Catalysis

Catalysis Volume 21 A Specialist Periodical Report Catalysis Volume 21 A Review of Recent Literature Editors James J...

Author: James J. Spivey | K.M. Dooley

296 downloads 2847 Views 7MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

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. Fernańdez-Garcıá, 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 [email protected] 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 field of catalysis enjoys significant scientific prominence due to its importance in areas that affect the general public—clean energy, environmental protection, and conversion of sustainable feedstocks, for example. These driving forces, among others, will guide research efforts in our field for the foreseeable future. This volume of the Royal Society of Chemistry’s 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 RassHansen, 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 significant challenges in developing processes for water purification can be addressed using photocatalytic reactions to remove both organic and inorganic pollutants. They point out the difficulties in studying the fundamentals of catalytic reactions in an aqueous medium, and the need to improve the typically low quantum yield in the processes—e.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 specific 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 first 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 final 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: [email protected]. E-mail: [email protected]

Catalysis, 2009, 21, 7–8 | 7 This journal is

c

The Royal Society of Chemistry 2009

how these liquids can be used in a number of different 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, identification of reaction intermediates can help us understand the reaction mechanism. The authors review surface measurement techniques and recent developments in spectroscopy that can help us examine these catalytic properties. We greatly appreciate the efforts 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, 7–8 This journal is

c


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 James J. Spivey* and Kerry M. Dooley*

7

Heterogeneous catalysis for production of value-added chemicals from biomass Kresten Egeblad, Jeppe Rass-Hansen, Charlotte C. Marsden, Esben Taarning and Claus Hviid Christensen* Introduction Setting a new scene Catalytic C–C bond breaking Catalytic hydrolysis Catalytic dehydrations Catalytic oxidations Catalytic hydrogenations Summary and outlook

13

Catalytic and photocatalytic removal of pollutants from aqueous sources J. A. Anderson* and M. Fernańdez-Garcıá General introduction Catalytic elimination of inorganics Photocatalytic removal of inorganics Catalytic and photocatalytic removal of organometallics Catalytic and photocatalytic removal of organics Removal of microorganisms

51

13 14 17 23 25 32 39 43

51 53 60 65 65 73

Catalysis, 2009, 21 | 9 This journal is

c


Nano-architecture and reactivity of titania catalytic materials. Part 2. Bidimensional nanostructured films Gabriele Centi and Siglinda Perathoner Introduction Outlooks for the development of catalysts based on the concept of nanostructured films Synthesis of titania nanostructured films Uses, with focus on catalysis Conclusions

82

82 89 100 106 119

Recent advances in heterogeneous catalysis enabled by first-principles methods Ye Xu Introduction Theory-aided catalyst design Molecular-level effects of reaction environment Outlook

131

Ionic liquids as catalysts, solvents and conversion agents Amit C. Gujar and Mark G. White Introduction Solubility of substrates in ionic liquids Physical and chemical properties of ionic liquids Demonstration of utility of RTILs as reaction solvents Review articles Synthesis of aluminum-containing ILs Alkoxy carbonylation Arene carbonylation Catalytic oxidations Diels–Alder reactions in ILs Dimerization of olefins in ILs Enzyme-catalyzed reactions Fischer esterifications in ILs Friedel–Crafts reactions in ILs Heck reaction Henry reactions in ILs Hydrogenation in ILs Hydroformylation Isomerization Metathesis of olefins Michael reaction Pechmann condensation in ILs Sonogashira reaction Sulfonation of arenes

154

10 | Catalysis, 2009, 21 This journal is

c


131 133 137 146

154 155 156 157 157 158 159 159 162 162 164 165 166 166 169 170 171 173 176 177 177 179 179 181

Supported analogs of ionic liquid catalysts Task specific ionic liquids (TSIL) Telomerization in ILs

181 182 184

Measurement techniques in catalysis for mechanism development: kinetic, transient and in situ methods Nora M. McLaughlin and Marco J. Castaldi Introduction Structure to kinetics Surface measurement techniques Current in situ measurement techniques Future directions

191

191 192 193 197 215

Catalysis, 2009, 21 | 11 This journal is

c


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 efficiently 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 20–30% 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 efficiency 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 efforts 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 different facets. One is the discovery and development of methods to use renewable resources to supply suitable energy carriers, in sufficient 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: [email protected]


c


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 carboncontaining 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 significant 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 C–C, C–H, C–O, and O–H 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 finds 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 valueadded chemicals from biomass using heterogeneous catalytic processes. 2. 2.1

Setting a new scene 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 significantly 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.9–11 It can be argued that if the amount of biomass available is too limited to substitute fossil resources in all its applications and if sufficiently efficient 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 significantly more valuable than transportation fuels. This can be illustrated in 14 | Catalysis, 2009, 21, 13–50 This journal is

c


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, significantly 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 efficiently 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.10–22 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 classified in Fig. 1 as primary renewable building blocks (of which only selected examples are given) with only one purification step. For example, ethanol can be obtained by fermentation of sucrose, glucose by hydrolysis of starch, glycerol by transesterification of triglycerides (or by fermentation of glucose), xylose by hydrolysis of hemi-cellulose, fructose by hydrolysis of sucrose (and by isomerization of glucose), and finally synthesis gas can be obtained directly by gasification 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, 13–50 | 15 This journal is

c


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 different 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 transformations in Fig. 1 are labeled with different arrows to illustrate specific 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 efficiently 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 sufficient 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, 13–50 This journal is

c


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 significant 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 efforts 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 five important reaction types, specifically, C–C 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 first 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. 3.1

Catalytic C–C bond breaking Introduction

This section concerns catalytic processes that transform chemicals from renewables by C–C bond breaking. Among these are thermochemical processes, such as pyrolysis and also gasification, catalytic reactions, such as catalytic cracking and different reforming reactions, and decarbonylation and decarboxylation reactions. Many of these reactions occur simultaneously, particularly in the thermochemical processes. Another technically important class of C–C bond breaking reactions is the fermentation processes, however, they will not be considered in this section since they do not involve heterogeneous catalysis. 3.2

C–C Bond breaking reactions involving bio-resources

3.2.1 Crude biomass. Next to combustion, gasification is probably the easiest and most primitive method for degradation of biomass. In the simplest form, gasification involves heating of biomass (or any other carbonaceous material) to temperatures around 800–900 1C, in an atmosphere with only little oxygen, until it thermally decomposes into smaller fragments. This partial oxidation process obviously requires a significant energy input and is not particularly selective; on the other hand, it is reasonably flexible since essentially all types of biomass can be gasified. Gasification, in particular of coal, has been known for long and was previously used to produce town gas. However, the gas resulting from gasification has a relatively low heating value of only 10–50% of that of Catalysis, 2009, 21, 13–50 | 17 This journal is

c


natural gas,23–25 and this was a major reason for replacing town gas with natural gas. During World War II, biomass gasification 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 gasification 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 different types of gasification furnaces but they generally work by having several different 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 (1–4) and steam reformed (5–6) or dry reformed (7–9) 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).23–26 CH4 + 1/2O2 = CO + 2H2 H2 + 1/2O2 = H2O CnHm + (n/2 + m/4)O2 = nCO + (m/2)H2O C + 1/2O2 = CO CnHm + nH2O = nCO + (n + m/2)H2 C + H2O = CO + H2 CnHm + nCO2 = 2nCO + (m/2)H2 C + CO2 = 2CO

(1) (2) (3) (4) (5) (6) (7) (8)

CH4 + CO2 = 2CO + 2H2

(9)

CO + H2O = CO2 + H2

(10)

The major challenge in gasification is to avoid the formation of tars, which have a tendency to clog filters and condense in end-pipelines. Tars are considered as the condensable fraction of the organic gasification products, and consist mainly of different 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.24–26 Alkali metal salts can be mixed directly with the biomass before entering the gasification furnace. They enhance the gasification reactions and lower the tar content, but recovery of the catalyst is difficult and costly making the alkali metals unattractive as catalysts for industrial use.25,26 Another family of catalysts, which can be used effectively for gasification, 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, 13–50 This journal is

c


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 effective in tar destruction, the reforming of hydrocarbons and in adjusting the composition of the synthesis gas by the water gas shift reaction (10). They are operated as secondary catalysts in a downstream reactor, which can be operated at conditions different from those in the gasifier. 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.24–26 Instead of gasifying biomass, it can be subjected to liquefaction in a pyrolysis process. Pyrolysis is actually one of the main processes occurring during gasification, 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 500–600 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) (250–425 1C, 10–35 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.28–32 Valenzuela et al.,33 however, were the first 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 400–500 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, 13–50 | 19 This journal is

c


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,36–38 and recently a review was published describing how biomass could be converted into fuels or chemicals in a conventional petrochemical refinery in FCC or hydrotreating refinery units.39 3.2.3 Carbohydrate resources. Carbohydrate resources, such as hydrolyzed 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 first attempts to conduct catalytic cracking of biomass. Carbohydrate resources have also been processed under hydrotreating conditions, i.e. high hydrogen pressures (35–300 bar) and high temperatures (300–600 1C) in the presence of Co–Mo or Ni–Mo-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 C–C 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.28–32 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 225–265 1C and 29–56 bar,28 as well as over a specially designed non-precious metal catalysts (Raney Ni–Sn).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, 13–50 This journal is

c


3.3.2 Steam reforming of ethanol. Steam reforming (SR) is probably the most investigated process for breaking C–C bonds in chemicals available from biomass. Particularly, ethanol SR for production of hydrogen has been extensively examined,42–44 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 800–950 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.42–44 Ideally, one mole of ethanol is converted into 6 moles of hydrogen (13). During SR, ethanol decomposes mainly through two different 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, 16–18).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 difficulty in ethanol SR is a too rapid catalyst deactivation due to coking. This can occur by several reactions, such as methane decomposition (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 efficiency fuel cell, compared to combusting the ethanol in a motor engine with a relatively low efficiency, the overall energy output could be significantly improved.50 Alternatively, the steam reforming reaction can be Catalysis, 2009, 21, 13–50 | 21 This journal is

c


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 effective 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 reflux conditions in pure furfural (165 1C), which achieves continuous removal of CO and furan from the reactor. However, a continuous flow system at 159–162 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 (300–400 1C). Typical catalysts are zinc-iron chromite or zinc–manganese 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 flavoring 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 (9–24%) 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 significantly 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. C–C and C–O bond breaking by hydrogenolysis of different polyols (glycerol, xylitol, erythritol and sorbitol) has been investigated by Montassier et al.59,60 Predominantly ruthenium and copper-based charcoal catalysts were studied at 210–260 1C and 1–6 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, 13–50 This journal is

c


ethylene glycol and propylene glycol using ruthenium on a range of different 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 CuO–ZnO 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 CuO–ZnO catalyst was superior in achieving a high C4+ selectivity (73% yield) compared to Ru and Ni catalysts which mainly yielded C1–C3 products. A commercial example of a hydrotreating technology is examined below. The IPCI (International Polyol Chemicals, Inc.) hydrogenolysis process is carried out at 100–300 1C and at hydrogen pressures of 70–300 bar.63 The hydrogenolysis process is used to cleave carbohydrates to smaller polyol fragments. Specifically, sorbitol and mannitol are reformed to propylene glycol and ethylene glycol as the main products, and to different 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 C6H14O6 + H2 = 2C3H8O3

(23) (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. 4.1

Catalytic hydrolysis 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 different path has been sought for. Catalysis, 2009, 21, 13–50 | 23 This journal is

c


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 diffusion and restricted accessibility, as well as the production of many by-products.68 In an effort 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.69–71 A common problem with solid acids, including ionexchange 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 selectivities 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 organic–inorganic 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 significant glucose degradation was observed simultaneously.76 It can be seen then that heterogeneous catalysis may find an opportunity for replacing the enzymatic catalysis of disaccharides to its monosaccharides, and thereby provide industry with a more efficient 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, 13–50 This journal is

c


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 Brøndsted 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 cationexchange 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 ionexchange resin (Amberlyst 15), nafion-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. 5.1

Catalytic dehydrations Introduction

There are several examples of dehydrations of chemicals derived by renewable 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 esterification 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 etherification reaction between two alcohol functionalities. Catalysis, 2009, 21, 13–50 | 25 This journal is

c


5.2 Dehydration reactions involving bio-resources and primary renewable building blocks 5.2.1 Bio-oils. Transesterification 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 different ways, as well as a hydrotalcite catalyst, were all reported to be effective catalysts for the transesterification 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 95–100%. In a similar study, experiments carried out at 180 1C showed a difference 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 (75–80%) 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 transesterification reactions have also been studied using alumina-supported solid base catalysts at methanol reflux 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 transesterification at about 65 1C in a study using cottonseed oil as the fatty acid source,83 and even poultry fat has been transesterified 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 Cu–Zn–Al-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.85–87 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 carboncontaining 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, 13–50 This journal is

c


solid dehydration catalysts have been tested for these reactions, and a comprehensive review of this field is beyond the scope of this review. The reader is referred elsewhere for reviews on these topics.88–91 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 finely 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.93–95 The best results were obtained with silicasupported 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 different 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 silicasupported 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 effective 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 copper–chromite as catalyst.

5.2.5 Xylose. Catalytic dehydration of xylose, which is the most abundantly 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 different 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.


c


Recently, several reports concerning this reaction have appeared in literature describing the use of zeolites,97 ion-exchange resins,98 sulfonic acid functionalized MCM-41,98 immobilized heteropolyacids,99–101 niobium silicates102 and exfoliated titanate and niobate nanosheet structures103 as solid acid catalysts. In 1998, Moreau et al. compared zeolites H–Y 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 H–Y 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 functionalizing 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 different 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 acidified 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 finally exfoliating the sheets with tetrabutylammonium 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 different products, however, some control of the dehydration products obtained can be achieved using different 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, significant achievements were made in the production of HMF by homogeneous catalytic approaches.107 Typically, 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, 13–50 This journal is

c


HMF. These were categorized into five classes of catalysts (organic acids, inorganic acids, salts, Lewis acids and other) by Cottier et al. who also categorized the different methods by which the dehydration reaction was carried out into five different 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 rehydration–decomposition 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 acidcatalyzed dehydration. Solid acid catalysts applied to facilitate the reaction are zeolites, ionexchange 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 different 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 dehydration. Using PK-216, a solution of water-DMSO-polyvinylpyrrolidone containing 10 wt% fructose was dehydrated to HMF after 8–16 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, 13–50 | 29 This journal is

c


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 Brøndsted acidic ILIS were effective for the transformation. 70% yield was obtained at 100% conversion after 4 min. of 200 W microwave irradiation in DMSO using Brøndsted acidic 3-allyl-1-(4-sulfobutyl)imidazolium trifluormethanesulfonate [ASBI][Tf], whereas a yield of 67% was achieved using Lewis acidic 3-allyl-1-(4-sulfurylchloride butyl)imidazolium trifluormethanesulfonate [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 (MxVO1–xPO4 2H2O, M being Cr, Mn, Fe, Al and Ga), it was shown that vanadyl phosphate, which contains both Brøndsted 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 significantly, 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 flow 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-anhydroglucose 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, 13–50 This journal is

c


was present in the reaction mixture after experiments starting from aqueous glucose. In 2000, Kro¨ger et al. reported that 2,5-furan-dicarboxylic acid (FDCA) could be produced in a combined one pot dehydration–oxidation reaction starting from fructose, however, only in 25% yield.121 Very recently, this was improved significantly, when Ribeiro and Schuchardt reported that also SiO2-gel containing Co(acac)3 could be used as catalyst for the combined dehydration–oxidation 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 specific 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 esterification 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.125–127 For dimethyl succinate, 70% isolated yield was achieved using Fe3+-exchanged montmorillonite after 4.5 h at methanol reflux temperature.127 For dibutyl succinate, the best results were obtained with Al-montmorillonite (94% yield after 8 h, also at reflux temperature), which also proved to be a good catalyst for other esterification reactions, e.g. for diesterification 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 esterification 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 dehydration–decomposition of fructose. The synthesis of diethyl levulinate was recently reported using sulfonated Starbon-400-SO3H. The Catalysis, 2009, 21, 13–50 | 31 This journal is

c


selectivity towards the diester was 499% at 85% conversion after 6 h at 80 1C.128 In the paper esterification 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 functionalized 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 dehydration–esterification reaction.130 6. 6.1

Catalytic oxidations 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 efforts 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 films and in polyvinyl acetate for wood glue. Moreover, it also finds use in the food industry as an acidity regulator, as the additive E260. The oxidation of ethanol to acetic acid was among the first 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 effective 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, 13–50 This journal is

c


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 significant attention, and there are several recent reviews with entire sections devoted to this particular topic.136–138 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 transesterification,136 there is a significant 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 therefore 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 2–4, platinum on charcoal showed low catalytic activity for oxidation of the secondary hydroxy group of glycerol with a dihydroxyacetone 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 catalyst.

Oxidation of glycerol in acid media leads to dihydroxyacetone, using Pt–Bi/C as

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 fixed bed process using a Pt–Bi catalyst supported on charcoal. Here, a DHA selectivity of 80% at a conversion of 80% was obtained.140 Catalysis, 2009, 21, 13–50 | 33 This journal is

c


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 catalyst.

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

During the last few years, significant 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 different 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 C–C bond breakage and a resulting loss of selectivity. This was independently supported by the fact the Au–Pd catalysts showed higher selectivity to glycerate than the monometallic Au catalyst, which was shown to be related to the higher efficiency 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 effect 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.148–150 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, 13–50 This journal is

c


acid (Scheme 9). This represents one of the first 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, Pd–Bi/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, specifically 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 modified palladium catalyst. Thus, a Pd–Bi/C catalyst was found to be capable of oxidizing glucose to gluconic acid with excellent selectivity (95–100%) 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.160–164 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 first example of gold catalysts being able conduct aerobic oxidations of Catalysis, 2009, 21, 13–50 | 35 This journal is

c


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 90–100% at 80–94% 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 today’s 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 hydroxypuric is afforded 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 bismuthpromoted platinum catalyst under acidic conditions (pH 3–4) 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 difficult to obtain pyruvic acid directly from 36 | Catalysis, 2009, 21, 13–50 This journal is

c


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 oxidative 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 available from dehydration of pentoses, and it can be produced in a very large scale if necessary. Furfural finds limited use today, though, which is reflected 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 different 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, 13–50 | 37 This journal is

c


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 acetylacetonate encapsulated in sol–gel 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 conditions—25 1C, 1 bar O2 and 8% NaOMe—in 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 pressure—130 1C and 4 bar O2—2,5-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 effect 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 significant 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, 13–50 This journal is

c


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. 7.1

Catalytic hydrogenations 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 significant role in today’s industry, and holds great promise for further developments. Here, selected examples of heterogeneously 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 homogenous 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 different oxidizing agents, thereby catalyzing the hydrolysis part of the reaction. Selectivities to mannitol of 37–40% 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 40–50% 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 affords 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 different study, the 1-4-b glucose dimer cellubiose was used as a model substrate for cellulose. Using a polymer supported ruthenium Catalysis, 2009, 21, 13–50 | 39 This journal is

c


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 modification to the procedure was reported recently, in which cellulose is hydrolyzed in superheated water (200–250 1C) and reduced using a carbonsupported 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 different metal catalysts including nickel, copper, copper-chromite, ruthenium, rhodium palladium and platinum.192–195 For these reactions, the difficulties 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.193–196 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 effective than standard Raney nickel, which can be problematic due to leaching of nickel.197–200 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 effectively 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, 13–50 This journal is

c


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).197–199,203,204 Ruthenium supported on carbon is among the most studied catalysts for this reaction,205 and it is, in fact, also effective 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 fermentation 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 fibers industry in the production of stain resistant carpets etc. Aqueous solutions of 3-hydroxypropanal were reduced using TiO2 supported ruthenium catalysts at 40–60 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, 13–50 | 41 This journal is

c


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 140–220 1C with a hydrogen pressure of 0.1–0.72 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 different study, aqueous phase hydrogenation of lactic acid was achieved using a carbon supported ruthenium catalyst. The hydrogenation is operated at temperatures from 100–170 1C with a hydrogen pressure of 7–14 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 pentoses. 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 hydrogenation with copper containing catalysts, e.g. copper–barium–chromium oxide, copper oxide supported on silica or alumina, copper–chromium oxide and copper–cobalt 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 landfill sites.211 In recent years, several new catalytic systems have been demonstrated 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 carbon—activated carbon, diamond and graphitized fibers—was 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, 13–50 This journal is

c


higher number of surface copper sites and the defective sites at the copper and magnesium oxide interfacial region. Similarly, Cu–Ca/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 hydrogenation, 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 100–150 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 20.3–30.4 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 hydrogen 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 functionalized molecules are very different 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 hydrocarbon feedstock. Due to the quite low reactivity of the hydrocarbons; it has Catalysis, 2009, 21, 13–50 | 43 This journal is

c


been possible to develop efficient catalytic processes that operate satisfactorily 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 different. Still the desirable transformations are entirely dependent on catalysis, but now it is often a question of controlled removal of superfluous chemical functionality under sufficiently mild conditions 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 significant 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 significant 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 heterogeneously 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 significant challenge to discover and develop heterogeneous catalysts that exhibit sufficient 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 offers some very promising opportunities but that there clearly exists a great need for further discoveries and developments in this emerging field.

References 1 2 3 4 5 6 7 8 9 10 11

12

C. H. Christensen and J. K. Nørskov, J. Chem. Phys., 2008, 128, 182503. S. Shaik, Angew. Chem., Int. Ed., 2003, 42, 3208. S. Senkan, Angew. Chem., Int. Ed., 2001, 40, 312. I. Maxwell, Stud. Surf. Sci. Catal., 1996, 101, 1. Handbook of Heterogeneous Catalysis, eds. G. Ertl, H. Kno¨zinger, F. Schu¨th and J. Weitkamp, Wiley-VCH, Weinheim, 2008. B. A. Tokay, Biomass Chemicals, in Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH, Weinheim, 2005. C. H. Christensen, J. Rass-Hansen, C. C. Marsden, E. Taarning and K. Egeblad, ChemSusChem., 2008, 1, 283. J. Goldemberg, Science, 2007, 315, 808. G. W. Huber, S. Iborra and A. Corma, Chem. Rev., 2006, 106, 4044. A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107, 2411. 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. D. R. Dodds and R. A. Gross, Science, 2007, 318, 1250.


c


13 P. Ma¨ki-Arbela, B. Holmbom, T. Salmi and D. Y. Murzin, Catal. Rev. Sci. Eng., 2007, 49, 197. 14 Top Value Added Chemicals From Biomass—Results 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 Biorefineries—Industrial 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. Jørgensen 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. O’Connor, 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, 13–50 | 45 This journal is

c


48 S. Czernik, R. Evans and R. French, Catal. Today, 2007, 129, 265. 49 J. R. Rostrup-Nielsen, J. Sehested and J. K. Nørskov, Adv. Catal., 2002, 47, 65. 50 J. Rass-Hansen, C. H. Christensen, J. Sehested, S. Helveg, J. R. RostrupNielsen 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 Ullmann’s 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. Meńe´zo, 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, 9–11 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. Theóleyre, Ind. Crops Prod., 2000, 11, 237. 75 A. Abbadi, K. F. Gotlieb and H. van Bekkum, Starch/Sta¨rke, 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, 13–50 This journal is

c


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. Romań-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. A—General, 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. A—General, 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, 13–50 | 47 This journal is

c


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. Kro¨ger, U. Pru¨sse 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. Jørgensen, 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. Jørgensen, 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. Griffin 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, 13–50 This journal is

c


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. 169–180. 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. Pru¨sse, 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, 13–50 | 49 This journal is

c


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, 8714–8715. 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. Hoffer, E. Crezee, P. R. M. Mooijman, A. D. van Langeveld, F. Kapteijn and J. A. Moulijn, Catal. Today, 2003, 79–80, 35–41. 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, 117–124. 207 R. D. Cortright, M. Sanchez-Castillo and J. A. Dumesic, Appl. Catal. B, 2002, 39, 353–359. 208 Z. Zhang, J. E. Jackson and D. J. Miller, Appl. Catal. A, 2001, 219, 89–98. 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, 13–50 This journal is

c


Catalytic and photocatalytic removal of pollutants from aqueous sources J. A. Anderson*a and M. Fernańdez-Garcıáb 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 afford safe clean drinking water. Humanity’s needs are met by only 1% of the Earth’s 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 filtration, 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 effects 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 purification 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 affects the chemical composition. Furthermore undesirable ions, such as nitrates, remain unaffected 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 specific to water treatments. The first is the challenge presented by operating using water as a reaction medium if a full a b

Surface Chemistry and Catalysis Group, King’s College, University of Aberdeen, Aberdeen, Scotland, UK AB24 3FX. E-mail: [email protected] Instituto de Catalisis y Petoleoquimica (CSIC), Cantobalanco, Madrid 28049, Spain


c


kinetic mechanism including identification of reaction intermediates and active sites is to be established. Application of most spectroscopic techniques 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 difficult 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 treatment 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 final 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 applications which have been developed since their first descriptions in the scientific literature. Despite the difference 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 understood that we are referring to any chemical process catalyzed by a solid where the external energy source is an electromagnetic field with wavenumbers in the UV-visible range.2–7 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 filled valence band and an empty conduction band. Among these possible semiconductors, TiO2 is the most used photocatalytic material as it fulfils all of these requirements and exhibits adequate conversion values.8 However, in spite of the high conversion 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 differing by the distortion of each octahedron and the 52 | Catalysis, 2009, 21, 51–81 This journal is

c


assembly pattern of such octahedra chains. The Ti–Ti distances in the anatase structure are greater than in rutile, whereas Ti–O distances are shorter.12 These structural differences cause different mass densities and lead to different 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 different 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 Metal–Oxygen bond have been discussed in a more general context for several photocatalytic oxides by Wiswanathan.14 In any case, these differential structural features which distinguish anatase and rutile are presumably responsible for the difference 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 influence the electron-hole charge separation process.2,6,7,15 Another methodology, central in current research, considers the extension of the solid lightabsorption 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 (sulfides, nitrides, and other non-Ti-based oxides), have been tested in pursuing this goal.2–9,16 The majority of studies involving TiO2 aimed at improving its optical absorption and photocatalytic performance, are focused on anatase modification by cation and/or anion doping.2,5–7,17 Doping may also result in surface modification 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 anataseTiO2 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 significant quantities of dissolved inorganic salts need to be considered.18 Dissolved inorganic species, both cations and anions, can strongly influence the removal of organic species and thus must be carefully considered. Catalysis, 2009, 21, 51–81 | 53 This journal is

c


Inorganic residues present in water can be simply classified 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 specific 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 different approaches used to treat the individual contaminants.

2.1

Nitrate elimination

The potentially harmful effects 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). Additionally, 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 effects 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 agricultural 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 methods. Physical-chemical processes are based on ion-exchange, electrodialysis or reverse osmosis procedures that have the benefit of being able to remove the pollutant without the introduction of other substances. However these processes are not effective 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, 51–81 This journal is

c


Biological treatments offer good selectivity when combined with effective sludge processing. Although bacteria provide an economically viable solution for water decontamination,19–21 the process is slow and its efficiency 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 summarised in a book chapter.25 The effective 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 comparing 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. Welldispersed 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 configuration, 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 modified Pd/alumina catalysts have been ascribed to the presence of highly dispersed alloyed Pd–Sn 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 Pd–In, activity was increased as a function of In addition although a 6:1 ratio was Catalysis, 2009, 21, 51–81 | 55 This journal is

c


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.30–32 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 influenced by the nature of the reducing and oxidising pretreatments which will subsequently influence 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 Pd–Cu have involved TiO2,34–36 niobia37 and activated carbons38,39 Hydrotalcite, modified 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 hydrotalcite’s 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 monometallic Pd catalyst for effective 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.


c


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 effective solution buffering, 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 suffer from poisoning when CO2, was used as a pH buffer. 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 determined for Pd–Cu/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 first order as nitrate levels were depleted. The same authors found that the kinetics could be described in terms of a Langmuir–Hinshelwood 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 confirmed the selection of Pd–Cu catalysts for the process. The same authors45 proposed that the kinetics of nitrate removal were consistent with a mechanism involving ionic intermediates, which required participation of different 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 different sites by the reactants, as predicted by the Langmuir adsorption model. Infrared spectroscopic studies confirm 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 final reduction to N2, N2O or NH4+ takes place at a Pd monometallic site where NO may be the key intermediate. A recent spectroscopic Catalysis, 2009, 21, 51–81 | 57 This journal is

c


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 Pd–Cu.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 N–H bond.29 While an increase of hydrogen availability may have benefits 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 flow rate leads to a reduction in ammonium formation24 albeit with a loss in activity. It has also been proposed that specific 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 significantly 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 controlling undesired ammonium formation, is through the use of formic acid as a reductant which can be added in stoichiometric amounts to avoid overhydrogenation.25,51 However, this serves an additional purpose since the 58 | Catalysis, 2009, 21, 51–81 This journal is

c


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)

2NO3 (aq) + 5H2(g) = N2(g) + 2OH (aq) + 4H2O(l)

(1) (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 buffer25 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 defining the ‘‘right place,’’ it has been shown by the use of supports of a range of pore sizes, that diffusion of hydroxide ions from pores may be influential 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 buffers 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 consequently 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 diffusion 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 influences 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 denitrification 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 sulfide is also Catalysis, 2009, 21, 51–81 | 59 This journal is

c


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 modifications 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 fixed-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 sufficiently 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 selectivity appeared to be the main difficulty.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 first step is conducted at pH 4 11 over Pd–Cu 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. 3.1

Photocatalytic removal of inorganics 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 significantly 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 difficult 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 inefficient 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, 51–81 This journal is

c


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.61–63 Other anions display unique features with respect to the chemical behaviour previously mentioned. Among those, two in particular, merit discussion. Halogens and nitrates differ 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 effective under near-UV excitation. Bromide and iodide radicals are both essentially ineffective 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 specific 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.67–69 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.70–72 In addition to the selection of the metal loaded on the semiconductor oxide, photocatalytic reduction reactions may be expected2 to be influenced by a number of factors, including irradiation time, solution pH and the chemical nature of the sacrificial agent and such sensitivities to these factors for nitrate and nitrite reduction are reported.73–76 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 concentrations 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, 51–81 | 61 This journal is

c


further investigation. In the main, the majority of studies involving nitrate reduction have involved less complex sacrificial hole scavenger molecules added with a view to enhancing photocatalytic efficiency. These molecules include methanol,79–81 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 efficiency 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 findings regarding the relative efficiencies 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 effective 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 Pd–Cu based hydrogenation catalysts. Similar performances have subsequently been obtained in our laboratory76 where it was confirmed that optimum performance is achieved with ca. 1 wt% Ag, (Fig. 3) consistent with findings 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


c


appeared to demonstrate greater performance for nitrate elimination.76 However, as is the case with hydrogenation catalysts, samples showed poorer conversion efficiency 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 infinite 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 countries and typically, maximum concentrations of 0.001–0.0001 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 filters based on using different 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 effects. 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 classified as non-hazardous. Using EDDS may reduce environmental impact of chelates in formulations while maintaining the chelate performance. Catalysis, 2009, 21, 51–81 | 63 This journal is

c


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. sacrificial 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 sacrificial hole trapping agents such as acids and alcohols.93,94 Depending on pH and other experimental conditions (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 sacrificial 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 modification of 64 | Catalysis, 2009, 21, 51–81 This journal is

c


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 application 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 influence of UV irradiation, however, as noted in the review of semiconductor photocatalysis 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 190–250 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 investigated. 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 efficiency with both naturally occurring organic matter (NOM) such as humic acids77 and those resulting from Catalysis, 2009, 21, 51–81 | 65 This journal is

c


anthropogenic activity. Often, combinations of these procedures enhance operational efficiency. For example, membrane filtration is a reliable and accepted means of treating drinking water as these are effective 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 carbon–carbon 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 purification. 5.1

Catalytic removal of organics

Catalytic wet air oxidation (CWAO) is an effective 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 (2–20 MPa) and temperatures (200–320 1C) and there are drivers for replacing these methodologies with improved catalytic wet oxidation processes. Potentially useful results have been published which involve CWO of dyeing and printing wastewater.105–109 Catalysts based upon the use of copper salts appear to be the most active homogeneous catalysts, which entails a subsequent separation step to remove the toxic copper ion from the final effluent.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,112,114–116 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 ineffective 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 effectively 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/Al2O3–CeO2 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 effective than metal oxide catalysts.104,112,117 All of 66 | Catalysis, 2009, 21, 51–81 This journal is

c


these catalytic oxidation processes require the use of elevated temperatures, however there are recent studies in the treatment of petroleum refinery 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 specific 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 effluents. The feasibility of hydrodechlorination for treating groundwater in situ has been demonstrated 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 form128–131 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 suffer 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 hydrogenation of sulfites/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 308–448 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 Henry’s 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 affected by operation under these conditions as an approximately 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 conventional 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, 51–81 | 67 This journal is

c


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-defined regimes associated with two different reaction pathways. At low concentrations (below 103–104 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 significant 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 deactivation 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 affect photoactivity in a milder manner. - Photodegradation of para derivates depends mainly on the effect of the substituent on the aromatic ring, the reaction rate being increased by electrondonating 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 significant 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 effective 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, 51–81 This journal is

c


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 floating 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 photocatalysts is not free from difficulties 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 C–X (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-effective energy source, may provide a lucrative alternative to current technology.7,17

5.2.1 Application of photocatalysis to specific reagent groups. Pesticides contain a wide variety of compounds arising from industrial effluents, agricultural runoffs, and chemical spills.166 Such chemicals are toxic, having potential effects such as carcinogenesis, neurotoxicity, and effects 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 classified 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 different technologies are current used and investigated in order to degrade these compounds. The main degradation processes and technologies are detailed below. Catalysis, 2009, 21, 51–81 | 69 This journal is

c


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 significant number of EDCs are not efficiently removed by these treatments and novel technologies are sought. Photocatalysis offers interesting areas of research and application in the field 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 mineralization 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 interconnected pathways. Some of the longer-lived intermediates which are detected can be classified into five 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 specific reactor configurations adequate for specific 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, 51–81 This journal is

c


In the case of degradation of insecticides, by considering their chemical nature it is possible to distinguish between organo-phosphates (OPs), organochlorine (OCs) and carbamate compounds. Photocatalytic degradation of DDVP (dimethyl-2,2-dichlorovynil phosphate), DEP (dimethyl-2,2,2trichloro-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 efficiencies. 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 influence 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 suffer direct photolysis, possibly due to the low bond energy of P–S and C–S bonds.173–177 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.175–177 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 important 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. 5–6 months in fish. Studies of its photodegradation show complete mineralization through a complex, multi-step mechanism.180 Other carbamates whose degradation is efficiently 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-2methylphenoxyacetic acid). The reaction rate is, in all cases, related to the Hammett constant and adsorbability (measured by the partition coefficient 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, 51–81 | 71 This journal is

c


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 significantly enhanced in the case of lanthanide-doped titania.170,182,184–186 Finally, some fungicides such as metaxyl (N-(2,6-dimethylphrnyl)-N(methoxyl-acetyl)-D,L-alanine methyl ester), pyrimethanil (N-(4,6-dimethylpirimidin-2-yl)aniline), chlorothalonil (tetrachloroisophtalonitrile), and dichlofuanid (N-dichlorofluoromethylthio-N0 ,N0 -dimethyl-N-phenylsulfamide) were subjected to photocatalytic degradation, leading to complete mineralization following Langmuir-Hinshelwood kinetic models.187–189 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 influences the degradation rate by affecting the electric charge suffered 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 artificial 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 significant number of pollutant families previously described, the combined use of ozonization, sonolysis and other methods together with photocatalysis typically leads to synergistic effects.193 5.2.4 Photocatalytic removal of other molecules. While there an enormous 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, 51–81 This journal is

c


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 competitive alternative in terms of time and economical requirements.194 Also, it is interesting to mention the elimination of specific 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 efficient pathway for its elimination and it seems that the acid–base properties of titania play a significant 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 chemical or photochemical damage or physical removal of microorganisms by filtration. Chlorination is a universally practiced water-disinfection process, which can prevent waterborne infectious diseases, but it is not efficient in the inactivation of spores, cysts and some viruses.197 Chlorination, combined with ozonization, significantly improves the disinfection efficiency 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 significant number of microorganisms including Gram-negative bacteria (Escherichia coli, Enterobacter cloacae, Erwinia Caratovira, Salmonella typhimurimm, enterica, and faecalis Pseudonomas earuginosa, and fluorescens, 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).6,7,199–206 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, fluences (power time) and irradiation conditions (pulsed, continuous) used in experiments with titania as they all significantly influence 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 effective killing, B. pumilis inactivation is better obtained by using intermittent illumination.207–209 Catalysis, 2009, 21, 51–81 | 73 This journal is

c


- The second factor concerns the state and nature of the biocidal oxide; TiO2 works by surface/near-surface contact and display significant variability in efficiency 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 microorganisms by phagocytosis, causing rapid cellular damage in addition to that caused by photo-activity.209 - The third is the effect of temperature as typically the inactivation rate increases/decreases with temperature for gram-positive/gram-negative bacteria. 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 specific 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 efficiencies, close to those considered useful for bacteria disinfection (4–5 log reduction in the temporal range of minutes) are customarily reported in studies reviewed here. It must be noted, however, that wastewater with significant 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 biofilms 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 effectively.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 films: (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.


c


proposals suggest that photokilling mechanisms imply an oxidation of the intracellular coenzyme A (CoA), inhibiting cell respiration. Subsequent direct TiO2—microorganism contact would result in cell death.207 However, the photokilling sensitivity to the microorganism structural surface properties, 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 suffer a continuous collapse by which the photoprocess 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 efficiency of the microorganism 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 Aguëra, A. Sutherland, P. Sampedro-Tejedor, A. Kubacka, and C. Coloń and to acknowledge financial 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. Hoffmann, 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. Coloń, C. Belver and M. Fernańdez-Garcı´ a, ‘‘Nanostructured oxides in Photocatalysis’’, in ‘‘Synthesis Properties and Applications of Solid Oxides’’, eds. J. A. Rodrı´ guez and M. Fernańdez-Garcı´ 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, ‘‘Tecnologıá de Fotocata´lisis Solar’’, Cuadernos Monogra´ficos, 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, 51–81 | 75 This journal is

c


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. Fernańdez-Garcı´ a and G. Coloń, 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. Ho¨rold, K.-D. Vorlop, T. Tacke and M. Sell, Catal. Today, 1993, 17, 21. 24 S. Ho¨rold, T. Tacke and K.-D. Vorlop, Environmental Technol., 1993, 14, 931. 25 K.-D. Vorlop and U. Pru¨sse, in Environmental Catalysis, eds. F. J. J. G. Janssen and R. A. van Santen, ICP, London, 1999, vol. 1, p. 195. 26 U. Pru¨be, S. Ho¨rold 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. Pine´da 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, 93–95, 333. 36 J. Sa´, T. Berger, K. Fo¨ttinger, 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. Lamy–Pirata and J. Barbier, Catal. Today, 2002, 5, 35. 39 N. Barrabe´s, J. Just, A. Dafinov, 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. D’Arino, 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. Wa¨rna˘, I. Turunen, T. Salmi and T. Maunula, Chem. Eng. Sci., 1994, 49, 5763. 48 S. D. Ebbesen, B. L. Mojet and L. Lefferts, 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. Pru¨be, M. Kro¨ger 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, 51–81 This journal is

c


53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91

F. Ruiz-Bevia´, J. Catal., 2004, 227, 563. A. Pintar, M. Sˇetinc and J. Levec, J. Catal., 1998, 174, 72. A. Pintar and J. Batista, Catal. Today, 1999, 53, 35. A. Pintar, J. Batista and I. Musˇ evicˇ, Appl. Catal. B, 2004, 52, 49. A. Pintar J. Batista and J. Levec, Chem. Eng. Sci., 2001, 56, 1551. A. Corma, A. E. Palomares, F. Rey and J. G. Prato, J. Catal., 2004, 227, 561. M. Sell, M. Bischoff and D. Bonse, Vom Wasser, 1992, 79, 129. A. Pintar and J. Batista, Appl. Catal. B, 2006, 63, 150. M. Abdullah, G. K. C. Low and R. W. Mattheus, J. Phys. Chem., 1990, 94, 6820. O. J. Jung, Korean Chem. Soc., 2001, 22, 1183. G. Coloń, M. C. Hidalgo and J. A. Navı´ o, Appl. Catal. B, 2003, 45, 39. M. M. Lewandowski and D. F. Ollis, J. Catal., 2003, 217, 38. P. B. Amama, K. Itoh and M. Murabayashi, Appl. Catal. B, 2002, 37, 321. H. Thiele, Ber., 1907, 40, 4914. W. T. Anderson, J. Am. Chem. Soc., 1924, 46, 797. D. S. Villars, J. Am. Chem. Soc., 1927, 49, 326. P. Warneck and C. Wurzinger, J. Phys. Chem., 1968, 72, 389. A. Kudo, K. Domen, K.-I. Maruya and T. Onishi, Chem. Lett., 1987, 1019. A. Kudo, K. Domen, K.-I. Maruya and T. Onishi, J. Catal., 1990, 135, 300. K. T. Ranjit, T. K. Varadarajan and B. Viswanathan, J. Photochem. Photobiol. A, 1995, 89, 67. K. T. Ranjit, R. Krishbamoorthy and B. Viswanathan, J. Photochem. Photobiol. A, 1994, 81, 55. K. T. Ranjit, R. Krishbamoorthy, T. K. Varadarajan and B. Viswanathan, J. Photochem. Photobiol. A, 1995, 86, 185. F. Zhang, R. Jin, J. Chen, C. Shao, W. Gao, L. Li and N. Guan, J. Catal., 2005, 223, 424. J. Sa´, C. Alcaraz Aguëra, S. Gross and J. A. Anderson, Appl. Catal. B., Environmental, DOI: 10.1016/j.apcatb.2008.07.014. A. Bansal, S. Madhavi, T. T. Yang Tan and T. M. Lim, Catal. Today, 2008, 131, 250. B. Bems, F. C. Jentoft and R. Schlo¨gl, Appl. Catal. B, 1999, 20, 155. K. T. Ranjit, T. K. Varadarajan and B. Viswanathan, J. Photochem. Photobiol. A, 1997, 108, 73. T. Mori, J. Suzuki, K. Fujimoto, M. Watanabe and Y. Hasegawa, Appl. Catal. B., 1999, 23, 283. S. Tawkaew, S. Yin and T. Sato, Int. J. Inorg. Mater., 2001, 3, 855. Y. Li, Q. Peng and F. Wasgestian, Chin. J. Catal., 1999, 20, 378. L. L. Perissinotti, M. A. Brusa and M. A. Grela, Langmuir, 2001, 17, 8422. H. Kominami, F. Akitoshi, M. Shinya, I. Hiroyuki, K. Yoshiya and B. Ohtani, Catal. Lett., 2001, 76, 31. Y. Li and F. Wasgestian, J. Photochem. Photobiol. A, 1998, 112, 255. R. Jin, W. Gao, J. Chen, H. Zeng, F. Zhang, Z. Liu and N. Guan, J. Photochem. Photobiol. A, 2004, 162, 585. M. Penpolcharoen, R. Amal and M. Brungs, J. Nanopart. Res., 2001, 3, 289. W. Gao, R. Jin, J. Chen, H. Zeng, F. Zhang, Z. Liu and N. Guan, Catal. Today, 2004, 90, 331. D. M. Stanbury, Adv. Inorg. Chem., 1979, 83, 652. W. Salomons, U. Forstner and P. Mader, in ‘‘Heavy Metals, Problems and Solutions’’, Springer, Berlin, 1988. EC Directive Related with Drinking Water Quality Intended for Human Consumption, Brussels, Belgium, 98/83, 1998. Catalysis, 2009, 21, 51–81 | 77 This journal is

c


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. Martı´ 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. Go´mez-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, 102–103, 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. Dafinov, 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. Stu¨ber, 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. Fo¨ger, 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. Cuauhte´moc, G. Del Angel, G. Torres and V. Bertin, Catal. Today, 2008, 133–135, 588. 78 | Catalysis, 2009, 21, 51–81 This journal is

c


123 Y. Sun, Y. Zhang and X. Quan, Separation and Purification 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. Schu¨th 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. Rodrı´ guez, Appl. Catal. B, 2008, 78, 259. 129 L. Calvo, M. A. Gilarranz, J. A. Casas, A. F. Mohedano and J. J. Rodrı´ guez, Appl. Catal. B, 2006, 67, 68. 130 C. Schu¨th, S. Disser, F. Schu¨th 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 Meńorval, 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. Stafford, 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. Aranã, E. Tello-Redoń, J. Donã-Rodrı´ guez, J. A. Herrera-Meliań and M. Pe´rez-Penã, Appl. Catal. B, 2001, 30, 1. 139 G. Coloń, M. C. Hidalgo, G. Munuera, I. Ferino, M. C. Cutrufello and J. A. Navı´ 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. Garcı´ a-Molina, M. A. Banõs, J. Gimeńez 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. Aylloń, 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. Adań, A. Bahamonde, M. Fernańdez-Garcı´ a and A. Martı´ 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, 51–81 | 79 This journal is

c


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. Official 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. Fernańdez, J. Banco and S. Malato, J. Photochem. Photobiol. A, 2007, 183, 354. 178 G. Mills and M. R. Hoffmann, 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. Fernańdez-Alba, S. Malato and M. I. Maldonado, Environ. Sci. Technol., 2000, 34, 1563. 80 | Catalysis, 2009, 21, 51–81 This journal is

c


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. Adań, A. Martı´ nez-Arias, M. Fernańdez-Garcia and A. Bahamonde, Appl. Catal. B, 2007, 77, 395. 196 C. Adań, A. Bahamonde, A. Martı´ nez-Arias, M. Fernańdez-Garcia, L. A. Pe´rez-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. Fernańdez-Garcı´ a and M. Fernańdez-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. Fernańdez-Garcı´ a, F. Fernańdez-Martı´ n, R. J. Jimeńez-Rioboo, A. de Andre´s, A. Kubacka, M. Ferrer and M. Fernańdez-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. Rincoń and C. Pulgarin, Appl. Catal. B, 2003, 44, 268. 210 K. Sunada, T. Watanabe and K. Hasimoto, J. Photochem. Photobiol. A, 2003, 156, 227.


c


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.1–3 They find 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.4–5 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,1–9 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 influences 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 influenced 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.11–13 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-defined 3D structure (both on short and longrange), 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: [email protected]; Fax: +39-090-391518


c


microstrains, oxygen vacancies or metal ions in unusual coordination states. A 3D environment for adsorption/transformation may significantly 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 relationship 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 Schlo¨gl13 suggested the use of epitaxial films with defined surface structure. Peden et al.18 used molecular beam epitaxy (MBE) systems to grow model metal-oxide films. Rainer and Goodman19 also stressed the relevance of thin film oxides for structure-activity investigations of heterogeneous catalysts. A general issue in these studies is that the preparation method is quite different 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 films 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 films with an ordered nanostructure offers several new opportunities for preparing novel catalysts. These possibilities include (i) nanostructuring the surface in the form of catalytic nano-reactors, (ii) nanoconfinement 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-confined nano-drops. These will be discussed in the following section. Nano-architecture in oxide-type materials In general terms, building a defined nano-architecture in oxide-type materials further extends the concept of nanocatalysis,23 e.g. when the electrons are confined, 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, 82–130 | 83 This journal is

c


nanometer length scale showing that the presence of quantum size effects and structural fluxionality (capability of small clusters to exhibit several structural forms of comparable energy) markedly influence the reactivity. Further confinement of these metal nanoparticles within host ordered oxides could lead, from one side, to further promote these effects 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 offer 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 significantly altered, changing the surface-to-volume ratio, and particle-to-support interaction.28 The misfit between the oxide particle and the host ordered porous material could stabilize defects, create unusual coordination states, and in general terms significantly alter the metal-oxide properties and reactivity.26,27 Nanostructures within porous materials show different 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 first 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 effect in terms of creation of reactive sites, but a negative effect 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 differ 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 different reaction of selective oxidation (toluene to benzaldehyde).33 84 | Catalysis, 2009, 21, 82–130 This journal is

c


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 development 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 significantly different 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 films is defined. These ordered array materials find 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 specific 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 different requirements. Ordered mesoporous oxides, obtained for example by block copolymertemplated,36 micelle-templated37 or other methods, such as inverted opals, colloidal templating or double templating procedures,38 or cooperative selfassembly methodologies,39 offer 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 differentiate 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 influences significantly their catalytic behaviour. An elegant demonstration was given by Ueda et al.40 recently. Schlo¨gl et al.41 showed how 1D vanadium oxide nanostructures, induced by nucleation over carbon nanotubes, have different 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 different 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 beneficial presence of a higher amount of oxidized gold species and surface oxygen vacancies resulting from the strong interaction Catalysis, 2009, 21, 82–130 | 85 This journal is

c


between Au and the well-defined 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 specific 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. Significant effort has been made in recent years on the development of novel methodologies to synthesize oxide nano-wires, -tubes and -rods.30,48–52 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.55–59 In most of the cases, research attention was given to synthesis, while less studies were specifically 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 specifically investigated. In the first 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 films, e.g. bidimensional ordered arrays of 1D nanostructures of TiO2. In fact, even if various reviews55–59 have reported advances in the field, 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 offer real advantages or when the effect 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-effectiveness 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 offer 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 films, ordered arrays of nanotubes or nano-rods/-wires, nanobowl array, but with some comments also on 86 | Catalysis, 2009, 21, 82–130 This journal is

c


analogous 2D materials such as nano-membranes (called also nanohole array) and bidimensional nanosponges. We could call these materials as nanostructured titania films. We define here as ‘‘film’’ a 2D oxide layer, where the aspect ratio (considering the film as a parallelepiped, the aspect ratio is the ratio between base to height, the latter being the thickness of the film) is higher than 300–500, and as ‘‘nanostructured’’ a film 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 flat oxide film) should be higher than 10 000. Based on this definition, we caution the reader that often materials indicated in literature as ‘‘nanostructured oxides’’ do not belong to the category we define. We also remark that following this definition monolayer oxides, e.g. for example V2O2/TiO2, do not fall within this definition of ‘‘nanostructured’’ oxides. Why focus on titania? We focus attention here on titania (TiO2) for the following reasons. The first 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 multifunctional properties, such as Lewis and Bro¨nsted sites, redox centres, etc. The third is that it has several applications both as a catalyst and an advanced material for coating, sensors, functional films, etc. The fourth is its high photocatalytic activity61–64 which make titania unique materials. TiO2-based materials and thin films find 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 applications 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 titaniabased 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 offers general advantages for preparing catalysts, even if there are still many problems to solve, as discussed later. Titania nanostructured films There are many methods to prepare titania thin films,67–73 the principal of which are metallorganic chemical vapour deposition (MOCVD), sol-gel, Catalysis, 2009, 21, 82–130 | 87 This journal is

c


vacuum thermal evaporation, plasma or magnetron sputtering. Some of these produce ordered or pseudo-ordered nanostructured films, but in general a randomly arranged assembly of nanocrystals is present. When used for photocatalytic applications, or to produce novel photovoltaic materials (by photosensitivization of the titania by dyes, the well known Gratzel cells,72–74 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 differences between TiO2 thins films and aligned nanorod array of TiO2.


c


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 diffusion coefficient 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 efficient fuel production (e.g., solar hydrogen production) or electricity (photochemical solar cells).76–80 We will not discuss here aspects related to the use of TiO2 nanostructured films for dye-sensitivized photovoltaic cells, even though this is a major field 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 field. The number of publications in this field 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 photocatalysts. Nevertheless, this demonstrates that a large body of knowledge is available for its use in preparing advanced catalysts. Some of the possibilities in this field are outlined in the following section. 2. Outlooks for the development of catalysts based on the concept of nanostructured films Although the discussion reported here will be mainly in reference to titania nanostructured films, the opportunities and problems analyzed are of more general interest and regard how to use the concept of bidimensional nanostructured oxide films to prepare novel catalysts with tailored nano-architectures. Catalysis, 2009, 21, 82–130 | 89 This journal is

c


Possibilities offered from structuring the surface as array of nanoreactors Fig. 3 shows a conceptual example of the possibilities offered 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 offered from this nanoreactor structure are also outlined in Fig. 3 and may be summarized as follows: Nano-Confinement. There are limited, but interesting studies, regarding the confinement 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 confined within the mesoporous cavities.83 To evidence also that there are many studies of confinement in mesoporous materials in the polymer diffusion and membrane literature, but they refer essentially to entropic effects 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 confined 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 first 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 confined in MCM-41 and SBA-15. The effect was related to the influence of the inner surfaces of mesoporous silica. Other authors have further demonstrated the influence of confinement on the adsorption and properties of fluids within an ordered mesoporous material.85–87 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 offered by this concept (in particular in terms of realizing multifunctional catalysts for cascade reactions in nanoconfined liquids) and a SEM image of an array of TiO2 nanotubes produced by anodic oxidation of Ti foils. Source: Centi et al.81


c


first showing that an enthalpic excess is present during the adsorption processes, due to a wall effect (confinement) in mesoporous materials. Confinement effect was shown to influence 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 identified (the position is depending on the pressure), once liquid scCO2 is confined inside a periodically organized mesoporous material.90 The effect is thus not related to geometrical constraints induced on complexes anchored in mesoporous channels (sometimes also called as confinement effect, even if this definition is not properly correct),91 neither to shape-selectivity effects as possible in zeolites, since the size of mesoporous channels is much larger than those of micro-porous materials. Instead, an effective modification on the characteristics of the fluids is observed due to the electrostatic field generated by the channel walls. This is an enthalpic effect versus an entropic effect as observed when the modification is instead related to limitations in the translation modes of molecules. Recently, it was also demonstrated that wall curvature influence the molecular orientation of the transition states.92 The confinement in mesoporous materials thus influences the properties of fluids 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 effect is particularly enhanced in the case of the mesoporous material. The effect is equivalent to a higher virtual pressure within the mesoporous channels which influences from one side the partial pressure of H2 and O2 (virtual solubility) and from the other hand the fluid viscosity. In fact, by increasing the total pressure Pd-SiO2 also shows similar productivities and selectivities. In the absence of confinement effect, e.g. different fluid characteristics or higher virtual pressure induced by the effect of the wall in a mesoporous material, above effects cannot be explained. We have also shown that the characteristics of the wall (hydrophobicity, for example), as well of the fluid, influence largely the confinement effect. Therefore, reaction in confined 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 mesoporous 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 difficult to realize a mesoporous silica film 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 diffusion limitations. We may also comment that the local electrostatic field within a TiO2 nanotube is significantly different from that of an equivalent SiO2 Catalysis, 2009, 21, 82–130 | 91 This journal is

c


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 different from those of silica particles due to their different modality of preparation), as well as the hydrophilic characteristics, number and strength of Bro¨nsted sites, presence of Lewis acid or basic sites, etc. In addition, electron delocalization, and mobility of adspecies is quite different from TiO2 (or other redox and/or semiconductor oxides) with respect to silica. Therefore the characteristics of a fluid confined in a TiO2 nanotube and the influence of this effect on the catalytic reactivity are different 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 different from that possible for mesoporous silica. There are various studies which show that the confinement 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 confinement of immobilized, single-site chiral catalysts to enhance enantioselectivity. Santiso et al.96 reviewed some possible confinement effects (influence of steric hindrance on the equilibrium and kinetics, influence 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, significantly extends the range of possible design of novel catalysts having a defined and tailored 3D architecture of the active centres. In addition, nanostructured TiO2 films as those shown in Fig. 3 offer new possibilities to design improved photocatalysts. In a recent review97 the progresses in photocatalysis on mesostructured systems, which behave as spatially confined micro- and nanoreactors, were presented. Mesoporous oxides and hybrid photocatalysts based on molecularorganized assemblies offer new possibilities for converting low reactant 92 | Catalysis, 2009, 21, 82–130 This journal is

c


levels, to control the reaction environment, and realizing host–guest catalysis of light-driven reactions. An increase of selectivity can be attained through functionalization of spatially confined photo-reactors which allow coupling of different photo- and redox-catalysts, and to create hybrid micro-photoreactors including those which mimic the natural photosynthetic 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 efficient 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 offer new possibilities to realize robust and recyclable materials for cascade catalytic reactions. In a very recent review99 catalytic reactions within a confined 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 offers also new possibilities for nano-building of catalytic components, e.g. acting as nanoreactors for spatially confined synthesis of composite materials. Shchukin et al.101 have discussed the use of hollow nanostructures (nanotubes and capsules) for the synthesis of nanoparticles 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 different from that of Au on anatase or rutile TiO2 catalysts.105 Ag nanoparticles in nanotubes also showed a different recativity.106 Therefore, the TiO2 nanotubes could be used as nanoreactors for a controlled preparation of other guest nanoparticles which reactivity could be different 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 offer interesting possibilities to integrate in a defined nanospace organometallic complexes, metal particles, other active species (Bro¨nsted 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 effective 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, 82–130 | 93 This journal is

c


photoelectrodes were fabricated through immobilization of bacterial photosynthetic 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 films such as those schematically represented in Fig. 3 offer 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 photoanodes for conversion of light energy into electricity. The specific nanostructure of TiO2 allows to improve the performances with respect to films composed by porous TiO2 particles. In addition, materials prepared by anodization permit the growth of the oxide nanostructured film 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-confined nanodrops. 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 film 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 film 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 advantages 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 film 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 films of ionic liquids show enhanced enantioselectivity for chiral hydrogenation. Immobilization of organometallic complexes in thin supported films of ionic liquids allow generation of a new class of hydroamination catalysts.121 The concept of immobilizing organometallic complexes in a thin film 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)ferrocene—DPPF), and strong Bro¨nsted acidic functions.122 The materials showed exceptional catalytic activity for the 94 | Catalysis, 2009, 21, 82–130 This journal is

c


addition of aniline to styrene, providing the Markownikoff product under kinetically controlled conditions and mainly the anti-Markownikoff product in the thermodynamic region. Heterogeneization of liquid phase reactions in surface-confined nano-drops, e.g. inside TiO2 nanoreactors, offers additional possibilities to control and improve this class of reactions. Therefore, the concept of bidimensional nanostructured film presented in Fig. 3 offers several opportunities to develop novel catalysts. However, for practical applications a major problem is related to the fact that if a thin film on a flat surface is used, the amount of catalyst is quite low. For one cm2 of surface and a thickness of the titania film 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 film is deposited. A possible solution is to use high surface area foams over which the nanostructured titania film is then created. Another solution is to create a multi-stack of thin substrates over which the nanostructured titania film is present. Other solutions are possible, but the use of the concept presented in Fig. 3 requires first 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 film 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 film thickness. One of the active directions of research in this area, in fact, is to maintain the nanostructure, but increase the film 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 nanotube array growth to approximately 15 mm/h, representing a growth rate approximately 750–6000% higher than that seen, respectively, in other polar organic or aqueous based electrolytes previously used to form TiO2 nanotu be arrays. Catalysis, 2009, 21, 82–130 | 95 This journal is

c


The same authors also reported recently that self-aligned hexagonally closepacked 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 films up to 250 mm,125 but in general film thickness is few microns or below. For a film 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 nanomembranes. First attempts to prepare self-organized, free-standing TiO2 nanotube membranes for flow-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 floating in the etching solution. After being rinsed with methanol and distilled water, the layers were placed (closed tube side down) 1–2 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, flow-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 biofiltration 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.129–133 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, 82–130 This journal is

c


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 flux 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 field of air and water purification,137 purification of drinking water,138 novel membrane for high temperature PEM fuel cells,139,140 Li-ion batteries,141–143 advanced nanoelectrode arrays (NEA)144 and nanofiltration and pervaporation.145 The possibility of free-standing titania membranes with thickness of about 150–200 microns and straight channels of 100–200 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 film, 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, 82–130 | 97 This journal is

c


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 Energy’’146 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 specific peculiar aspect of nanostructured TiO2 films for catalytic applications 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 first 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, 82–130 This journal is

c


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 semiconductor, it is possible also to apply an external bias and further control the characteristics of the supported metal particles by this effect. In other words, with respect to alumina or silica, the catalytic properties of metal particles deposited over TiO2 can be modified 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 specific 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 field which influences 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 field (c). Source: Centi and Perathoner.147


c


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 influenced by the presence of a local magnetic field. Quite interesting, this observation opens an exciting opportunity to develop catalysts with photo-switchable reactivity. 3.

Synthesis of titania nanostructured films

We refer here only to the synthesis of 2D titania nanostructured films, 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 nanotubes55–58,65,136 and can be used for further understanding 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 different specific characteristics related to the nanostructure. For example, O2 adsorption is different 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 different 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 first attempts have been made using as a template nanotube aluminum oxide129,151,152 which is prepared by anodic oxidation of Al films153 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 difference between the two techniques is that casting is a filling 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) films 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 electrophoretic deposition allows the formation of an ordered packing.158 100 | Catalysis, 2009, 21, 82–130 This journal is

c


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 influence of an electrical field.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 nanoneedles 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 nanoflake array films could be produced by leaching of a solution phase from the glass surface through hydrothermal processes.164 Electron microscope observations revealed that the nanoflakes 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 films shows that a variety of nanostructures and types of array packing can be produced by the different methods, but that in general the preparation methods are either complex, costly or difficult 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 different 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, 82–130 | 101 This journal is

c


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 first who showed that a very high degree of order can be achieved for these porous geometries. Zwilling et al.171 first reported the porous surface of titania films electrochemically formed in fluorinated 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,173–175 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, microfluidics, molecular filtration, 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 1–5 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 specific 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 fluoride 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 fluoride 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 fluoride complexes, TiO2 + 6F - [TiF6]2 4+

Ti

+ 6F - [TiF6]

2

(2) (3)

Reaction (1) describes the oxide growth on an anodized metal surface, as schematically depicted in Fig. 8a in a fluoride-free electrolyte. Oxidized metal species react with O2 ions (from H2O) to form an oxide layer. 102 | Catalysis, 2009, 21, 82–130 This journal is

c


Fig. 8 Schematic representation of the Ti anodization (a) in absence of fluorides (results in flat layers), and (b) in presence of fluorides (results in the tube growth). F = Field Strength. Source: Schmuki et al.136

Further oxide growth is controlled by field-aided ion transport (O2 and Ti4+ ions) through the growing oxide. As the system is under a constant applied voltage, the field within the oxide is progressively reduced by the increasing oxide thickness, the process is self-limiting. In the presence of fluoride 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 field (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 film 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 significantly thinner walls than at their bottoms.182,183 The reason for separation into tubes probably should be ascribed to accumulation of fluoride 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 film 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 15–120 nm in the Catalysis, 2009, 21, 82–130 | 103 This journal is

c


potential range between 1 and 25 V.184 For mixed glycerol–water 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 fluoride ions.186,187 Recent results reported a thickness close to 1000 mm.124 Nanotube arrays greater than several tens of microns in length (film thickness) are mechanically robust to form of self-standing membranes. Usually the nanotube arrays have been made from a titanium thick film 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 film sputtered onto a variety of substrates, such as silicon and fluorine 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 film 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 difficult 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 influences 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 difference in growth characteristics was attributed to the inhibitory effect 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 closepacked structure191 similar to that observed in porous anodic alumina layers. Several factors strongly influence 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 first tube layer 104 | Catalysis, 2009, 21, 82–130 This journal is

c


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 first parameter set may lead to a first geometry, then underneath a second layer of tubes can be grown with a different 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 films 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.194–196 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 significant 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 temperature 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. TiO2–Nb2O5199 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 different concentration levels. While ClO4 ions are hardly incorporated, SO42 and particularly PO43 are incorporated into the entire tube to significant levels (some few atomic %).136 It is clear that significant amounts of F (about 1–5 at.%) are entering the TiO2 structure. Recent TEM investigations indicate that fluorides 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 fluorides 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 films, show that significant 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. Significant 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, 82–130 | 105 This journal is

c


mixed-oxide nanostructured films. 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 defined geometry resulting in a narrow distribution of diffusion 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 flat surface. In the previous section it was shown that significant 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 effects, 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.202–210 In general, it was found that annealed TiO2 nanotubes show higher photocatalytic efficiency in the degradation of dyes or pollutants than reference films 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 efficiency 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, 82–130 This journal is

c


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 films formed by an array of nanotubes (prepared by anodic oxidation) has still not provided convincing evidences for superior performance with respect to the difficulties in implementing this new technology of preparation. However, recent progress in the preparation and in particular the use of mixed oxide nanostructured films 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 efficiency in methylene blue degradation. The use of hybrid composites of MWCNT (multiwalled carbon nanotube) (20% wt.) and TiO2 nanotubes significantly 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 films. Recently, a two-dimensional TiO2/carbon nanowall composite material was fabricated by growing carbon nanowalls on a Ti sheet with hot filament chemical vapor deposition, followed by metal-organic chemical vapor deposition 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, 82–130 | 107 This journal is

c


application to Li ion batteries220 and for novel electrodes for methanol electrooxidation in fuel cells.221 An alternative interesting approach is that defined as tube-in-tube nanostructures.222 By annealing TiO2 nanotube arrays in the presence of carbon volatile precursors, carbon nanotubes were grown inside the inorganic nanotubes. Compared with unmodified TiO2, a hybrid C-TiO2 photocatalyst shows an enhanced efficiency 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 benefits 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 nanoparticles, resulting in an increase in its photocatalytic activity for oxidation of acetaldehyde. Enhancing photobehavior with visible-light A further significant 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 significantly higher visible-light photocatalytic activity for the degradation of methylene blue than pure TiO2 nanotubes or CdS nanoparticles.226 The highest photodegradation efficiency after 6 h irradiation was 84.5%. CdS quantum dots (QDs) sensitized TiO2 nanotube-array photoelectrodes were studied for their photoelectrochemical (PEC) performance.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 efficiency (up to 4.15%). These results demonstrate that the unique nanotube structure can facilitate the propagation 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 superior 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 nanoparticles 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 influenced. A 3-order enhancement in the charge injection rate was achieved by decreasing the CdSe particle 108 | Catalysis, 2009, 21, 82–130 This journal is

c


diameter from 8 nm to 2.4 nm.231 The TiO2-CdSe composite nanostructures show a maximum IPCE (photon-to-charge carrier generation efficiency) for 3 nm diameter CdSe nanoparticles of 35% using TiO2 nanoparticles and 45% for a tubular TiO2 morphology, demonstrating the role of TiO2 nanostructure. 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 modification 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 efficiency of CdS/TiO2 on ITO increased 50.0% in the presence of single-walled carbon nanotubes due to the improved charge-collecting efficiency 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 semiconductor 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 efficiency (IPCE) has been observed using such composite architecture. These results indicate significant progress in the design of nanostructure hybrid architectures for next generation photomaterials.235 The photoactive semiconductor nanoparticles (TiO2 or ZnO) when dispersed on conducting scaffolds of SWCNT exhibit improved photoconversion efficiency (IPCE) as the carbon network facilitates charge collection and charge transport in semiconductor 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 applications, 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. Modification 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. Significant activity has been made recently to produce doped titania nanotubes. Annealing titanium metal foils and titanium oxide films in a hydrocarbon flame forms carbondoped titania with a significantly enhanced full spectrum photoresponse. On annealing Ti metal foils in a natural gas flame at 850 1C, Khan et al.236 found the diffuse reflectance spectra of these samples to be significantly Catalysis, 2009, 21, 82–130 | 109 This journal is

c


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-modified TiO2 nanotubes prepared by wet chemical methods exhibited higher photocatalytic activity for the degradation of methylene blue than unmodified titanate nanotubes under artificial 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 efficient (i.e., gives high photocurrent at a low external bias, e.g. about 2.5–2.8 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 hexagonallyordered 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 flame) showed a 2 fold increase in photocurrent with respect to an undoped TiO2 nanotube film. 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 illumination 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 Efficiency) data. Murphy240 thus concluded that there is no convincing evidence in the literature of effective 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 films 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 films were synthesized using ZnO 110 | Catalysis, 2009, 21, 82–130 This journal is

c


nanorods as the template and doped with urea at 623 K.243 Under UV and visible light irradiation, the efficiencies 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 efficient electron field emission.245 Such a titania nanotube array shows very good stability at high field emission current (fluctuation o3% at field emission current of 160 mA within 4 h) and low turn-on and threshold fields (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 field emission into efficient materials through the introduction of acceptor and donor states above the valence band maximum and below the conduction band minimum (in the band gap of titania), respectively. 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 Vo–NO–Ti 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 effective reliability of the results regarding photoactivity with visible-light. There is instead good evidence that N is incorporated in TiO2 and modifies 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 photocatalytic 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, 82–130 | 111 This journal is

c


However, B-doped samples displayed stronger absorption in both UV and visible range and higher photoelectrocatalytic efficiency 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 flow 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 modification of the anodization bath, e.g. during the preparation of the TiO2 nanostructured film 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 efficiently 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 photobehavior with visible-light is an exciting and challenging direction. Several promising results have been reported, but sometimes not with the necessary scientific 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 fields 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).262–266 The titania nanotube array architecture results in a large effective surface area in close proximity with the electrolyte thus enabling diffusive transport of photogenerated holes to oxidizable species in the electrolyte. Separation of photogenerated charges is assisted by action of the electric field in the depletion region.267 In addition, due to light scattering within a porous structure, the incident photons are more effectively absorbed than on a flat electrode.268 A maximum conversion efficiency 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 320–400 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 efficiency 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, 82–130 This journal is

c


induction time, the amount of photogenerated H2 increases nearly linear, even if data for times longer than 2000 s were not given. Using UV (320–400 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 580–620 1C range the efficiency in H2 generation was about 12.2%. The increase in photocurrent and efficiency 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 densification 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 efficiency were observed to increase with the nanotube array film thickness.270 Fig. 9 evidences that a linear relationship was observed between film thickness (in the 2–6 micron range) and rate of H2 photo-generation.270 For a 6 mm thickness of the nanotube array film annealed at 600 1C the quantum efficiency 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 efficiency clearly indicates that the incident light is effectively utilized by the nanotube arrays for charge carrier generation. Using CdS/TiO2 nanotube array photoanodes (diameter of the nanotube of 80–100 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-film 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 films of different lengths annealed at 530 1C. Electrode area of 1 cm2; 100 mW/cm2 visible light. In the inset FESEM crosssectional image of 2.8 mm long TiO2 nanotube array prepared by anodic oxidation of a titanium foil in an electrolyte containing potassium fluoride (KF; 0.1 M), sodium hydrogen sulfate (1 M), trisodium citrate (0.2 M) and sodium hydroxide. Elaborated from Grimes et al.270


c


In a fluoride-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 0.1–0.5 M H2O2 (anodization potentials between 10–23 V).271 The use of ethylene glycol as the electrolyte medium significantly alters the anodization kinetics and resulting film 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 confirmed 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 potentiostatic anodization of Ti foil in fluoride 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 sonoelectrochemical method and its application for photoelectrochemical splitting of water was reported by Mohapatra.273 Self-ordered arrays of TiO2 nanotubes of 30–100 nm in diameter and 300–1000 nm in length (thus relatively thin films) can be rapidly synthesized under an applied potential of 5–20 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 efficient for incorporating carbon into TiO2 nanotubes. The photoelectrocatalytic 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 different solutions at the anode and cathode side (strong acid for H2 generation and strong basic for O2 generation) has significant consequences and the reaction mechanism is different. The same authors237 also discussed the optimal design of a photoelectrochemical (PEC) cell using C-doped TiO2 nanotube arrays as the photoanode and Pt nanoparticles incorporated in TiO2 nanotube arrays as 114 | Catalysis, 2009, 21, 82–130 This journal is

c


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 finally 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 significant progress has been made on the photoelectrocatalytic 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 films 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 confirm the results and have more extended verification, as well as a clear identification 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 film materials directly as catalysts, even if the cited progress in producing thicker films 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 briefly comment also on some selected results to show how to use the TiO2 array thin film 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 different oxygen adsorption.150 This feature is clearly relevant for catalysis. On TiO2 nanotubes array which consist of a 80/20% mixture Catalysis, 2009, 21, 82–130 | 115 This journal is

c


of the anatase/rutile, molecular O2 adsorption, influenced 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 CO–H2 and CO–H2O reactions over TiO2 nanotubes filled with Pt metal nanoparticles was investigated by Sato et al.281 Pt-TiO2 nanotube catalysts were prepared by employing fine fiber shaped crystals of [Pt(NH3)4](HCO3)2 complex as a structure templating material. The turnover 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 CO–H2 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 CO–H2 reaction, a similar promotion effect 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 field 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 electrochemical 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 specific 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, 82–130 This journal is

c


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 films, together with the possibility of stabilization of metal particles against sintering. There are thus an increasing number of publications on the methanol and ethanol electro-oxidation over these materials. Schmuki et al.286 first 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 significantly 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 amorphous 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 opportunities are offered 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 electrooxidation in acidic media was studied by Song et al.288 The results demonstrated 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 electrocatalytic 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 electrodeposited 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, 82–130 | 117 This journal is

c


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 diffusivity 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 effect 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 diffusion of adsorbed CO. The lattice relaxation and strong ligand effects 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 effects for Pt2Fe2 clusters were estimated. It must be finally 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 efficient electrocatalysts 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 efficient 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 flat 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 specific characteristics of the nanostructure. Various results have been also published indicating that metal or metal oxides supported on TiO2 nanotubes have different characteristics than when supported over conventional TiO2 particles or other type of supports. We briefly mention here some selected examples, because in principle different characteristics indicate also possible differences 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 different 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, 82–130 This journal is

c


conditions.299 Ni nanocylinders can be grown in titania nanotubes.300 Intergrown binary metal oxide nanotubes could be also produced. Selforganized porous TiO2-ZrO2301 and TiO2–Nb2O5199 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 films and be used as bio-catalyst.305 It was also observed that adhesion, spreading, growth, and differentiation of mesenchymal stem cells are critically dependent on the tube diameter.306 Conclusions Research in the field of TiO2 nanostructured films has progressed significantly 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 nanofibres, 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 field 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 photoelectrocatalytic production of H2 or the elimination of pollutants, and in developing advanced electrodes for fuel cells, particularly for direct methanol or ethanol oxidation. Nevertheless, also in this case the field 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 defined. However, there are clear indications that this will be a major area of research not only for this specific field, 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 electrocatalytic 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, 82–130 | 119 This journal is

c


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 offered 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 efficiently 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. Trifiro`, 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. Schlo¨gl, 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. Trifiro`, The Technology of Catalytic Oxidations, Editions TECHNIP, Paris, France, 2001. 10 G. Centi and S. Perathoner, Catal. Today, 2003, 79–80, 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 Schlo¨gl, Angew. Chem., Int. Ed., 2005, 44, 4704. 13 W. Weiss. R. Schlo¨gl, 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, 82–130 This journal is

c


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. Schlo¨gl, 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 Scientific 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, 44–45, 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 Scientific 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 Scientific 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. Trifiro`, A. Aboukais, C. F. Aı¨ ssi and M. Guelton, J. Phys. Chem., 1992, 96, 2617. 32 G. Bellussi, G. Centi, S. Perathoner and F. Trifiro`, 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 Scientific 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. Schlo¨gl, Mater. Res. Bull., 2007, 42, 354. Catalysis, 2009, 21, 82–130 | 121 This journal is

c


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 Scientific 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 Scientific 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, 336–338(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 Scientific 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, 82–130 This journal is

c


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 12–16th, 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, PreprintsAmerican 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, 82–130 | 123 This journal is

c


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 Scientific 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. Ballauff, 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. Pandolfi, 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, 82–130 This journal is

c


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. D’Epifanio, 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 6–8, 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, 82–130 | 125 This journal is

c


158 S. J. Limmer, T. P. Chou and G. Z. Cao, J. Sol-Gel Science and Techn., 2005, 36, 183. 159 X. Wang, C. Neff, 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, 82–130 This journal is

c


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 Scientific 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, 434–435, 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, 82–130 | 127 This journal is

c


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 6–10, 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 6–10, 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, 313–314, 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 6–10, 2008, PHYS-189. 235 P. V. Kamat, A. Kongkanand, Abstracts of Papers, 235th ACS National Meeting, New Orleans, LA, United States, April 6–10, 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, 82–130 This journal is

c


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, 82–130 | 129 This journal is

c


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.


c


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 figures2), much of which due to heterogeneous catalysis. Traditionally heterogeneous 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 scientifically 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 finds the cost of raw material and energy rapidly rising, traditional feedstock showing signs of depletion, and pollution taking a toll on our planet’s 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 topdown mode and is limited in its ability to meet these new demands efficiently. 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 offer a direct, spatially resolved view of catalytic surfaces, important clues for the nature of individual adsorbates and adsorbate-surface interaction, and sometime glimpses of molecular 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


c


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-defined clusters; mesoporous structures; nanotubes and nanowires; metal-organic frameworks; organic and organometallic 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 approximationbased exchange–correlation functionals (e.g. PW91,9 and PBE10); pseudopotential 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 finally enable calculations involving extended surfaces and large clusters, especially those containing metals. Together with quantum chemistry, energy minimization algorithms, transition state-seeking algorithms,15–17 mean-field and statistical thermodynamic and kinetic modeling techniques,18–20 d-band theory,21 bond order conservation and adsorption energy scaling relations,22–24 and BrønstedEvans-Polanyi (BEP) linear energy relations25–27 have led to significant 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 determination of catalyst structures, various energetic quantities, properties of transient intermediates and transition states, and reaction mechanisms. Once embodied in computational algorithms, theory offers 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 characterization 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 briefly survey two cross-cutting themes that have in recent years benefited substantially from theoretical insights, especially what first-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 effects of reactive physiochemical environments on catalytic reactivity. The relatively simple and predictable structure of metals 132 | Catalysis, 2009, 21, 131–153 This journal is

c


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 identified, 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 sufficiently elucidated. 2.1

Targeting activity and selectivity

Reactions involving simple reactants and products often offer simple descriptions of activity or selectivity. The Brønsted-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 calculations26–28 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 sufficient 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; (b–d) 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 coefficients (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.


c


functional dependence on DE when combined with a Langmuir-Hinshelwood (LH) type of mechanism.28,29 Jens Nørskov and co-workers have adeptly exploited the BEP relation to improve upon several established technological catalysts. An early success came with the identification of Co–Mo and Ni–Mo 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 identified several Fe–Ni alloys, which experiments confirmed 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 identified several promising candidates, including a Pt-based Bi–Pt surface alloy. Surprisingly, although Bi adsorbed on Pt strongly poisons Pt, the Bi–Pt 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 sulfides and nitrides). A correlation between the reaction rate and the strength of the M–X 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 sulfides when plotted as a function of calculated DEM–S.28 To further optimize the HDS catalysts, which are currently promoted Mo and W sulfides,37 structural databases of inorganic compounds have been screened, and a number of promising candidates, some quite complex sulfides, have been identified in the desirable range of DEM–S.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 different transition states (TS’s) 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, 131–153 This journal is

c


computationally screened several Ag-based binary alloys and identified a EO Cu–Ag alloy which had a DEa = EAC a –Ea ca. 0.1 eV larger than on pure Ag and should therefore be more selective toward EO. Experimental measurements of several Cu-modified Ag catalysts showed up to 50% increase in selectivity.40 Nørskov and co-workers tackled a different challenge when they identified Ni–Zn alloys to replace the more expensive Pd–Ag 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 calculations 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 Pd–Ag does between desorbing ethylene (weak DEC2H4) and hydrogenating acetylene (strong DEC2H2). Based on these criteria, Ni–Zn was picked for its low cost and predicted stability against segregation (Fig. 2). A Ni25Zn75 alloy catalyst was verified in the laboratory to have even greater selectivity than Pd–Ag. Linic and co-workers provided two additional examples of modifying selectivity for the ethylene epoxidation reaction. They demonstrated the promotional effect 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 field that interacts with the different dipoles of the two TS’s.42 More recently Christopher et al. synthesized and tested (100) facet-dominated Ag nanowire 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 catalysts, 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.


c


recently a theory-inspired new synthesis recipe enabled the production of anatase TiO2 single crystals with a significantly higher percentage of the more reactive {001} facets than conventional polycrystalline anatase particles possess.44 Hydrofluoric 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, OH–OH 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 Ru–Pt core-shell nanoparticles that were more active for the preferential oxidation (PROX) of CO in H2 than either pure Pt or Pt–Ru 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’ findings 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 different PtxAu1x surface alloys and identified 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 OH’s 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


c


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 Pd–Au(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.52–54

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 verified. 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 effects or the adsorption of reactive species; whether surface or bulk species form; and whether a catalytic material dissolves because of solvent, pH, and potential effects 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 effects of reaction environment

During catalysis, catalysts come into intimate contact with a microscopic world teeming with different chemical species (reactants, products, intermediates, spectators, solvents) and different forms of energy (thermal, electromagnetic). Their active states may be very different from as-prepared or observed states in UHV, which often serve as the basis of modeling. Researchers have long been aware of the significant effects that the reaction environment may have on the structure, chemical state, and reactivity of heterogeneous catalysts, but the molecular origin of such effects 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 reactions60–66 in situ, and first-principles methods are playing an important role in interpreting the experimental results and explaining environmental effects at the molecular level. Catalysis, 2009, 21, 131–153 | 137 This journal is

c


3.1

First-principles modeling of environmental effects

Given that quantum chemistry calculations directly provide electronic energies, which formally correspond to zero temperature and pressure, ways for connecting to finite, realistic temperature and pressure are needed. One method is first-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 ! X 1 surface GG gðT; p; fNi gÞ ¼ 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 flexibility 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,69–75 metal compounds,67,76–80 and finite particles.81–83 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 Brønsted (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.


c


Grand canonical Monte Carlo (MC) simulations provide a more direct and efficient way than the FPT-DFT approach to assess phase behavior on a fixed lattice and is particularly advantageous in handling phases with large periodicity or multiple adsorbate species. Cluster expansion can be combined with MC simulations to significantly reduce the number of first-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 segregation 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 environment change, the coverages of these species and their derivatives on the catalytic surface will respond. Co-adsorbates can affect 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 sufficiently destabilized at high coverage, resulting in a weaker differential 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 different reaction channels. For NO oxidation, Smeltz et al. have recently hypothesized a NO-assisted O2 activation pathway, via a peroxynitrite (NO–OO) 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 intermediate. 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 Ru–Pt 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 effect 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, 131–153 | 139 This journal is

c


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 dissociation 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-field kinetic modeling and kinetic Monte Carlo, and concluded that the neglect of fluctuations in the probability of finding 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 efforts to understand the actual action of the catalyst. Surface oxide phases, for example, have been identified or suggested on all platinum-group metals and noble metals,75,100–109 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, offer 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 efforts aimed at explaining the location of the excess oxygen atoms led to the identification of a O–Ru–O 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 confirmed 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 identified COcus + Ocus 140 | Catalysis, 2009, 21, 131–153 This journal is

c


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.73,117–119 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. 500–600 K and 10–30 atm, remains unclear, although the reaction mechanism has been elucidated in single-crystal studies.39 A p(4 4) (‘‘p4’’) superstructure was first identified by Rovida et al. using LEED after exposing Ag(111) to O2 between 400–500 K.120 It was not a typical chemisorbed O structure and was suggested to be a single O–Ag–O trilayer.121 The first 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 thermodynamically 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 = 1–5, correspond to the symmetrically different 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.


c


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 effective 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 classified 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 A˚2 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 A˚2 patches of the p(4 4) and c(3 5O3)rect overlayers, 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.


c


through painstaking STM, sXRD, CLS, and DFT work, also concluded that models based on the same Ag6 units fit the data best.108 They pointed out that the calculated energies of the new and old models were sufficiently close that it was beyond the limit of the current mainstream DFT formulations 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 effect 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 Cu–Ag 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 Ag–Cu alloy.

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


c


3.4

Electrochemical environment

Several important energy-related applications, including hydrogen production, fuel cells, and CO2 reduction, have thrust electrocatalysis into the forefront of catalysis research recently. Electrocatalysis involves several physiochemical environmental effects,129,130 which poses substantial challenges for the theoreticians. First, there is the electric potential which can affect 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 field at the interface that further affects interfacial reactivity. One approach to accounting for the potential and solution effects 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 first 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 field 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 effects of surface charging, interfacial electric field, 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 filling the interslab 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 first 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 differently charged slabs,135 the as-calculated energy (EDFT) needs to be corrected:

E ¼ EDFT þ

Zq

hVidQ þ qU:

0

is the volume-average electrostatic potential in the simulation cell Here hVi that arises from the metal–water interface and the background charge, and 144 | Catalysis, 2009, 21, 131–153 This journal is

c


q is the number of excess electrons added. This E amounts to a potentialdependent free energy. The kinetics of electron-transfer reactions, which is also affected by the electrode potential and the metal–water interface, is more difficult 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 techniques such as the NEB method16 are not designed to search for minimum-energy reaction paths subject to a constant potential. Approximations that allow affordable quantum chemistry calculations to get around this limitation have been proposed, ranging from the electron affinity/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 0.7–0.9 V, respectively). On Cu(111), sufficiently high potentials were found to cause the dissolution of Cu2+ ions and the formation of a passivating Cu2O overlayer.130 Rossmeisl et al. demonstrated 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 different 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.


c


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 suffers from large overpotential loss even on the best monometallic catalyst, Pt, which necessitates high Pt loadings to compensate. Intense effort has been spent to understand the mechanism of the ORR and the origin of the overpotential, and to find a lower-cost alternative to pure Pt catalysts. Nørskov 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 field effects into consideration and found that Pt is located near the transition between the O2 activation-limited regime (Ag and Au) and the protonationlimited 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 confirmed 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 electrocatalytic oxidation of methanol and CO, have likewise been studied taking potential and solution effects into consideration.130,143 4.

Outlook

Impressive progress in heterogeneous catalysis has been made with the help of theory and first-principles methods. The Pd–Au alloy catalyst created by Goodman and co-workers for vinyl acetate synthesis represents an encouraging 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 simultaneous consideration of multiple descriptors. For instance, Janik et al. have identified a set of criteria for designing improved anode catalysts for direct methanol fuel cells.130 These include preferential breaking of C–H vs. O–H 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., finite 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 specific transformations. The local structures of these catalysts and 146 | Catalysis, 2009, 21, 131–153 This journal is

c


pertinent reaction mechanisms are being elucidated with the help of theory.144–146 The kind of surprise that local nanometer-scale structures can hold is illustrated by the MoS2 HDS catalysts and supported sizeselected 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 sizedependent electronic structure and cluster-support interaction strongly influence their reactivity.152,153 Indeed, nanomaterials, whose structures are well-defined at the nanometer or a¨ngstrom 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 effects 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 identified for Pd and Ag interacting with oxygen, which suggests that the catalyst surfaces may be in a state of flux 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, Scheffler and co-workers suggest that the state of the surface should be expected to fluctuate 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 A˚2; 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 figure with permission from Bollinger et al., Physical Review Letters, 2001, 87, 196803.147 r American Physical Society.


c


obstacle for understanding dynamic behavior of catalysts is the lack of an efficient 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 fly, 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 effects) 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 Laboratory’s Center for Nanophase Materials Sciences, which is sponsored by the Scientific User Facilities Division, Office of Basic Energy Sciences, US Department of Energy. References 1 M. W. Roberts, Catalysis Letters, 2000, 67, 1. 2 D. Filmore, Today’s Chemist At Work, 11/2002, 33. 3 J. H. Sinfelt, Bimetallic Catalysts: Discoveries, Concepts, and Applications, Wiley, New York, 1983. 4 C. N. Satterfield, 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. Roeffaers, 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. Schlu¨ter and C. Chiang, Physical Review Letters, 1979, 43, 1494. 12 D. H. Vanderbilt, Physical Review B, 1990, 41, 7892. 13 P. E. Blo¨chl, Physical Review B, 1994, 50, 17953. 14 G. Kresse and D. Joubert, Physical Review B, 1999, 59, 1758. 148 | Catalysis, 2009, 21, 131–153 This journal is

c


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. Scheffler, 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. Nørskov, 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. Nørskov, Physical Review Letters, 2007, 99, 016105. 25 V. Pallassana and M. Neurock, Journal of Catalysis, 2000, 191, 301. 26 J. K. Nørskov, 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. Nørskov, 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. Nørskov, Journal of the American Chemical Society, 2001, 123, 8404. 31 A. Logadottir, T. H. Rod, J. K. Nørskov, 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. Nørskov, 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. Nørskov and C. H. Christensen, Topics in Catalysis, 2007, 45, 9. 34 J. Greeley, T. F. Jaramillo, J. Bonde, I. B. Chorkendorff and J. K. Nørskov, 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. Nørskov, M. Scheffler 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. Nørskov, 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, 131–153 | 149 This journal is

c


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. Clearfield 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. Nørskov and F. Besenbacher, Physical Review B, 1997, 56, 5822. 58 A. V. Ruban, H. L. Skriver and J. K. Nørskov, Physical Review B, 1999, 59, 15990. 59 J. Greeley and J. K. Nørskov, 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. Nørskov, 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. Stampfl, Catalysis Today, 2005, 105, 17. 70 F. Faglioni and W. A. Goddard, Journal of Chemical Physics, 2005, 122, 014704. 71 D. Loffreda, 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. Scheffler, Physical Review Letters, 2007, 98, 046101. 74 H. Q. Shi and C. Stampfl, Physical Review B, 2008, 77, 094127. 150 | Catalysis, 2009, 21, 131–153 This journal is

c


75 R. B. Getman, Y. Xu and W. F. Schneider, Journal of Physical Chemistry C, 2008, 112, 9559. 76 Z. Lodziana, J. K. Nørskov 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. Scheffler, 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. Nørskov, 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. Scheffler, Physical Review Letters, 2004, 93, 116105. 99 B. Temel, H. Meskine, K. Reuter, M. Scheffler 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. Stampfl, 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. Scheffler, 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. Stampfl, Physical Review B, 2006, 73, 165424. Catalysis, 2009, 21, 131–153 | 151 This journal is

c


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. Scheffler 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. Stampfl, 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. Stampfl and M. Scheffler, Physical Review Letters, 1997, 78, 1500. 113 A. Bo¨ttcher, 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. Stampfl and M. Scheffler, 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. Stampfl and M. Scheffler, Physical Review B, 2003, 68, 165412. 123 W. X. Li, C. Stampfl and M. Scheffler, 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. Loffreda, P. Sautet, A. Alavi and D. A. King, Journal of the American Chemical Society, 2003, 125, 5620. 126 M. L. Bocquet and D. Loffreda, Journal of the American Chemical Society, 2005, 127, 17207. 127 A. Michaelides, K. Reuter and M. Scheffler, Journal of Vacuum Science & Technology A, 2005, 23, 1487. 128 S. Piccinin, C. Stampfl and M. Scheffler, 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. Nørskov, 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. Nørskov, 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, 131–153 This journal is

c


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. Nørskov, 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. Nørskov, 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 ChemieInternational 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. Nørskov, 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. Topsøe, K. W. Jacobsen, E. Laegsgaard, J. K. Nørskov 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.


c


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 specific 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 first research efforts involved the synthesis of ethylammonium nitrate.1 Hurley and Wier at the Rice Institute in Texas, 1948, developed the first ionic liquids with chloroaluminate ions as bath solutions for electroplating aluminum.2 These liquids were studied primarily for their applications as electrolytes in electrochemistry 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 different 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 applications 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


c


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 influence of the cation is shown by investigations of the solubility of 1-octene in different 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 homogeneouslycatalyzed reactions. In this review, we focus our attention on those physical

Fig. 2 Solubility of 1-octene in four different tri-n-alkylmethylammonium tosylate melts at 80 1C, n (C) = number of C atoms of the alkyl residue.


c


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-alkylammonium salts have limited thermal stability, owing to decomposition via the Hoffmann 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 noncoordinating solvents. They can be expected, therefore, to have a strong rate-enhancing effect on reactions involving cationic intermediates. - Ionic liquids containing chloroaluminate ions are strong Lewis, Franklin and Brønsted 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 offer 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, 154–190 This journal is

c


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 first 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 trifluoromethanesulfonate anion, X = halide, and OMs = methanesulfonate salt. The reactions observed are described in Scheme 1.

Scheme 1

When the IL anion was trifluoromethanesulfonate (triflate), 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, 154–190 | 157 This journal is

c


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 first ever international meeting devoted to room-temperature ionic liquids,’’ noted ionic liquids expert Kenneth R. Seddon of Queen’s University, Belfast, Northern Ireland. ‘‘And it is the first 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 briefly 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 properties 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 industrial 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 comprehensive 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 aluminumcontaining species such as aluminum halides, especially aluminum chlorides. It seems appropriate to briefly 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 different cations. The patent does not report any reaction results. 158 | Catalysis, 2009, 21, 154–190 This journal is

c


In the following we consider the use of ILs as solvents for homogeneous 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 amidocarbonylation 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 liquid–catalyst 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 ammonium 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 significant palladium leaching into the products. Arene carbonylation Carbonylation of the arene system can be achieved using highly acidic media such as triflic 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, 154–190 | 159 This journal is

c


Lewis acid catalysts, such as AlCl3, and hazardous Brønsted acids such as HF. The ionic liquid can function as both a catalyst and a solvent for acid catalyzed processes. Since Lewis and Brønsted 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 first 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 affords the tetrachloroaluminate, [emim][AlCl4] which is referred to as being neutral and is not active as a Friedel–Crafts catalyst. In contrast, the 2+1 adduct, [emim][Al2Cl7] is strongly acidic and was shown to be very active in Friedel–Crafts 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) afforded 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 Friedel–Crafts alkylation of benzene with long chain a-olefins 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 alkylbenzenes using chloroaluminate ionic liquids has been described.32 The potential to retrofit existing installations with the ionic liquid catalyst offers enormous benefits 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-alkylimidazolium (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 (1–3 bar) dramatically improved the reactivity of the system, which was modeled quite effectively 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, 154–190 This journal is

c


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 (B1–3 MPa), reaction times 5–8 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 first. The effect 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 fit by a linear relationship with an intercept of zero, one could conclude that an irreversible reaction

Fig. 3

Effect of CO partial pressure upon yields.


c


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 olefins with the urea–H2O2 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 olefins 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 Schiff-base complex, has been conducted in [bmim][PF6].42 However, dichloromethane was required as a cosolvent, as the ionic liquid solidifies at the reaction temperature (0 1C), which nullifies a primary incentive for using an ionic liquid. The ionic liquid containing the catalyst could be recovered and recycled four times, albeit with a significant loss in yield. A more recent and very exciting development is the electro-assisted biomimetic activation of molecular oxygen by a chiral Mn Schiff-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 olefin. This would appear to offer potential for clean, electro-catalytic oxidations with molecular oxygen in ionic liquid media. Diels–Alder reactions in ILs It was reported44 that Sc(OTf)3 catalyzes Diels–Alder reactions in [bmim][X] (X = BF4, SbF6 or OTf), in this case at much lower catalyst loadings (0.2 m%). In contrast to the Friedel–Crafts 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, 154–190 This journal is

c


Scheme 4

The Diels–Alder 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-methylimidazoliumtetrafluoroborate, [hmim][BF4].45

Scheme 5

The potential activation of different Lewis acid catalysts and their load effect 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 (R1QMe–CQO; R2 = R3 = H; and R1QEt–CQO; 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 (16–17%) 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 effect 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 triflate-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 triflates are strong Lewis acids known to be quite effective 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 triflate group. On the other hand, reaction times of 1 hour were required for Catalysis, 2008, 21, 154–190 | 163 This journal is

c


complete transformation when either triflic or triflamic 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 triflate, 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 Diels–Alder reactions. In fact, cerium triflate performs better than any other catalyst tested, although scandium triflate 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 tetrafluoroborate with cerium triflate, as well as with scandium triflate, 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 five 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 olefins 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 hexafluorophosphate 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 triflate 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, 154–190 This journal is

c


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 influence 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 olefin and copper triflate. This polarized complex then reacts with the Pd(0)-p-complex with the substrate to form the final 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 confirm 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 effect 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, 154–190 | 165 This journal is

c


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 esterifications in ILs Brønsted 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 esterification of alcohols with acids.50 These agents were characterized using FTIR and NMR spectroscopy. The ionic liquids as catalysts afforded good alcohol conversion and excellent ester selectivity. The Fischer esterification of acetic acid with benzyl alcohol using different Brønsted 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 significant loss in catalytic activity was observed on recycling. Friedel–Crafts reactions in ILs Chloroaluminate ionic liquids modified with HCl were recently shown to give higher rates and more favorable product distributions in Friedel–Crafts alkylations, and these results were attributed to the superacidities of these media. In this context it is also worth mentioning the work of Ho¨lderich et al.51 who showed that ionic liquids immobilized on inorganic supports (SiO2, Al2O3, TiO2, ZrO2) were effective catalysts for the Friedel–Crafts 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 afforded the monoalkylated product in 98% selectivity at 99% conversion at 80 1C. Hardacre et al. report the Friedel–Crafts 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 significantly higher using ionic liquids compared with organic solvents.52 Continuous-flow 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


c


The acid liberated was quantified using both titration experiments and ion-exchange experiments with sodium-exchanged zeolites. Seddon and coworkers53 studied the Friedel–Crafts 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 triflates, in particular Sc(OTf)3, have been widely studied as water-tolerant Lewis acids in a variety of transformations, including Friedel–Crafts alkylations and acylations.54 Song and coworkers55 have recently shown that Sc(OTf)3 catalyzes the Friedel–Crafts alkylation of aromatics with olefins in hydrophobic 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, afforded 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 afford 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 offers potential environmental and economic benefits.

Scheme 11


c


A new protocol was developed for the synthesis of N-substituted thioamides, employing arenes and isothiocyanates in 1-butyl-3-methylimidazolium chloroaluminate 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 effect of Lewis acidity and the stoichiometry 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 Brønsted acid-demanding reactions.33–38 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 effect 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, 154–190 This journal is

c


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 assignments 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 Friedel–Crafts benzylation reaction using ILs of 1-butyl-3-methylimidazolium chloride-ZnCl2 ([bmim][ZnCl3]), 1-butyl-3-methylimidazolium chloride–FeCl3 ([bmim][FeCl4]), 1-butyl-3-methylimidazolium chloride-FeCl2 ([bmim][FeCl3]) and 1-butyl-3methylimidazolium chloride–AlCl3 ([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 effect of the Lewis acid upon the reactivity was as follows: AlCl3 4 FeCl3 4 FeCl2 4 ZnCl2. The effects of reaction temperature, reaction time, and the ratio of metal chloride to [bmim]Cl, as well as the amount of IL on the Friedel–Crafts 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 C–C coupling reaction where an unsaturated hydrocarbon or arene halide/triflate/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 effect 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, hexafluorophosphate or tetrafluoroborate 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, 154–190 | 169 This journal is

c


Scheme 14

The authors showed that the above reaction carried out in N-hexylpyridinium 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 efficient in the ILs when compared to conventional solvents.

Henry reactions in ILs The Henry reaction is an important carbon–carbon 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


c


substrates. Low yields of the product (2–30%; 35–40 h reaction time) were observed for equimolar amounts of the chloroaluminate and 1-butylpyridinium chloride, [bp][Cl]. But much higher yields were recorded (60–80%) in shorter reaction times (15–16 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 offered no explanation for the apparent order of reactivity with these ILs, and they could be reused up to five 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 effective catalyst for the biphasic hydrogenation of olefins, with turnover frequencies up to 540 h1.66 Similarly, (bmim)3-Co(CN)5 dissolved in [bmim][BF4] catalyzed the hydrogenation of butadiene to but-1-ene, with 100% selectivity at complete conversion. More recently, the ruthenium-catalyzed hydrogenation of sorbic acid to cishex-3-enoic acid, Scheme 16, was achieved in a biphasic bmim-PF6–methyl tert-butyl ether (MTBE) system.67 The ruthenium cluster [H4Ru(Z6-C6H6)4] [BF4]4, in [bmim][BF4], was shown to be an effective 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, 154–190 | 171 This journal is

c


Scheme 17

bmim-SbF6–isopropyl alcohol, Scheme 17.69 The observed enantioselectivity 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 enantioselective 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 afforded the anti-inflammatory 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 significant 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 DiPFc–Rh and the chiral EtDuPHOS–Rh 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 IL–supercritical CO2, scCO2, for catalytic hydrogenation.72,73 Tumas and coworkers showed that the catalytic hydrogenation of olefins 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 hydrogenation of tiglic acid, Scheme 19, and the precursor of the anti-inflammatory drug, ibuprofen, Scheme 20, using Ru(OAc)2(tol-BINAP) as the catalyst. 172 | Catalysis, 2009, 21, 154–190 This journal is

c


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 different solvent effects 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, exemplified by tiglic acid, give higher enantioselectivities at lower H2 concentrations. Hydroformylation Hydroformylation of propene in an aqueous biphasic system, using a watersoluble rhodium complex of the sodium salt of trisulfonated triphenylphosphine (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 olefins owing to the very low solubility of higher olefins in water. Hence, one can envisage that the use of an appropriate IL could provide the basis for biphasic hydroformylation of higher olefins. As noted earlier, Parshall showed in 1972 that platinum-catalyzed hydroformylations could be performed in tetraethylammonium trichlorostannate melts.76 More recently, Waffenschmidt and Wasserscheid77 studied the platinumcatalyzed 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


c


The ruthenium- and cobalt-catalyzed hydroformylation of internal and terminal olefins in molten tetra-n-butylphosphonium bromide was reported by Knifton in 1987.78 More recently, the rhodium-catalyzed hydroformylation of hex-1-ene was conducted in molten phosphonium tosylates, e.g. Bu3PEt+TsO2 and Ph3PEt+TsO2, having melting points of 81–83 1C and 94–95 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-modified diphosphine ligands containing a xanthene backbone.81 Xanthene-based diphosphine ligands with large bite angles (P–metal–P bond angles B1101) are known to give high selectivities (Z98%) towards the linear aldehyde.82 Biphasic hydroformylation of oct1-ene, Scheme 22, using rhodium complexes of these ligands in [bmim][PF6], afforded 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 coordinating anions, for the biphasic hydroformylation of hex-1-ene catalyzed by rhodium complexes of modified 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 modification of the ligands with cationic (guanidinium or pyridinium) or anionic (sulfonate) 174 | Catalysis, 2009, 21, 154–190 This journal is

c


groups. Another interesting recent development is the rhodium-catalyzed biphasic hydroformylation of oct-1-ene in [bmim][PF6]–scCO2 in a continuous flow process.84 Because of the low solubility of Rh–tppms and Rh–tppts complexes in the IL, [pmim][Ph2PC6H4SO3] (pmim = 1-propyl-3-methylimidazolium) was synthesized and used together with Rh2(OAc)4 as the catalyst precursor. 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 significant 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 hexafluorophosphate ([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-Hamilton’s group has also developed a continuous flow 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(trifluoromethanesulfonyl)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 flow 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, 154–190 | 175 This journal is

c


with CO and H2 in an scCO2 system. They conclude that a proper design of catalyst ligands was required to maintain the catalyst’s 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 cobaltbased catalyst in a non-aqueous liquid with improved catalyst recycling. A process for hydroformylating olefinically 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 olefins or olefinically 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 paraffins 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 paraffin 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 effect of adding these Lewis acids to the reaction mixture showed that incremental yields per mole of additive are different. 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 effective as the ‘‘harder’’ acids in catalyzing the paraffin isomerization. 176 | Catalysis, 2009, 21, 154–190 This journal is

c


Fig. 4

Incremental yields for addition of Lewis acids to chloroaluminate ILs.

Metathesis of olefins Hembre et al.96 disclosed a process for which one olefin selected from certain a,b-dihydroxyalkenes and 4-(alkenyl)ethylenecarbonates was reacted with a second olefin to produce a metathesis product. When the first olefin 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 effective 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 effective 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 C–C and C–hetero atom bonds. Karodia’s group97 studied the use of the ionic liquid ethyltri-n-butylphosphonium tosylate (n-Bu3PEtOTs) as a solvent for Catalysis, 2008, 21, 154–190 | 177 This journal is

c


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 conditions 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 recyclability 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 significant 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 C–N 178 | Catalysis, 2009, 21, 154–190 This journal is

c


and C–S 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 effective 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 hexafluorophosphate 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 Brønsted 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 trifluoromethanesulfonate imidazolium (e.g., [MBsIm][CF3SO3]) IL was effective 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 Brønsted 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-trimethylN-propanesulfonic acid ammonium hydrogen sulfate [TMPSA][HSO4] was an effective catalyst for the Pechmann reaction of various phenols with ethyl acetoacetate.

Sonogashira reaction The Sonogashira reaction is a C–C 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, 154–190 | 179 This journal is

c


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 different 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 ½PdðZ3 -C3 H5 ÞCl2 =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, 154–190 This journal is

c


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 offered advantages over conventional 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 first 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 first 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 different 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 confirming the success of the quaternization reaction. This Catalysis, 2008, 21, 154–190 | 181 This journal is

c


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 specific ionic liquids (TSIL) Recently, reports have appeared on a class of ILs known as task specific ionic liquids (TSIL). The term was introduced by J. H. Davis, Jr’s group120 to refer to those ILs which have functional groups attached to them so as to give specific properties and functionalities. Thus, they not only perform specific 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.125–131 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 esterification giving the desired TSIL. Diels-Alder cycloaddition was carried out with 2,3-dimethylbutadiene and cyclopentadiene to give corresponding adducts. After washing with ether, transesterification resulted in cyclohexene derivatives, Scheme 29. 182 | Catalysis, 2009, 21, 154–190 This journal is

c


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 transesterified to give the final 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, 154–190 | 183 This journal is

c


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-tetramethylpiperidine-1-oxyl (TEMPO) appended to an imidazolium cation for the oxidation 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. Bazureau’s group reported extensive work with TSILs for synthesis of various 4-thiazolidinones, 1,4-dihydropyridines, 3,4-dihydropyrimidin-2(1H)-ones and polyhydroquinones.136–140 They functionalized imidazole compounds with esters, which were then reacted with reagents to afford 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 homocoupling 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. Significant 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 significant 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 filtration 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 trifluoromethanesulfonate, [mammim][TfO]. Telomerization in ILs Telomerization is the reaction of olefins having conjugated double bonds (conjugated dienes) in the presence of a nucleophile (telogen). The main 184 | Catalysis, 2009, 21, 154–190 This journal is

c


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 effective catalysts for telomerization.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 olefins 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 olefins could be either 1,3-butadiene or isoprene. Vijayaraghavan et al.150 completed a calorimetric study of the telomerization reaction of acrylonitrile and trichloroacetyl chloride to give an important intermediate for a commercial pesticide, using [C4mpyr][NTf2], N-butyl-N-methylpyrrolidinium bis(trifluoromethanesulfonyl)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 effectively. 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, 1291–1298. 2 (a) F. H. Hurley, US Patent 2,446,331, 1948; (b) F. H. Hurley and T. P. Wier, J. Electrochem. Soc., 1951, 98, 207–212. 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, 3772–3789. Catalysis, 2008, 21, 154–190 | 185 This journal is

c


8 P. Bonhoˆte, A. Dias, N. Papageorgiou, K. Kalyanasundaram and M. Gra¨tzeb, 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), 707–710. 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 liquids—from molten salts to neoteric solvents’’, Green Chemistry, 2002, 4, 73–80. 14 K. Holbrey and K. Seddon, ‘‘Ionic Liquids’’, Clean Products and Processes, 1999, 1, 223–236. 15 M. Earle and K. Seddon, ‘‘Ionic liquids. Green solvents for the future’’, Pure Appl. Chem., 2000, 72(7), 1391–1398. 16 P. Wasserscheid and W. Keim, ‘‘Ionic Liquids—New ‘Solutions’ for Transition Metal Catalysis’’, Angew. Chem. Int. Ed., 2000, 39, 3772–3789. 17 M. Freemantle, ‘‘Eyes on Ionic Liquids’’, Science/Technology, 2000, 78(20), 37–50. 18 M. Freemantle, ‘‘Ionic liquids in Organic synthesis—Room-temperature ionic liquids provide unique environment for organic reactions’’, Science/Technology, 2004, 82(45), 44–49. 19 R. Sheldon, ‘‘Catalytic reactions in ionic liquids’’, Chem. Comm., 2001, 2399–2407. 20 A. Corma and H. Garcia, ‘‘Lewis Acids: From Conventional Homogeneous to Green Homogeneous and Heterogeneous Catalysis’’, Chem. Rev., 2003, 103, 4307–65. 21 T. Welton, Chem. Rev., 1999, 99, 2071. 22 J. Dupont, R. de Souza and P. Suarez, Chem. Rev., 2002, 102, 3667–3692. 23 D. Zhao, M. Wu, Y. Kou and E. Min, Catalysis Today, 2002, 74, 157–189. 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. OlivierBourbigou 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. Bala´zs, C. Benedek and S. To˜ro¨s, J. Mol. Catal. A: Chemical, 2005, 244, 105–109. 31 J. Boon, J. Levisky, J. Pflug 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, 51–58. 34 E. Angueira and M. White, A.I.Ch.E. Journal, 2005, 51, 2778–2785. 35 E. Angueira and M. White, J. Mol. Catal. A: Chemical, 2005, 238, 163–174. 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 8–11, 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, 154–190 This journal is

c


41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75

G. Owens and M. Abu-Omar, Chem. Commun., 2000, 1165. C. Song and E. Roh, Chem. Commun., 2000, 837. L. Gaillon and F. Bedioui, Chem. Commun., 2001, 1458. C. E. Song, W. H. Shim, E. J. Roh, S. Lee and J. H. Choi, Chem. Commun., 2001, 1122. G. Silvero, M. Jose´, A. Jose´, L. Bravo, M. Valos, J. Jimeńez and I. Lo´pez, Tetrahedron, 2005, 61, 7105–7111. J. Peng, J. Li, H. Qui, J. Jiang, K. Jiang, J. Mao and G. Lai, J. Mol. Catal. A: Chemical, 2006, 255, 16–18. P. Wasserscheid and M. Eichmann, Catalysis Today, 2001, 66, 309–316. D. Thiele and R. de Souza, J. Mol. Catal. A: Chemical, 2005, 264, 293–298. M. Kim and J.-k. Lee, US Patent 7,005,282, February 28, 2006. T. Joseph, S. Sahoo and S. Halligudi, J. Mol. Catal. A: Chemical, 2005, 234, 107–110. C. DeCastro, E. Sauvage, M. Valkenberg and W. Ho¨lderich, J. Catal., 2000, 196, 86. C. Hardacre, S. Katdare, D. Milroy, P. Nancarrow, D. Rooney and J. Thompson, Journal of Catalysis, 2004, 227, 44–52. C. Adams, M. Earle, G. Roberts and K. Seddon, Chem. Commun., 1998, 2097. For reviews see: S. Kobayashi, Synlett, 1994, 689; S. Kobayashi, Eur. J. Org. Chem., 1999, 15. M. Yoon, J. Kim, D. Choi, U. Shin, J. Lee and C. Song, Adv. Syn. Cat., 2007, 349, 1725–1737. P. Naik, S. Nara, J. Harjani and M. Salunkhe, Can. J. Chem., 2003, 81, 1057–1060. S. Nara, J. Harjani and M. Salunkhe, J. Org. Chem., 2001, 66, 8616–8620. D. Yin, C. Li, L. Tao, N. Yu, S. Hu and D. Yin, J. Mol. Catal. A: Chemical, 2006, 245, 260–265. D. Kaufmann, M. Nouroozian and H. Henze, Synlett, 1996, 1091. A. Carmichael, M. Earle, J. Holbrey, P. McCormac and K. Seddon, Organic Letters, 1999, 1(1), 997–1000. J. Howarth and A. Dallas, Molecules, 2000, 5, 851–855. V. Calo´, A. Nacci, L. Lopez and A. Napola, Tetrahedron Lett., 2001, 42, 4701. V. Calo´, A. Nacci, A. Monopoli, L. Lopez and A. di Cosmo, Tetrahedron, 2001, 57, 6071–6077. A. Kumar and S. Pawar, J. Mol. Catal. A: Chemical, 2005, 235(1-2, 1), 244–248. P. Suarez, J. Dullins, S. Einloft, R. de Souza and J. Dupont, Polyhedron, 1996, 15, 1217. P. Suarez, J. Dullins, S. Einloft, R. de Souza and J. Dupont, Inorg. Chim. Acta, 1997, 255, 207. S. Steines, B. Drießen-Ho¨lscher and P. Wasserscheid, J. Prakt. Chem., 2000, 342, 348. P. Dyson, D. Ellis, D. Parker and T. Welton, Chem. Commun., 1999, 25. Y. Chauvin, L. Mussman and H. Olivier, Angew. Chem., Int. Ed. Engl., 1995, 34, 2698. A. Monteiro, F. Zinn, R. de Souza and J. Dupont, Tetrahedron: Asymmetry, 1997, 8, 177. K. Boyle, E. Lipsky and C. Kalberg, Tetrahed. Lett., 2006, 47, 1311–13. F. Liu, M. Abrams, R. Baker and W. Tumas, Chem. Commun., 2001, 433. R. Brown, P. Pollet, E. McKoon, C. Eckert, C. Liotta and P. Jessop, J. Am. Chem. Soc., 2001, 123, 1254. R. Noyori, Asymmetric Catalysis in Organic Synthesis, Wiley, New York, 1994. E. Kuntz, CHEMTECH, 1987, 570. Catalysis, 2008, 21, 154–190 | 187 This journal is

c


76 G. Parshall, J. Am. Chem. Soc., 1972, 94, 8716. 77 H. Waffenschmidt 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. Waffenschmidt and P. Wasserscheid, Organometallics, 2000, 19, 3818. 81 P. Wasserscheid, H. Waffenschmidt, 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. Waffenschmidt and P. Wasserscheid, J. Catal., 1999, 186, 481. 86 C. Mehnert, R. Cook, N. Dispenziere and E. Mozeleski, Polyhedron, 2004, 23, 2679–2688. 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, 15577–15588. 91 U. Hintermair, G. Zhao, C. Santini, M. Muldoon and D. Cole-Hamliton, Chem. Commun., 2007, 1462–1464. 92 A. Frisch, P. Webb, G. Zhao, M. Muldoon, P. Pogorzelec and D. Cole-Hamliton, Dalton Trans., 2007, 5531–5538. 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. Hoff, 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, 11039–11043. 98 L. Yang, L. Xu, W. Zhou, L. Li and C. Xia, Tetrahedron Lett., 2006, 47, 7723–7726. 99 D.-g. Gu, S.-j. Ji, H.-x. Wang and Q.-y. Xu, Synth. Commun., 2008, 38, 1212–1223. 100 J.-m. Xu, Q. Wu, Q.-y. Zhang, F. Zhang and X.-f. Lin, Eur. J. Org. Chem., 2007, 1798–1802. 101 Y. Sharma and M. Degani, J. Mol. Catal. A: Chemical, 2007, 277, 215–220. 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, 249–252. 104 V. Singh, S. Kaur, V. Sapehiyia, J. Singh and G. Lad, Cata. Comm., 2005, 6, 57–60. 105 V. Soares, M. Alves, E. Souza, I. Pinto, J. Rubim, C. Andrade and P. Suarez, Int. J. Mol. Sci., 2007, 8, 392–398. 106 Y. Gu, J. Zhang, Z. Duan and Y. Deng, Adv. Synth. Catal., 2005, 347, 512–516. 107 F. Dong, C. Jian, G. Kai, S. Qunrong and L. Zuliang, Catal. Lett., 2008, 121, 255–259. 188 | Catalysis, 2009, 21, 154–190 This journal is

c


108 T. Fukuyama, M. Shinmen, S. Nishitani, M. Sato and I. Ryu, Organic Letters, 2002, 4(10), 1691–1694. 109 A. Gholap, K. Venkatesan, R. Pasricha, T. Daniel, R. Lahoti and K. Srinivasan, J. Org. Chem., 2005, 70, 4869–4872. 110 Y. Li, J. Zhang, W. Wang, Q. Miao, X. She and X. Pan, J. Org. Chem., 2005, 70, 3285–3287. 111 A. Corma, H. Garcia and A. Leyva, Tetrahedron, 2005, 61, 9848–9854. 112 M. Rahman, T. Fukuyama, I. Ryu, K. Suzuki, K. Yonemura, P. Hughes and K. Nokihara, Tetrahedron Lett., 2006, 47, 2703–2706. 113 J.-C. Hierso, J. Boudon, M. Picquet and P. Meunier, Eur. J. Org. Chem., 2007, 583–587. 114 P. de Lima and O. Antunes, Tetrahedron Lett., 2008, 49, 2506–2509. 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. Ho¨lderich, US Patent 6,969,693 November 29, 2005. 117 M. Valkenberg, C. de Castro and W. Ho¨lderich, Spec. Publ.R. Soc. Chem., 2001, 266, 242. 118 M. Valkenberg, C. de Castro and W. Ho¨lderich, Top. Catal., 2001, 14, 139. 119 M. G. White, US Provisional Patent, filed, 2003. 120 J. Davis, Jr, Chemistry Letters, 2004, 33(9), 1072–1077. 121 A. Visser, R. Swatloski, W. Reichert, R. Mayton, S. Sheff, A. Wierzbicki, J. Davis, Jr and R. Rogers, Chem. Comm., 2001, 135, 135–136. 122 J. Davis, Jr and K. Forrester, Tetrahedron Lett., 1999, 40, 1621–1622. 123 Y. Gu, F. Shi and Y. Deng, J. Molecular Cat. A: Chemical, 2004, 212, 71–75. 124 E. Bates, R. Mayton, I. Ntai and J. Davis, Jr, J. Am. Chem. Soc., 2002, 124(6), 926–927. 125 A. Visser, R. Swatloski, W. Reichert, R. Mayton, S. Sheff, 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, 5–9 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. Waffenschmidt 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. Waffenschmidt and P. Wasserscheid, J. Organomet. Chem., 2001, 630, 177. 132 S. Anjaiah, S. Chandrasekhar and R. Greé, Tetrahedron Lett., 2004, 45, 569–571. 133 A. Kamal and G. Chouhan, Tetrahedron Lett., 2005, 46, 1489–1491. 134 J. Gui, Y. Deng, Z. Hu and Z. Sun, Tetrahedron Lett., 2004, 45, 2681–2683. 135 X.-E. Wu, L. Ma, M.-X. Ding and L.-X. Gao, Synlett, 2005, 4, 0607–0610. 136 J. Fraga-Dubreuil and J. Bazureau, Tetrahedron, 2003, 59, 6121–6130. 137 H. Hakkou, J. Eynde, J. Hamelin and J. Bazureau, Synthesis, 2004, 11, 1793–1798. 138 J. Legeay, J. Eynde and J. Bazureau, Tetrahedron, 2005, 61, 12386–12397. 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, 1058–1064. 141 B. Ranu and S. Banerjee, Lett. Org. Chem., 2006, 607–609. Catalysis, 2008, 21, 154–190 | 189 This journal is

c


142 J. Ma, X. Zhou, X. Zang, C. Wang, Z. Wang, J. Li and Q. Li, Aust. J. Chem., 2007, 60, 146–148. 143 L. Magna, Y. Chauvin, G. Niccolai and J.-M. Basset, Organometallics, 2003, 22, 4418–4425. 144 Z. Zhang, Y. Xie, W. Li, S. Hu, J. Song, T. Jiang and B. Han, Angew. Chem. Int. Ed., 2008, 47, 1127–1129. 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, 1133–1138. 147 R. Baker, D. Halliday and T. Smith, J. Organomet. Chem., 1972, 35, C61–C63; R. Baker, Chem. Rev., 1973, 73, 487–530; R. Baker, A. Cook and T. Smith, J. Chem. Soc., Perkin Trans. II, 1974, 1517–1524. 148 J. Dullius, P. Suarez, S. Einloft, R. de Souza, J. Dupont, J. Fischer and A. Cian, Organometallics, 1999, 17, 997–1000. 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, 1025–1028.


c


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 effect are attributed to Sir Humphrey Davy and Russian chemist Gottlieb Kirchhoff whom both observed catalytic effects by 1812. Davy was evidently the first to observe that platinum induced the oxidation of alcohol vapor in air1 and Kirchhoff was the first 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 efforts. The categories can be generally classified as experimental and theoretical and more specifically, the experimental consists of preparation techniques, activity-selectivity testing and kinetic and mechanistic elucidation. The specific 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 significant integration of these categories that has enabled a tremendous understanding of catalytic reactions, cycles and effect. One of the many significant contributions to the catalysis field was the observation that the catalyst structure changes during the reaction that it is affecting. Gabor Somorjai conducted the first experiments that compared surface diffraction 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 specific 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, structureactivity correlations and idealized modeling efforts. 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 different 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


c


chapter will focus on a few of the tools that have significantly advanced the field 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 field and the potential for significant 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 different 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 classified 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 effects, resonance, and inductive effects. SAR was observed and postulated in 19074,5 and has since been extended and formalized by Boudart.6–8 Currently there are two main categories of SAR: primary SAR, caused by geometric effects, and secondary SAR, caused by preferential poisoning of sites of a given geometry and coordination. Moreover, structure sensitive reactions typically occur on large multiple atom sites, which are very sensitive to alloy or poison effects and involve the activation of bonds (i.e. C–C or N–H). A significant 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 Information 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 afforded 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 quantified 192 | Catalysis, 2009, 21, 191–218 This journal is

c


this effect by defining 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 finally 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 field 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 influent to product effluent, 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, 191–218 | 193 This journal is

c


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 reactions. For example, mass and heat transport effects create error in the rate coefficients obtained via conventional differential 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.14–22 Generally, the intermediates cannot be isolated because the gases are in a regime where they continue to react before they can be identified. 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 effects to guiding catalyst design for near 100% selectivity for hydrogenation and dehydrogenation reactions on supported Pd particles, methanol oxidation at vanadium oxide model catalysts and, catalysts with a well defined charge state.25–28 Goodman and co-workers used model catalyst systems to understand the nature of size and shape, i.e. clusters to nanoparticles, on catalytic performance29–31 as well as model supports such as silica.32 This method is a significant 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 efforts. 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 competition with desorption.33 One method that has enabled rapid testing schemes is non-isothermal testing of an integral catalytic reactor called Polythermaltemperature-ramping reactor (PTR).34–36 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 different runs showing varying degrees of catalytic activity.34


c


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 differential 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 significant amount of time. While this technique has significantly advanced the understanding of catalyst reaction engineering, it cannot be applied to highly exothermic reactions and does not identify intermediate chemical species. Pfefferle 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


c


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 effluent. 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 Monsanto38–42 in the mid 1980’s. The first version had the entire system under vacuum conditions and a schematic is shown in Fig. 3. The first 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-defined surfaces to distinguishing between sequential or parallel reaction paths. The capabilities 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 modified, resulting in the TAP-2 reactor system.38 The main difference between the TAP and TAP-2 reactor systems is higher sensitivity and detection efficiency of reactor effluent 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 identified 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


c


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.44–50 In general, gas transport through the reactor test bed proceeds by either viscous flow, Knudsen flow or a combination of the two. Inertial flow effects, however, take effect at the onset of short gas pulses and at the beginning of the pulse’s viscous flow. 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 flux 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 significantly 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 find 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, 191–218 | 197 This journal is

c


dynamics, while providing nanoscale spatial and femto- to milli-second time resolution. Operando methodology aims to define 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 1–2). 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 modification 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


c


product effluent 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 spectroscopy 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 first IR spectrum, vibrational spectroscopy has become an important analytical research tool. This technique was then applied to the analysis of adsorbates on well-defined surfaces, subsequently moving towards heterogeneous reaction studies. Terenin and Kasparov (1940) made the first 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 configuration 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 parameters 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 difference 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 difference in IR spectra from before activation to during reaction. The black lines show the spectra first at the start of the reaction, and the next is at the maximum reaction rate.56


c


time-on-stream and it can be seen, for example, that at wave number 1602 there is a significant 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 different sites and varying degrees of hydration. Within the IR spectroscopy arena, the most frequently used techniques are transmission–absorption, diffuse reflectance, ATR, specular reflectance, and photoacoustic spectroscopy.57 A typical in situ IR system is shown in Fig. 7. Choosing appropriate probe molecules is important because it will influence 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 effect of different surfaces to very specific compounds (e.g. heavy water and deuterated 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 effect 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.57–59


c


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 flow conditions. Diffuse reflectance infrared spectroscopy (DRIFTS) In the IR spectroscopy field, a majority of experiments are performed in the transmission–absorption and the diffuse reflectance 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 diffuse reflection spectra are the result of the light interacting with the sample in every conceivable way, the spectra may exhibit features of transmission, external reflection, and/or internal reflection. 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. Diffuse reflectance is an excellent sampling tool for powdered or crystalline materials in the mid-IR and near-IR spectral ranges.60 Heated reaction chambers for diffuse reflectance allow the study of catalysis and oxidation reactions in situ, and can evaluate the effects 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 benefits 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 effect, 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 flow. Absorbance increases with time. (b) In situ Raman spectra measured while methanol flowing over catalyst.61


c


sample dilution. Reststrahlen bands, are generated when there is nearly 100% reflectivity 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 affect 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 conventional 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 first 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 difficult to obtain. Therefore, most of the reliable data are averages of many measurements 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 filling 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 difficult 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 O–H bond). An external electric field 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 carbon–carbon double bonds. The pi-electron cloud of the double bond is easily distorted in an external electric field. 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, 191–218 This journal is

c


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 different coke formation behavior, the reaction of propene on a zeolite was performed. The spectra obtained were (A) C3H6/He flow at 773 K for 3 h; (B) O2 flow at 773 K for 1 h and (C) O2 flow at 873 K for 1 h. This data shows that most of the carbon, identified as polyaromatic and pregraphite, can be removed at 773 K with oxygen. However there is still carbon present as identified 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 fluidized 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 fluorescence and luminescence can easily interfere with the data. This is particularly acute with surface carbon, which fluoresces 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


c


Fig. 10 Top spectrum: n-heptane/H-ZSM-5 recorded using the in situ fluidized 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 fluoresce 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 modification to the catalyst and produce high quality reliable data, a circulating fluidized bed was implemented to minimize the influence 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 modified to enable investigation of adsorbates on surfaces that were initially thought to be of insufficient sensitivity. Surface Enhanced Raman Scattering (SERS) is used to investigate the vibrational properties of adsorbed molecules. Metal surfaces must be of high reflectivity 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.65–69 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 spectrometer to determine the affects 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, 191–218 This journal is

c


modes were most sensitive in the visible spectra. The vibration modes of bridging V–O–Al and V–O–Fe bonds were observed for bulk AlVO4 and FeVO4 for the first time and demonstrated that use of fluidized 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 significant finding 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.70–74 For example, the oxidative dehydrogenation of ethane and propane was examined via UV–visible and Raman spectra. The study investigated the catalytic properties vanadia formulations that possessed a range of VOx surface species density (1.4–34.2 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 flat and clean, just a slice through a plane in the material’s 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 A˚ngstroms (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 flow 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 environment. Therefore, the information obtained relates to the first 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.76–79 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, significantly 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, 191–218 | 205 This journal is

c


Fig. 11 High pressure, high temperature STM. Sections: A-UHV sample manipulator with indirect heating and cooling, B-variable temperature STM, C-magnetically coupled linearrotary 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 modification 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 different mechanisms occurring at different pressures. For example at a CO pressure of 1.0 Torr induced mobility of Pt atoms to form clusters, indicating significant weakening of the surface Pt–Pt 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 fill 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 refinements and modifications to those techniques will be made that will provide higher fidelity 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, 191–218 This journal is

c


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 significance 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, 191–218 | 207 This journal is

c


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 filled 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 acknowledge 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 operando investigation is the DRIFT-MS apparatus of Burch’s 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 significantly 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 confirmed that the formate species was most likely the main surface intermediate 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 different. Therefore, the reaction mechanism based on data using different reaction mixtures or reaction temperatures, or even differently prepared 208 | Catalysis, 2009, 21, 191–218 This journal is

c


catalysts having the same nominal composition can be very difficult to generalize over a wide range of conditions. Multinuclear MRI has recently been used to analyze preparation techniques and reactions in heterogeneous catalysis.82 Similar to other optical techniques it is non-invasive and nondestructive and can have chemical specificity. It can measure diffusivity, pore size distribution and temperature among other parameters and thus can be used to understand the progress of a reaction affected 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 resolution of 5–10 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 finely resolve the thickness of the catalyst penetration into the particle, resulting in good control of the finished catalyst. This technique has been successfully used to monitor the liquid phase hydrogenation of unsaturated hydrocarbons over Pt/Al2O3 and Pd/Al2O3 catalysts with 2–3 second resolution time. The temperature of the reaction was scanned from 100–250 1C. The observations showed an oscillating regime in the pellet with constant external conditions. It was a clear sign of the coupling effect 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-filled 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 microscopy.83 Fluorescence microscopy is capable of studying diffusion and catalysis in zeolite pores with high 3D spatial and temporal resolution. This technique was successfully used to monitor acid catalyzed transformation of fluorogenic organics on a mordenite zeolite. The fluorescence microscopy technique is basically a far field microscope that can be deployed in two types of configurations. The first is confocal that focuses light to a diffraction 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, 191–218 | 209 This journal is

c


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 configuration is the wide field operation mode where the detector is a CCD camera that enables hundreds of mm2 to be monitored. The spots resolved are still diffraction limited but have a reduced depth dimension, thus lower resolution but much higher excitation volumes covering hundreds of mm2. Using appropriate fitting of the point spread function (PSF), many molecular events can be monitored simultaneously as well as achieving a precision below the diffraction 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.84–90 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 defined regimes in the absence of transport effects. 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 effects.


c


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 defined periods of time. The apparatus’s step change from ambient to desired reaction conditions eliminates transport effects 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 flow reactor for practical performance evaluation and scale up. Moreover, it has capability to conduct temperature- and pressurejump 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 defined 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 reflected 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 reflects 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 diffusion has an influence. Internal mass transfer effects 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 defined.


c


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 diffusion mass transfer resistance needs to be evaluated by calculating the Damko¨hler 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 diffusion 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 effective diffusivity based on the diffusion coefficient (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 effects 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 diffusion in the washcoat.


c


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 shock’s 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 different 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 off 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’’ configurations of the catalyst and the gas phase molecules approaching the surface. The calculations produced thermodynamic energetic and entropic effects as the reaction coordinate changed, bring a reactant closer to the Catalysis, 2009, 21, 191–218 | 213 This journal is

c


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.94–104 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 field by numerical differentiation 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 effect 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 first reduction peak was attributed to monomeric and polymeric vanadium oxide being reduced, which was confirmed 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 effect 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 V–K–O compounds at the highest vanadium and potassium loading on titania where the potassium modified surface developed a Raman band near 800 cm 1, which was in agreement with the DFT calculations. 214 | Catalysis, 2009, 21, 191–218 This journal is

c


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 nanostructures, 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 significant area of progress will be modification 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 first 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, 137–144. 4 J. Boeseken, Recl. Trav. Chim. Pays-Bas Belg., 1907, 26, 285–288. 5 J. Boeseken, Recl. Trav. Chim. Pays-Bas Belg., 1911, 29, 330–339. 6 M. Boudart, Actas Simp. Iberoam. Catal., 1984, 1, 3–18. 7 M. Boudart and G. Djega-Mariadasson, Kinetics of Heterogeneous Catalytic Reactions, 1984. 8 M. Boudart and D. G. Loffler, J. Phys. Chem., 1984, 88, 5763. 9 G. C. Bond, Acc. Chem. Res., 1993, 26, 490–495. 10 H. S. Taylor, Proc. R. Soc. London. Ser. A, 1925, 108, 105–111. 11 D. Farin and D. Avnir, J. Am. Chem. Soc., 1988, 110, 2039–2045. 12 J. J. Carberry, J. Catal., 1987, 107, 248–253. 13 R. D. Cortright and J. A. Dumesic, Adv. Catal., 2001, 46, 161–264. 14 M. J. Castaldi, A. M. Vincitore and S. M. Senkan, Combust. Sci. Technol., 1995, 107, 1–19. 15 M. J. Castaldi and S. M. Senkan, Combust. Sci. Technol., 1996, 107, 141–150. 16 M. J. Castaldi and S. M. Senkan, Ullman’s 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, 693–702. 18 M. J. Castaldi and S. M. Senkan, Combust. Sci. Technol., 1996, 116–117, 167–181. 19 M. J. Castaldi and S. M. Senkan, J. Air Waste Manage. Assoc., 1998, 48, 77–81. 20 N. M. Marinov, M. J. Castaldi, C. F. Melius and W. Tsang, Combust. Sci. Technol., 1998, 131, 295–342. Catalysis, 2009, 21, 191–218 | 215 This journal is

c


21 N. M. Marinov, W. J. Pitz, C. K. Westbrook, M. J. Castaldi and S. M. Senkan, Combust. Sci. Technol., 1996, 116–117, 211–287. 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, 192–213. 23 J. V. Lauritsen, R. T. Vang and F. Besenbacher, Catal. Today, 2006, 111, 34–43. 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, 107–08, 12–22. 25 H. J. Freund, Surf. Sci., 2007, 601, 1438–1442. 26 H. J. Freund, Top. Catal., 2008, 48, 137–144. 27 H. J. Freund, M. Baumer, J. Libuda, T. Risse, G. Rupprechter and S. Shaikhutdinov, J. Catal., 2003, 216, 223–235. 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, 291–293. 30 D. W. Goodman, Chem. Rev., 1995, 95, 523–536. 31 D. W. Goodman, J. Catal., 2003, 216, 213–222. 32 B. K. Min, A. K. Santra and D. W. Goodman, Catal. Today, 2003, 85, 113–124. 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, 209–226. 35 N. M. Rice and B. W. Wojciechowski, Catal. Today, 1997, 36, 191–207. 36 B. W. Wojciechowski, Catal. Today, 1997, 36, 167–190. 37 M. Lyubovsky and L. Pfefferle, Catal. Today, 1999, 47, 29–44. 38 J. T. Gleaves, G. S. Yablonskii, P. Phanawadee and Y. Schuurman, Appl. Catal., A, 1997, 160, 55–88. 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, 202–218. 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, 1878–1891. 45 P. Phanawadee, M. Phongaksorn, N. Chaimongkol, A. Jaree and J. Limtrakul, Chem. Eng. J. (Amsterdam, Neth.), 2005, 115, 51–62. 46 S. O. Shekhtman, G. S. Yablonsky, J. T. Gleaves and R. R. Fushimi, Chem. Eng. Sci., 2004, 59, 5493–5500. 47 D. Wang, J. Chin. Chem. Soc. (Taipei, Taiwan), 2003, 50, 551–555. 48 J. A. Delgado, T. A. Nijhuis, F. Kapteijn and J. A. Moulijn, Chem. Eng. Sci., 2002, 57, 1835–1847. 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, 1913–1923. 51 S. O. Shekhtman, G. S. Yablonsky, S. Chen and J. T. Gleaves, Chem. Eng. Sci., 1999, 54, 4371–4378. 52 O. M. F. Romero Sarria, P. Bazin, J. Saussey, J. Lesage, A. Guesdon and M. Daturi, Catal. Today, 2006, 113, 87–93. 53 R. P. Eischens, S. A. Francis and W. A. Pliskin, J. Phys. Chem., 1956, 60, 194–201. 54 J. Ryczkowski, Catal. Today, 2001, 68, 263–381. 216 | Catalysis, 2009, 21, 191–218 This journal is

c


55 N.-Y. Topsoe, Catal. Today, 2006, 113, 58–64. 56 F. C. J. B. S. KLose, P. Joshi, A. Trunschke, R. Schlog, I. R. Subbotina and V. B. Kazansky, Catal. Today, 2006, 116, 121–131. 57 J. A. Lercher, C. Grundling and G. EderMirth, Catal. Today, 1996, 27, 353–376. 58 J. A. Lercher, V. Veefkind and K. Fajerwerg, Vib. Spectrosc., 1999, 19, 107–121. 59 G. Mirth, F. Eder and J. A. Lercher, Appl. Spectrosc., 1994, 48, 194–197. 60 M. P. Fuller and P. R. Griffiths, Anal. Chem., 1978, 50, 1906–1910. 61 R. F. Howe, in In situ Spectroscopy in Heterogeneous Catalysis, ed. J. D. Haw, Wiley-VCH, Los Angeles, 2002, pp. 139–177. 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, 353–360. 64 Y. T. Chua and P. C. Stair, J. Catal., 2000, 196, 66–72. 65 T. V. M. Rao, G. Deo, J. M. Jehng and I. E. Wachs, Langmuir, 2004, 20, 7159–7165. 66 C. Resini, T. Montanari, G. Busca, J. M. Jehng and I. E. Wachs, Catal. Today, 2005, 99, 105–114. 67 H. J. Tian, I. E. Wachs and L. E. Briand, J. Phys. Chem. B, 2005, 109, 23491–23499. 68 I. E. Wachs, Catal. Today, 2005, 100, 79–94. 69 I. E. Wachs, Y. Chen, J. M. Jehng, L. E. Briand and T. Tanaka, Catal. Today, 2003, 78, 13–24. 70 S. W. Yang, E. Iglesia and A. T. Bell, J. Phys. Chem. B, 2006, 110, 2732–2739. 71 M. D. Argyle, K. D. Chen, E. Iglesia and A. T. Bell, J. Phys. Chem. B, 2005, 109, 2414–2420. 72 M. D. Argyle, K. D. Chen, C. Resini, C. Krebs, A. T. Bell and E. Iglesia, Chem. Commun., 2003, 2082–2083. 73 K. D. Chen, A. T. Bell and E. Iglesia, J. Catal., 2002, 209, 35–42. 74 K. D. Chen, A. T. Bell and E. Iglesia, J. Phys. Chem. B, 2000, 104, 1292–1299. 75 J. A. Jensen, K. B. Rider, Y. Chen, M. Salmeron and G. A. Somorjai, J. Vac. Sci. Technol. B, 1999, 17, 1080–1084. 76 R. Reichelt, S. Guenther, M. Roessler, J. Wintterlin, B. Kubias, B. Jakobi and R. Schloegl, Phys. Chem. Chem. Phys., 2007, 9, 3590–3599. 77 M. Montano, D. C. Tang and G. A. Somorjai, Catal. Lett., 2006, 107, 131–141. 78 K. B. Rider, 2001. 79 G. A. Somorjai, Appl. Surf. Sci., 1997, 121/122, 1–19. 80 M. A. Newton, A. J. Dent, S. G. Fiddy, B. Jyoti and J. Evans, Catal. Today, 2007, 126, 64–72. 81 F. C. Meunier, D. Tibiletti, A. Goguet, S. Shekhtman, C. Hardacre and R. Burch, Catal. Today, 2007, 126, 143–147. 82 I. V. Koptyug, A. A. Lysova, R. Z. Sagdeev and V. N. Parmon, Catal. Today, 2007, 126, 37–43. 83 M. B. J. Roeffaers, J. Hofkens, G. De Cremer, F. C. De Schryver, P. A. Jacobs, D. E. De Vos and B. F. Sels, Catal. Today, 2007, 126, 44–53. 84 M. J. Castaldi, US Pat., 6,746,657. 85 M. J. Castaldi, M. Lyubovsky, R. LaPierre, W. C. Pfefferle and S. Roychoudhury, SAE Technical Paper, 2003. 86 M. J. Castaldi, S. Roychoudhury, R. S. Boorse, H. Karim, R. LaPierre and W. C. Pfefferle, 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. Pfefferle, NSF National Meeting, January 2002. Catalysis, 2009, 21, 191–218 | 217 This journal is

c


88 W. C. Pfefferle, 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 7–11, 2002, 2002, CATL-018. 89 L. L. Smith, A. Karim, A. Hasan, A. Castaldi, J. Marco, S. Etemad, W. C. Pfefferle, V. K. Khanna and K. O. Smith, ASME Turbo Expo, Power, Land Sea, & Air, 2003. 90 M. J. Castaldi, R. LaPierre, M. Lyubovski, W. Pfefferle and S. Roychoudhury, Catal. Today, 2005, 99, 339–346. 91 H. S. Fogler, Elements of Chemical Reaction Engineering, Prentice Hall, Englewood Cliff, NJ, ch. 6. 92 F. H. Ribeiro, M. Chow and R. A. Dalla Betta, J. Catal., 1994, 146, 537–544. 93 K. Ramanathan, V. Balakotaiah and D. H. West, Chem. Eng. Sci., 2003, 58, 1381–1405. 94 Application: EP Pat., 2006-1167571876548, 2008. 95 X. Wang and X. Xu, J. Phys. Chem. C, 2007, 111, 16974–16981. 96 X. Lin, C. Zhao, C.-M. Che, Z. Ke and D. L. Phillips, Chem.–Asian J., 2007, 2, 1101–1108. 97 S. Ostrowski, J. C. Dobrowolski and J. Kijenski, Catal. Commun., 2007, 8, 1354–1360. 98 S. A. French, Platinum Met. Rev., 2007, 51, 54–62. 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, 34–41. 101 M. Okumura, Y. Irie, Y. Kitagawa, T. Fujitani, Y. Maeda, T. Kasai and K. Yamaguchi, Catal. Today, 2006, 111, 311–315. 102 M. B. Zellner, A. M. Goda, O. Skoplyak, M. A. Barteau and J. G. Chen, Surf. Sci., 2005, 583, 281–296. 103 R. Tokarz-Sobieraj, M. Witko and R. Grybos, Catal. Today, 2005, 99, 241–253. 104 M. Witko and R. Tokarz-Sobieraj, Catal. Today, 2004, 91–92, 171–176. 105 H. Si-Ahmed, M. Calatayud, C. Minot, E. L. Diz, A. E. Lewandowska and M. A. Banãres, Catal. Today, 2007, 126, 96–102. 106 C. J. Cramer, Essentials of Computational Chemistry: Theories and Models, John Wiley & Sons, New York, 2004.


c


Catalysis

Catalysis

Environmental Catalysis

Environmental Catalysis

Iridium Catalysis

Industrial Catalysis

Micellar Catalysis

Dendrimer Catalysis

Environmental Catalysis

Supramolecular Catalysis

Dendrimer Catalysis

ENVIRONMENTAL CATALYSIS

Dendrimer Catalysis

Bioinorganic Catalysis

Catalysis of organic reactions

Green Chemistry and Catalysis

Advances in Catalysis

Asymmetric Phase Transfer Catalysis

Heterogeneous Catalysis Research Progress

Electrochemical activation of catalysis

Catalysis of Organic Reactions

Catalysis of Organic Reactions

Surface and nanomolecular catalysis

Green Chemistry and Catalysis

Trends in Catalysis Research

Heterogeneous Catalysis Research Progress

Theoretical Heterogeneous Catalysis

Catalysis

Catalysis

Environmental Catalysis

Environmental Catalysis

Iridium Catalysis

Industrial Catalysis

Micellar Catalysis

Dendrimer Catalysis

Environmental Catalysis

Supramolecular Catalysis

Dendrimer Catalysis

ENVIRONMENTAL CATALYSIS

Dendrimer Catalysis

Bioinorganic Catalysis

Catalysis of organic reactions

Green Chemistry and Catalysis

Advances in Catalysis

Asymmetric Phase Transfer Catalysis

Heterogeneous Catalysis Research Progress

Electrochemical activation of catalysis

Catalysis of Organic Reactions

Catalysis of Organic Reactions

Surface and nanomolecular catalysis

Green Chemistry and Catalysis

Trends in Catalysis Research

Heterogeneous Catalysis Research Progress

Theoretical Heterogeneous Catalysis

Recommend Documents