou, J. Am. Chem. Soc.,
2006, 128, 87148715.
191 C. Luo, S. Wang and H. Liu, Angew. Chem., Int. Ed., 2007, 46, 7637.
192 M. A. Dasari, P. P. Kiatsimkul, W. R. Sutterlin and G. J. Suppes, Appl. Catal.
A, 2005, 281, 225.
193 D. G. Lahr and B. Shanks, J. Catal., 2005, 232, 386.
194 T. Miyazawa, Y. Kusunoki, K. Kunimori and K. Tomishige, J. Catal., 2006,
240, 213.
195 J. Chaminand, L. Djakovitch, P. Gallezot, P. Marion, C. Pinel and C. Rosier,
Green Chem., 2004, 6, 359.
196 T. Miyazawa, S. Koso, K. Kunimori and K. Tomishige, Appl. Catal. A, 2007,
329, 30.
197 B. W. Hoer, E. Crezee, P. R. M. Mooijman, A. D. van Langeveld,
F. Kapteijn and J. A. Moulijn, Catal. Today, 2003, 7980, 3541.
198 A. Perrard, P. Gallezot, J.-P. Joly, R. Durand, C. Baljou, B. Coq and P. Trens,
Appl. Catal. A, 2007, 331, 100.
199 H. C. M. Pijnenburg, B. F. M. Kuster and H. S. Van der Baan, Staerke, 1978,
30, 352.
200 J. Wisniak and R. Simon, Ind. Eng. Chem. Prod. Res. Dev., 1979, 18, 50.
201 J. Wisniak, M. Hershkowitz and S. Stein, Ind. Eng. Chem. Prod. Res. Dev.,
1974, 13, 232.
202 J. Wisniak, M. Hershkowitz, R. Leibowitz and S. Stein, Ind. Eng. Chem. Prod.
Res. Dev., 1974, 13, 75.
203 M. Makkee, A. P. G. Kieboom and H. Van Bekkum, Carbohydrate Res., 1985,
138, 225.
204 J. Wisniak and R. Simon, Ind. Eng. Chem. Prod. Res. Dev., 1979, 18, 50.
205 A. W. Heinen, J. A. Peters and H. van Bekkum, Carbohydrate Res., 2000, 328,
449.
206 M. Besson, P. Gallezot, A. Pigamo and S. Reifsnyder, Appl. Catal. A, 2003,
117124.
207 R. D. Cortright, M. Sanchez-Castillo and J. A. Dumesic, Appl. Catal. B, 2002,
39, 353359.
208 Z. Zhang, J. E. Jackson and D. J. Miller, Appl. Catal. A, 2001, 219, 8998.
209 B. W. Mao, Z. Z. Cai, M. Y. Huang and Y. Y. Jiang, Polym. Adv. Technol.,
2003, 14, 278.
210 G. Seo and H. Chon, J. Catal., 1981, 67, 424.
211 R. Rao, R. Dandekar, R. T. K. Baker and M. A. Vannice, J. Catal., 1997, 171,
406.
212 R. S. Rao, R. T. Baker and M. A. Vannice, Catal. Lett., 1999, 60, 51.
213 J. Wu, Y. Shen, C. Liu, H. Wang, C. Geng and Z. Zhang, Catal. Commun.,
2005, 6, 633.
214 X. Chen, H. Li, H. Luo and M. Qiao, Appl. Catal. A, 2002, 233, 13.
215 H. Y. Zheng, Y. L. Zhua, B. T. Teng, Z. Q. Bai, C. H. Zhang, H. W. Xiang
and Y. W. Li, J. Mol. Catal. A, 2006, 246, 18.
216 R. V. Christian, H. D. Brown and R. M. Hixon, J. Am. Chem. Soc., 1947, 69,
1961.
217 K. Folkers and H. Adkins, J. Am. Chem. Soc., 1932, 54, 1145.
218 V. Schiavo, G. Descotes and J. Mentech, Bull. Soc. Chim. Fr., 1991, 704.

50 | Catalysis, 2009, 21, 1350


This journal is
c The Royal Society of Chemistry 2009
Catalytic and photocatalytic removal
of pollutants from aqueous sources
J. A. Anderson*a and M. Fernandez-Garcab
DOI: 10.1039/b601303c

1. General introduction
On the 9th of February, 2004, in its resolution A/RES/58/217, the General
Assembly of the United Nations proclaimed the period from 2005 to 2015 as
the International Decade for Action, Water for Life, which commenced on
World Water Day, 22nd March 2005. 2015 was also a deadline year set by
world leaders at the United Nations Millennium Summit where a pledge
was made to reduce by 50% the proportion of people unable to reach or to
aord safe clean drinking water. Humanitys needs are met by only 1% of
the Earths total water and so careful use of this resource is essential. The
writing of this review coincides with the release of an RSC report on
sustainable water which highlights how increasing global population,
climate change and pollution will only exacerbate issues associated with
access to fresh water. The report also highlights the key role which can be
played by chemistry in aspects of sustainable water. In this review we
address only chemical solutions based on catalytic and photocatalytic
methods. While many of the issues associated with provision of potable
water in the developing world may be resolved by the use of simple physical
methodologies such as ltration, many of the issues associated with water
purity in the developed world involve complex, stable molecules present at
low concentrations but non the less capable of producing toxic eects in
plants and animals and which require removal technologies which are more
demanding. It is with these reagents that this chapter is principally
concerned.
The purication of drinking water from various water sources generally
combines a series of physical and chemical steps to eliminate the solid
fraction, kill bacteria and reduce the level of chemical pollutants. Many
processes include an oxidation step, based on the use of a strong oxidizing
agent such as chlorine, sodium hypochlorite, ozone or hydrogen peroxide.
These treatments generally lead to an acceptable level of micro-organisms in
water, however the process also aects the chemical composition. Further-
more undesirable ions, such as nitrates, remain unaected by such oxidative
treatments.

1.1 Catalytic and photocatalytic methods


The fundamental principles of catalysis are not required for the intended
readership of this article. However, there are several key points worthy of
attention which are fairly specic to water treatments. The rst is the
challenge presented by operating using water as a reaction medium if a full
a
Surface Chemistry and Catalysis Group, Kings College, University of Aberdeen, Aberdeen,
Scotland, UK AB24 3FX. E-mail: j.anderson@abdn.ac.uk
b
Instituto de Catalisis y Petoleoquimica (CSIC), Cantobalanco, Madrid 28049, Spain

Catalysis, 2009, 21, 5181 | 51


This journal is 
c The Royal Society of Chemistry 2009
kinetic mechanism including identication of reaction intermediates and
active sites is to be established. Application of most spectroscopic techni-
ques is limited in the presence of water thereby posing a real challenge for
those who wish to conduct in situ and operando type investigations. The
second concerns the additional species which are present in waste water
which may well act as poisons to the supported base and noble metal
catalysts which predominate the studies which have been performed in this
area. Additionally, as the composition of waste water varies from source to
source, it is dicult to predict how a laboratory or event pilot plant test
catalyst system may perform when the composition of the wastewater is
changed. Thirdly, the large volumes of reactant which may require treat-
ment make the need to perform reaction at elevated temperatures unrealistic
in terms of energy cost and thus high activity (and selectivity) at the
temperature of the local water temperature, which may be only a few
degrees above freezing, is desirable. The nal point worth considering is that
most catalytic methods of pollutant control, involve the use of hydrogen or
oxygen as reagents. As the Henrys constants in water are 7.8  104 and
1.3  103 mol L1 atm1 at 298 K, respectively,1 and most kinetic studies
report positive reaction orders with respect to the gases listed, pollutant
conversion may be slow if performed using atmospheric pressures of
reactant gases.
Photo-induced processes are studied in several industrial-oriented appli-
cations which have been developed since their rst descriptions in the
scientic literature. Despite the dierence in character and utilization, all
photo-induced processes have the same origin. A semiconductor can be
excited by light energy higher than the band gap inducing the formation of
energy-rich electron-hole pairs. By photo-catalysis it is commonly under-
stood that we are referring to any chemical process catalyzed by a solid
where the external energy source is an electromagnetic eld with wave-
numbers in the UV-visible range.27 Customarily, photo-catalysts are solid
semiconductors which are (i) able to absorb visible and/or UV light,
(ii) chemically and biologically inert and photostable, (iii) inexpensive and
(iv) non-toxic. TiO2, ZnO, SrTiO3, CeO2, WO3, Fe2O3, CdS and ZnS can
act as photoactive material for redox processes due to their electronic
structure which is characterized by a lled valence band and an empty
conduction band. Among these possible semiconductors, TiO2 is the most
used photocatalytic material as it fulls all of these requirements and
exhibits adequate conversion values.8 However, in spite of the high conver-
sion values obtained for TiO2, the calculated quantum yield for the studied
reactions is appreciably low; certainly well below 10% for most degradation
processes.9 TiO2 occurs in nature in three crystallographic phases: rutile,
anatase and brookite, anatase being the most commonly employed in
photocatalytic applications due to its inherent superior photo-catalytic
properties.7,8,10 Anatase is the least thermodynamically stable polymorph
of TiO2 as a bulk phase, although from energy calculations, it appears as the
more likely phase when the grain size is below 15 nm.11 High surface TiO2
materials would thus present the anatase polymorph as a general rule. The
crystalline structure of the TiO2 oxides can be described in terms of TiO6
octahedral chains diering by the distortion of each octahedron and the

52 | Catalysis, 2009, 21, 5181


This journal is 
c The Royal Society of Chemistry 2009
assembly pattern of such octahedra chains. The TiTi distances in the
anatase structure are greater than in rutile, whereas TiO distances are
shorter.12 These structural dierences cause dierent mass densities and
lead to dierent electronic structure of the bands. The anatase phase is 9%
less dense than rutile, presenting more pronounced localization of the Ti 3d
states and, therefore, a narrower 3d band. Also the O 2p-Ti 3d hybridisation
is dierent in the two structures (more covalent mixing than in rutile), with
anatase exhibiting a valence and conduction band with, respectively, more
pronounced O 2p-Ti 3d characters and less non-bonding self-interaction
between similar ions (e.g. anion-anion and cation-cation interactions).13
The importance of the covalent vs. ionic contributions to the MetalOxygen
bond have been discussed in a more general context for several photo-
catalytic oxides by Wiswanathan.14 In any case, these dierential structural
features which distinguish anatase and rutile are presumably responsible for
the dierence in the mobility of charge carriers during light excitation.
In the quest to optimise the photo-activity of anatase-TiO2 systems,
several paths have been pursued. Almost from the conception of the use
of photocatalysis, the photo-activity of anatase-TiO2 systems was typically
improved with the addition of surface noble metals such as Pt, Pd, and Ag
which act as electron trapping centres and/or oxide-oxide contact using
SnO, ZrO2 and others with appropriate electronic structure to positively
inuence the electron-hole charge separation process.2,6,7,15 Another metho-
dology, central in current research, considers the extension of the solid light-
absorption spectrum to the visible region. This would facilitate the use of
sunlight, an inexpensive, renewable energy source, as the excitation energy
of the photocatalytic processes. Several alternatives, mainly based on the
use of (solid inorganic and polymer materials, molecular) sensitizers,
hypothetical cubic-type TiO2 polymorphs, or the use of other inorganic
solids (suldes, nitrides, and other non-Ti-based oxides), have been tested in
pursuing this goal.29,16 The majority of studies involving TiO2 aimed at
improving its optical absorption and photocatalytic performance, are
focused on anatase modication by cation and/or anion doping.2,57,17
Doping may also result in surface modication as well as changes to the
electron-hole charge handling properties of the anatase-TiO2 systems. All of
these systems typically require deposition on appropriate supports to
facilitate separation of the aqueous media and reuse.6,7,9 In brief, all of
these systems are customarily used to enhance photo-activity of anatase-
TiO2 systems and constitute materials currently working on the elimination/
degradation of water pollutants.

2. Catalytic elimination of inorganics


All forms of water, whether wastewater, ground water, untreated reservoir
water or water treated for human consumption, are complex and may
contain inorganic and/or organic species. Particularly important in this
respect is the use of seawater for production of drinking water, where
signicant quantities of dissolved inorganic salts need to be considered.18
Dissolved inorganic species, both cations and anions, can strongly inuence
the removal of organic species and thus must be carefully considered.

Catalysis, 2009, 21, 5181 | 53


This journal is 
c The Royal Society of Chemistry 2009
Inorganic residues present in water can be simply classied in terms of the
anions such as phosphates, sulfates, nitrates, chlorides, etc. and cations such
as transition metals and alkaline and alkaline earths. As the methodology
for removal of these species tends to be somewhat specic and reported
studies tend to tackle these on an individual basis, the following
sections look at these on a species by species basis. However, it is clear
that remediation technologies must involve integration of the dierent
approaches used to treat the individual contaminants.

2.1 Nitrate elimination


The potentially harmful eects of exposure to high concentrations of
nitrates in drinking water result from reduction to nitrites, which combine
with haemoglobin to form methaemoglobin (blue baby disease). Addition-
ally, nitrosamine formation can cause cancer and hypertension. In nature,
high levels of nutrients, such as nitrates, lead to eutrophication of water
sources, which in, severe cases, lead to the extermination of the other
aquatic life due the decreased levels of oxygen and luminosity.
Nitrogen pollution from agricultural sources is characterized by temporal and
spatial variability, depending on the interplay between anthropogenic eects
and physical processes occurring in nature (climate, soil and topography). The
main European legislation dealing with protection of water resources from
nitrogen based agricultural pollution, is the Nitrate Directive (ND) (EC, 1991).
The Water Framework Directive (WFD) (EC, 2000) is a broader concept for
sustainable management of water resources. Limits within the EU are 50, 0.1
and 0.5 ppm for NO3, NO2 and NH4+, respectively. Monitoring data
show that a nitrate concentration of 50 ppm is exceeded in around one third
of the European groundwater bodies for which information is currently
available (EEA, 2003). The problem is particularly pronounced in agri-
cultural areas where, according to the Dobris Assessment, 87% of areas have
nitrate concentrations in groundwater above the 25 ppm guide level, and
approximately 20% of these exceed the 50 ppm limit.
Potential abatement technologies of nitrates from aqueous environments
include physical-chemical, electrochemical, biological and catalytic meth-
ods. Physical-chemical processes are based on ion-exchange, electrodialysis
or reverse osmosis procedures that have the benet of being able to remove
the pollutant without the introduction of other substances. However these
processes are not eective in attempting to target particular anions in
isolation and a potential impact on water quality may result. Normally, a
secondary step is required for the destruction of the sludge containing the
removed species since the process only removes the contaminants from the
water. Electrochemical, catalytic reduction and bacterial degradation are
more promising processes because they convert nitrates to N2. The main
disadvantages of the electrochemical methods are the low rate of the
reduction and the production of undesirable side-products such as nitrites,
ammonia and N2O (a greenhouse gas). Biological denitration is a more
sustainable process than the electrochemical processes because it is based on
existing technologies, namely organic matter decontamination.

54 | Catalysis, 2009, 21, 5181


This journal is 
c The Royal Society of Chemistry 2009
Biological treatments oer good selectivity when combined with eective
sludge processing. Although bacteria provide an economically viable solu-
tion for water decontamination,1921 the process is slow and its eciency
relies on maintaining a rather constant level of nitrates in water, which
makes it less suitable for applications in regions where seasonal variation in
nitrate levels exist. Additionally, there is a minor risk of sulfate reduction to
H2S when working at low nitrate levels which may occur when treated water
is used for dilution of heavily contaminated water reservoirs.
The potential for catalytic removal of nitrate was introduced in the late eighties
and bears some similarity to the electrochemical process in as much as both
involve stepwise reduction of the nitrogen oxidation state via formation of nitrites,
ammonia and N2. As with the electrochemical process, the major drawback is the
potential for formation of undesired products such nitrite and ammonium which
must be depleted at more stringent levels than the nitrate itself.

2.1.1 Catalyst formulation-selection of active metals. Catalytic removal


of nitrates was proposed in the eighties22 and the work involved in catalyst
selection developed in subsequent publications23,24 and more recently sum-
marised in a book chapter.25 The eective reduction of nitrite and subsequent
intermediates is plausible over a monometallic catalyst whereas a bimetallic
catalyst was viewed as being more appropriate for nitrate reduction. As
avoidance of high levels of intermediate nitrite is crucial, several commercial
catalysts were screened for the reduction of this anion.24 Pd containing
samples showed good activity and high selectivity towards N2 with low rates
of ammonium formation, whereas catalysts containing Ru, Ir and Rh showed
poor activity with ammonia as the main reaction product. Pt catalysts were
also active but generated unacceptable levels of ammonium. Studies compar-
ing PdCl2, Pd(NO3)2 and Pd(NH3)4(OH)2, showed the latter to give best
results and deposition-precipitation gave better results than impregnation for
alumina support whereas the opposite was found in the case of silica. Well-
dispersed Pd catalysts, with uniform particle distribution as prepared by
deposition-precipitation, show little variation in activity or ammonium
formation as a function of loading in the range 0.2 to 5%.25
As nitrate reduction was found to be favoured in a bimetallic congura-
tion, screening was performed to select the most appropriate partner for Pd
in this catalyst.25,26 Cu, Zn, Sn and In addition all showed promise although
the former and latter gave the most active materials. The use of Fe, Mn, Pb,
Ni, Au, Pt or Rh all led to poorly active catalysts with high levels of
ammonium formation.25 Favourable catalytic performances of Sn modied
Pd/alumina catalysts have been ascribed to the presence of highly dispersed
alloyed PdSn species, with decreased electron density on Pd due to the Sn
leading to the observed increase in catalyst selectivity.27 A 4:1 weight ratio
for Pd/Sn is reported as giving the most selective catalysts although activity
could be enhanced by increasing the ratio to 4:2.5.26 In the case of Pd/Cu, a
weight ratio of 4:1 gave the most active and selective alumina supported
catalyst.24 A Cu/Pd surface atom ratio of 5:1 has been suggested as
optimum for selective nitrate-to-nitrite reduction.28,29 In the case of PdIn,
activity was increased as a function of In addition although a 6:1 ratio was

Catalysis, 2009, 21, 5181 | 55


This journal is 
c The Royal Society of Chemistry 2009
the most favourable in terms of minimising ammonium formation. In
terms of the nature of the 2 metal component systems, it is suggested
that an active bimetallic site requires only close proximity between the
two metals such that the spill-over hydrogen originating on the noble metal
sites can access the copper sites by simple surface migration.3032 This is
supported by in situ X-Ray Absorption Spectroscopy results which show
that whereas Pd retains its metallic state throughout, copper is oxidized
during the course of the nitrate reduction process and only in the presence
of Pd and hydrogen can the Cu be returned to its active, reduced metallic
state.33 The extent of interactions between components in the bimetallic
system may be inuenced by the nature of the reducing and oxidising
pretreatments which will subsequently inuence the activity for nitrate
reduction.32

2.1.2 Catalyst formulation-selection of support. In addition to the use


of silica and alumina as supports, studies using PdCu have involved
TiO2,3436 niobia37 and activated carbons38,39 Hydrotalcite, modied by
addition of Cu to its formulation was found to be more active and selective
catalyst than Pd/Cu/Al2O3.40 The results were ascribed to the hydrotalcites
capacity to concentrate anions between its layers. Although the materials
exhibited a low surface area during reaction, the nitrate concentration near the
active sites was expected to be high as a consequence of ionic forces in the layer
spacings.
A number of studies have highlighted the possibility of the use of mono-
metallic Pd catalyst for eective nitrate/nitrite reduction if a reducible support
is selected.36,41,42 Epron et al.41 proposed a reaction mechanism for Pd/CeO2,
(Fig. 1) in which oxygen vacancies at the Pd-support interface were active sites
for the reduction of nitrate. Although the selectivity to nitrogen was poor, the

Fig. 1 Nitrate reduction mechanism using oxygen vacancies of the support.

56 | Catalysis, 2009, 21, 5181


This journal is 
c The Royal Society of Chemistry 2009
paper was followed by a number of articles where partially reducible supports
were used as Pd promoters in monometallic and bimetallic forms.
Gavagnin et al.42,43 using SnO2 as support, obtained activities comparable
to those of bimetallic catalysts and with a higher selectivity. The latter was
attributed to the low support surface area and low support porosity which
prevented the occurrence of mass transfer limitations and ensured an
eective solution buering, i.e., removal of hydroxide groups from the
vicinity of catalyst active sites.
Titania36 as with the ceria,41 led to Pd promotion in terms of activity for
nitrate degradation but yielded catalyst which exhibited poorer selectivity than
bimetallic catalysts. However, unlike ceria based systems,41 these catalysts36
did not suer from poisoning when CO2, was used as a pH buer. Nitrate
reduction over Pd/TiO2 catalysts was associated with partially reduced
titania species which migrated over the Pd particles during the reductive
pre-treatment. The latter also led to the formation of a Pd b-hydride phase
which was also thought to play a role in the reduction process.

2.1.3 Reduction kinetics, mechanism and selectivity. A rate law deter-


mined for PdCu/Al2O3 at 283 K indicates an order of 0.7 with respect to
nitrate and zero with respect to hydrogen pressure, if the latter is greater
than 1.0 bar.44 Pintar et al.45 noted however, that at higher nitrate
concentrations, the reaction was zero order but became rst order as nitrate
levels were depleted. The same authors found that the kinetics could be
described in terms of a LangmuirHinshelwood type mechanism, which
accounted for both non-competitive and equilibrium adsorption of nitrate
and dissociative hydrogen adsorption steps, as well as an irreversible
bimolecular surface process that controlled the overall reaction. The
enthalpy of adsorption for nitrate was 22 kJ mol1 and the apparent
activation energy was determined as 47 kJ mol1. The low activation energy
conrmed the selection of PdCu catalysts for the process. The same
authors45 proposed that the kinetics of nitrate removal were consistent
with a mechanism involving ionic intermediates, which required participa-
tion of dierent types of active sites. The possibility of heterogeneous
homogenous free radical mechanisms was discarded. The bimolecular
reaction between adsorbed species was considered to be the rate-limiting
step which was thought to occur via heterolytic electron transfer.45 The rate
of nitrate removal is directly proportional to the fractional coverage of the
dierent sites by the reactants, as predicted by the Langmuir adsorption
model. Infrared spectroscopic studies conrm that adsorbed forms of
nitrate, nitrite and NO can be detected and that all are reactive in the
presence of hydrogen.46
A stepwise reaction mechanism which involves adsorption of nitrate at a
bimetallic site, reduction to nitrite, desorption in to the aqueous phase and
re-adsorption at a monometallic (e.g. Pd) site has been proposed47 and is
supported by theoretical prediction.29 A reaction scheme based on the use of
a bimetallic catalyst is illustrated in Fig. 2.
The nal reduction to N2, N2O or NH4+ takes place at a Pd mono-
metallic site where NO may be the key intermediate. A recent spectroscopic

Catalysis, 2009, 21, 5181 | 57


This journal is 
c The Royal Society of Chemistry 2009
Fig. 2 Potential reaction scheme for the catalytic hydrogenation of nitrates over a supported
bimetallic catalyst showing catalysed steps and the formation of products and intermediates.

study48 suggests that NO on Pd is hydrogenated directly to produce


nitrogen whereas adsorbed NH2 is the precursor of the ammonium anion.
On the other hand, it has been suggested49 that the overall selectivity of the
nitrate reduction reaction is governed by the nitrite hydrogenation step.
Consequently, catalyst formulation must be made to ensure that the rate of
intermediate reduction is much greater than nitrate reduction. Nitrous
oxide was only detected for Pd/TiO2 when high levels of intermediate nitrite
ion were introduced.50 Similarly, N2O formation is much more prevalent
over PdCu than PdSn or PdIn where the former reduces the intermediates at
a slower rate during nitrate hydrogenation.25 Theoretical studies29 suggest
that atomic oxygen is produced at each step in the process, which is strongly
adsorbed and blocks active sites on the catalyst and reduces the activity.
High oxygen coverage also hinders the recombination of nitrogen atoms to
form N2. The model ascribed the removal of the hydroxyl groups from the
copper surface as the rate-limiting step in nitrate reduction over PdCu.29
Continuous removal of oxygen and hydroxyl groups from the surface is
achievable by increasing the hydrogen partial pressure. In respect to
ammonia formation, the rate-determining step is the association of atomic
nitrogen with atomic hydrogen to form the NH bond.29
While an increase of hydrogen availability may have benets for activity if
removal of hydroxyl groups and strongly adsorbed oxygen atoms are rate
limiting,29 the probability of reaction between nitrogen and hydrogen atoms
rather than reaction between two N atoms increases,25,46 thus enhancing
undesired ammonium formation. Reducing the hydrogen ow rate leads to
a reduction in ammonium formation24 albeit with a loss in activity. It has
also been proposed that specic atoms within the Pd surface, may be those
responsible for ammonium formation. Yoshinaga et al.49 describe low
coordinated Pd atoms, such as those atoms in high index planes, and
isolated Pd atoms, formed due to dilution by a second metal, as being
responsible for the formation of ammonium ion. However attempts to
block such sites by selective deposition of Bi failed to signicantly impact on
ammonium formation and it was concluded46 that hydrogen availability in
the immediate environment of the active Pd site, was probably more
important that the coordination number on the metal atom.
An alternative means of limiting hydrogen availability, and thus control-
ling undesired ammonium formation, is through the use of formic acid as a
reductant which can be added in stoichiometric amounts to avoid over-
hydrogenation.25,51 However, this serves an additional purpose since the

58 | Catalysis, 2009, 21, 5181


This journal is 
c The Royal Society of Chemistry 2009
decomposition of formic acid over a metal containing catalyst yields one
mole of CO2 to accompany each mole of liberated hydrogen (eqn (1))
HCO2H(aq) = H2(g) + CO2(g) (1)
 
2NO3 (aq) + 5H2(g) = N2(g) + 2OH (aq) + 4H2O(l) (2)

Given that the reduction of nitrate by hydrogen to yield nitrogen must also
generate hydroxide ion (eqn (2)), this leads to loss of activity as negatively
charged hydroxide ions compete for adsorption sites with nitrate37 (and
nitrite anions), but also leads to a loss in selectivity as the pH is driven up.
The selectivity to ammonium shows a strong correlation with the solution
pH23,25,52 and the addition of a source of CO2 is often added25,42,46 to
maintain the solution at slightly acidic levels (eqn (3)).
CO2(g) + 2H2O(l) = H3O+ + HCO3(aq) (3)

The use of formic acid (or acetate)43 can be seen as an in situ buer25 where
release of CO2 takes place at the right time and place at the point where it is
required to neutralise OH ions.43 In terms of dening the right place, it
has been shown by the use of supports of a range of pore sizes, that diusion
of hydroxide ions from pores may be inuential in determining selectivity as
high localised concentrations of OH may build up due to mass transport
limitations.42,43
In addition to the use of buers such as CO2, either added directly to the
solution or through in situ generation via decomposition of an appropriate
organic reagent,25,42,43 high localised OH concentrations, and conse-
quently high levels of ammonium formation, can be avoided though
appropriate selection of support. This consists of the use of materials with
low surface areas and/or large pore sizes. This has been exploited in
particular in studies using SnO2 as support43,46 while in the case of layered
hydrotalcite materials the low ammonia formation was attributed to a
decrease of the catalyst diusion limitations where positive or neutral are
forced out of the layer spacings by ionic forces.40

2.1.4 Practical considerations and implementation. Most investigations


involve the use of distilled/deionised water with KNO3 as the nitrate ion
source thereby avoiding any potential impact of water hardness and dissolved
salts on the catalytic removal of nitrates. It has been pointed out53 that in the
presence of anions such as SO42 and bicarbonates, which may be present in
tap-water at concentrations of above 90 ppm, reduced nitrate reduction rates
are to be expected as a result of competitive anion adsorption. Pintar and
co-workers54,55 have indicated that nitrate removal rates are reduced when
using drinking water as opposed to distilled water. Chloride ion is known to
reduce the rate of nitrate removal56 while the choice of cation as counter ion
inuences the rate in the order, K+ o Na+ o Ca2+ o Mg2+ o Al3+.54
This may be of relevance if a possible implementation mode involves
integrated ion-exchange/catalytic denitrication process where NaCl is used
as background electrolyte.57 High chloride levels may also lead to catalysts
poisoning, possibly due to metal corrosion.25 Although sulde is also

Catalysis, 2009, 21, 5181 | 59


This journal is 
c The Royal Society of Chemistry 2009
expected to act as a poison, it is not produced by sulfate reduction under
conditions of nitrate reduction. Lecloux28 tested the feasibility of the catalytic
reduction processes in the treatment of drinking water in a pilot plant reactor
and concluded that catalytic nitrate reduction could be scaled up if issues such
as the presence of micro-organisms could be dealt with. These interfere with
the catalytic reduction causing modications to the hydrodynamic behaviour
of the reactors, which may lead to uncontrolled biomass production in
the reactors.
Corma et al.58 have indicated that the use of a xed-bed reactor with 3 m3
of catalyst operating at an LHSV of 350 h1 would be adequate to treat the
water requirements of a town of 100 000 inhabitants if a catalyst of
suciently high selectivity were available, although this level of optimism
has been received with some scepticism.53 Pilot plants based on nitrate
hydrogenation procedures have been tested,59 although inadequate selec-
tivity appeared to be the main diculty.25 As nitrate reduction takes place
through a series of consecutive reactions, systems based on a two stage
reactor set-up have been envisaged,25 supported by theoretical models29
which predict improved overall performance when a system based on Pd/Cu
to convert nitrate to nitrite and then a monometallic Pd catalyst to reduce
nitrite to nitrogen is employed. The use of an integrated ion exchange/
catalytic process based on a two stage reactor design whereby the rst step is
conducted at pH 4 11 over PdCu based catalyst and then a second reactor
operating at ca pH 4.5 over monometallic Pd catalyst was found to give
considerably improved selectivity over a single reactor based system.60

3. Photocatalytic removal of inorganics

3.1 Anion removal


Anionic inorganic groups correspond to a large group of pollutants which,
in numerous cases, strongly interact with titanium oxide. In fact, blocking
of titania surface sites is a signicantly widespread phenomenon as it is
observed, for example, in the cases of Cl, ClO4, CO32, and HCO3.
Phosphates and sulfates correspond to chemical groups which are dicult
to photodegrade due to the fact that they adsorb strongly on semiconductor
surfaces, particularly titania, even at concentrations as low as 1 mM. In the
case of phosphates, the binding with the oxide is so strong that removal with
water is inecient and alkali washing is necessary.61,62 Phosphates and
sulfates may, however, display singular features as they form reactive
species under UV illumination in accordance with eqns (4) and (5):

h+ + SO4 - SO4d (4)

h+ + H2PO42 - H2PO4d (5)

These radicals may initiate oxidation reactions with organic species. The
reaction of S-containing radicals with organic moieties leading to CO2
formation is faster than the corresponding reaction with P-containing
radicals.61 However, as previously indicated, these anionic species may be

60 | Catalysis, 2009, 21, 5181


This journal is 
c The Royal Society of Chemistry 2009
very strongly adsorbed at the surface of the oxide, typically leading to a
decrease in activity of the system in the elimination of organic pollutants.6163
Other anions display unique features with respect to the chemical
behaviour previously mentioned. Among those, two in particular, merit
discussion. Halogens and nitrates dier from the species describe above in
that they may play active roles in the degradation reaction of organic
pollutants and can be simultaneously eliminated from the aqueous phase.
Halogens, including the Cl-containing species, are also held strongly by the
surface of oxide semiconductors, including titania and typically have a
variable impact on the elimination of organic matter. This behaviour has
been rationalized using thermodynamic calculations of the corresponding
negatively changed radicals formed upon photo excitation.64 Chloride
radicals are predicted to oxidize, for example, linear hydrocarbons but
not aromatics. Fluorine radicals are capable of reacting with almost all
organic matter but their formation is not eective under near-UV excita-
tion. Bromide and iodide radicals are both essentially ineective when at the
surface of the oxide which may result in reaction inhibition. Essentially, all
these species are consumed and/or eliminated from the surface in the
presence of organic matter.65
Nitrates correspond to the second group of inorganic materials with can
be photodegraded as a result of their specic physico-chemical interaction
with oxide semiconductors. Very early reports involving photoreduction of
nitrates from aqueous solution in the absence of a catalyst66 found that the
main product was of photoreduction was nitrite, although subsequent
interpretation of data in the literature is not entirely consistent regarding
the product distribution and possible nitrate/nitrite equilibrium.6769 More
recent studies70,71 using Pt/TiO2 as catalyst report that nitrate (and nitrite)
reduction lead to ammonia formation with negligible amounts of dinitrogen
detected. A quantum yield of ca. 2% for NH3 formation at 330 nm in 1 M of
nitric acid was calculated. Similar behaviour to that described for Pt in
respect to ammonia formation was found for other noble metals.7072
In addition to the selection of the metal loaded on the semiconductor oxide,
photocatalytic reduction reactions may be expected2 to be inuenced by a
number of factors, including irradiation time, solution pH and the chemical
nature of the sacricial agent and such sensitivities to these factors for nitrate
and nitrite reduction are reported.7376 As indicated above, nitrate removal
can be combined with the simultaneous photodegradation of organic species
which may be present in aqueous systems which require treatment. One
example is humic acids which are a soluble component formed from plant
residue which contribute to colour and odour of water and may be precursors
to trihalomethanes. These may be mineralised photocatalytically over TiO2
and the process is accelerated when the latter is promoted by silver.77 In
attempting to simultaneously remove nitrates, Bems et al.78 employed con-
centrations of humic acids typically found in lakes and rivers. They found
enhanced nitrate reduction with reduced formation of nitrite and ammonia as
the main products, which were attributed to the numerous reactions that the
humic acids can undergo during reaction. The complexity of the system may
explain why this line of research has not been pursued although given its
relevance to practical application, it would appear to be an avenue worthy of

Catalysis, 2009, 21, 5181 | 61


This journal is 
c The Royal Society of Chemistry 2009
further investigation. In the main, the majority of studies involving nitrate
reduction have involved less complex sacricial hole scavenger molecules
added with a view to enhancing photocatalytic eciency. These molecules
include methanol,7981 ethanol,75,79 formic acid,75,76,82,83 oxalic acid,75,84,85
sodium oxalate,75,86 EDTA,79 and sucrose.87 Although there is some
consensus regarding the selection of formic acid in terms of its eciency as
a hole scavenger,75,76 the relative merits of these reagents is a subject of
debate.76 For example, Jin et al.86 and Gao et al.88 report contrasting ndings
regarding the relative eciencies of sodium oxalate and oxalic acid as hole
scavengers with the latter attributing the better performance of the acid to
the creation of a positively charged titania surface at pH o 6.25 (i.e. below
the PZC) which should facilitate adsorption of the nitrate or nitrite anions.
The literature is equivocal regarding the role of surface charge and evidence
for improved reaction rates for photocatalytic reactions at higher and lower
pH can be found.8 Although these reagents are generally easier to handle than
hydrogen, the photocatalytic performances of the catalysts are still in the
main unsatisfactory due to the high concentration of undesirable nitrites and
ammonia, with the latter generally being the principle end product rather
than dinitrogen.
Despite the apparent potential of the process, reported activities and
selectivities to N2 are still far from optimal. Recent reports75,88 involving
Ag/TiO2, where silver is exploited as an eective electron sink,77 provide
promise in terms of both activity and selectivity. A total nitrate conversion
of 98% could be obtained at a removal rate of 24 mmolNO3g1Ag min1 and
with a selectivity to nitrogen of ca 100%.75 The activity was higher, when
compared in terms of metal used, than most of the PdCu based hydro-
genation catalysts. Similar performances have subsequently been obtained
in our laboratory76 where it was conrmed that optimum performance is
achieved with ca. 1 wt% Ag, (Fig. 3) consistent with ndings for similar
catalysts for the photodegradation of organics.77 When compared on the
basis of mass, the titania with smaller crystallite size (higher surface area)

Fig. 3 Results for the photocatalytic reduction of nitrate in the presence of 0.04 M HCOOH as
hole scavenger over titania and Ag/TiO2 catalysts prepared by impregnation (IMP) or
photodeposition (Photo) and comparing use of Degussa P25 (49 m2 g1) with Hombikat UV
100 (250 m2 g1) as supports.76

62 | Catalysis, 2009, 21, 5181


This journal is 
c The Royal Society of Chemistry 2009
appeared to demonstrate greater performance for nitrate elimination.76
However, as is the case with hydrogenation catalysts, samples showed
poorer conversion eciency in the presence of anions such as SO42,
CO32 and HCO3, which could be attributed to competition for sites
which would otherwise be available for nitrate adsorption on the titania
surface.75 A proposed mechanism involves reduction of nitrate and/or
nitrite by CO2d species which are generated by reaction of the formate
ion with photogenerated holes in accordance with eqn (6).
HCO2 + h+ = CO2d (6)

As these species exhibit a stronger reduction potential (E1CO2/CO2d =


1.8 V)89 when compared with photogenerated electrons (E1 = 0.29 V) or
the formate anion (E1 CO2/HCOO = 0.2 V) they might be expected to
play a key role in nitrate reduction either directly to nitrogen (E1NO3/N2 =
1.25 V) or via formation of nitrite (E1 NO3/NO2 = 0.94 V) and subsequent
reduction to nitrogen (E1NO2/N2 = 1.45 V).

3.2 Cation removal


Metals can play a variety of roles in advanced elimination processes. They
can play active roles in the degradation of organic mater in processes such
as the Fenton or Photo-Fenton processes but in the context of their
elimination, discussion should focus here on the use of chelating agents,
as well as electrochemical and photochemical methods. Metals are generally
non-degradable, have innite lifetime and progressively built up their
concentration in food chains leading to toxic levels. It is well known that
Hg2+, Pd2+, Cd2+, Ag+, Ni2+ and Cr6+ ions are very toxic. Their
maximum concentration in drinking water is regulated in developed coun-
tries and typically, maximum concentrations of 0.0010.0001 ppm are
allowed for the above mentioned cations.90 The European Commission
recommends an upper limit of 50 mg l1 for manganese in drinking water.91
Current manganese removal methods generally required the use of strong
oxidizing agents such as potassium permanganate, chlorine, hypochlorite,
chlorine dioxide or ozone. However attempts to use biological oxidation
methods using trickling lters based on using dierent fractions of silicic
gravel treated with inoculums has been piloted.92
The recovery of metals, as opposed to transformation to a less toxic state
through oxidation, is not only desirable for preventing metal pollution but
as a resource of conservation due to their multiple and demanding industrial
uses. Increasing levels of EDTA in household and industrial waste may lead
to extraction of heavy, toxic metals from mud and sediment and cause
remobilization of the metals in the environment that may have adverse long
term eects. EDDS (Ethylene diaminedisuccinate) is a naturally occurring
chelant and is a structural isomer of EDTA, which, owing to its two chiral
centres, is readily biodegradable and is completely mineralised in the
environment. EDDS therefore cannot remobilize heavy metals as it is
normally biodegraded before discharge into river systems. EDDS is classi-
ed as non-hazardous. Using EDDS may reduce environmental impact of
chelates in formulations while maintaining the chelate performance.

Catalysis, 2009, 21, 5181 | 63


This journal is 
c The Royal Society of Chemistry 2009
Fig. 4 Illustration of valence and conduction band edge positions of anatase at pH 0 and
Mn+/M0 reduction potential energy region for Au, Ag, Hg, Cr, Cu, and Fe. Energy scale takes
vacuum as zero. See text for details.

Recovery of metals by light-driven reduction processes is possible when


the Mn+/M reduction potential is less negative than the bottom of the
conduction band energy level of the semiconductor, as schematically
illustrated in Fig. 4. Once deposited on the surface of the semiconductor,
metals are subsequently extracted from the slurry by mechanical and/or
chemical methods. Metals whose reduction potentials do not permit this
procedure, can be oxidized and deposited over titania as insoluble oxides.6,7
In the absence of organic matter (e.g. sacricial agents and/or additional
pollutants) the conjugated oxidation reaction of metal-ions is the oxidation
of water, a kinetically slow process. Therefore, the reduction of metals is
typically carried out in presence of sacricial hole trapping agents such as
acids and alcohols.93,94 Depending on pH and other experimental condi-
tions (such as metal concentration, dissolved gases etc.), the photogenerated
electrons departing from the oxide surface may reduce protons, water,
dissolved oxygen, etc. Therefore, the photocatalytic reduction of metal ions
is favoured at higher pH and under deoxygenated conditions, and the use of
sacricial agents is frequently required to optimize the reaction rate. Work
under such conditions can decrease metal concentrations down to the
thermodynamic limit, which can be as low as 1012 M. Above this level,
photo-reduction of Au3+, Cr6+, Hg2+, Hg22+, Ag+, Fe3+, Cu+, and
Cu2+ is thermodynamically feasible although Fe3+ and Cr6+ can be only
be reduced to Fe2+ and Cr3+, respectively. Cd2+, Fe2+, and Cr3+ can not
be photocatalytically reduced because their reduction potentials are close or
more negative than that of the photogenerated holes. A possible solution for
the incomplete-reduction and remaining metal cases, is the modication of

64 | Catalysis, 2009, 21, 5181


This journal is 
c The Royal Society of Chemistry 2009
TiO2 with some functional groups such as carboxylic and thiol or amino,
which, through chelation, modify the electrochemical properties of the
metal ions and/or introduce deep trapping sites physically separated from
the oxide and allowing improved electron-hole charge separation.95

4. Catalytic and photocatalytic removal of organometallics


The use of organometalics and their subsequent release into the aqueous
environment represents a particular concern and their removal and applica-
tion of particular removal procedures must be carefully considered to avoid
the formation of by-products and intermediates which are often more toxic
than the starting reagent. For example, the toxicity of organomercury
compounds such as methyl or phenylmercuric salts are greater than
inorganic mercury species. There are a number of studies which have
considered the degradation of organometalics under the inuence of UV
irradiation, however, as noted in the review of semiconductor photo-
catalysis by Mills and Le Hunte,96 several of these studies represent
examples of homogeneous photochemistry rather than photocatalysis and
are often not covered in overviews and reviews on the subject.
Tributyltin (TBT) has been used in antifouling paints, as a fungicide,
bactericide, insecticide and preservative for wood, while dibutyltin (DBT)
and monobutyltin (MBT) are used as stabilizers for PVC and other plastics.
Although inorganic tin compounds are basically harmless, some organotin
compounds are very toxic to both animal and vegetable life. Photolytic
degradation of butyltin compounds is plausible and several studies have
indicated that the degradation process occurs by a stepwise debutylation
mechanism97,98 leading to less toxic compounds. The UV photoassisted
degradation of triphenyltin (TPT) in water also takes place by sequential
dephenylation via diphenyltin (DPT) and monophenyltin (MPT).99 Similar
sequential dephenylation has been noted for diphenyl and monophenyl
mercury compounds which lead to the release of inorganic forms of mercury
along with other organic pollutants.100
However, photolytic degradation is limited as a consequence of the fact
that, for example, butyltin compounds exhibit a maximum absorption
wavelength in the range 190250 nm and therefore these species are degraded
very slowly by natural sunlight (t1/2 4 89 days). Due to this high resistance to
photolytic degradation, the use of TiO2101,102 and Fe(III)98 have been investi-
gated. The former leads to the simultaneous photodeposition of tin onto the
TiO2 surface. Similarly, Hg(0) and Hg2Cl2 are deposited from phenylmercury
salts in aqueous solutions when the acetate or chloride salts respectively, are
exposed to UV in the presence of titania.102 Unlike the photodegradation
process,100 formation of dangerous methyl- or ethylmercury species was not
observed for the heterogeneous photocatalysis process using TiO2 using the
acetate salt. Phenol was detected as a product of the reaction in both cases.

5. Catalytic and photocatalytic removal of organics


An array of physical, chemical and biological methodologies exist which
may deal at various levels of eciency with both naturally occurring organic
matter (NOM) such as humic acids77 and those resulting from

Catalysis, 2009, 21, 5181 | 65


This journal is 
c The Royal Society of Chemistry 2009
anthropogenic activity. Often, combinations of these procedures enhance
operational eciency. For example, membrane ltration is a reliable and
accepted means of treating drinking water as these are eective for the
removal of particles, microorganisms and natural organic matter (NOM).
As NOM has been implicated as a major cause of fouling of membranes,
combining this procedure with the use of a chemical treatment can alleviate
this problem. For example, ozone is a powerful oxidant that preferentially
oxidizes electron-rich moieties containing carboncarbon double bonds and
aromatic alcohols and can lead to reduced membrane fouling.103 In the
following sections we will focus on the use of catalytic and photocatalytic
methodologies which might be employed either in isolation or in combination
with other physical methods of water purication.

5.1 Catalytic removal of organics


Catalytic wet air oxidation (CWAO) is an eective process to treat high
concentrated organic wastewater, which is too concentrated for practical
biological remediation and too dilute for economical incineration and
recovery. The topic of industrial catalytic wet air oxidation processes has
been reviewed by Luck.104 Traditional methods of wet oxidation for the
removal of organic pollutants have involved high pressures (220 MPa) and
temperatures (200320 1C) and there are drivers for replacing these metho-
dologies with improved catalytic wet oxidation processes. Potentially useful
results have been published which involve CWO of dyeing and printing
wastewater.105109 Catalysts based upon the use of copper salts appear to be
the most active homogeneous catalysts, which entails a subsequent separa-
tion step to remove the toxic copper ion from the nal euent.110,111 Clearly
the use of a heterogeneous based catalyst system alleviates this
issue104,105,112 and these have been studied in the removal of aniline,113
phenols,112,114116 acetic acid,112 polyethylene glycol117 N,N- and dimethyl
formamide (DMF)118 amongst others.104 Unfortunately, reported cases of
using heterogeneous catalyst still involve the use of elevated (ca. 200 1C)
temperatures and often prove ineective as a result of deactivation by
sintering, poisoning or fouling.110,119 Additionally, loss of the active
components into the aqueous phase as a contact of exposure to conditions
of elevated temperature and acidic pH is reported.119 There is a bias towards
work on azo based dyes at the expense of CWO studies of anthraquinone
based dyes. Recent studies suggest that the latter can be dealt with
eectively using a catalyst system based on FeCl3/NaNO2 with oxygen
at 0.5 M Pa as oxidant.120
The use of oxygenated additives such as methyl tert-butyl ether (MTBE),
ethyl tert-butyl ether (ETBE) and tert-amylmethyl ether (TAME) with
octane numbers 4 100 improve the quality of gasoline, however, these
compounds are highly soluble in water, and MTBE, ETBE and TAME
have been all been detected in rivers, lakes and groundwater.121 Catalytic
wet oxidation using supported metal catalysts such as Rh/Al2O3 and
Rh/Al2O3CeO2 catalysts has also been reported as a method of removal
of such gasoline oxygenates.122 There is a general consensus that noble
metal catalysts are more eective than metal oxide catalysts.104,112,117 All of

66 | Catalysis, 2009, 21, 5181


This journal is 
c The Royal Society of Chemistry 2009
these catalytic oxidation processes require the use of elevated temperatures,
however there are recent studies in the treatment of petroleum renery
wastewater123 and phenolic pollutants124 which advocate the use of
microwave assisted CWO to reduce energy requirements/reaction times.
One particular subset of organic molecules which require specic
mention, are the halogenated and in particular, chlorinated hydrocarbons.
These may be treated by hydrodechlorination, which, although not leading
to the complete destruction of the pollutants, leads to a convenient
transformation such that the toxicity of the streams can be reduced
dramatically giving rise to more biodegradable euents. The feasibility of
hydrodechlorination for treating groundwater in situ has been demon-
strated by McNab et al.125 A number of contaminants have been shown
to be potentially amenable to catalytic hydrodechlorination, including
chlorinated benzenes and biphenyls, and g-hexachlorocyclohexane
(Lindane),126 even though the reaction rates may be slower than those for
trichloroethylene and perchloroethylene.127 In terms of the active phases,
Pd is widely employed, either alone in a supported form128131 or in the form
of bimetallics132,133 or present in perovskite type structures.131,134 Although
catalysts based on Cu and Ni have also been studied, they are often found to
be less active than Pd-based catalysts and also suer from deactivation due
to leaching of the active phase.128 Pd based catalysts can be promoted by
addition of Au where hydrodechlorination rate of trichloroethane was
maximised at ca. 70% Pd surface coverage. As Pd catalysts are sensitive
for example to poisoning by sulfur species potentially created by hydro-
genation of sultes/sulfates in solution, attempts have been made to
improve the resistance of the catalyst against poisoning by anionic species,
by embedding Pd particles within a hydrophobic zeolite with microporous
structure that excluded ions but was permeable for hydrophobic solutes.130
Although there are examples of successful conversion at room temperature
and atmospheric pressure of hydrogen,130,132 most testing has been
performed at either elevated temperatures134 or both elevated temperature
and pressure128,129,133 with temperatures in the range 308448 K and
pressures in the range 0.2 to 2.8 MPa employed. Clearly any application
should seek to avoid the need to heat huge volumes of wastewater for
treatment purposes and although the Henrys constant for hydrogen in
water (7.8  104 mol L1 atm1 at 298 K)1 permits hydrogenation
reactions to proceed at atmospheric pressure,135 the removal kinetics may
be adversely aected by operation under these conditions as an approxi-
mately linear dependence of initial rates on hydrogen pressure at all
temperatures is generally observed.133

5.2 Photocatalytic removal of organics


Photocatalysis treatment of wastewater has focused attention on attaining
the degradation of reagents which are not readily eradicated by conven-
tional or other advanced degradation methods. Among the two groups of
pollutants, phenol and its derivates and hydrocarbons are important
examples.6,7,10

Catalysis, 2009, 21, 5181 | 67


This journal is 
c The Royal Society of Chemistry 2009
Degradation of phenol and related molecules is important due to the large
number of industrial process where these molecules become chemical waste
products.136,137 Photochemical degradation of phenol with titania has
yielded reasonably good results but the process shows two well-dened
regimes associated with two dierent reaction pathways. At low concentra-
tions (below 103104 ppm) direct OH-radical attack is favoured and the
reaction rate is typically limited by electron-hole recombination.138,139 At
higher concentrations, degradation takes place at the surface of the oxide by
formation of peroxocompounds. These are UV absorbers which limit
OH-radical formation and holes appear to directly attack the adsorbed
intermediates.138 In this latter case, degradation proceeds through
O2d radicals140 and a signicant number of intermediates are formed such
as hydroquinone and maleic acid, and adsorbed formates and carbonates,
which are responsible for deactivation of the catalyst as they act as hole
trapping centres.141,142 Other studies indicate the presence of biphenols and
other water insoluble products could play similar key roles in the deactiva-
tion processes of the titania catalyst.143 Many studies have dealt with
substituted phenols and a summary of the main results should highlight
the following points:
- for phenol derivates containing hydroxyl rather than alkoxyl groups,
degradation seems to proceed through mechanisms favouring ring opening
vs. hydroxylation and dimethoxylation steps.144 Chloro-phenols degradation
proceeds initially though para-hydroxylation of the aromatic ring.145,146 Ring
opening typically occurs with formation of carboxylic acids.
- Some substituent groups such as NO2 or COOH seem to aect photo-
activity in a milder manner.
- Photodegradation of para derivates depends mainly on the eect of the
substituent on the aromatic ring, the reaction rate being increased by electron-
donating groups and decreased by electron-withdrawing groups.147,148 An
exception to this rule is hydroquinone.
- Photodegradation rates of ortho derivates present good correlation
with the thermodynamic stability of sigma-complexes formed between
the aromatic ring and the surface OH-radicals. Rates decrease in the
order OCH3 (guiacol) 4 Cl (2-chlorophenol) E H (phenol) 4 OH
(catechol).149
As previously mentioned, a key point in the optimization of the catalysts
photodegradation of phenol and its derivates, is the minimization of the
electron-hole recombination and the intimate connection of this process
with the anion vacancies present in the size-limited, nanometric oxide
particles. Minimization of the overall amount of oxide defects has a
signicant impact on the reaction rate.139,150 Traditional methods for
improving electron-hole charge separation beyond what can be obtained
with bare titania, involves doping mainly with Fe,151,152 although surface
noble-metal15,153 and oxide154,155 contact have been also proved to be
eective in the enhancement of the reaction rate. In all cases, changes in
the surface properties, particularly surface acidity, may also favour activity.
Additionally, it should be mentioned that Keggin-type compounds such as
Na4W10O32 materials have shown better performances than titania-based
systems, at least for the case of chloro-phenol(s) degradation.156

68 | Catalysis, 2009, 21, 5181


This journal is 
c The Royal Society of Chemistry 2009
The other family of chemicals mentioned in the introduction of this
section corresponds to hydrocarbons. Two main groups of pollutants are
concerned within this family in respect to photocatalytic remediation
treatments; the spillage of oil derivates and industrial solvents. Oil derivates
belong to the most dangerous compounds for the environment, as they
possess large migration abilities both in water and on land. Oil spills on
water can be cleaned up using TiO2-coated glass or ceramic micro-beads
oating at the oil-water interface.157 Also, TiO2-deposited on graphite is
able to pump up heavy oils into the macropores of the carbonaceous
material and subsequently decompose these into the oxide component.158
The photodegradation mineralization can reach levels as high as 90% over
titania-based catalysts and products are typically less harmful than those
produced by oil decomposition by weathering, which produces phenols,
polyphenols and eventually tar.159,160 However, long term use of photo-
catalysts is not free from diculties as polymeric intermediates that would
strongly adsorb on the catalyst surface may eventually lead to severe
deactivation. Amongst the numerous industrial solvents, halo-hydrocarbons
such as trichloroethylene are very important due to the limited applicability
of current clean-up technologies. In addition, such solvents are extremely
carcinogenic, toxic and mobile in the environment. The stability of the CX
(X = Cl, F, Br, I) bond in halo-hydrocarbons is responsible for their
toxicity and persistence in biological environments.161,162 Rates for full
photo-degradation are rather slow and produce very toxic intermediates
such as dichloroacetyl chloride and phosgene (COCl2).163,164 Additionally,
the initial intermediate formed is typically accumulated on the surface,
inhibiting further reaction.165 Other hydrocarbons used as solvents in
industrial chemical processes where photo-degradation has been frequently
employed are benzene and toluene. However, although safe and adequate
disposal procedures for removal of these components are possible using
conventional technologies, the use of visible-light active systems, potentially
allowing solar light as a cost-eective energy source, may provide a lucrative
alternative to current technology.7,17

5.2.1 Application of photocatalysis to specic reagent groups. Pesticides


contain a wide variety of compounds arising from industrial euents, agri-
cultural runos, and chemical spills.166 Such chemicals are toxic, having
potential eects such as carcinogenesis, neurotoxicity, and eects on cell
reproduction and cell development, particularly at the early stage of life. In
the main they are stable to natural decomposition, and persistent in the
environment,167 and consequently correspond to a world-wide problem.
Maximum permitted concentrations and their degradation products are
established by regulations.91,168 Pesticides are customary classied either by
their chemical nature, such as organo-phosphates, and organo-chlorinated
compounds, or by their mode of action, such as insecticides, herbicides, and
fungicides. As this is a large family of chemicals, a wide grouping of dierent
technologies are current used and investigated in order to degrade these
compounds. The main degradation processes and technologies are detailed
below.

Catalysis, 2009, 21, 5181 | 69


This journal is 
c The Royal Society of Chemistry 2009
Several chemical pollutants which are foreign to the body of living
organisms and resistant to environment degradation are generating a major
concern as they may mimic hormones and other endocrine chemical
molecules with impact on biological life, and in particular, on humans.
The main type of pollutants playing a role in this context are polycyclic
aromatic hydrocarbons (PAHs), alkylphenols (APs), and many others such
as alkylbenzene sulfonates which are used as synthetic detergents and which
may break down to endocrine disrupting chemicals (EDCs).169 Current
technologies to treat these are based upon; (i) adsorption onto suspended
solids or association with fats and acids; (ii) aerobic and anaerobic
degradation; (iii) chemical (abiotic) degradation by processes such as
hydrolysis; (iv) volatilization. However, a signicant number of EDCs
are not eciently removed by these treatments and novel technologies
are sought.
Photocatalysis oers interesting areas of research and application in the
eld of pesticides and endocrine disruptor degradation due to the limited
applicability of conventional technologies. Also, some other areas of
photocatalytic research may consider commercial, particularly textile, dye
photobleaching but, unfortunately, complete degradation or mineralization
is far from being achieved with these organic chemicals.7 For this reason,
here we will only detail the cases of pesticides, endocrine disruptors (EDs),
and other molecules where photocatalysis may add some interesting input
to the current technologies.

5.2.2 Photocatalytic removal of pesticides. Until now, complete miner-


alization of pesticides is observed, as a general rule, only after extended
irradiation times. An exception to this is the case of s-triazine herbicide
where degradation produces highly stable triazine nuclei, refractory to
photocatalytic elimination. At the end of this process, cyanuric acid is
formed, which is very stable, but fortunately, non-toxic.170 Also, typically, a
large number of compounds are detected during degradation due to the
complexity of degradation routes which involve multi-step and inter-
connected pathways. Some of the longer-lived intermediates which are
detected can be classied into ve groups: (i) hydroxylated products and
derivates occurring after dehalogenation of corresponding pesticides;
(ii) oxidation products of side chains; (iii) ring-opening products in the case
of aromatic pesticides; (iv) decarboxylation products; (v) isomerisation and
cyclisation products. Advantages of the photocatalytic removal with respect
to traditional treatments includes: (i) complete oxidation, even in the ppb
range; (ii) minimization of polycyclised products; (iii) and availability of
cheap catalysts for specic reactor congurations adequate for specic
pollutants. Additional general points of interest are the fact that pesticides
are used in formulations which include non-ionic and anionic surfactants
which interfere in the degradation processes (possibly by competitive
adsorption) and that solar light degradation and mineralization of some
pesticides is faster than under UV light (possibly due to a contribution from
direct or homogeneous photolysis).7,171,172

70 | Catalysis, 2009, 21, 5181


This journal is 
c The Royal Society of Chemistry 2009
In the case of degradation of insecticides, by considering their chemical
nature it is possible to distinguish between organo-phosphates (OPs), organo-
chlorine (OCs) and carbamate compounds. Photocatalytic degradation of
DDVP (dimethyl-2,2-dichlorovynil phosphate), DEP (dimethyl-2,2,2-
trichloro-1-hydroxyethil phosphonate), methamidophos, phorate, malathion,
diazion, primiphos-methyl, fenamiphos, chlorfenvinphos, and others was
observed until quasi-complete mineralization but with variable reaction rates
and eciencies. Mechanistic analysis of the degradation processes showed a
general applicability of Langmuir-Hinshelwood kinetics. The presence of
anions such as carbonates, phosphates, sulfates, nitrates and cations such as
Ca, Mg, or Fe was found to inuence the reaction rate as were a number of
other parameters, including the nature of the pollutant, pH and others. Some
pollutants such as methamidophos and phorate suer direct photolysis,
possibly due to the low bond energy of PS and CS bonds.173177 The
synergistic combination of H2O2 or photo-Fenton and TiO2, the competition
between such technologies, as well as the use of co-catalysis such as activated
carbon to enhance adsorption steps, have received attention in recent
literature.175177 In the case of OCs, the main focus of research considers
the degradation of lindane (gamma-isomer 1,2,3,4,5,6-hexacholorocyclohexane)
and its derivates, DDT (1,1,1-trichloro-2,2-bis(4-chlorophenyl)ethane),
methoxychlor (and 1,1,1-trichloro-2,2-bis(4-methoxyphenyl)ethane), and
chloro-phenols. The latter has been previously discussed in the section
dealing with phenols. Essentially, the mechanism occurs with complete
mineralization through OH or hole-radical attack to form hydroxylated
products followed by a slow ring opening and fragmentation. Lindane and
DDT appear to be the most problematic pollutants due to the limited
reaction rates obtained.2,146,178,179 Carbamates constitute the third impor-
tant class of insecticides, which are widely used against pests in vast forest
areas because of their rapid action and relatively low persistence in
the environment. The most important member of the family is carbaryl
(1-naphtyl N-methylcarbamate). Although it has a relatively low lifetime of
weeks while in soil, its biological half-life is much longer, e.g. 56 months in
sh. Studies of its photodegradation show complete mineralization through
a complex, multi-step mechanism.180 Other carbamates whose degradation
is eciently mineralized using titania-based photocatalysts are; XMC
(3,5-xyxyl methylcarbamate), MPMC (3,4-dimethylphenyl methyl carbamate),
oxamyl (N,N-dimethyl-2-methylcarbamoyloxyimino-2-(metylthio) acetamide),
EPTC (S-ethyl-N,N-dipropyl thiocarbamate), and MCP (4-chloro-2-
methylphenoxyacetic acid). The reaction rate is, in all cases, related to the
Hammett constant and adsorbability (measured by the partition coecient
between octanol and water).172,181,182 Only in the case of vapam (monoalkyl
dithiocarbamate) is the possible deactivation of titania by S-containing
species formed during degradation, thought to prevent complete
mineralization of the substance.183
The photocatalytic degradation of widely used herbicides such as 2,4-D
(2,4-dichlorophenoxyecetic acid), 2,4,5-T (2,4,5-trichlorophenoxyacetic acid),
bentazon (3-isopropyl-1 H-2,1,3-benzothiadiazin-4(3H)-one 2,2-dioxide),
S-tiazines, carbetamide ((R)-1-(ethylcarbamoyl)ethyl carbanilate), and
monouron (3-(4-chlorophenyl)-1-methoxy-l-methylurea) in water have been

Catalysis, 2009, 21, 5181 | 71


This journal is 
c The Royal Society of Chemistry 2009
extensively studied. With the exception of the previously mentioned case of
S-tiazines, complete mineralization is observed even though the reaction rate
depends on the chemistry of the compound. Reaction rates appear in the
order 2,4,5-T 4 2,4-D and propazine 4 atrazine 4 cyanacine. In the case of
aromatic-based herbicides, the aliphatic part of the molecule (when present) is
readily split from the aromatic ring and subsequent polyhydroxylation of the
aromatic ring results in ring opening and formation of aliphatic acids.
S-tiazine decomposition may present contributions from heterogeneous and
UV-light homogeneous photolysis, leading to formation of the photo-stable
cyanuric acid. It appears that the degradation rate of these pollutants may be
signicantly enhanced in the case of lanthanide-doped titania.170,182,184186
Finally, some fungicides such as metaxyl (N-(2,6-dimethylphrnyl)-N-
(methoxyl-acetyl)-D,L-alanine methyl ester), pyrimethanil (N-(4,6-dimethyl-
pirimidin-2-yl)aniline), chlorothalonil (tetrachloroisophtalonitrile), and
dichlofuanid (N-dichlorouoromethylthio-N0 ,N0 -dimethyl-N-phenylsulfamide)
were subjected to photocatalytic degradation, leading to complete
mineralization following Langmuir-Hinshelwood kinetic models.187189

5.2.3 Photocatalytic removal of endocrine disruptors (EDs). The family


of biphenol xenoestrogens has been subjected to several photocatalytic
essays of degradation. Bisphenol A degradation using titania was found
to be strongly pH dependent, reaching maximum values for low, acidic pH
but forming less toxic intermediates at pH 10. Substitution of the two
methyl groups (binding the C atom bridging the two phenyl groups in
Bisphenol A) of the molecule inuences the degradation rate by aecting the
electric charge suered by the central C atom bridging phenyl groups as well
as the C-ring atoms closer to the bridge position. Depending of the
substituents, OH-radical attack occurs at the phenyl ring but in the case
of the absence of methyl groups (4,4 0 -methylenebisphenol), the OH-adduct
intermediates are only formed after ring cleavage.190,191 Other works
investigated the photocatalytic degradation of natural estrogens such as
17b-estradiol, estrone and estriol, synthetic estrongens as 17a-ethynyloestradiol,
and xenoestrogens as resorcinol and 2,4-dichlorophenol. In all cases, close
to complete degradation (mineralization was not generally analyzed) was
observed at near neutral pH but rates are favoured in
alkali media due to the enhancement of both adsorption and oxidation
steps.192,193 Other potential EDs by themselves or by some of their natural
or articial degradation products correspond to several pesticides already
discussed in the previous section such as cyclohexane-, phenol-, uracil-,
imidazole- and phthalate-based compounds among which lindane, atrazine,
malthion, and chlorophenols appear as rather important cases. As is the
case with a signicant number of pollutant families previously described,
the combined use of ozonization, sonolysis and other methods together with
photocatalysis typically leads to synergistic eects.193

5.2.4 Photocatalytic removal of other molecules. While there an enor-


mous number of chemical targets where degradation has been pursued
using photocatalysis, we mention only a few here and in particular, have

72 | Catalysis, 2009, 21, 5181


This journal is 
c The Royal Society of Chemistry 2009
highlighted those which may not be readily degradated using other
conventional methods. Without trying to provide an extensive list, we
mention methyl tert-butyl ether as it is resistant to microbiological
decomposition, air stripping or selective adsorption technologies. The
combination of photocatalysis with sonolysis appears as a rather compe-
titive alternative in terms of time and economical requirements.194 Also, it
is interesting to mention the elimination of specic resistant contaminants
such as ethidium bromide (3,8-diamino-6-phenyl-5-ethylphenanthidium
bromide) which is used as a DNA intercalating agent in biotechnological
processes. Photocatalysis has been shown to be a rather ecient pathway
for its elimination and it seems that the acidbase properties of titania play
a signicant role in the process due to the strong adsorption of the
pollutant.195,196

6. Removal of microorganisms
The current disinfection technologies of water supplies apply either chemi-
cal or photochemical damage or physical removal of microorganisms by
ltration. Chlorination is a universally practiced water-disinfection process,
which can prevent waterborne infectious diseases, but it is not ecient in the
inactivation of spores, cysts and some viruses.197 Chlorination, combined
with ozonization, signicantly improves the disinfection eciency for all
pathogens such as bacteria, viruses, as well as cystforming protozoan
parasites.198 A key problem for these technologies (and possibly all
technologies) is related to regrowth of microorganisms.

6.1 Photocatalytic removal of microorganisms


Photocatalytic disinfection has experienced a boost in the last decades and
promising results have been obtained with a signicant number of micro-
organisms including Gram-negative bacteria (Escherichia coli, Enterobacter
cloacae, Erwinia Caratovira, Salmonella typhimurimm, enterica, and faecalis
Pseudonomas earuginosa, and uorescens, Listeria monocytogenes, Klebsiella
pneumonae, and Microbacterium sp.), Gram-positive bacteria (Staphylococuss
aureus, Streptococuss sobrinus AHT, Bacillus stearopthermophilus,
pumilus, subtilis and sp., Lactobacillus helveticus, and plantarum), yeasts
(Zygosaccharomices rouxii, and Pichia jadini), fungus (Aspergillus niger,
Candida albicans, Fusarium solani), protozoa (Giardia lambia, Acanthamoeba
polyphaga), and viruses (Lactobacillus case phage PL-1, Bacteriophage
MS2, polvirus 1, Avian A/H5N2).6,7,199206 Titania is used as a
powder6,7,199,200,202,203,205 but also supported6,7,201,204,206 on plastics, metals,
ceramics and other materials which facilitate its use and recovery. These
studies showed that photokilling performance is sensitive to several
experimental factors, among which those of importance are:
- excitation energies, uences (power  time) and irradiation conditions
(pulsed, continuous) used in experiments with titania as they all signicantly
inuence the photokilling performance, possibly in a more pronounced way
than in the photodegradation of organic waste. For example, while E. coli
and B. fragilis require prolonged illumination for eective killing, B. pumilis
inactivation is better obtained by using intermittent illumination.207209

Catalysis, 2009, 21, 5181 | 73


This journal is 
c The Royal Society of Chemistry 2009
- The second factor concerns the state and nature of the biocidal oxide;
TiO2 works by surface/near-surface contact and display signicant varia-
bility in eciency while used as powders or are immobilized on a support.
As is well known, as a powder it may present typically a 1-fold order of
magnitude greater activity (for example, in the case of E. coli) due to the fact
that nanoparticles in suspension/powders can be ingested by micro-
organisms by phagocytosis, causing rapid cellular damage in addition to
that caused by photo-activity.209
- The third is the eect of temperature as typically the inactivation rate
increases/decreases with temperature for gram-positive/gram-negative bac-
teria. The exception to this rule are coliforms. All microorganisms display,
in any case, very narrow temperature ranges where photocatalytic
disinfection activity reaches maximum values.6,7,209
- Titania photokilling performance seems to decrease in the order virus 4
gram-negative 4 gram-positive 4 bacterial spores E yeasts 4 fungus.6,7
This appears to be connected with the increasing cell wall complexity which
goes from the thin peptidoglycan layer of gram-negative bacteria, to the
thicker and more compact walls of gram-positive bacteria and cocci, and
ending with the thick eukaryotic cell membrane containing sugar polymers
for yeasts and/or complex peptidoglycan chains for fungus.
- The last, minor point is specic to the measurements of cell inactivation
reaction rates as they show a linear relationship with the initial bacteria
concentration, a fact that should be considered when comparing results.
In spite of these problems, good eciencies, close to those considered
useful for bacteria disinfection (45 log reduction in the temporal range of
minutes) are customarily reported in studies reviewed here. It must be
noted, however, that wastewater with signicant amounts of solids in
suspension and dissolved organic matter may complicate the functionality
of titania photocatalysts. Tests of titania performance with real municipal
wastewater has been proven nevertheless successful.7 Also, the presence of
microorganism aggregation forming biolms is another point of relevance
due to the limited use of current technologies. In this aspect, only a few
titania-based photocatalysts show adequate performance for reducing
microorganism population eectively.204
The mechanism leading to cell death appears as key information in order
to optimize the photokilling process but is not yet fully understood. Early

Fig. 5 SEM images of P. aeruginosa cells sited in TiO2-coated polymer lms: (A) absence of
light; (B) after 30 min (1 kJ m2) UV-treatment. Lysed cells are in a box while cell debris are
marked with arrows.

74 | Catalysis, 2009, 21, 5181


This journal is 
c The Royal Society of Chemistry 2009
proposals suggest that photokilling mechanisms imply an oxidation of the
intracellular coenzyme A (CoA), inhibiting cell respiration. Subsequent
direct TiO2microorganism contact would result in cell death.207 However,
the photokilling sensitivity to the microorganism structural surface proper-
ties, particularly to the chemical complexity and thickness of the cell wall,
suggests that it is initiated by a cell wall and/or cytoplasmatic membrane
attack. Kinetic studies with E. coli spheroplats, which do not have cell walls,
strongly support such a mechanism.210 Recent microscopy studies (Fig. 5),
indicate that cell walls suer a continuous collapse by which the photo-
process leads initially to lysed, round-shaped cells with a restricted number
of breaks in their walls. This decreases the cell volume by a factor of 2/3 but
the microorganism may subsequently re-grow while under dark conditions.
In a subsequent step, viability of cells is however fully lost by further attack
by TiO2-derived radicals.204 The viability loss would be a function of the cell
wall chemical complexity and thickness and of the eciency of the micro-
organism repair/protection mechanisms using the superoxide dismutase (an
enzyme that dismutates superoxide radicals to H2O2 and O2) and catalase
(an enzyme that reduces intracellular concentration of H2O2 and converts it
to H2O and O2) enzymes.

Acknowledgements
We would like to thank researchers and colleagues who have worked in our
laboratories in the area of water treatments, including R. P. K. Wells, J. Sa,
C. Alcaraz Aguera, A. Sutherland, P. Sampedro-Tejedor, A. Kubacka, and
C. Colon and to acknowledge nancial support under project CICYT
CTQ2007-60480/BQU (Spain) and the Royal Society (London)-CSIC
(Joint international project grant). We are also grateful to the reviewers
of this article for their suggestions of additional sections and references.

References
1 http://www.mpch-mainz.mpg.de/Bsander/res/henry.html.
2 M. R. Homann, B. T. Martin, W. Choi and D. W. Bahnemann, Chem. Rev.,
1995, 95, 69.
3 A. L. Linsebigler, G. Lu and J. T. Yates, Chem. Rev., 1995, 95, 735.
4 D. W. Bahnemann, Solar Energy, 2004, 77, 445.
5 N. Serpone, J. Phys. Chem. B, 2006, 110, 24287.
6 G. Colon, C. Belver and M. Fernandez-Garc a, Nanostructured oxides in
Photocatalysis, in Synthesis Properties and Applications of Solid Oxides,
eds. J. A. Rodr guez and M. Fernandez-Garc a, Whiley, NY, 2007.
7 O. Carp, C. L. Huisman and A. Rellr, Prog. Solid State Chem., 2004, 32, 33.
8 M. A. Fox and M. T. Dulay, Chem. Rev., 1993, 93, 341.
9 J. Blanco and S. Malato, Tecnologa de Fotocatalisis Solar, Cuadernos
Monogracos, CIEMAT, 1996.
10 A. Fujishima, T. N. Rao and D. A. Tryk, J. Photochem. Photobiol. C, 2000, 1, 1.
11 U. Dietbold, N. Ruzycki, G. S. Herman and A. Selloni, Catal. Today, 2003,
85, 93.
12 J. K. Burdett, T. Hughbands, J. M. Gordon, J. W. Richarson and J. Smith,
J. Am. Chem. Soc., 1987, 109, 3639.

Catalysis, 2009, 21, 5181 | 75


This journal is 
c The Royal Society of Chemistry 2009
13 R. Asahi, Y. Taga, W. Mannstadt and A. J. Freem, Phys. Rev. B, 2000, 61,
7459.
14 B. Wiswanathan, Bull. Catal. Soc. India, 2003, 2, 71.
15 S. K. Lee and A. Mills, Plat. Metals Rev., 2003, 47, 61.
16 M. Mattesini, J. S. Almeida, L. Dubrovinsky, N. Dubrovinskaia, B. Johansson
and R. Ahuja, Phys. Rev. B, 2004, 70, 115101.
17 A. Kubacka, M. Fernandez-Garc a and G. Colon, J. Catal., 2008, 254, 272.
18 K. Takeda and K. Fujiwara, Water Res., 1996, 30, 323.
19 V. Mateju, S. Cizinska, J. Krejci and J. Thomas, Enzyme Microb. Technol.,
1992, 14, 170.
20 S. Aslan and A. Turkman, Water Sci. Technol., 2003, 48, 489.
21 B. U. Bae, C. H. Kim and Y. I. Kim, Water Sci. Technol., 2004, 49, 413.
22 K.-D. Vorlop and T. Tacke, Chem. Ing. Tech., 1989, 61, 836.
23 S. Horold, K.-D. Vorlop, T. Tacke and M. Sell, Catal. Today, 1993, 17, 21.
24 S. Horold, T. Tacke and K.-D. Vorlop, Environmental Technol., 1993, 14, 931.
25 K.-D. Vorlop and U. Prusse, in Environmental Catalysis, eds. F. J. J. G.
Janssen and R. A. van Santen, ICP, London, 1999, vol. 1, p. 195.
26 U. Prube, S. Horold and K.-D. Vorlop, Chem., Ing. Tech., 1997, 69, 93.
27 J. Sa, D. Gasparovicova, K. Hayek, E. Halwax, J. A. Anderson and H. Vinek,
Catal. Letts., 2005, 105, 209.
28 A. J. Lecloux, Catal. Today, 1999, 53, 23.
29 A. Klesing, D. Labrenz and R. A. Van Santen, J.C.S. Faraday Trans., 1998,
94, 3229.
30 F. Epron, F. Gauthard, C. Pineda and J. Barbier, J. Catal., 2001, 198, 309.
31 J. Sa and H. Vinek, Appl. Catal. B, 2004, 57, 247.
32 F. Epron, F. Gauthard and J. Barbier, Appl. Catal., 2002, 237, 253.
33 A. Eldemann, W. Schieber, H. Vinek and A. Jentys, Catal. Letts., 2000, 69, 11.
34 W. Gao, N. Guan, J. Chen, X. Guan, R. Jin, H. Zeng, Z. Liu and F. Zhang,
Appl. Catal. B, 2003, 46, 341.
35 W. Gao, J. Chen, X. Guan, R. Jin, F. Zhang and N. Guan, Catal. Today, 2004,
9395, 333.
36 J. Sa, T. Berger, K. Fottinger, A. Riss, J. A. Anderson and H. Vinek, J. Catal.,
2005, 234, 282.
37 M. P. Maia, M. A. Rodrigues and F. B. Passos, Catal. Today, 2007, 123, 171.
38 S. Kerkeni, E. LamyPirata and J. Barbier, Catal. Today, 2002, 5, 35.
39 N. Barrabes, J. Just, A. Danov, F. Medina, J. L. G. Fierro, J. E. Sueiras,
P. Salagre and Y. Cesteros, Appl. Catal. B, 2006, 62, 77.
40 A. E. Palomares, J. G. Prato, F. Rey and A. Corma, J. Catal., 2004, 221, 62.
41 F. Epron, F. Gauthard and J. Barbier, J. Catal., 2002, 206, 363.
42 R. Gavagnin, L. Biasetto, F. Pinna and G. Strukul, Appl. Catal. B, 2002, 38,
91.
43 M. DArino, F. Pinna and G. Strukul, Appl. Catal. B., 2004, 53, 161.
44 T. Tacke and K.-D. Vorlop, Chem. Ing. Tech., 1993, 65, 1500.
45 A. Pintar, J. Batista, J. Levec and T. Kajiuchi, Appl. Catal. B, 1996, 11, 81.
46 J. Sa, J. Montero, E. Duncan and J. A. Anderson, Appl. Catal. B, 2007, 73, 98.
47 J. Warna, I. Turunen, T. Salmi and T. Maunula, Chem. Eng. Sci., 1994, 49,
5763.
48 S. D. Ebbesen, B. L. Mojet and L. Leerts, J. Catal., 2008, 256, 15.
49 Y. Yoshinaga, T. Akita, I. Mikami and T. Okuhara, J. Catal., 2002, 207, 37.
50 J. Sa and J. A. Anderson, Appl. Catal. B, 2008, 77, 409.
51 U. Prube, M. Kroger and K.-D. Vorlop, Chem. Ing. Tech., 1997, 69, 87.
52 G. Strukul, F. Pinna, M. Marella, L. Meregalli and M. Tomaselli, Catal.
Today, 1996, 27, 209.

76 | Catalysis, 2009, 21, 5181


This journal is 
c The Royal Society of Chemistry 2009
53 F. Ruiz-Bevia, J. Catal., 2004, 227, 563.
54 A. Pintar, M. Setinc and J. Levec, J. Catal., 1998, 174, 72.
55 A. Pintar and J. Batista, Catal. Today, 1999, 53, 35.
56 A. Pintar, J. Batista and I. Mus evic, Appl. Catal. B, 2004, 52, 49.
57 A. Pintar J. Batista and J. Levec, Chem. Eng. Sci., 2001, 56, 1551.
58 A. Corma, A. E. Palomares, F. Rey and J. G. Prato, J. Catal., 2004, 227, 561.
59 M. Sell, M. Bischo and D. Bonse, Vom Wasser, 1992, 79, 129.
60 A. Pintar and J. Batista, Appl. Catal. B, 2006, 63, 150.
61 M. Abdullah, G. K. C. Low and R. W. Mattheus, J. Phys. Chem., 1990, 94,
6820.
62 O. J. Jung, Korean Chem. Soc., 2001, 22, 1183.
63 G. Colon, M. C. Hidalgo and J. A. Nav o, Appl. Catal. B, 2003, 45, 39.
64 M. M. Lewandowski and D. F. Ollis, J. Catal., 2003, 217, 38.
65 P. B. Amama, K. Itoh and M. Murabayashi, Appl. Catal. B, 2002, 37, 321.
66 H. Thiele, Ber., 1907, 40, 4914.
67 W. T. Anderson, J. Am. Chem. Soc., 1924, 46, 797.
68 D. S. Villars, J. Am. Chem. Soc., 1927, 49, 326.
69 P. Warneck and C. Wurzinger, J. Phys. Chem., 1968, 72, 389.
70 A. Kudo, K. Domen, K.-I. Maruya and T. Onishi, Chem. Lett., 1987, 1019.
71 A. Kudo, K. Domen, K.-I. Maruya and T. Onishi, J. Catal., 1990, 135, 300.
72 K. T. Ranjit, T. K. Varadarajan and B. Viswanathan, J. Photochem. Photobiol.
A, 1995, 89, 67.
73 K. T. Ranjit, R. Krishbamoorthy and B. Viswanathan, J. Photochem. Photobiol.
A, 1994, 81, 55.
74 K. T. Ranjit, R. Krishbamoorthy, T. K. Varadarajan and B. Viswanathan,
J. Photochem. Photobiol. A, 1995, 86, 185.
75 F. Zhang, R. Jin, J. Chen, C. Shao, W. Gao, L. Li and N. Guan, J. Catal.,
2005, 223, 424.
76 J. Sa, C. Alcaraz Aguera, S. Gross and J. A. Anderson, Appl. Catal. B.,
Environmental, DOI: 10.1016/j.apcatb.2008.07.014.
77 A. Bansal, S. Madhavi, T. T. Yang Tan and T. M. Lim, Catal. Today, 2008,
131, 250.
78 B. Bems, F. C. Jentoft and R. Schlogl, Appl. Catal. B, 1999, 20, 155.
79 K. T. Ranjit, T. K. Varadarajan and B. Viswanathan, J. Photochem. Photobiol.
A, 1997, 108, 73.
80 T. Mori, J. Suzuki, K. Fujimoto, M. Watanabe and Y. Hasegawa, Appl. Catal.
B., 1999, 23, 283.
81 S. Tawkaew, S. Yin and T. Sato, Int. J. Inorg. Mater., 2001, 3, 855.
82 Y. Li, Q. Peng and F. Wasgestian, Chin. J. Catal., 1999, 20, 378.
83 L. L. Perissinotti, M. A. Brusa and M. A. Grela, Langmuir, 2001, 17, 8422.
84 H. Kominami, F. Akitoshi, M. Shinya, I. Hiroyuki, K. Yoshiya and B. Ohtani,
Catal. Lett., 2001, 76, 31.
85 Y. Li and F. Wasgestian, J. Photochem. Photobiol. A, 1998, 112, 255.
86 R. Jin, W. Gao, J. Chen, H. Zeng, F. Zhang, Z. Liu and N. Guan,
J. Photochem. Photobiol. A, 2004, 162, 585.
87 M. Penpolcharoen, R. Amal and M. Brungs, J. Nanopart. Res., 2001, 3, 289.
88 W. Gao, R. Jin, J. Chen, H. Zeng, F. Zhang, Z. Liu and N. Guan, Catal.
Today, 2004, 90, 331.
89 D. M. Stanbury, Adv. Inorg. Chem., 1979, 83, 652.
90 W. Salomons, U. Forstner and P. Mader, in Heavy Metals, Problems and
Solutions, Springer, Berlin, 1988.
91 EC Directive Related with Drinking Water Quality Intended for Human
Consumption, Brussels, Belgium, 98/83, 1998.

Catalysis, 2009, 21, 5181 | 77


This journal is 
c The Royal Society of Chemistry 2009
92 A. G. Tekerlekopoulou, I. A. Vasiliadou and D. V. Vayenas, Biochemical
Engineering Journal, 2008, 38, 292.
93 M. R. Prairie, L. R. Evans, B. M. Strange and S. C. L. Mart nez, Environ. Sci.
Technol., 1993, 27, 1776.
94 S. Wang, Z. Wang and Q. Zhung, Appl. Catal. A, 1992, 1, 257.
95 R. S. Kubal, N. K. Meshkov, T. Rajh and M. C. Thurnauer, J. Photochem.
Photobiol. A, 2002, 148, 393.
96 A. Mills and S. Le Hunte, J. Photochem. Photobiol. A. Chem., 1997, 108, 1.
97 J. A. Navio, F. J. Marchena and C. J. Cerrillos, J. Photochem. Photobiol. A.
Chem., 1993, 71, 97.
98 G. Mailhot, M. Astruc and M. Bolte, Appl. Organometal. Chem., 1999, 13, 53.
99 J. A. Navio, C. Cerrillos, M. A. Pradera, E. Morales and J. L. Gomez-Ariza,
J. Photochem. Photobiol. A Chemistry, 1997, 108, 59.
100 C. Cerrillos, M. A. Pradera Adrian and J. A. Navio, J. Photochem. Photobiol.
A Chemistry, 1994, 84, 299.
101 J. A. Navio, C. Cerrillos, F. J. Marchena, F. Pablos and M. A. Pradera,
Langmuir, 1996, 12, 2007.
102 E. M. de la Fournie`re, A. G. Leyva, E. A. Gautier and M. I. Litter, Chemosphere,
2007, 69, 682.
103 J. Kim, S. H. R. Davies, M. J. Baumann, V. V. Tarabara and S. J. Masten,
J. Membrane Sci., 2008, 311, 165.
104 F. Luck, Catal. Today, 1996, 27, 195.
105 Y. Liu and D. Sun, Appl. Catal. B, 2007, 72, 205.
106 J. Donlagic 0 and J. Levec, Environ. Sci. Technol., 1998, 32, 1294.
107 G. Chen, L. Lei and P.-L. Yue, Ind. Eng. Chem. Res., 1999, 38, 1837.
108 L. Lei, Q. Dai, M. Zhou and X. Zhang, Chemosphere, 2007, 68, 1135.
109 J. Garcia, H. T. Gomes, P. Serp, P. Kalck, J. L. Figueiredo and J. L. Faria,
Catal. Today, 2005, 102103, 101.
110 S. T. Kolaczkowski, P. Plucinski, F. J. Beltran, F. J. Rivas and D. B. McLurgh,
Chem. Eng. J., 1999, 73, 143.
111 S. Imamura, Ind. Eng. Chem. Res., 1999, 38, 1743.
112 D. Duprez, F. Delano, J. Barbier, Jr, P. Isnard and G. Blanchard, Catal.
Today, 1996, 29, 317.
113 J. Garcia, H. T. Gomes, Ph. Serp, Ph. Kalck, J. L. Figueiredo and J. L. Faria,
Carbon, 2006, 44, 2384.
114 J. Wang, W. Zhu, S. Yang, W. Wang and Y. Zhou, Appl. Catal. B, 2008, 78,
30.
115 P. Massa, A. Danov, F. Medina Cabello and R. Fenoglio, Catal. Commun.,
2008, 9, 1533.
116 A. Ayude, T. Rodriguez, J. Font, A. Fortuny, C. Bengoa, A. Fabregat and
F. Stuber, Chem. Eng. Sci., 2007, 62, 7351.
117 D. Mantzavinos, R. Hellenbrand, A. G. Livingston and I. S. Metcalfe, Appl.
Catal. B, 1996, 11, 99.
118 G. Sun, A. Xu, Y. He, M. Yang, H. Du and C. Sun, J. Hazard. Mater., 2008,
156, 335.
119 S. K. Bhargava, J. Tardio, J. Prasad, K. Foger, D. B. Akolekar and S. C. Grocott,
Ind. Eng. Chem. Res., 2006, 45, 1221.
120 Y. Peng, D. Fu, R. Liu, F. Zhang, X. Xue, Q. Xu and X. Liang, Appl. Catal. B,
2007, 79, 163.
121 F. Welsch, B. Elwisck, R. A. James, M. C. Marr, C. B. Myers and R. W. Tyt,
J. Appl. Toxicol., 2003, 23, 387.
122 I. Cuauhtemoc, G. Del Angel, G. Torres and V. Bertin, Catal. Today, 2008,
133135, 588.

78 | Catalysis, 2009, 21, 5181


This journal is 
c The Royal Society of Chemistry 2009
123 Y. Sun, Y. Zhang and X. Quan, Separation and Purication Technology, 2008,
62, 565.
124 L. Bo, X. Quan, X. Wang and S. Chen, J. Hazard. Mater., 2008, 157, 179.
125 W. W. McNab, Jr, R. Ruiz and M. Reinhard, Environ. Sci. Technol., 2000, 34,
149.
126 C. Schuth and M. Reinhard, Appl. Catal. B, 1998, 18, 215.
127 G. Lowry and M. Reinhard, Environ. Sci. Technol., 1999, 33, 1905.
128 L. Calvo, M. A. Gilarranz, J. A. Casas, A. F. Mohedano and J. J. Rodr guez,
Appl. Catal. B, 2008, 78, 259.
129 L. Calvo, M. A. Gilarranz, J. A. Casas, A. F. Mohedano and J. J. Rodr guez,
Appl. Catal. B, 2006, 67, 68.
130 C. Schuth, S. Disser, F. Schuth and M. Reinhard, Appl. Catal. B, 2000, 28,
147.
131 A. Morato, C. Alonso, F. Medina, J. L. Garreta, J. E. Sueiras, Y. Cesteros,
P. Salagre, D. Tichit and B. Coq, Catal. Lett., 2001, 77, 141.
132 M. O. Nutt, K. N. Heck, P. Alvarez and M. S. Wong, Appl. Catal. B, 2006, 69,
115.
133 P. D. Vaidya and V. V. Mahajani, Appl. Catal. B, 2004, 51, 21.
134 A. Morato, F. Medina, J. E. Sueiras, Y. Cesteros, P. Salagre, L. C. de Menorval,
D. Tichit and B. Coq, Appl. Catal. B, 2003, 42, 251.
135 J. A. Anderson, F.-M. McKenna, A. Linares-Solano and R. P. K. Wells,
Catal. Letts., 2007, 119, 16.
136 U. Staord, K. A. Gray and P. V. Kamet, J. Catal., 1997, 167, 25.
137 K. Wang, Y. Sich, M. Chou and C. Chang, Appl. Catal. B, 1999, 21, 1.
138 J. Arana, E. Tello-Redon, J. Dona-Rodr guez, J. A. Herrera-Melian and
M. Perez-Pena, Appl. Catal. B, 2001, 30, 1.
139 G. Colon, M. C. Hidalgo, G. Munuera, I. Ferino, M. C. Cutrufello and
J. A. Nav o, Appl. Catal. B, 2006, 63, 45.
140 I. Ilisz, Z. Laslo and A. Doml, Appl. Catal. A, 1999, 180, 35.
141 M. A. Fox and C. C. Chen, Tetrahedron Lett., 1983, 24, 547.
142 M. A. Fox and J. N. Younathan, Tetrahedron, 1986, 42, 6235.
143 A. Subczynski, L. Ducznd and W. Zmudzin, J. Mol. Catal. A, 2004, 213, 225.
144 X. Li, J. W. Cubbage and W. S. Jenks, J. Photochem. Photobiol. A, 2001, 143,
69.
145 R. W. Matheus and D. F. Sangler, J. Phys. Chem., 1967, 71, 4056.
146 M. Pera-Titus, V. Garc a-Molina, M. A. Banos, J. Gimenez and S. Esplugas,
Appl. Catal. B, 2004, 219, 256.
147 L. Almarich, C. Guillard, E. Blanc-Brude and P. Pitchat, Water Res., 1996, 30,
1137.
148 S. Parra, J. Olivero, L. Pacheco and C. Pulgarin, Appl. Catal. B, 2003, 43, 293.
149 A. M. Peiro, J. A. Ayllon, J. Peral and X. Domenech, Appl. Catal. B, 2001, 30,
359.
150 K. Arata, Appl. Catal. A, 1996, 146, 3.
151 B. Pal, T. Hata, K. Goto and G. Nogami, J. Mol. Catal. A, 2001, 169, 147.
152 C. Adan, A. Bahamonde, M. Fernandez-Garc a and A. Mart nez-Arias, Appl.
Catal. B, 2007, 72, 11.
153 A. Mills, R. H. Davies and D. Worsley, Chem. Soc. Rev., 1993, 22, 417.
154 R. A. Doong, C. H. Che, R. A. Marthreepale and S. M. Chan, Water Res.,
2001, 35, 2873.
155 N. Serpone and V. Emelie, Int. J. Photoenergy, 2002, 4, 91.
156 I. Texier, Ch. Giannoti, S. Malato, Ch. Ritcher and J. Delaire, Catal. Today,
1999, 54, 297.
157 A. Heller, Acc. Chem. Res., 1995, 28, 503.

Catalysis, 2009, 21, 5181 | 79


This journal is 
c The Royal Society of Chemistry 2009
158 T. Tsumura, N. Kojitani, I. Izumi, N. Iwashita and M. inagaki, J. Mater.
Chem., 2002, 12, 1391.
159 R. L. Ziolli and W. F. Jardim, J. Photochem. Photobiol. A, 2002, 147, 205.
160 A. Heller, J. Schwitzgebel, M. Pishko and J. G. Ekerdt, in Waste Water
Treatment. Proceedings in Environmental Catalysis, eds. T. L. Rose, O. Murphy,
E. Rudd and B. E. Conway, The American Electrochemical Society, NY,
1994, p. 1.
161 K. R. Krijgsheld and A. van der Gen, Chemosphere, 1986, 15, 825.
162 H. Rogues, Chemical Water Treatment, VCH-Verlag, Weienheim,
Germany, 1996.
163 H. Wang, C. Peticci and O. Raferty, J. Am. Chem. Soc., 1998, 120, 4388.
164 M. D. Dressen, A. L. Goodman, T. M. Miller, G. A. Zacharis and V. H. Grasian,
J. Phys. Chem. B, 1998, 102, 349.
165 S. Yamaki-Nishida, X. Fu, M. A. Anderson and K. Hori, J. Photochem.
Photobiol. A, 1996, 97, 175.
166 L. Muszkat, D. Raucher, M. Magaritz and D. Roenen, in Groundwater
Contamination and Control, ed. U. Zoller, Marcel Dekker, NY, 1994, p. 287.
167 H. D. Burrous, M. L. Candle, J. A. Santaballa and S. Steenken, J. Photochem.
Photobiol. A, 2002, 67, 71.
168 World Health Organization. Guidelines for drinking water quality. Geneva,
WHO, 1993.
169 Directive 2000/60/EC. Ocial Bulletion of the European Community. Brussel,
20/10/2000.
170 C. Minero, V. Maurino and E. Pelizzetti, Res. Chem. Intermed., 1997, 23, 291.
171 I. K. Konstantinou and T. A. Albanis, Appl. Catal. B, 2003, 42, 319.
172 S. Devipriya and S. Yesodharan, Solar Energ. Mater. Solar Cells, 2005, 86,
309.
173 K. Harada, T. Hisanaga and K. Tanaka, Water Res., 1990, 24, 1415.
174 M. C. Lu, G. D. Roan, J. N. Chen and C. P. Chuang, J. Photochem.
Photobiool. A, 1993, 76, 103.
175 R. Doong and W. Chang, J. Photochem. Photobiol. A, 1997, 107, 239.
176 J. M. Hermann and C. Guillard, C.R. Acad. Sci. Paris, Serie II-Chemie, 2000,
31, 417.
177 M. I. Maldonado, D. C. Passarinho, I. Oller, W. Gernjak, P. Fernandez,
J. Banco and S. Malato, J. Photochem. Photobiol. A, 2007, 183, 354.
178 G. Mills and M. R. Homann, Environ. Sci. Technol., 1993, 27, 1681.
179 A. Zaleska, J. Hupka, M. Wiergowski and M. Bizuik, J. Photochem. Photobiol.
A, 2000, 135, 213.
180 C. E. Peris, J. Terol, A. R. Maviri, M. de la Guardia and E. Pramaura,
J. Environm. Sci., 1993, 28, 431.
181 B. Sangehakr, T. Hisinaga and K. Tanaka, J. Photochem. Photobiol. A, 1995,
85, 187.
182 K. Tanaka and K. S. N. Reddy, Appl. Catal. B, 2002, 39, 305.
183 A. Vidal and M. A. M. Luengo, Appl. Catal. B, 2001, 32, 1.
184 M. Rejto, S. Saltzman, A. J. Acher and L. Muskat, J. Agric. Food Chem., 1983,
31, 138.
185 J. C. D 0 Oliveira, C. Guillard, C. Maillard and P. Pitchat, J. Environm. Sci.
Health A, 1993, 28, 941.
186 K. Ranjit, I. Willner, S. H. Bossmann and A. M. Braun, Environ. Sci. Technol.,
2001, 35, 1544.
187 A. Tapalar, D. Molnar-Gabon and J. Csanadi, Water Res., 1999, 33, 1371.
188 A. Aguera, E. Almansa, A. Tejedor, A. R. Fernandez-Alba, S. Malato and
M. I. Maldonado, Environ. Sci. Technol., 2000, 34, 1563.

80 | Catalysis, 2009, 21, 5181


This journal is 
c The Royal Society of Chemistry 2009
189 V. A. Sakkas and T. A. Albanis, Appl. Catal. B, 2003, 46, 175.
190 N. Watanabe, S. Horikoshi, H. Kaeabe, Y. Sugie, J. Zhao and H. Hidaka,
Chemosphere, 2003, 52, 851.
191 K. Chiang, T. M. Lim, L. Tsen and R. Amal, Appl. Catal. A, 2004, 261, 225.
192 H. M. Coleman, K. Chiang and R. Amal, Chem. Eng. J., 2005, 113, 65.
193 V. Belgiorno, L. Rizzo, D. Fatta, C. Della Rocca, G. Lofrano, A. Nikolau,
V. Naddeo and S. Meric, Desalination, 2007, 215, 166.
194 E. Selli, C. L. Bianchi, C. Pirola and M. Bertelli, Ultrasonics Sonochem., 2005,
12, 395.
195 C. Adan, A. Mart nez-Arias, M. Fernandez-Garcia and A. Bahamonde, Appl.
Catal. B, 2007, 77, 395.
196 C. Adan, A. Bahamonde, A. Mart nez-Arias, M. Fernandez-Garcia,
L. A. Perez-Estrada and S. Malato, Catal. Today, 2007, 129, 79.
197 U. Zszewsyk, R. U. Zszewsyk, W. Manz and K. H. Scheifer, Ann. Envirn.
Microbiol., 2000, 54, 81.
198 A. Abarnou and L. Miossec, Sci. Total Environ., 1992, 126, 173.
199 J. M. C. Robertson, P. K. J. Robertson and L. A. Lawton, J. Photochem.
Photobiol. A, 2005, 175, 51.
200 J. Lonnen, S. Kilvington, S. C. Kehoe, F. Al-Touati and K. G. McGuigan,
Water Res., 2005, 39, 877.
201 A. Vohra, D. Y. Goswami, D. A. Desphande and S. S. Block, Appl. Catal. B,
2006, 65, 57.
202 A. Pal, S. O. Pehkonen, L. E. Yu and M. B. Ray, J. Photochem. Photobiol. A,
2007, 186, 335.
203 D. Mitoraj, A. Janczyk, M. Strus, H. Kisch, G. Stochel and P. B. Heeczko,
Photochem. Photobiol. Sci., 2007, 6, 642.
204 A. Kubacka, C. Serrano, M. Ferrer, H. Lundsford, P. Bielecki, M. L. Cerrarda,
M. Fernandez-Garc a and M. Fernandez-Garcia, Nano Lett., 2007, 7, 2529.
205 C. Guillard, T. H. Bui, C. Felix, V. Moules, B. Lina and P. Lejeune, Comptes
Rendus Chim., 2008, 11, 103.
206 M. L. Cerrarda, C. Serrano, M. Fernandez-Garc a, F. Fernandez-Mart n,
R. J. Jimenez-Rioboo, A. de Andres, A. Kubacka, M. Ferrer and
M. Fernandez-Garcia, Adv. Funct. Mater., 2008, 18, 1949.
207 T. Matsunga, R. Tomato and T. Nakajima, FEMS Microbiol. Lett., 1985, 29,
211.
208 M. N. Pan, T. McDowell and E. Wilkins, J. Environ. Sci. Health A, 1995, 30,
627.
209 A. G. Rincon and C. Pulgarin, Appl. Catal. B, 2003, 44, 268.
210 K. Sunada, T. Watanabe and K. Hasimoto, J. Photochem. Photobiol. A, 2003,
156, 227.

Catalysis, 2009, 21, 5181 | 81


This journal is 
c The Royal Society of Chemistry 2009
Nano-architecture and reactivity of titania
catalytic materials. Part 2. Bidimensional
nanostructured films
Gabriele Centi and Siglinda Perathoner
DOI: 10.1039/b712663h

1. Introduction
Metal oxides are an important class of heterogeneous catalysts.13 They nd
direct application in a variety of reactions, from acid-base to redox
reactions, in photocatalytic processes, and as catalysts for environmental
protection. In addition, they are widely used as supports for other active
components (metal particles or other metal oxides), although often they act
not only as a support, but actively participate in the reaction mechanism.45
A key aspect of metal oxides is that they possess multiple functional
properties: acid-base, electron transfer and transport, chemisorption by s
and p-bonding of hydrocarbons, O-insertion and H-abstraction, etc. This
multi-functionality allows them to catalyze complex selective multistep
transformations of hydrocarbons,19 as well as other catalytic reactions
(NOx conversion, for example). The control of the catalyst multi-functionality
requires the ability to control not only the nanostructure, e.g. the nano-scale
environment around the active site,10 but also the nano-architecture, e.g. the
3D spatial organization of nano-entities. The active site is not the only relevant
aspect for catalysis. The local area around the active site orients or assists the
coordination of the reactants, and may induce sterical constrains on the
transition state, and inuences short-range transport (nano-scale level).
Therefore, it plays a critical role in determining the reactivity and selectivity
in multiple pathways of transformation. In addition, there are indications
pointing out that the dynamics of adsorbed species, e.g. their mobility during
the catalytic processes which is also an important factor determining
the catalytic performances in complex surface reaction,3,4 is inuenced by the
nanoarchitecture.
The nano-architecture is thus an important aspect to consider for the
design of novel catalysts and a critical element to consider also in analyzing
how to bridge the gap between model and real catalysts. In fact, in addition
to the issues of pressure and material gap, the complexity gap exists.1113
Goodman14 over ten years ago pointed out that despite the successes in
modelling catalysts with single crystals, there is a clear need to develop
models with higher levels of complexity and which take into account the 3D
nanoarchitecture.
Most of the studies on real nanostructured oxides are based on
materials not having a well-dened 3D structure (both on short and long-
range), being composed of irregularly shaped nano-crystals. These materials
are polycrystalline, and show several nano-interfaces, which stabilize

Dept. of Industrial Chemistry and Engineering of Materials and ELCASS (European


Laboratory for Catalysis and Surface Science), University of Messina, Salita Sperone 31,
98166, Messina, Italy. E-mail: centi@unime.it; Fax: +39-090-391518

82 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
microstrains, oxygen vacancies or metal ions in unusual coordination states.
A 3D environment for adsorption/transformation may signicantly modify
the adsorption of reactants and induce stabilization of transition state
complexes, a well-known concept in enzymes, but typically not considered
for solid catalysts.
Although several studies have been dedicated to analyzing the relation-
ship between nanostructure of supported metal particles and catalytic
behaviour, fewer studies have been dedicated to growing controlled oxide
nanostructures and their relationship to catalytic reactivity.
Henry15 proposed to understand the intrinsic heterogeneity of real
catalyst using supported model catalysts prepared by epitaxial growth on
oxide single crystals. Barteau et al.16 investigated ordered arrays of discrete,
reactive oxide molecules, such as heteropolyanions (polyoxometalates)
deposited to form ordered monolayers that permit site-by-site mapping of
chemical functions on the surface. Datye17 proposed to use oxide particles
of simple geometric shape. Weiss and Schlogl13 suggested the use of
epitaxial lms with dened surface structure. Peden et al.18 used molecular
beam epitaxy (MBE) systems to grow model metal-oxide lms. Rainer
and Goodman19 also stressed the relevance of thin lm oxides for
structure-activity investigations of heterogeneous catalysts.
A general issue in these studies is that the preparation method is quite
dierent from ones used to prepare real catalysts to be tested under practical
conditions. This is an important issue, because there is the need to link the
micro-kinetic and surface mechanism studies to the catalytic behaviour
under real conditions, and to use the knowledge generated by the
fundamental investigations to prepare industrially relevant catalysts.
A relevant aspect to consider for this goal is to develop metal-oxide
catalysts showing ordered interfaces. Onishi and Iwasawa20 remarked the
role of the interfacial chemistry on metal-oxides on the reactivity. Also in
metal nanoparticles supported on metal-oxides, the interface between the
two plays a relevant catalytic role.21
We suggested that interesting materials for these investigation are based on
thin lms showing on ordered arrays of 1D nanostructures (nanorods,
nanotubes). They are a suitable model materials in studies for bridging material
and complexity gap in catalysis.22 The possibility to synthesize metal-oxide thin
lms with an ordered nanostructure oers several new opportunities for
preparing novel catalysts. These possibilities include (i) nanostructuring the
surface in the form of catalytic nano-reactors, (ii) nanoconnement and
3D geometrical architectures of active sites, (iii) nano-building of catalytic
components, (iv) integration of homogeneous, heterogeneous and bio-catalytic
elements, and (v) heterogeneization of liquid phase reactions in surface-conned
nano-drops. These will be discussed in the following section.

Nano-architecture in oxide-type materials


In general terms, building a dened nano-architecture in oxide-type materials
further extends the concept of nanocatalysis,23 e.g. when the electrons are
conned, and physical and chemical properties are not scalable from the bulk
properties. Studies have been made mainly on clusters/metal particles in the

Catalysis, 2009, 21, 82130 | 83


This journal is
c The Royal Society of Chemistry 2009
nanometer length scale showing that the presence of quantum size eects and
structural uxionality (capability of small clusters to exhibit several structural
forms of comparable energy) markedly inuence the reactivity. Further
connement of these metal nanoparticles within host ordered oxides could
lead, from one side, to further promote these eects and, from the other side,
allow to use these principles to prepare real catalysts.
A well known class of oxides having an ordered nanoporous structure is
that of silica micro- and meso-porous materials.24,25 They oer the advantage
of a great variety of possible ordered nanostructures. Although heteroatoms
and functional groups could be introduced within the framework or anchored
to surface, these systems do not possess the multifunctionality characteristics
of oxide-based materials. Oxide nanoparticles could be introduced within the
ordered mesopore structure,26,27 but they acts as discrete units. As the particle
size is scaled down to a few nanometers, and stabilized within an ordered
micro- or meso-porous structure, the microstructure could be signicantly
altered, changing the surface-to-volume ratio, and particle-to-support inter-
action.28 The mist between the oxide particle and the host ordered porous
material could stabilize defects, create unusual coordination states, and in
general terms signicantly alter the metal-oxide properties and reactivity.26,27
Nanostructures within porous materials show dierent properties which
could be used in a number of applications from electronics, to information
storage devices, photonics and sensors.28,29
However, for application as catalysts, there are some problems. The rst
regards the fact that often the catalytic cycle involves the exchange of electrons
between the reacting molecule and the active centres. This is a critical aspect,
for example, in selective oxidation reactions where there is the possibility of
delocalization of these electrons far from the adsorbed intermediate.3 This long
range transport is also an important aspect for the fast regeneration of the
active site, e.g. for turnover. Decreasing the size of oxide particles, therefore,
has a positive eect in terms of creation of reactive sites, but a negative eect in
terms of long-range transport properties which play also a relevant role. This is
the reason why 1D-type nanostructures such as nanowires and nanotubes are
often preferable for oxides with respect to 0D-type nanostructures, e.g.
spherical nanoparticles below few nanometers. Nesper et al.30 have discussed
in detail how oxidic nanotubes and nanorods represent anisotropic modules
for a future nanotechnology, although with focus on the synthesis aspects.
They clearly show how these anisotropic materials dier from those of the
corresponding bulk material and those of isotropic nanoparticles.
The second limit of oxide nanoparticles within ordered nanoporous
matrices is related to the fact that due to the wetting characteristics of
silica, small oxide nanoparticles, e.g. below about 2 nm, could be stabilized
only for very low loading of the oxide. For example, the loading of
vanadium oxide in silicalite should be lower than about 1% wt. to obtain
such a small nanoparticles.31 Although such small particles show a higher
turnover number than larger particles and also higher selectivity in propane
oxidative dehydrogenation,32 overall catalyst productivity per volume of
catalyst is low. The same conclusions were obtained studying vanadium
oxide nanoparticles within micro- and meso-porous silica materials and a
dierent reaction of selective oxidation (toluene to benzaldehyde).33

84 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
A third problem is related to the slow back-desorption of the products of
reaction, when they form on metal-oxide nanoparticles within a host ordered
porous silica matrix. For example, in toluene oxidation to benzaldehyde over
Fe-Mo-oxide nanoparticles stabilized within a silicalite matrix, the slow rate
of reoxidation of the reduced Fe-Mo-oxide, due to the low nanoparticle size,
increases the presence of reduced molybdenum sites, which, interacting with
the carbonyl group of benzaldehyde, slow down the desorption and enhance
the rate of the consecutive oxidation.34
These motivations strengthen the interest for catalysis towards the devel-
opment of ordered assemblies of 1D nanostructures for oxide materials, e.g.
metal-oxide catalysts in which the 3D macro-structure is constituted by an
ordered assembling of regular 1D structures with nanometric size. Note that
this type of structure is signicantly dierent from that of metal-oxide
supported over other metal-oxides, such as monolayer-type V2O2/TiO2
materials. See also later, when the concept of nanostructured metal-oxide
lms is dened.
These ordered array materials nd interest not only in catalysis, but in
several other applications, from optical materials, sensors, low-k materials,
ionic conductors, photonic crystals, and bio-mimetic materials.35 However,
with respect to these applications, catalysis requires additional specic
characteristics, such as the presence of a thermally stable nanostructure,
the minimization of grain boundaries where side reactions may occur, and
the presence of a porous structure which guarantees a high surface area
coupled to low heat and mass transfer limitations. An ordered assembly of
1D nanostructures for oxide materials could, in principle, meet these
dierent requirements.
Ordered mesoporous oxides, obtained for example by block copolymer-
templated,36 micelle-templated37 or other methods, such as inverted opals,
colloidal templating or double templating procedures,38 or cooperative self-
assembly methodologies,39 oer also interesting opportunities as novel materials
for catalytic applications, but they often lack the necessary thermal stability,
crystalline long-range order, and in particular the 1D-type nanostructure
characteristics which could dierentiate the properties of nano-oxides from
those of the corresponding bulk material.

Role of dimensionality of oxides


There are increasing demonstrations that dimensionality of oxides inuences
signicantly their catalytic behaviour. An elegant demonstration was given by
Ueda et al.40 recently. Schlogl et al.41 showed how 1D vanadium oxide
nanostructures, induced by nucleation over carbon nanotubes, have dierent
characteristics from bulk vanadium oxide and they are selective in n-butane
oxidation to maleic anhydride, while bulk vanadium oxide produces only
carbon oxides. Wang et al.42 showed that gold nanoparticles supported on
b-MnO2 having nanorods or conventional particle shapes have dierent
catalytic behavior in the liquid-phase aerobic oxidation of benzyl alcohol.
The enhanced catalytic activity of the Au/MnO2-nanorod catalyst was
attributed to the benecial presence of a higher amount of oxidized gold
species and surface oxygen vacancies resulting from the strong interaction

Catalysis, 2009, 21, 82130 | 85


This journal is
c The Royal Society of Chemistry 2009
between Au and the well-dened reactive surface of MnO2 nanorods.
Rajeswari et al.43 showed that Pt deposited on WO3 nanorods shows
enhanced properties in methanol oxidation. Zhong et al.44 reported that
gold deposited on a-Fe2O3 nanorods exhibited higher catalytic activity in
CO oxidation than a Au/a-Fe2O3 nanoparticle catalyst. Tungsten-oxide
deposited over titania nanotubes shows enhanced catalytic properties in the
oxidation of dibenzothiophene.45 LaSrCuO4 nanowires, prepared using
carbon nanotubes as template, show higher specic activity in CO oxidation
with respect to LaSrCuO4 nanoparticles prepared by a conventional route.46
TiO2 nano-tube supported Cu species show higher activity in the reduction
of NO by ammonia in the presence of O2 in comparison with similar 2 wt%
Cu/TiO2 catalysts prepared using TiO2 nano-particle supports.47
These results, even if preliminary and not systematic, show that in general
a 1D-type of nanostructure for oxide, as such or as support for metal or
metal-oxide particles, corresponds to higher catalytic performances not only
related to a higher external surface area.
Signicant eort has been made in recent years on the development of novel
methodologies to synthesize oxide nano-wires, -tubes and -rods.30,4852 Most of
these studies were focused on the preparation of 1D type materials of zinc, tin,
vanadium or molybdenum-oxide, while fewer were related to ternary oxides54 or
TiO2,53 even if research activities on the latter or similar materials (titanate) are
fast growing.5559 In most of the cases, research attention was given to synthesis,
while less studies were specically dedicated to the analysis of the functional
properties (behavior as sensors, for example) and nanostructural-functional
properties relationships. The use as catalysts was often mentioned, but much
less specically investigated. In the rst part of the work on nano-architecture
and reactivity of titania catalytic materials,60 we reviewed the state-of-the-art on
the nano-architecture and reactivity of titania catalytic materials, with focus on
1D nanostructures. In this second part, attention is given instead on 2D
nanostructured lms, e.g. bidimensional ordered arrays of 1D nanostructures
of TiO2. In fact, even if various reviews5559 have reported advances in the eld,
they were not focused on the perspectives and possibilities of the use of these
nanostructure for catalysis, or on the issues and problems for their application as
catalysts. In fact, aspects such as the possible scale-up, thermal stability, purity,
etc. were often not considered.
On the other hand, it should be mentioned that often these materials are
used as catalysts only because they are novel, and little attention is given to
clearly discriminate, when they oer real advantages or when the eect is
only apparent, e.g. due to a higher surface area. A higher activity does not
necessarily imply a better catalyst, because, for example, when the activity is
higher due to a double surface area, but the cost of preparation is four times
higher, the higher active catalyst is not preferable, e.g. cost-eectiveness is
the parameter for catalyst evaluation. There are many other aspects to
consider (stability, scaling-up, etc.), but in general it should be analyzed in
detail when the use of new methods to synthesize nanostructured materials
oer real advantages for the preparation of novel catalysts.
In this second part, we extend this approach to 2D titania ordered
nanostructures, e.g. columnar-type lms, ordered arrays of nanotubes or
nano-rods/-wires, nanobowl array, but with some comments also on

86 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
analogous 2D materials such as nano-membranes (called also nanohole
array) and bidimensional nanosponges. We could call these materials as
nanostructured titania lms.
We dene here as lm a 2D oxide layer, where the aspect ratio (consider-
ing the lm as a parallelepiped, the aspect ratio is the ratio between base to
height, the latter being the thickness of the lm) is higher than 300500, and as
nanostructured a lm composed of distinguished, but analogous 1D units
(nanotubes, nanorods, etc.) arranged in an ordered array. The ratio between
diameter of these 1D units to geometrical surface (e.g. surface of the at oxide
lm) should be higher than 10 000. Based on this denition, we caution the
reader that often materials indicated in literature as nanostructured oxides
do not belong to the category we dene. We also remark that following this
denition monolayer oxides, e.g. for example V2O2/TiO2, do not fall within
this denition of nanostructured oxides.

Why focus on titania?


We focus attention here on titania (TiO2) for the following reasons. The
rst is that titania is a widely used oxide support for both metal particles
and metal oxides, and used in some cases also directly as catalyst
(Claus reaction, for example). The second is that it possesses multi-
functional properties, such as Lewis and Bronsted sites, redox centres,
etc. The third is that it has several applications both as a catalyst and
an advanced material for coating, sensors, functional lms, etc. The
fourth is its high photocatalytic activity6164 which make titania unique
materials.
TiO2-based materials and thin lms nd application in a quite broad
range of uses: (i) active coating (self-cleaning coatings and paints), (ii) water,
air and soil disinfection and decontamination, (iii) novel sensors and
membranes, (iv) photovoltaic cells and photoelectron devices, (v) pigments,
(vi) corrosion-protecting layers, (vii) novel textiles, and (viii) niche applica-
tions such as dielectric coatings, carriers for drugs in (nano)medicine,
spin- and opto-electronics, etc., in addition to catalysis and photocatalysis.
The further motivation to choose titania derives from the fact that there
are already indications that the control of the nano-architecture of titania-
based materials is a fundamental key to improve the behaviour for
challenging reactions such as the production of H2 and O2 by water
splitting.65,66 The need to have an ordered array of 1D nanostructures
was also shown, in order to improve light harvesting, minimize grain
boundary, allow a fast and vectorial-type charge transport to minimize
e /h+ recombination, minimize defects, etc. However, we believe that the
presence of an ordered array of 1D nanostructures oers general advan-
tages for preparing catalysts, even if there are still many problems to solve,
as discussed later.

Titania nanostructured lms


There are many methods to prepare titania thin lms,6773 the principal of
which are metallorganic chemical vapour deposition (MOCVD), sol-gel,

Catalysis, 2009, 21, 82130 | 87


This journal is
c The Royal Society of Chemistry 2009
vacuum thermal evaporation, plasma or magnetron sputtering. Some of
these produce ordered or pseudo-ordered nanostructured lms, but in
general a randomly arranged assembly of nanocrystals is present.
When used for photocatalytic applications, or to produce novel photo-
voltaic materials (by photosensitivization of the titania by dyes,
the well known Gratzel cells,7274 or analogous systems75) a main problem
regards the high grain/crystal interface, because this enhances the presence
of traps, defects and adsorbed species which act as recombination
centers for the holes and electrons produced by light absorption in the
TiO2 semiconductor particles. A columnar nanostructure with a fast
mechanism of transport and collection of electrons would enhance the
charge separation and increase also the light harvesting, as schematically
shown in Fig. 1.
Some mechanisms could mitigate charge recombination in Gratzel-type
cells:75 a dye bonded to the TiO2 surface passivates recombination
centers, and suppression of trapping-detrapping events at the surface

Fig. 1 Schematic model of the dierences between TiO2 thins lms and aligned nanorod array
of TiO2.

88 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
Fig. 2 Number of publications (only in English) in the last 8 years found by SciFinder using as
keyword TiO2 nanotube or TiO2 nanotube array: as entered and concept.

increases the diusion coecient of the electrons through the


nanocrystal matrix facilitating electron transport to the back contact.
Nevertheless, the nanostructure architectures for solar energy
conversion play a fundamental role to achieve ecient fuel production
(e.g., solar hydrogen production) or electricity (photochemical solar
cells).7680
We will not discuss here aspects related to the use of TiO2 nano-
structured lms for dye-sensitivized photovoltaic cells, even though this
is a major eld of use of these materials, because we will focus the
discussion on catalysis and related aspects. However, several of the aspects
discussed, particularly for photo-reactivity applications, are common also
to this eld.
The number of publications in this eld is exponentially raising, as shown
in Fig. 2 which reports as example the number of publications found by
SciFinder in the last 8 years using as keyword TiO2 nanotube or
TiO2 nanotube array. The largest part of them is related to aspects
connected to preparation and characterization, or use mainly as photo-
catalysts. Nevertheless, this demonstrates that a large body of knowledge is
available for its use in preparing advanced catalysts. Some of the possibilities
in this eld are outlined in the following section.

2. Outlooks for the development of catalysts based on the concept


of nanostructured lms
Although the discussion reported here will be mainly in reference to titania
nanostructured lms, the opportunities and problems analyzed are of more
general interest and regard how to use the concept of bidimensional
nanostructured oxide lms to prepare novel catalysts with tailored
nano-architectures.

Catalysis, 2009, 21, 82130 | 89


This journal is
c The Royal Society of Chemistry 2009
Possibilities oered from structuring the surface as array of nanoreactors
Fig. 3 shows a conceptual example of the possibilities oered from
structuring the surface of oxides in terms of array of TiO2 nanotubes which
can be viewed as an ensemble of nanoreactors. The SEM image in the inset
of Fig. 3 shows an example of the TiO2 nanostructures obtained in the case
of anodic oxidation of titanium foils.82
Some of the possibilities oered from this nanoreactor structure are also
outlined in Fig. 3 and may be summarized as follows:
 Nano-Connement. There are limited, but interesting studies, regarding
the connement in ordered mesoporous materials. First observations were
made on nematic liquids within mesoporous SBA-15 host materials which
showed a change in the phase transition, when conned within the
mesoporous cavities.83 To evidence also that there are many studies of
connement in mesoporous materials in the polymer diusion and
membrane literature, but they refer essentially to entropic eects due to
restricted motion of these materials inside the ordered mesoporous materials
which in enhanced by more hydrophobic and less polar surfaces.
This is especially true as the molecules become larger, because the
number of conformations the molecule can adopt in a conned space is
limited. We refer here, on the contrary, to aspects relevant for catalysis
and in which thus the dimensions of the molecules (of the order of 0.1 nm) is
far below the dimensions of the cavities (around 5 nm for SBA-15, for
example).
The rst demonstration, to our knowledge, that the properties of a gas or
liquid within a mesoporous cavity change was made by Dosseh et al.84
studying the properties of cyclohexane and benzene conned in MCM-41
and SBA-15. The eect was related to the inuence of the inner surfaces of
mesoporous silica. Other authors have further demonstrated the
inuence of connement on the adsorption and properties of uids within
an ordered mesoporous material.8587 However, Fajula` et al.88,89 were the

Fig. 3 Concept of nanoengineering of oxide catalytic surface in terms of nanoreactor array,


some of the possibilities oered by this concept (in particular in terms of realizing multi-
functional catalysts for cascade reactions in nanoconned liquids) and a SEM image of an
array of TiO2 nanotubes produced by anodic oxidation of Ti foils. Source: Centi et al.81

90 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
rst showing that an enthalpic excess is present during the adsorption
processes, due to a wall eect (connement) in mesoporous materials.
Connement eect was shown to inuence the catalytic behaviour.
In addition, it was observed that monitoring the position of IR vibrations
of supercritical CO2 (scCO2) a virtual higher pressure than the applied
one could be identied (the position is depending on the pressure),
once liquid scCO2 is conned inside a periodically organized mesoporous
material.90
The eect is thus not related to geometrical constraints induced on
complexes anchored in mesoporous channels (sometimes also called as
connement eect, even if this denition is not properly correct),91 neither
to shape-selectivity eects as possible in zeolites, since the size of mesoporous
channels is much larger than those of micro-porous materials. Instead, an
eective modication on the characteristics of the uids is observed due to the
electrostatic eld generated by the channel walls. This is an enthalpic eect
versus an entropic eect as observed when the modication is instead related
to limitations in the translation modes of molecules. Recently, it was also
demonstrated that wall curvature inuence the molecular orientation of the
transition states.92
The connement in mesoporous materials thus inuences the properties
of uids within the cavities and in turn the catalytic behaviour. An example
is given in Fig. 4 which reports a comparison of the behaviour of Pd
supported on silica or deposited inside the channels of a mesoporous silica
(SBA-15) in the direct synthesis of H2O2 from H2/O2 using methanol or
CO2-expanded methanol.93,94 The use of CO2-expanded methanol increases
the productivity and the selectivity to H2O2 for both catalysts, but the eect
is particularly enhanced in the case of the mesoporous material. The eect is
equivalent to a higher virtual pressure within the mesoporous channels
which inuences from one side the partial pressure of H2 and O2 (virtual
solubility) and from the other hand the uid viscosity. In fact, by increasing
the total pressure Pd-SiO2 also shows similar productivities and selectivities.
In the absence of connement eect, e.g. dierent uid characteristics or
higher virtual pressure induced by the eect of the wall in a mesoporous
material, above eects cannot be explained. We have also shown that the
characteristics of the wall (hydrophobicity, for example), as well of the uid,
inuence largely the connement eect. Therefore, reaction in conned
environments could further enhance the performances of CO2-expanded
solvents to make possible the direct synthesis of H2O2 in milder and safer
conditions.
A bidimensional array of oxide nanotubes as that schematically shown in
Fig. 3 shows some advantages over mesoporous silica. In fact, in mesopor-
ous silica the elongation factor of the channels, e.g. the ratio between length
and diameter is typically a factor at least 100 or higher, while it is lower in a
bidimensional oxide nanotube array as schematically shown in Fig. 3.
Furthermore, it is dicult to realize a mesoporous silica lm with the
channels vertically aligned with respect to the surface. Both these aspects are
relevant for full accessibility of the inner surface, e.g. to eliminate diusion
limitations. We may also comment that the local electrostatic eld within a
TiO2 nanotube is signicantly dierent from that of an equivalent SiO2

Catalysis, 2009, 21, 82130 | 91


This journal is
c The Royal Society of Chemistry 2009
Fig. 4 Selectivity and productivity in H2O2 direct synthesis at room temperature using Pd-SiO2
and Pd-SBA-15 catalysts and CO2-free or CO2-expanded methanol. See Centi et al.93,94 for details
on the reaction conditions.

nanotube (note that characteristics os silica nanotubes are well dierent


from those of silica particles due to their dierent modality of preparation),
as well as the hydrophilic characteristics, number and strength of Bronsted
sites, presence of Lewis acid or basic sites, etc. In addition, electron
delocalization, and mobility of adspecies is quite dierent from TiO2
(or other redox and/or semiconductor oxides) with respect to silica.
Therefore the characteristics of a uid conned in a TiO2 nanotube and
the inuence of this eect on the catalytic reactivity are dierent with
respect to mesoporous silica.
 3D geometrical architectures of the active centres. As commented in the
previous point, a nanostructure such as that shown in Fig. 3 shows a
number of features dierent from that possible for mesoporous silica. There
are various studies which show that the connement of complexes within
mesoporous silica opens a number of possibilities. For example, Thomas
and Raja95 recently discuss how to exploit the mesoporous nanospace for
asymmetric catalysis, including the opportunities given by the connement
of immobilized, single-site chiral catalysts to enhance enantioselectivity.
Santiso et al.96 reviewed some possible connement eects (inuence of
steric hindrance on the equilibrium and kinetics, inuence of electrostatic
interactions with the supporting material on the reaction mechanism and
equilibrium yield for reactions involving a charge transfer) and how to use
this knowledge for catalyst design. A redox oxide such as TiO2, due to its
multifunctional characteristics, signicantly extends the range of possible
design of novel catalysts having a dened and tailored 3D architecture of
the active centres. In addition, nanostructured TiO2 lms as those shown in
Fig. 3 oer new possibilities to design improved photocatalysts. In a recent
review97 the progresses in photocatalysis on mesostructured systems, which
behave as spatially conned micro- and nanoreactors, were presented.
Mesoporous oxides and hybrid photocatalysts based on molecular-
organized assemblies oer new possibilities for converting low reactant

92 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
levels, to control the reaction environment, and realizing hostguest
catalysis of light-driven reactions. An increase of selectivity can be attained
through functionalization of spatially conned photo-reactors which allow
coupling of dierent photo- and redox-catalysts, and to create hybrid
micro-photoreactors including those which mimic the natural photo-
synthetic centers. Fig. 3 shows that TiO2 nanotubes can host metal
nanoparticles, anchored as metallo-organic complexes, and enzymes as
well, to develop a multifunctional and highly ecient nanoreactor, easily
accessible by reactants. Interesting recent results have been reported for
cascade catalytic reactions in polymeric nanoreactors.98 Catalysts based on
TiO2 nanotube arrays could oer new possibilities to realize robust and
recyclable materials for cascade catalytic reactions. In a very recent review99
catalytic reactions within a conned nanospace (molecular capsules,
zeolites and micelles) have been discussed, with reference in particular to
self-assembled nanocapsules based on non-covalent bonds and used as
nanoreactors for various types of organic and metal catalyzed
reactions. The concept of nanoreactor has been also patented recently by
Somorjai et al.100
 Nano-building of catalytic components. The nanostructure of TiO2
nanotube arrays oers also new possibilities for nano-building of catalytic
components, e.g. acting as nanoreactors for spatially conned synthesis
of composite materials. Shchukin et al.101 have discussed the use of
hollow nanostructures (nanotubes and capsules) for the synthesis of nano-
particles of metals and metal oxides. Li and Zeng102 described the use of
hollow core-shell Au-TiO2 nano-composites to form nanoreactors that
contain a catalyst core with a shell that is permeable to reactants.
Chung and Rhee103 have shown that Pt-Pd bimetallic nanoparticles
prepared within a dendrimer nanoreactor show enhanced catalytic
activity in the partial hydrogenation of 1,3-cyclooctadiene. Highly active
platinum catalyst nanoparticles could be prepared in polymers
nanoreactors.104 Gold nanoparticles deposited within TiO2 nanotubes
have shown an activity for CO and H2 oxidation dierent from that of
Au on anatase or rutile TiO2 catalysts.105 Ag nanoparticles in nanotubes
also showed a dierent recativity.106 Therefore, the TiO2 nanotubes could
be used as nanoreactors for a controlled preparation of other guest
nanoparticles which reactivity could be dierent from that shown by
deposition of the same guest nanoparticles on TiO2 (or other materials)
particles.
 Integration of homogeneous, heterogeneous and bio-catalytic elements.
Previous comments have already shown that TiO2 nanotubes oer inter-
esting possibilities to integrate in a dened nanospace organometallic
complexes, metal particles, other active species (Bronsted or Lewis sites,
transition metal ions coordination sites, etc.) and bio-catalysts. Enzymes
immobilized within nanoreactors could be better reused and show enhanced
reactivity.107 A TiO2 nanotube array was shown to be a very eective
substrate to host Horseradish Peroxidase for preparing biosensing
devices,108,109 and immobilize glucose oxidase for electrochemical
sensors.110,111 Hydrogenase enzymes coupled to TiO2 showed interesting
results in the photo-induced hydrogen production.112 Bio-nano-composite

Catalysis, 2009, 21, 82130 | 93


This journal is
c The Royal Society of Chemistry 2009
photoelectrodes were fabricated through immobilization of bacterial photo-
synthetic proteins on a nanocrystalline TiO2.113 Saccharide-O2 biofuel cells
were prepared using TiO2 sensitivized with zinc chlorin-e6 (ZnChl-e6).114
Porphyrin-sensitized TiO2 was studied as a photoanode material for a new
cell that converts light energy into electricity.115 These examples show how
bidimensional nanostructured TiO2 lms such as those schematically
represented in Fig. 3 oer various opportunities not only for catalysis to
host homo- hetero- and bio-catalysts in a single nanoreactor to realize
cascade catalytic reactions,116 but also to develop a number of other
interesting applications, from novel electrodes for biofuel cells to photo-
anodes for conversion of light energy into electricity. The specic nano-
structure of TiO2 allows to improve the performances with respect to lms
composed by porous TiO2 particles.
In addition, materials prepared by anodization permit the growth of the
oxide nanostructured lm over a conductive substrate (Ti foil), an important
aspect for preparation of robust electrodes. As discussed later, electrical
contact with the conductive substrate could be further improved using carbon
nanotubes.
 Heterogeneization of liquid phase reactions in surface-conned nano-
drops. A further opportunity given by the TiO2 nano-architecture shown in
Fig. 3, is related to the possibility of using capillary condensation of a liquid
within the nanotube to have a liquid solution inside, while outside the
nanotube, the partial pressure is lower than that necessary for liquid lm
condensation. In carbon nanotubes capillary condensation is used for
development of capacitive humidity sensor.117 Kim et al.118 showed the
presence of capillary condensation at room temperature within carbon
nanotubes for glycerine, ethylene glycol, and distilled water, for partial
pressures where no formation of a liquid lm over oxide particles could be
expected. Therefore, using this concept, it is possible to realize TiO2
nanoreactors inside of which a liquid solvent is present, but outside the
temperature/pressure conditions do not allow liquid condensation. This
allows combining the properties of homogeneous catalysts to the advan-
tages of heterogeneous catalysts, e.g. no need of separation (the system may
operate as an heterogeneous catalyst), no mass transport limitation, and
higher productivity. Supported liquid phase catalyst (SLPC) is one way to
support homogeneous catalysts, by using organometallic complexes which
are dissolved in a small quantity of liquid phase dispersed in the form of isle
or lm on the surface of supports.119 The SLPC has successfully been
applied for several catalytic chemical transformations. Fow et al.120 recently
demonstrated that chiral organometallic complexes immobilized in silica
supported thin lms of ionic liquids show enhanced enantioselectivity for
chiral hydrogenation. Immobilization of organometallic complexes in thin
supported lms of ionic liquids allow generation of a new class of hydro-
amination catalysts.121 The concept of immobilizing organometallic
complexes in a thin lm of supported ionic liquids was used to synthesize
novel bi-functional catalysts combining soft Lewis acidic (a palladium
complex [Pd(DPPF)(CF3CO2)2], prepared in situ from Pd(CF3CO2)2 and
1,1-bis(diphenylphosphino)ferroceneDPPF), and strong Bronsted acidic
functions.122 The materials showed exceptional catalytic activity for the

94 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
addition of aniline to styrene, providing the Markowniko product under
kinetically controlled conditions and mainly the anti-Markowniko
product in the thermodynamic region. Heterogeneization of liquid
phase reactions in surface-conned nano-drops, e.g. inside TiO2 nano-
reactors, oers additional possibilities to control and improve this class of
reactions.
Therefore, the concept of bidimensional nanostructured lm presented in
Fig. 3 oers several opportunities to develop novel catalysts. However, for
practical applications a major problem is related to the fact that if a thin
lm on a at surface is used, the amount of catalyst is quite low. For
one cm2 of surface and a thickness of the titania lm of 1 micron, the
amount of TiO2 is about 1 mg assuming a density of 1 g/cm3, while density
is even lower. This means that to develop practical applications it is
necessary to improve the geometrical area over which the nanostructured
thin lm is deposited. A possible solution is to use high surface area foams
over which the nanostructured titania lm is then created. Another solution
is to create a multi-stack of thin substrates over which the nanostructured
titania lm is present. Other solutions are possible, but the use of the
concept presented in Fig. 3 requires rst solving the issue on how to prepare
materials with enough amount of the nanostructured oxide layer to be used
as heterogeneous catalyst. As shown later, however, very recent studies have
shown that it is possible to reach a lm thickness up to several hundred
microns, and not few microns or less as in the initial studies. Therefore, the
issue of the amount of oxide per cm2 of geometrical area of the catalyst is
now a less critical aspect. However, there is still the problem of suitable
shapes to be used in catalytic reactors. There is thus still the need to transfer
this new knowledge to the preparation of catalysts, for example in the form
of multistack monoliths or foams.

Electro- and photo-catalytic nanostructured materials


For applications as electro- or photo-catalysts, or for other applications as
sensors, solar cells, electrodes in biosensing and biofuel cells, etc. the problem
of the amount of titania is less critical, but there is still the need to optimize
the lm thickness. One of the active directions of research in this area, in
fact, is to maintain the nanostructure, but increase the lm thickness.
A new benchmark for TiO2 nanotube arrays was recently reached.123
A self-standing 720 mm thick TiO2 nanotube membrane prepared by
anodization of a 250 mm thick Ti foil sample was obtained. A double sided
electrochemical oxidation of Ti in an electrolyte comprised of H2O, NH4F,
and ethylene glycol was used to produce 2 highly ordered TiO2 nanotube
arrays 360 mm in length that are separated by a thin compact oxide layer. The
individual nanotubes have an aspect ratio (ratio between length to diameter)
of approximately 2200. The potentiostatic anodization of Ti in an ethylene
glycol, NH4F, and H2O electrolyte dramatically increases the rate of nano-
tube array growth to approximately 15 mm/h, representing a growth rate
approximately 7506000% higher than that seen, respectively, in other polar
organic or aqueous based electrolytes previously used to form TiO2 nanotu
be arrays.

Catalysis, 2009, 21, 82130 | 95


This journal is
c The Royal Society of Chemistry 2009
The same authors also reported recently that self-aligned hexagonally close-
packed TiO2 nanotube arrays of over 1000 mm in length and aspect ratio higher
than 10 000 could be produced by potentiostatic anodization of Ti.124 Other
authors have reported results on lms up to 250 mm,125 but in general lm
thickness is few microns or below. For a lm thickness of 1000 mm, the amount
of titania is of the order of one gram per cm2. Therefore, there are good
prospects that the issue of catalyst amount is overcome, and probably within
relatively short time the application of these materials as practical catalysts in a
number of reactions will rapidly increase.

Titania nano-membranes
Another attractive area of development is to prepare catalytic nano-
membranes. First attempts to prepare self-organized, free-standing TiO2
nanotube membranes for ow-through photocatalytic applications was
reported by Schmuki et al.126 who showed a method to prepare by anodic
oxidation (ethylene glycol+0.2 M HF at 100 V for 10 h anodization)
robust, dense and free-standing membranes consisting of vertically oriented,
both-side-open TiO2 nanotubes. The array consists of regular tubes with a
diameter of 160  30 nm and a wall thickness of 20  5 nm. The layer is
overall 145 mm thick and has very smooth walls typical of nanotubes grown
in organic electrolytes.127 The TiO2 nanotube layers were separated from
the Ti substrates by selective metal dissolution. The TiO2/Ti specimens were
immersed into a mixture of Br2 and dry methanol for 12 h under a dry N2
atmosphere. This leads to a free standing nanotube layer oating in the
etching solution. After being rinsed with methanol and distilled water, the
layers were placed (closed tube side down) 12 cm above an open HF 48%
bottle for 30 min. This leads to HF condensation at the bottom and
preferential etching of the tube bottoms, i.e. the procedure opens the tubes.
After this treatment an opaque free-standing layer is obtained. The bottom
of the tubes is closed. Fig. 5 reports top and cross-section scanning electron
images (SEM) of this membrane prepared by Schmuki et al.126 for
size-selective, ow-through photocatalytic reactions (methylene blue
decoloration).
Similar TiO2 nanotube array membranes of uniform pore size distribution
were prepared also recently by Paulose et al.128 and tested in bioltration
applications. The size of the membranes was 12.5 cm2, a size limited by the
processing equipment. These membranes can be used in lab-scale tests, but
are fragile.
It should be mentioned that alternative possibilities to prepare similar
membranes include the use of a porous alumina membrane as matrix,
with the titania nanotubes grown in the channels.129133 Nanoporous
alumina membranes are commercial products, also synthesized by anodic
oxidation. The commercial Whatman Corporation anodic membrane has
holes of about 20-nm diameter at the top of the membrane and about
200-nm diameter at the bottom of membrane.133 Within these pores
TiO2 nanotubes fabricated by template synthesis and water vapour
hydrolysis could be grown, but non-uniform membrane characteristics
are obtained due to the non-uniform pores of the commercial alumina

96 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
Fig. 5 SEM images of the TiO2 nanotube membrane prepared by Schmuki et al.126

membrane. More recently, silica/titania nanotubes composite membranes


for photocatalytic applications were prepared also using porous alumina
as a matrix.134 It was reported that the silica/titania nanotubes composite
membrane had the multiple functions of separation, degradation,
and improvement of membrane ux in photo-oxidation of organic
contaminants in wastewater. Although the use of a matrix could improve
robustness of the membrane, the approach of preparing free-standing TiO2
nanotube membrane seems preferable. However, it may be cited that
macroporous silicon membranes could be prepared by photo-assisted
electrochemical etching and subsequent opening of the pores from the back
side.135 This silica membrane is characterized by larger channels (about
3 mm in diameter and 200 mm in length) and could be a better host for inner
titania nanotubes, giving at the same time a better robustness to the
material.

Catalytic nanofactories
There is increasing interest in preparing TiO2 nanomembranes136 both for
advanced photocatalytic processes in the eld of air and water purication,137
purication of drinking water,138 novel membrane for high temperature PEM
fuel cells,139,140 Li-ion batteries,141143 advanced nanoelectrode arrays
(NEA)144 and nanoltration and pervaporation.145
The possibility of free-standing titania membranes with thickness of
about 150200 microns and straight channels of 100200 nm opens new
possibilities for catalysis. A grand-challenge for catalysis and sustainable
chemistry is to realize new materials able to perform multistep complex
reactions with ideally 100% selectivity. One approach to go in this direction
is the possibility to develop a tailored sequence of active centres inserted in a
channel (nanotube) of an ordered membrane to perform selectively the
conversion of molecules passing through this porous lm, minimizing at the
same time the possibility of side reactions. This concept was indicated as
catalytic nanofactories, because it resembles that of a production chain in a
factory, where multiple sequential operations are made in a chain-like
sequence. Fig. 6 shows schematically this concept.81 Although knowledge

Catalysis, 2009, 21, 82130 | 97


This journal is
c The Royal Society of Chemistry 2009
Fig. 6 Concept of catalytic nanofactories. Adapted from Centi et al.81

has to be still acquired to implement this concept, it should be considered a


vision to develop new generation of catalysts.
The concept presented in Fig. 6 could use also other type of ordered
mesoporous membranes, based on silica for example. As discussed before,
oxides such as TiO2 provide better multi-functionalities for the design
of such a type of nanofactory catalysts. Worth to note is that in the
cover picture of the recent US DoE report Catalysis for Energy146 a very
similar concept was reported. This cover picture illustrates the concept, in
part speculative, that to selectively convert biomass-derived molecules to
fuels and chemicals, it is necessary to insert a tailored sequence of enzyme,
metal complexes on metal nanoparticles in a channel of a mesoporous
oxide.

Light-driven control of reactivity


A specic peculiar aspect of nanostructured TiO2 lms for catalytic applica-
tions is related to the semiconductor character of TiO2. Upon light irradiation
of TiO2 with enough energy (about 3 eV), an electron is promoted from the
valence to the conduction band, thus creating a corresponding hole. This is
the rst event of photocatalytic processes, because then the electron and holes
on the surface of TiO2 may react with the incoming molecules to give rise to

98 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
the sequence of reactions involved in the photochemical transformation.
However, the electrons generated in the process may interact also with
supported metal particles which acts as electron acceptors (Pt, Au or Ag
nanoparticles, for example), thus modifying their properties. The noble metal,
which acts as a reservoir for photo-generated electrons, promotes an
interfacial charge-transfer process. In addition, being titania a semi-
conductor, it is possible also to apply an external bias and further control
the characteristics of the supported metal particles by this eect. In other
words, with respect to alumina or silica, the catalytic properties of metal
particles deposited over TiO2 can be modied and in part controlled by
illumination and an external bias. Several studies on photoinduced electron
transfer processes for metal particles supported on titania were published
(see for example, the review of Kamat76 on nanostructure architectures for
solar energy conversion), but more limited results are reported on how to use
these concepts to tune the catalytic performance, because attention is
generally focused on the photocatalytic process itself and not on how light
irradiation of TiO2 could modify the reactivity of supported metal particles.
By light irradiation and/or application of an external bias, it is possible to
induce transient modulations in the catalytic performances, opening new
interesting perspectives on how to externally control the reactivity properties
of a catalyst.
A specic and peculiar additional possibility was suggested147 from the
observation that an ordered array of TiO2 nanocoils forms in particular
conditions of synthesis by anodic oxidation.82,148 The interesting aspect is
that in such nanostructures, photo-induced current generates a local
magnetic eld which inuences the reactivity properties of catalytic

Fig. 7 FESEM images of titania nanocoil produced by anodic oxidation (a). The cartoon
shows schematically the photocurrent generated by light irradiation of nanocoil containing a
catalyst particle (b) and the associated magnetic eld (c). Source: Centi and Perathoner.147

Catalysis, 2009, 21, 82130 | 99


This journal is
c The Royal Society of Chemistry 2009
nanoparticles sitting at the bottom of these nanocoils or anchored
at the inner walls (Fig. 7).147 Both the structure of the catalytic
particles and chemisorption of reactants and desorption of the
products (and transition state as well) may be inuenced by the presence
of a local magnetic eld. Quite interesting, this observation opens
an exciting opportunity to develop catalysts with photo-switchable
reactivity.

3. Synthesis of titania nanostructured lms


We refer here only to the synthesis of 2D titania nanostructured lms,
because several aspects regarding the synthesis of TiO2 1D nanostructures
(nano-tubes, -rods, -wires, etc.) were discussed in part 1.60 In addition,
this section will not provide a systematic analysis of the methodologies
of synthesis of these materials, or of their characteristics, because only
selected aspects relevant for the objective of discussion on their
use as catalysts will be analyzed. A number of reviews were published
recently on TiO2 nanotubes5558,65,136 and can be used for further under-
standing of the aspects related to the preparation of these materials. Further
information can be also found on a web site dedicated to titania
nanotubes.149
As pointed out in part 1,60 1D oxide nanostructures do not represent only
particles with high elongation ratio, but possess dierent specic characteristics
related to the nanostructure. For example, O2 adsorption is dierent from that
observed for TiO2 single crystals150 and this has consequences on the catalytic
behaviour. In order to exploit the properties of 1D oxide nanostructure,
however, it is essential to orient nanotubes on substrates and to create ordered
arrays.

Methods to prepare ordered arrays of 1D nanostructures


Many dierent approaches have been explored. They are based on
lithography, use of nano-tools (e-beam, X-ray, ion beam, STM, AFM),
or rely on self-alignment processes. The rst attempts have been made using
as a template nanotube aluminum oxide129,151,152 which is prepared by
anodic oxidation of Al lms153 or polymer mold, on which titanium oxide
was deposited electrochemically.154 These methods belong to the general
class of procedures indicated as nanocasting and nanocoating.155 The
dierence between the two techniques is that casting is a lling of the
porous structure of the material, whereas coating results in a layer of
the inorganic substance on the polymer structure. Following formation of
the hybrid, the organic template can be removed, yielding a structured
inorganic material.
TiO2 periodic microstructures can be prepared using a biomimetic
approach with poly(methylmethacrylate) (PMMA) lms with ordered
microcavities as a mold.156 Fabrication of well-ordered high-aspect-ratio
nanopore arrays in TiO2 single crystals is possible by nanolithography
involving swift heavy ion bombardment through a porous anodic alumina
mask.157 The growth of oxide nanorod arrays by template sol electro-
phoretic deposition allows the formation of an ordered packing.158

100 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
Patterned nanorod arrays of TiO2 for photonic applications could be
produced by coating patterned and aligned ZnO nanorod arrays.159
A TiO2 rutile nano-rod array on a-Ti surfaces was prepared by coating a
layer of sodium diborate on titanium substrates and subsequent thermal
treatment. The rods in the arrays grow almost perpendicular to the
substrate surface.160 Aligned TiO2 nanoneedles with diameters in the tens
of nanometers can be grown in solution from patterned nanocavities under
the inuence of an electrical eld.161 The latter, applied perpendicular to the
substrate plane, drove the precursor solution into the cavities by
overcoming the surface tension encountered and oriented the TiO2 nano-
needles during the growth process. Aligned TiO2 nanowire arrays were
fabricated onto Si wafers by a thermal deposition (PVD) method.162
TiO2-based composite nanotube arrays could be prepared via layer-by-layer
assembly.163 Oriented TiO2/glass nanoake array lms could be produced
by leaching of a solution phase from the glass surface through hydrothermal
processes.164 Electron microscope observations revealed that the nanoakes
formed a continuous porous three-dimensional-network array with a large
surface-to-volume ratio. Dense nanorods array with approximately 200 nm
diameter and 10 mm length were obtained from colloidal TiO2 solution
under centrifugation.165 Dense arrays of vertically aligned titania nanotubes
are created directly on silicon substrates by combining atmospheric
layer deposition with an alumina template-based fabrication approach.166
Nanotubes were fabricated with tube walls thin enough (o3 nm) to exhibit
a wall-thickness-dependent blue-shift in the optical absorption spectra of
the arrays. Titania nanohole arrays could be prepared using a nanoporous
alumina and combining the dissolution of the anodic alumina to a process
of deposition of the titania.167 Dense and well-oriented rutile TiO2 nanorod
arrays were synthesized on a titanium substrate using dibutyltin dilaurate
as the oxygen source in the oxidation of Ti at 850 1C.168 Polycrystalline
TiO2 grains were formed at 800 1C; in contrast, TiO2 micro-whiskers were
grown on the Ti substrate at 900 1C. Ordered TiO2 nanobowl arrays could
be produced starting from a self-assembled monolayer of polystyrene
spheres, which is used as a template for atmospheric layer deposition
of a TiO2 layer.169 After ion-milling, toluene-etching, and annealing of
the TiO2-coated spheres, ordered arrays of nanostructured TiO2 nanobowls
have been fabricated.
This brief overview of some of the methods used to prepare bidimensional
TiO2 lms shows that a variety of nanostructures and types of array packing
can be produced by the dierent methods, but that in general the preparation
methods are either complex, costly or dicult to reproduce and control in
order to prepare catalysts or materials for other applications. For this
reason studies have been focused recently on a methodology which allows
to prepare in a cheap and reproducible way (under right conditions) ordered
titania nanostructures, e.g. the anodization technique. Besides to these
advantages, it was shown148 that dierent types of nanostructures, from
nanotube to nanorod arrays, and to spongy type thin layer can be synthetized
by changing the reaction parameters, but using a similar procedure. This
allows a much better analysis of the role of the nanostructure on the catalytic
performance.

Catalysis, 2009, 21, 82130 | 101


This journal is
c The Royal Society of Chemistry 2009
Anodization technique
It is known from decades, particularly for Al, that porous oxide layers can
be grown by anodization typically in acidic electrolytes, while anodization
in neutral electrolytes typically leads to a compact oxide layer. However,
Masuda et al.170 were the rst who showed that a very high degree of order
can be achieved for these porous geometries. Zwilling et al.171 rst reported
the porous surface of titania lms electrochemically formed in uorinated
electrolyte by titanium anodization, but only a decade later Grimes et al.172
showed that the nanostructure is constituted by uniform titania nanotube
arrays.
This result has opened a large range of possible applications, because
vertically oriented, highly ordered TiO2 nanotube arrays have been found to
possess outstanding charge transport properties enabling a variety of
advanced applications: sensors,173175 dye sensitized solar cells,176,177 and
in hydrogen generation by water photoelectrolysis.178,179 Furthermore
materials with aligned porosity in the sub-micron regime are of great
interest for application in organic electronics, microuidics, molecular
ltration, drug delivery, and tissue engineering.180
First studies were focused at controlling nanotube morphology, length
and pore size, and wall thickness. Typically growth occurs proportional to
the applied potential with a growth factor of about 15 nm/V181 up to a
voltage, where dielectric breakdown of the oxide occurs. The structure of
the as-grown oxide can be amorphous or crystalline, strongly depending
on the specic electrochemical parameters used such as the applied
potential, the time of anodization, or the sweep rate of the potential ramp.
Depending on the anodizing conditions the crystal structure has been
reported to be anatase, a mixture of anatase and rutile, or rutile. However,
if uoride ions are present in electrolytes and suitable anodization
conditions are used, ordered nanotube nanotube/nanoporous structures
of TiO2 can be formed. Electrolyte composition, and its pH, determines
both the rate of nanotube array formation, as well as the rate at which the
resultant oxide is dissolved. In all cases, a uoride ion-containing electrolyte
is needed for nanotube array formation. The reaction mechanism is
schematically reported in Fig. 8.
The mechanism is determined from the competition between anodic oxide
formation:
Ti + 2H2O - TiO2 + 4H+ + 4e (1)

and chemical dissolution of the oxide and metal ions as soluble uoride
complexes,
TiO2 + 6F - [TiF6]2 (2)
Ti 4+
+ 6F - [TiF6] 2
(3)

Reaction (1) describes the oxide growth on an anodized metal surface, as


schematically depicted in Fig. 8a in a uoride-free electrolyte. Oxidized
metal species react with O2 ions (from H2O) to form an oxide layer.

102 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
Fig. 8 Schematic representation of the Ti anodization (a) in absence of uorides (results in at
layers), and (b) in presence of uorides (results in the tube growth). F = Field Strength. Source:
Schmuki et al.136

Further oxide growth is controlled by eld-aided ion transport (O2 and


Ti4+ ions) through the growing oxide. As the system is under a constant
applied voltage, the eld within the oxide is progressively reduced by the
increasing oxide thickness, the process is self-limiting. In the presence of
uoride ions, water-soluble TiF62 complexes form. Due to the small ionic
radius, they enter the growing TiO2 lattice and can be transported through
the oxide by the applied eld (thus competing with O2 transport). The
complex formation leads to a permanent chemical attack (dissolution) of
formed TiO2 and prevents Ti(OH)xOy precipitation, because the Ti4+ ions
at the oxide/solution interface are solubilized as TiF62 before precipitating
as Ti(OH)xOy.
The shape, feature and array density of the titania 1D nanostructures
depend on the relative rates between these processes which in turn depend
on the procedure of anodization, time, solvent, etc. During anodization
continuous growth of oxide takes place at the inner interface, and chemical
dissolution of the oxide layer occurs simultaneously. Steady state is
established when the pore growth rate at the metal oxide interface is
identical to the thickness reducing dissolution rate of the oxide lm at the
outer interface. It should be remarked that the chemical dissolution of TiO2
occurs over the entire tube length. As a consequence, the tubes assumes
progressively a v-shaped morphology, i.e., at the top the tubes possess
signicantly thinner walls than at their bottoms.182,183 The reason for
separation into tubes probably should be ascribed to accumulation of
uoride species at the tube bottom and thus to establishment of an anion
containing weaker (and more soluble) TiO2 structure between neighbouring
pores/tubes.

Factors controlling nanostructure and oxide lm thickness


The key factor controlling the tube diameter is the anodization voltage.184 For
anodization experiments carried out in 1 M H3PO4+0.3 wt% HF it has been
shown that the tube diameter can be grown in the range of 15120 nm in the

Catalysis, 2009, 21, 82130 | 103


This journal is
c The Royal Society of Chemistry 2009
potential range between 1 and 25 V.184 For mixed glycerolwater electrolytes
containing 0.27 M NH4F, the tube diameter range was further extended from
20 up to 300 nm in the potential range between 2 and 40 V.136 In non-aqueous
electrolytes, typically higher voltages are reported to grow tubes of a given
diameter. In KF and NaF aqueous electrolytes the maximum nanotube length
was few microns with a growth rate of approximately 0.25 mm h 1.185 However,
in non-aqueous electrolyte, the fabrication of TiO2 nanotube arrays up to
222 mm in length and with growth rates of up to 15 mm h 1 have been reported.
A variety of polar organic electrolytes have been used. The most common were
ethylene glycol (EG), formamide (FA), N-methylformamide (NMF), and
dimethyl sulfoxide (DMSO) in combination with HF, KF, NaF, NH4F, Bu4NF,
or BnMe3NF to provide uoride ions.186,187 Recent results reported a thickness
close to 1000 mm.124 Nanotube arrays greater than several tens of microns
in length (lm thickness) are mechanically robust to form of self-standing
membranes.
Usually the nanotube arrays have been made from a titanium thick lm or
foil, in which case the resulting nanotubes rest upon an underlying Ti
substrate as separated by a barrier layer. The nanotube arrays have also
been fabricated from a titanium thin lm sputtered onto a variety of
substrates, such as silicon and uorine doped tin oxide (FTO) coated
conductive glass. This extends the possibility for preparing technical
catalysts by deposing a thin Ti layer over a substrate (a foam, for example)
and then inducing the formation of the nanostructured titania lm by
anodic oxidation.188,189
In general, in order to produce very long titania nanotube arrays using
polar organic electrolytes it is necessary to minimize the water content to less
than 5%.179 With organic electrolytes the donation of oxygen is more dicult
in comparison to water and results in a reduced tendency to form oxides,190
while the reduction in water allows the formation of thinner or lower quality
barrier layers through which ionic transport may be enhanced. Grimes
et al.186 recently demonstrated how the electrolyte cation strongly inuences
both the nanotube growth rate and the resulting nanotube length, with the
length and aspect ratio of the nanotubes increasing with the cation size.
Under similar conditions, electrolytes containing a tetrabutylammonium
cation resulted in the longest nanotubes (about 95 mm), while the shortest
nanotubes (about 3 mm) were obtained when H+ ions were the sole cationic
species in the anodization electrolyte. The dierence in growth characteristics
was attributed to the inhibitory eect of the quaternary ammonium ions that
restrict the thickness of the interfacial (barrier) oxide layer; a thinner
interfacial oxide layer facilitates ionic transport thus enhancing nanotube
growth.
The tubes grown in ethylene glycol electrolytes show a hexagonal close-
packed structure191 similar to that observed in porous anodic alumina
layers. Several factors strongly inuence the degree of ordering: (i) the
anodization voltage (the highest possible voltage just below dielectric
breakdown is the most appropriate), and (ii) the purity of the material
(certain ordering faults can be eliminated by using a high purity Ti).
Furthermore, repeated anodization, as in the case of Al, can clearly improve
the ordering. By using this approach, the bottom imprints a rst tube layer

104 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
in the underneath Ti acting as pre-ordering guides for a subsequent
anodic tube initiation and growth.
Using this approach it is possible to prepare multi-layer stacks. First a layer
under a rst parameter set may lead to a rst geometry, then underneath a
second layer of tubes can be grown with a dierent parameter set. The
underneath layer may be initiated at the bottom of a tube (to break through
the bottom of a tube),192 or in the spaces between the tubes.193 Such
multilayer stacks may be quite interesting in developing novel catalysts.

Crystallinity of the lms and purity


As-formed TiO2 tubes typically have an amorphous structure, but by annealing
at temperatures of about 300 1C anatase structure can be detected, while at
higher temperature (about 450 1C) both anatase and rutile structures are
present.194196 Most recently, there are indications that already in the as-formed
tubes under certain conditions nano-crystallites can be present.136 Usually crystal
growth starts at the nanotube bottom via interface nucleation, due to the larger
space available for crystal growth than in the side wall. For certain tubes it was
found that essentially over the entire length only one plane(101)is present
along the walls.136 Annealing under oxygen-free conditions (e.g. in argon
atmosphere) leads to a blackening of the tube layers due to a signicant
reduction of the Ti(IV)-species in the oxide to Ti(III) and worsening of the
mechanical properties.194 Introduction of Nb197 or C198 into the TiO2 increases
the temperature of the anatase-to-rutile conversion, and shifts also the tempera-
ture of structural collapse to a higher value. These aspects are relevant for the
design of catalysts and indicates also the possibility to prepare nanostructured
mixed oxides. TiO2Nb2O5199 and TiO2-ZrO2200 mixed oxide nanotubes have
been prepared, but other mixed oxides are possible. Also this is an important
step forward in using these materials for catalytic applications.
A problem is related to the observation that some of the ions of the
electrolyte during the anodization enter into the tube structure at dierent
concentration levels. While ClO4 ions are hardly incorporated, SO42 and
particularly PO43 are incorporated into the entire tube to signicant levels
(some few atomic %).136 It is clear that signicant amounts of F (about
15 at.%) are entering the TiO2 structure. Recent TEM investigations
indicate that uorides indeed are accumulated (before annealing) at the
metal/oxide interface and to a certain extent between the individual tubes.
Annealing leads to almost complete loss of the uorides at around 300 1C195
and clearly the amount of surface hydroxides is reduced.201

Conclusions on the synthesis


These results, even if not covering systematically all aspects related to the
synthesis of nanostructured bidimensional titania lms, show that signicant
progress has been made in the last few years on this topic. Several parameters
have to be quite carefully controlled to have reproducible samples and the
uniformity on a large scale (above few cm2) is still an issue. Signicant
progress has also been made in terms of control of the thickness and array
characteristics. Less progress, has been made in the preparation of

Catalysis, 2009, 21, 82130 | 105


This journal is
c The Royal Society of Chemistry 2009
mixed-oxide nanostructured lms. Therefore, the major limits in the synthesis
of these materials to be used as catalysts have been removed, and interesting
perspectives exists on their use, as discussed in the next section, but still
further research is necessary to overcome the barriers in the applicability of
the methodologies for the development of practical industrial catalysts.

4. Uses, with focus on catalysis


The main advantages of regular tube arrays are the large surface area and
the dened geometry resulting in a narrow distribution of diusion paths
not only for entering the tubular depth (e.g., reactants to be transported to
the tube bottom), but also for species to be transported through the tube
wall, e.g., electrons, holes, ions. These aspects are all of particular interest
for developing catalysis, but as pointed out in the previous section, the main
obstacle for catalytic uses was the limited amount of TiO2 (of the order of
mg per cm2) and limitations in producing these materials in forms suitable
for catalytic uses. In fact, in the largest part of the cases, these materials
were produced as a thin nanostructured layer over a at surface. In the
previous section it was shown that signicant progress has been made to
remove these limitations, but it is not surprising that applications have been
focused up to now on aspects for which these limitations are less severe.
Applications of titania nanotube arrays have been focused up to now on
(i) photoelectrochemical and water photolysis properties, (ii) dye-sensitized
solar cells, (iii) photocatalysis, (iv) hydrogen sensing, self-cleaning sensors,
and biosensors, (v) materials for photo- and/or electro-chromic eects, and
(vi) materials for fabrication of Li-batteries and advanced membranes
and/or electrodes for fuel cells. A large part of recent developments in these
areas have been discussed in recent reviews.58,65,136,202 We focus here on the
use of these materials as catalysts, even though results are still limited, apart
from the use as photocatalysts for which more results are available.

Photocatalytic behaviour of TiO2 nanotube arrays


Photocatalytic activity of TiO2 nanotubes has been studied by several
groups.202210 In general, it was found that annealed TiO2 nanotubes show
higher photocatalytic eciency in the degradation of dyes or pollutants than
reference lms of TiO2 particles. The performance depends, however, on the
TiO2 characteristics and crystallinity. For example, Khan et al.210 reported that
crystalline TiO2 nanotubes prepared by hydrogen peroxide treatment of low
crystalline titania nanotubes show about 2-fold higher activity than those that
are non-crystalline, compared to TiO2-P25 (Degussa) in the photocatalytic
oxidation of trimethylamine gas under UV irradiation. Schmuki et al.204 also
showed that annealed TiO2 nanotubes have an activity about two times higher
than a compacted Degussa P25 layer in the degradation of Acid Orange 7 using
UV irradiation. This eciency can be even further accelerated by depositing
Ag nanoparticles on TiO2 nanotube arrays.136 However, it must be noted that
for photocatalytic conversion of pollutants, either in gas or liquid phase, the
results up to now are not still particularly relevant and no careful study was
made in terms of use of relevant pollutants, choice of the optimal reactions
conditions, and reference photocatalysts. In fact, it is known that depending on

106 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
the substrate, Degussa P25 is not always the best reference for photocatalytic
studies and a number of additional factors have to be considered in comparison
with the catalytic performances and techno-economical feasibility. In terms of
market perspectives, the use of nanoparticle coatings methodologies still
appears to be preferable for pollutant degradation applications, due to their
better adaptability and scaling-up of the methods.

New directions and possibilities


The use of nanostructured TiO2 lms formed by an array of nanotubes
(prepared by anodic oxidation) has still not provided convincing evidences
for superior performance with respect to the diculties in implementing this
new technology of preparation. However, recent progress in the preparation
and in particular the use of mixed oxide nanostructured lms or of hybrid
materials could open new perspectives. For example, the synthesis and
photocatalytic property of SnO2/TiO2 nanotubes composites has been
reported.211 The results showed that the SnO2/TiO2 nanotubes composite
photocatalyst with 5 wt% SnO2 loading had the highest photocatalytic
eciency in methylene blue degradation. The use of hybrid composites of
MWCNT (multiwalled carbon nanotube) (20% wt.) and TiO2 nanotubes
signicantly enhances the performance for photodegradation of humic
substances (HSs) in water, a relevant practical problem.212 It is known that
the combination of carbon nanotubes (CNTs) and TiO2 promotes the
performance of the latter, for example in the degradation of azo dyes,213
because the excited e in the conduction band of TiO2 may migrate into
CNTs, and be transported a long distance. Thus, the possibility of the
recombination of e /h+ pairs decreases. Meanwhile, O2 adsorbed on
the surface of CNTs may accept e and form O2 , which then leads to
the formation of hydroxyl radicals, promoting the rate of degradation of the
dyes. This is an example of the synergy possible in preparing hybrid
materials also in the case of titania nanostructured lms.
Recently, a two-dimensional TiO2/carbon nanowall composite material
was fabricated by growing carbon nanowalls on a Ti sheet with hot lament
chemical vapor deposition, followed by metal-organic chemical vapor deposi-
tion using titanium isopropoxide as TiO2 precursor and argon as carrier
gas.214 SEM showed that TiO2 was uniformly coated on the entire carbon
nanowall producing a TiO2/carbon nanowall composite. As a result of this
heterojunction, enhanced separation of photogenerated electrons and holes
was observed as well as a higher photocatalytic activity than TiO2 nanotubes
for the degradation of phenol under UV light irradiation. Yao et al.215
also demonstrated that TiO2/CNT nanotube composites show charge
recombination and enhance reactivity in phenol degradation. Alternative
methods to produce carbon-nanotube/TiO2 hybrid materials were recently
described by Eder and Windle,216 Gend et al.,217 Castro et al.218 and Byrappa
et al.218 It was also recently shown that carbon nanotubes synergistically
enhance the photocatalytic activity of TiO2 by a factor of about 400 in
photocatalytic H2 gas production from water/alcohol mixtures.219 These
materials show interesting properties also for nanocomposite electrodes for

Catalysis, 2009, 21, 82130 | 107


This journal is
c The Royal Society of Chemistry 2009
application to Li ion batteries220 and for novel electrodes for methanol
electrooxidation in fuel cells.221
An alternative interesting approach is that dened as tube-in-tube nano-
structures.222 By annealing TiO2 nanotube arrays in the presence of carbon
volatile precursors, carbon nanotubes were grown inside the inorganic
nanotubes. Compared with unmodied TiO2, a hybrid C-TiO2 photocatalyst
shows an enhanced eciency of photodecomposition of methyl orange due to
the increasing carrier rate and enhanced adsorption properties as well as the
unique nanostructure. Furthermore, the transition from anatase to rutile was
suppressed by carbon, resulting in a high content of the photoactive anatase,
which also benets the high catalytic activity of C-TiO2 photocatalyst.
Doping with small amounts of noble metals was often used to promote
the photobehavior. However, the localization of these metal nanoparticles
within the TiO2 nanotubes was recently shown to have an important role.
Nishijima et al.223 investigated site-selective deposition of Pt nanoparticles
on a titania nanotube (TNT). When Pt nanoparticles were deposited only
inside the TNT, active sites on the TNT were not covered by Pt nano-
particles, resulting in an increase in its photocatalytic activity for oxidation
of acetaldehyde.

Enhancing photobehavior with visible-light


A further signicant step could be the promotion of the activity with visible
light. Enhanced photocatalytic activity in composites of TiO2 nanotubes and
CdS nanoparticles has been reported.224,225 Bamboo-like CdS/TiO2 nanotube
composites show signicantly higher visible-light photocatalytic activity for
the degradation of methylene blue than pure TiO2 nanotubes or CdS
nanoparticles.226 The highest photodegradation eciency after 6 h irradiation
was 84.5%. CdS quantum dots (QDs) sensitized TiO2 nanotube-array
photoelectrodes were studied for their photoelectrochemical (PEC) perfor-
mance.227 The CdS QDs deposited in the pores of the TiO2 nanotube arrays
can increase the liquid junction PEC short-circuit photocurrent (from 0.22 to
7.82 mA/cm2) and increase cell eciency (up to 4.15%). These results
demonstrate that the unique nanotube structure can facilitate the propaga-
tion and kinetic separation of photogenerated charges. CdS/TiO2 nanotube
arrays used as photoanode in photoelectrocatalytic hydrogen generation
allowed a rate of hydrogen generation of 245.4 mL h 1  cm 2, which opens
new perspectives for photoelectrocatalytic hydrogen generation by using
CdS/TiO2 nanotube arrays.228
TiO2 nanotubes functionalized with PbS quantum dots showed also super-
ior photocatalytic behavior in comparison to Degussa P25 catalyst for the
degradation of organic dyes.229 The photoresponse of TiO2 nanoarrays has
been extended into the visible by attaching CdSe quantum dots.230 Upon
bandgap excitation, CdSe quantum dots inject electrons into TiO2 nano-
particles and nanotubes, thus enabling the generation of photocurrent in a
photoelectrochemical solar cell. The photosensitization of TiO2 nanoparticles
by quantized CdSe nanoparticles is dependent on the particles size, because
the rate of electron transfer is directly inuenced. A 3-order enhancement in
the charge injection rate was achieved by decreasing the CdSe particle

108 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
diameter from 8 nm to 2.4 nm.231 The TiO2-CdSe composite nanostructures
show a maximum IPCE (photon-to-charge carrier generation eciency) for
3 nm diameter CdSe nanoparticles of 35% using TiO2 nanoparticles and 45%
for a tubular TiO2 morphology, demonstrating the role of TiO2 nano-
structure. The photo-response of a CdSe sensitized TiO2 nanotube array is
strongly dependent on the CdSe QDs size.231,232
The performance of these materials could be further enhanced by
modication of CdS/TiO2 quantum-dot sensitized solar cells by using
single-walled carbon nanotubes (SWCNT) on indium-doped tin oxide
(ITO) electrodes.233 The presence of the single-walled carbon nanotube
layers on an ITO electrode was found to increase the short-circuit current
under irradiation and also reduce the charge recombination process under
dark conditions. The power conversion eciency of CdS/TiO2 on ITO
increased 50.0% in the presence of single-walled carbon nanotubes due to
the improved charge-collecting eciency and reduced recombination.
Kamat et al.234 also proposed an advanced architecture for light energy
conversion based on the combination of a SWCNT network and TiO2
nanotube arrays on electrode surfaces. Upon band gap excitation of the
SWCNT-TiO2 composite the electrons are quickly transferred from semi-
conductor particles into carbon nanotubes as these two systems undergo
charge equilibration. The carbon nanotube network plays an important role
in facilitating charge collection and charge transport to the collecting
electrode surface. A two-fold increase of photoconversion eciency (IPCE)
has been observed using such composite architecture.
These results indicate signicant progress in the design of nanostructure
hybrid architectures for next generation photomaterials.235 The photoactive
semiconductor nanoparticles (TiO2 or ZnO) when dispersed on conducting
scaolds of SWCNT exhibit improved photoconversion eciency (IPCE) as the
carbon network facilitates charge collection and charge transport in semi-
conductor nanostructures. The photoresponse of TiO2 nanoarrays can be
extended into the visible by attaching CdSe quantum dots. It has been indicated
that this advanced nanoarchitecture is well suited for photocatalytic applica-
tions, particularly in the area of solar hydrogen production.235 However, the
instability of CdS or CdSe nanoparticles in an oxidation environment is an issue
and thus stability of operations of photoelectrochemical cells to generate
hydrogen from aqueous electrolytes has still to be solved. Similar doubts exist
in relation to the use of these materials as photocatalysts under practical
reaction conditions.

Modication of TiO2 band gap by doping


An alternative to the use of semiconductor quantum dots to improve the
behavior of TiO2 nanotube arrays with visible light is to modify by doping
the TiO2 band gap towards the visible region. Signicant activity has been
made recently to produce doped titania nanotubes. Annealing titanium
metal foils and titanium oxide lms in a hydrocarbon ame forms carbon-
doped titania with a signicantly enhanced full spectrum photoresponse. On
annealing Ti metal foils in a natural gas ame at 850 1C, Khan et al.236
found the diuse reectance spectra of these samples to be signicantly

Catalysis, 2009, 21, 82130 | 109


This journal is
c The Royal Society of Chemistry 2009
broadened; in addition to a shift in the primary absorption threshold from
414 to 440 nm, a second optical absorption threshold appeared at 535 nm,
which was used to extract a band gap of 2.32 eV. In applying the method to
nanotube array samples, carefully controlled conditions have to be used to
avoid destruction of the nanostructure.65 Care should be also made in
considering that the introduction of C in TiO2 also creates a number of
defects in the material which increase the holes/electron recombination rate
and, when on the surface, quench the photoactivity.
Geng et al.217 reported that carbon-modied TiO2 nanotubes prepared by
wet chemical methods exhibited higher photocatalytic activity for the
degradation of methylene blue than unmodied titanate nanotubes under
articial solar light. Mohapatra et al.237 reported the use of C-doped TiO2
nanotube arrays as the photoanode and Pt nanoparticles incorporated in
TiO2 nanotube arrays as the cathode in a photoelectrochemical (PEC) cell.
The PEC cell is highly ecient (i.e., gives high photocurrent at a low
external bias, e.g. about 2.52.8 mA cm 2 at 0.4 V vs. Ag/AgCl) and was
tested for 80 h in H2 generation by H2O splitting under solar-simulated
conditions. As-synthesized TiO2 nanotubes (self-organized hexagonally-
ordered TiO2 nanotubes produced by anodization) are annealed under a
reducing atmosphere (H2), which converts the amorphous nanotube arrays
to the photoactive anatase phase and favors the doping by C (carbon
derives from the reduction of adsorbed ethylene glycol) to give the TiO2-C
type photoanode. The Pt nanoparticle cathode was prepared by reduction
to Pt0 of a Pt salt which was added to the TiO2 nano-tubular arrays by
incipient impregnation wetness method. Xu et al.238 reported that C-doped
TiO2 nanotubes (annealed in air and natural gas ame) showed a 2 fold
increase in photocurrent with respect to an undoped TiO2 nanotube lm.
The band gap of TiO2 was reduced to 2.84 eV and an additional intra-gap
band was introduced at 1.30 eV above the valence band, both contributing
to extend activity in the visible light regions. Carbon-doped TiO2 nanotubes
show also improved performance, when used as an anode material for
lithium-ion batteries.239
Interesting results have been thus reported recently on C-doped TiO2
nanotubes, but it should be reported that doubts exist on the results.
Murphy240 recently indicated that the results reported on the water splitting
by carbon-doped titanium dioxide photoelectrodes under visible illumina-
tion are due to an artifact in the measurements. He also pointed out that the
mechanism proposed for water splitting under visible illumination has no
physical basis and the photocurrents results are inconsistent with the IPCE
(Incident Photon to Current Eciency) data. Murphy240 thus concluded
that there is no convincing evidence in the literature of eective water
splitting under visible light in carbon-doped TiO2.
Doping TiO2 with nitrogen is another area under active development.
Asahi et al.241 reported that doping TiO2 with nitrogen by sputtering in a
nitrogen containing gas mixture improves the photoelectrochemical reactivity
of TiO2 lms towards organic molecules under visible light illumination.
Chen et al.242 reported superior photocatalytic performance of N-doped TiO2
nanotubes in the degradation of methyl orange under UV irradiation.
Nitrogen-doped TiO2 nanotube thin lms were synthesized using ZnO

110 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
nanorods as the template and doped with urea at 623 K.243 Under UV and
visible light irradiation, the eciencies for photocatalytic degradation of
methylene blue was about 30%, 10% of toluene (after 120 min) and less than
8% for 1-hexadecene. The latter two molecules were considered as model
molecules for using these materials in the photocatalytic degradation of spill
oils. Recently the possibility to produce N-doped TiO2 nanotubes by
anodization of a TiN alloy was also shown.244 It should be also remarked
that a nitrogen-doped titania nanotube array vertically aligned on a titanium
substrate exhibits ecient electron eld emission.245 Such a titania nanotube
array shows very good stability at high eld emission current (uctuation
o3% at eld emission current of 160 mA within 4 h) and low turn-on and
threshold elds (11.2 and 24.4 V mm 1, respectively), because of the
coexistence of doped nitrogen and concomitant oxygen vacancies in titania
nanotubes. This work demonstrates the possibility of converting pure titania
nanotubes without eld emission into ecient materials through the intro-
duction of acceptor and donor states above the valence band maximum and
below the conduction band minimum (in the band gap of titania), respec-
tively. These acceptor and donor states are associated with nitrogen and
concomitant oxygen vacancies, respectively. Recent results246 indicate that N
atoms incorporated into anatase TiO2 are present in the form of radical NO
molecules coupled to an oxygen vacancy. A relation was observed between
VoNOTi concentration (Vo indicates single-electron-trapped oxygen
vacancies) and visible light photoactivity. The highest photoactivity was
obtained for the catalyst treated with NH3 at 600 1C. At higher temperature
(700 1C) the formation of TiN was observed leading to a decrease in the
photoactivity.
As a general comment regarding N-doped TiO2, it should be noted that
after the paper of Asahi241 an intense activity was made worldwide on
N-doped TiO2 materials, but also several questions were raised on the
eective reliability of the results regarding photoactivity with visible-light.
There is instead good evidence that N is incorporated in TiO2 and modies
the properties of titania.
Other doping species such as transition metals,247,248 or non-metals like
phosphorus,249 sulphur,250 boron251 have been introduced into TiO2 compact
layers or powders. After several studies,252 it was recognized that ion
implantation of transition metals is not a suitable method and may produce
also structural damage,253 even if ion implantation with Cr254 improve the
visible light response of TiO2 nanotube arrays. Cr may be introduced by
hydrothermal ion-intercalation.255 Samples of this material were used to
investigate the photoelectrochemical H2O splitting to H2 and O2 under visible
light. The photoelectrochemical activity of Cr-doped TiO2 was higher than
that of the undoped sample and the performance depends on the amount of
chromium doping. A red shift in the band gap was induced by Cr doping of
TiO2. At high Cr concentrations, however, the formation of a secondary
Cr-oxide phase decreases the photoelectrochemical activity. The photo-
catalytic activity of these materials was also reported to depend on the Cr
doping.256
Chemical vapor deposition (CVD) was applied to introduce boron into
TiO2 nanotubes arrays,257 but alteration of the morphology was observed.

Catalysis, 2009, 21, 82130 | 111


This journal is
c The Royal Society of Chemistry 2009
However, B-doped samples displayed stronger absorption in both UV and
visible range and higher photoelectrocatalytic eciency in methyl orange
degradation. Highly ordered boron-doped TiO2 nanotube arrays were
fabricated also by electrodeposition.258 The rate of the photoelectrocatalytic
conversion of phenol under simulated solar irradiation was found to be
about 28% higher using boron-doped TiO2 nanotube arrays with respect to
the corresponding undoped sample. The combination of doping with B of
TiO2 with a coating with carbon in order to improve the performance
(see previous comments) was also reported recently.259 These materials,
indicated as TiO2-B@C Core-Shell nanoribbons, showed interesting
electrochemical intercalation properties.
Sulfur-doped highly ordered TiO2 nanotube arrays, produced by annealing
in a ow of H2S at 380 1C, were also reported to show higher photobehavior
(photocurrent) under visible light irradiation up to 650 nm.260
It should be mentioned also that an interesting opportunity is that
the dopant introduction can occurs also via modication of the
anodization bath, e.g. during the preparation of the TiO2 nanostructured
lm itself.65
Another direction of investigation is to extend the nanotube array
architecture to other metal oxides, most noticeably a-Fe2O3261 and mixed
FeTi oxides,58 to develop materials capable of eciently responding to the
visible light spectrum, while maintaining the outstanding charge transport
properties demonstrated by the TiO2 nanotube arrays.
In conclusion, doping TiO2 nanotube arrays in order to improve photo-
behavior with visible-light is an exciting and challenging direction. Several
promising results have been reported, but sometimes not with the necessary
scientic precautions for reliable data.

Use in photoelectrochemical devices for H2 generation and pollutants


conversion
Photoelectrochemical application of titania nanotube photoanodes is one of
the most relevant elds of development, particularly for H2 photogeneration,
but also for the elimination of pollutants, particularly of low bio-reactive
compounds (indicated often in literature as recalcitrant molecules).262266 The
titania nanotube array architecture results in a large eective surface area in
close proximity with the electrolyte thus enabling diusive transport of
photogenerated holes to oxidizable species in the electrolyte. Separation of
photogenerated charges is assisted by action of the electric eld in the depletion
region.267 In addition, due to light scattering within a porous structure, the
incident photons are more eectively absorbed than on a at electrode.268
A maximum conversion eciency of 6.8% was obtained at a temperature of
5 1C and a potential (versus reference Ag/AgCl electrode) of about 0.4 V
(under 320400 nm, 100 mW/cm2 illumination; 1M KOH electrolyte). The
hydrogen generation rate was 960 mmol/h.W, or 24 mL/h.W.65,269 These rates
and eciency are among the highest reported, but it should be observed that
the contribution from photoelectrolysis was not considered. In addition, H2
photogeneration starts only after 400 s, indicating that a transformation of the
TiO2 photoanode probably occurs in the strong basic medium. After the

112 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
induction time, the amount of photogenerated H2 increases nearly linear, even
if data for times longer than 2000 s were not given.
Using UV (320400 nm) illumination the photocurrent increases with
the annealing temperature up to 675 1C, after which it reduces with
samples annealed at 700 1C due to disruption of the nanotube array
architecture. For a sample annealed in the range 580620 1C range the
eciency in H2 generation was about 12.2%. The increase in photo-
current and eciency are due to the increased crystallinity of the nanotube
walls, with the reduction of the amorphous regions and grain boundaries
which reduce the number of charge carrier recombination centers. In
addition, at temperatures near 675 1C the densication of the bottom
part of the nanotubes starts isolating the undestroyed nanotubes from the
metal electrode, reducing the number of charge carriers which reach the
electrode.
Both photocurrent magnitude and photoconversion eciency were
observed to increase with the nanotube array lm thickness.270 Fig. 9
evidences that a linear relationship was observed between lm thickness
(in the 26 micron range) and rate of H2 photo-generation.270
For a 6 mm thickness of the nanotube array lm annealed at 600 1C the
quantum eciency was calculated to be 81% and 80%, for wavelengths of
337 nm (3.1 mW cm 2) and 365 nm (89 mW cm 2), respectively.65 The high
quantum eciency clearly indicates that the incident light is eectively
utilized by the nanotube arrays for charge carrier generation.
Using CdS/TiO2 nanotube array photoanodes (diameter of the nano-
tube of 80100 nm and length about 550 nm; CdS nano-particles deposited
by chemical deposition in an ammonia-thiourea system) the rate of
photoelectrocatalytic hydrogen generation was 245.4 mL h 1  cm 2.228
Chromium-doped titanium dioxide thin-lm photoanodes were also
observed to have a high activity in visible-light-induced water cleavage.255

Fig. 9 Rate of hydrogen generation from nanotube arrays lms of dierent lengths annealed
at 530 1C. Electrode area of 1 cm2; 100 mW/cm2 visible light. In the inset FESEM cross-
sectional image of 2.8 mm long TiO2 nanotube array prepared by anodic oxidation of a titanium
foil in an electrolyte containing potassium uoride (KF; 0.1 M), sodium hydrogen sulfate
(1 M), trisodium citrate (0.2 M) and sodium hydroxide. Elaborated from Grimes et al.270

Catalysis, 2009, 21, 82130 | 113


This journal is
c The Royal Society of Chemistry 2009
In a uoride-free HCl electrolytes containing H2O2 highly ordered
nanotube arrays up to 860 nm in length, 15 nm inner pore diameter, and
10 nm wall thickness can be obtained for one hour anodizations using a
0.5 M HCl aqueous electrolyte containing 0.10.5 M H2O2 (anodization
potentials between 1023 V).271 The use of ethylene glycol as the electrolyte
medium signicantly alters the anodization kinetics and resulting lm
morphologies. For anodization potentials between 8 V and 18 V only a
few minutes are needed in order to obtain nanotube of several microns in
length. The nanotube arrays obtained from the ethylene glycol electrolytes
show relatively higher photocurrents, E0.8 mA cm 2. Under 100 mW cm 2
illumination a 500 1C annealed nanotube array sample (one cm2), obtained
by anodization of a Ti foil sample in ethylene glycol+0.5 M HCl+0.4 M
H2O2 electrolyte, demonstrates a hydrogen evolution rate of approximately
391 mL h 1 by water photoelectrolysis. Time-power normalized evolution
rate was 3.9 mL W 1 h 1, with the water splitting conrmed by the 2:1 ratio
of evolved hydrogen to oxygen.
Grimes et al.272 recently reported also the fabrication of highly-ordered
TiO2 nanotube up to 220 mm in length, with a length-to-outer diameter
aspect ratio of E1400, for hydrogen production by water photoelectrolysis.
These highly-ordered TiO2 nanotube arrays were fabricated by potentio-
static anodization of Ti foil in uoride ion containing baths in combination
with non-aqueous organic polar electrolytes including N-methylformamide,
DMSO, formamide, or ethylene glycol. Depending upon the anodization
voltage, the inner pore diameters of the resulting nanotube arrays range
from 20 to 150 nm.
A novel method for the synthesis of titania nanotubes using a sono-
electrochemical method and its application for photoelectrochemical splitting
of water was reported by Mohapatra.273 Self-ordered arrays of TiO2 nano-
tubes of 30100 nm in diameter and 3001000 nm in length (thus relatively
thin lms) can be rapidly synthesized under an applied potential of 520 V.
The rate of formation of the TiO2 nanotubes by the sonoelectrochemical
method was found to be almost twice faster than the magnetic stirring
method. It was also demonstrated that high-quality nanotubes can be
prepared using high viscous solvents like ethylene glycol under ultrasonic
treatment. The TiO2 nanotubes prepared in the organic electrolytes (ethylene
glycol) are then annealed under H2 atmosphere to give TiO2-C type material
having a band gap of around 2.0 eV. This process was found to be highly
ecient for incorporating carbon into TiO2 nanotubes. The photoelectro-
catalytic activity of these materials to generate H2 by water splitting was
found to be promising at 0.2 V vs. Ag/AgCl.
As commented before, also in this case it is more correct to indicate
photoelectrolysis instead that photocatalytic splitting. It is not only a
semantic question, because the chemical potential created by the dierent
solutions at the anode and cathode side (strong acid for H2 generation and
strong basic for O2 generation) has signicant consequences and the
reaction mechanism is dierent.
The same authors237 also discussed the optimal design of a photo-
electrochemical (PEC) cell using C-doped TiO2 nanotube arrays as the
photoanode and Pt nanoparticles incorporated in TiO2 nanotube arrays as

114 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
the cathode. They also reported274 that TiO2 nanotubes prepared by
the sonoelectrochemical anodization method can be functionalized with
2,6-dihydroxyanthraquinone (anthrafavic acid). The functionalization takes
place by chemical condensation of the Ti-OH hydroxyl groups present on
the TiO2 nanotube surface with the phenolic hydroxyl groups of anthrafavic
acid forming an inorganic-organic hybrid material. The condensation
results in an intramolecular ligand-to-metal charge-transfer transition that
leads to an enhanced absorption band in the visible region and an increase
in the photoelectrochemical generation of hydrogen from H2O up to 30%.
It should be nally cited that H2 could be produced photochemically from
these materials also from diluted bioresources. Lin et al.275,276 reported that
1% Pt-doped TiO2 nanotubes produce 20% more H2 than 1% Pt/Degussa
P-25 TiO2 in the photocatalytic dehydrogenation of ethanol after 2 h of UV
light irradiation.
We may conclude that signicant progress has been made on the photo-
electrocatalytic generation of H2, although it is not always possible to
clearly identify when results due to are photoelectrolysis or photocatalytic
splitting. The strong basic medium needed for photoelectrolysis further
stresses the need to evaluate stability in long-term operations.
Several attempts have also been made to promote the activity with visible
light, noting that care should be taken to verify that they are not artefacts.
Also for these materials stability is an issue. The availability of novel
methods to prepare nanostructured thin lms with several hundred micron
thickness is highly promising, because as shown in Fig. 9, the performance
depends strongly on this factor, even if the optimal thickness still has to be
established.
This is a highly dynamic area of research and the cited results are all quite
recent. There is thus the need to conrm the results and have more extended
verication, as well as a clear identication of the limits of the technology.
However, it may be said that this is a probable breakthrough area for the
production of renowable H2.

Use in catalytic and electrocatalytic reactions


The low amount of titania per cm2 of geometrical surface area of the sample
has limited the use of TiO2 array thin lm materials directly as catalysts,
even if the cited progress in producing thicker lms has overcome this
problem and thus a higher number of applications is expected in the near
future. Up to now the results should be thus considered mainly as
exploratory studies more than as real examples of advantages of these
materials. Some of the studies, however, can be seen in preparation of future
studies, and we will thus briey comment also on some selected results to
show how to use the TiO2 array thin lm material as substrate to depositing
other active elements, when the peculiar nanostructure could provide
advantages with respect to the use of TiO2 nanoparticles.
A characteristic of TiO2 nanotube arrays is the stabilization of oxygen
vacancies due to the peculiar nanostructure, consistent with the fact that
they show a dierent oxygen adsorption.150 This feature is clearly relevant
for catalysis. On TiO2 nanotubes array which consist of a 80/20% mixture

Catalysis, 2009, 21, 82130 | 115


This journal is
c The Royal Society of Chemistry 2009
of the anatase/rutile, molecular O2 adsorption, inuenced by defects, as
well as CO2 formation at low temperatures (approximately 100 K) were
observed.277 Supporting copper oxide over TiO2 nanotubes resulted also in
highly active catalysts for CO oxidation.278 The synthesis of gold-, gold and
copper-doped TiO2 nanotubes by impregnation-reduction method and their
behavior in CO oxidation was also analyzed.279,280
The COH2 and COH2O reactions over TiO2 nanotubes lled with Pt
metal nanoparticles was investigated by Sato et al.281 Pt-TiO2 nanotube
catalysts were prepared by employing ne ber shaped crystals of
[Pt(NH3)4](HCO3)2 complex as a structure templating material. The turn-
over frequencies (TOF) of these nanotube catalysts were more than one
order of magnitude higher than those of the conventional impregnated
Pt/TiO2 catalysts, and the selectivity for methanol in COH2 reaction was
extraordinarily high compared to the impregnated catalysts. The XPS and
XRD analyses of the nanotubes revealed characteristic electronic state of
reduced TiO2 (Ti3+ in rutile structure) with zerovalent Pt even after the
calcination at 773 K. In WGS reaction, electron rich Ti3+ on the nanotube
walls may play an important role to activate water molecules for the
oxidation of CO. In COH2 reaction, a similar promotion eect of Ti3+
species may operate for selective methanol formation by supplying active
OH(a). These results are not only interesting for the development of new
materials for methanol or WGS reactions, but also for the design of new
electrocatalysts for methanol and ethanol conversion, a major eld of
interest discussed later.
Besides CO oxidation, few other uses of TiO2 nanotube arrays in oxidation
reactions have been reported. Pd/titanium dioxide nanotubes catalysts were
investigated for hydrazine oxidation.282 Compared to pure Pd particles and
Pd/TiO2 particles, Pd/TiO2 nanotube catalyst show a much higher electro-
chemical activity. WOx/TiO2 nanotubes catalysts were investigated for the
oxidation of dibenzothiophene,45 although the annealing at 500 1C of these
samples destroyes the nanotube structure and yields anatase nanoparticles
decorated by tungsten nanoparticles on its surface. Ruthenium(III) hydrated
oxide supported on TiO2 nanotube was investigated in the selective oxidation
of alcohols by oxygen.283 A turnover frequency of 450 h 1 was found which
was higher than that reported for a Ru(III)/Al2O3 catalyst (TOF = 335 h 1).
The ruthenium loading could exceed 8 wt% without loss of specic catalytic
activity.
Copper-oxide supported on TiO2 nanotubes was also investigated for
selective NO reduction with NH3.284 The nanotubes were found to be more
thermally stable than nanoparticles and the Cu species supported on the
nanotubes showed a higher catalytic activity than those supported on the
nanoparticles.
A further peculiar characteristic of TiO2 nanotube arrays is the high H2
adsorption and mobility of H-species, which was successfully applied to
develop highly sensitive sensors for H2.58,202 These properties are of interest
for developing hydrogenation catalysts. Coville et al.285 studied the selective
hydrogenation of o-chloronitrobenzene (o-CNB) over palladium supported
nanotube titanium dioxide catalysts which gave complete conversion
(100%) of o-CNB with the selectivity to ortho-chloroaniline of 86%.

116 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
A strong palladium-support interaction was found to be crucial to the
overall activity of the catalyst.
A major area of investigation of TiO2 nanotube based catalysts is the
development of novel electrocatalysts for fuel cells, which take advantage of
both the peculiar nanostructure and the high transport characteristics of the
nanostructured lms, together with the possibility of stabilization of metal
particles against sintering. There are thus an increasing number of publica-
tions on the methanol and ethanol electro-oxidation over these materials.
Schmuki et al.286 rst explored the oxidative electrocatalytic properties of
a system consisting of bimetallic Pt/Ru nanoparticles dispersed over a
nanotube self-organized TiO2 matrix. The nanotube TiO2 layers consist of
individual tubes of 100 nm diameter, 500 nm length and 15 nm wall
thickness. This nanotube TiO2 support provides a high surface area and
the electrocatalytic activity of Pt/Ru for MeOH oxidation is signicantly
enhanced (relative to the performance of Pt/Ru at the same loading, but
immobilized on a conventional compact TiO2 support). After annealing to
form crystalline anatase, the TiO2 nanotube support exhibits a higher
MeOH electrooxidation activity than when used in the as-formed amor-
phous structure. The overall electrocatalytic activity of the system can be
further increased by illumination with UV-light (wavelength 325 nm),
demonstrating the concept reported in Section 2 that considerable oppor-
tunities are oered by investigating how the catalytic behavior of
TiO2-based catalysts could be promoted by electrons and holes generated
by light irradiation.
Hu et al.287 investigated Pd electrocatalyst supported on carbonized TiO2
nanotubes for ethanol oxidation in alkaline media. The conducting treatment
of a TiO2 nanotube was performed by carbonization using an organic
polymer. The Pd/TiO2-C electrocatalyst with a 1:1 mass ratio of Pd to
TiO2-C gives the best performance. The Pd/TiO2-C electrocatalyst was found
not only superior in activity for EtOH oxidation, but also more stable during
constant current polarization in alkaline media in comparison with a Pd/C
electrocatalyst. The behavior of similar catalysts in the ethanol electro-
oxidation in acidic media was studied by Song et al.288 The results demon-
strated that the titania nanotubes can greatly enhance the catalytic activity of
Pt for ethanol oxidization and increase the utilization rate of platinum. CO
stripping tests showed that the titania nanotube can shift the CO oxidation
potential to lower values with respect to TiO2. This was indicated as helpful
factor for increasing the rate of ethanol oxidation. Similar Pt-TiO2/CNTs,
e.g. a combination of titania and carbon nanotubes, were used in the direct
ethanol fuel cells (DEFCs). Pt-TiO2/CNT electrodes have higher electro-
catalytic activity and CO-tolerance for ethanol oxidation than Pt/C (20 wt%
Pt, E-TEK) and Pt/CNTs catalyst in acid solution.289 The Pt/TiO2 optimal
molar ratio was found to be equal to 1:1.
The catalytic properties of Pt and Ru particles on TiO2-CNT catalysts in
the methanol electro-oxidation were investigated by He et al.290 Pt and Ru
nanoparticles, approximately of 3 nm in diameter, were uniformly electro-
deposited on the as synthesized TiO2-supported C nanotubes. An enhanced
and stable catalytic activity was obtained in the electro-oxidation of
methanol due to the uniformly dispersed Pt and Ru nanoparticles on the

Catalysis, 2009, 21, 82130 | 117


This journal is
c The Royal Society of Chemistry 2009
nanotube array which facilitates the electron transfer. The optimal
temperature of treatment of these materials was about 500 1C.291 The
electro-oxidation of methanol on TiO2 nanotube supported platinum
electrodes was studied also by Maiyalagan et al.292
Hepel et al.293 showed, by quantum mechanical calculations of supported
clusters (in terms of catalyst-support interaction) and of the surface
diusivity of adsorbed intermediates, new aspects of MeOH oxidation
which could explain why bimetallic particles on TiO2 nanotubes show
enhanced properties. They found a strong enhancement of the ligand eect
for very small bimetallic catalysts attributed to a decreased local density of
states near the Fermi level of Pt atoms neighboring the Ru metal. This
enhancement resulted in a decreased barrier for surface diusion of
adsorbed CO. The lattice relaxation and strong ligand eects in small
supported particles lead to lower adsorption energy for COad and thus, to
higher reactivity and mobility of reactants and intermediates. Using as
model a TiO2 nano-ring (representing the top open part of a TiO2 nanotube)
unusually strong electron delocalization eects for Pt2Fe2 clusters were
estimated.
It must be nally shown that Song et al.294 reported that the structural
water in TiO2 nanotubes co-catalyzes ethanol oxidation and improves the
tolerance to poisoning, but the mechanism is not fully clear.
Ordered TiO2 nanotubes were also used to develop ecient electro-
catalysts for O2 reduction, e.g. the cathode side instead of the anode side
as above. Schmuki et al.295 reported that Au nanoparticles dispersed over a
self-organized nanotube TiO2 matrix can be used as a highly ecient
catalyst for the electrochemical oxygen reduction reaction in aqueous
solutions. For the same loading of Au nanoparticles, the nanotube support
provides a large increase in the reaction rate in comparison with a at TiO2
support, or a pure Au sheet electrode. CoTMPP-TiO2NT composite
electrocatalyst (Co-tetramethoxyphenylporphirin on C-supported TiO2
nanotubes) shows a higher catalytic activity and better stability than
CoTMPP supported directly on carbon (Black Pearls) for oxygen reduction
reaction.296
In conclusion, although results on the use of TiO2 nanotube array
catalysts or electrocatalysts are still limited, the future is promising. In
several cases, unusual behavior was demonstrated and associated with the
specic characteristics of the nanostructure.
Various results have been also published indicating that metal or metal
oxides supported on TiO2 nanotubes have dierent characteristics than
when supported over conventional TiO2 particles or other type of supports.
We briey mention here some selected examples, because in principle
dierent characteristics indicate also possible dierences in the catalytic
behavior.
TiO2 nanotubes were used to support MoO3 observing a spontaneous
dispersion of molybdenum-oxide on the surface of nanotubes, which was
dierent from that observed on titania particles.297 Supporting tungsten
oxides a preferential orientation of the (002) planes was observed.298
Vanadium-oxide in the form of nanorods could be prepared using the
titania nanotube as structure-directing template under hydrothermal

118 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
conditions.299 Ni nanocylinders can be grown in titania nanotubes.300
Intergrown binary metal oxide nanotubes could be also produced. Self-
organized porous TiO2-ZrO2301 and TiO2Nb2O5199 could be produced by
anodization.
Finally, it should be cited that the particular nanostructure of TiO2
nanotube arrays is well suited for hosting heteropoly compounds to be
used either to produce advanced electrodes or catalysts.302,303
Deposition of metals may lead to well dispersed metal nanoparticles, as
discussed in the previous section, but also to special metal structures. Using
a TiO2 nanotube array prepared by anodic oxidation as a template and
electrodepositing Au onto the template, Au nanonets could be prepared.304
The TiO2 nanotube arrays are also well suited for hosting and stabilizing
bio-catalysts. Hemoglobin could be immobilized in TiO2 nanotube
lms and be used as bio-catalyst.305 It was also observed that adhesion,
spreading, growth, and dierentiation of mesenchymal stem cells are
critically dependent on the tube diameter.306

Conclusions
Research in the eld of TiO2 nanostructured lms has progressed signicantly
over the last two years. This review integrates and extends to bidimensional
nanostructured systems (e.g. array of 1D nanostructures) the analysis
reported in part 160 on quasi-1D TiO2 nanostructures, e.g. nanorods,
nanowires and nanobres, nanotubes and nanopillars. However, it must be
commented that in the last two years, e.g. approximately the time spam from
the preparation of part 1,60 the progress in this eld has been so rapid that
several of the doubts presented in the earlier review have been in large part
solved. The aim of these two reviews is to highlight the potential and issues in
using these materials for catalysis, in particular the opportunities due to the
nanostructure. Still the number of examples of application of these materials
as solid catalysts in gas or liquid phase reactions is limited, but it was shown
that several of the barriers hindering their larger use have been solved or
nearly solved. We thus expect a large growth of examples and probably
applications in the near future.
Better established is their use as photocatalysts, in the photoelectro-
catalytic production of H2 or the elimination of pollutants, and in devel-
oping advanced electrodes for fuel cells, particularly for direct methanol or
ethanol oxidation. Nevertheless, also in this case the eld can be still
considered to be at an earlier stage. It has been shown how several of the
results have to be further demonstrated, and issues and limits better dened.
However, there are clear indications that this will be a major area of
research not only for this specic eld, but in general for all catalysis.
The recent US DoE report Catalysis for Energy also indicates that the
development of better tailored nanostructures for photo- and electro-
catalytic applications, particularly for better use of renewable resources, is
one of the priority areas of research in catalysis and in general of science.
We have limited the discussion here to TiO2 materials, but various
examples have been shown to indicate that both TiO2 nanotubes are a well
suited support for other catalytic elements, including bio-catalysts, and that

Catalysis, 2009, 21, 82130 | 119


This journal is
c The Royal Society of Chemistry 2009
mixed-oxides with TiO2 or other oxides, but with analogous nanostructures,
could be produced. Therefore, although focused on titania, the discussion
presented here, particularly in the second section, on the possibilities oered
by ordered 1D nanostructures for catalysis has a larger validity. This is thus
a fascinating area of research which has been driven up to now mainly by
experts in the area of synthesis of nanomaterials and electrochemistry. It is
time for experts in catalysis to use these knowledge to develop advanced
catalysts for tailored behaviour which can address some of the challenges
for catalysis, such as develop ideally 100% selective catalysts, use
more eciently renewable resources, better integrate homo-, hetero- and
bio-catalysis to realize new sustainable chemical processes, and develop
light-tunable catalysts. We have shown in this review that in principle these
challenges could be all addressed by this new class of materials which
further highlights the interest in them.

Acknowledgements
This work was realized in the frame of the EU Network of Excellence
IDECAT (Integrated Design of Catalytic Nanomaterials for a Sustainable
Production; NMP3-CT-2005-011730) and of the EU project NATAMA
(NMP3-CT-2006-032583) which are gratefully acknowledged.

References
1 Metal Oxide Catalysis, eds. S. D. Jackson and J. S. J. Hargreaves, Wiley VCH,
Weinheim, Germany, 2008, vols. 1 and 2.
2 Metal Oxides. Chemistry and Applications, ed. J. L. G. Fierro, CRC Taylor &
Francis, Boca Raton, US, 2006.
3 G. Centi, F. Cavani and F. Triro`, Selective Oxidation by Heterogeneous
Catalysis. Recent Developments, Kluwer/Plenum Publishing Corporation,
New York & London, 2001.
4 G. Centi and S. Perathoner, Curr. Opin. Solid State Mater. Sci., 1999, 4, 74.
5 Y. Ono, Catal. Today, 2003, 81, 3.
6 M. Misono, Catal. Today, 2005, 100, 95.
7 R. Schlogl, A. Knop-Gericke, M. Havecker, U. Wild, D. Frickel, T. Ressler,
R. E. Jentoft, J. Wienold, G. Mestl, A. Blume, O. Timpe and Y. Uchida, Top.
Catal., 2001, 15, 219.
8 B. K. Hodnett, Heterogeneous Catalytic Oxidation, J. Wiley & Sons,
New York, US, 2000.
9 P. Arpentinier, F. Cavani and F. Triro`, The Technology of Catalytic
Oxidations, Editions TECHNIP, Paris, France, 2001.
10 G. Centi and S. Perathoner, Catal. Today, 2003, 7980, 3.
11 H.-J. Freund, H. Kuhlenbeck, J. Libuda, G. Rupprechter, M. Baumer and
H. Hamann, Topics in Catal., 2001, 15, 201.
12 T. Ressler, B. L. Kniep, I. Kasatkin and Robert Schlogl, Angew. Chem., Int.
Ed., 2005, 44, 4704.
13 W. Weiss. R. Schlogl, Top. Catal., 2000, 13, 75.
14 D. W. Goodman, Chem. Rev., 1995, 95, 523.
15 C. R. Henry, in Catalysis and Electrocatalysis at Nanoparticle Surfaces,
eds. A. Wieckowski, E. R. Savinova and C. G. Vayenas, 2003, p. 239.
16 M. A. Barteau, J. E. Lyons and I. K. Song, J. Catal., 2003, 216, 236.
17 A. K. Datye, Top. Catal., 2000, 13, 131.

120 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
18 C. H. F. Peden, G. S. Herman, I. Z. Ismagilov, B. D. Kay, M. A. Henderson,
Y.-J. Kim and S. A. Chambers, Catal. Today, 1999, 51, 513.
19 D. R. Rainer and D. W. Goodman, NATO ASI Series, Series E, Applied
Sciences, 1997, 331(Chemisorption and Reactivity on Supported Clusters and
Thin Films), 27.
20 H. Onishi and Y. Iwasawa, in Interfacial Science, ed. M. W. Roberts,
Blackwell, Oxford, UK, 1997, p. 57.
21 L. Guczi, Z. Paszti, K. Frey, A. Beck, G. Peto and Cs. S. Daroczy, Top. Catal.,
2006, 39, 137.
22 G. Centi, R. Passalacqua, S. Perathoner, D. S. Su, G. Weinberg and
R. Schlogl, Phys. Chem. Chem. Phys., 2007, 9, 4930.
23 S. Abbet and U. Heiz, Nanocatalysis, in Encyclopedia of Nanoscience and
Nanotechnoloy, ed. H. S. Nalwa, American Scientic Publishers, Valencia, CA,
US, 2004, vol. 6, 161.
24 A. Corma, Chem. Rev., 1997, 97, 2373.
25 F. Schuth and W. Schmidt, Advanced Materials (Weinheim, Ger.), 2002, 14,
629.
26 F. Schuth, A. Wingen and J. Sauer, Microporous and Mesoporous Materials,
2001, 4445, 465.
27 G. Centi, S. Perathoner and F. Vazzana, in Catalysis by Unique Ion Structures
in Solid Matrices, NATO Science Series II: Mathematics Physics and
Chemistry, eds. A. Bell, G. Centi and B. Wichterlowa, Kluwer/Academic
Press Pub., New York, 2001, ch. 11, p. 165.
28 K. P. Jayadevan and T. Y. Tseng, Oxide Nanoparticles, in Encyclopedia of
Nanoscience and Nanotechnology, ed. H. S. Nalwa, American Scientic
Publishers, Valencia, CA, US, 2004, vol. 8, 333.
29 Y. Kumzerov and S. Vakhrushev, Nanostructures within Porous Materials, in
Encyclopedia of Nanoscience and Nanotechnoloy, ed. H. S. Nalwa, American
Scientic Publishers, Valencia, CA, US, 2004, vol. 7, 811.
30 G. R. Patzke, F. Krumeich and R. Nesper, Angew. Chem., Int. Ed., 2002, 41,
2446.
31 G. Centi, S. Perathoner, F. Triro`, A. Aboukais, C. F. A ssi and M. Guelton,
J. Phys. Chem., 1992, 96, 2617.
32 G. Bellussi, G. Centi, S. Perathoner and F. Triro`, in Selective Oxidation
Catalysts ACS Symposium Series, eds. T. Oyama and J. Hightower, American
Chemical Society Pub, Washington, 1992, ch. 21, p. 281.
33 G. Centi, F. Fazzini, L. Canesson and A. Tuel, Stud. Surf. Sci. Catal., 1997,
110, 893.
34 G. Centi, S. Perathoner and S. Tonini, Catal. Today, 2000, 61, 211.
35 H. Maekawa, Materia (Nihon Kinzoku Gakkai), 2006, 45, 464.
36 G. D. de A. A. Soler-Illia, E. L. Crepaldi, D. Grosso and C. Sanchez, Curr.
Opin. Colloid Interface Sci., 2003, 8, 109.
37 F. Di Renzo, A. Galarneau, P. Trens and F. Fajula, in Handbook of Porous
Solids, eds. F. Schueth, K. S. W. Sing and J. Weitkamp, Wiley-VCH,
Weinheim, Germany, 2002, 1311.
38 F. Iskandar, Mikrajuddin and F. Okuyama, Ordered Nanoporous Particles, in
Encyclopedia of Nanoscience and Nanotechnoloy, ed. H. S. Nalwa, American
Scientic Publishers, Valencia, CA, US, 2004, vol. 8, p. 259.
39 G. J. de Soler-Illia, C. Sanchez, B. Lebeau and J. Patarin, Chem. Rev., 2002,
102, 4093.
40 W. Ueda, M. Sadakane and H. Ogihara, Catal. Today, 2008, 132, 2.
41 X.-W. Chen, Z. Zhu, M. Haevecker, D. Su and R. Schlogl, Mater. Res. Bull.,
2007, 42, 354.

Catalysis, 2009, 21, 82130 | 121


This journal is
c The Royal Society of Chemistry 2009
42 L.-C. Wang, Y.-M. Liu, M. Chen, Y. Cao, H.-Y. He and K.-N. Fan, J. Phys.
Chem. C, 2008, 112, 6981.
43 J. Rajeswari, B. Viswanathan and T. K. Varadarajan, Mate. Chem. Phys.,
2007, 106, 168.
44 Z. Zhong, J. Ho, J. Teo, S. Shen and A. Gedanken, Chem. Mater., 2007, 19,
4776.
45 M. A. Cortes-Jacome, M. Morales, C. Angeles Chavez, L. F. Ramirez-Verduzco,
E. Lopez-Salinas and J. A. Toledo-Antonio, Chem. Mater., 2007, 19, 6605.
46 F. Teng, B. Gaugeu, S. Liang and Y. Zhu, Appl. Catal., A: General, 2007, 328,
156.
47 J.-N. Nian, S.-A. Chen, C.-C. Tsai and H. Teng, J. Phys. Chem. B, 2006, 110,
25817.
48 G. Cao and S. J. Limmer, Oxide Nanowires and Nanorods, in Encyclopedia
of Nanoscience and Nanotechnoloy, ed. H. S. Nalwa, American Scientic
Publishers, Valencia, CA, US, 2004, vol. 8, 377.
49 S. Sharma, H. Li, H. Chandrasekaran, R. C. Mani and M. K. Sunkara,
Synthesis of Inorganic Nanowires and Nanotubes, in Encyclopedia of Nanoscience
and Nanotechnoloy, ed. H. S. Nalwa, American Scientic Publishers, Valencia,
CA, US, 2004, vol. 10, 327.
50 H. M. Shang, Y. Wang and G. Cao, Key Engineering Materials (Trans Tech
Publications Ltd), 2007, 336338(High-Performance Ceramics IV), 2122.
51 Y. Wang and G. Cao, Chem. Mater., 2006, 18, 2787.
52 X. Wang, X. Sun, J. Xu and Y. Li, in Nanowires and Nanobelts, ed. Z. L. Wang,
Kluwer Academic Publishers, Norwell, US, 2003, vol. 2, 173.
53 W.-T. Yao and S.-H. Yu, Int. J. Nanotechnology, 2007, 4, 129.
54 Y. Mao, T.-J. Park and S. S. Wong, Chem. Commun., 2005, 5721.
55 X. Chen and S. S. Mao, J. Nanosci. Nanotechnol., 2006, 6, 906.
56 O. K. Varghese and C. A. Grimes, J. Nanosci. Nanotechnol., 2003, 3, 277.
57 D. V. Bavykin, J. M. Friedrich and F. C. Walsh, Adv. Mater. (Weinheim,
Ger.), 2006, 18, 2807.
58 C. A. Grimes, J. Mater. Chem., 2007, 17, 1451.
59 Q. Chen and L.-M. Peng, Int. J. Nanotechnology, 2007, 4, 44.
60 G. Centi and S. Perathoner, in Catalysis, ed. J. J. Spivey, Royal Society of
Chemistry, Cambridge, UK, 2007, vol. 20, 367.
61 K. Hashimoto, H. Irie and A. Fujishima, J. Appl. Phys., Part 1, 2005, 44, 8269.
62 A. G. Agrios and P. Pichat, J. Appl. Electrochem., 2005, 35, 655.
63 H. J.-M. Herrmann, Top. Catal., 2005, 34, 49.
64 M. Anpo and M. Takeuchi, J. Catal., 2003, 216, 505.
65 G. K. Mor, O. K. Varghese, M. Paulose, K. Shankar and C. A. Grimes,
Sol. Energy Mater. Sol. Cells, 2006, 90, 2011.
66 M. Adachi, J. Jiu and S. Isoda, Current Nanoscience, 2007, 3, 285.
67 H. Gnaser, B. Huber and C. Ziegler, Nanocrystalline TiO2 for Photocatalysis,
in Encyclopedia of Nanoscience and Nanotechnoloy, ed. H. S. Nalwa,
American Scientic Publishers, Valencia, CA, US, 2004, vol. 90, 669.
68 M. Anpo, S. Dohshi, M. Kitano and Y. Hu, in Metal Oxides. Chemistry and
Applications, ed. J. L. G. Fierro, CRC Taylor & Francis, Boca Raton, US,
2006, p. 595.
69 A. Borras, A. Barranco and A. T. Gonzalez-Elipe, J. Mater. Sci., 2006, 41,
5220.
70 S. Tanemura, L. Miao, W. Wunderlich, M. Tanemura, Y. Mori, S. Toh and
K. Kaneko, Science and Technology of Advanced Materials, 2005, 6, 11.
71 M. L. Hitchman and S. E. Alexandrov, Electrochemical Society Interface,
2001, 10, 40.

122 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
72 F. Campus, O. Bonhote, M. Gratzel, S. Heinen and L. Walder, Sol. Energy
Mater. Sol. Cells, 1999, 56, 281.
73 M. Gratzel, Curr. Opin. Colloid Interface Sci., 1999, 4, 314.
74 P. V. V. Jayaweera, A. G. U. Perera and K. Tennakone, Inorg. Chim. Acta,
2008, 361, 707.
75 Y. Amao, Current Nanoscience, 2008, 4, 45.
76 P. V. Kamat, J. Phys. Chem. C, 2007, 111, 2834.
77 A. J. Frank, N. Kopidakis and J. van de Lagemaat, Coord. Chem. Rev., 2004,
248, 1165.
78 A. S. Arico, P. Bruce, B. Scrosati, J.-M. Tarascon and W. van Schalkwijk,
Nature Materials, 2005, 4, 366.
79 N. S. Lewis, Science, 2007, 315, 798.
80 G. Chen, T. Zeng, T. Borca-Tasciuc and D. Song, Mater. Sci. Eng., 2000,
A292, 155.
81 G. Centi and S. Perathoner, Catal. Today, 2008, in press, (presented at Paul
Sabatier 5th European Conference on Catalysis, Ottrott (France), September
1216th, 2007).
82 S. Perathoner, R. Passalacqua, G. Centi, D. S. Su and G. Weinberg, Catal.
Today, 2007, 122, 3.
83 L. Frunza, S. Frunza, A. Schonhals, U. Bentrup, R. Fricke, I. Pitsch and
H. Kosslick, Studies in Surface Science and Catalysis, 2002, 142B(Impact of
Zeolites and Other Porous Materials on the New Technologies at the Beginning
of the New Millennium), 1323.
84 G. Dosseh, Y. Xia and C. Alba-Simionesco, J. Phys. Chem. B, 2003, 107, 6445.
85 G. Rother, D. Woywod, M. Schoen and G. Findenegg, J. Chem. Phys., 2004,
120, 11864.
86 N. Floquet, J. P. Coulomb, N. Dufau and R. Kahn, Studies in Surface Science
and Catalysis, 2004, 154B(Recent Advances in the Science and Technology of
Zeolites and Related Materials), 1804.
87 N. Floquet, J. P. Coulomb, G. Andre and R. Kahn, Studies in Surface Science
and Catalysis, 2007, 160(Characterization of Porous Solids VII), 375.
88 P. Trens, N. Tanchoux, A. Galarneau and F. Fajula, J. Phys. Chem. B, 2005,
109, 16415.
89 S. Pariente, P. Trens, F. Fajula, F. Di Renzo and N. Tanchoux, Appl. Catal.,
A: General, 2006, 307, 51.
90 F. Goettmann and C. Sanchez, J. Mater. Chem., 2007, 17, 24.
91 K. Malek, C. Li and R. A. van Santen, J. Mol. Catal. A: Chem., 2007, 271, 98.
92 M. K. Kidder and A. C. III Buchanan, J. Phys. Chem. C, 2008, 112, 3027.
93 S. Abate, K. Barbera, P. Lanzafame, S. Perathoner and G. Centi, Preprints-
American Chemical Society, Div. of Petroleum Chemistry, 2008, 53, 227.
94 G. Centi and S. Perathoner, Catal. Today, 2008, submitted. (plenary lecture at
the International Symposium on Catalysis for Clean Energy and Sustainable
Chemistry (CCESC), Madrid (Spain), June 2008).
95 J. M. Thomas and R. Raja, Acc. Chem. Res., 2008, 41, 708.
96 E. E. Santiso, M. K. Kostov, A. M. George, M. B. Nardelli and K. E. Gubbins,
Appl. Surface Science, 2007, 253, 5570.
97 D. G. Shchukin, D. V. Sviridov and J. Photochem., Photobiology, C:
Photochem. Reviews, 2006, 7, 23.
98 B. Gall, M. Bortenschlager, O. Nuyken and R. Weberskirch, Macromol.
Chemistry and Physics, 2008, 209, 1152.
99 T. S. Koblenz, J. Wassenaar and J. N. H. Reek, Chem. Soc. Rev., 2008, 37, 247.

Catalysis, 2009, 21, 82130 | 123


This journal is
c The Royal Society of Chemistry 2009
100 A. P. Alivisatos, Y. Yin, R. M. Rioux and G. A. Somorjai, PCT Int. Appl.
(WO 2006033666), 2006, assigned to The Regents of the University of
California, USA.
101 D. G. Shchukin, G. B. Sukhorukov and H. Moehwald, in Physics Chemistry
and Application of Nanostructures, eds. V. E. Borisenko, S. V. Gaponenko and
V. S. Gurin, World Scientic Publishing Co., Singapore, 2005, p. 493.
102 J. Li and H. C. Zeng, Angew. Chem., Int. Ed., 2005, 44, 4342.
103 Y.-M. Chung and H.-K. Rhee, Catal. Lett., 2003, 85, 159.
104 S. Mayavan, N. R. Choudhury and N. K. Dutta, Adv. Mater. (Weinheim,
Ger.), 2008, 20, 1819.
105 T. Akita, M. Okumura, K. Tanaka, K. Ohkuma, M. Kohyama, T. Koyanagi,
M. Date, S. Tsubota and M. Haruta, Surface and Interface Analysis, 2005, 37,
265.
106 H. Li, X. Duan, G. Liu and X. Liu, J. Mater. Sci., 2008, 43, 1669.
107 T. Neumann, B. Haupt and M. Ballau, Macromolec. Bioscience, 2004, 4, 13.
108 S. Liu and A. Chen, Langmuir, 2005, 21, 8409.
109 P. Xiao, B. B. Garcia, Q. Guo, D. Liu and G. Cao, Electrochem. Commun.,
2007, 9, 2441.
110 A. Curulli, A. Cusma, S. Kaciulis, G. Padeletti, L. Pandol, F. Valentini and
M. Viticoli, Surface and Interface Analysis, 2006, 38, 478.
111 S.-J. Bao, C. M. Li, J.-F. Zang, X.-Q. Cui, Y. Qiao and J. Guo, Adv. Funct.
Mater., 2008, 18, 591.
112 G. M. Mura, G. Galli, C. Pratesi, G. Frascotti, L. Serbolisca, P. Pedroni and
G. Grandi, J. Marine Biotechn., 1996, 4, 68.
113 Y. Lu, Y. Liu, J. Xu, C. Xu, B. Liu and J. Kong, Sensors, 2005, 5, 258.
114 Y. Amao and T. Komori, Biosensors & Bioelectronics, 2004, 19, 843.
115 A. Brune, G. Jeong, P. A. Liddell, T. Sotomura, T. A. Moore, A. L. Moore
and D. Gust, Langmuir, 2004, 20, 8366.
116 T. Kieboom, in Catalysis for Renewables: From Feedstock to Energy Production,
eds. G. Centi and R. A. Van Santen, Wiley-VCH, Weinheim, Germany, 2007,
p. 273.
117 J. T. W. Yeow and J. P. M. She, Nanotechnology, 2006, 17, 5441.
118 B. M. Kim, S. Sinha and H. H. Bau, Nano Letters, 2004, 4, 2203.
119 F. Zhao, S.-i. Fujita and M. Arai, Curr. Org. Chem., 2006, 10, 1681.
120 K. L. Fow, S. Jaenicke, T. E. Mueller and C. Sievers, J. Mol. Catal. A: Chem,
2008, 279, 239.
121 C. Sievers, O. Jimenez, R. Knapp, X. Lin, T. E. Mueller, A. Tuerler,
B. Wierczinski and J. A. Lercher, J. Mol. Catal. A: Chem., 2008, 279, 187.
122 O. Jimenez, T. E. Mueller, C. Sievers, A. Spirkl and J. A. Lercher, Chem.
Commun., 2006, 2974.
123 H. E. Prakasam, K. Shankar, M. Paulose, O. K. Varghese and C. A. Grimes,
J. Phys. Chem. C, 2007, 111, 7235.
124 M. Paulose, H. E. Prakasam, O. K. Varghese, L. Peng, K. C. Popat,
G. K. Mor, T. A. Desai and C. A. Grimes, J. Phys. Chem. C, 2007, 111, 14992.
125 S. P. Albu, A. Ghicov, J. M. Macak and P. Schmuki, Phys. Status Solidi RRL:
Rapid Res Lett., 2007, 1, R65.
126 S. P. Albu, A. Ghicov, J. M. Macak, R. Hahn and P. Schmuki, Nano Letters,
2007, 7, 1286.
127 J. M. Macak, H. Tsuchiya, L. Taveira, S. Aldabergerova and P. Schmuki,
Angew. Chem., Int. Ed., 2005, 44, 7463.
128 M. Paulose, L. Peng, K. C. Popat, O. K. Varghese, T. J. LaTempa, N. Bao and
T. A. Desai, C.A. J. Membrane Science, 2008, 319, 199.

124 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
129 H. Imai, Y. Takei, K. Shimizu, M. Matsuda and H. Hirashima, J. Mater.
Chem., 1999, 9, 2971.
130 J. Nishino, S. Teekateerawej and Y. Nosaka, J. Mater. Sci. Lett., 2003, 22,
1007.
131 S. Teekateerawej, J. Nishino and Y. Nosaka, Adv. in Technology of Materials
and Materials Processing J., 2003, 5, 80.
132 T. Peng, H. Yang, G. Chang, K. Dai and K. Hirao, Chem. Lett., 2004, 33, 336.
133 S. Lee, C. Jeon and Y. Park, Chem. Mater., 2004, 16, 4292.
134 H. Zhang, X. Quan, S. Chen and H. Zhao, Env. Science & Techn., 2006, 40,
6104.
135 V. Lehmann, J. Electrochem. Soc., 1993, 140, 2836.
136 J. M. Macak, H. Tsuchiya, A. Ghicov, K. Yasuda, R. Hahn, S. Bauer and
P. Schmuki, Curr. Opin. Solid State Mater. Sci., 2007, 11, 3.
137 R. Thiruvenkatachari, S. Vigneswaran and I. S. Moon, Korean J. Chem. Eng.,
2008, 25, 64.
138 N. Agoudjil and T. Benkacem, Desalination, 2007, 206, 531.
139 M. V. Fedkin, E. Chalkova, S. Komarneni, D. J. Wesolowski and S. N. Lvov,
ECS Transactions, 2006, 1(6, Proton Exchange Membrane Fuel Cells V), 215.
140 A. Sacca, A. Carbone, E. Passalacqua, A. DEpifanio, S. Licoccia,
E. Traversa, E. Sala, F. Traini and R. Ornelas, J. Power Sources, 2005, 152, 16.
141 K. Wang, M. Wei, M. A. Morris, H. Zhou and J. D. Holmes, Adv. Mater.
(Weinheim, Ger.), 2007, 19, 3016.
142 J. Xu, C. Jia, B. Cao and W. F. Zhang, Electrochim. Acta, 2007, 52, 8044.
143 R. Hahn, A. Ghicov, H. Tsuchiya, J. M. Macak, A. G. Munoz and
P. Schmuki, Physica Status Solidi A: Appl. and Materials Science, 2007, 204,
1281.
144 C. Laberty-Robert, M. Kuemmel, J. Allouche, C. Boissiere, L. Nicole, D. Grosso
and C. Sanchez, J. Mater. Chem., 2008, 18, 1216.
145 J. Sekulic, J. E. ten Elshof and D. H. A. Blank, Adv. Mater. (Weinheim, Ger.),
2004, 16, 1546.
146 A. T. Bell, B. C. Gates, and D. Ray, Basic Research Needs Catalysis for
Energy (PNNL-17214), U.S. Department of Energy, Report from a
Workshop held in August 68, 2007, in Bethesda, Maryland (US)
(www.sc.doe.gov/bes/reports/list.html).
147 G. Centi and S. Perathoner, Stud. Surf. Sci. Catal., 2007, 172(Science and
Technology in Catalysis 2006, Kodansha/Elsevier), 79.
148 S. Perathoner, R. Passalacqua, G. Centi, D. S. Su and G. Weinberg, Stud.
Surf. Sci. Catal., 2007, 172(Science and Technology in Catalysis 2006,
Kodansha/Elsevier), 437.
149 http://titaniananotubes.com.
150 S. Funk, B. Hokkanen, U. Burghaus, A. Ghicov and P. Schmuki, Nano
Letters, 2007, 7, 1091.
151 M. Zhang, Y. Bando and K. Wada, J. Mater. Sci. Lett., 2001, 20, 167.
152 S. M. Liu, L. M. Gan, L. H. Liu, W. D. Zhang and H. C. Zeng, Chem. Mater.,
2002, 14, 1391.
153 Z. L. Wang, R. P. Gao, J. L. Gole and J. D. Stout, Adv. Mater. (Weinheim,
Ger.), 2000, 12, 1938.
154 P. Hoyer, Langmuir, 1996, 12, 1411.
155 R. A. Caruso, Top. Curr. Chem., 2003, 226(Colloid Chemistry I), 91.
156 S. Yamabi, H. Imai and K. Awazu, Chem. Lett., 2002, 714.
157 R. Sanz, A. Johansson, M. Skupinski, J. Jensen, G. Possnert, M. Boman,
M. Vazquez and K. Hjort, Nano Letters, 2006, 6, 1065.

Catalysis, 2009, 21, 82130 | 125


This journal is
c The Royal Society of Chemistry 2009
158 S. J. Limmer, T. P. Chou and G. Z. Cao, J. Sol-Gel Science and Techn., 2005,
36, 183.
159 X. Wang, C. Ne, E. Graugnard, Y. Ding, J. S. King, L. A. Pranger,
R. Tannenbaum, Z. L. Wang and C. J. Summers, Adv. Mater. (Weinheim,
Ger.), 2005, 17, 2103.
160 Y. Liu, K. Okamoto, S. Hayakawa, K. Tsuru and A. Osaka, in Proceedings
of the International Congress on Glass, (Kyoto, Japan, Sept. 2004), (2004),
ed. T. Yoko, Ceramic Society of Japan, Tokyo, Japan, 16.022/1.
161 C.-C. Weng, C.-P. Chen, C.-H. Ting and K.-H. Wei, Chem. Mater., 2005, 17,
3328.
162 B. Xiang, Y. Zhang, Z. Wang, X. H. Luo, Y. W. Zhu, H. Z. Zhang and
D. P. Yu, J. Physics D: Appl. Phys., 2005, 38, 1152.
163 Y.-G. Guo, J.-S. Hu, H.-P. Liang, L.-J. Wan and C.-L. Bai, Adv. Funct.
Mater., 2005, 15, 196.
164 W. Ho, J. C. Yu and J. Yu, Langmuir, 2005, 21, 3486.
165 T. Wen, J. Zhang, T. P. Chou, S. J. Limmer and G. Cao, J. Sol-Gel Science and
Techn., 2005, 33, 193.
166 M. S. Sander, M. J. Cote, W. Gu, B. M. Kile and C. P. Tripp, Adv. Mater.
(Weinheim, Ger.), 2004, 16, 2052.
167 T. Hamaguchi, M. Uno, K. Kurosaki and S. Yamanaka, J. Alloys Compd.,
2005, 386, 265.
168 X. Peng and A. Chen, Appl. Phys. A: Materials Science & Processing, 2005, 80,
473.
169 X. D. Wang, E. Graugnard, J. S. King, Z. L. Wang and C. J. Summers, Nano
Letters, 2004, 4, 2223.
170 H. Masuda and K. Fukuda, Science, 1995, 268, 1466.
171 V. Zwilling, M. Aucouturier and E. Darque-Ceretti, Electrochim. Acta, 1991,
45, 921.
172 D. Gong, C. A. Grimes, O. K. Varghese, W. Hu, R. S. Singh, Z. Chen and
E. C. Dickey, J. Mater. Res., 2001, 16, 3331.
173 O. K. Varghese, X. Yang, J. Kendig, M. Paulose, K. Zeng, C. Palmer, K. G. Ong
and C. A. Grimes, Sens. Lett., 2006, 4, 120.
174 O. K. Varghese, D. Gong, M. Paulose, K. G. Ong and C. A. Grimes, Sens.
Actuators B, 2003, 93, 338.
175 C. A. Grimes, K. G. Ong, O. K. Varghese, X. Yang, G. Mor, M. Paulose,
E. C. Dickey, C. Ruan, M. V. Pishko, J. W. Kendig and A. J. Mason, Sensors,
2003, 3, 69.
176 M. Paulose, K. Shankar, O. K. Varghese, G. K. Mor, B. Hardin and
C. A. Grimes, Nanotechnology, 2006, 17, 1446.
177 M. Paulose, K. Shankar, O. K. Varghese, G. K. Mor and C. A. Grimes,
J. Phys. D: Appl. Phys., 2006, 39, 2498.
178 O. K. Varghese, M. Paulose, K. Shankar, G. K. Mor and C. A. Grimes,
J. Nanosci. Nanotechnol., 2005, 5, 1158.
179 M. Paulose, K. Shankar, S. Yoriya, H. E. Prakasam, O. K. Varghese, G. K. Mor,
T. A. Latempa, A. Fitzgerald and C. A. Grimes, J. Phys. Chem. B, 2006, 110,
16179.
180 K. C. Popat, E. E. Leary-Swan and T. A. Desai, Biomaterials, 2005, 26, 1969.
181 J. W. Schultze and M. M. Lohrengel, Electrochim. Acta, 2000, 45, 2499.
182 K. Yasuda and P. Schmuki, Electrochim. Acta, 2007, 52, 4053.
183 K. Yasuda, J. M. Macak, S. Berger, A. Ghicov and P. Schmuki, J. Electrochem
Soc., 2007, 154, C472.
184 S. Bauer, S. Kleber and P. Schmuki, Electrochem Commun., 2006, 8, 1321.

126 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
185 J. M. Macak, H. Tsuchiya and P. Schmuki, Angew. Chem., Int. Ed., 2005, 44,
2100.
186 K. Shankar, G. K. Mor, A. Fitzgerald and C. A. Grimes, J. Phys. Chem. C,
2007, 111, 21.
187 C. M. Ruan, M. Paulose, O. K. Varghese, G. K. Mor and C. A. Grimes,
J. Phys. Chem. B, 2005, 109, 15754.
188 G. K. Mor, O. K. Varghese, M. Paulose and C. A. Grimes, Adv. Funct. Mater.,
2005, 15, 1291.
189 G. K. Mor, O. K. Varghese, M. Paulose, K. G. Ong and C. A. Grimes, Thin
Solid Films, 2006, 496, 42.
190 H. Foll, S. Langa, J. Carstensen, M. Christophersen and I. M. Tiginyanu,
Advanced Materials (Weinheim, Ger.), 2003, 15, 183.
191 J. M. Macak, S. P. Albu and P. Schmuki, Phys. Stat. Sol. (RRL), 2007, 1,
R181.
192 J. M. Macak, S. P. Albu, D. H. Kim, I. Paramasivam, S. Aldabergerova and
P. Schmuki, Electrochem Solid-State Lett., 2007, 10, K28.
193 K. Yasuda and P. Schmuki, Electrochem. Commun., 2007, 9, 615.
194 A. Ghicov, H. Tsuchiya, J. M. Macak and P. Schmuki, Phys. Stat. Sol. A,
2006, 203, R28.
195 J. M. Macak, S. Aldabergerova, A. Ghicov and P. Schmuki, Phys. Stat. Sol. A,
2006, 203, R67.
196 J. Zhao, X. Wang, T. Sun and L. Li, Nanotechnology, 2005, 16, 2450.
197 J. Arbiol, J. Cerda, G. Dezanneau, A. Cicera, F. Peiro and A. Cornet, J. Appl.
Phys., 2002, 92, 853.
198 B. Tryba, A. W. Moravski and M. Inagaki, Appl. Catal. B: Environ., 2003, 52,
203.
199 A. Ghicov, S. Aldabergerova, H. Tsuchiya and P. Schmuki, Angew. Chem.,
Int. Ed., 2006, 45, 6993.
200 K. Yasuda and P. Schmuki, Adv. Mater. (Weinheim, Ger.), 2007, 19, 1757.
201 J. Kunze, A. Ghicov, H. Hildebrand, J. M. Macak, L. Taveira and P. Schmuki,
Z. Phys. Chem., 2006, 129, 1561.
202 O. K. Varghese and C. A. Grimes, Metal oxide nanostructures as gas sensors,
in Encyclopedia of Nanoscience and Nanotechnoloy, ed. H. S. Nalwa, American
Scientic Publishers, Valencia, CA, US, 2004, vol. 5, 505.
203 Y. Xie, Adv. Funct. Mater., 2006, 16, 1823.
204 J. M. Macak, M. Zlamal, J. Krysa and P. Schmuki, Small, 2007, 3, 300.
205 Y. Lai, L. Sun, Y. Chen, H. Zhuang, Ch. Lin and J. W. Chin, J. Electrochem.
Soc., 2006, 153, D123.
206 H. Zhuang, Ch. Lin, Y. Lai and L. Sun, J. Environ. Sci. Technol., 2007, 41,
4735.
207 J. Zhao, X. Wang, T. Sun and L. Li, J. Alloys Compd., 2007, 434435, 792.
208 F. Jiang, S.-r. Zheng, Z. Zheng, Z.-y. Xu and Y.-j. Wang, J. Env. Sciences,
Beijing, China, 2006, vol. 18, 783.
209 J. Yu, H. Yu, B. Cheng and C. Trapalis, J. Mol. Catal. A: Chem., 2006, 249,
135.
210 M. A. Khan, H.-T. Jung and O.-B. Yang, J. Phys. Chem. B, 2006, 110, 6626.
211 Y. Lin-Rui and P. Y. Chang-Zhou, J. Hazardous Materials, 2007, 139, 310.
212 Z. Zhu, Y. Zhou, H. Yu, T. Nomura and B. Fugetsu, Chem. Lett., 2006, 35,
890.
213 Y. Yu, J. C. Yu, C.-Y. Chan, Y.-K. Che, J.-C. Zhao, L. Ding, W.-K. Ge and
P.-K. Wong, Appl. Catal., B: Env., 2005, 61, 1.
214 H. Wang, X. Quan, H. Yu and S. Chen, Carbon, 2008, 46, 1126.

Catalysis, 2009, 21, 82130 | 127


This journal is
c The Royal Society of Chemistry 2009
215 Y. Yao, G. Li, S. Ciston, R. M. Lueptow and K. A. Gray, Env. Science &
Techn., 2008, 42, 4952.
216 D. Eder and A. H. Windle, Adv. Mater. (Weinheim, Ger.), 2008, 20, 1787.
217 J. Geng, Z. Jiang, Y. Wang and D. Yang, Scripta Materialia, 2008, 59, 352.
218 K. Byrappa, A. S. Dayananda, C. P. Sajan, B. Basavalingu, M. B. Shayan,
K. Soga and M. Yoshimura, J. Mater. Sci., 2008, 43, 2348.
219 B. Ahmmad, Y. Kusumoto, S. Somekawa and M. Ikeda, Catal. Commun.,
2008, 9, 1410.
220 D.-H. Lee, J.-G. Park, K. J. Choi, H.-J. Choi and D.-W. Kim, Eur. J. Inorg.
Chem., 2008, 878.
221 H. Song, X. Qiu and F. Li, Electrochim. Acta, 2008, 53, 3708.
222 M. R. S. Castro, E. D. Sam, M. Veith and P. W. Oliveira, Nanotechnology,
2008, 19, 105704/1.
223 K. Nishijima, T. Fukahori, N. Murakami, T.-a. Kamai, T. Tsubota and T. Ohno,
Appl. Catal., A: General, 2008, 337, 105.
224 L. Yang, S. Luo, S. Liu and Q. Cai, J. Phys. Chem. C, 2008, 112, 8939.
225 J. C. Kim, J. Choi, Y. B. Lee, J. H. Hong, J. I. Lee, J. W. Yang, W. I. Lee and
N. H. Hur, Chem. Commun., 2006, 5024.
226 J. Zhu, D. Yang, J. Geng, D. Chen and Z. Jiang, J. Nanoparticle Research,
2008, 10, 729.
227 H.-Y. Pan, X.-F. Gao, Q. Chen and L.-M. Peng, J. Am. Chem. Soc., 2008, 130,
1124.
228 J. L. Zhang, X. W. Zhang and L. C. Lei, Chinese Science Bull., 2008, 53, 1929.
229 C. Ratanatawanate, C. Xiong, K.J. Balkus, Abstracts of Papers, 235th
ACS National Meeting, New Orleans, LA, United States, April 610, 2008,
INOR-984.
230 P. V. Kamat, A. Kongkanand, K. Tvrdy, K. Takechi, M. Kuno, Abstracts
of Papers, 235th ACS National Meeting, New Orleans, LA, United States,
April 610, 2008, COLL-049.
231 A. Kongkanand, K. Tvrdy, K. Takechi, M. Kuno and P. V. Kamat, J. Am.
Chem. Soc., 2008, 130, 4007.
232 H.-Y. Si, Z.-H. Sun and H.-L. Zhang, Colloids and Surfaces, A: Physicochem.
and Eng. Aspects, 2008, 313314, 604.
233 W. Lee, J. Lee, S. Lee, W. Yi, S.-H. Han and B. W. Cho, Appl. Phys. Lett.,
2008, 92, 153510/1.
234 P. V. Kamat, P. Brown, A. Kongkanand, Abstracts of Papers, 235th ACS
National Meeting, New Orleans, LA, United States, April 610, 2008,
PHYS-189.
235 P. V. Kamat, A. Kongkanand, Abstracts of Papers, 235th ACS National
Meeting, New Orleans, LA, United States, April 610, 2008, PETR-029.
236 S. U. M. Khan, M. Al-Shahry and W. B. Ingler, Science, 2002, 297, 2243.
237 S. K. Mohapatra, M. Misra, V. K. Mahajan and K. S. Raja, J. Phys. Chem. C,
2007, 111, 8677.
238 C. Xu, Y. A. Shaban, W. B. Ingler and S. U. M. Khan, Sol. Energy Mater. Sol.
Cells, 2007, 91, 938.
239 J. Xu, Y. Wang, Z. Li and W. F. Zhang, J. Power Sources, 2008, 175, 903.
240 A. B. Murphy, Sol. Energy Mater. Sol. Cells, 2008, 92, 363.
241 R. Asahi, T. Morikawa, T. Ohwaki, A. Aoki and Y. Taga, Science, 2001, 293,
269.
242 Y. Chen, S. Zhang, Y. Yu, H. Wu, S. Wang, B. Zhu, W. Huang and S. Wu,
J. Dispersion Science and Techn., 2008, 29, 245.
243 Y.-Y. Hsu, T.-L. Hsiung, H. P. Wang, Y. Fukushima, Y.-L. Wei and J.-E. Chang,
Marine Pollution Bull., 2008, 57, 873.

128 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
244 D. Kim, S. Fujimoto, P. Schmuki and H. Tsuchiya, Electrochem. Commun.,
2008, 10, 910.
245 G. Liu, F. Li, D.-W. Wang, D.-M. Tang, C. Liu, X. Ma, G. Q. Lu and
H.-M. Cheng, Nanotechnology, 2008, 19, 025606/1.
246 C. Feng, Y. Wang, Z. Jin, J. Zhang, S. Zhang, Z. Wu and Z. Zhang, New J.
Chem., 2008, 32, 1038.
247 K. Wilke and H. D. Breuer, J. Photochem. Photobiol. A, 1999, 127, 107.
248 L. Lin, W. Lin, Y. Zhu, B. Zhao and Y. Xie, Chem. Lett., 2005, 34, 284.
249 M. Anpo, Catal. Surv. Jpn., 1997, 1, 169.
250 T. Ohno, T. Mitsui and M. Matsumura, Chem. Lett., 2003, 32, 364.
251 W. Zhao, W. Ma, C. Chen, J. Zhao and Z. Shuai, J. Am. Chem. Soc., 2004,
126, 4782.
252 M. Anpo and M. Takeuchi, J. Catal., 2003, 216, 505.
253 H. Tsuji, H. Sugahara, Y. Gotoh and J. Ishikawa, Nucl. Instr. Meth. Phys.
Res. B, 2003, 206, 249.
254 A. Ghicov, B. Schmidt, J. Kunze and P. Schmuki, Chem. Phys. Lett., 2007,
433, 323.
255 C.-C. Tsai and H. Teng, Appl. Surf. Sci., 2008, 254, 4912.
256 S. Zhang, Y. Chen, Y. Yu, H. Wu, S. Wang, B. Zhu, W. Huang and S. Wu,
J. Nanoparticle Res., 2008, 10, 871.
257 Y. Su, S. Han, X. Zhang, X. Chen and L. Lei, Mater. Chem. Phys., 2008, 110,
239.
258 J. Li, N. Lu, X. Quan, S. Chen and H. Zhao, Ind. Eng. Chem. Res., 2008, 47,
3804.
259 J. Zhang, B. Liu, M. Li, S. Yu, L. Wang, Z. Li, D. Liu, Y. Hou, Y. Zou,
B. Zou, T. Cui and G. Zou, Cryst. Growth Des., 2008, 8, 1812.
260 D. Li, J. Phys. Chem. C, 2008, 112, 5405.
261 H. E. Prakasam, M. Paulose, O. K. Varghese, G. K. Mor and C. A. Grimes,
Nanotechnology, 2006, 17, 4285.
262 Y. Xie, Electrochim. Acta, 2006, 51, 3399.
263 S. Yang, X. Quan, X. Li and C. Sun, Photochem. Photobiological Sciences,
2006, 5, 808.
264 S. Yang, Y. Liu and C. Sun, Appl. Catal., A: General, 2006, 301, 284.
265 M. Tian, G. Wu, B. Adams, J. Wen and A. Chen, J. Phys. Chem. C, 2008, 112,
825.
266 Z. Liu, X. Zhang, S. Nishimoto, M. Jin, D. A. Tryk, T. Murakami and
A. Fujishima, J. Phys. Chem. C, 2008, 112, 253.
267 J. P. H. Sukamto, C. S. Mcmillan and W. Smyrl, Electrochim. Acta, 1993, 38,
15.
268 F. I. Marin, M. A. Hamstra and D. Vanmaekelbergh, J. Electrochem. Soc.,
1996, 143, 1137.
269 G. K. Mor, K. Shankar, M. Paulose, O. K. Varghese and C. A. Grimes, Nano
Lett., 2005, 5, 191.
270 M. Paulose, G. K. Mor, O. K. Varghese, K. Shankar and C. A. Grimes,
J. Photochemistry and Photobiology A: Chem, 2006, 178, 8.
271 N. K. Allam, K. Shankar and C. A. Grimes, J. Mater. Chem., 2008, 18, 2341.
272 K. Shankar, G. K. Mor, H. E. Prakasam, S. Yoriya, M. Paulose, O. K. Varghese
and C. A. Grimes, Nanotechnology, 2007, 18, 065707/1.
273 S. K. Mohapatra, M. Misra and V. K. Mahajan, J. Catal., 2007, 246, 362.
274 S. K. Mohapatra and M. Misra, J. Phys. Chem. C, 2007, 111, 11506.
275 H.-L. Kuo, C.-Y. Kuo, C.-H. Liu, J.-H. Chao and C.-H. Lin, Catal. Lett.,
2007, 113, 7.

Catalysis, 2009, 21, 82130 | 129


This journal is
c The Royal Society of Chemistry 2009
276 C.-H. Lin, C.-H. Lee, J.-H. Chao, C.-Y. Kuo, Y.-C. Cheng, W.-N. Huang,
H.-W. Chang, Y.-M. Huang and M.-K. Shih, Catal. Lett., 2004, 98, 61.
277 S. Funk and U. Burghaus, Catal. Lett., 2007, 118, 118.
278 B. Zhu, X. Zhang, S. Wang, S. Zhang, S. Wu and W. Huang, Microporous and
Mesoporous Materials, 2007, 102, 333.
279 B. Zhu, Q. Guo, S. Wang, X. Zheng, S. Zhang, S. Wu and W. Huang, Reaction
Kinetics and Catalysis Letters, 2006, 88, 301.
280 B. Zhu, Q. Guo, X. Huang, S. Wang, S. Zhang, S. Wu and W. Huang, J. Mol.
Catal. A: Chem, 2006, 249, 211.
281 Y. Sato, M. Koizumi, T. Miyao and S. Naito, Catal. Today, 2006, 111, 164.
282 B. Dong, B.-L. He, J. Huang, G.-Y. Gao, Z. Yang and H.-L. Li, J. Power
Sources, 2008, 175, 266.
283 D. V. Bavykin, A. A. Lapkin, P. K. Plucinski, J. M. Friedrich and F. C. Walsh,
J. Catal., 2005, 235, 10.
284 S.-A. Chen, J.-N. Nian, C.-C. Tsai and H. Teng, J. Air & Waste Management
Assoc., 2007, 57, 600.
285 L. M. Sikhwivhilu, N. J. Coville, B. M. Pulimaddi, J. Venkatreddy and
V. Vishwanathan, Catal. Commun., 2007, 8, 1999.
286 J. M. Macak, P. J. Barczuk, H. Tsuchiya, M. Z. Nowakowska, A. Ghicov,
M. Chojak, S. Bauer, S. Virtanen, P. J. Kulesza and P. Schmuki, Electrochem.
Commun., 2005, 7, 1417.
287 F. Hu, F. Ding, S. Song and P. K. Shen, J. Power Sources, 2006, 163, 415.
288 H. Song, X. Qiu, X. Li, F. Li, W. Zhu and L. Chen, J. Power Sources, 2007,
170, 50.
289 H. Song, X. Qiu, F. Li, W. Zhu and L. Chen, Electrochem. Commun., 2007, 9,
1416.
290 D. He, L. Yang, S. Kuang and Q. Cai, Electrochem. Commun., 2007, 9, 2467.
291 H. Song, L. Qiu and F. Li, Electrochim. Acta, 2008, 53, 3708.
292 T. Maiyalagan, B. Viswanathan and U. V. Varadaraju, J. Nanosci. Nanotechnol.,
2006, 6, 2067.
293 M. Hepel, I. Dela, T. Hepel, J. Luo and C. J. Zhong, Electrochim. Acta, 2007,
52, 5529.
294 H. Song, X. Qiu, D. Guo and F. Li, J. Power Sources, 2008, 178, 97.
295 J. M. Macak, F. Schmidt-Stein and P. Schmuki, Electrochem. Commun., 2007,
9, 1783.
296 X.-Y. Xie, Z.-F. Ma, X. Wu, Q.-Z. Ren, X. Yuan, Q.-Z. Jiang and L. Hu,
Electrochim. Acta, 2006, 52, 2091.
297 W. Wang, J. Zhang, H. Huang, Z. Wu and Z. Zhang, Appl. Surf. Sci., 2008,
254, 1725.
298 Y. Guo, X. Quan, N. Lu, H. Zhao and S. Chen, Env. Science & Techn., 2007,
41, 4422.
299 L. Yu and X. Zhang, Mater. Chem. Phys., 2004, 87, 168.
300 V. M. Prida, M. Herna`ndez-Ve`lez, K. R. Pirota, A. Mene`ndez and M. Va`zquez,
Nanotechnology, 2005, 16, 2696.
301 H. Tsuchiya, J. M. Macak, A. Ghicov, L. Taveira and P. Schmuki, Corrosion
Science, 2005, 47, 3324.
302 Y. Xie, Adv. Funct. Mater., 2006, 16, 1823.
303 Y. Xie, Nanotechnology, 2006, 17, 3340.
304 L. Yang, Q. Cai and Y. Yu, Inorganic Chemistry, 2006, 45, 9616.
305 W. Zheng, Y. F. Zheng, K. W. Jin and N. Wang, Talanta, 2008, 74, 1414.
306 J. Park, S. Bauer, K. von der Mark and P. Schmuki, Nano Letters, 2007, 7,
1686.

130 | Catalysis, 2009, 21, 82130


This journal is
c The Royal Society of Chemistry 2009
Recent advances in heterogeneous catalysis
enabled by first-principles methods
Ye Xu
DOI: 10.1039/b712659j

1. Introduction
The development of the catalytic science since its advent in the early 19th
century1 laid the foundation for the tremendous growth of the petroleum and
chemical industries in the 20th century, one of the factors directly responsible
for the substantial increase in the standards of living in the industrialized world.
Today, catalysts are responsible for over $3.3 trillion in global GDP (2002
gures2), much of which due to heterogeneous catalysis. Traditionally hetero-
geneous catalysis has been an empirical science. Commercially successful
catalyst formulations have emerged through thousands of experiments, many
of which involving complex mixtures of metals, metal compounds, promoters
and inhibitors, and functional supports.3,4 The development process is often
time- and resource-intensive and discoveries often serendipitous. Therefore,
researchers have always sought detailed understanding of how their catalysts
work, in order to approach catalyst development in a rational and scientically
well-grounded manner. The wedding of surface science methods, including
various spectroscopic, microscopic, and temperature-programmed techniques,
to catalysis in the latter half of the 20th century, an approach pioneered by such
prominent scientists as Gerhardt Ertl and Gabor Somorjai, is a huge step
forward for the catalytic science.
The beginning of the new millennium nds the cost of raw material and
energy rapidly rising, traditional feedstock showing signs of depletion, and
pollution taking a toll on our planets eco-system. Researchers are once
again being called upon not only to improve existing catalytic processes but
to develop entirely new ones in order to meet the continuously changing
demands for energy and chemicals production, which now also need to
reduce energy consumption, reduce or eliminate pollutants and toxic
wastes, and use alternative feedstock. The mainstream approach to catalyst
synthesis, testing, and characterization, however, remains largely in a top-
down mode and is limited in its ability to meet these new demands
eciently.
A new wave of research tools have emerged at the end of the 20th century that
hold the promise of ushering in a new area in heterogeneous catalysis research.
New imaging techniques such as STM and AFM oer a direct, spatially
resolved view of catalytic surfaces, important clues for the nature of individual
adsorbates and adsorbate-surface interaction, and sometime glimpses of mole-
cular transformation events, for a growing number of systems. These and many
spectroscopic techniques are also being continuously enhanced to provide better
time/spatial resolutions and to operate in increasingly realistic conditions (e.g.
high pressure or liquid phase instead of the traditional UHV).5,6 At the outer

Center for Nanophase Materials Sciences, Oak Ridge National Laboratory,


Oak Ridge TN 37831, USA

Catalysis, 2009, 21, 131153 | 131


This journal is 
c The Royal Society of Chemistry 2009
limits of science, sub-angstrom, attosecond, and single turnover resolutions
have been achieved in limited circumstances.7,8 Information about catalysts and
catalytic reactions are now being generated at an unprecedented level of detail.
Meanwhile, synthetic chemistry has been creating an ever broader menagerie of
nano-scale objects with diverse and increasingly well-controlled compositions
and structures. A non-exhaustive list may include: surface and near-surface
metal alloys; core-shell particles; oxide layers or clusters supported on metals
and oxides; etched particles; size-dened clusters; mesoporous structures;
nanotubes and nanowires; metal-organic frameworks; organic and organo-
metallic compounds anchored to inorganic surfaces. Many of these motifs are
already being explored for use as catalysts.
The same decade or so has also seen a tremendous growth in the application
of quantum chemistry methods, in particular electronic structure calculations
based on density functional theory (DFT), to heterogeneous catalysis and
surface science in general. This growth has been actuated both by a rapid
increase in computer power and by new developments in theory and its
algorithmic implementations, such as generalized gradient approximation-
based exchangecorrelation functionals (e.g. PW91,9 and PBE10); pseudo-
potential approximations11,12 and the projector-augmented wave method;13,14
and plane-wave formulations; all of which have made it possible to strike the
necessary balance between accuracy and cost to nally enable calculations
involving extended surfaces and large clusters, especially those containing
metals.
Together with quantum chemistry, energy minimization algorithms,
transition state-seeking algorithms,1517 mean-eld and statistical thermo-
dynamic and kinetic modeling techniques,1820 d-band theory,21 bond order
conservation and adsorption energy scaling relations,2224 and Brnsted-
Evans-Polanyi (BEP) linear energy relations2527 have led to signicant
progress in explaining and predicting surface property and reactivity and in
providing molecular- and atomic-level information about heterogeneous
catalysis that remains challenging for experiment, including the determina-
tion of catalyst structures, various energetic quantities, properties of
transient intermediates and transition states, and reaction mechanisms.
Once embodied in computational algorithms, theory oers the tremendous
(albeit virtual) advantage of controlling matter at the atomic level and
investigating hypothetical catalysts all conveniently on a computer. Until
real matter can be as readily manipulated and probed, theory will remain an
indispensable complement to increasingly atomistic synthesis and charac-
terization techniques. Jointly they have already begun to revolutionize the
catalytic science.
This chapter is not intended to furnish a comprehensive review of the
latest theoretical developments across all types of reaction and catalyst in
heterogeneous catalysis. It will instead briey survey two cross-cutting
themes that have in recent years beneted substantially from theoretical
insights, especially what rst-principles DFT calculations have been able to
explain and predict, and that are commanding substantial current interest:
the design of new catalysts aided by theoretical insights; and the elucidation
of the molecular level eects of reactive physiochemical environments on
catalytic reactivity. The relatively simple and predictable structure of metals

132 | Catalysis, 2009, 21, 131153


This journal is 
c The Royal Society of Chemistry 2009
has to a large extent facilitated the application of theory along these lines, so
many of the examples given below involve metal catalysts. The chapter will
close with an outlook for opportunities and challenges for theory in
heterogeneous catalysis.

2. Theory-aided catalyst design


Given its ability to generate mechanistic insight and perform virtual, in silico
experiments, it is no surprise that theory has already been seized upon to aid
catalyst development. Before entire catalytic reactions can be simulated at
multiple time and length scales with just a few keystrokes, however, theory
will remain most productively employed when used to identify key factors
controlling e.g. activity or selectivity. The desired states of these factors and
the appropriate ways for achieving them will then need to be identied,
likely again with the help of theory, following which new catalysts are
proposed. Finally the promising candidates are synthesized and tested in
the laboratory. This approach requires the pertinent reaction mechanisms
to be suciently elucidated.

2.1 Targeting activity and selectivity


Reactions involving simple reactants and products often oer simple descrip-
tions of activity or selectivity. The Brnsted-Evans-Polanyi (BEP) relation,
for instance, states that the activation barrier, Ea, and the heat of reaction,
DE, of elementary dissociation reactions are often linearly correlated (Fig. 1).
Extensive DFT calculations2628 have supported this empirical relation for a
large number of elementary dissociation reactions and suggest that it is due to
the structural similarities between the transition state and the product states.
When such a dissociation step is rate-limiting, knowing DE is sucient to
capture the activity of the overall reaction, which exhibits volcano-shaped

Fig. 1 Plots of activation energy (Ediss


a ) against reaction enthalpy (DH) for (a) more than
50 elementary reactions; (bd) three subclasses of steps: (b) discrete adsorbate dehydrogenation
(class I); (c) diatomic dissociation and hydrocarbon cracking (class II); and (d) triatomic
dissociation (class III). The correlation coecients (r) have been determined from linear
regression. Reprinted with permission from Michaelides et al., Journal of the American
Chemical Society, 2003, 125, 3704.27 r 2003, American Chemical Society.

Catalysis, 2009, 21, 131153 | 133


This journal is 
c The Royal Society of Chemistry 2009
functional dependence on DE when combined with a Langmuir-Hinshelwood
(LH) type of mechanism.28,29
Jens Nrskov and co-workers have adeptly exploited the BEP relation to
improve upon several established technological catalysts. An early success
came with the identication of CoMo and NiMo alloys (actually nitrides)
based on the binding energy of N (DEN) to replace the more expensive Ru
for ammonia synthesis (N2 + 3H2 - 2NH3).30,31 More recently, Andersson
et al.32 screened over a hundred elemental metals and bi-metallic alloys for
the methanation reaction (CO + 3H2 - CH4 + H2O). DFT calculations
had shown a BEP relation between the barrier and the reaction energy of
CO dissociation (DECO) across metals, and suggested that the optimum
monometallic catalysts were in fact Co and Ru. The traditional Ni catalyst
was a compromise between the more active but expensive Co and Ru and
the less active but much cheaper Fe. Pareto analysis based on DECO and
catalyst cost identied several FeNi alloys, which experiments conrmed
to be more active toward both CO and CO2 conversion than either Ni or Fe
alone.33 Another example of massive screening was provide by Greeley
et al., who screened over 700 binary alloys of metals and semi-metals for the
electrocatalytic hydrogen evolution reaction (HER; H+ + e - 1/2H2),34
based on the free energy of adsorption of H, DGH. The maximum HER
activity had been previously established around DGH = 0.35 Pareto analysis
based on DGH and stability criteria identied several promising candidates,
including a Pt-based BiPt surface alloy. Surprisingly, although Bi adsorbed
on Pt strongly poisons Pt, the BiPt surface alloy was ca. 50% more active
than pure Pt.
Linear energy relations also extend to metal compounds. Many such
compounds (MX) mediate the exchange of the non-metal elements (X) between
reactants and products via the Mars-Van Krevelen (MvK) mechanism
(e.g. partial oxidation catalyzed by metal oxides such as vanadia, molybdates,
and mixed oxides based thereon; hydrodesulfurization (HDS) catalyzed by
metal suldes and nitrides). A correlation between the reaction rate and
the strength of the MX bond has been considered a sign of the MvK
mechanism.36 As an example, kinetic modeling based on the linear energy
relation have nicely reproduced the volcano trend in the HDS activities of a
number of transition metal suldes when plotted as a function of calculated
DEMS.28 To further optimize the HDS catalysts, which are currently promoted
Mo and W suldes,37 structural databases of inorganic compounds have been
screened, and a number of promising candidates, some quite complex suldes,
have been identied in the desirable range of DEMS.38
Modifying the selectivity for a particular product is a more challenging
task. To understand why Ag is the most selective catalyst for ethylene
epoxidation, an highly important reaction practiced industrially for decades,
Linic et al. performed detailed spectroscopic and kinetic isotope experiments
and DFT calculations, and they concluded that the selectivity between the
partial and total oxidation of ethylene on Ag(111) is controlled by the relative
stability of two dierent transition states (TSs) that are both accessible to a
common oxametallacycle intermediate: One results in the closure of the
epoxide ring and ethylene oxide (EO), while the other leads to acetaldehyde
(AC) via intra-molecular H shift and eventually combustion.39 The authors

134 | Catalysis, 2009, 21, 131153


This journal is 
c The Royal Society of Chemistry 2009
computationally screened several Ag-based binary alloys and identied a
CuAg alloy which had a DEa = EAC EO
a Ea ca. 0.1 eV larger than on pure Ag
and should therefore be more selective toward EO. Experimental measure-
ments of several Cu-modied Ag catalysts showed up to 50% increase in
selectivity.40
Nrskov and co-workers tackled a dierent challenge when they identied
NiZn alloys to replace the more expensive PdAg catalysts that are currently
used to remove trace amounts of acetylene from ethylene via hydrogenation,
a necessary step prior to ethylene polymerization.41 It would be ideal if
the adsorption of ethylene could be made as weak as possible when
simultaneously the adsorption of acetylene as strong as possible, but calcula-
tions showed that the two quantities were correlated through the adsorption
energy of the methyl group and could not be independently varied. The task
then became to identify a less expensive material that would make a similar
compromise as PdAg does between desorbing ethylene (weak DEC2H4) and
hydrogenating acetylene (strong DEC2H2). Based on these criteria, NiZn was
picked for its low cost and predicted stability against segregation (Fig. 2). A
Ni25Zn75 alloy catalyst was veried in the laboratory to have even greater
selectivity than PdAg.
Linic and co-workers provided two additional examples of modifying
selectivity for the ethylene epoxidation reaction. They demonstrated the
promotional eect of Cs on EO formation using DFT calculations: Cs
atoms increase DEa by up to 0.2 eV vs. Ag only, via an induced electric eld
that interacts with the dierent dipoles of the two TSs.42 More recently
Christopher et al. synthesized and tested (100) facet-dominated Ag nano-
wire catalysts, based on the DFT results that DEa is ca. 0.1 eV larger
on Ag(100) than on Ag(111), because of the extra elongation of the
Ag-adsorbate bonds required to form the TS to AC on Ag(100). The Ag
nanowire catalysts were indeed more selective than conventional Ag cata-
lysts, in which Ag particles mainly exposes the (111) facet.43 Incidentally,

Fig. 2 2006 price of 70 binary intermetallic compounds plotted against the calculated methyl
binding energies. The left region represents low reactivity and high selectivity, and the right
region represents high reactivity and low selectivity, for acetylene hydrogenation. Adapted
from Studt et al., Science, 2008, 320, 1320.41 r 2008, Reprinted with permission from AAAS.

Catalysis, 2009, 21, 131153 | 135


This journal is 
c The Royal Society of Chemistry 2009
recently a theory-inspired new synthesis recipe enabled the production of
anatase TiO2 single crystals with a signicantly higher percentage of the
more reactive {001} facets than conventional polycrystalline anatase parti-
cles possess.44 Hydrouoric acid was used as morphology-controlling agent
because DFT calculations suggested that F atoms could make the surface
energy of the {001} facets lower than that of the {101} facets.
Theoretical work by Mavrikakis and co-workers provides further evidence
that nano-structured materials can open huge opportunities to heterogeneous
catalysis. Their DFT calculations suggested that the surface properties of Pt
could be tuned by laying just a layer of it on another metal. Zhang et al.
deposited single layers of Pt on the (111) or (0001) facets of Au, Ir, Pd, Rh,
and Ru and found that their activity for the oxygen reduction reaction
(ORR), the main reaction that takes place at the cathode of proton-exchange
membrane fuel cells (PEMFCs), showed a volcano dependence on the oxygen
binding energy (DEO), with Pt/Pd(111) being more active than Pt(111), the
most active monometallic catalyst for ORR. DFT calculations further
suggested that if a third metal alloyed into the Pt surface preferentially
attracted OH, which had been suspected to poison Pt under PEMFC
operation conditions, OHOH repulsion could free up Pt sites for ORR.45
Several ternary Pt/M/Pd(111) surfaces were synthesized and indeed showed
additional improvement in ORR activity over Pt/Pd(111) (Fig. 3). In a
separate study, Alayoglu et al. synthesized novel RuPt core-shell nano-
particles that were more active for the preferential oxidation (PROX) of CO
in H2 than either pure Pt or PtRu alloys.46 This was anticipated as DFT
calculations suggested that the Pt surfaces of these nano-particles promoted
COads + Oads - CO2 relative to Oads + Hads - OHads and were less
poisoned by CO than pure Pt, hence the higher PROX activity.
Neurock and co-workers ndings challenged material engineers when
they investigated alloying for enhancing NO decomposition to N2 by using
DFT-based kinetic Monte Carlo (kMC) simulations coupled with a bond
order conservation model.48 They screened a number of dierent PtxAu1x
surface alloys and identied Pt50Au50 as approximately the optimal

Fig. 3 Measured kinetic current densities at 0.80 V as a function of the calculated repulsion
energy between two OHs on (Pt3M)ML/Pd(111) (M = Au, Pt, Ir, Pd, Rh, and Ru) or between
OH and O (M = Re and Os). All energies are in eV. With kind permission from Springer
Science+Business Media: A. U. Nilekar et al., Topics in Catalysis, 2007, 46, 276, Figure 7.47

136 | Catalysis, 2009, 21, 131153


This journal is 
c The Royal Society of Chemistry 2009
composition. Furthermore, the turnover frequency depended sensitively on
the arrangement of the metal atoms: Square Pt4 and Au4 surface ensembles
did not improve over pure Pt, whereas + shaped Pt5 surface ensembles
dispersed among Au atoms were predicted to be twice as active. Although
their study did not suggest ways to achieve the desired surface ensembles of
atoms, in a more recent study Han et al. have determined experimentally the
preparation conditions that maximize the concentration of isolated Pd
monomer pairs in the PdAu(100) surface.49 Commercially Pd is alloyed
with a large amount of Au for the synthesis of vinyl acetate (VA) from the
oxidative coupling of acetic acid and ethylene. Both the reactants and
the product are prone to decompose on Pd.50,51 Diluting Pd prevents the
decomposition, but non-contiguous Pd monomer pairs are required for
catalyzing the coupling.5254

2.2 Stability considerations


Besides activity and selectivity, additional selection criteria have also been
articulated, in particular regarding the stability of catalytic materials
designed in silico based on mechanistic insights. Whether the desired
catalytic structure and composition are stable in the intended reaction
environment needs to be veried. Factors to be considered include whether
a component of a bulk alloy segregates to the surface, or whether de-mixing
and islanding occur in a surface alloy, because of entropic eects or the
adsorption of reactive species; whether surface or bulk species form; and
whether a catalytic material dissolves because of solvent, pH, and potential
eects in aqueous and electrochemical applications.34,55,56 Databases such
as binary alloy segregation energies57,58 and dissolution potentials of
alloys59 help more precisely and speedily identify practical new catalysts.

3. Molecular-level eects of reaction environment


During catalysis, catalysts come into intimate contact with a microscopic
world teeming with dierent chemical species (reactants, products, inter-
mediates, spectators, solvents) and dierent forms of energy (thermal,
electromagnetic). Their active states may be very dierent from as-prepared
or observed states in UHV, which often serve as the basis of modeling.
Researchers have long been aware of the signicant eects that the reaction
environment may have on the structure, chemical state, and reactivity of
heterogeneous catalysts, but the molecular origin of such eects is often
unclear. Indeed, terms such as pressure gap and materials gap have
been coined for cases where extrapolation of results obtained under low
pressure, fails to explain catalytic activity of the same catalyst under high
pressure. Enhanced experimental techniques tools have now begun to
generate a tremendous amount of new information about the nature of
many catalysts and catalytic reactions6066 in situ, and rst-principles
methods are playing an important role in interpreting the experimental
results and explaining environmental eects at the molecular level.

Catalysis, 2009, 21, 131153 | 137


This journal is 
c The Royal Society of Chemistry 2009
3.1 First-principles modeling of environmental eects
Given that quantum chemistry calculations directly provide electronic
energies, which formally correspond to zero temperature and pressure,
ways for connecting to nite, realistic temperature and pressure are needed.
One method is rst-principles thermodynamics (FPT), the basic concept of
which is that the thermodynamically prevailing state of a surface is the one
that minimizes the surface free energy, g, subject to external conditions such
as temperature and the chemical potentials of the various components of the
system:18,67
!
1 X
surface
gT; p; fNi g GG  Ni m i
A i

The exchange of material with the environment, e.g., via the adsorption of
gas-phase species or the segregation of a component into/from the bulk of a
mixture, is accounted for by including the chemical potentials (Nimi) of
the independent chemical species involved. The chemical potential of
e.g. adsorbed oxygen, can be related to that of an O2 gas, which can
be calculated from tabulated thermodynamic properties,18,67 through the
assumptions of chemical equilibrium between surface O and gas-phase O2.
The resulting information can be graphically represented by phase diagrams
like those shown in Fig. 4.
For its simplicity and exibility FPT has been applied to many systems
relevant to heterogeneous catalysis, including adsorbates ranging from simple
gasses such as hydrogen and oxygen, to water, H2S, and larger compounds
such as ethylene and acrolein, on metals,6975 metal compounds,67,7680 and
nite particles.8183 For adsorption systems, the FPT framework is used to
compare the thermodynamic stability of disparate but related structures
(e.g., chemisorption, surface reconstructions, surface and bulk compounds)
on the same basis, thereby predicting surface coverage and possible phase
change. FPT has also been used to provide a measure of the acidity/basicity
of metal oxides through the evaluation of the concentration of Lewis
(unsaturated metal centers) and Brnsted (surface hydroxyls) sites in thermal
and chemical equilibrium with water.84,85

Fig. 4 Normalized surface energy as a function of temperature at (a) 1012 atm and (b) 15 atm,
for various phases of O on Ag(111). (c) The most stable phases plotted in the (T, pO2) space.
Reprinted with permission from Michaelides et al., Chemical Physics Letters, 2003, 367, 344.68
r Elsevier.

138 | Catalysis, 2009, 21, 131153


This journal is 
c The Royal Society of Chemistry 2009
Grand canonical Monte Carlo (MC) simulations provide a more direct
and ecient way than the FPT-DFT approach to assess phase behavior on
a xed lattice and is particularly advantageous in handling phases with large
periodicity or multiple adsorbate species. Cluster expansion can be
combined with MC simulations to signicantly reduce the number of
rst-principles calculations required. The coupled-cluster approach enables
the simultaneous equilibration of adsorption and alloy segregation, which
has been used to directly demonstrate the oxygen-induced surface segrega-
tion on Pt(1x)Rux alloys86 and Ag3Pd(111).87 Currently there is no reliable
method for predicting phase transitions involving structural changes in
metals ab initio.

3.2 Local chemical environment and reactivity


As the temperature and partial pressures of various species in the environ-
ment change, the coverages of these species and their derivatives on the
catalytic surface will respond. Co-adsorbates can aect the activation barrier
of a rate-determining step. Honkala et al. have shown that the activation
barrier of N2 dissociation varies considerably with the local chemical
environment at the Ru step edge: H, N, or NHx species co-adsorbed in close
proximity of an N2 molecule variously increase its dissociation barrier.88 NO
oxidation is another good example: The overall NO + O2 reaction is
marginally exothermic and provides little thermodynamic driving force.
Adsorbed O atoms, which participate directly in the reaction, easily become
an energy sink until they are suciently destabilized at high coverage,
resulting in a weaker dierential binding energy that makes them more
reactive.75,89 Subsurface O atoms, which may be kinetically stabilized and
develop locally, have been shown to facilitate the dissociation of molecules
such as of H2, O2, and NO on Ag(111).90,91
Co-adsorbates may also open up dierent reaction channels. For NO
oxidation, Smeltz et al. have recently hypothesized a NO-assisted O2
activation pathway, via a peroxynitrite (NOOO) intermediate on Pt(111),
based on experimental and DFT evidence, which may be operative at high
surface coverage because of hindered O2 dissociation.92 In a combined
theoretical and experimental study, Andersson et al. have examined CO
dissociation on Ni surfaces over a range of CO and H2 pressures through
extensive DFT calculations and thermodynamic and kinetic modeling.93
They report that, while CO dissociates directly with a high barrier under
UHV conditions, the catalytic activities under high hydrogen pressure are
best explained by hydrogen-assisted dissociation through a COH inter-
mediate. Alayoglu et al. have likewise proposed a hydrogen-assisted O2
dissociation mechanism (via OOH) to explain O2 activation near room
temperature for the PROX of CO on the compressed, less reactive Pt
surfaces of their RuPt core-shell nano-particles.46 The in situ generation of
OOH or HOOH appears to be crucial to some selective oxidation processes,
including the direct gas-phase epoxidation of propylene over supported Au
catalysts.94,95 The promotional eect of moisture on the low-temperature
CO oxidation activity of Au nanoparticles96 may similarly be due to the
formation of an OOH species from O2 and water.97

Catalysis, 2009, 21, 131153 | 139


This journal is 
c The Royal Society of Chemistry 2009
The relevance of the alternate reaction channels depends on the probability
with which particular local chemical environments occur. The lowest barrier
pathway may not be the one that contributes most to the overall reactivity
under reaction conditions. For instance, Honkala et al. have reported that
near the entrance of a simulated ammonia synthesis reactor, the N2 dissocia-
tion pathway with the lowest barrier (Ea = 0.49 eV), in which an N2 molecule
has no nearest neighbors, contributes only about 40% as much to the overall
rate as N2 dissociating with a neighboring hydrogen atom (Ea = 0.59 eV).88
For CO oxidation on RuO2 (see section 3.3), Reuter et al. have similarly
reported that CO2 formation is dominated by the COcus + Ocus pathway
(Ea = 0.9 eV), not the lowest-barrier CObr + Ocus pathway (Ea = 0.8 eV),
which contributes about 1/3 as much as the former to the overall rate, at
steady state.98 Temel et al. have recently demonstrated discrepancies between
the results of mean-eld kinetic modeling and kinetic Monte Carlo, and
concluded that the neglect of uctuations in the probability of nding a
reactant pair located next to each other can produce an accumulative error in
the predicted reaction rates.99

3.3 Adsorbate-induced material change


The morphological and chemical changes that sometimes develop in the
surface of a catalyst under reaction conditions add another obstacle for
eorts to understand the actual action of the catalyst. Surface oxide phases,
for example, have been identied or suggested on all platinum-group metals
and noble metals,75,100109 many of which bear little resemblance to the
corresponding bulk oxides. Pd hydrides that form during hydrogenation
reactions remain poorly characterized.66 As examples, the Ru-oxygen and
Ag-oxygen systems, because of the substantial attention that they have
received in recent years, oer an in-depth look at how the interplay of
experiment and theory has advanced our understanding of these complex
systems, and where the limits of our knowledge of the microscopic world lie.
Ru, in both supported and single-crystal forms, was found to exhibit superior
CO oxidation activity to other Pt group metals when exposed to atmospheric O2
pressure,110 but much less active under UHV conditions.111 This unusual
phenomenon was originally explained by Goodman and co-workers in terms
of a full monolayer of adsorbed oxygen, Ru(0001)(1  1)O, that resulted a
loosely bound, reactive oxygen species, which was corroborated by DFT
results.112 However, Over and co-workers demonstrated that CO oxidation
on (1  1)O was orders of magnitude slower than on a Ru(0001) surface that
took up more than 3 ML of oxygen.113 Modeling eorts aimed at explaining the
location of the excess oxygen atoms led to the identication of a ORuO
trilayer, proposed to be the precursor to complete oxidation to RuO2.114 LEED
and STM experiments in conjunction with DFT calculations discovered
RuO2(110) in co-existence with O-covered Ru(0001),115 and autocatalytic
growth of RuO2 was suggested.116 First-principles thermodynamic analysis
conrmed bulk RuO2 to be thermodynamically stable when surface coverage
of adsorbed oxygen atoms exceeded ca. 0.75 ML.69 kMC simulations of CO
oxidation on RuO2(110) based entirely on DFT and statistical mechanics
closely reproduced the experimental reaction rates and identied COcus + Ocus

140 | Catalysis, 2009, 21, 131153


This journal is 
c The Royal Society of Chemistry 2009
as the dominant reaction pathway, as mentioned before.98 The understanding of
this reaction system, however, has continued to evolve. The more recent
experimental and theoretical evidence appear to suggest that the highest CO
oxidation rate occurs under conditions where a surface oxide phase coexists
with O-chemisorbed metallic Ru; or perhaps more precisely, where the structure
of the surface oscillates between the two phases.73,117119 Excess CO and
oxygen, which reduces or over-oxidizes the oxide phase respectively, are both
detrimental to CO oxidation activity.
The nature of the active Ag surface during the commercial ethylene
epoxidation process, which operates at ca. 500600 K and 1030 atm, remains
unclear, although the reaction mechanism has been elucidated in single-crystal
studies.39 A p(4  4) (p4) superstructure was rst identied by Rovida et al.
using LEED after exposing Ag(111) to O2 between 400500 K.120 It was not a
typical chemisorbed O structure and was suggested to be a single OAgO
trilayer.121 The rst STM images of the p4 structure were reported by Carlisle
et al. (Fig. 5).106 It appeared as bright features in a honeycomb arrangement,
which simulations suggested to be an Ag2O-like trilayer with 1/12 of the Ag
atoms missing, resulting in a stoichiometry of Ag1.83O. FPT-DFT studies by
both Michaelides et al. and Li et al. concluded that Ag1.83O was thermo-
dynamically most stable under a wide range of conditions (Fig. 3.1).68,122,123 In
addition, the experimentally deduced O content and measured Ag 3d core

Fig. 5 Montage image combining an STM image of the Ag oxide structure (from bottom left)
superimposed over the proposed oxide structure (from top right). The numbers, n = 15,
correspond to the symmetrically dierent positions within the middle silver layer sandwiched
between two O layers. Ag1 and Ag2 have metallic character, as they are exclusively bonded to
silver atoms in the substrate below, whereas Ag3, Ag4, and Ag5 are directly bonded to oxygen
inside the oxide rings and are ionic in nature. Both Ag4 and Ag5 sites sit above threefold sites of
the underlying (111) lattice atoms, whereas Ag3 occupies a top site. Reprinted with permission
from Bocquet et al., Journal of the American Chemical Society, 2003, 125, 3119.65 r 2003,
American Chemical Society.

Catalysis, 2009, 21, 131153 | 141


This journal is 
c The Royal Society of Chemistry 2009
level shift were also consistent with the Ag1.83O model.121,124 The adsorption
and epoxidation of ethylene on the Ag1.83O trilayer was subsequently
modeled.65,125,126 A process via an oxametallacycle was found to operate
on both Ag(111) and Ag1.83O with similar overall activation barriers. It
was concluded that both structures were eective catalysts and that
ethylene epoxiation could occur over a wide range of oxygen potential.125
The pressure and materials gaps were apparently closed for the Ag
epoxidation catalyst.
Michaelides et al., however, soon cast doubts on the Ag1.83O model after
re-consideration of the existing evidence and new DFT results.127 They
pointed out that the Ag1.83O trilayer was a meta-stable structure and should
not be able to cover large surface areas with little defect. More recently,
several experimental and theoretical studies appeared that threaten to
invalidate the Ag1.83O model altogether. Schnadt et al. obtained new
STM images of the p4 structure at higher resolutions, which contradicted
the original assignment of the bright features to individual Ag atoms.107
Instead, through the comparison of the p4 structure with coincident
p(4  5O3)rect and c(3  5O3)rect structures (Fig. 6), and through
simulations of a series of plausible models, the authors concluded that the
bright features were in fact triangular Ag6 units with O atoms intercalated
between adjacent units. Though having a stoichiometry of Ag2O, the new
model bore no resemblance to bulk Ag2O and could not even be strictly
classied as oxides. An independent and concurrent study by Schmid et al.,

Fig. 6 STM images and partial structural models of the Ag-oxygen overlayers. (a) A 200 A2 patch
of Ag(111) covered by the p(4  4), c(3  5O3)rect, and p(4  5O3)rect overlayers, measured at
0.51 nA and 131.5 mV. (b) The Ag1.83O model for the p(4  4) proposed by Carlisle et al.106
Solid (open) large gray balls represent the overlayer (substrate) Ag atoms, and small balls the
oxygen atoms. (c, e) STM images of 35 A2 patches of the p(4  4) and c(3  5O3)rect over-
layers, respectively. (c) is at 0.42 nA and 21.7 mV, and (e) is at 0.40 nA and 34.2 mV.
The inset in (e) displays the six-atom structural element of the c(3  5O3)rect phase measured
with a tip state that allowed the resolution of the central parts of the proposed Ag6 triangles
(0.42 nA, 21.7 mV). (d, f) The proposed Ag6-based models for the p(4  4) and c(3  5O3)rect
phases, respectively. Figure adapted with permission from Schnadt et al., Physical Review Letters,
2006, 96, 146101.107 r American Physical Society.

142 | Catalysis, 2009, 21, 131153


This journal is 
c The Royal Society of Chemistry 2009
through painstaking STM, sXRD, CLS, and DFT work, also concluded
that models based on the same Ag6 units t the data best.108 They pointed
out that the calculated energies of the new and old models were suciently
close that it was beyond the limit of the current mainstream DFT formula-
tions to distinguish which was the most stable. Van der Waals-type
interactions between Ag atoms, which were neglected by the local (LDA)
and semi-local (GGA) density functionals, were suggested to stabilize the
Ag6 models vs. the more open oxides.
Like CO oxidation on Ru, the understanding for ethylene epoxidation on
Ag has continued to evolve. Many questions remain open, including the
reaction mechanism on the Ag6 structures, and the role of intercalated
oxygen atoms. Another dimension that is little explored so far is the surface
states in a combined oxygen-ethylene atmosphere. Greeley et al. have
reported recently that an ethylenedioxy intermediate may be present at
appreciable coverage under industrial reaction conditions,72 the eect of
which on the structure of the surface is unknown. More importantly, the
implication of a dynamic co-existence of various surface oxides under
reaction conditions107 for the reaction mechanism needs to be explored
and understood at greater depth.
As alluded to before, the adsorption of atoms and molecules may also
induce segregation in alloys. Upon revisiting the thermodynamic behavior
of the improved CuAg alloy catalysts for ethylene epoxidation synthesized
by Linic et al., (section 2.1) Piccinin et al. calculated that, while in the
absence of oxygen Cu prefers to stay in the subsurface layers, oxygen
adsorption causes it to segregate to the surface which then phase-separates
into clean Ag(111) and various Cu surface oxides under typical industrial
conditions (Fig. 7).128 This casts doubt on the active state of the previous
Cu-Ag catalysts being a well-mixed surface AgCu alloy.

Fig. 7 Surface phase diagram indicating the most stable surface structures for CuAg
bimetallic surface as a function of the Cu surface content and oxygen chemical potential.
Reprinted gure with permission from Piccinin et al., Physical Review B, 2008, 77, 075426.128
r American Physical Society.

Catalysis, 2009, 21, 131153 | 143


This journal is 
c The Royal Society of Chemistry 2009
3.4 Electrochemical environment
Several important energy-related applications, including hydrogen produc-
tion, fuel cells, and CO2 reduction, have thrust electrocatalysis into the
forefront of catalysis research recently. Electrocatalysis involves several
physiochemical environmental eects,129,130 which poses substantial
challenges for the theoreticians. First, there is the electric potential which
can aect the thermodynamics of the system and the kinetics of the electron
transfer reactions. The electrolyte, which is usually aqueous, contains water
and ions that can interact directly with a surface and charged/polar
adsorbates, and indirectly with the charge in the electrode to form the
electrochemical double layer, which sets up an electric eld at the interface
that further aects interfacial reactivity.
One approach to accounting for the potential and solution eects is to
treat them separately (the additive approach). The thermodynamics due to
an electric potential is described by the Nernst equation, which is, to the rst
order of approximation,131,132

DG(U) = DG(U0) + ne(U  U0),

for an n-electron transfer step at an electrode potential of U. With the


standard hydrogen electrode (SHE) as reference, DG for the reaction
1/2H2 - H+ + e is zero at U0 = 0 V, T = 298 K, pH2 = 1 bar, and
solution pH = 0, which facilitates the treatment of protonation + electron
transfer steps. Deviation from zero pH is corrected by kT ln[H+], and the
water-metal/adsorbate interaction is modeled by water bilayer(s).132,133 The
electric eld due to the double layer can be directly modeled in DFT.134
Neurock and co-workers have developed a more elaborate approach, the
double reference method, that simultaneously takes the eects of surface
charging, interfacial electric eld, and water into account.135 The periodic
metal slab model is augmented with a certain number of added or
subtracted electrons, a compensating background charge spread uniformly
throughout the simulation cell, and water at bulk density lling the inter-
slab space. This approach hinges on the assumption that the potential of a
charged metal-water interface is determined by the Fermi level of the slab
when properly referenced. The potential of the charged interface is
rst referenced to that of the same interface but uncharged, which is
then referenced to the vacuum level by inserting a vacuum layer in the
water region. This technique allows the potential of the charged interface to
be referenced to the vacuum level and in turn to SHE. To directly compare
dierently charged slabs,135 the as-calculated energy (EDFT) needs to be
corrected:

Zq
E EDFT 
hVidQ qU:
0

 is the volume-average electrostatic potential in the simulation cell


Here hVi
that arises from the metalwater interface and the background charge, and

144 | Catalysis, 2009, 21, 131153


This journal is 
c The Royal Society of Chemistry 2009
q is the number of excess electrons added. This E amounts to a potential-
dependent free energy.
The kinetics of electron-transfer reactions, which is also aected by the
electrode potential and the metalwater interface, is more dicult and complex
to treat than the thermodynamic aspects. While the theoretical development for
electron transfer kinetics began decades ago,136 a practical implementation for
surface reactions is still unavailable. Popular transition state-searching techni-
ques such as the NEB method16 are not designed to search for minimum-energy
reaction paths subject to a constant potential. Approximations that allow
aordable quantum chemistry calculations to get around this limitation have
been proposed, ranging from the electron anity/ionization potential matching
method137,138 to heuristic arguments based on interpolations.130,132
Electrochemical activation of water is of prime importance because
aqueous solutions are the most common electrolytes. Filhol et al. applied
the double reference method to study the interaction of water with
Pd(111).139 They calculated the potential-dependent free energies of three
phases: un-dissociated water, OHads plus solvated protons, and Hads plus
solvated hydroxides, and used them to construct phase diagrams that
predicted that adsorbed water would dissociate to form Hads at potentials
(vs. SHE) more cathodic than 0.5 V and form OHads at potentials more
anodic than 1.1 V, in close agreement with available experimental results
(transitions at 0.4 V and 0.70.9 V, respectively). On Cu(111), suciently
high potentials were found to cause the dissolution of Cu2+ ions and the
formation of a passivating Cu2O overlayer.130 Rossmeisl et al. demon-
strated that the additive approach and the double reference method
produced very similar phase diagrams for water interacting with Pt(111)
(Fig. 8), that agreed with the reversible potential for OH adsorption of 0.8 V
from experiment.140 Hansen et al. since extended the potential-dependent
phase diagrams to 2D potential-and-pH phase diagrams for water on
Pt(111), Ag(111), and Ni(111).141

Fig. 8 Phase diagram showing the free energy of dierent surface structures of water at
pH = 0 in contact with Au(111), Pt(111), and Ni(111). The lowest line represents the
thermochemically most stable phase. The crossing of the two bottom lines indicates a phase
change. Reprinted with permission from Rossmeisl et al., Journal of Physical Chemistry B,
2006, 110, 21833.140 r 2006, American Chemical Society.

Catalysis, 2009, 21, 131153 | 145


This journal is 
c The Royal Society of Chemistry 2009
The phase behavior of water is directly relevant to the oxygen reduction
reaction (ORR; O2 + 4H+ + 4e - 2H2O), another frequently studied
electrocatalytic reaction. The interest in ORR stems from the fact that it suers
from large overpotential loss even on the best monometallic catalyst, Pt, which
necessitates high Pt loadings to compensate. Intense eort has been spent to
understand the mechanism of the ORR and the origin of the overpotential, and
to nd a lower-cost alternative to pure Pt catalysts. Nrskov et al. performed a
kinetic analysis of the ORR on Pt and other metals using the additive approach
and demonstrated that proton transfer to Oads or OHads is rate-limiting at
relevant potentials.132 More recent work by the same group included electric
eld eects into consideration and found that Pt is located near the transition
between the O2 activation-limited regime (Ag and Au) and the protonation-
limited regime (Ir, Rh, Ru, etc.). Furthermore, on metals that bind O less
strongly than Pt, e.g. Ag and Au, high potential and low surface coverage favor
the dissociative mechanism (O2 - 2O), whereas the opposite conditions favor
the associative mechanism (O2 + H+ + e - OOH).134 These results not
only conrmed that oxygenates derived from O2 and water inhibit ORR
kinetics142 but also revealed the complexity of this seemingly simple reaction.
Other reactions that are directly relevant to PEMFCs, including the electro-
catalytic oxidation of methanol and CO, have likewise been studied taking
potential and solution eects into consideration.130,143

4. Outlook
Impressive progress in heterogeneous catalysis has been made with the help
of theory and rst-principles methods. The PdAu alloy catalyst created by
Goodman and co-workers for vinyl acetate synthesis represents an encoura-
ging trend in which mechanistic insights from experiment and theory merge
to become a general way of approaching catalysis research. Tremendous
opportunities as well as challenges lie ahead. Take the ab initio design of
catalysts as an example. The successes to date have largely focused on
optimizing a single step, and have been made possible in part by the
predictability of the geometric and electronic structures of metals and their
alloys. More complex reactions and applications will demand the simulta-
neous consideration of multiple descriptors. For instance, Janik et al. have
identied a set of criteria for designing improved anode catalysts for direct
methanol fuel cells.130 These include preferential breaking of CH vs. OH
bonds, weakened binding of CO vs. Pt, ability to activate water and oxidize
CO, and comparable rate of methanol oxidation to CO to Pt. Their
calculations suggest the ternary surface alloy PtRuAu/Ru(0001) as a
promising candidate, which remains to be tested. The complexity in design
variables and the amount of data that theory will generate may one day rise
to a level where they pose a challenging optimization and organization
problem.
A comparable paradigm of success is still to be established for more
complex catalytic materials (e.g., nite metal/compound clusters, mixed
oxides, zeolites), many of which are highly important technological catalysts
and often contain multiple active centers that are required to achieve a series
of specic transformations. The local structures of these catalysts and

146 | Catalysis, 2009, 21, 131153


This journal is 
c The Royal Society of Chemistry 2009
pertinent reaction mechanisms are being elucidated with the help of
theory.144146 The kind of surprise that local nanometer-scale structures
can hold is illustrated by the MoS2 HDS catalysts and supported size-
selected metal clusters. The active MoS2 particles consist of single layers of
MoS2 exposing the S-terminated (0001) basal plane, usually thought to be
inactive. However, thiophene and its hydrogenated derivatives adsorb on
this plane because of unexpected metallic edges (Fig. 9).147,148 Size-selected
metal clusters containing just a handful of atoms (e.g. Aun,149 Ptn,150 and Irn151)
supported on oxides and zeolites) have been shown to exhibit catalytic
activity that can change markedly with one more or fewer atom. Metal
aggregates in this size regime are no longer metallic, and unique size-
dependent electronic structure and cluster-support interaction strongly
inuence their reactivity.152,153 Indeed, nanomaterials, whose structures
are well-dened at the nanometer or angstrom scale, and whose properties
are functions of potentially a large number of adjustable parameters, give
substantial freedom to create catalytic reactivity. However, much work
remains to be done to understand and control the relationship between
structure and reactivity. This is where theory can be expected to play an
increasingly important role.
Dynamic eects are a potentially important but easily overlooked aspect
of heterogeneous catalysis that can nonetheless impact the accuracy of our
knowledge and predictions. For example, multiple co-existing meta-stable
surface oxide phases have been identied for Pd and Ag interacting with
oxygen, which suggests that the catalyst surfaces may be in a state of ux
under reaction conditions, adding new uncertainty to the nature of the
active species.103,107 In their analysis of CO oxidation on Ru and Pd,
Scheer and co-workers suggest that the state of the surface should be
expected to uctuate readily under the conditions that give the highest
activity.73,98 Other transient phases, such as host metal atoms being
transported across the surface, may have a catalytic role as well. One

Fig. 9 STM image (48  53 A2; Vt = 5.3 mV; It = 1.28 nA) showing a MoS2 nanocluster
synthesized on a reconstructed Au(111) surface. The bright brim extending all the way around
the edge is due to localized metallic edge states. Reprinted gure with permission from Bollinger
et al., Physical Review Letters, 2001, 87, 196803.147 r American Physical Society.

Catalysis, 2009, 21, 131153 | 147


This journal is 
c The Royal Society of Chemistry 2009
obstacle for understanding dynamic behavior of catalysts is the lack of an
ecient and reliable method for predicting structural transition ab initio.
Such a method would have to tackle the challenge of locating the global
minimum and other low-lying minima on a free energy surface in the grand
canonical ensemble, without prior knowledge of structure, composition, or
periodicity. A promising technique is the adaptive kinetic Monte Carlo
proposed by Henkelman et al.,154,155 which uses the dimer method to
discover reaction pathways on the y, without the use of a pre-determined
event table. For the time being, researchers must rely on meticulous
experiment (which may or may not have the necessary in situ capability
or resolution) and computer modeling to tease out the intimate roles of the
catalysts in the reactions that they catalyze.
All in all, the continued growth of computer power and development
of theory and its algorithmic implementation will enable more physics
(e.g. van der Waals interactions, electron self-interaction, entropic eects)
to be correctly captured and enable ultimately more speedy, accurate, and
comprehensive (multi-scale and multi-phasic) simulations. Theory will be an
ever more indispensible ally of experiment and continue to supply detail
knowledge on molecular transformations, thereby propelling the growth of
heterogeneous catalysis and enriching our world in years to come.

Acknowledgements
The author acknowledges support by Oak Ridge National Laboratorys
Center for Nanophase Materials Sciences, which is sponsored by the
Scientic User Facilities Division, Oce of Basic Energy Sciences, US
Department of Energy.

References
1 M. W. Roberts, Catalysis Letters, 2000, 67, 1.
2 D. Filmore, Todays Chemist At Work, 11/2002, 33.
3 J. H. Sinfelt, Bimetallic Catalysts: Discoveries, Concepts, and Applications,
Wiley, New York, 1983.
4 C. N. Sattereld, Heterogeneous Catalysis in Industrial Practice, McGraw-Hill,
1991.
5 J.-D. Grunwaldt and A. Baiker, Physical Chemistry Chemical Physics, 2005, 7,
3526.
6 MRS Bulletin, 2007, 32.
7 S. Baker, J. S. Robinson, C. A. Haworth, H. Teng, R. A. Smith, C. C. Chirila,
M. Lein, J. W. G. Tisch and J. P. Marangos, Science, 2006, 312, 424.
8 M. B. J. Roeaers, B. F. Sels, H. Uji-i, F. C. De Schryver, P. A. Jacobs,
D. E. De Vos and J. Hofkens, Nature, 2006, 439, 572.
9 J. P. Perdew, J. A. Chevary, S. H. Vosko, K. A. Jackson, M. R. Pederson,
D. J. Singh and C. Fiolhais, Physical Review B, 1992, 46, 6671.
10 J. P. Perdew, K. Burke and M. Ernzerhof, Physical Review Letters, 1996, 77,
3865.
11 D. R. Hamann, M. Schluter and C. Chiang, Physical Review Letters, 1979, 43,
1494.
12 D. H. Vanderbilt, Physical Review B, 1990, 41, 7892.
13 P. E. Blochl, Physical Review B, 1994, 50, 17953.
14 G. Kresse and D. Joubert, Physical Review B, 1999, 59, 1758.

148 | Catalysis, 2009, 21, 131153


This journal is 
c The Royal Society of Chemistry 2009
15 G. Henkelman and H. Jonsson, Journal of Chemical Physics, 1999, 111, 7010.
16 G. Henkelman, B. P. Uberuaga and H. Jonsson, Journal of Chemical Physics,
2000, 113, 9901.
17 P. Baron, H. Andreas, A. T. Bell and A. Chakraborty, Journal of Chemical
Physics, 2004, 120, 7877.
18 K. Reuter and M. Scheer, Physical Review B, 2002, 65, 035406.
19 R. D. Cortright and J. A. Dumesic, in Advances in Catalysis, 2002, vol. 46,
p. 161.
20 M. Neurock and D. H. Mei, Topics in Catalysis, 2002, 20, 5.
21 B. Hammer and J. K. Nrskov, Surface Science, 1995, 343, 211.
22 E. Hansen and M. Neurock, Surface Science, 1999, 441, 410.
23 H. Sellers and E. Shustorovich, Surface Science, 2002, 504, 167.
24 F. Abild-Pedersen, J. Greeley, F. Studt, J. Rossmeisl, T. R. Munter, P. G. Moses,
E. Skulason, T. Bligaard and J. K. Nrskov, Physical Review Letters, 2007, 99,
016105.
25 V. Pallassana and M. Neurock, Journal of Catalysis, 2000, 191, 301.
26 J. K. Nrskov, et al., Journal of Catalysis, 2002, 209, 275.
27 A. Michaelides, Z.-P. Liu, C. J. Zhang, A. Alavi, D. A. King and P. Hu,
Journal of the American Chemical Society, 2003, 125, 3704.
28 H. Toulhoat and P. Raybaud, Journal of Catalysis, 2003, 216, 63.
29 T. Bligaard, J. K. Nrskov, S. Dahl, J. Matthiesen, C. H. Christensen and
J. Sehested, Journal of Catalysis, 2004, 224, 206.
30 C. J. H. Jacobsen, S. Dahl, B. S. Clausen, S. Bahn, A. Logadottir and
J. K. Nrskov, Journal of the American Chemical Society, 2001, 123, 8404.
31 A. Logadottir, T. H. Rod, J. K. Nrskov, B. Hammer, S. Dahl and
C. J. H. Jacobsen, Journal of Catalysis, 2001, 197, 229.
32 M. P. Andersson, T. Bligaard, A. Kustov, K. E. Larsen, J. Greeley,
T. Johannessen, C. H. Christensen and J. K. Nrskov, Journal of Catalysis,
2006, 239, 501.
33 J. Sehested, K. E. Larsen, A. L. Kustov, A. M. Frey, T. Johannessen,
T. Bligaard, M. P. Andersson, J. K. Nrskov and C. H. Christensen, Topics
in Catalysis, 2007, 45, 9.
34 J. Greeley, T. F. Jaramillo, J. Bonde, I. B. Chorkendor and J. K. Nrskov,
Nature Materials, 2006, 5, 909.
35 R. Parsons, Transactions of the Faraday Society, 1958, 54, 1053.
36 C. Doornkamp and V. Ponec, Journal of Molecular Catalysis A: Chemical,
2000, 162, 19.
37 P. Raybaud, Applied Catalysis A-General, 2007, 322, 76.
38 J. K. Nrskov, M. Scheer and H. Toulhoat, MRS Bulletin, 2006, 31, 669.
39 S. Linic and M. A. Barteau, Journal of the American Chemical Society, 2003,
125, 4034.
40 S. Linic, J. Jankowiak and M. A. Barteau, Journal of Catalysis, 2004, 224, 489.
41 F. Studt, F. Abild-Pedersen, T. Bligaard, R. Z. Sorensen, C. H. Christensen
and J. K. Nrskov, Science, 2008, 320, 1320.
42 S. Linic and M. A. Barteau, Journal of the American Chemical Society, 2004,
126, 8086.
43 P. Christopher and S. Linic, Journal of the American Chemical Society, 2008,
130, 11264.
44 H. G. Yang, C. H. Sun, S. Z. Qiao, J. Zou, G. Liu, S. C. Smith, H. M. Cheng
and G. Q. Lu, Nature, 2008, 453, 638.
45 J. Zhang, M. B. Vukmirovic, K. Sasaki, A. U. Nilekar, M. Mavrikakis and
R. R. Adzic, Journal of the American Chemical Society, 2005, 127, 12480.

Catalysis, 2009, 21, 131153 | 149


This journal is 
c The Royal Society of Chemistry 2009
46 S. Alayoglu, A. U. Nilekar, M. Mavrikakis and B. Eichhorn, Nature Materials,
2008, 7, 333.
47 A. U. Nilekar, Y. Xu, J. Zhang, M. B. Vukmirovic, K. Sasaki, R. R. Adzic and
M. Mavrikakis, Topics in Catalysis, 2007, 46, 276.
48 L. D. Kieken, M. Neurock and D. H. Mei, Journal of Physical Chemistry B,
2005, 109, 2234.
49 P. Han, S. Axnanda, I. Lyubinetsky and D. W. Goodman, Journal of the
American Chemical Society, 2007, 129, 14355.
50 Y. F. Han, D. Kumar, C. Sivadinarayana, A. Cleareld and D. W. Goodman,
Catalysis Letters, 2004, 94, 131.
51 F. Calaza, D. Stacchiola, M. Neurock and W. T. Tysoe, Surface Science, 2005,
598, 263.
52 M. Neurock, Journal of Catalysis, 2003, 216, 73.
53 D. Stacchiola, F. Calaza, L. Burkholder, A. W. Schwabacher, M. Neurock
and W. T. Tysoe, Angewandte Chemie-International Edition, 2005, 44, 4572.
54 M. S. Chen, D. Kumar, C. W. Yi and D. W. Goodman, Science, 2005, 310,
291.
55 J. Greeley and M. Mavrikakis, Nature Materials, 2004, 3, 810.
56 Z. H. Gu and P. B. Balbuena, Journal of Physical Chemistry A, 2006, 110,
9783.
57 A. Christensen, A. V. Ruban, P. Stoltze, K. W. Jacobsen, H. L. Skriver,
J. K. Nrskov and F. Besenbacher, Physical Review B, 1997, 56, 5822.
58 A. V. Ruban, H. L. Skriver and J. K. Nrskov, Physical Review B, 1999, 59,
15990.
59 J. Greeley and J. K. Nrskov, Electrochimica Acta, 2007, 52, 5829.
60 H. Over, Y. D. Kim, A. P. Seitsonen, S. Wendt, E. Lundgren, M. Schmid,
P. Varga, A. Morgante and G. Ertl, Science, 2000, 287, 1474.
61 B. L. M. Hendriksen, S. C. Bobaru and J. W. M. Frenken, Surface Science,
2004, 552, 229.
62 B. L. M. Hendriksen and J. W. M. Frenken, Physical Review Letters, 2002, 89,
046101.
63 M. D. Ackermann, et al., Physical Review Letters, 2005, 95, 255505.
64 M. Baumer, J. Libuda, K. M. Neyman, N. Rosch, G. Rupprechter and
H. J. Freund, Physical Chemistry Chemical Physics, 2007, 9, 3541.
65 M. L. Bocquet, P. Sautet, J. Cerda, C. I. Carlisle, M. J. Webb and D. A. King,
Journal of the American Chemical Society, 2003, 125, 3119.
66 M. Morkel, G. Rupprechter and H. J. Freund, Surface Science, 2005, 588,
L209.
67 M. V. Bollinger, K. W. Jacobsen and J. K. Nrskov, Physical Review B, 2003,
67, 085410.
68 A. Michaelides, M. L. Bocquet, P. Sautet, A. Alavi and D. A. King, Chemical
Physics Letters, 2003, 367, 344.
69 C. Stamp, Catalysis Today, 2005, 105, 17.
70 F. Faglioni and W. A. Goddard, Journal of Chemical Physics, 2005, 122,
014704.
71 D. Loreda, F. Delbecq and P. Sautet, Chemical Physics Letters, 2005, 405,
434.
72 J. Greeley and M. Mavrikakis, Journal of Physical Chemistry C, 2007, 111,
7992.
73 J. Rogal, K. Reuter and M. Scheer, Physical Review Letters, 2007, 98,
046101.
74 H. Q. Shi and C. Stamp, Physical Review B, 2008, 77, 094127.

150 | Catalysis, 2009, 21, 131153


This journal is 
c The Royal Society of Chemistry 2009
75 R. B. Getman, Y. Xu and W. F. Schneider, Journal of Physical Chemistry C,
2008, 112, 9559.
76 Z. Lodziana, J. K. Nrskov and P. Stoltze, Journal of Chemical Physics, 2003,
118, 11179.
77 B. Meyer, Physical Review B, 2004, 69, 045416.
78 M. V. Ganduglia-Pirovano and J. Sauer, Physical Review B, 2004, 70, 045422.
79 C. S. Lo, K. S. Tanwar, A. M. Chaka and T. P. Trainor, Physical Review B,
2007, 75, 075425.
80 E. Krebs, B. Silvi and P. Raybaud, Catalysis Today, 2008, 130, 160.
81 Y. Xu, W. A. Shelton and W. F. Schneider, Journal of Physical Chemistry B,
2006, 110, 16591.
82 S. Laursen and S. Linic, Physical Review Letters, 2006, 97, 026101.
83 J. G. Wang and B. Hammer, Topics in Catalysis, 2007, 44, 49.
84 M. Digne, P. Sautet, P. Raybaud, P. Euzen and H. Toulhoat, Journal of
Catalysis, 2004, 226, 54.
85 C. Chizallet, G. Costentin, M. Che, F. Delbecq and P. Sautet, Journal of
Physical Chemistry B, 2006, 110, 15878.
86 B. C. Han, A. Van der Ven, G. Ceder and B.-J. Hwang, Physical Review B,
2005, 72, 205409.
87 J. R. Kitchin, K. Reuter and M. Scheer, Physical Review B, 2008, 77, 075437.
88 K. Honkala, A. Hellman, I. N. Remediakis, A. Logadottir, A. Carlsson, S. Dahl,
C. H. Christensen and J. K. Nrskov, Science, 2005, 307, 555.
89 S. Ovesson, B. I. Lundqvist, W. F. Schneider and A. Bogicevic, Physical
Review B, 2005, 71, 115406.
90 Y. Xu, J. Greeley and M. Mavrikakis, Journal of the American Chemical
Society, 2005, 127, 12823.
91 A. B. Mohammad, I. V. Yudanov, K. H. Lim, K. M. Neyman and N. Rosch,
Journal of Physical Chemistry C, 2008, 112, 1628.
92 A. D. Smeltz, R. B. Getman, W. F. Schneider and F. H. Ribeiro, Catalysis
Today, 2008, 136, 84.
93 M. P. Andersson, et al., Journal of Catalysis, 2008, 255, 6.
94 T. V. Choudhary and D. W. Goodman, Topics in Catalysis, 2002, 21, 25.
95 A. M. Joshi, W. N. Delgass and K. T. Thomson, Journal of Physical Chemistry
B, 2005, 109, 22392.
96 H. H. Kung, M. C. Kung and C. K. Costello, Journal of Catalysis, 2003, 216,
425.
97 A. Bongiorno and U. Landman, Physical Review Letters, 2005, 95, 106102.
98 K. Reuter, D. Frenkel and M. Scheer, Physical Review Letters, 2004, 93,
116105.
99 B. Temel, H. Meskine, K. Reuter, M. Scheer and H. Metiu, Journal of
Chemical Physics, 2007, 126, 204711.
100 J. Gustafson et al., Physical Review Letters, 2004, 92, 126102.
101 H. Zhang, A. Soon, B. Delley and C. Stamp, Physical Review B, 2008, 78,
045436.
102 E. Lundgren, J. Gustafson, A. Mikkelsen, J. N. Andersen, A. Stierle, H. Dosch,
M. Todorova, J. Rogal, K. Reuter and M. Scheer, Physical Review Letters,
2004, 92, 046101.
103 J. Klikovits, E. Napetschnig, M. Schmid, N. Seriani, O. Dubay, G. Kresse and
P. Varga, Physical Review B, 2007, 76, 045405.
104 N. Seriani, W. Pompe and L. C. Ciacchi, Journal of Physical Chemistry B,
2006, 110, 14860.
105 A. Soon, M. Todorova, B. Delley and C. Stamp, Physical Review B, 2006, 73,
165424.

Catalysis, 2009, 21, 131153 | 151


This journal is 
c The Royal Society of Chemistry 2009
106 C. I. Carlisle, D. A. King, M. L. Bocquet, J. Cerda and P. Sautet, Physical
Review Letters, 2000, 84, 3899.
107 J. Schnadt, A. Michaelides, J. Knudsen, R. T. Vang, K. Reuter, E. Laegsgaard,
M. Scheer and F. Besenbacher, Physical Review Letters, 2006, 96, 146101.
108 M. Schmid, et al., Physical Review Letters, 2006, 96, 146102.
109 H. Q. Shi and C. Stamp, Physical Review B, 2007, 76, 075327.
110 C. H. F. Peden and D. W. Goodman, Journal of Physical Chemistry, 1986, 90,
1360.
111 H. I. Lee and J. M. White, Journal of Catalysis, 1980, 63, 261.
112 C. Stamp and M. Scheer, Physical Review Letters, 1997, 78, 1500.
113 A. Bottcher, H. Niehus, S. Schwegmann, H. Over and G. Ertl, Journal of
Physical Chemistry B, 1997, 101, 11185.
114 K. Reuter, M. V. Ganduglia-Pirovano, C. Stamp and M. Scheer, Physical
Review B, 2002, 65, 165403.
115 H. Over and M. Muhler, Progress in Surface Science, 2003, 72, 3.
116 H. Over, A. P. Seitsonen, E. Lundgren, M. Schmid and P. Varga, Surface
Science, 2002, 515, 143.
117 J. Assmann, et al., Angewandte Chemie-International Edition, 2005, 44, 917.
118 R. Blume, M. Havecker, S. Zafeiratos, D. Teschner, E. Kleimenov,
A. Knop-Gericke, R. Schlogl, A. Barinov, P. Dudin and M. Kiskinova,
Journal of Catalysis, 2006, 239, 354.
119 H. Over, M. Muhler and A. P. Seitsonen, Surface Science, 2007, 601, 5659.
120 G. Rovida, F. Pratesi, M. Maglietta and E. Ferroni, Surface Science, 1974, 43,
230.
121 C. T. Campbell, Surface Science, 1985, 157, 43.
122 W. X. Li, C. Stamp and M. Scheer, Physical Review B, 2003, 68, 165412.
123 W. X. Li, C. Stamp and M. Scheer, Physical Review Letters, 2003, 90,
256102.
124 V. I. Bukhtiyarov, V. V. Kaichev and I. P. Prosvirin, Journal of Chemical
Physics, 1999, 111, 2169.
125 M. L. Bocquet, A. Michaelides, D. Loreda, P. Sautet, A. Alavi and D. A. King,
Journal of the American Chemical Society, 2003, 125, 5620.
126 M. L. Bocquet and D. Loreda, Journal of the American Chemical Society,
2005, 127, 17207.
127 A. Michaelides, K. Reuter and M. Scheer, Journal of Vacuum Science &
Technology A, 2005, 23, 1487.
128 S. Piccinin, C. Stamp and M. Scheer, Physical Review B, 2008, 77, 075426.
129 C. D. Taylor and M. Neurock, Current Opinion in Solid State & Materials
Science, 2005, 9, 49.
130 M. J. Janik, C. D. Taylor and M. Neurock, Topics in Catalysis, 2007, 46, 306.
131 R. M. Roques and A. B. Anderson, Journal of the Electrochemical Society,
2004, 151, E85.
132 J. K. Nrskov, J. Rossmeisl, A. Logadottir, L. Lindqvist, J. R. Kitchin,
T. Bligaard and H. Jonsson, Journal of Physical Chemistry B, 2004, 108, 17886.
133 J. G. Wang and B. Hammer, Journal of Chemical Physics, 2006, 124, 184704.
134 G. S. Karlberg, J. Rossmeisl and J. K. Nrskov, Physical Chemistry Chemical
Physics, 2007, 9, 5158.
135 C. D. Taylor, S. A. Wasileski, J. S. Filhol and M. Neurock, Physical Review B,
2006, 73, 165402.
136 R. A. Marcus, Journal of Chemical Physics, 1956, 24, 966.
137 R. A. Sidik and A. B. Anderson, Journal of Electroanalytical Chemistry, 2002,
528, 69.

152 | Catalysis, 2009, 21, 131153


This journal is 
c The Royal Society of Chemistry 2009
138 A. B. Anderson, J. Roques, S. Mukerjee, V. S. Murthi, N. M. Markovic and
V. Stamenkovic, Journal of Physical Chemistry B, 2005, 109, 1198.
139 J. S. Filhol and M. Neurock, Angewandte Chemie-International Edition, 2006,
45, 402.
140 J. Rossmeisl, J. K. Nrskov, C. D. Taylor, M. J. Janik and M. Neurock,
Journal of Physical Chemistry B, 2006, 110, 21833.
141 H. A. Hansen, J. Rossmeisl and J. K. Nrskov, Physical Chemistry Chemical
Physics, 2008, 10, 3722.
142 J. X. Wang, N. M. Markovic and R. R. Adzic, Journal of Physical Chemistry
B, 2004, 108, 4127.
143 P. Ferrin, A. U. Nilekar, J. Greeley, M. Mavrikakis and J. Rossmeisl, Surface
Sciences, 2008, 602, 3424.
144 J. Macht, M. J. Janik, M. Neurock and E. Iglesia, Angewandte Chemie-
International Edition, 2007, 46, 7864.
145 J. Joubert, F. Delbecq and P. Sautet, Journal of Catalysis, 2007, 251, 507.
146 A. Goodrow and A. T. Bell, Journal of Physical Chemistry C, 2007, 111, 14753.
147 M. V. Bollinger, J. V. Lauritsen, K. W. Jacobsen, J. K. Nrskov, S. Helveg
and F. Besenbacher, Physical Review Letters, 2001, 87, 196803.
148 J. V. Lauritsen, M. Nyberg, R. T. Vang, M. V. Bollinger, B. S. Clausen,
H. Topse, K. W. Jacobsen, E. Laegsgaard, J. K. Nrskov and F. Besenbacher,
Nanotechnology, 2003, 14, 385.
149 S. Lee, C. Fan, T. Wu and S. L. Anderson, Journal of the American Chemical
Society, 2004, 126, 5682.
150 U. Heiz, A. Sanchez, S. Abbet and W. D. Schneider, Chemical Physics, 2000,
262, 189.
151 A. M. Argo, J. F. Odzak and B. C. Gates, Journal of the American Chemical
Society, 2003, 125, 7107.
152 U. Landman, B. Yoon, C. Zhang, U. Heiz and M. Arenz, Topics in Catalysis,
2007, 44, 145.
153 C. Zhang, B. Yoon and U. Landman, Journal of the American Chemical
Society, 2007, 129, 2228.
154 G. Henkelman and H. Jonsson, Journal of Chemical Physics, 2001, 115, 9657.
155 L. Xu and G. Henkelman, Journal of Chemical Physics, 2008, 129, 114104.

Catalysis, 2009, 21, 131153 | 153


This journal is 
c The Royal Society of Chemistry 2009
Ionic liquids as catalysts, solvents
and conversion agents
Amit C. Gujar and Mark G. White
DOI: 10.1039/b712677h

The literature of room temperature ionic liquids (RTILs) was reviewed to


select and report on those RTILs involved directly in chemical reactions
either as a solvent for a catalyst, a conversion agent, or a task specic ionic
liquid. Special emphasis was placed on manuscripts appearing in the
literature in the last ten years.

Introduction
The development of ionic liquids dates to 1914. The rst research eorts
involved the synthesis of ethylammonium nitrate.1 Hurley and Wier at the
Rice Institute in Texas, 1948, developed the rst ionic liquids with chloro-
aluminate ions as bath solutions for electroplating aluminum.2 These liquids
were studied primarily for their applications as electrolytes in electro-
chemistry technologies such as electroplating, batteries and alloy
preparations.
An ionic liquid (IL) is a substance that is composed entirely of ions, and is a
liquid at room temperature. Frequently the ionic liquid consists of organic
cations and inorganic anions, although it is not limited to these combinations.
While some people have said that the ionic liquid can have a high melting
temperature such as in the case of the molten salt form of NaCl, the most
commonly held understanding of this term is one that has a melting point of
less than 100 1C, more preferably less than 50 1C. For example, many
preferred ionic liquids are liquid at room temperature, or less.
The cations of the ionic liquid include organic and inorganic cations.
Examples of cations include dialkylimidazolium ion, tetra-alkylphosphonium
ion, etc. and these cations can be associated with a number of dierent
anions (Fig. 1). The anion includes organic and inorganic anions such as
PF6, CF3SO3, CF3COO, etc. and Lewis acids such as AlCl4, GaCl4,
etc.
Many ionic liquids have been widely investigated with regard to applica-
tions other than as liquid solvents: such as electrolytes, phase-transfer
reagents,3 surfactants,4 and fungicides and biocides.5,6 The physical and
chemical properties of ionic liquids can be varied over a wide range by the
selection of suitable cations and anions. Some of the properties that
depend on the cation and anion selection includes: melting point, viscosity,
density, acidity and coordination ability, solvation strength and solubility
characteristics.7
Changes in ion types, substitution, and composition produce new ionic
liquid systems, each with a unique set of properties that can be explored.

Dave C. Swalm School of Chemical Engineering, Mississippi State University, Mississippi


State, MS 37962

154 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
Fig. 1 Examples of common cations and anion pairs used in the formation of ionic liquids,
and general progression of changes in IL properties with anion type.

With the potential large matrix of both anions and cations, it becomes clear
that it will be impossible to screen any particular reaction in all possible
ionic liquids. Work is clearly needed to determine how the properties of
ionic liquids vary as functions of anion/cation and establish which, if any,
properties change in a systematic way.

Solubility of substrates in ionic liquids


By changing the nature of the ions present in an IL, it is possible to change
the resulting properties of the IL. For example, the miscibility with water can
be varied from complete miscibility to almost total immiscibility, by changing
the anion from Cl to [PF6]. The inuence of the cation is shown by
investigations of the solubility of 1-octene in dierent tosylate melts (Fig. 2).7
By increasing the nonpolar character of the cation, the solubility of 1-octene
in the melts increases markedly. Thus, one may be able to design an IL to
produce the solvation properties appropriate to the task.
Room temperature ionic liquids continue to attract interest by both
fundamental and applied researchers. Several general review articles have
been published in recent years that describe not only their physical properties
but also discuss how these physical properties can be applied for solvents used
in separations and as replacement for organic solvents for homogeneously-
catalyzed reactions. In this review, we focus our attention on those physical

Fig. 2 Solubility of 1-octene in four dierent tri-n-alkylmethylammonium tosylate melts at


80 1C, n (C) = number of C atoms of the alkyl residue.

Catalysis, 2008, 21, 154190 | 155


This journal is 
c The Royal Society of Chemistry 2009
and chemical properties that make them attractive for use in catalytic
conversions.

Physical and chemical properties of ionic liquids


Consider the many physical and chemical properties of ILs, which make
them potentially attractive media for homogeneous catalysis:
- They have very low vapor pressures, i.e. they do not evaporate and are
easy to contain.
- They generally have reasonable thermal stability. While tetra-alkyl-
ammonium salts have limited thermal stability, owing to decomposition via
the Homann elimination, [emim][BF4] is reportedly stable up to 300 1C
and emim-(CF3SO2)2N up to 400 1C.8 In other words many ionic liquids
have liquid ranges of more than 300 1C, compared to the 100 1C liquid range
of water.
- They are able to dissolve a wide range of organic, inorganic and
organometallic compounds.
- The solubility of gases, e.g. H2, CO and O2, is generally good which
makes them attractive solvents for catalytic hydrogenations, carbonylations,
hydroformylations, and aerobic oxidations.
- They are immiscible with some organic solvents, e.g. alkanes, and,
hence, can be used in two-phase systems. Similarly, lipophilic ionic liquids
can be used in aqueous biphasic systems.
- Polarity and hydrophilicity/lipophilicity can be readily adjusted by a
suitable choice of cation/anion (see earlier) and ionic liquids have been
referred to as designer solvents.9
- They are often composed of weakly coordinating anions, e.g. BF410
and PF611 and, hence, have the potential to be highly polar yet non-
coordinating solvents. They can be expected, therefore, to have a strong
rate-enhancing eect on reactions involving cationic intermediates.
- Ionic liquids containing chloroaluminate ions are strong Lewis, Franklin
and Brnsted acids. Protons present in [emim][AlCl4] have been shown to be
superacidic with Hammett acidities up to 18.12 Such highly acidic ionic
liquids are, nonetheless, easily handled and oer potential as non-volatile
replacements for hazardous acids such as HF in several acid-catalyzed
reactions.
One can envisage various scenarios for catalysis in and/or by ionic liquids:
- Monophasic systems in which the catalyst and substrate are dissolved in
the ionic liquid.
- Monophasic systems in which the ionic liquid acts as both the solvent
and the catalyst, e.g. dialkylimidazolium chloroaluminates as Friedel
Crafts catalysts (see later).
- Biphasic systems in which the catalyst resides in the ionic liquid and the
substrate/product in the second phase or vice versa.
- Mono- or biphasic systems in which the anion of the ionic liquid acts as
a ligand for the homogeneous catalyst, e.g. a sulfonated phosphine ligand
(see later).
- Triphasic systems comprising, for example, an ionic liquid, water and an
organic phase in which the catalyst resides in the ionic liquid, the substrate

156 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
and product in the organic phase and salts formed in the reaction are
extracted into the aqueous phase, e.g. in Heck reactions.

Demonstration of utility of RTILs as reaction solvents


Earle et al.11 reported the rst proof of the paradigm that the outcome
of a chemical reaction can be radically altered by the choice of the ionic
liquid as the solvent for the catalyst. The reactants were toluene and nitric
acid in HCl for the following three ILs: (1) [bmim][OTf], (2) [bmim][X],
and (3) [bmim][OMs]; where OTf is the triuoromethanesulfonate anion,
X = halide, and OMs = methanesulfonate salt. The reactions observed are
described in Scheme 1.

Scheme 1

When the IL anion was triuoromethanesulfonate (triate), the NO2 group


added to the ring with the o/p/m selectivity determined by the reaction
conditions. When the anion was a halide, it substituted on the ring for a
hydrogen atom, again with the regioselectivity determined by the reaction
conditions. Finally, when the anion is a methanesulfonate salt [OMs], the
methyl group was oxidized to the organic acid.

Review articles
We begin with a brief summary of some of the review articles that have been
written on the subject of ionic liquids. Wilkes13 wrote a short history of ionic
liquids describing the chronological development of ionic liquids with an
emphasis on listing the names and pictures of those involved in the research.
Holbrey and Seddon14 and Earle and Seddon15 reviewed the literature of
ionic liquids composed entirely of ions which were mainly of interest to
electrochemists. Recently, however, it has become apparent that, inter alia,
their lack of measurable vapor pressure characterizes them as green solvents,
and that a wide range of chemical reactions (reviewed here) can be performed
in them. Wassercheid and Keim16 reviewed the literature of ionic liquids, not
only the synthesis and physical properties of the ILs, but also their use as

Catalysis, 2008, 21, 154190 | 157


This journal is 
c The Royal Society of Chemistry 2009
solvents in separations with an emphasis on their use as solvents for transition
metal catalysts. A recent NATO advanced workshop held in Heraklion,
Crete, Green Industrial Applications of Ionic Liquids,17 aimed to provide
some answers to this question by bringing together chemists and chemical
engineers from academia and industry with expertise in ionic liquids and in
green chemistry and engineering. The workshop is the rst ever international
meeting devoted to room-temperature ionic liquids, noted ionic liquids
expert Kenneth R. Seddon of Queens University, Belfast, Northern Ireland.
And it is the rst time representatives from industry and universities have
had an opportunity to express their views in open forum about the potential
of these liquids. Fremantle18 described the use of ionic liquids in organic
synthesis. Sheldon19 described very briey the literature of ILs used in
catalytic reactions. In particular, he focused upon the advantages of using
ILs as solvents for homogeneous catalysts and described the unique proper-
ties of ILs that act as both solvent and conversion agent as in the case
of the chloroaluminate ILs. Corma and Garcia20 reviewed the literature of
Lewis acids to describe the use of ionic liquids as conversion agents. Welton21
described the many organic reactions that have been completed in ILs
showing excellent chemical and/or enantio-selectivity. Dupont et al.22
reviewed the literature describing ionic liquids as immobilizing agents for
organometallic catalysis. They have compared the behavior of reactions in
ILs with those under homogeneous conditions and under aqueous biphasic
organometallic conditions. Zhao et al.23 reviewed many reactions of indus-
trial interest and listed those reactions with their details and advantages in
tabular fashion. Apart from the above reviews, a few chapters in books
have been published on ionic liquids which primarily focus on catalysis in
non-traditional media.24,25 Wasserscheid and Welton26 edited a comprehen-
sive book on ionic liquids which included chapters ranging from IL synthesis
and physicochemical properties to their use as catalysts and solvents.
A large family of ionic liquids may be derived from the aluminum-
containing species such as aluminum halides, especially aluminum chlorides.
It seems appropriate to briey examine the patent literature for a description
of the synthesis of some members of this family of ILs.

Synthesis of aluminum-containing ILs


Moulton27 reports novel ionic liquids comprising a Lewis acid anion such as
AlyR3y+1 wherein y is greater than 0 and where the group R is independently
selected from the group consisting of an alkyl group and halogen group. The
cation of the ionic liquids can be selected from ammonium, sulfonium, and
phosphonium cations having less than 14 total carbon atoms. The anion may
contain an organic bridge to bond neighboring aluminum atoms that would
otherwise be susceptible to leaching aluminum trichloride. The ionic liquids
are useful in many applications and particularly as catalysts. The patent
literature illustrates the synthesis of this family of ILs in six examples using
AlCl3 as the source of the aluminum and for which R was derived from a
family of halide compounds containing dierent cations. The patent does not
report any reaction results.

158 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
In the following we consider the use of ILs as solvents for homo-
geneous catalysis as well as their use as catalysts. The information will
be organized according to the general category of the reaction, e.g.,
oxidation, hydroformylation, or by a named reaction such as Diels-Alder,
Friedel-Crafts, etc.

Alkoxy carbonylation
Much less attention has been focused on carbonylation reactions in ionic
liquids. The biphasic palladium-catalyzed alkoxycarbonylation of styrene,
Scheme 2, in [bmim][BF4]cyclohexane has been reported.28

Scheme 2

Very high regioselectivities (Z 99.5% iso) were obtained, using


PdCl2(PhCN)2 in combination with (+)-neomethyldiphenylphosphine and
toluene-p-sulfonic acid, under mild conditions (70 1C and 10 bar). More
recently, the palladium-catalyzed alkoxycarbonylation and amido-
carbonylation of aryl bromides and iodides in [bmim][BF4] and [bmim][PF6]
has been described.29 Enhanced reaction rates were observed compared to
conventional media and the ionic liquidcatalyst could be recycled.
Bala`zs et al. reported the hydroalkoxycarbonylation of styrene in quaternary
ammonium salts using palladium-based catalysts.30 The reactions were carried
out at 110 1C using Pd(PPh3)Cl2 or PdCl2/PPh3/CuCl2 in quaternary ammo-
nium salts such as tetrabutylammonium bromide, tetrabutylphosphonium
bromide, etc. Very high conversions and good regioselectivities were observed
for these systems. Reactions carried out in other systems such as [bmim][BF4]
and [bmim][Cl] showed poor performance with a palladium black. They
observed varying conversions and selectivities depending upon the halide
counterion used. The use of NBu4Br with PdCl2/CuCl2/PPh3 gave 100%
conversion with 78% regioselectivity; whereas, the use of NBu4Cl in the same
catalyst system gave 82% conversion and 96% regioselectivity. They also
reported that the reaction media could be reused many times without
signicant palladium leaching into the products.

Arene carbonylation
Carbonylation of the arene system can be achieved using highly acidic media
such as triic acid or sulfated zirconia. Such highly acidic conditions can also
be created in certain ionic liquids derived from the chloroaluminate anions
and the appropriate organic cations, such as the dialkylimidazolinum
cations or pyridinium cations. Ionic liquids containing chloroaluminate
(AlCl4, Al2Cl7) anions are strong Lewis acids and if protons are present
they are superacidic. Coupled with the fact that they are relatively easy to
handle makes these materials attractive non-volatile alternatives for standard

Catalysis, 2008, 21, 154190 | 159


This journal is 
c The Royal Society of Chemistry 2009
Lewis acid catalysts, such as AlCl3, and hazardous Brnsted acids such as
HF. The ionic liquid can function as both a catalyst and a solvent for acid
catalyzed processes. Since Lewis and Brnsted acid-mediated processes
generally involve cationic intermediates, e.g. carbenium and acylium ions,
one would also expect to see substantial rate enhancements in ionic liquids.
Indeed, some of the rst reactions to be studied in ionic liquids were Friedel
Crafts alkylations and acylations. Wilkes and coworkers showed that ionic
liquids derived from the reaction of [emim][Cl] with AlCl3 exhibit a wide
range of Lewis acidity depending on the molar ratio of reactants. A 1+1
mixture aords the tetrachloroaluminate, [emim][AlCl4] which is referred to
as being neutral and is not active as a FriedelCrafts catalyst. In contrast, the
2+1 adduct, [emim][Al2Cl7] is strongly acidic and was shown to be very
active in FriedelCrafts alkylations and acylations.31 For example, a mixture
of benzene, acetyl chloride and [emim][Al2Cl7] in the molar ratio 1.1/1.0/0.5
(i.e., less than a stoichiometric amount of the ionic liquid) aorded complete
conversion to acetophenone in less than 5 minutes at room temperature.
Spectral evidence suggested the formation of the free acylium cation,
CH3CO+, in the ionic liquid medium. The FriedelCrafts alkylation of
benzene with long chain a-olens is used industrially for the manufacture
of more than two million tons of linear alkylbenzenes worldwide. The
products are the precursors of the corresponding alkylbenzene sulfonates
which are widely used as surfactants. Traditionally the reaction is performed
using liquid HF or AlCl3 as the catalyst. The production of linear alkyl-
benzenes using chloroaluminate ionic liquids has been described.32 The
potential to retrot existing installations with the ionic liquid catalyst oers
enormous benets with regard to reduced catalyst consumption, ease of
product separation and elimination of caustic quenching associated with
catalyst leaching.
A family of ILs derived from chloroaluminates and methyl-alkyl-
imidazolium (R = ethyl, n-butyl, n-hexyl, benzyl, n-octyl, and n-dodecyl)
were found to be active for toluene carbonylation at room temperature and
at carbon monoxide partial pressures of B70 bar.33 Addition of gaseous
HCl at modest pressures (13 bar) dramatically improved the reactivity of
the system, which was modeled quite eectively using simple quantum
mechanical procedures and equilibrium thermodynamics.34 Quite surprising
was the systematic change in the reactivity, which depended upon the
R-group in the imidazolium cation.35 Further investigation of these systems
showed that the molecular environment of the aluminum anions changed in
a systematic manner with changes in the R-group of the imidazolium
cation.36 The authors interpreted these results in the light of the invariant
HCl absorption isotherms37 to suggest that the population of highly
acidic protons in the ILs depended upon the molecular environment of
the aluminum anions. This speculation was supported in part by the
13
C-NMR of C-labeled acetone (2-13C-acetone) which acted as a weak base
to titrate only the most acidic protons in the system. Subsequent work on
the methyl-n-butylimidazolium chlorometallate (MCl3: M = Al, Ga, and
In) showed that the activity of the toluene carbonylation reaction correlated
well with the Hammett acidity function developed from a consideration of
the chemical shift of the 13C-NMR of labeled acetone.38 One could conclude

160 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
that the hardness of the chlorometallate regulated the strength of the
proton in the IL derived from dissolved HCl gas.
Wang et al.,39 report the carbonylation of nitroaromatics with aromatic
amines in a catalytic system containing elemental sulfur in an ionic liquid,
Scheme 3.

Scheme 3

The reaction conditions were as follows: temperature = 150 1C, pressurized


CO (B13 MPa), reaction times 58 h. The S-catalyst for this system requires
an organic base to function properly, e.g., triethylamine, so that CO can be
captured by S to form the intermediate: SQCQO. Interestingly, however,
is that an optimum amount of Et3N can be observed. Increasing the amount
of S initially present in the IL results in increased yields of product. Finally, the
yields depend upon the order of introduction of the aniline to the reaction
mixture, being higher when the amine was introduced rst.
The eect of CO partial pressure, Fig. 3, was examined in more detail by
examining the yield as a function of CO partial pressure when all other
reaction variables were held constant (time = 8 h, 1.5 mmol S, 5 mmol each
of nitrobenzene and aniline, 5 mmol of Et3N, 1.5 g of [bmim][BF4], 150 1C).
The yield increased proportionally to the CO partial pressure; however the
intercept is greater than zero. Had these data been t by a linear relationship
with an intercept of zero, one could conclude that an irreversible reaction

Fig. 3 Eect of CO partial pressure upon yields.

Catalysis, 2008, 21, 154190 | 161


This journal is 
c The Royal Society of Chemistry 2009
had occurred between the CO and a species in the liquid. If the interaction
were reversible between CO and a species in the liquid, then saturation
kinetics would have been observed which was not the case here.
When the reaction was completed in other solvents (e.g., DMF, THF,
toluene and methanol) and the same reaction conditions, the yields were
much lower. Clearly, the reactivity was much higher in the ionic liquids
indicating that the polar nature of the ILs facilitated the formation of
intermediates in the reaction mechanism.

Catalytic oxidations
Considering the commercial importance of catalytic oxidations, and the fact
that ionic liquids are expected to be relatively inert towards autoxidation
with O2, surprisingly little attention has been devoted to performing such
reactions in ionic liquids. The Ni(acac)2-catalyzed aerobic oxidation of
aromatic aldehydes, to the corresponding carboxylic acids, in [bmim][PF6]
has been described.40 However, rather high (3 mol%) catalyst loadings
were used and this can hardly be considered a challenging oxidation. The
methyltrioxorhenium (MTO)-catalyzed epoxidation of olens with the
ureaH2O2 adduct (UHP) in [emim][BF4] has been reported.41 Both
the UHP and the MTO are soluble in [emim][BF4] and the medium remains
homogeneous throughout the reaction. It should be noted, however, that
the substrates were generally highly reactive olens and when the more
challenging dec-1-ene was used, a long reaction time (72 h) was needed for
moderate conversion (46%), using 2 equivalents of oxidant. When 30% aq.
H2O2 was used as the oxidant, there was ring opening of sensitive epoxides.
The asymmetric Jacobsen-Katsuki epoxidation with NaOCl, catalyzed by a
chiral Mn Schi-base complex, has been conducted in [bmim][PF6].42
However, dichloromethane was required as a cosolvent, as the ionic liquid
solidies at the reaction temperature (0 1C), which nullies a primary
incentive for using an ionic liquid. The ionic liquid containing the catalyst
could be recovered and recycled four times, albeit with a signicant loss in
yield. A more recent and very exciting development is the electro-assisted
biomimetic activation of molecular oxygen by a chiral Mn Schi-base
complex in [bmim][PF6], described by Gaillon and Bedioui.43 Evidence
was provided for the formation of the highly reactive oxomanganese(V)
intermediate that could transfer its oxygen to an olen. This would appear
to oer potential for clean, electro-catalytic oxidations with molecular
oxygen in ionic liquid media.

DielsAlder reactions in ILs


It was reported44 that Sc(OTf)3 catalyzes DielsAlder reactions in
[bmim][X] (X = BF4, SbF6 or OTf), in this case at much lower catalyst
loadings (0.2 m%). In contrast to the FriedelCrafts alkylation (see below)
the product did not form a separate phase and was recovered by extraction
with ether. It was shown, however, that the ionic liquid containing the
catalyst could be recycled eleven times without loss of activity, Scheme 4.
Furthermore, improved endo/exo selectivities were observed with cyclic
dienes.

162 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
Scheme 4

The DielsAlder reaction between cyclopentadiene (1) and a series of


dienophiles (2), Scheme 5, was examined in a typical alkylimidazolium ionic
liquid at room temperature: 1-hexyl-3-methylimidazoliumtetrauoroborate,
[hmim][BF4].45

Scheme 5

The potential activation of dierent Lewis acid catalysts and their load
eect when used in combination with this solvent were explored, in order to
determine the improvement of rates and selectivity to the endo and exo
isomers. The list of Lewis acid catalysts included: Li(OTf), Li(NTf2), ZnI2,
AlCl3, BF3, HOTf, HNTf2, Ce(OTf)4*5H2O, Y(OTf)3, Sc(OTf)3, Sc(NTf2)
and a blank without any Lewis acid. The reaction conditions were as follows:
2.2 mmol of cyclopentadiene + 2.0 mmol of dienophile + 0.2 mol% of
catalyst in 2 mL [hmim][BF4]. When no catalyst was added, the two ketones
(R1QMeCQO; R2 = R3 = H; and R1QEtCQO; R2 = R3 = H) showed
modest activity (B50% in 1 h) with endo/exo selectivity = 85/15. Whereas
acrolein showed modest activity (59% conversion in 2 h), methacrolein
and crotonaldehyde were inert without a Lewis acid catalyst. Acrylonitrile
and methyl acrylate underwent low conversions in 1 h (1617%) whereas,
N-phenylmaleimide, maleic anhydride and 2-methyl-1,4-benzoquinone
showed complete reaction in 5 min with high endo isomer yields.
In subsequent studies, methyl vinyl ketone (2.0 mmole) was chosen as the
dienophile so as to determine the combined eect of the ionic liquid (2 mL)
and the Lewis acids (0.2 and 0.5 wt%) upon the yield and selectivity. Without
the Lewis acid catalyst, this system demonstrated a 52% conversion of the
cyclopentadiene (2.2 mmol) in 1 h with the endo/exo selectivity being 85/15.
The cerium triate-catalyzed reaction was quantitative in 5 min and the
endo:exo selectivity was very good for this experiment as well (94:6, endo:exo).
Also with the scandium or yttrium salts tested, reactions came to completion
in a short time with high stereo-selection. Cerium, scandium and yttrium
triates are strong Lewis acids known to be quite eective catalysts in the
cycloadditions of cyclopentadiene with acyclic aldehydes, ketones, quinones
and cycloalkenones. These compounds are expected to act as strong Lewis
acids because of their hard character and the electron-withdrawing triate
group. On the other hand, reaction times of 1 hour were required for

Catalysis, 2008, 21, 154190 | 163


This journal is 
c The Royal Society of Chemistry 2009
complete transformation when either triic or triamic acids were used,
resulting also in a good endo:exo ratio. The rest of the Lewis acids tested
showed catalytic activity, yet not as good as the ones previously described.
Enhancement of the endo/exo ratio was achieved in some cases. It is known
that traditional Lewis acids are sensitive to water, and therefore they turn
inactive when used in lower quantity than the residual water content of the
reaction medium. This could account for the poor catalytic activity shown in
some experiments. The best Lewis acid catalyst was cerium triate, judged by
its highest activity and best endo/exo ratio. Ce(IV), Sc(III) or Y(III) salts are
extraordinarily active when used in [hmim][BF4] in DielsAlder reactions. In
fact, cerium triate performs better than any other catalyst tested, although
scandium triate is usually considered the most active in the literature. The
scope of this procedure has been extended to a wide variety of dienophiles.
The combination of 1-hexyl-3-methylimidazolium tetrauoroborate with
cerium triate, as well as with scandium triate, gave excellent results not
only in terms of reaction rates, but also in enhanced stereo-selectivity. This
protocol competes favorably with others reported previously. It was possible
to recycle the IL medium up to six runs without any loss of activity. Catalytic
systems consisting on Sc(OTf)3 plus [hmim][BF4] and Ce(OTf)4  5H2O plus
[hmim][BF4] can also be recycled and reused after extraction of the products
for at least ve times without loss of activity and endo/exo selectivity. It was
speculated that the reuse of these catalysts was possible because the catalyst
remained in the IL phase during its extraction with ether to remove the
products.

Dimerization of olens in ILs


Peng et al.46 report the dimerization of styrene in ILs using palladium-Lewis
acid catalysts and compare these results for the same dimerization in
organic solvents. The stoichiometry of the reaction was given by the
following, Scheme 6:

Scheme 6

The reaction conditions were room temperature and pressure and the IL
was developed from 1-butyl-3-methylimidazolium hexauorophosphate
using 1-4 mol% Pd(OAc)2, 1 mol % Lewis acid, and 0.5 mL of IL. The
reaction times were from 2.5 to 4.5 h. The Lewis acids were chosen from the
following: Cu(OTf)2, Cu(CF3CO2)2, Zn(OTf)2, and In(OTf)3.
Complete conversion of the substrate was observed with reaction times as
short as 2.5 h when copper triate was the Lewis acid, and the selectivity
was usually greater than 98% towards the desired compound. Upon reuse
of the same catalyst system with fresh substrate, it was reported that the
conversion was 80% and 76% for the second and third reuse of the catalyst.

164 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
No reaction was observed when either the Pd complex, the Lewis acid or the
IL was absent. Only very low yields were observed when PdCl2 was used
with Cu(OTf)2 and the IL, and no reaction was observed when Pd(PPh)3Cl2
was the catalyst. Thus, it appears that weakly coordinating ligands favor
the dimerization reaction. The reaction was also not observed with
Pd(OAc)2/Cu(OTf)2 when ILs with the anion BF4 were used. Moreover,
the choice of solvent appears to have an inuence upon the reactivity of the
preferred catalyst system, since no reaction was observed when the IL was
replaced by either 1,4-dioxane or DMSO; however a low conversion
(B30%) of the substrate was observed when the solvent was CH3NO2. It
is clear that a solvent that stabilizes ions favors the reaction.
The authors speculated that Pd(II) was reduced by reaction with the IL,
followed by formation of sigma complex between the olen and copper
triate. This polarized complex then reacts with the Pd(0)-p-complex with
the substrate to form the nal product as shown by the scheme below,
Scheme 7.
[bmim][PF6] + Pd(OAc)2 - Pd(0)

Scheme 7

By this scheme, the Pd and Cu complexes were consumed with repeated use
and thus the conversion should decrease accordingly. More evidence must
be presented to conrm this mechanism.
Wasserscheid and Eichmann47 have investigated the dimerization of
1-butene in chloroaluminate ILs using (cod)Ni(hfacac) catalyst. This
biphasic mode of operation had several advantages over running the
reaction in toluene as the solvent, such as high activity even at low
temperatures, high turnover frequency and no detectable leaching of
the catalyst. Thiele and de Souza48 have studied the eect of adding a
co-catalyst (AlEtCl2) when [Ni(MeCN)6][BF4] was used as a catalyst in
chloroaluminate IL. They concluded that the neutral IL was the best
medium for the dimerization reaction.

Enzyme-catalyzed reactions
Kim and Lee49 disclosed an ionic liquid-coated enzyme that remarkably
improves enzyme functions, such as enantioselectivity and stability, when
the enzyme, which may be lipase, is coated with an ionic liquid. Further,
even in the case where the ionic liquid-coated enzyme is reused, the

Catalysis, 2008, 21, 154190 | 165


This journal is 
c The Royal Society of Chemistry 2009
enantiomeric excess, enantioselectivity, and activity do not degenerate. The
coated enzyme is usable as a catalyst for providing a chiral intermediate
required in the synthesis of chiral pesticides, medicines, natural chemicals,
and so on.

Fischer esterications in ILs


Brnsted acidic, ionic liquids containing nitrogen-based organic cations,
such as 1-methylimidazolium and 1-butyl-3-methylimidazolium, and
inorganic anions such as BF4, PF6 and PTSA have been synthesized
in good yields and used as catalysts and reaction media for the Fischer
esterication of alcohols with acids.50 These agents were characterized using
FTIR and NMR spectroscopy. The ionic liquids as catalysts aorded good
alcohol conversion and excellent ester selectivity. The Fischer esterication
of acetic acid with benzyl alcohol using dierent Brnsted acidic ILs gave a
maximum substrate conversion of 100% and product selectivity of 100%
using [bmim][PTSA] as catalyst over a period of 2 h. The ester was easily
separated from the reaction mixture and the ionic liquid was reused four
times after removing water. No signicant loss in catalytic activity was
observed on recycling.

FriedelCrafts reactions in ILs


Chloroaluminate ionic liquids modied with HCl were recently shown to
give higher rates and more favorable product distributions in FriedelCrafts
alkylations, and these results were attributed to the superacidities of these
media. In this context it is also worth mentioning the work of Holderich
et al.51 who showed that ionic liquids immobilized on inorganic supports
(SiO2, Al2O3, TiO2, ZrO2) were eective catalysts for the FriedelCrafts
alkylation of aromatics. Activities were higher than those observed with a
conventional zeolite catalyst and no leaching of the ionic liquid from the
surface was observed. Reactions were performed in batch, continuous
liquid-phase and continuous gas-phase operation. For example, alkylation
of benzene with dodecene aorded the monoalkylated product in 98%
selectivity at 99% conversion at 80 1C.
Hardacre et al. report the FriedelCrafts benzoylation of anisole with
benzoic anhydride to yield 4-methoxybenzophenone with various ILs and
zeolite catalysts (USY, HZSM-5, H-beta, and H-mordenite). The rates of
reaction were found to be signicantly higher using ionic liquids compared
with organic solvents.52 Continuous-ow studies of successful ionic liquid
systems indicate that the bulk of the catalysis is due to the formation of an
acid via the ion exchange of the cation with the protons of the zeolite as
shown in the following reaction, Scheme 8.

Scheme 8

166 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
The acid liberated was quantied using both titration experiments and
ion-exchange experiments with sodium-exchanged zeolites.
Seddon and coworkers53 studied the FriedelCrafts acylations of toluene,
chlorobenzene and anisole with acetyl chloride in [emim][Al2Cl7] and
obtained excellent regioselectivities to the para isomer, Scheme 9. Similarly,
the fragrance chemical, traseolide, was obtained in 99% yield as a single
isomer, Scheme 10. It should be noted, however, that the question of
product recovery from the reaction medium still needs to be addressed in
these systems.

Scheme 9

Scheme 10

As in conventional AlCl3-promoted acylations the ketone product forms a


strong complex with the chloroaluminate IL. Lanthanide triates, in parti-
cular Sc(OTf)3, have been widely studied as water-tolerant Lewis acids
in a variety of transformations, including FriedelCrafts alkylations and
acylations.54 Song and coworkers55 have recently shown that Sc(OTf)3
catalyzes the FriedelCrafts alkylation of aromatics with olens in hydro-
phobic ionic liquids, e.g. [bmim][PF6] and [bmim][SbF6]. In contrast, no
reaction was observed in common organic solvents, water or hydrophilic
ionic liquids such as [bmim][BF4] or [bmim][OTf]. For example, the reaction of
benzene with cyclohexene, Scheme 11, aorded cyclohexylbenzene in 92%
yield at 99% cyclohexene conversion in [bmim][SbF6] at 20 1C for 12 h. The
product formed a separate layer and, after phase separation, the IL phase
containing the catalyst was recycled to aord 92% yield of cyclohexylbenzene
at 4 99% conversion. Although high catalyst loadings (20 mol%) were used,
the ease of separation and recycling of the catalyst oers potential environ-
mental and economic benets.

Scheme 11

Catalysis, 2008, 21, 154190 | 167


This journal is 
c The Royal Society of Chemistry 2009
A new protocol was developed for the synthesis of N-substituted thioamides,
employing arenes and isothiocyanates in 1-butyl-3-methylimidazolium chloro-
aluminate IL, [bmim][AlCl4], as a homogenous Lewis acid catalyst and
solvent,56 Scheme 12.

Scheme 12

Under similar reaction conditions, higher conversions were realized when


R was aryl rather than alkyl. The eect of Lewis acidity and the stoichio-
metry of the IL on the extent of product formation showed (1) that a
progressive increase in yields was observed with increasing Lewis acidity,
and that (2) two equivalents of [bmim][AlCl4] was the optimal amount for
the reaction. The reactivity of the reaction appeared to increase with ratio of
AlCl3/bmim-Cl = N over the range 0.5 o N o 0.67. Furthermore, the
conversion of reactant depended upon the ratio of IL/substrate, with the
highest conversion obtained when the initial reaction mixture contained 2
molar equivalents of the IL to the substrates. A distinct para selectivity for
the incoming thioamido group on activated arenes was observed under
ambient conditions. These results, activity dependence upon the value of N
and upon the initial IL/substrate ratio, and the high para selectivity, were
observed by others studying Brnsted acid-demanding reactions.3338 This
similarity is not unexpected since the thiocyanato group is believed to accept
a proton at the C to form a carbocation and thus engage the arene substrate
in an electrophilic substitution reaction. However, unlike the earlier results
for arene carbonylation, complete conversion of the substrate was not
observed when the IL/substrate ratio was Z 2 for N = 2.
1-Butyl-3-methylimidazolium chloroaluminate ionic liquids have been
employed as an unconventional reaction medium and Lewis acid catalyst
for the Friedel-Crafts sulfonylation reaction of benzene and substituted
benzenes with 4-methyl benzenesulfonyl chloride,57 Scheme 13.

Scheme 13

The substrates exhibited enhanced reactivity, giving almost complete yields


of diaryl sulfones under ambient conditions. Studies concerning the eect of
Lewis acidity of the ionic liquid on the initial extent of conversion of this
reaction have been completed using 27Al-NMR spectroscopy as a tool to
investigate the mechanistic details of the reaction. 27Al-NMR spectral studies

168 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
show the predominance of [Al2Cl7] species in [bmim][AlCl4], when the ratio
of AlCl3/bmim-Cl, N, was equal to 0.67. But in the presence of 4-methyl
benzenesulfonyl chloride, and after reaction with the aromatic hydrocarbon,
the [AlCl4] species predominates in the acidic IL. This change in speciation
of aluminum can be attributed to the interaction of the Lewis acidic
species [Al2Cl7] of the IL with the HCl formed during the sulfonylation
reaction, as evidenced by a control experiment. Preliminary investigations on
Friedel-Crafts acylation further substantiate this argument. However, these
assignments of 27Al-NMR peaks with the chloroaluminate species are not
accepted by some investigators. Part of the evidence against these assign-
ments are: (1) thermodynamic calculations34 which show that the AlCl4
species are not favored even at high HCl concentrations in the gas phase;
(2) quantum mechanical calculations37 that show HCl favors associating with
the dimeric chloroaluminate over the AlCl4 species.
Diphenylmethane and its derivatives were prepared via FriedelCrafts
benzylation reaction using ILs of 1-butyl-3-methylimidazolium chloride-ZnCl2
([bmim][ZnCl3]), 1-butyl-3-methylimidazolium chlorideFeCl3 ([bmim][FeCl4]),
1-butyl-3-methylimidazolium chloride-FeCl2 ([bmim][FeCl3]) and 1-butyl-3-
methylimidazolium chlorideAlCl3 ([bmim][AlCl4]) as both reaction media
and Lewis acid catalysts.58 In comparison with the reaction performed in
conventional organic solvents, faster reaction rates and higher selectivity to
target products were achieved in such IL media. The eect of the Lewis acid
upon the reactivity was as follows: AlCl3 4 FeCl3 4 FeCl2 4 ZnCl2. The
eects of reaction temperature, reaction time, and the ratio of metal chloride to
[bmim]Cl, as well as the amount of IL on the FriedelCrafts benzylation were
also investigated. It was found that increasing reaction temperature led to high
catalytic activity and selectivity, and that the excess amount of Lewis acidity in
the IL was detrimental to the reaction selectivity. Moreover, these ILs could be
conveniently recovered for recycling; in particular [bmim][ZnCl3], which shows
moisture-stability, could be reused at least eight times without loss of catalytic
activity.

Heck reaction
The Heck reaction is a CC coupling reaction where an unsaturated
hydrocarbon or arene halide/triate/sulfonate reacts with an alkene in
presence of a base and Pd(0) catalyst so as to form a substituted alkene.
Kaufmann et al.59 showed that the Heck reaction carried out in presence
of ILs such as tetra-alkyl ammonium and phosphonium salts without
the phosphine ligands, resulted in high yields of product. They attributed
the activity to the stabilizing eect of ammonium and phosphonium salts on
Pd(0) species. Carmichael et al.60 used ionic liquids containing either
N,N 0 -dialkylimidazolium and N-alkylpyridinium cations with anions
such as halide, hexauorophosphate or tetrauoroborate to carry out
reactions of aryl halide and benzoic anhydride with ethyl and butyl
acrylates in presence of Pd catalyst. An example of iodobenzene
reacting with ethyl acrylate to give trans-ethyl cinnamate is shown in
Scheme 14.

Catalysis, 2008, 21, 154190 | 169


This journal is 
c The Royal Society of Chemistry 2009
Scheme 14

The authors showed that the above reaction carried out in N-hexyl-
pyridinium salts gave higher yields than imidazolium salts, that the addition
of a phosphine ligand to Pd reduced the yield in pyridinium salts, and that
higher reaction temperatures were required to obtain high yields. Howarth
and Dallas61 studied the use of [bmim][PF6] as an IL solvent for the reaction
of aryl halides and methyl acrylate. They found that the yields of the
products were comparable with that obtained with DMF, and that the
solvent and catalyst could be reused several times.
Calo`s group62,63 have reacted both a-substituted and b-substituted
acrylates with haloaromatics in presence of a Pd-benzothiazole carbene
complex using tetrabutyl ammonium bromide as the solvent. The reactions
were found to be fast and ecient in the ILs when compared to
conventional solvents.

Henry reactions in ILs


The Henry reaction is an important carboncarbon bond forming reaction
having wide synthetic applications. In this reaction, a coupling between a
carbonyl compound and an alkylnitro compound takes place with the help of
an organic or inorganic base, quaternary ammonium salts and organic
solvents. Examples of the Henry reaction are shown below, Scheme 15,
where the carbonyl substrate is either an aldehyde or a ketone. Kumar and
Pawar64 showed how these reactions can be catalyzed by ILs derived from
chloroaluminates and either dialkylimidazolium chloride or pyridinium
chloride. Since the fundamental chemistry shows this reaction to be catalyzed
by a base, one then expects that high reactivity should be observed in ILs that
are bases. In the case of the chloroaluminate ILs, the higher reactivity should
be observed in mixtures when the organic cation is present in excess. Such
results were observed by Kumar and Pawar for the aldehyde and ketone

Scheme 15

170 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
substrates. Low yields of the product (230%; 3540 h reaction time) were
observed for equimolar amounts of the chloroaluminate and
1-butylpyridinium chloride, [bp][Cl]. But much higher yields were recorded
(6080%) in shorter reaction times (1516 h) when only 5% excess of the
organic cation was present initially. The reactions of these substrates were
repeated with the following organic cations: (1) 1-butyl-4-methylpyridinium
chloride, [bmp][Cl], (2) dialkylimidazolium chloride with methyl/ethyl alkyl
groups, [emim][Cl], ethyl/methyl [meim][Cl], and methyl/butyl [bmim][Cl].
The reactivity changed with the type of organic cation in the following
manner:
[emim] 4 [meim] 4 [bp] 4 [bmp] 4 [bmim].

The authors oered no explanation for the apparent order of reactivity with
these ILs, and they could be reused up to ve times with only small losses of
apparent reactivity.

Hydrogenation in ILs
Researchers65 performed the biphasic hydrogenation of cyclohexene with
Rh(cod)2 BF4 (cod = cycloocta-1,5-diene) in ILs. They observed roughly
equal reaction rates, reported as turnover frequencies of ca. 50 h1, in either
[bmim][BF4] or [bmim][PF6]. The presumption here was that the [bmim][BF4]
was free from chloride. In a separate report, the same group showed that
RuCl2(Ph3P)3 in [bmim][BF4] was an eective catalyst for the biphasic
hydrogenation of olens, with turnover frequencies up to 540 h1.66
Similarly, (bmim)3-Co(CN)5 dissolved in [bmim][BF4] catalyzed the hydro-
genation of butadiene to but-1-ene, with 100% selectivity at complete
conversion.
More recently, the ruthenium-catalyzed hydrogenation of sorbic acid to cis-
hex-3-enoic acid, Scheme 16, was achieved in a biphasic bmim-PF6methyl
tert-butyl ether (MTBE) system.67 The ruthenium cluster [H4Ru(Z6-C6H6)4]
[BF4]4, in [bmim][BF4], was shown to be an eective catalyst for the
hydrogenation of arenes to the corresponding cycloalkanes at 90 1C and
60 bar.68 The cycloalkane product formed a separate phase, which was
decanted and the IL phase, containing the catalyst, could be repeatedly
recycled.

Scheme 16

Enantioselective hydrogenation in ILs is of particular interest as it


could provide a means for facile recycling of metal complexes of expensive
chiral ligands. In their original study, Chauvin et al. reported that
[Rh (cod)(2)-(diop)][PF6] catalyzed the enantioselective hydrogenation of
a-acetamidocinnamic acid to (S)-phenylalanine with 64% ee, in a biphasic

Catalysis, 2008, 21, 154190 | 171


This journal is 
c The Royal Society of Chemistry 2009
Scheme 17

bmim-SbF6isopropyl alcohol, Scheme 17.69 The observed enantio-


selectivity is what one would expect with diop, which is not a particularly
good ligand for this reaction. The product, contained in the isopropyl
alcohol, could be separated quantitatively and the recovered IL, containing
the catalyst, reused. Similarly, Dupont and coworkers70 extended their
studies of ruthenium-catalyzed hydrogenations in ILs to enantio-
selective reactions. The chiral [RuCl2(S)-BINAP]2NEt3 complex was shown
to catalyze the asymmetric hydrogenation of 2-phenylacrylic acid and
2-(6-methoxy-2-naphthyl) acrylic acid in [bmim][BF4]isopropyl alcohol.
The latter aorded the anti-inammatory drug, (S)-naproxen, in 80% ee,
Scheme 18.

Scheme 18

The product could be quantitatively separated and the recovered IL


catalyst solution recycled several times without any signicant change in
activity or selectivity. The RTIL [emim][OTf] was employed as the sole
reaction solvent for the asymmetric hydrogenation of methyl a-benzamido
cinnamate. Near-quantitative conversions were observed At 60 psi
hydrogen partial pressure and 50 1C for 24 h, using both the achiral
DiPFcRh and the chiral EtDuPHOSRh catalysts.71 Enantiomeric excess
of 89% was observed for hydrogenations carried out with the chiral
catalyst.
An interesting recent development is the use of a biphasic ILsupercritical
CO2, scCO2, for catalytic hydrogenation.72,73 Tumas and coworkers showed
that the catalytic hydrogenation of olens could be conducted in a biphasic
[bmim][PF6]scCO2 system. The IL phase containing the catalyst was
separated by decantation and reused in up to four consecutive batches.
Jessop and coworkers extended this concept to the asymmetric hydrogena-
tion of tiglic acid, Scheme 19, and the precursor of the anti-inammatory
drug, ibuprofen, Scheme 20, using Ru(OAc)2(tol-BINAP) as the catalyst.

172 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
Scheme 19

Scheme 20

They found that reaction 19 was more selective in a [bmim][PF6]water


biphasic mixture while reaction 20 gave poor enantioselectivities in the wet
IL. In this case the best result (85% ee) was obtained using methanol as a
cosolvent at 100 bar H2 pressure. In both cases the product was separated
by scCO2 extraction when the reaction was complete. The dierent solvent
eects observed with the two substrates were assumed to be due to the
solubility of H2 in the reaction mixture. The hydrogen concentration
dependence of asymmetric catalytic hydrogenation with ruthenium BINAP
complexes is known to be dependent on the substrate.74 Class I substrates
such as the ibuprofen precursor give higher enantioselectivities at higher H2
concentration, while class II substrates, exemplied by tiglic acid, give
higher enantioselectivities at lower H2 concentrations.

Hydroformylation
Hydroformylation of propene in an aqueous biphasic system, using a water-
soluble rhodium complex of the sodium salt of trisulfonated triphenyl-
phosphine (tppts), forms the basis of the Ruhr Chemie-Rhone Poulenc
process for the manufacture of butanal.75 Unfortunately this process is
limited to C2 to C5 olens owing to the very low solubility of higher olens
in water. Hence, one can envisage that the use of an appropriate IL could
provide the basis for biphasic hydroformylation of higher olens. As noted
earlier, Parshall showed in 1972 that platinum-catalyzed hydroformylations
could be performed in tetraethylammonium trichlorostannate melts.76
More recently, Waenschmidt and Wasserscheid77 studied the platinum-
catalyzed hydroformylation of oct-1-ene in [bmim][SnCl3], Scheme 21,
which is a liquid at room temperature. Despite the limited solubility of
oct-1-ene in the IL, high activities (TOF = 126 h1) were observed together
with a remarkably high regioselectivity (n/iso = 19). The product was
recovered by phase separation and no leaching of platinum was observed.

Scheme 21

Catalysis, 2008, 21, 154190 | 173


This journal is 
c The Royal Society of Chemistry 2009
The ruthenium- and cobalt-catalyzed hydroformylation of internal and
terminal olens in molten tetra-n-butylphosphonium bromide was reported
by Knifton in 1987.78 More recently, the rhodium-catalyzed hydro-
formylation of hex-1-ene was conducted in molten phosphonium tosylates,
e.g. Bu3PEt+TsO2 and Ph3PEt+TsO2, having melting points of 8183 1C
and 9495 1C, respectively, at 120 1C and 40 bar.79 The higher melting
points of these (ionic liquids) were helpful to decant the product from the
solid catalyst medium at room temperature. Chauvin and coworkers19
investigated the rhodium-catalyzed, biphasic hydroformylation of
pent-1-ene in [bmim][PF6]. Higher activities were observed in the IL
(TOF = 333 h1 compared to 297 h1 in the conventional solvent, toluene)
with the neutral Rh(CO)2(acac)Ph3P as the catalyst precursor, but some
leaching of the catalyst into the organic phase occurred. This catalyst
leaching could be avoided by using Rh(CO)2acac with tppts or tppms
(monosulfonated triphenylphosphine) as the catalyst precursor, although
at the expense of lower reaction rates (TOF = 59 h1 with tppms). Higher
activities (TOF = 810 h1) and high regioselectivity (n/iso = 16) were
observed in the biphasic hydroformylation of oct-1-ene in [bmim][PF6]
using cationic cobaltocenium diphosphine ligands, but some catalyst
leaching (o0.5%) was observed.80 Better results were obtained with
cationic guanidine-modied diphosphine ligands containing a xanthene
backbone.81 Xanthene-based diphosphine ligands with large bite angles
(PmetalP bond angles B1101) are known to give high selectivities
(Z98%) towards the linear aldehyde.82 Biphasic hydroformylation of oct-
1-ene, Scheme 22, using rhodium complexes of these ligands in [bmim][PF6],
aorded high regioselectivities and the catalyst could be recycled ten times
(resulting in an overall turnover number of 3500) without detectable
(o0.07%) leaching of Rh to the organic phase.

Scheme 22

The group of Olivier-Bourbignou83 recently explored the use of a wide range


of ILs based on imidazolium and pyrrolidinium cations and weakly coordina-
ting anions, for the biphasic hydroformylation of hex-1-ene catalyzed by
rhodium complexes of modied phosphine and phosphite ligands. The latter
are unstable in aqueous biphasic media. The rate and regioselectivity could be
optimized by choosing a suitable combination of cation, anion and phosphine
or phosphite ligand. Rhodium leaching was minimized by modication of
the ligands with cationic (guanidinium or pyridinium) or anionic (sulfonate)

174 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
groups. Another interesting recent development is the rhodium-catalyzed
biphasic hydroformylation of oct-1-ene in [bmim][PF6]scCO2 in a continuous
ow process.84 Because of the low solubility of Rhtppms and Rhtppts complexes
in the IL, [pmim][Ph2PC6H4SO3] (pmim = 1-propyl-3-methylimidazolium)
was synthesized and used together with Rh2(OAc)4 as the catalyst precur-
sor. Aldehydes were produced at a constant rate for 72 h albeit with
moderate regioselectivity (n/iso = 3.8). Analysis of recovered products
revealed that o1 ppm Rh was leached into the organic phase. The
monophasic hydroformylation of methylpent-3-enoate in [bmim][PF6] has
been reported85 to yield the linear aldehyde product, Scheme 23, which is a
precursor to adipic acid in an alternative butadiene-based route. The
product was removed by distillation (0.2 mbar/110 1C) and the IL recycled
ten times without signicant loss in activity.

Scheme 23

Mehnert et al.,86 completed biphasic hydroformylation studies on hexene-1


using Rh-based catalysts in ILs such as [bmim][PF6], [bmim][BF4] and
1,2-dimethyl-3-butyl-imidazolium hexauorophosphate ([bdmim][PF6]).
They used high-pressure NMR studies to elaborate the active Rh species.
They also found out that increasing the amount of phosphine in the system
decreased the amount of leaching of Rh into the organic phase, and that
increasing the polarity of the organic phase leads to increasing loss of
metal from the ionic phase. Their results indicate that n/iso ratio of the
product increases only slightly with increase in the P/Rh ratio. The turnover
frequency (TOF) of the IL system was much less than that of the
homogenous system. All these factors led the authors to conclude that much
work needed to be done in order to make the IL system commercially
feasible.
The hydroformylation of alkenes in scCO2 combined with ILs was
reported by Cole-Hamilton87 while others reported on the hydrovinylation
of styrene also in the combined solvent of ILs and scCO2.88 In a recent
study, the role of scCO2 was elaborated using IR spectroscopy of the IL in
contact with scCO2.89 These IR data showed that the anion interacted with
the scCO2 and in this way its chemical properties altered. Cole-Hamiltons
group has also developed a continuous ow system for hydroformylation of
alkenes in scCO2.90 They reported a high rate of reaction with low rhodium
leaching into the products using 1-alkyl-3-methylimidazolium bis(triuoro-
methanesulfonyl)amides. They also compared this reaction system with
various commercial hydroformylation systems to show that the turnover
numbers and space time yields were comparable; the only disadvantage was
the need for high pressure. The same group has also reported the use of
supported ILs as catalysts for hydroformylation of 1-octene using silica as
the support.91 They have also developed a solventless continuous ow
hydroformylation system for 1-octene where the rhodium catalyst and IL
was dissolved in a mixture of nonanal and 1-octene which was then reacted

Catalysis, 2008, 21, 154190 | 175


This journal is 
c The Royal Society of Chemistry 2009
with CO and H2 in an scCO2 system. They conclude that a proper design of
catalyst ligands was required to maintain the catalysts solubility in the
reaction system, and that the operating range for the system they tested was
narrow.92
Magna et al.93 describe a hydroformylation process employing a cobalt-
based catalyst in a non-aqueous liquid with improved catalyst recycling. A
process for hydroformylating olenically unsaturated compounds by means
of a cobalt-based catalyst was carried out in a non-aqueous IL which is a
liquid at a temperature below 90 1C and comprises one ammonium and/or
phosphonium and/or sulfonium cation with one anion. Catalyst recycling
was improved by using a ligand selected from the group consisting of Lewis
bases and employing a depressurization step between the pressurized
reaction step and the phase separation step, by decanting. The IL containing
the catalyst was re-used.
Bohnen et al.94 report a hydroformylation method to convert olens or
olenically unsaturated compounds in the presence of at least one rhodium
compound and sulfonated arylphosphines in ILs based on a quaternary
ammonium ion or the equivalent of a multiply charged ammonium ion and
organic sulfonates or sulfates.

Isomerization
Herbst et al.,95 report a process for the conversion of linear and/or branched
parans based on the use of an IL catalyst in combination with a metal salt
additive, which provides a catalytic composition of increased activity
compared with the IL alone. Under suitable reaction conditions this
conversion leads to paran hydrocarbon fractions having higher octane
numbers.
The hydrocarbon feed used for the isomerization experiments was a mixture
of 19 wt% n-heptane, 21 wt% 2-methylhexane, 21 wt% 3-methylhexane,
36 wt% methyl-cyclohexane, 1 wt% 2,4-dimethylpentane, 1.7 wt%
2,3-dimethylpentane and 0.3 wt% of other C7 isomer compounds. The ILs
described here were developed from excess AlCl3 and trimethylamine chloride.
Anhydrous metal chlorides, such as CuCl2, CuOHCl, CuSO4, CoCl2, FeCl3,
MnCl2, MoCl5, NbCl5, NiCl2, TiCl3, or ZnCl4, were then added to the IL
before the linear alkane was added. Data for the eect of adding these Lewis
acids to the reaction mixture showed that incremental yields per mole of
additive are dierent. Adding a small amount of AlCl3 (B2 g) to the mixture
results in the highest incremental yield increase; whereas, adding more AlCl3
(5 g) results in lower incremental yields, i.e., diminishing returns are observed.
Co, Zr and Nb chlorides were also shown to have positive incremental yields
upon their addition; however, smaller amounts of AlCl3 showed the highest
incremental yields. On the other hand, negative incremental yields were
observed for the addition of Ni, Mn, and Ti chlorides (Fig. 4).
AlCl3 4 ZrCl4 4 TiCl3 B NbCl5 B MnCl2 B CoCl2 B NiCl2

Apparently, these metal chlorides that are softer acids than AlCl3 are
therefore not as eective as the harder acids in catalyzing the paran
isomerization.

176 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
Fig. 4 Incremental yields for addition of Lewis acids to chloroaluminate ILs.

Metathesis of olens
Hembre et al.96 disclosed a process for which one olen selected from certain
a,b-dihydroxyalkenes and 4-(alkenyl)ethylenecarbonates was reacted with a
second olen to produce a metathesis product. When the rst olen reactant
was an optically enriched or enantiomerically pure a,b-dihydroxyalkene,
cross metathesis reactions result in products having the same optical purity.
The a,b-dihydroxyalkenes and the 4-(alkenyl)ethylene carbonates may be
converted to hydrogenated products, and the 4-(alkenyl)ethylenecarbonates
may be decarboxylated to provide the corresponding epoxides. The
decarboxylation process may be carried out in a solvent in the presence of
a decarboxylation catalyst. Halide anions are eective decarboxylation
catalysts and thus a catalyst may be selected from group IA halide salts.
The use of a solvent to favor the solubility of these salts or the addition of
agents such as crown ethers to enhance the solvation of a salt in a given
solvent may be used to increase their catalytic activity. For similar reasons,
one or more ammonium or phosphonium salts of the halides may be used as
decarboxylation catalysts. In this case, the ammonium and/or phosphonium
salts are known as ionic liquids and can serve both as the reaction solvent
and the decarboxylation catalyst. Because ILs have very low vapor pressures,
they can be a very eective catalyst/solvent medium for the production of
relatively low-boiling epoxides, which can be removed by fractionation.

Michael reaction
The Michael reaction is the nucleophilic addition of a carbanion to
a,b-unsaturated carbonyl compounds. It is a useful way to make CC
and Chetero atom bonds. Karodias group97 studied the use of the ionic
liquid ethyltri-n-butylphosphonium tosylate (n-Bu3PEtOTs) as a solvent for

Catalysis, 2008, 21, 154190 | 177


This journal is 
c The Royal Society of Chemistry 2009
the acid-catalyzed hetero-Michael reaction. They reacted a variety of
Michael donors such as alcohols, amines and thiols with methyl vinyl
ketone in presence of p-toluenesulfonic acid monohydrate (TsOH*1H2O)
as the acid catalyst. An example involving ethanol as the Michael donor is
shown in Scheme 24.

Scheme 24

They showed that use of the phosphonium IL resulted in milder condi-


tions and yields comparable with conventional solvents. Both aliphatic and
aromatic amines reacted well in presence of phosphonium ILs, in contrast
with ammonium ILs where only aliphatic amines acted as Michael donors.
Yang et al.98 demonstrated the use of aromatic amines and N-heterocycles
as Michael donors in presence of a basic IL such as [emim]OH. They reacted
cyclohexenone with aniline in presence of [emim]OH at room temperature
to give the corresponding Michael adduct, Scheme 25.

Scheme 25

They reported a yield of 63% after 9 h using 10 mole% IL as the catalyst.


This yield increased to 90% when 50 mole% of catalyst was used. The same
authors also tried using aromatic N-heterocylces such as imidazole and
pyrazole and observed moderate to good yields of Michael adducts.
Gu et al.99 used an [hmim][HSO4] acidic IL to synthesize b-indolylketones
in excellent yields from indoles and a,b-unsaturated ketones. The recycl-
ability of the IL was poor with yields dropping in the second and third
cycles. The authors attribute this loss of activity to the loss of acidic IL
during the extraction process.
Xu et al.100 reported using [bmim][OH] as a basic IL for the Aza-Michael
addition reaction between various amines and a,b-unsaturated carbonyl
compounds and nitriles. The products were obtained in high yields and the
IL could be recycled eight times without signicant change in the product
yields.
Sharma and Degani101 have used 2-hydroxyethylammonium formate as a
low cost alternative to imidazolium based ILs in hetero-Michael reactions.
They have synthesized many Michael addition products containing CN

178 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
and CS bonds using the IL in high yields. They also observed that reactions
involving long chain acrylates were generally slow and that faster reactions
occurred in presence of a phase transfer catalyst, such as TBAB.

Pechmann condensation in ILs


Chloroaluminate ILs derived from 1-butyl-3-methylimidazolium chloride
having two equivalents of aluminum chloride were eective catalysts for the
Pechmann condensation of phenol with ethylacetoacetonate, Scheme 26.102

Scheme 26

While this reaction to form coumarin derivatives can be completed in mineral


acids, research shows that the reaction was much faster in ILs even at room
temperature. The same group103 used 1-butyl-3-methylimidazolium hexauoro-
phosphate IL at high temperatures without employing any acid catalyst. The
yields were comparable to chloroaluminate ILs with catalytic amounts of acid
at room temperature. They also concluded that Brnsted acidity (produced by
HF when [bmim][PF6] contacts water) was not responsible for the observed
activity. Singh et al.104 have used 1-butyl-3-methylimidazolium hydrogen sulfate
IL in combination with microwave irradiation. They were able to synthesize
coumarins in quantitative yields with drastic reduction in reaction times. Soares
et al.105 have used [bmim][NbCl6] IL to perform the Pechmann reaction using
various phenols with ethyl acetoacetate to produce coumarin in moderate yields
(B35%).
Gu et al.106 tested four non-chloroaluminate acidic Ils for the reaction of
phenols and methyl acetoacetate and found that SO3H-functionalized
triuoromethanesulfonate imidazolium (e.g., [MBsIm][CF3SO3]) IL was
eective for the Pechmann condensation reaction. Low IL loadings
(5 mol%) led to high yields of products. They were able to recover the IL
easily and recycle it three times with only a small loss in activity. They
concluded that the Brnsted acidities of the ILs play an important role in
accelerating the rate of the reaction. Dong et al.107 have reported similar
reactions using non-chloroaluminate ILs and showed that N,N,N-trimethyl-
N-propanesulfonic acid ammonium hydrogen sulfate [TMPSA][HSO4] was
an eective catalyst for the Pechmann reaction of various phenols with ethyl
acetoacetate.

Sonogashira reaction
The Sonogashira reaction is a CC coupling reaction of terminal alkynes
with aryl or vinyl halides in presence of Pd(0) metal and/or Cu(I) catalyst.
These compounds are useful in synthesizing species having pharmaceutical

Catalysis, 2008, 21, 154190 | 179


This journal is 
c The Royal Society of Chemistry 2009
and optoelectronic value. An example of Sonogashira reaction between
iodobenzene and phenylacetylene is shown in Scheme 27.

Scheme 27

Fukuyama et al.108 were among the earliest researchers to use ILs as


solvents for the Sonogashira reaction. They found that the reaction
performed in [bmim][PF6] with PdCl2(PPh3)2 but no Cu co-catalyst gave
good yields of product. For example, the reaction of iodobenzene with
phenylacetylene using PdCl2(PPh3)2 catalyst gave 95% yield in [bmim][PF6]
as compared to 48% yield for reactions conducted in a toluene solvent.
Gholap et al.109 compared the performance of the Sonogashira reaction in
copper and ligand-free conditions using ILs and a molecular solvent
(acetone) at ambient conditions under ultrasonic radiation. They showed
that using acetone gave good reaction times but poor recyclability. On the
other hand, using ILs as a solvent for a PdCl2 catalyst gave higher yields and
good recyclability at the expense of longer reaction times. They concluded
that ultrasonic conditions resulted in stable, nanocrystalline Pd(0) species
which promoted the catalytic activity. Li et al.110 synthesized annulenes by
the Sonogashira reaction using ILs, obtaining good yields and low catalyst
consumption. They were able to reduce the use of CuI in the system, thus
reducing the undesirable homocoupling reaction between the alkynes.
Corma et al.111 compared the use of dierent IL solvents and polyethylene
glycol (PEG) using a carbopalladacycle complex as catalyst for Suzuki and
Sonogashira reactions. They showed that the dialkyl-substituted imidazolium
compounds had poor stability, reactivity and recyclability when compared
with trialkyl substituted imidazolium compounds and PEG. They concluded
that this result could be attributed to the stabilization of the Pd nanoparticles
in the solvent. They showed that PEG was a better solvent since it gave better
yields, had good stability, low cost and low toxicity.
Rahman et al.112 used a novel high throughput reactor to produce
substituted acetylene by the Sonogashira reaction and the Mizoroki-Heck
reaction in series using the same IL in a one-pot operation. The products
were obtained in good yields and the contamination from the previous
reaction was not carried forward to the next.
Hierso et al.113 reported a copper-free, Sonogashira reaction for a number
of activated and deactivated aryl halides with alkyl-/aryl acetylenes and
using a variety of metallic precursors, bases and tertiary phosphanes in
[bmim][BF4]. They found that a combination of PdZ3 -C3 H5 Cl2 =PPh3
with 1% pyrrolidine in the absence of copper showed the highest activity.
De Lima et al.114 described a copper-free Sonogashira reaction using
palladium catalysts with butylpyridinium (C4Py) ILs. They reported that
C4Py-based ILs required an induction period in ethanol to ensure the
formation of Pd(0) species. It is possible to avoid Cu(I) and PPh3 when
C4PyPF6 is used as solvent and C4PyNO3 is used in the reaction of
phenylacetylene and iodoarenes having electron-withdrawing groups.

180 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
Sulfonation of arenes
Earle and Katdare115 reported a process for the sulfonation of an aromatic
compound in the presence of an IL. The method for the sulfonation of
aromatic compounds in water-stable ILs oered advantages over conven-
tional sulfonation reactions: (1) no by-products were formed; (2) the IL was
not consumed, and (3) the sulfonating agent (e.g., SO3) was relatively
inexpensive.

Supported analogs of ionic liquid catalysts


ILs can be immobilized on a functionalized support which contains one
component of the IL or a precursor to such a component.116 The IL may be
immobilized via the anion by treating a support with an anion source, e.g.,
an inorganic halide, before the IL is applied or formed. Alternatively, the IL
may be immobilized by having the cation covalently bound to the support,
e.g., through silyl groups, or incorporated in the support by synthesizing the
support in the presence of a suitable base. The immobilized ILs are of use as
catalysts, for example for the Friedel-Crafts reaction.
Two schemes can be imagined wherein: (1) the component forming the
anion, e.g., AlCl3 is reacted rst then followed by the organic cation
component (dialkylimidazolium chloride); or (2) the surface is reacted with
a derivative of the cation component which shows a reactive tail for the
metal oxide surface, such as trialkoxysilane, and then the anion component is
reacted with this immobilized component. For example, using conventional
chemistry for decorating metal oxide surfaces with reactive compounds, silica
has been decorated rst with a dialkylimidazolium chloride and then
this surface compound was reacted with AlCl3 to form an immobilized
chloroaluminate/imidazolium complex.117,118 Metal oxides other than silica
have been decorated such as zirconia, H-b zeolite, titania, and alumina, with
the silica showing the highest retention of the IL (B30%) when using
[bmim][AlCl4] as the IL system. Each approach showed similar reactivities
towards the test reaction, benzene alkylation with dodecene at 80 1C. Very
high dodecene conversions and high selectivities to the alkylated product
were obtained.
We tried a dierent approach to create a supported chloroaluminate IL
on silica to be used for the toluene carbonylation reaction. A functionalized
silica (2-(2-pyridyl)ethyl-functionalized silica gel (Sigma Aldrich #53,798-5)
was reacted with one equivalent of n-propyl chloride in toluene (0.2 M) for
12 h at 140 1C to form the quaternary ammonium chloride.119 The pyridyl
concentration on the silica was 1.3 mmol/g silica per the vendor. Prior to
reaction with the n-propyl chloride, the solid was evacuated at 150 1C for
1 h. The supernatant liquid was decanted, then the solid was evacuated at
room temperature overnight. A small portion of this solid was examined by
13
C-MAS-NMR to determine how well the quaternization reaction had
occurred. The spectra of the sample before and after treatment with the
n-propyl chloride were compared to show the same peaks in both samples:
1.42, 12.22, 29.87, 47.81, 110.5, 120.8, 134.9, 147.4, and 163.5 ppm;
however, the treated sample showed the peak at 110.5 ppm grew after the
treatment conrming the success of the quaternization reaction. This

Catalysis, 2008, 21, 154190 | 181


This journal is 
c The Royal Society of Chemistry 2009
functionalized silica was treated with 3.9 mmol of AlCl3 in 2.6 mmol of
toluene dissolved in 5 mL of n-octane. This mixture was stirred for 1 h at
room temperature in a Fisher-Porter tube. Carbon monoxide was added to
the Fisher-Porter tube at a pressure of 225 psig and was stirred for 4 h. The
system was depressurized and small amounts of p-tolualdehyde were
observed in the supernatant liquid.

Task specic ionic liquids (TSIL)


Recently, reports have appeared on a class of ILs known as task specic
ionic liquids (TSIL). The term was introduced by J. H. Davis, Jrs group120
to refer to those ILs which have functional groups attached to them so as to
give specic properties and functionalities. Thus, they not only perform
specic functions like metal ion extraction,121 catalysis122,123 and capture of
CO2124 but also maintain the desired physical characteristics such as
physical state, non-volatility, viscosity, etc. The implementation of TSILs
further enhances the versatility of classical ILs where both reagent and
medium are coupled.125131 The union of reagent with medium has been
found to be a viable alternative approach toward modern synthetic
chemistry, especially when considering the growing environmental demands
being placed on chemical processes.
4- and 5-methyl thioazole-based ILs have been used in conversion of
benzaldehyde to benzoin after alkylating the thiozole with n-butyl bromide
followed by anion exchange with BF4 salt.122 This reaction gave
about 80% conversion of benzaldehyde to benzoin in toluene solution.
The product was easily removed by decantation of the toluene phase,
Scheme 28.

Scheme 28

Electrophilic alkenes have been appended to imidazolium-type ILs for


use in the Diels-Alder cycloaddition, 1,4-addition, Heck and Stetter
reactions.132 Electrophilic alkenes containing Wang-type linkers were
alkylated to imidazole followed by ion exchange and esterication giving
the desired TSIL. Diels-Alder cycloaddition was carried out with
2,3-dimethylbutadiene and cyclopentadiene to give corresponding adducts.
After washing with ether, transesterication resulted in cyclohexene
derivatives, Scheme 29.

182 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
Scheme 29

Nucleophilic additions were studied using the same TSIL with pyrrolidine
and thiophenol as models. As with the Diels-Alder reaction above, the reaction
gave the required adducts which were then transesteried to give the nal
products. Heck coupling catalyzed by a transition metal and the Stetter
reaction, Scheme 30, to prepare 1,4-dicarbonyl compounds were also studied
by the same group using similar TSILs.

Scheme 30

Kamal and Chouhan133 prepared the TSIL [bmim][SCN] to convert alkyl


halides to alkyl thiocyanates. This [bmim][SCN] was prepared by ion
exchange of [bmim][Cl] with KSCN in acetone. Alkyl halides were reacted
with this TSIL to give quantitative amounts of the corresponding alkyl
thiocyanates with high yields, Scheme 31. The product was then easily
extracted with ether.
Gui et al.134 reported the Beckmann rearrangement of ketoximes using a
sulfonyl chloride-containing imidazolium-based TSIL to furnish e-caprolactam,
which was immiscible in the IL, resulting in easy separation. Similar ILs were

Catalysis, 2008, 21, 154190 | 183


This journal is 
c The Royal Society of Chemistry 2009
Scheme 31

found to have high activity for the Beckmann rearrangement of cyclohexanone


to e-caprolactam, with good yields of the resulting product. The authors also
tested the same TSIL for various other oxime conversions to the corresponding
amine or cyanide. They too gave yields which were better than previously
reported.
Wu et al.135 have synthesized a TSIL having 2,2,6,6-tetramethylpiper-
idine-1-oxyl (TEMPO) appended to an imidazolium cation for the oxida-
tion of alcohols to corresponding carbonyl compounds. These reactions
have shown high yields similar to non-supported TEMPO, but with the
additional advantage of easy separation of products.
Bazureaus group reported extensive work with TSILs for synthesis of various
4-thiazolidinones, 1,4-dihydropyridines, 3,4-dihydropyrimidin-2(1H)-ones and
polyhydroquinones.136140 They functionalized imidazole compounds with
esters, which were then reacted with reagents to aord the target compound.
They report that microwave dielectric heating is a useful aid to rapidly produce
the desired product in high yields.
Ranu and Banerjee141 developed a [bmim][OH] TSIL for oxidative homo-
coupling of terminal alkynes to 1,4-disubstituted 1,3-diynes in atmospheric
conditions using Cu(II) without using either palladium catalyst, amines,
oxidants or organic solvents. Signicant advantages stated by the authors
include fast kinetics, high yields and mild reaction conditions.
Ma et al.142 used TSIL [Rmim]HSO4 to synthesize xanthene derivatives
like 9-Aryl-3,4,5,6,7,9-hexahydroxanthene-1,8-dione. They also found that
ILs such as [bmim][HSO4] gave much higher yields of the product than
[bmim][Br], [bmim][BF4] and [bmim][PF6]. The TSIL could easily be
recovered and reused many times without any signicant loss of activity.
Magna et al.143 developed a one-pot synthesis to make b-thiocyanato
ketones by using the TSIL [bmim][SCN] at room temperature. After product
separation, the resulting IL [bmim][OH] was used to regenerate [bmim][SCN].
An interesting application of TSIL was developed by Zhang et al.144 for the
catalytic hydrogenation of carbon dioxide to make formic acid. Ruthenium
immobilized on silica was dispersed in aqueous IL solution for the reaction. H2
and CO2 were reacted to produce formic acid in high yield and selectivity. The
catalyst could easily be separated from the reaction mixture by ltration and the
reaction products and the IL were separated by simple distillation. The TSIL
developed for this reaction system was basic with a tertiary amino group
(N(CH3)2) on the cation 1-(N,N-dimethylaminoethyl)-2,3-dimethylimidazolium
triuoromethanesulfonate, [mammim][TfO].

Telomerization in ILs
Telomerization is the reaction of olens having conjugated double bonds
(conjugated dienes) in the presence of a nucleophile (telogen). The main

184 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
products obtained are compounds made up of two equivalents of the diene
and one equivalent of the nucleophile. A typical reaction is butadiene with
methanol to give 2,7-octadien-1-yl methyl ether as the primary product.
The temperature at which the telomerization reaction is carried out is in the
range 10 to 180 1C. The reaction pressure is from 1 to 300 bar. Halogen-free
palladium(0) and palladium(II) compounds, such as catalyzed by Pd(acac)2
in triphenylphosphine, have been found to be eective catalysts for telo-
merization.145 In addition, compounds of other transition metals, e.g.,
cobalt,146 rhodium, nickel147 and platinum have also been used as catalysts.
However, the latter systems are inferior to palladium complexes in activity
and selectivity. Magna et al.143 found that using 1,3-dialkylimadazolium
salts for palladium catalyzed reaction of butadiene and methanol led to
deactivation of the catalyst by the formation of palladium imidazolylidene
complexes; whereas the use of pyridinium and 1,2,3-trialkylimidazolium
salts gave high activity and selectivity without deactivation of the catalysts.
The telomerization of butadiene by means of water in ILs was described
by Dullius et al.148 Rottger et al. report a process for the telomerization of
acyclic olens having at least two conjugated double bonds, or their
mixtures, using a palladium-carbene complex as catalyst in an IL solvent.149
The nucleophiles included water, alcohols, phenols, polyols, carboxylic
acids, ammonia and primary and secondary amines. The acycylic olens
could be either 1,3-butadiene or isoprene.
Vijayaraghavan et al.150 completed a calorimetric study of the telomer-
ization reaction of acrylonitrile and trichloroacetyl chloride to give an
important intermediate for a commercial pesticide, using [C4mpyr][NTf2],
N-butyl-N-methylpyrrolidinium bis(triuoromethanesulfonyl)amide. The
reaction was highly exothermic, often resulting in temperature and pressure
excursions. The authors found that the use of the IL led to an intrinsically
safe process, as the IL was able to remove the heat of reaction eectively. As
the moderator for the reaction, the IL had ideal properties such as low
vapor pressure and high thermal stability.

Acknowledgements
The authors acknowledge helpful conversations with Dr E. J. Angueira
(Ciba Specialty Chemicals, McIntosh, AL) in selecting the literature to be
reviewed. We also acknowledge support from the Earnest W. Deavenport, Jr
endowed chair.

References
1 S. Sugden and H. Wilkins, J. Chem. Soc., 1929, 12911298.
2 (a) F. H. Hurley, US Patent 2,446,331, 1948; (b) F. H. Hurley and T. P. Wier,
J. Electrochem. Soc., 1951, 98, 207212.
3 D. Albanese, D. Landini, A. Maia and M. Penso, J. Mol. Catal. A, 1999, 150,
113.
4 E. Blackmore and G. Tiddy, J. Chem. Soc., Faraday Trans. 2, 1988, 84, 1115.
5 J. Pernak, J. Krysinski and A. Skrzypczak, Pharmazie, 1985, 40, 570.
6 J. Pernak, A. Czepukowicz and R. Pozniak, Ind. Eng. Chem. Res., 2001, 40,
2379.
7 P. Wasserscheid and W. Keim, Angew. Chem. Int. Ed., 2000, 39, 37723789.

Catalysis, 2008, 21, 154190 | 185


This journal is 
c The Royal Society of Chemistry 2009
8 P. Bonhote, A. Dias, N. Papageorgiou, K. Kalyanasundaram and M. Gratzeb,
Inorg. Chem., 1996, 35, 1168.
9 M. Freemantle, C&EN, March 30, 1998, p. 32; May 15, 2000, p. 37; January 1,
2001, p. 21; H. Carmichael, Chem. Brit., January 2000, p. 36.
10 R. Sheldon, Pure Appl. Chem., 2000, 72, 1233; R. Sheldon, Chem. Ind.
(London), 1997, 12.
11 M. Earle, S. Katdare and K. Seddon, Organic Letters, 2004, 6(5), 707710.
12 G. Smith, A. Dworkin, R. Pagni and S. Zingg, J. Am. Chem. Soc., 1989, 111,
525.
13 J. S. Wilkes, A short history of ionic liquidsfrom molten salts to neoteric
solvents, Green Chemistry, 2002, 4, 7380.
14 K. Holbrey and K. Seddon, Ionic Liquids, Clean Products and Processes,
1999, 1, 223236.
15 M. Earle and K. Seddon, Ionic liquids. Green solvents for the future, Pure
Appl. Chem., 2000, 72(7), 13911398.
16 P. Wasserscheid and W. Keim, Ionic LiquidsNew Solutions for
Transition Metal Catalysis, Angew. Chem. Int. Ed., 2000, 39, 37723789.
17 M. Freemantle, Eyes on Ionic Liquids, Science/Technology, 2000, 78(20),
3750.
18 M. Freemantle, Ionic liquids in Organic synthesisRoom-temperature ionic
liquids provide unique environment for organic reactions, Science/Technology,
2004, 82(45), 4449.
19 R. Sheldon, Catalytic reactions in ionic liquids, Chem. Comm., 2001, 23992407.
20 A. Corma and H. Garcia, Lewis Acids: From Conventional Homogeneous to
Green Homogeneous and Heterogeneous Catalysis, Chem. Rev., 2003, 103,
430765.
21 T. Welton, Chem. Rev., 1999, 99, 2071.
22 J. Dupont, R. de Souza and P. Suarez, Chem. Rev., 2002, 102, 36673692.
23 D. Zhao, M. Wu, Y. Kou and E. Min, Catalysis Today, 2002, 74, 157189.
24 D. Adams, P. Dyson and S. Tavener, Chemistry in Alternative Reaction Media,
Wiley, 2003.
25 B. Cornils, W. Herrmann, I. Horvath, W. Leitner, S. Mecking, H. Olivier-
Bourbigou and D. Vogt, Multiphase Homogeneous Catalysis, Wiley-VCH, 2005.
26 Ionic Liquids in Synthesis, eds. P. Wasserscheid and T. Welton, Wiley-VCH
Verlag, 2008, vol. 1 and 2.
27 R. Moulton, US Patent 7,053,232 (May 30, 2006).
28 D. Zim, R. de Souza, J. Dupont and A. Monteiro, Tetrahedron Lett., 1998, 39,
7071.
29 E. Mizushima, T. Hayashi and M. Tanaka, Green Chem., 2001, 3, 76.
30 A. Balazs, C. Benedek and S. Toros, J. Mol. Catal. A: Chemical, 2005, 244,
105109.
31 J. Boon, J. Levisky, J. Pug and J. Wilkes, J. Org. Chem., 1986, 51, 480.
32 J. Holbrey and K. Seddon, Clean Prod. Process., 1999, 1, 223.
33 E. Angueira and M. White, J. Mol. Catal. A: Chemical, 2005, 227/1-2, 5158.
34 E. Angueira and M. White, A.I.Ch.E. Journal, 2005, 51, 27782785.
35 E. Angueira and M. White, J. Mol. Catal. A: Chemical, 2005, 238, 163174.
36 E. Angueira and M. White, Catalysis of Organic Reactions, ed. M. Prunier,
Marcell-Dekker, Inc., 2006, ch. 17.
37 E. Angueira, PhD Thesis, Georgia Institute of Technology, 2005.
38 E. Angueria and M. White, Green Solvents for Processes, October 811,
2006, Lake Constance, Friedrichshafen, Germany.
39 X. Wang, P. Li, X. Yuan and S. Lu, J. Mol. Catal. A: Chemical, 2006, 255, 25.
40 J. Howarth, Tetrahedron Lett., 2000, 41, 6627.

186 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
41 G. Owens and M. Abu-Omar, Chem. Commun., 2000, 1165.
42 C. Song and E. Roh, Chem. Commun., 2000, 837.
43 L. Gaillon and F. Bedioui, Chem. Commun., 2001, 1458.
44 C. E. Song, W. H. Shim, E. J. Roh, S. Lee and J. H. Choi, Chem. Commun.,
2001, 1122.
45 G. Silvero, M. Jose, A. Jose, L. Bravo, M. Valos, J. Jimenez and I. Lopez,
Tetrahedron, 2005, 61, 71057111.
46 J. Peng, J. Li, H. Qui, J. Jiang, K. Jiang, J. Mao and G. Lai, J. Mol. Catal. A:
Chemical, 2006, 255, 1618.
47 P. Wasserscheid and M. Eichmann, Catalysis Today, 2001, 66, 309316.
48 D. Thiele and R. de Souza, J. Mol. Catal. A: Chemical, 2005, 264, 293298.
49 M. Kim and J.-k. Lee, US Patent 7,005,282, February 28, 2006.
50 T. Joseph, S. Sahoo and S. Halligudi, J. Mol. Catal. A: Chemical, 2005, 234,
107110.
51 C. DeCastro, E. Sauvage, M. Valkenberg and W. Holderich, J. Catal., 2000,
196, 86.
52 C. Hardacre, S. Katdare, D. Milroy, P. Nancarrow, D. Rooney and J. Thompson,
Journal of Catalysis, 2004, 227, 4452.
53 C. Adams, M. Earle, G. Roberts and K. Seddon, Chem. Commun., 1998, 2097.
54 For reviews see: S. Kobayashi, Synlett, 1994, 689; S. Kobayashi, Eur. J. Org.
Chem., 1999, 15.
55 M. Yoon, J. Kim, D. Choi, U. Shin, J. Lee and C. Song, Adv. Syn. Cat., 2007,
349, 17251737.
56 P. Naik, S. Nara, J. Harjani and M. Salunkhe, Can. J. Chem., 2003, 81, 10571060.
57 S. Nara, J. Harjani and M. Salunkhe, J. Org. Chem., 2001, 66, 86168620.
58 D. Yin, C. Li, L. Tao, N. Yu, S. Hu and D. Yin, J. Mol. Catal. A: Chemical,
2006, 245, 260265.
59 D. Kaufmann, M. Nouroozian and H. Henze, Synlett, 1996, 1091.
60 A. Carmichael, M. Earle, J. Holbrey, P. McCormac and K. Seddon, Organic
Letters, 1999, 1(1), 9971000.
61 J. Howarth and A. Dallas, Molecules, 2000, 5, 851855.
62 V. Calo, A. Nacci, L. Lopez and A. Napola, Tetrahedron Lett., 2001, 42, 4701.
63 V. Calo, A. Nacci, A. Monopoli, L. Lopez and A. di Cosmo, Tetrahedron,
2001, 57, 60716077.
64 A. Kumar and S. Pawar, J. Mol. Catal. A: Chemical, 2005, 235(1-2, 1), 244248.
65 P. Suarez, J. Dullins, S. Einloft, R. de Souza and J. Dupont, Polyhedron, 1996,
15, 1217.
66 P. Suarez, J. Dullins, S. Einloft, R. de Souza and J. Dupont, Inorg. Chim. Acta,
1997, 255, 207.
67 S. Steines, B. Drieen-Holscher and P. Wasserscheid, J. Prakt. Chem., 2000,
342, 348.
68 P. Dyson, D. Ellis, D. Parker and T. Welton, Chem. Commun., 1999, 25.
69 Y. Chauvin, L. Mussman and H. Olivier, Angew. Chem., Int. Ed. Engl., 1995,
34, 2698.
70 A. Monteiro, F. Zinn, R. de Souza and J. Dupont, Tetrahedron: Asymmetry,
1997, 8, 177.
71 K. Boyle, E. Lipsky and C. Kalberg, Tetrahed. Lett., 2006, 47, 131113.
72 F. Liu, M. Abrams, R. Baker and W. Tumas, Chem. Commun., 2001, 433.
73 R. Brown, P. Pollet, E. McKoon, C. Eckert, C. Liotta and P. Jessop, J. Am.
Chem. Soc., 2001, 123, 1254.
74 R. Noyori, Asymmetric Catalysis in Organic Synthesis, Wiley, New York,
1994.
75 E. Kuntz, CHEMTECH, 1987, 570.

Catalysis, 2008, 21, 154190 | 187


This journal is 
c The Royal Society of Chemistry 2009
76 G. Parshall, J. Am. Chem. Soc., 1972, 94, 8716.
77 H. Waenschmidt and P. Wasserscheid, J. Mol. Catal. A: Chemical, 2000, 164,
61.
78 J. Knifton, J. Mol. Catal., 1987, 43, 65.
79 N. Karodia, S. Guise, C. Newlands and J. Anderson, Chem. Commun., 1998,
2341.
80 C. Brasse, U. Englert, A. Salzer, H. Waenschmidt and P. Wasserscheid,
Organometallics, 2000, 19, 3818.
81 P. Wasserscheid, H. Waenschmidt, P. Machnitzki, K. Kottsieper and O. Steltzer,
Chem. Commun., 2001, 451.
82 P. van Leeuwen, P. Kamer, J. Reek and P. Dierkes, Chem. Rev., 2000, 100.
83 F. Favre, H. Olivier-Bourbigou, D. Commereuc and L. Saussine, Chem.
Commun., 2001, 1360.
84 M. Sellin, P. Webb and D. Cole-Hamilton, Chem. Commun., 2001, 781.
85 W. Keim, D. Vogt, H. Waenschmidt and P. Wasserscheid, J. Catal., 1999,
186, 481.
86 C. Mehnert, R. Cook, N. Dispenziere and E. Mozeleski, Polyhedron, 2004, 23,
26792688.
87 D. Cole-Hamilton, Chem. Commun., 2001, 781.
88 W. Leitner and P. Wasserscheid, Angew. Chem., Int. Ed., 2001, 40, 2697.
89 S. Kazarian, B. Briscoe and T. Welton, Chem. Commun., 2000, 2047.
90 P. Webb, M. Sellin, T. Kunene, S. Williamson, A. Slawin and
D. Cole-Hamliton, J. Am. Chem. Soc., 2003, 125, 1557715588.
91 U. Hintermair, G. Zhao, C. Santini, M. Muldoon and D. Cole-Hamliton,
Chem. Commun., 2007, 14621464.
92 A. Frisch, P. Webb, G. Zhao, M. Muldoon, P. Pogorzelec and D. Cole-Hamliton,
Dalton Trans., 2007, 55315538.
93 L. Magna, H. Olivier-Bourbigou, L. Saussine and V. Kruger-Tissot, US Patent
7,060,860 (June 13, 2006).
94 H. Bohnen, J. Herwig, D. Ho, P. Wasserscheid and R. van Hal, US Patent
6,995,293, February 7, 2006.
95 K. Herbst, J. Houzvicka, B. Jespersen and J. Zavilla, US Patent 7,053,261
(May 30, 2006).
96 R. Hembre and J. Penney, US Patent 7,002,049, February 21, 2006.
97 N. Karodia, X. Liu, P. Ludley, D. Pletsas and G. Stevenson, Tetrahedron,
2006, 62, 1103911043.
98 L. Yang, L. Xu, W. Zhou, L. Li and C. Xia, Tetrahedron Lett., 2006, 47,
77237726.
99 D.-g. Gu, S.-j. Ji, H.-x. Wang and Q.-y. Xu, Synth. Commun., 2008, 38,
12121223.
100 J.-m. Xu, Q. Wu, Q.-y. Zhang, F. Zhang and X.-f. Lin, Eur. J. Org. Chem.,
2007, 17981802.
101 Y. Sharma and M. Degani, J. Mol. Catal. A: Chemical, 2007, 277, 215220.
102 M. Potdar, S. Mohile and M. Salunkhe, Tetrahedron Lett., 2001, 42, 9285.
103 M. Potdar, M. Rasalkar, S. Mohile and M. Salunkhe, J. Mol. Catal. A:
Chemical, 2005, 235, 249252.
104 V. Singh, S. Kaur, V. Sapehiyia, J. Singh and G. Lad, Cata. Comm., 2005, 6,
5760.
105 V. Soares, M. Alves, E. Souza, I. Pinto, J. Rubim, C. Andrade and P. Suarez,
Int. J. Mol. Sci., 2007, 8, 392398.
106 Y. Gu, J. Zhang, Z. Duan and Y. Deng, Adv. Synth. Catal., 2005, 347, 512516.
107 F. Dong, C. Jian, G. Kai, S. Qunrong and L. Zuliang, Catal. Lett., 2008, 121,
255259.

188 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
108 T. Fukuyama, M. Shinmen, S. Nishitani, M. Sato and I. Ryu, Organic Letters,
2002, 4(10), 16911694.
109 A. Gholap, K. Venkatesan, R. Pasricha, T. Daniel, R. Lahoti and K. Srinivasan,
J. Org. Chem., 2005, 70, 48694872.
110 Y. Li, J. Zhang, W. Wang, Q. Miao, X. She and X. Pan, J. Org. Chem., 2005,
70, 32853287.
111 A. Corma, H. Garcia and A. Leyva, Tetrahedron, 2005, 61, 98489854.
112 M. Rahman, T. Fukuyama, I. Ryu, K. Suzuki, K. Yonemura, P. Hughes and
K. Nokihara, Tetrahedron Lett., 2006, 47, 27032706.
113 J.-C. Hierso, J. Boudon, M. Picquet and P. Meunier, Eur. J. Org. Chem., 2007,
583587.
114 P. de Lima and O. Antunes, Tetrahedron Lett., 2008, 49, 25062509.
115 M. Earle and S. Katdare, US Patent, 7,009,077, March 7, 2006.
116 E. Sauvage, M. Valkenberg, C. De Castro-Moreira and W. Holderich,
US Patent 6,969,693 November 29, 2005.
117 M. Valkenberg, C. de Castro and W. Holderich, Spec. Publ.R. Soc. Chem.,
2001, 266, 242.
118 M. Valkenberg, C. de Castro and W. Holderich, Top. Catal., 2001, 14, 139.
119 M. G. White, US Provisional Patent, led, 2003.
120 J. Davis, Jr, Chemistry Letters, 2004, 33(9), 10721077.
121 A. Visser, R. Swatloski, W. Reichert, R. Mayton, S. She, A. Wierzbicki,
J. Davis, Jr and R. Rogers, Chem. Comm., 2001, 135, 135136.
122 J. Davis, Jr and K. Forrester, Tetrahedron Lett., 1999, 40, 16211622.
123 Y. Gu, F. Shi and Y. Deng, J. Molecular Cat. A: Chemical, 2004, 212, 7175.
124 E. Bates, R. Mayton, I. Ntai and J. Davis, Jr, J. Am. Chem. Soc., 2002, 124(6),
926927.
125 A. Visser, R. Swatloski, W. Reichert, R. Mayton, S. She, A. Wierzbicki,
J. Davis, Jr and R. Rogers, Chem. Commun., 2001, 135.
126 A. Wierzbicki and J. Davis, Jr, in Proceedings of the Symposium on Advances
in Solvent Selection and Substitution for Extraction, 59 March 2000, Atlanta,
GA.
127 C. Mathews, P. Smith, T. Welton, A. White and C. Williams, Organometallics,
2001, 20, 3848.
128 J. Fraga-Dubreuil and J. Bazureau, Tetrahedron Lett., 2001, 42, 6097.
129 C. Brasse, U. Englert, A. Salzer, H. Waenschmidt and P. Wasserscheid,
Organometallics, 2000, 19, 3818.
130 K. Kottsieper, O. Stelzer and P. Wasserscheid, J. Mol. Catal. A: Chemical,
2001, 285.
131 D. Brauer, K. Kottsieper, C. Liek, O. Stelzer, H. Waenschmidt and
P. Wasserscheid, J. Organomet. Chem., 2001, 630, 177.
132 S. Anjaiah, S. Chandrasekhar and R. Gree, Tetrahedron Lett., 2004, 45,
569571.
133 A. Kamal and G. Chouhan, Tetrahedron Lett., 2005, 46, 14891491.
134 J. Gui, Y. Deng, Z. Hu and Z. Sun, Tetrahedron Lett., 2004, 45, 26812683.
135 X.-E. Wu, L. Ma, M.-X. Ding and L.-X. Gao, Synlett, 2005, 4, 06070610.
136 J. Fraga-Dubreuil and J. Bazureau, Tetrahedron, 2003, 59, 61216130.
137 H. Hakkou, J. Eynde, J. Hamelin and J. Bazureau, Synthesis, 2004, 11, 17931798.
138 J. Legeay, J. Eynde and J. Bazureau, Tetrahedron, 2005, 61, 1238612397.
139 J. Legeay, J. Goujon, J. Eynde, L. Toupet and J. Bazureau, J. Comb. Chem.,
ASAP article, 31st August 2006.
140 A. Arfan, L. Paquin and J. Bazureau, Russ. J. Org. Chem., 2007, 43, 10581064.
141 B. Ranu and S. Banerjee, Lett. Org. Chem., 2006, 607609.

Catalysis, 2008, 21, 154190 | 189


This journal is 
c The Royal Society of Chemistry 2009
142 J. Ma, X. Zhou, X. Zang, C. Wang, Z. Wang, J. Li and Q. Li, Aust. J. Chem.,
2007, 60, 146148.
143 L. Magna, Y. Chauvin, G. Niccolai and J.-M. Basset, Organometallics, 2003,
22, 44184425.
144 Z. Zhang, Y. Xie, W. Li, S. Hu, J. Song, T. Jiang and B. Han, Angew. Chem.
Int. Ed., 2008, 47, 11271129.
145 A. Behr, in Aspects of Homogeneous Catalysis, ed. R. Ugo, D. Reidel
Publishing Company, Doordrecht/Boston/Lancaster, 1984, vol. 5, p. 3.
146 R. Baker, A. Onions, R. Popplestone and T. Smith, J. Chem. Soc., Perkin
Trans. II, 1975, 11331138.
147 R. Baker, D. Halliday and T. Smith, J. Organomet. Chem., 1972, 35, C61C63;
R. Baker, Chem. Rev., 1973, 73, 487530; R. Baker, A. Cook and T. Smith,
J. Chem. Soc., Perkin Trans. II, 1974, 15171524.
148 J. Dullius, P. Suarez, S. Einloft, R. de Souza, J. Dupont, J. Fischer and
A. Cian, Organometallics, 1999, 17, 9971000.
149 D. Rottger, M. Beller, R. Jackstell, H. Klein and K.-d. Wiese, US Patent
7,026,523, April 11, 2006.
150 R. Vijayaraghavan, M. Surianarayanan and D. MacFarlane, Ind. Eng. Chem.
Res., 2007, 46, 10251028.

190 | Catalysis, 2009, 21, 154190


This journal is 
c The Royal Society of Chemistry 2009
Measurement techniques in catalysis for
mechanism development: kinetic, transient
and in situ methods
Nora M. McLaughlin and Marco J. Castaldi
DOI: 10.1039/b712662j

1. Introduction
The initial observations regarding a catalytic eect are attributed to Sir
Humphrey Davy and Russian chemist Gottlieb Kirchho whom both
observed catalytic eects by 1812. Davy was evidently the rst to observe
that platinum induced the oxidation of alcohol vapor in air1 and Kirchho
was the rst to observe the liquid phase catalysis of acid breaking down
starch.2 That was the likely beginning of the experimentally based study of
catalysis that continues today. Over the years, catalysis research developed
broad categories that still govern most current research eorts. The
categories can be generally classied as experimental and theoretical and
more specically, the experimental consists of preparation techniques,
activity-selectivity testing and kinetic and mechanistic elucidation. The
specic theoretical techniques are recently more capable of predicting
electronic and mechanical bulk properties of more complex systems based
on electronic wave function or electron density calculations.
There has been signicant integration of these categories that has enabled
a tremendous understanding of catalytic reactions, cycles and eect. One of
the many signicant contributions to the catalysis eld was the observation
that the catalyst structure changes during the reaction that it is aecting.
Gabor Somorjai conducted the rst experiments that compared surface
diraction patterns of a crystal both before and after it was exposed to
reactive gases.3 That work set the stage for the molecular investigation of
catalysis.
The grand challenge to catalysis has become the a-priori design of a
catalyst for a given reaction. That is, select the specic chemicals (i.e. Pt, Re,
Ni, La, CeO) in the correct ratios, prepare the catalyst, and deposit it onto a
support (i.e. g-Al2O3) and substrate (e.g. monolith), put the substrate into
the reactor and achieve a predicted result. This is not possible currently.
Thus we must resort to sets of preparation-activity correlations, structure-
activity correlations and idealized modeling eorts. The combination of
these sets has and will continue to provide a very detailed understanding of
catalytic reaction engineering and development. Not until we can accurately
and completely describe the electronic nature of irregular surface structures
containing several dierent atoms will we be able to come close to this
a priori ability. Therefore, the frontier of catalysis today is the development
of analysis tools to help us see the reactions occurring during the catalytic
cycle along with the modeling tools to match that data generated. This

Department of Earth & Environmental Engineering, Columbia University, New York NY


10027, USA

Catalysis, 2009, 21, 191218 | 191


This journal is
c The Royal Society of Chemistry 2009
chapter will focus on a few of the tools that have signicantly advanced the
eld of catalysis and will likely do so for the foreseeable future.
While in situ techniques encompass all characterization/spectroscopic
methods that can be used to probe the surface chemistry of an operating
practical catalyst, its entirety is too large to cover in any real detail.
Therefore the methods covered in this review were chosen based on their
prevalence and use in the eld and the potential for signicant observations
during reaction cycles. Some methods, such as ATR, TAP and catalytic
shock tube, were chosen based on the potential of these methods and the
likelihood that they will become more widely used as they are integrated
with evolving spectroscopic techniques.

2. Structure to kinetics
The relationship between the morphology of the catalytic surface and the
performance of a system is referred to as the structure-activity relationship
(SAR). When a catalytic reaction takes place, there exists intermediate
chemical species, which are dierent from both the reactants and products.
Intermediate state species are the result of inter- and intra-molecular
processes and bond changes when the reacting substance and the catalyst
interact. This change enables faster reaction rates and higher activity.
Compounds are often classied together if they have similar structural
characteristics, including shape, size, stereochemical arrangement, and
distribution of functional groups. Factors contributing to SAR include
chemical reactivity, electronic eects, resonance, and inductive eects. SAR
was observed and postulated in 19074,5 and has since been extended and
formalized by Boudart.68 Currently there are two main categories of
SAR: primary SAR, caused by geometric eects, and secondary SAR,
caused by preferential poisoning of sites of a given geometry and coordina-
tion. Moreover, structure sensitive reactions typically occur on large multi-
ple atom sites, which are very sensitive to alloy or poison eects and involve
the activation of bonds (i.e. CC or NH). A signicant number of studies
have been done on various catalytic metals ranging from base metals such as
Ni, Cu and Fe to precious metals such as Au, Pt, Pd and Ir.
The use of two SAR categories is a very powerful methodology to
organize catalyst performance. It should be recognized, however, that
SAR performance is not just a two category phenomena; rather, it is a
spectrum where the catalysts have varying degrees of behavior.9 Informa-
tion concerning the mechanism and kinetics of a reaction can be determined
by knowing the intermediate chemical species formed on the surface by
activating bonds. The connection to kinetics is the impact of structure on
activity, manifested in the evaluation of the Turnover Frequency (TOF).7
The TOF is the intrinsic value of the rate of reaction on a given active site
aorded by its structure. The active site is only expected to stay constant
when it is comprised of a single metal atom. The number of surface metal
atoms (counted via titration or other techniques) is equated to the number
of points at which the reaction can proceed, varying the capacity of a
catalyst to adsorb gas and promote reaction.10 Researchers have quantied

192 | Catalysis, 2009, 21, 191218


This journal is
c The Royal Society of Chemistry 2009
this eect by dening a reaction dimension of particle size related to catalyst
activity using fractal theory.11
As noted by Carberry in 1987,12 only phenomenological values can be
measured in the laboratory since it is not possible to a priori distinguish
between A (the catalytic area) and A (exposed measurable area), per volume
of catalyst agent. This yields a structure-sensitive reaction that is dependent
on crystallite size. While it is clear that a mechanism cannot be determined
from purely kinetic measurements, a proposed mechanism is only accepted
after it can predict the observed kinetic measurements. The dominant issue
of the observed measurements is whether A or A is being measured. This
correct measurement will yield the proper intrinsic kinetics, but will not
reveal much insight into the mechanism. Thus, it is imperative to identify
and obtain as much information as possible on the nature of intermediate
chemical species.
While many techniques have evolved to evaluate surface intermediates, as
will be discussed below, it is equally important to also obtain information
on gas phase intermediates, as well. While the surface reactions are
interesting because they demonstrate heterogeneous kinetic mechanisms,
it is the overall product yield that is nally obtained. As presented in a text
by Dumesic et al.,13 one must approach heterogeneous catalysis in the way it
has been done for gas phase systems, which means using elementary
reaction expressions to develop a detailed chemical kinetic mechanism
(DCKM). DCKMs develop mechanisms in which only one bond is broken
or formed at each step in the reaction scheme. The DCKM concept was
promoted and used by numerous researchers to make great advances in the
eld of gas phase model predictions.
In the case of heterogeneous catalysis, a DCKM or microkinetic model
must incorporate the added dimension of adsorbed chemical species as well
as active versus non-active sites. To obtain the full predictive capability from
reactant inuent to product euent, all possible reactions in the system,
both heterogeneous and homogeneous, must be accounted for. To properly
understand the catalytic reaction sequence, it is possible that seemingly
unimportant intermediates on the surface may be what initiate gas phase
reactions. To begin this elucidation, the surface chemical species and their
properties must be known.
Catalytic reactions involve a complex set of related and linked events.
Both the catalyst structure and the type of reactants, intermediates and
products adsorbed on the surface impact the functionality of a catalyst.
Therefore, an understanding of the various events must be synthesized with
the intrinsic kinetics to produce a reaction mechanism that accurately
models the catalyst surface. The goal is to characterize and understand
the catalyst surface well enough to predict the performance either for a class
of reactions or within the tested range.

3. Surface measurement techniques


The early indications and correlations of catalyst surface impacts on
performance were obtained via ex situ techniques in which the catalyst is
removed from the actual operating system. The catalysts studied were either

Catalysis, 2009, 21, 191218 | 193


This journal is
c The Royal Society of Chemistry 2009
placed in reacting environments that were far from the targeted operating
conditions, or tested simply by probing the potentially active site portions of
the catalyst.
There exist limitations in the current systems of studying surface reac-
tions. For example, mass and heat transport eects create error in the rate
coecients obtained via conventional dierential reactors. This primarily
occurs when gradients of temperature and species concentration exist
between the gas and surface during reaction. Another limitation is the
inability to quench reactive intermediates, which are essential in developing
a mechanistic understanding of a process.1422 Generally, the intermediates
cannot be isolated because the gases are in a regime where they continue to
react before they can be identied.
Other examples are the ultra-high vacuum (UHV) and scanning tunneling
microscopy (STM) methods, where surfaces are examined. The UHV
literature, such as work by Besenbacher,23,24 discuss developments in the
use of model supported catalysts to aid in future catalyst development, such
as steam reforming with MgAl2O3/Ni/Au catalysts. Freund and co-workers
have used model catalyst systems to probe the interactions of atomic
scale eects to guiding catalyst design for near 100% selectivity for hydro-
genation and dehydrogenation reactions on supported Pd particles, methanol
oxidation at vanadium oxide model catalysts and, catalysts with a well
dened charge state.2528 Goodman and co-workers used model catalyst
systems to understand the nature of size and shape, i.e. clusters to nano-
particles, on catalytic performance2931 as well as model supports such as
silica.32 This method is a signicant step in progressing towards synthesizing
new catalyst formulations. However, the success of such a technique relies on
a multitude of existing work to direct the exploration eorts. Recently, Soft
X-ray absorption spectroscopy was used to investigate water dissociation on
Ru(001). It was found water wets Ru(001) either nondissociatively or
dissociatively, this enables a partially dissociated overlayer to be favored
energetically. Further dissociation is kinetically hindered due to the competi-
tion with desorption.33 One method that has enabled rapid testing schemes is
non-isothermal testing of an integral catalytic reactor called Polythermal-
temperature-ramping reactor (PTR).3436
This method involves the measurement of reactant concentrations and
temperatures at the reactor exit for various contact times as shown in Fig. 1.

Fig. 1 Raw and smoothed data of conversion (x) versus clock time (t) for dierent runs
showing varying degrees of catalytic activity.34

194 | Catalysis, 2009, 21, 191218


This journal is
c The Royal Society of Chemistry 2009
This data can be acquired in non-isothermal reactor operation, which is
advantageous, but radial gradients must be non-existent. To evaluate
kinetic data, it is necessary that the varying contact times correspond to
dierential changes of exit concentrations and temperatures. With this
precondition, the relationship between measured exiting concentration,
temperature and contact time can be described by a simple approximated
interpolation function. However, since the precision of rate data is generally
lower than that of concentration data (i.e. rate data is noisier), reducing this
data may consume a signicant amount of time. While this technique has
signicantly advanced the understanding of catalyst reaction engineering, it
cannot be applied to highly exothermic reactions and does not identify
intermediate chemical species.
Pfeerle and Lyubovsky37 executed types of measurements that yielded
critical information between active Pd phases for catalytic combustion using
pure a-alumina plates with zero porosity as a support for the catalyst. This
procedure uniformly covers the plate with metal particles on the top surface
where they are easily available for the reaction gases and optical analysis.
This type of experimental procedure has shown that in high-temperature
methane oxidation the reduced form of the supported palladium catalyst is
more active than the oxidized form. The temperature at which the PdO 3 Pd
transition occurs depends on parameters including the gas phase oxygen
pressure and interaction between metal and support. A schematic of the
system used is shown in Fig. 2.
Somewhere in between the model catalysts used in UHV systems and the
technique just discussed, there were other attempts to freeze the catalyst
surface morphology. One technique was dropping the catalyst test sample
into a liquid nitrogen bath, intending to quench all reactions, while
maintaining the catalyst surface in its reacting state. Since both the reaction
time and the time to transfer the sample from reactor to bath take place on
the order of microseconds, the surface studied is likely somewhere between a
reacting and non-reacting state.

Fig. 2 Reactor system used to quench the catalyst surface after testing and that allows for
in situ measurements during reaction.37

Catalysis, 2009, 21, 191218 | 195


This journal is
c The Royal Society of Chemistry 2009
Temporal analysis of products (TAP) reactor systems enable fast
transient experiments in the millisecond time regime and include mass
spectrometer sampling ability. In a typical TAP experiment, sharp pulses
shorter than 2 milliseconds, e.g. a Dirac Pulse, are used to study reactions
of a catalyst in its working state and elucidate information on surface
reactions. The TAP set-up uses quadrupole mass spectrometers without a
separation capillary to provide fast quantitative analysis of the euent.
TAP experiments are considered the link between high vacuum molecular
beam investigations and atmospheric pressure packed bed kinetic studies.
The TAP reactor was developed by John T. Gleaves and co-workers at
Monsanto3842 in the mid 1980s. The rst version had the entire system
under vacuum conditions and a schematic is shown in Fig. 3. The rst
review of TAP reactors systems was published in 1988.
Over the years, TAP systems have been used in catalyst industry research
ranging from determining selectivity and activity of well-dened surfaces to
distinguishing between sequential or parallel reaction paths. The capabil-
ities of the TAP system are varied and can perform applications such as
TPD, TPSR, TPO and TPR. In pulse experiments, the adsorption of
reactants provides detailed information on the thermal stability of adsorbed
intermediates. Fig. 4 shows the general parameters of the TAP system.
The original TAP system has been modied, resulting in the TAP-2
reactor system.38 The main dierence between the TAP and TAP-2 reactor
systems is higher sensitivity and detection eciency of reactor euent
because the microreactor and the detector are physically much closer in
TAP-2 than TAP. Moreover, TAP-2 can perform experiments using much
smaller input pulses than the original TAP system. Gleaves et al.38 have
identied the next set of adaptations for the TAP-2 reactor system by
combining the TAP-2 experimental apparatus with IR and Raman

Fig. 3 Schematic of TAP reactor.41

196 | Catalysis, 2009, 21, 191218


This journal is
c The Royal Society of Chemistry 2009
Fig. 4 Parameters of the TAP reactor system.

spectroscopy. This will enable simultaneous monitoring of the catalyst and


the gas phase during reaction. The new TAP system is an in situ analysis
method that simultaneously investigates surface adsorbed species, desorbed
reaction products, and the catalyst structure during catalytic reactions.
Currently, there is a TAP-3 system with an improved pulse valve head
design that makes modifying and changing the valves easier. BASF has
recently taken receipt of this new TAP-3 reactor system.43
The interpretation of TAP data relies largely on the modeling of the test
bed studied.4450 In general, gas transport through the reactor test bed
proceeds by either viscous ow, Knudsen ow or a combination of the two.
Inertial ow eects, however, take eect at the onset of short gas pulses and
at the beginning of the pulses viscous ow. The choice of the reactor
model is critical to data interpretation. The models range from one-zone
(thin-zone) with a Dirac delta function as a boundary condition where
concentration gradients can be neglected51 to a three-zone model using
linearly decreasing input or ux functions as the boundary conditions. The
three-zone unit has the catalyst test sample positioned between an upstream
and downstream inert packing material, thus zone 1 is inert, zone 2 has
catalyst and zone 3 has inert.
These experiments have made tremendous advances of in situ catalyst
reactor system analysis. A discussion of all of them is outside the scope of
this chapter. However, descriptions of the techniques that have contributed
signicantly to the rational design and analysis of catalysts are presented in
the next section. It should be recognized that it is not an exhaustive list but,
rather, should be considered a starting point to help in understanding the
various in situ techniques both in use currently and at the forefront of
development. In addition, review papers are referenced for the reader to nd
more information on techniques that are and are not discussed.

4. Current in situ measurement techniques


In depth understanding of the individual steps that determine rates of
reaction is required to design a new catalyst that can facilitate the necessary
reactions without imparting a negative impact on other elementary steps. The
catalysis science and engineering community is developing tools to study
catalysts under real-time working conditions, i.e. operando. Operando tools
measure catalyst activity/selectivity in terms of structure, kinetics, and

Catalysis, 2009, 21, 191218 | 197


This journal is
c The Royal Society of Chemistry 2009
dynamics, while providing nanoscale spatial and femto- to milli-second time
resolution.
Operando methodology aims to dene and characterize structure/function
relationships which must be interfaced with rate and dynamics measurements
of the elementary steps. Recent years have shown a marked increase in the
presence of spectroscopic investigations of catalytic reactions in literature
(see Catalysis Today, 113 issues 12). For example, operando techniques were
used to determine the temperature stability range of two NOx reduction
catalyst types, (NH4)[Co(H2O)2]Ga(PO4)3 vs. (NH4)[Mn(H2O)2]Ga(PO4)3.52
Fig. 5 shows that the catalyst with manganese changes in structural stability
around 673 K. Inspection of the catalyst with cobalt shows that there is no
structure modication at a temperature below 673 K.
It is imperative to monitor the details of chemical reactions at the molecular
level, which operando data can contribute to immensely through the use of
spectroscopy. Spectroscopic techniques have been used in the past mostly to
characterize fresh or used catalysts, obtaining structural information relating
to the bulk and surface of the solids. In addition, on-line gas analysis of

Fig. 5 The top graph represents surface evolution of (NH4)[Mn(H2O)2]Ga(PO4)3 catalyst


during the TPD to determine the stability temperature range. The bottom graph is the surface
behavior during NOx reduction on (NH4)[Co(H2O)2]Ga(PO4)3 at 623 K.52

198 | Catalysis, 2009, 21, 191218


This journal is
c The Royal Society of Chemistry 2009
product euent will ensure that the activity data obtained in the operando
reactor are consistent with those observed in a conventional reactor.
Since 2003, there has been an explosion of research using many and varied
operando techniques. Below is a discussion of three very powerful and
widely used techniques. They are FT-IR (and DRIFTS), Raman spectro-
scopy and high temperature, high pressure scanning tunneling microscopy.
While these are not the only techniques, they are a very good starting point
to gain a better understanding of the types of experimental tools used for
in situ catalyst study, because they represent the tools of choice employed by
many catalytic research groups.

Infrared spectroscopy
Since 1905, when Coblentz obtained the rst IR spectrum, vibrational
spectroscopy has become an important analytical research tool. This
technique was then applied to the analysis of adsorbates on well-dened
surfaces, subsequently moving towards heterogeneous reaction studies.
Terenin and Kasparov (1940) made the rst attempt to employ IR in
adsorption studies using ammonia adsorbed on a silica aerogel containing
dispersed iron. This led to a prediction by Eischens et al. from Beacon
Laboratories in 195653 that the IR technique would prove to be extremely
important in the study of adsorption and catalysis. For an excellent review
article in IR spectroscopy, see Ryczkowski54 and references therein; and for
a more recent review with applications, see Topsoe.55
IR spectroscopy is likely the most frequently used technique in studying
heterogeneous catalysis. Its conguration and use is fairly straight forward,
making it relatively simple to carry out in situ studies. Transmission
spectroscopy involves passing infrared radiation through a sample and then
measuring the extent of absorption. The information obtained provides
insight into important parameters, ranging from characterization of
supports to the nature of adsorbed molecules and reaction intermediates.
IR studies have helped establish correlations between observed IR para-
meters and the catalyst activity, allowing researchers to move towards a
more rational design of industrial catalysts. Fig. 6 shows a typical data set
that is obtained. The data shows the dierence in the spectra going from an
as-prepared, non-reacting sample to the sample in the working state. TOS is

Fig. 6 FT-IR spectra of catalytic surface during n-butane isomerization at 323 K taken every
16 minutes. This shows the dierence in IR spectra from before activation to during reaction.
The black lines show the spectra rst at the start of the reaction, and the next is at the maximum
reaction rate.56

Catalysis, 2009, 21, 191218 | 199


This journal is
c The Royal Society of Chemistry 2009
time-on-stream and it can be seen, for example, that at wave number 1602
there is a signicant growth in the absorption signal, which indicates the
formation of increasing amounts of water. The presence of water at more
than one frequency is due to water adsorption on dierent sites and varying
degrees of hydration.
Within the IR spectroscopy arena, the most frequently used techniques
are transmissionabsorption, diuse reectance, ATR, specular reectance,
and photoacoustic spectroscopy.57 A typical in situ IR system is shown in
Fig. 7. Choosing appropriate probe molecules is important because it will
inuence the obtained characteristics of the probed solid and the observed
structure-activity relationship. Thus, the probe molecules cover a range
from the very common to the very rare, in order to elucidate the eect of
dierent surfaces to very specic compounds (e.g. heavy water and deuter-
ated acetonitrile, CD3CN). The design of the IR cell is extremely important
and chosen to suit the purposes of each particular study. For catalytic
reactions, the exposure of catalytic metals must be eliminated in cell
construction, otherwise the observed eect of the catalyst may not be
accurate.

Fig. 7 Schematic IR reactor system enables time-resolved kinetic and IR spectroscopic data
measurements: (a) IR cell; (b) total set up.5759

200 | Catalysis, 2009, 21, 191218


This journal is
c The Royal Society of Chemistry 2009
Currently, in situ IR analysis is at the point where it is available as a
commercial product. Systems have been developed that operate over a wide
pressure range from 15 atm to 10 5 torr.54 The catalyst heating is isolated
and well controlled from room temperature up to 450 1C. It is low volume,
equipped with exchangeable windows that enable data collection over a
wide IR frequency range during static or ow conditions.

Diuse reectance infrared spectroscopy (DRIFTS)


In the IR spectroscopy eld, a majority of experiments are performed in the
transmissionabsorption and the diuse reectance IR spectroscopy
(DRIFT) mode. This has been done routinely in the UV-Vis and IR, using
integrating spheres and ellipsoids/paraboloids, respectively, to collect the
scattered radiation. Since diuse reection spectra are the result of the light
interacting with the sample in every conceivable way, the spectra may exhibit
features of transmission, external reection, and/or internal reection.
Spectra, generated from powders that range in size from 2 to 5 mm, can be
recorded by illuminating rough surfaces and scattered light can be collected
with appropriate optics for spectral analysis.
Diuse reectance is an excellent sampling tool for powdered or crystalline
materials in the mid-IR and near-IR spectral ranges.60 Heated reaction
chambers for diuse reectance allow the study of catalysis and oxidation
reactions in situ, and can evaluate the eects of temperature and catalyst
behavior. Scratching sample surfaces with abrasive paper and then measuring
the spectra of the particles adhering to the paper allows for analysis of
intractable solids. Perhaps one of the greatest additional benets is that this
system is amenable to automation.
As with transmission analysis, DRIFTS test samples are generally ground
and mixed with an IR transparent salt, such as potassium bromide (KBr),
prior to sampling. However, to ensure that there is no catalytic eect, most
systems use CaF2 or powdered crystalline carbon (diamond). One drawback
is the presence of reststrahlen bands whose contrast can be reduced by

Fig. 8 (a) DRIFTS spectra of catalyst surface in NO ow. Absorbance increases with time.
(b) In situ Raman spectra measured while methanol owing over catalyst.61

Catalysis, 2009, 21, 191218 | 201


This journal is
c The Royal Society of Chemistry 2009
sample dilution. Reststrahlen bands, are generated when there is nearly
100% reectivity in a frequency, prohibiting light propagation into the
sample being studied. In addition, sample particle size and angle of
illumination and/or observation and surface roughness can aect the
observed spectra.
An advantage of DRIFTS is that it is a fast, non-destructive technique
and does not require the preparation procedures associated with conven-
tional IR transmission methods. DRIFTS is also well suited to analyze
strongly absorbing species with a resolution into the nanograms, with the
ability to probe only the rst few layers of the sample, operations with
samples as little as micrograms. The technique is very quantitative yet
require proper calibration, which is very time consuming and dicult to
obtain. Therefore, most of the reliable data are averages of many measure-
ments on multiple samples to minimize error from physical property
variations. For sensitivity and other reasons, it is imperative to optimize
the instrument parameters for the system being studied. Factors such as
resolution, apodization and zero lling need to be optimized for the physical
state of the targeted sample otherwise the signal to noise ratio will be poor.62

Raman spectroscopy
The previous discussion focused on IR spectroscopy, a method that involves
probing molecules with photons that resonate its vibrational frequencies.
Raman spectroscopy, however, is based on detection of scattered photons
from absorbed species which interfere weakly with signals from the gas phase.
This makes Raman analysis complementary to IR analyses. It is an excellent
in situ technique for catalyst structure, but is not very surface sensitive.
Moreover, the common alumina supports of most catalysts are weak Raman
scatterers. This positions Raman to be well suited for in situ analysis because
of its wide frequency range, from 50 to about 5000 cm 1 and high resolution
of nearly 1 cm 1. Since the gas phase yields almost no signal, Raman is
practical for studying surface reactions. The information obtained from a
single measurement gives insight into the molecular reactants and the solid
catalyst surface. Raman has been most valuable in analysis of oxide catalysts,
since the metal-oxygen vibration frequency is dicult to probe with IR
because it is typically between B400 and 1000 cm 1.
Not all vibrations that get excited by the incident light are observable,
such as the weak vibrations of a highly polar moiety (i.e. the OH bond). An
external electric eld is not capable of inducing large changes in the dipole
moment, and stretching or bending the bond does not change this. A
vibration will be Raman active if the molecule changes shape, thus changing
the polarizability of the molecule. Typical strong Raman scatterers have
distributed electron clouds, such as carboncarbon double bonds. The
pi-electron cloud of the double bond is easily distorted in an external
electric eld. Bending or stretching the bond changes the distribution of
electron density substantially, and causes a large change in induced dipole
moment.
Raman typically provides information on a catalysts crystallinity,
coordination of site structure and spatial distribution of phases through

202 | Catalysis, 2009, 21, 191218


This journal is
c The Royal Society of Chemistry 2009
the sample. These characteristics have rendered Raman very useful for
catalyst formulations, especially using molybdenum, vanadium and
tungsten. The primary focus has been to identify the surface species formed
during catalyst testing, pretreatment, and preparation. The information
obtained has established a good understanding of hydro-treating catalysts
and some partial oxidation catalysts.
Recently, UV laser stimulation of catalyst samples has been developed to
overcome the problem of interference by coke (carbon deposition) on
catalysts.63 Fig. 9 shows a typical Raman data set that was obtained for
carbon deposition as a function of temperature. To explore dierent coke
formation behavior, the reaction of propene on a zeolite was performed.
The spectra obtained were (A) C3H6/He ow at 773 K for 3 h; (B) O2 ow at
773 K for 1 h and (C) O2 ow at 873 K for 1 h. This data shows that most of
the carbon, identied as polyaromatic and pregraphite, can be removed at
773 K with oxygen. However there is still carbon present as identied by the
broad band at 1610 cm 1 suggesting that carbon is in a more inert form
such as coke. Not until the temperate is taken to 873 K with oxygen is that
carbon removed.
This initial setup proved to be problematic because of the strong UV local
heating. Subsequently, Prof. Stair developed a uidized bed to avoid local
heating. The result of that work is shown in Fig. 10.
Similarities in the spectra of n-heptane and the catalyst adsorbate suggest
that n-heptane is the dominant species on the H-ZSM-5 surface. Small peaks
on the catalyst surface spectra indicate the presence of coke formation.64
Raman scattering cross sections are small (B10 28 cm2), thus uores-
cence and luminescence can easily interfere with the data. This is parti-
cularly acute with surface carbon, which uoresces strongly when the
Raman is excited by visible wavelength lasers. In addition, catalyst

Fig. 9 Fluidized bed UV Raman scattering spectra of adsorbates on HZSM5 zeolite


catalyst.63

Catalysis, 2009, 21, 191218 | 203


This journal is
c The Royal Society of Chemistry 2009
Fig. 10 Top spectrum: n-heptane/H-ZSM-5 recorded using the in situ uidized bed reactor.
Laser power, 2 mW. Collection time, 7000 s. Bottom spectrum: Liquid n-heptane. 2 mW; 60 s.64

components that are typically considered stabilizers or promoters can


strongly uoresce and washout the Raman signal. This has led to the use
of Raman on precious metal catalysts supported by alumina. It was found
that the UV incident laser was modifying the catalyst by initiating reactions
thermally and photochemically. To avoid modication to the catalyst and
produce high quality reliable data, a circulating uidized bed was
implemented to minimize the inuence of the incident beam. For a good
discussion on the capability of this technique, see In situ Spectroscopy in
Heterogeneous Catalysis, chapter 5.
Raman techniques were modied to enable investigation of adsorbates on
surfaces that were initially thought to be of insucient sensitivity. Surface
Enhanced Raman Scattering (SERS) is used to investigate the vibrational proper-
ties of adsorbed molecules. Metal surfaces must be of high reectivity and suitable
roughness. This method revealed that certain molecules and appropriately
prepared metal surfaces, such as silver and gold, could display Raman scattering
cross-sections many orders of magnitude greater than isolated molecules.
Professor Israel Wachs of Lehigh University has employed operando
characterization for surface-structure-activity relationship over metal oxide
catalysts and has shown numerous results for a range of conditions and
applications.6569 The Raman spectra (532, 442 nm and 325 nm) of bulk
mixed metal oxides ((Fe2(MoO4)3, Cr2(MoO4)3, and Al2(MoO4)3 and
FeVO4, AlVO4, CrVO4, and AgVO3) were compared on the same spectro-
meter to determine the aects of excitation energy variation impacts the
resulting Raman spectra.
The studies showed asymmetric stretching of some bulk MO4 coordinated
sites were UV sensitive and bulk terminal MQO vibrations and bending

204 | Catalysis, 2009, 21, 191218


This journal is
c The Royal Society of Chemistry 2009
modes were most sensitive in the visible spectra. The vibration modes of
bridging VOAl and VOFe bonds were observed for bulk AlVO4 and
FeVO4 for the rst time and demonstrated that use of uidized bed sample
systems was not necessary.67 In addition, the advancement and application
of spectroscopic techniques for understanding mixed metal oxide catalysts
has yielded insights into the electronic and molecular structures of the active
catalytic sites. One very signicant nding has been the catalytically active
sites of mixed metal oxides is the amorphous metal oxide phases that
have been found to be present for many applications.68
Professors Alex Bell and Enrique Iglesia of the University of California,
Berkeley have used UV-Vis DRIFTS and Raman spectroscopy to elucidate
the role of many catalytic systems ranging from mixed metal oxides to
precious metal formulations for applications ranging from dehydrogenation
of hydrocarbons to oxidation of alkanes to the role of exposed species on
dispersed surfaces.7074
For example, the oxidative dehydrogenation of ethane and propane was
examined via UVvisible and Raman spectra. The study investigated the
catalytic properties vanadia formulations that possessed a range of VOx surface
species density (1.434.2 V/nm2) on an Al2O3 support. The observations
showed increased surface densities, greater than 2.3 V/nm2, favored
two-dimensional polyvanadates. At lower surface densities, ca. 2.3 V/nm2,
predominately isolated monovanadate species were observed. Further
increasing surface densities to more than 7.0 V/nm2 yielded V2O5 crystallites.71

High-pressure, high-temperature STM


A scanning tunneling microscope (STM) looks at individual atoms on the
surface of a material with the use of a current. The ideal sample is at and
clean, just a slice through a plane in the materials crystal structure, and it
must conduct electricity. The measuring tip is a needle, so sharp that it
terminates in a single atom. The tip does not actually touch the sample,
rather it is approximately a few Angstroms (Batomic diameters) away and
is held at zero voltage, or ground. Meanwhile, a bias voltage is placed on
the sample, on the order of a few millivolts to a few volts, inducing a
tunneling current to ow between the tip and the sample. This current is
exponentially dependent upon the distance between the tip, therefore a
small change in the distance between the tip and the sample greatly changes
the current.
The proximate placement of the tip over the surface allows measurements to
probe the surface without impacting it, while ignoring the gas-phase environ-
ment. Therefore, the information obtained relates to the rst few metal surface
layers and any absorbed species. Jensen et al.75 have developed this technique
over the temperature range 25 to 400 1C and the pressure range 5  10 10 Torr
to 1 atm, falling in the range of many operating conditions.7679
The apparatus combines a high pressure, high temperature scanning
tunneling microscope (HPHT-STM) with a UHV surface analysis chamber.
This combined system is coupled with an improved heating system,
signicantly improving in situ study by allowing sample transfer without
contamination by exposure to air. The system is divided into three parts: a

Catalysis, 2009, 21, 191218 | 205


This journal is
c The Royal Society of Chemistry 2009
Fig. 11 High pressure, high temperature STM. Sections: A-UHV sample manipulator with
indirect heating and cooling, B-variable temperature STM, C-magnetically coupled linear-
rotary transfer arm, D-optical access for analyzers, E-pneumatic air legs and suspension frame,
F-turbo molecular pump, G-ion pump.75

UHV surface preparation and analysis chamber, a variable pressure STM


chamber, and a load lock and sample transfer system. Fig. 11 shows a
schematic of the system.
The main modication that enables the system to analyze in situ reactions
is the custom built chamber for the STM stage with indirect heating via an
electron beam. Therefore, the sample can be brought to the desired
temperature and pressure without disturbing surface interactions. While
this technique is primarily used for model catalysts to be studied, it provides
very good insight into the mechanisms present over a range of pressures.
Fig. 12 is representative of the results obtained with this technique for a
Pt(111) single crystal cleaned in UHV via Ar+ sputtering. It shows that
there are very dierent mechanisms occurring at dierent pressures. For
example at a CO pressure of 1.0 Torr induced mobility of Pt atoms to form
clusters, indicating signicant weakening of the surface PtPt interactions.
Another test (not shown) indicates a Rh(111) surface at 100 1C and 20 Torr
showed that the Rh was migrating from the step edges to ll in nearby holes.
These observations can give very insightful clues into the mechanisms that
are occurring on the surface during reaction and provide insight into
performance changes. This allows researchers to identify likely reasons
for the change, and take measures to facilitate desired changes.
The selection of tip material is a very important consideration for this
technique. Tungsten was shown to be the best choice for most application,
but becomes unstable in oxidizing conditions. Gold tips are useful for room
temperature STM studies in any gas environment, but are unstable when
heated even slightly above room temperature. Therefore, the development
of a tip material that is good for all conditions will broaden the use of this
very powerful technique. The techniques just discussed present a snapshot
of the state-of-the-art tools the catalysis community uses to elucidate a
mechanistic understanding. It is certain that renements and modications
to those techniques will be made that will provide higher delity data with
enhanced resolution. In addition, new techniques will be developed to
complement the tools in existence. To date, there is no one technique that

206 | Catalysis, 2009, 21, 191218


This journal is
c The Royal Society of Chemistry 2009
Fig. 12 Looking from scans (a) to (d) one can observe induced mobility of Pt away from the
step edges. The Pt forms islands shown by the right arrow in (d). The formation of the island
results in leaving a single atom deep hole as shown by the left arrow in (d).75

can answer all questions about the catalyst structure activity and structure
selectivity relationship.
As noted above, it should be realized that understanding the activity on a
surface is not the only issue needed to fully characterize the catalytic
system. There must be an understanding of how the surface reactions
produce intermediates that can desorb and initiate gas phase reactions close
to the surface. To that extent, the author is developing a catalytic shock
tube technique to probe the interface of the catalyst surface and the
immediate gas phase layer, bridging the gap between surface mechanisms
and gas phase mechanisms.

Combined techniques and condensed phase


It has been quickly recognized that the individual operando techniques can
be combined to yield a more complete picture of the catalytic reaction
sequence. In addition, since many reactions of industrial signicance occur
in the liquid phase, it is important that techniques are developed to probe
and monitor those systems under conditions that at least keep the reactants,
intermediates and products in their actual operating states or phases. This
has resulted in researchers utilizing a multitude of techniques, some in situ
and ex situ, to obtain a more complete understanding of the entire catalytic
cycle.
Newton et al. developed a combination of DRIFTS, EXAFS and electron
impact mass spectrometry (MS) which enables structure, function and

Catalysis, 2009, 21, 191218 | 207


This journal is
c The Royal Society of Chemistry 2009
reactivity in a dynamic mode to be studied.80 It has recently been used to
determine the redox cycle for alumina supported Rh catalysts in powder
form. Transmission EDE and DRIFTS were simultaneously applied to the
sample area of the catalyst and both systems penetrated 500 mm into the
sample. To obtain structural data on very small changes and normalization
procedure using a reference sample, as opposed to air, was developed. The
reference sample was 0.5 wt% Pt/AL2O3 prepared the same way the Rh test
sample was prepared, thus eliminating any spectrum from the structure.
The authors describe a very nice sequence of reactions that occurs at
interface/defect sites associated with oxidized Rh, however are not involved
in their formation. They make a compelling case that the NO+ species
appears to be removed to facilitate subsequent reduction of the Rh oxide
phase. However they do not observe NO production in the MS, thus initial
reaction of the NO+ species needed to be determined. Synthesizing all the
information provided by the combined equipment the authors postulate a
Rh(NO+) species is formed due to continued penetration of oxygen into
the remainder of the Rh particle after the completion of an oxidized layer.
The continued penetration of oxygen into the particles yields defect sites
that are lled via the Rh(NO+) entity. This results in a mechanism for
promoting Rh oxidation and simultaneously encouraging the formation of
N2O. The transient observations from the MS indicated this nicely. The
chosen temperature of 673 K was limited by sealing and authors acknow-
ledge that the Rh/Al2O3 sample was not very challenging material to study
and are uncertain of the performance of the method on more challenging
materials.
Another technique that combines MS with optical monitoring for oper-
ando investigation is the DRIFT-MS apparatus of Burchs laboratory.
Meunier et al.81 published a very nice study on the role of intermediate
species for the WGS reaction using DRIFT-MS with isotope probing. The
technique of steady-state isotopic transient kinetic analysis (SSITKA) showed
how the intermediate chemical species formate transitions from a spectator
species to a major player at temperatures near 473 K over ceria supported
catalysts. The investigation used 13CO, monitored by DRIFTS-MS, to probe
the conversion of CO during the WGS reaction and identify reaction
intermediates and their role in the conversion process. The isotopic exchange
of the formate was found to be signicantly slower than that of the CO2 (g)
product at 433 K. This lead to the conclusion that the formate species
observed by DRIFT measurements did participate in the reaction pathway.
However, at 493 K the 50% exchange time for the formate intermediate and
the CO2 (g) product species were found to be identical. This observation
conrmed that the formate species was most likely the main surface inter-
mediate in the formation of CO2 (g).
Using the powerful combination of MS coupled to DRIFTS allowed
isotopic analysis to very convincingly show that over a narrow temperature
range formates change from being spectator species to being active reaction
intermediates. This clearly shows that catalytic surfaces can abruptly change
resulting in the reaction mechanism of a surface species being dramatically
dierent. Therefore, the reaction mechanism based on data using dierent
reaction mixtures or reaction temperatures, or even dierently prepared

208 | Catalysis, 2009, 21, 191218


This journal is
c The Royal Society of Chemistry 2009
catalysts having the same nominal composition can be very dicult to
generalize over a wide range of conditions.
Multinuclear MRI has recently been used to analyze preparation techni-
ques and reactions in heterogeneous catalysis.82 Similar to other optical
techniques it is non-invasive and nondestructive and can have chemical
specicity. It can measure diusivity, pore size distribution and temperature
among other parameters and thus can be used to understand the progress of
a reaction aected by heat and mass transport. Unlike many optical
techniques it monitors processes over the entire sample, thus getting the
aggregate performance of the system. Using typical MRI a spatial resolu-
tion of 510 mm is achieved and thus cannot observe pores. Yet, using a
technique that combines relaxation mediated sensitivity of the observed
signal, it is possible to monitor pore structures. Combination of NMR
porosimetry techniques with MRI allows one to obtain local pore size
distributions of various porous objects while simultaneously observing the
deposition process during catalyst preparation. This is an advantage over
many techniques. For example STM cannot penetrate pores since it must
use conductive surfaces.
This technique was successfully used to monitor the deposition of catalyst
within a support pellet and clearly distinguish the amount of penetration
there was into the pellet. Therefore, one can quickly obtain a time required
to produce a catalyst pellet that is of the egg-shell type or completely
saturated. Moreover, it is possible to very nely resolve the thickness of the
catalyst penetration into the particle, resulting in good control of the
nished catalyst.
This technique has been successfully used to monitor the liquid phase
hydrogenation of unsaturated hydrocarbons over Pt/Al2O3 and Pd/Al2O3
catalysts with 23 second resolution time. The temperature of the reaction
was scanned from 100250 1C. The observations showed an oscillating
regime in the pellet with constant external conditions. It was a clear sign of
the coupling eect between heat and mass transport of a liquid reacting
system within a catalyst pellet. The authors interpreted the observations to
be phase transfer within the catalyst pellet during the reaction cycle.
The images obtained during an experiment demonstrated the ignition of
individual catalyst beads throughout the reactor. These beads disappear
from the image completely or partially since only the liquid-lled catalyst
particles and their parts are observed in the images. The oscillations of the
liquid front were observed to persist for long periods of time and were only
possible in the Operando mode.
Another technique that bridges the phase gap is Fluorescence micro-
scopy.83 Fluorescence microscopy is capable of studying diusion and cata-
lysis in zeolite pores with high 3D spatial and temporal resolution. This
technique was successfully used to monitor acid catalyzed transformation of
uorogenic organics on a mordenite zeolite. The uorescence microscopy
technique is basically a far eld microscope that can be deployed in two types
of congurations. The rst is confocal that focuses light to a diraction limited
spot. Fluorescence light is screened by a pin hole which only enables in-focus
light to pass. That light is then subsequently measured by a photomultiplier
tube. Typical signal to noise ratios are near 10 to 1 and time resolution is on

Catalysis, 2009, 21, 191218 | 209


This journal is
c The Royal Society of Chemistry 2009
the order of microseconds. Usually emission spectra can be measured with a
CCD combined with a polychromator to monitor single molecular events, the
target area can be scanned at a high frequency. This approach is useful to get
highly temporally resolved dynamic data on individual molecules and can
provide three dimensional structure information.
The second conguration is the wide eld operation mode where the
detector is a CCD camera that enables hundreds of mm2 to be monitored.
The spots resolved are still diraction limited but have a reduced depth
dimension, thus lower resolution but much higher excitation volumes
covering hundreds of mm2. Using appropriate tting of the point spread
function (PSF), many molecular events can be monitored simultaneously as
well as achieving a precision below the diraction barrier for the molecules
localized. For this operation mode the signal to noise ratio is less than the
confocal mode and near 5 to 1. Due to the limitations of the processing time
of the CCD unit, typical time resolutions are in the millisecond range.

Catalytic shock tube


This new technique incorporates a catalyzed short contact time (SCT)
substrate into a shock tube, Fig. 13. These SCT reactors are currently used
in industry for a variety of applications, including fuel cell reformers and
chemical synthesis.8490 The combination of a single pulse shock tube with
the short contact time reactor enables the study of complex heterogeneous
reactions over a catalyst for very well dened regimes in the absence of
transport eects. These conditions initiate reaction in a real environment
then abruptly terminate or freeze the reaction sequence. This enables
detection of intermediate chemical species that give insight into the reaction
mechanism occurring in the presence of the chosen catalyst. There is no
limitation in terms of the catalyst formulations the technique can study.
Shock tubes have long been used to determine both reaction rate data and
mechanistic information for various homogeneous reactions. The main
attribute of a shock tube apparatus is the ability to rapidly, B0.5 ms, raise

Fig. 13 Schematic of proposed apparatus for catalyst surface reaction studies in the absence of
transport eects.

210 | Catalysis, 2009, 21, 191218


This journal is
c The Royal Society of Chemistry 2009
the temperature and pressure to reacting conditions, and drop them back to
ambient conditions almost as rapidly. This nearly provides a step function
where reacting conditions can be obtained for precisely dened periods of
time.
The apparatuss step change from ambient to desired reaction conditions
eliminates transport eects between catalyst surface and gas phase reactants.
Using catalytic reactors that are already used in industry enables easy transfer
from the shock tube to a ow reactor for practical performance evaluation and
scale up. Moreover, it has capability to conduct temperature- and pressure-
jump relaxation experiments, making this technique useful in studying
reactions that operate near equilibrium. Currently there is no known
experimental, gas-solid chemical kinetic method that can achieve this.
The primary objective of the shock tube tests is to enable only reactions
that occur at the catalyst surface, which is done by removing all aspects of
mass transport and to quickly stop all reactions in precisely dened time
duration. The right choice of a test mixture composition can help to isolate
the chemical reaction under investigation, while varying initial conditions
allows for a wide range of concentration, pressures and temperatures.
The shock tube can be operated in one of two modes, either as a shock
expansion tube (SET) or reected shock tube (RST). Fig. 14 shows the wave
diagrams (left) of an RST operating mode and temperature/pressure time
trace (right) for the system developed.
The incident shock wave moves down the tube, heating and accelerating
the test gas. In RST mode, the shock hits an end plate and reects back to
the test gas, further heating the gas and initiating stagnation conditions.
Subsequently, a rarefaction wave travels down the tube and quenches all
further reactions.
To ensure the system is probing reactions in a kinetically controlled regime, the
reaction conditions must be calculated to determine the value of the Wiesz-Prater
criterion.91 This criterion uses measured values of the rate of reaction to
determine if internal diusion has an inuence. Internal mass transfer eects
can be neglected for values of the dimensionless number lower than 0.1. For
example, taking a measured CPOX rate92 of 5.9  10 6 molCH4 s 1 g 1 results

Fig. 14 A wave diagram showing how the test time is dened.

Catalysis, 2009, 21, 191218 | 211


This journal is
c The Royal Society of Chemistry 2009
in a value of 0.014 at the conditions targeted to be studied. The results of
incorporating this measured rate into the Wiesz-Prater calculation yields a value
of 0.0011 in the most severe conditions related to concentration and temperature.
In addition, the external diusion mass transfer resistance needs to be evaluated
by calculating the Damkohler number. All test conditions have shown it is
negligible even in the most stringent conditions of temperature and pressure.
To further ensure there will be no contribution of transport within the
washcoat, calculations done based on a method of transient analysis for
catalytic ignition by Balakotaiah should be done.93 Fig. 15 shows the results
of the calculation in comparison to where washcoat diusion occurs. By
maintaining a nominal washcoat thickness of 7 micron, the catalyst will
operate in the kinetic regime for all reaction conditions considered.
These calculations were done using the typical thickness of a washcoated
SCT substrate (B7 micron) and the eective diusivity based on the
diusion coecient (4.148  10 6 m2 s 1) calculated for a real gas matrix
and taking representative values of the tortuosity (3.0), porosity (0.4) and
constriction (0.8) factors.
Fig. 16 can help visualize the location where the reaction is likely to take
place based on the above discussion and calculations. While the entire
washcoat is impregnated with catalyst, which must be done to achieve a
robust support on the SCT, the calculations indicate the reaction will only
take place at the surface sites near the top on the sketch and on the SEM.
To date, the only shock tube apparatus equipped to study surface
reactions is the KIST facility at ATK GASL in New York. The tests done
so far have studied methane oxidation, CH4 + 2O2 - CO2 + 2H2O on the
surface of an SCT ferrous-based reactor impregnated with platinum based
catalyst. To isolate the eects of the screen and the catalyst on the reaction,
three types of tests were run: catalyzed screen with combustible gases,

Fig. 15 Analysis showing experimental conditions do not have diusion in the washcoat.

212 | Catalysis, 2009, 21, 191218


This journal is
c The Royal Society of Chemistry 2009
Fig. 16 Schematic and SEM image of the catalyst surface studied to visualize the region the
catalytic shock tube probes.

Fig. 17 Representative data from catalytic shock tube: pressure trace of a Mach 1.3 shock.

un-catalyzed screen with combustible gases, and catalyzed screen with inert
gases. The pressure traces in Fig. 16 are examples of the test data, showing
the pressure at a point just downstream of the SCT. The data shows the
shocks arrival and followed by rapid return to non-reacting steady state
conditions. The three tests enable comparison between non-reacting and
reacting conditions. The resulting pressure traces show that both the
catalyzed screen with inert gas and the un-catalyzed screen with combustible
gases behave the same. The reactive mix of the catalyzed screen with
combustible gases exhibits dierent pressure behaviors, holding the pressure
longer after an observed ignition delay. As evidenced from Fig. 17 the
tests show that after the ignition delay time, there was nearly 1% conversion
of the methane, giving o heat and manifesting as a temperature increase
(20 K) and pressure increase (1.5 psi).

Computational techniques
To be sure all experimental methods need to be complemented by theoretical
techniques. The calculational techniques started with ab-initio and quantum
calculational methods, such as MOPAC and GAMESS. These methods focus
on the solution for the wave functions of the system being modeled. Those
computations enabled calculations to be done in a sequence of frozen
congurations of the catalyst and the gas phase molecules approaching the
surface. The calculations produced thermodynamic energetic and entropic
eects as the reaction coordinate changed, bring a reactant closer to the

Catalysis, 2009, 21, 191218 | 213


This journal is
c The Royal Society of Chemistry 2009
surface or attempting to remove an adsorbed species from the surface. The
largest hurdle in the early days of modeling was the accurate simulation of the
surface.
The development of density functional theory (DFT) since then has
become very powerful and is widely used.94104 DFT methods solve for
the electron density and like ab-initio methods requires a basis set to
represent the atomic and molecular orbitals. The main idea of DFT is to
describe an interacting system by its electron density and not its wave
function (ab initio). For N electrons in a solid, which obey the Pauli
principle and repulse each other via the Coulomb potential, this means
that the basic variable of the system depends only on three dimensional
Cartesian coordinates x, y, and z. While DFT in principle gives a good
description of ground state properties, practical applications of DFT are
based on approximations for exchange-correlation potential. DFT has
provided a good description of the geometries for a wide range of transition
metal catalysts. It is imperative to predict the activation energies of the
reactions of interest to within a fraction of the actual value otherwise errors
in predicting absolute activities will be a factor of 10 or more.
Si-Ahmed et al. reported on the use of DFT calculations to support and
interpret in situ TP_Raman and TPR experiments.105 Since the role of alkali
dopants on the structure-activity relationship of supported vanadia catalysts is
not fully understood, the investigation focused on providing a molecular
understanding via DFT with complementary experimental evidence. The
Raman intensity was computed with respect to the electric eld by numerical
dierentiation of dipole derivatives. Setting the potassium to vanadium ratio to
0.5, whereas in the prepared experimental sample it was 0.45, enabled various
initial structures to be investigated. It was shown the main eect of the
potassium in the vicinity of two VQO bonds was an elongation from 1.58
to 1.62 A. The DFT calculations guided the interpretation of hydrogen TPR
spectra. The rst reduction peak was attributed to monomeric and polymeric
vanadium oxide being reduced, which was conrmed by Raman spectroscopy
of the prepared sample. The adsorption of hydrogen was calculated to be
exothermic in all tests with the undoped model being favored over the K-doped
system. For example the heat of adsorption for one H was 4.08 eV for V2O5
and 2.94 eV for the K-doped surface. This helped explain that the eect of
potassium was to decrease the reducibility of the system that was also observed
in the experimental TPR. That is during tests, the TPR temperatures had to
increase to achieve reduction with systems doped with potassium.
Absolute values, especially for high frequencies, were overestimated using
DFT and the formal oxidation state could not be determined due to the
delocalized description of the charge density in the methodology. It was
possible to compare frequency shifts between potassium doped and bare
surfaces. In fact other experiments of 51V NMR-MAS spectra showed
potassium addition strongly changed the vanadium sites enough to form
alkali vanadates on V/TiO2. The results of this study showed the incipient
formation of bulk VKO compounds at the highest vanadium and
potassium loading on titania where the potassium modied surface devel-
oped a Raman band near 800 cm 1, which was in agreement with the DFT
calculations.

214 | Catalysis, 2009, 21, 191218


This journal is
c The Royal Society of Chemistry 2009
An overview of the basic principles of DFT, advantages and disadvantages
as well as comparison to using molecular orbital simulations can be found in
the text Essentials of Computational Chemistry: Theories and Models.106

5. Future directions
The review just presented has provided an overview of the more well used
techniques for catalyst mechanism investigation. As faster computer power
becomes available and more precise engineering, i.e. fabrication of nano-
structures, becomes more developed catalyst science will progress to the next
level. The next 30 years will include coupling of multiple investigation
techniques with sophisticated computational analysis to further understand
the role of catalyst surfaces for many industrially relevant reaction sequences.
It is likely the next signicant area of progress will be modication of
existing catalysts to achieve better selectivity toward desired reactions.
Activity improvements will continue but at a slower pace. The advent of
nano-engineering with unprecedented visualization of surfaces is leading in
the direction of a priori designed catalysts. As our understanding of multiple
classes of reactions grows, it is very likely within the next 50 years a catalyst
will be designed from rst principles to achieve a desired selectivity. Once
that occurs, the next frontier will be to design a catalyst for a desired
durability with predetermined activity and selectivity.

References
1 The Collected Works of Sir Humphry Davy, ed. J. Davy, Smith, Elder, and Co.,
Cornhill, London, 1840.
2 L. Gmelin, Hand-book of Chemistry, Harrison and Sons, London, 1862.
3 S. Westerberg, C. Wang and G. A. Somorjai, Surf. Sci., 2005, 582, 137144.
4 J. Boeseken, Recl. Trav. Chim. Pays-Bas Belg., 1907, 26, 285288.
5 J. Boeseken, Recl. Trav. Chim. Pays-Bas Belg., 1911, 29, 330339.
6 M. Boudart, Actas Simp. Iberoam. Catal., 1984, 1, 318.
7 M. Boudart and G. Djega-Mariadasson, Kinetics of Heterogeneous Catalytic
Reactions, 1984.
8 M. Boudart and D. G. Loer, J. Phys. Chem., 1984, 88, 5763.
9 G. C. Bond, Acc. Chem. Res., 1993, 26, 490495.
10 H. S. Taylor, Proc. R. Soc. London. Ser. A, 1925, 108, 105111.
11 D. Farin and D. Avnir, J. Am. Chem. Soc., 1988, 110, 20392045.
12 J. J. Carberry, J. Catal., 1987, 107, 248253.
13 R. D. Cortright and J. A. Dumesic, Adv. Catal., 2001, 46, 161264.
14 M. J. Castaldi, A. M. Vincitore and S. M. Senkan, Combust. Sci. Technol.,
1995, 107, 119.
15 M. J. Castaldi and S. M. Senkan, Combust. Sci. Technol., 1996, 107, 141150.
16 M. J. Castaldi and S. M. Senkan, Ullmans Encyclopedia, 6th edn., 2002.
17 M. J. Castaldi, N. M. Marinov, C. F. Melius, J. Huang, S. M. Senkan,
W. J. Pitz and C. K. Westbrook, Symposium (International) on Combustion,
[Proceedings], 1996, 26th, 693702.
18 M. J. Castaldi and S. M. Senkan, Combust. Sci. Technol., 1996, 116117,
167181.
19 M. J. Castaldi and S. M. Senkan, J. Air Waste Manage. Assoc., 1998, 48, 7781.
20 N. M. Marinov, M. J. Castaldi, C. F. Melius and W. Tsang, Combust. Sci.
Technol., 1998, 131, 295342.

Catalysis, 2009, 21, 191218 | 215


This journal is
c The Royal Society of Chemistry 2009
21 N. M. Marinov, W. J. Pitz, C. K. Westbrook, M. J. Castaldi and S. M. Senkan,
Combust. Sci. Technol., 1996, 116117, 211287.
22 N. M. Marinov, W. J. Pitz, C. K. Westbrook, A. M. Vincitore, M. J. Castaldi,
S. M. Senkan and C. F. Melius, Combust. Flame, 1998, 114, 192213.
23 J. V. Lauritsen, R. T. Vang and F. Besenbacher, Catal. Today, 2006, 111, 3443.
24 H. Topsoe, B. Hinnemann, J. K. Norskov, J. V. Lauritsen, F. Besenbacher, P.
L. Hansen, G. Hytoft, R. G. Egeberg and K. G. Knudsen, Catal. Today, 2005,
10708, 1222.
25 H. J. Freund, Surf. Sci., 2007, 601, 14381442.
26 H. J. Freund, Top. Catal., 2008, 48, 137144.
27 H. J. Freund, M. Baumer, J. Libuda, T. Risse, G. Rupprechter and
S. Shaikhutdinov, J. Catal., 2003, 216, 223235.
28 R. Meyer, C. Lemire, S. K. Shaikhutdinov and H. Freund, Gold Bull., 2004,
37, 72.
29 M. S. Chen, D. Kumar, C. W. Yi and D. W. Goodman, Science, 2005, 310,
291293.
30 D. W. Goodman, Chem. Rev., 1995, 95, 523536.
31 D. W. Goodman, J. Catal., 2003, 216, 213222.
32 B. K. Min, A. K. Santra and D. W. Goodman, Catal. Today, 2003, 85, 113124.
33 K. Andersson, A. Nikitin, L. G. M. Pettersson, A. Nilsson and H. Ogasawara,
Phys. Rev. Lett., 2004, 93, 196101.
34 S. P. Asprey, N. M. Rice and B. W. Wojciechowski, Catal. Today, 1997, 36,
209226.
35 N. M. Rice and B. W. Wojciechowski, Catal. Today, 1997, 36, 191207.
36 B. W. Wojciechowski, Catal. Today, 1997, 36, 167190.
37 M. Lyubovsky and L. Pfeerle, Catal. Today, 1999, 47, 2944.
38 J. T. Gleaves, G. S. Yablonskii, P. Phanawadee and Y. Schuurman, Appl.
Catal., A, 1997, 160, 5588.
39 Application: US Pat., 93-561245376335, 1994.
40 Application: US Pat., 90-5132885039489, 1991.
41 J. T. Gleaves, A. G. Sault, R. J. Madix and J. R. Ebner, J. Catal., 1990, 121,
202218.
42 Application: US Pat., 84-6820284626412, 1986.
43 S. Fabricant, in Student Life: The Independent Newspaper of Washington
University in St. Louis Since 1878, St. Louis, 2006, vol. 128.
44 D. Constales, S. O. Shekhtman, G. S. Yablonsky, G. B. Marin and J. T. Gleaves,
Chem. Eng. Sci., 2006, 61, 18781891.
45 P. Phanawadee, M. Phongaksorn, N. Chaimongkol, A. Jaree and J. Limtrakul,
Chem. Eng. J. (Amsterdam, Neth.), 2005, 115, 5162.
46 S. O. Shekhtman, G. S. Yablonsky, J. T. Gleaves and R. R. Fushimi, Chem.
Eng. Sci., 2004, 59, 54935500.
47 D. Wang, J. Chin. Chem. Soc. (Taipei, Taiwan), 2003, 50, 551555.
48 J. A. Delgado, T. A. Nijhuis, F. Kapteijn and J. A. Moulijn, Chem. Eng. Sci.,
2002, 57, 18351847.
49 T. A. Nijhuis, F. Kapteijn and J. A. Moulijn, Chem. Eng. Sci., 2001, 56, 3927.
50 D. Constales, G. S. Yablonsky, G. B. Marin and J. T. Gleaves, Chem. Eng.
Sci., 2001, 56, 19131923.
51 S. O. Shekhtman, G. S. Yablonsky, S. Chen and J. T. Gleaves, Chem. Eng.
Sci., 1999, 54, 43714378.
52 O. M. F. Romero Sarria, P. Bazin, J. Saussey, J. Lesage, A. Guesdon and
M. Daturi, Catal. Today, 2006, 113, 8793.
53 R. P. Eischens, S. A. Francis and W. A. Pliskin, J. Phys. Chem., 1956, 60, 194201.
54 J. Ryczkowski, Catal. Today, 2001, 68, 263381.

216 | Catalysis, 2009, 21, 191218


This journal is
c The Royal Society of Chemistry 2009
55 N.-Y. Topsoe, Catal. Today, 2006, 113, 5864.
56 F. C. J. B. S. KLose, P. Joshi, A. Trunschke, R. Schlog, I. R. Subbotina and
V. B. Kazansky, Catal. Today, 2006, 116, 121131.
57 J. A. Lercher, C. Grundling and G. EderMirth, Catal. Today, 1996, 27, 353376.
58 J. A. Lercher, V. Veefkind and K. Fajerwerg, Vib. Spectrosc., 1999, 19, 107121.
59 G. Mirth, F. Eder and J. A. Lercher, Appl. Spectrosc., 1994, 48, 194197.
60 M. P. Fuller and P. R. Griths, Anal. Chem., 1978, 50, 19061910.
61 R. F. Howe, in In situ Spectroscopy in Heterogeneous Catalysis, ed. J. D. Haw,
Wiley-VCH, Los Angeles, 2002, pp. 139177.
62 J. S. Eric Lichtfouse and Didier Robert, Environmental Chemistry: Green
Chemistry And Pollutants In Ecosystems, Springer, 2005.
63 C. Li and P. C. Stair, Catal. Today, 1997, 33, 353360.
64 Y. T. Chua and P. C. Stair, J. Catal., 2000, 196, 6672.
65 T. V. M. Rao, G. Deo, J. M. Jehng and I. E. Wachs, Langmuir, 2004, 20,
71597165.
66 C. Resini, T. Montanari, G. Busca, J. M. Jehng and I. E. Wachs, Catal. Today,
2005, 99, 105114.
67 H. J. Tian, I. E. Wachs and L. E. Briand, J. Phys. Chem. B, 2005, 109,
2349123499.
68 I. E. Wachs, Catal. Today, 2005, 100, 7994.
69 I. E. Wachs, Y. Chen, J. M. Jehng, L. E. Briand and T. Tanaka, Catal. Today,
2003, 78, 1324.
70 S. W. Yang, E. Iglesia and A. T. Bell, J. Phys. Chem. B, 2006, 110, 27322739.
71 M. D. Argyle, K. D. Chen, E. Iglesia and A. T. Bell, J. Phys. Chem. B, 2005,
109, 24142420.
72 M. D. Argyle, K. D. Chen, C. Resini, C. Krebs, A. T. Bell and E. Iglesia,
Chem. Commun., 2003, 20822083.
73 K. D. Chen, A. T. Bell and E. Iglesia, J. Catal., 2002, 209, 3542.
74 K. D. Chen, A. T. Bell and E. Iglesia, J. Phys. Chem. B, 2000, 104, 12921299.
75 J. A. Jensen, K. B. Rider, Y. Chen, M. Salmeron and G. A. Somorjai, J. Vac.
Sci. Technol. B, 1999, 17, 10801084.
76 R. Reichelt, S. Guenther, M. Roessler, J. Wintterlin, B. Kubias, B. Jakobi and
R. Schloegl, Phys. Chem. Chem. Phys., 2007, 9, 35903599.
77 M. Montano, D. C. Tang and G. A. Somorjai, Catal. Lett., 2006, 107, 131141.
78 K. B. Rider, 2001.
79 G. A. Somorjai, Appl. Surf. Sci., 1997, 121/122, 119.
80 M. A. Newton, A. J. Dent, S. G. Fiddy, B. Jyoti and J. Evans, Catal. Today,
2007, 126, 6472.
81 F. C. Meunier, D. Tibiletti, A. Goguet, S. Shekhtman, C. Hardacre and R. Burch,
Catal. Today, 2007, 126, 143147.
82 I. V. Koptyug, A. A. Lysova, R. Z. Sagdeev and V. N. Parmon, Catal. Today,
2007, 126, 3743.
83 M. B. J. Roeaers, J. Hofkens, G. De Cremer, F. C. De Schryver, P. A. Jacobs,
D. E. De Vos and B. F. Sels, Catal. Today, 2007, 126, 4453.
84 M. J. Castaldi, US Pat., 6,746,657.
85 M. J. Castaldi, M. Lyubovsky, R. LaPierre, W. C. Pfeerle and S. Roychoudhury,
SAE Technical Paper, 2003.
86 M. J. Castaldi, S. Roychoudhury, R. S. Boorse, H. Karim, R. LaPierre and
W. C. Pfeerle, Fuel Processing Session I, Proceedings from the 2003 Spring
National Meeting and Process Industries Exposition, New Orleans, LA, 2003.
87 M. J. Castaldi, Boorse, R. Samuel, S. Roychoudhury, P. Menacherry and
W. C. Pfeerle, NSF National Meeting, January 2002.

Catalysis, 2009, 21, 191218 | 217


This journal is
c The Royal Society of Chemistry 2009
88 W. C. Pfeerle, M. Castaldi, S. Etemad, H. Karim, M. Lyubovsky,
S. Roychoudhury and L. Smith, Abstracts of Papers, 223rd ACS National
Meeting, Orlando FL United States, April 711, 2002, 2002, CATL-018.
89 L. L. Smith, A. Karim, A. Hasan, A. Castaldi, J. Marco, S. Etemad,
W. C. Pfeerle, V. K. Khanna and K. O. Smith, ASME Turbo Expo, Power,
Land Sea, & Air, 2003.
90 M. J. Castaldi, R. LaPierre, M. Lyubovski, W. Pfeerle and S. Roychoudhury,
Catal. Today, 2005, 99, 339346.
91 H. S. Fogler, Elements of Chemical Reaction Engineering, Prentice Hall,
Englewood Cli, NJ, ch. 6.
92 F. H. Ribeiro, M. Chow and R. A. Dalla Betta, J. Catal., 1994, 146, 537544.
93 K. Ramanathan, V. Balakotaiah and D. H. West, Chem. Eng. Sci., 2003, 58,
13811405.
94 Application: EP Pat., 2006-1167571876548, 2008.
95 X. Wang and X. Xu, J. Phys. Chem. C, 2007, 111, 1697416981.
96 X. Lin, C. Zhao, C.-M. Che, Z. Ke and D. L. Phillips, Chem.Asian J., 2007, 2,
11011108.
97 S. Ostrowski, J. C. Dobrowolski and J. Kijenski, Catal. Commun., 2007, 8,
13541360.
98 S. A. French, Platinum Met. Rev., 2007, 51, 5462.
99 G. Pacchioni, Abstracts of Papers, 233rd ACS National Meeting, Chicago, IL,
United States, March 25-29, 2007, COLL-154.
100 J. N. J. van Lingen, O. L. J. Gijzeman, B. M. Weckhuysen and J. H. Van
Lenthe, J. Catal., 2006, 239, 3441.
101 M. Okumura, Y. Irie, Y. Kitagawa, T. Fujitani, Y. Maeda, T. Kasai and
K. Yamaguchi, Catal. Today, 2006, 111, 311315.
102 M. B. Zellner, A. M. Goda, O. Skoplyak, M. A. Barteau and J. G. Chen, Surf.
Sci., 2005, 583, 281296.
103 R. Tokarz-Sobieraj, M. Witko and R. Grybos, Catal. Today, 2005, 99, 241253.
104 M. Witko and R. Tokarz-Sobieraj, Catal. Today, 2004, 9192, 171176.
105 H. Si-Ahmed, M. Calatayud, C. Minot, E. L. Diz, A. E. Lewandowska and
M. A. Banares, Catal. Today, 2007, 126, 96102.
106 C. J. Cramer, Essentials of Computational Chemistry: Theories and Models,
John Wiley & Sons, New York, 2004.

218 | Catalysis, 2009, 21, 191218


This journal is
c The Royal Society of Chemistry 2009