Environmentally Friendly Catalytic Methods James H. Clark and Duncan J. Macquarrie Department of Chemistry University of York Heslington York UK YO1 5DD 1 Introduction Increasingly demanding environmental legislation public and cor- porate pressure and the resulting drive towards clean technology in the chemical industry with the emphasis on reduction of waste at source will require a level of innovation and new technology that the industry has not seen in many years.I Established chemical processes that are often based on technology developed in the first half of the 20th century may no longer be acceptable in these envi- ronmentally conscious days. ‘Enviro-economics’ will become the driving force behind new products and processes. The cost of clean- ing up chemical processes and plants and adopting the best envi- ronmental option will be high and could even exceed current R&D expenditure within the European Union. This level of expenditure brings with it an unprecedented opportunity for applied research aimed at developing new and more environmentally friendly chem- ical processes and for the introduction of new technology. Catalysts played a major role in establishing the economic strength of the chemical industry in the first half of the 20th century and the clean technology revolution in the industry will provide new opportunities for catalysis and catalytic processes. While the overwhelming majority of chemical processes introduced in the last SO years depend on catalysis the market growth potential for catalysis is still considerable and especially in the fine and special- ity chemicals industries where catalysts are relatively rarely used or where homogeneous catalysts are difficult to separate and require additional processing stages. Some of the major goals of clean technology in the chemical industry are to increase process selectivity to maximise the use of starting materials (aiming for 100% atom efficiency) to replace stoichiometric reagents with cat-alysts and to facilitate easy separation of the final reaction mixture including the efficient recovery (and hopefully reuse) of the cata- lyst. The use of solid mostly inorganic catalysts often based on common porous support materials promises to go a long way towards achieving these goals in many important chemical processes where current technology is very inefficient or leads to unacceptable levels of waste. Their use is also a good example of ‘heterogenisation’ whereby the inorganic reagents or catalysts are segregated from the liquid phase facil itating their separation recovery and reuse. Supported reagents based on inorganic materials have been known for almost 30 years and the development of the subject is apparant from the steady increase in the number of research articles James Clark obtained his BSc and PhD degrees from King5 College London. Following postdoctoral research appointments in Canada and England he joined the academic stafS at York in 1979. He was appointed to the Chair in Industrial and Applied Chemistry at York in 1994. His research interests cover Clean Synthesis Supported Reagents Materials Chemistry and Fluorine Chemistry. He has written or edited four books in these areas and his research has led to awards from the SCI RSC RSA and the EU. He cur- rently holds a Royal Academy of EngineeringlEPSRC Clean Technology Fellowship. the appearance of several books ,2-the first international symposia and the first industrial applications.6 Useful inorganic support mate- rials generally have a large surface area (typically > 100 m2 gg1) and are often porous. They include zeolites although their micro- porosity may make them less amenable to liquid phase processes commonly used in fine and speciality chemicals manufacturing than vapour phase processes (where they are well established) due to poor diffusion rates and pore blockage by larger molecules. Silica gels aluminas and clays are often used as supports and newer meso- porous inorganic supports such as the MCMs7 are likely to become increasingly important in this context. This review will focus on supported reagent type catalysts based on porous inorganic support materials and liquid phase organic reactions in which they are used. 2 Supported Reagents The original principle behind the use of supported reagents was to achieve an increase in the effective surface area and hence activity of potentially useful but insoluble inorganic reagents via their dis- persion over high surface area inert support materials. The majority of supported reagents reported since then have been stoichiomet- ric21.8in their chemistry so that at best only the support material could be recovered and reused. Increasingly the emphasis has been shifted towards truly catalytic materials which can involve dis- persed acid or basic sites or other reagents that are catalytic in their action or have been rendered catalytic by interaction or reaction with the support. Thus Bronsted and Lewis acids can be supported (eg.silica-H,PO clay-ZnCI,) as can bases (e.g.alumina-NaOH alumina-KF) but it is also possible to prepare other catalytic sup- ported reagents (e.g. supported phase-transfer catalysts) and to prepare catalytic forms of species that are normally stoichiometric in their chemistry (e.g.alumina-Crv’).5 The factors likely to be considered in preparing a supported reagent are (i)Choice of support material. (ii) Support pretreatment (including drying to remove loosely bonded water and treatment with aqueous HCI to maximise the cov- erage of surface OH groups). Duncan Macquarrie received his BSc and PhD degrees from the University of Strathclyde Glasgow. In 1985 he moved to the University of York as a Postdoctoral Fellow under the supervision of Prof. James Clark researching into the phase transfer catalvsis of nucleophilic juorination reactions. He then moved to Contract Chemicals where he was a member of the team which developed the range of Envirocat cata-lvsts. After spells in research at Ciba Geigy (Traflord Park) arid R+ D at Lonza (Switzerland) lie returned to York in October 1995 to take up a position as Royal Socie8 UniverJitv Research Fellow where hir current reseurch interests are in novel heterogeneous cutul y.5 ts via chemical surface modijca- tion. He is co-author of a hook on supported reagents und i the author of several paperr and patents. 303 Table 1 Methods of preparing supported reagents Method Comments linpregnation (evuporation) -filling Widely used the pores of a support with a solution Good control over dispersion of the reagent followed by and loading evaporation of the solvent Requires an appropriate solvent Precrpitatronlcoprecipitution-of the Valuable for poorly soluble reagent on to the support reagents Difficult to control AdJorption from solutiori -selective Easy to carry out removal of the reagent from the May be inefficient solution Intiinate solids mixing Simple to carry out Avoids the use of any other chemical Unlikely to be efficient loti e rchatige Simple and effective for materials with exchangeable ions (largely zeolites and clays) Sol-gel techniques -starting from a Does not require a pre formed functionalised silane monomer support Silane monomer must be synthesised Difficult to control material structure Srl~latronof the support Utilises the known structure of a pre formed support Silane monomer must be synthesised Resulting surface groups can be easily lost Chlorinution and denvatisation Efficient surface chlorination of the upp port is quite easy Six1 groups are highly reactive Further reaction may require an organometalI ic (m)Reagent loading (often assumed to be at best at monolayer coverage which enables the amount of reagent to be calculated but physisorbed reagents will not be completely dispersed and the loading of chemisorbed functions is in practice variable there are also examples where unexpected high loadings or low loadings can apparantly be beneficial) (w)Method of preparation (see Table 1) (v) Supported reagent post-treatment (e g calcination to fix the reagent) Generally the aim is to achieve maximum dispersion so as to achieve maximum activity per unit area of support Other factors may also prove to be important such as ensuring that the reagent is fixed to the support and will not be removed under the conditions of any reaction or separation For industrial applications long-term catalyst stability may be more important than initial activity if that early activity is followed by rapid decay The successful application of supported reagent catalysts requires a reasonable understanding of the bulk and surface struc- tures of the material Stability surface area porosity and the dis- persion and nature of the active sites are all important factors that can critically affect the catalytic value of the supported reagent Fortunately there are numerous spectroscopic and non-spectro- SCOPJC techniques available for the study of solids and supported surfaces The application of supported reagents can be hindered or limited by instability due to (I) Thermal decomposition of active sites (most common sup ports are stable up to very high temperatures although a few less CHEMICAL SOCIETY REVIEWS 19% commonly used materials such as some pillared clays may break down at moderate temperatures) (10Reaction of active sites with the atmosphere (solid bases for example can rapid1 y adsorb C0,-forming carbonates) (ill) Removal of active sites in the course of a rewtion (a common problem which causes contamination of the organic mixture and prevents reuse of the catalyst) Thermal analysis techniques (TGA DSC DTA) can be used to study the thermal stability of supported reagents which generally show a low temperature loss of loosely held water (and other solvent molecules) a gradual loss of chemisorbed water at higher temperatures (from surface hydioxyls) and the thermal decompo- sition of catalytic sites Chemisorbed organic functions will nor-mally decompose at ca 300 "C or higher which is adequate for most liquid phase reactions Low temperature thermal events from changes to inorganic reagents are less common but can be impor tant leg the decomposition of supported Fe(NO,) the reaction between CuO and the charcoal support or phase changes in sup-ported ZnCI,] The thermal behaviour of the superficially simple supported fluorides such as the widely used solid base KF-alumina is particularly complex and very important (Figure 1) OH- -F(H20)K+ HEAT OH- -FK+ OH-F(H20)K+ -H20 OH-FK+ Figure 1 Changes in the nature of KF-alumina on heating While dispersion of the reagenuactive sites over the surface of the support material is often only one of several factors affecting the performance of a supported reagent it remains a very important one Generally high surface areas are preferred although the choice of support material ISunlikely to depend on that alone Surface areas of common supports are in the 100-1000m2 g range with the new MCM materials having particularly high values (even to > 1000 m2 g I) Most commonly used support materials are either micro- porous (pore diameters 3-20 A)or mesoporous (20-500 A)with mesoporosity likely to give a reasonable balance between good dif- fusion rates and useful in-pore effects (e g high local concentra- tions of reagent sites to enhance reaction) lo Zeolites are the best known microporous solids although swelling clays also commonly have interlamellar spacings of less than 10 8 Pillared clays and non-aluminosilicate molecular sieves (1 e porous solids with regular structures) have pore diameters in the 10-20 A range Acid treated clays can contain mesopores as a result of the break up and reorganisation of the aluminosilicate structure (dealumination) I I Until recently mesoporous solids were restricted to amorphous materials with broad pore size distribution (silicas and aluminas) In 1992 the first truly mesoporous molecular sieves were reported These 'MCM' materials opened the way to the synthesis of ordered porous solids with high surface areas and tunable pore diameters bridging the zeolites and the common amorphous silicas Little is known about the use of these materials as supports but the potential IS considerable Reactive reagents notably hydroxides and fluorides will corrode the surfaces of common support materials so that the actual surface species present may be more complex than might be expected (see Figure 1 for example) and surface areas can be much reduced Even the simplest methods of supported reagent preparation can give good dispersion of most reagents as witnessed by techniques such as electron microscopy X-ray diffraction and where the reagent is amenable diffuse reflectance Fourier transform infra red (FTIR) spectroscopy ENVIRONMENTALLY FRIENDLY CATALYTIC METHODS-J H CLARK AND D J MACQUARRIE 305 One of the most interesting features of supported reagent chem- istry which has become particularly significant in recent years is that the activity of the composite material (ASR) is not usually a simple sum of the component parts (A +AR) The value of this in catalysis is when A >A +A This increase in activity is a result of a synergistic effect between the reagent and the support The exact nature of this effect when it occurs is variable In extreme cases it can be due to actual reaction between the support and the reagent such as in the case of some sup- ported fluorides In other cases it is more subtle with an increase in the number of available sites and/or the local in-pore concentration of sites (mini reaction vessels) being responsible such as is the case for supported zinc chloride lo The nature of an adsorbed reagent can be investigated by numer- ous techniques Diffuse reflectance FTIR and magic angle spinning (MAS) NMR spectroscopies are especially popular The useful information that can be obtained by such studies incI udes (I) Identification of the surface species (eg to confirm that the original reagent is intact or to identify new species formed via support-reagen t reaction) (II) Information on the strength of the interaction between the support and the reagent ([if) Identification of sites including Lewis and Brgnsted acid sites and determination of their relative strength (eg through the use of spectroscopic titration techniques such as those based on the use of pyridine") (~v)Determination of any changes to the bulk support structure as a result of corrosion of the support by attack of the reagent Through the use of spectroscopic and non-spectroscopic tech- niques it is possible not only to understand the behaviour of sup- ported reagent catalysts but also to help optimise their performance in organic reactions 3 Catalysis using Supported Reagent Solid Acids*~~,~ Solid acids are the most widely studied and commonly used het- erogeneous catalysts They are used in many important large-scale vapour phase manufacturing processes such as catalytic cracking (X and Y zeolites) alkylation (zeolites Si0,-H,PO,) and paraffin iso- merisations (chlorinated Pt-AI,O,) with the scale of larger reac- tions exceeding lo9kg per year All solid acids are characterised by the presence of surface protons or coordinatively unsaturated cationic centres which give Brcbnsted or Lewis acidity The number of acid sites gives the total surface acidity while their structure determines the acid strength Lewis and Brcbnsted acid sites are often present together and may be sufficiently strong to justify the term 'solid superacids' (H <-12) which can enable the catalysis of demanding reactions such as transformations of alkanes Suitable solid acid catalysts for reactions under more moderate conditions than typical petrochemical type reactions and in the liquid phase include those based on clays silicas and zeolites In many cases they are sought as replacements for inexpensive liquid or soluble acids such as H,SO HF and AlCl and relatively low cost solids are required The environmental unacceptability of using large volumes of corrosive acids and of the waste resulting from the work-up of such reactions which usually requires the neutralisa- tion/decomposition of the acid is a strong driving force for devel- oping environmentally friendly processes based on solid acids The range of acid-catalysed liquid phase reactions is enormous and includes Friedel-Crafts reactions halogenations and nitrations with relevance to almost all sectors of the fine speciality and inter- mediates chemical industries Friedel-Crafts reactions probably represent the most important range of 'named' reactions in organic chemistry They include acy- lations benzoylations alkylations and sulfonylations giving an enormous range of useful products including ketones alcohols alkylaromatics and sulfones Many batch processes operating in a very large number of companies use AlCI as the soluble acid cata- lyst The reagent is inexpensive and very reactive being one of the most powerful Lewis acids Unfortunately it is difficult to handle being readily hydrolysed by water (and therefore unstable to the atmosphere) giving health and safety and storage problems The work-up reactions using AICI present further problems with the usual water quench creating an acidic aluminium-rich waste stream The problem is particularly acute with reactions involving products that are capable of acting as Lewis bases such as ketones which complex the AlCl In these cases at least stoichiometric quantities of AlCI are required indeed in the case of sulfone- forming reactions up to three mole equivalents are used to ensure good conversion The quantity of waste generated in these reactions greatly exceeds the amount of product' Other problems with AICI 'catalysed' reactions include lack of selectivity with polyalkylation being a particular problem Alternative homogeneous reagents fare little better Hydrogen fluoride has useful activity but presents its own special hazards due to its extremely corrosive nature Solid acid supported reagents based on inexpensive inorganic solids notably clays and silica show promise as Friedel-Crafts cat- alysts as do some modified zeolites Many clays have been investi- gated including bentonites vermiculites halloysites and kaolinites Ion-exchanged clays can show much improved activity over the raw materials and especially those based on acid treated clays such as the commercial material KIO (which can itself be a useful solid Bronsted acid) The activity of ion-exchanged clays is very depen- dent on the cation so that in the benzylation of benzene using benzyl chloride for example (I) the order is Fellr >Zn" >Cull >ZrIV> PhH +PhCH,CI -PhCHzPh (1) TiiVTaV>All1'>Coil >K10 >NbV The order does not corre- spond to the order of Lewis acidities for homogeneous cations with the low activity for All1' (very active in solution) and the high activ- ity for Zn" (weakly active in solution) being especially noteworthy It is likely that at least part of the activity is due to the polarisation of water molecules by the cations within the highly polarising environ- ment of the interlamellar regions of the clays At best ion exchanged clays can show activities in reaction (1) 20 times greater than the simple acid-treated clay Selectivity to the monoalkylated product was little improved by ion exchange with only 57% isolated diphenylmethane from the fastest reaction and never better than 66% Alcohols and alkenes can also be used as alkylating agents but the rates of reaction are significantly reduced and the orders of activ- ity of the clay are not the same although the Ti1" exchanged clay in particular is active with all of the alkylating agents It is also impor- tant to note that these catalysts are reusable at least to some extent Perhaps the greatest breakthrough in the use of solid acid cata- lysts in Friedel-Crafts reactions came with the discovery that sup- ported (as opposed to ion exchanged) zinc chloride on K 10 was an extremely active and reusable catalyst for benylation reactions such as (1) The activity is several orders of magnitude greater than that of the ion-exchanged materials with the model reaction being com- plete in minutes at room temperature and giving a particularly high yield of diphenylmethane (80%) Like the ion-exchanged materials. the activity of the K10 supported reagents does not correlate with solution phase activities of the reagents so that supported ZnCI and CuCI are especially active while supported AlCl is a poor catalyst 'Clayzic' has been the subject of intense research since its dis covery in 1989 and also forms the basis of an industrial catalyst It is a particularly striking example of a supported reagent that is con siderably more active than its constituent parts and this is in part explained by a structural change to the support Acid treatment of the montmorillonite clay causes a breakdown in the lamellar struc- ture and the creation of mesopores which are occupied by the ZnCI (Figure 2) lo Thermal activation of clayzic results in further struc- tural changes but the catalytic activity of the material does not cor- relate with its surface areal4 -a good example of a supported reagent catalyst where surface area is not the most important factor Spectroscopic titration of the active sites on clayzic reveal the pres- ence of weak Lewis acid sites but little Bronsted acidity (KlO itself shows BrGnsted acid sites only).1° indeed mesoporous silica sup- ported ZnCI which can be as active as clayzic can be a pure solid Lewis acid The remarkable activity of supported ZnC1 cannot be M 00 oo om a0 00 M"+ M"+ b 0. 0. ee .a@ M"+ M"+ c 0. 00 a0 00 "+M M"+ M "+ severe acid ZnCl2 treatment1 Figure 2 Structural representation of the formation of clayzic due to strong acid sites -rather it is likely to be due to a high con- centration of sites within the constrained in-pore environments Clayzic has been used to catal yse various Friedel-Crafts reac-tions including those of aromatic substrates with alkyl halides alde- hydes and alcohols In the commercial manufacture of diphenylmethanes for example product yields of over 75% can be achieved by using this catalyst which is also easily recovered and can be reused Use of homogeneous AICI leads to product yields of less than 50% and the work-up procedure is difficult and destroys the AICI Other applications include the preparation of benzothio- phenes by cyclisations of phenylthioacetals (normal catalysts can cause extensive polymerisation of the thiophenes and the pores in clayzic are believed to favour the desired intramolecular cyclisation at the expense of the polymerisation -Figure 3)," and the olefina- tion of benzaldehyde (involving a previously unknown reaction mechanism -Figure 4) l6 Interestingly when two or more possible substrates are present ClayxicII I Figure 3 Formation of benzothiophenes using clayzic CHEMICAL SOCIETY REVIEWS 1996 *Y EPZ10 MeN02 room temp Figure 4 Olefination of benzaldehyde using clayzic (in its commercial form ' EPZl 0') reaction occurs first with the more polar substrate whatever its normal relative reactivity Thus with a mixture of an alcohol and an alkyl halide the alcohol will always react first although the alkyl halide is more reactive l7The discrimination can be more subtle so that for haloaromatics the order of activity in clayzic-catalysed alkylations is PhBr > PhCI > PhF PhH which parallels the polar- isibility of the substrate although it is the exact opposite of the activ- ities in solution phase reactions These observations are consistent with the presence of highly polar pores in the catalyst and with mol- ecular sieving on the basis of molecular polarity/polarisibil~ty The widespread use of AICI in Friedel-Crafts and other acid- catalysed reactions along with its environmental unacceptability makes it an obvious target for heterogenisation via a supported reagent Early attempts to prepare an active form of supported alu- minium chloride had limited success with promising results in vapour phase processes such as long chain alkane isomerisations and hydrocarbon cracking reactions but poor activity in liquid phase reactions l9 More recently a new form of the supported reagent has been prepared and found to be highly active in some liquid phase Friedel-Crafts reactions 2o Mesoporous silica and acid- treated montmorillonite supported aluminium chloride (prepared by reaction of the support with either AlCI or RAICl in an aromatic solvent) is particularly effective in catalysing the reactions of alkenes with aromatics such as that of oct-l-ene with benzene eqn (2) -remarkably the activity of the solid acid is comparable to PhH + CH,(CH,),CH=CH -Ph(CH,),CH (2)(complete conversion of the alkene within 2 h at room temp and up to 80% selectivity to the monoalkylate) that of homogeneous AICl but its selectivity towards monoalkyla- tion is significantly superior Greater selectivity is also observed in the dodecene-benzene reaction with the desired 2-alkyl isomer (this isomer gives the best emulsibility characteristics in detergency applications) being produced in 47% yield (32% with homogeneous AICI 13-20% with HF) As with clayzic and other supported reagents the o timum pore size is at the low end of the mesoporous range (ca 70 R) Alkyl halides also react quite well with aromatics in the presence of supported aluminium chloride so that dichloromethane for example reacts with benzene to give diphenyl- methane in a yield of 62% dfter 2 5 h at 40 "C A comparison of sup-ported aluminium chloride with supported zinc chloride two of the more promising solid acids for Friedel-Crafts alkylations would suggest that the former is considerably more active (strong BrGnsted and Lewis acid sites are revealed by spectroscopic titra- tion) but the latter is an extremely useful mild Lewis acid catalyst Supported reagents have been used rather less to catalyse aro- matic halogenations althougth zeolites have been proved to give enhanced para-selectivity in for example the chlorination and bromination of alkylaromatics (up to 75% in the chlorination of toluene using elemental chlorine21 and up to 95% in the bromina- tion of toluene using elemental bromine22) Aromatic nitration a particularly wasteful and hazardous indus- trial process has benefited relatively little from the use of supported reagents with very low conversions usually accompanying any good isomer selectivities (eg using zeolites) or good conversion and selectivity requiring inconvenient reaction times at high dilu- tion The combination of the H+ form of zeolite beta and acetyl nitrite as the nitrating agent can however give good selectivities (e8 79% para for toluene) in fast reactions 23 Numerous other typically acid-catal ysed reactions have been ENVIRONMENTALLY FRIENDLY CATALYTIC METHODS -J H CLARK AND D J MACQLJARRIE successfully carried out in the presence of supported reagent-type solid acids where the solids have replaced conventional acids such as mineral acids AICl BF and FeCI among others Some of the more interesting recent examples include the single-pot synthesis of methyl tert-butyl ether from tert-butyl alcohol and methanol using dodecatungstophosphoric acid supported on clay2 and the use of pillared acid-activated clays for Bronsted catalysed processes such as alkene alkylations and alcohol dehydrations 25 It is also worth noting the early successes emerging from studies on the use of MCM type materials as solid acids'" -this area seems likely to develop quickly 4 Oxidation Catalysis using Supported Reagents The partiaJ oxidation of organic substrates provides routes to a wide range of important functionalised molecules including alcohols aldehydes ketones epoxides and carboxyl ic acids Traditional methods of oxidation often involve the use of stoichiometric quan- tities or large excesses of poisonous high oxidation state chromium manganese and osmium reagents Environmental and economic factors make the use of these reagents increasingly unacceptable Oxidation processes based on lower oxidation state transition metals such as Coil Mnii and Cult in acetic acid media are also known and some are catalytic in the metal using molecular oxygen as the consumable oxidant but the conditions are often harsh the reagent mixture is corrosive (bromide is used as a promotor) and the chemistry is rarely selective Environmentally acceptable cat- alytic partial oxidations of inexpensive substrates (including hydro- carbons) that operate under moderate conditions in the liquid phase (most suitable for many of the industrial beneficiaries) with a high degree of selectivity are clearly desirable A large number of supported reagents have been used In the liquid phase partial oxidation of organic substrates The low cost of the support (commonly chromatographic materials such as silica gel) the mesoporosity of many of these supports and the other general advantages of supported reagents (ease of handling use and recovery low toxicity and the avoidance of solvents) make them very attractive in the context of clean synthesis However in oxi-dations supported reagents have generally acted as stoichiometric reagents being effectively dispersed forms of traditional oxidants There are notable exceptions to this which promise much for the future of supported reagent oxidation catalysis These are the new molecular sieves which incorporate active metal centres (notably Ti and V) in their structures and chemically modified support materi- als which have active metal centres on their surfaces While established aluminosilicate zeolites have proved popular in high temperature oxidation processes,' -their value in selective oxidations typically carried out in the liquid phase is less There has however been considerable early success in the use of other mole- cular sieves notably the titanium silicates such as TS-1 which are already being used in commercial units 27 The catalysts are syn- theised from typical sol-gel preparations involving tetraethyl- orthotitanate and tetraethylorthosilicate High Si Ti ratios minimise titanium centres with titanium nearest neighbours and maximise activity The pore diameter of TS-1 is only 5 5 A which is very restrictive in terms of accessible substrates and products but despite this it has been successfully used in the hydroxylation of aromat- ics the epoxidation of alkenes in ammoxidation amine oxidation and the oxidation of alcohols and thioethers In the hydroxylation of toluene for example the selectivity to para-cresol is an impres- sive 8 1% (Figure 5) The most important application for TS- 1 to date IS probably the hydroxylation of phenol giving mixtures of hydroquinone and catechol This represents a very clean option giving excellent conversion to product and very little waste The TS-1 process is not only cleaner than the alternatives (avoiding the use of strong acids or soluble transition metal catalysts) it also out- performs them particularly in terms of conversion (because of the much lower amounts of tars which result from side reactions and overoxidation) Catalytically active Si-0-Ti sites can also be formed via treatment of preformed silica with an active source of Ti such as TiF TS-1 lH20* y3$.Q'OH+ @OH OH 81YO 7% 12% Figure 5 Hydroxylation of toluene catalysed by TS 1 Vanadium silicate molecular sieves are capable of selectively oxidising 4-chlorotoluene to 4-chlorobenzaldehyde using hydrogen peroxide as the source of oxygen in acetonitrile solvent (Figure 6) 2x CH3I Cl vs-2f H*02 373 K i MeCNI CHO CH20H CH3I 1 I CI CI CI 66% 21Yo ao? Figure 6 Oxidation of chlorotoluenecatalysed by a vanadium silicate Vanadium has also been incorporated into the structure of MCM materials to give efficient catalysts for the oxidation of large mole- cules such as cyclodecane with hydrogen peroxide 29 Molecular sieves containing structural chromium can also be active in oxida-tions notably those of amines to nitro compounds using ?err-butyl hydroperoxide as the oxidant 30 While original forms of supported reagents involving high oxi-dation state metal oxidants were stoichiometric e g KMn0,-silica and K,Cr,O,-silica genuinely catalytic materials derived from the reaction of dichromate and permanganate with alumina were reported in 1989 3i The materials are prepared from aqueous solu- tion by careful control of pH and other reaction conditions The con- centrations of metal centres at the support surface are very low and while their identity is unknown a surface structure for the supported chromium species based on chemically bonded CrVi has been sug- gested as part of a proposed mechanistic pathway in catalytic oxi- dations (Figure 7) The supported chromium catalyst is active in alkyl benzene oxidations using only air as the source of oxygen and the neat substrate as the reaction medium -the ideal system from an environmental point of view In a typical oxidation that of ethyl- benzene to acetophenone the only other product is water and the catalyst can be used in very small quantities and is reusable The major drawback with the catalyst and the manganese analogue are their low activities -rates of oxidation of alkylaromatics of only II IMeH'CH Me' 'Ph 1-b0 l-l-O\ /O 02 Ct HO' 'H PhCOMe Figure 7 Possible mechanism for oxidations catalysed by supported chromium (VI) 1-2% h are possible at > 100 "C so that prohibitively long reac tion times are required for good overall substrate conversions Recent reports of oxidation catalysts based on chemically modi- fied support materials that can complex metal ions with useful redox properties including cobalt copper and iron may well represent a way forward in this field Such materials can be more robust than simpler metal-support materials and can show activity very similar to their homogeneous analogues Effective catalysts include cobalt immobilised on silica which has been derivatised with carboxylic acid functions (Figure 8) 32 This will catalyse the epoxidation of H' iH201 Figure 8 Formation of a supported cobalt catalyst based on a chemically modified silica alkenes using air and a sacrificial aldehyde Significant features of the catalysis include the high selectivity normally leading to the epoxide product only (diols commonly formed in homogeneous peracid reactions are not observed) and the ability of the catalyst to retain its metal ion even under harsh conditions 5 Catalysis using Supported Reagent Solid base^^-^ In contrast to the areas of heterogeneous oxidation catalysis and solid acid catalysis the use of solid base catalysis in liquid phase reactions has not seen the same level of major breakthroughs This is partly because the negative environmental impact of chemical processes using conventional acids and metal oxidants has attracted CHEMICAL SOCIETY REVIEWS 1996 more attention than those based on such as NaOH Base catalysis is however widely used both on a laboratory and industrial scale and the handling separation treatment and disposal of basic reagents and basic waste can all be troublesome Several different supports have been used for preparing solid bases with alumina based catalysts being the most widely studied in organic synthesis Most common basic reagents have been sup- ported including alkali metals (eg alumina-Na and silica-K) alkali metal hydroxides (eg alumina-KOH) and metal alkoxides (eg xonotlite-KOBu') Zeolite materials that have been exchanged with alkali metal cations and Cs+ in particular can also act as solid bases being particularly useful for higher temperature vapour phase processes Of these some of the more interesting are the immobilised alkali metals which can be prepared in a variety of ways including treatment of the support with a solution of the metal in liquid ammonia The solids are often brightly coloured which is due to the formation of colour centres of one electron donor char- acter Na + [2+] -Na+ + !+I Na + 0 -Na+ + 0' Na + OH (or 20Hj -ONa + H2(or H,Oj These materials have been referred to as solid superbases (estimated H > 37) and they are capable of promoting reactions of hydrocarbons such as the isomerisation of 5 vinylbicy-~1012 2 1 Ihept-2-ene to 5-ethylidenebicyclo(2 2 1 Ihept-2-ene (used as a comonomer in the production of synthetic rubber) Surprisingly perhaps the most widely studied supported reagent solid base in the context of organic synthesis is not based on a con- ventional base Attempts to support ionic fluorides and hence render them more active for nucleophilic fluorinations have been largely unsuccessful but have led to the discovery of remarkably useful solid bases459 Simple metal fluorides such as KF are known to be weak bases but their dispersion over a support is not enough to explain the often powerful basicity exhibited by KF-alumina for example The surface chemistry is in fact quite complex (see Figure 1) and oxide and hydroxide sites are likely to contribute to the basic properties When at their most active sup- ported fluorides are capable of adorbing large quantities of carbon dioxide from the atmosphere which reduces their activity through the formation of surface carbonates These facts help to explain the diversity of claims in the literature over their activity which has been variously described as weakly basic strongly basic and even superbasic 1 Supported fluorides have been used in a wide range of typically base catalysed reactions (Table 2) as well as several stoichiometric reactions KF-alumina is certainly the most widely studied although important variables such as loading and supported reagent post-treatment remain contentious issues The basic catalyst is espe- cially effective in carbon-carbon bond forming Michael reactions (eg reaction 4) and it is interesting to note that some of these have been translated into continuous flow reactions based on fixed cata lyst beds Table 2 Some of the reactions catalysed by KF-alumina Reaction Example Oxidation of alkylaromatics Ph,CH -PhCO Alkylations PhOH + MeOH -z PhOMe Condensations EtCHO + MeNO? -EtCHOHCH2N0 Rearrangements ArCH,CH =CH -ArCH =CHMe Michael reacti on s AcCH=CH + EtNO -Ac(CH2),CHMeN02 Additions CHCl + m O,NC,H,CHO -m O2NC6H,CH(OH)CCI ENVIRONMENTALLY FRIENDLY CATALYTIC METHODS-J H CLARK AND D J MACQUARRIE 309 6 Other Supported Reagent-catalysed Reactions and Future Trends Numerous other types of supported reagents (including those based on organic polymers which are beyond the scope of this review) have been developed and applied to liquid phase organic reactions These include catalysts for Diels-Alder reactions which enable the reaction to be applied beyond the normal electron rich dienes and electron poor dienophiles and in some cases allow the use of water as the solvent (a particularly important goal in clean synthesis) Perhaps the best solid catalyst for these reactions is Fe3+ -mont-morillonite which can result in a dramatic improvement in reaction rates (e g see Figure 9) Conditions Isolated product yield (9%) Fe7+-K10clay,p-ButC,H,0H (IO%),O”C lh 77 200 “C 20 h no catalyst 30 Figure 9 Diels-Alder reaction catalysed by Fe3+ -montmorillonite Apart form solid acid catal ysed halogenations other supported reagents have been used in halogenation reactions These include supported phase transfer catalysts such as alumina-Bu,PBr which will catalyse the halogen scrambling reactions between halo- alkanes eqn (5) The immobilisation of phase transfer catalyst may CH,CI + C,H,Br -CH,CIBr + CH,Br + C,H,CI (5) intuitively seem surprising but the concept of ‘triphase catalysis’33 has in fact been known for some 20 years The insoluble supported catalyst can act very effectively at the organic-aqueous interface and offers the usual advantages of supported reagents notably easy separation and reuse It is often physically difficult (due to very good solubility in most solvents and limited thermal stability) and rarely economic to recover a soluble phase transfer catalyst These facts have restricted their use and added to the waste generated from chemical processes Chemically fixing a catalytic structure such as an organometallic complex to a support material is a simple illustration of ‘hetero- genisation’ whereby the useful properties exhibited by the catalyst in solution are hopefully maintained in the solid base Apart from applications relevant to acid catalysis and oxidation catalysis out- lined earlier the principle has been successfully applied to other important reactions in the liquid phase such as hydrogenation asymmetric reduction and the deprotection of acetals Very sig- nificantly the direct grafting of an organometallic complex onto the inner walls of a mesoporous silica has recently been used to prepare via removal of the organic ligands by calcination a new form of supported Ti4+ This method is superior to other more conventional forms of grafting as it does not lead to undesired titano-oxospecies The catalyst is active in epoxidations with tert-butyl hydroperoxide as the oxidant 34 Supported reagent type catalysts have already proved their value in many organic reactions as environmentally friendly replace- ments for established reagents and catalysts that through their cor- rosive or toxic nature or difficulty in separation and recovery from reactions lead to unacceptable chemical waste The scope of these catalysts in liquid phase reactions is expanding on many fronts and commercial catalysts and industrial processes based on their use are now a reality 35 While serendipitous discoveries will continue to be made our understanding of the problems associated with clean synthesis and of the nature of solid catalysts will enable an increasingly more logical approach to the subject Much remains to be done including improvements in catalyst activity selectivity and stability particularly in areas such as Friedel-Crafts acylation nitration and base catalysis Process engineering aspects of the subject such as separation techniques and the translation of batch- type processes to continuous processes are also very important issues There are many interesting new developments emerging including the availability of controlled pore materials that may enable the right balance between activity and selectivity in liquid phase reactions to be achieved Chemical surface modification also has much to offer and research into ways of achieving high surface coverage and robust structures is very important if the catalytic potential of these materials is to be fully realised The way is clear for a greener future’ Acknowledgements The authors would like to express their sincere thanks to the many researchers at York and elsewhere who have contributed to the clean synthesis programme at York None of that would have been possible without the generous support of our spon- sors of whom Contract Chemicals the EPSRC Clean Technology Unit and the Royal Society (for a University research fellowship to D J M ) deserve special mention 7 References 1 ‘Chemistry of Waste Minimisation,’ ed J H Clark Chapman and Hall London 1995 2 ‘Preparative Chemistry using Supported Reagents,’ ed P Laszlo Academic San Diego 1987 3 ‘Solid Supports and Catalysts in Organic Synthesis,’ ed K Smith Ellis Horwood Chichester 1992 4 J H Clark A P Kybett and D J Macquarrie ‘Supported Reagents Preparation Analysis and Applications.’ VCH New York 1992 5 J H Clark ‘Catalysis of Organic Reactions using Supported Inorganic Reagents,’ VCH New York 1994 6 TW Bastock and J H Clark in ’Speciality Chemicals,’ ed B Pearson Elsevier London. 1992 7 J S Beck J C Vartuli W J Roth M E Leonowicz C T Kresge K D Schmitt C T W Chu D H Olson. E W Sheppard S B McCullen. J B Higgins and J L Schlenker. J Atn Chem Soc 1992,114 10834 8 A W McKillop and K W Young S\nthesis 1979,401 and 48 I 9 T Ando. S J Brown J H Clark D G Cork. T Hanafusa J Ichira J M Miller and M J Robertson J Chem Soc. Perkiti Truiis 2 1986 1133 10 J H C1ark.S R Cullen,S J Barlow andT W Bastock J Chern Snc Perkin Trans 2.1994 I 1 17 11 J P ButrilleandT J P~nnavaia.Cutaf Todab 1992,14. 141 12 12 P Laszlo and A Mathy Heir Chitn Actu ,1987,70,577 13 J H Clark A P Kybett D J Macquarrie S J Barlow and P Landon J Chem Soc Chem Comrnun . 1989,1353 14 D R Brown Geofogica Carpathica Series Cfavs (Brurisiuvu),1994 1 45 15 P D Clark,A Kirk and J G K Yee.J Org Chem . 1995,60.1936 16 H P van Shaik. R J Vjin and F Bickelhaupt Angeu Chem Int Ed Engf 1994,33 1611 17 A Cornelis. C Dony. P Laszlo and K M Nsunda. Tetruhedrorr Lett 1993,2901 and 2903 18 J H Clark,S R Cullen,S J Barlow andT W Bastock J Chern Soc Perkin Truns 1994,41 1 19 R S Drago S C Petrosius and P B Kaufman J Mol Cut 1994 89 317 20 J H Clark. K Martin A J Teadale and S J Barlow J Chein Six Chern Commun . 1995.2037 21 A P Singh and S B Kumar. Appl Cut A General 1995.126,27 22 K Smith and D Bahzad J Chern Soc Chetn Comtnuti . 1996,467 23 K Smith A Musson and G A DeBoon. J Chein Soc Chem Cornmun 1996.469 24 G D Yadav and N Kirthivasan J Chetn Soc Chem Cornmiin 1995 203 25 R Mokaya and W Jones. J Chetn Sac Chetn Cornrnirn 1994,929 26 J Aguad0.D P Serrano,M D RomeroandJ M Escola J Chrm Soc Chem Commun ,1996,725 27 A W Ramaswarmy. S Sivbasanker and P Ratnasamy. Mic roporoict Mater 1994,2.451 28 T SelvamandA P Singh,J Chern Soc Chem Corntnun 1995,883 29 K M Reddy. I Moudrakovski and A Sayari J Chem Soc Chem Cotntnun 1994. 1059 30 B Jayachandran M Sasldharan A Sudalai and T Ravindranathan J Chetn SOC Chetn Cotntnun 1995. 1523 CHEMICAL SOCIETY REVIEWS 1996 3 1 J. H. Clark A. P. Kybett P. Landon D. J. Macquarrie and K. Martin J. 34 T. Maschmeyer F. Rey G. Sankar and J. M. Thomas Nature 1995,378 Chem. SOC.,Chem. Commun . 1989,1355. 159. 32 A.J. Butterworth J. H. Clark,S. J. Barlow and T. W. Bastock UK Patent 35 Envirocats Contract Catalysts Knowsley Industrial Park Prescot Appl. 1996. Merseyside UK L34 9HY. 33 S. L. Regen J. Am. Chrm. SOC. 1975,97,5956.