《有机化学》课程教学资源（文献资料）The Baeyer-Villiger Reaction New Developments toward Greener Procedures.pdf_大学文库

The Baeyer−Villiger Reaction: New Developments toward Greener Procedures G.-J. ten Brink, I. W. C. E. Arends, and R. A. Sheldon* Laboratory for Biocatalysis and Organic Chemistry, Department of Biotechnology, Delft University of Technology, Julianalaan 136, 2628 BL Delft, The Netherlands Received February 10, 2004 Contents 1. Introduction 4105 2. H2O2 and O2 as Green Oxidants 4106 2.1. Hydrogen Peroxide 4106 2.2. Dioxygen 4106 3. Theoretical Considerations 4107 3.1. Mechanism of the Reaction 4107 3.2. Electrophilic Activation of Substrate 4108 3.3. Electrophilic Activation of Intermediate 4108 3.4. Nucleophilic Activation of Intermediate 4109 3.5. Nucleophilic Activation of Hydrogen Peroxide 4109 3.6. Electrophilic Activation of Hydrogen Peroxide 4110 4. Catalytic Reactions 4111 4.1. Homogeneous Catalysts 4111 4.1.1. Oxidation of Aldehydes 4111 4.1.2. Oxidation of Cycloalkanones 4112 4.1.3. Oxidative Ring Contraction 4113 4.1.4. Oxidation of Linear Ketones 4113 4.1.5. Enantioselective Oxidations 4113 4.2. Biocatalysis 4116 4.2.1. Lipases 4116 4.2.2. BVMOs 4116 4.2.3. Enantioselective Reactions 4117 4.2.4. Regioselective Reactions 4118 4.2.5. Chemoselectivity 4118 4.3. Heterogeneous Oxidation 4118 4.3.1. Solid Peracids 4118 4.3.2. Solid Lewis Acid Catalysts 4118 4.3.3. Solid Catalysts for in Situ Formation of Peracids 4120 5. Outlook 4121 6. Abbrevations 4121 7. References 4121 1. Introduction In 1899, Adolf Baeyer and Victor Villiger reported the oxidation of menthone to the corresponding lactone (Figure 1) using a mixture of sodium persulfate and concentrated sulfuric acid (Caro’s acid).1 The persulfuric acid was subsequently replaced by an organic peracid, and the Baeyer-Villiger (BV) reaction became one of the most well-known and widely applied reactions in organic synthesis.2,3 Its success is largely due to its versatility: (i) A variety of carbonyl compounds can be oxidized; that is, ketones are converted into esters, cyclic ketones into lactones, benzaldehydes into phenols, or carboxylic acids and R-diketones into anhydrides. (ii) A large number of functional groups are tolerated. (iii) The regiochemistry is highly predictable with the migratory aptitude being tertiary alkyl > cyclohexyl > secondary alkyl > benzyl > phenyl > primary alkyl > CH3. 4 (iv) The reaction is generally stereoselective; that is, the migrating group retains its configuration. (v) A wide range of oxidants may be used with their activity decreasing in the order: CF3CO3H > monopermaleic acid > monoperphthalic acid > 3,5-dinitroperbenzoic acid > p-nitroperbenzoic acid > m-CPBA ∼ HCO3H > C6H5CO3H > CH3CO3H . H2O2 > t-BuOOH. Although more than a century has gone by since its discovery, the BV reaction is far from being at the end of its development. The standard protocol for a BV oxidation suffers from several disadvantages. The use of an organic peracid results in the formation of one equivalent of the corresponding carboxylic acid salt as waste, which has to be recycled or disposed of (returning it to the manufacturer of the peracid is generally not an option). Moreover, organic peracids are expensive and/or hazardous (because of shock sensitivity), which limits their commercial application. For example, the transport and storage of peracetic acid have been severely curtailed, making its use prohibitive. Consequently, increasing attention has been focused on the in situ generation of organic peracids, via reaction of either the corresponding aldehyde with oxygen or the carboxylic acid with hydrogen peroxide. In an alternative approach, the use of an organic peracid is dispensed with altogether by employing hydrogen peroxide in the presence of a catalyst. A prerequisite for success is that the method should be amenable to the use of commercially available (30 or 60%) aqueous hydrogen peroxide and preferably avoid the use of environmentally unattractive solvents such as chlorinated hydrocarbons. If successful, such a method would circumvent both the environmental and the safety issues associated with the classical BV oxidation. In short, there is a definite need for a green BV oxidation procotol, which utilizes aqueous hydrogen peroxide as the stoichiometric oxidant in an environmentally attractive solvent or (preferably) under solvent-free conditions. From the viewpoint of scope in organic synthesis, the method should also exhibit high degrees of chemo-, regio-, and enantioselectivity and broad substrate specificity. Chem. Rev. 2004, 104, 4105−4123 4105 10.1021/cr030011l CCC: $48.50 © 2004 American Chemical Society Published on Web 08/14/2004

Consequently, in this review, we will focus on green BV reactions using hydrogen peroxide. In the first part, some general features of reactions with hydrogen peroxide and dioxygen are delineated. In the second part, the mechanism of the BV reaction is analyzed to identify ways in which a catalyst might improve the reaction. In the third part, reactions catalyzed by homogeneous catalysts, biocatalysts, and heterogeneous catalysts are discussed. 2. H2O2 and O2 as Green Oxidants 2.1. Hydrogen Peroxide The above-mentioned drawbacks of the classical BV reaction have stimulated considerable activitys especially in the past few yearssin the development of catalysts that employ hydrogen peroxide as a clean oxidant.3b,5 The use of hydrogen peroxide has many advantages: it is safe and cheap, the active oxygen content is high, it does not require a buffer, and it is clean, since the byproduct formed is water. These points make the use of hydrogen peroxide extremely interesting from an industrial (large-scale) point of view. However, there are some disadvantages concerning the use of hydrogen peroxide.6 (i) Because water is always present in solution, hydrolysis of the product esters may occur, and not all substrates are therefore compatible with water. (ii) Hydrogen peroxide is one of the weakest oxidants of a wide range of available peroxides and peracids (see above), and a catalyst is required to activate it. (iii) Some catalysts show a low selectivity on hydrogen peroxide. This may cause the formation of unselective hydroxy or hydroperoxy radicals. Furthermore, pure dioxygen may evolve from H2O2 decomposition, causing a build-up of pressure and creating a potentially unsafe combination with flammable organic solvents. (iv) High concentrations of hydrogen peroxide (>40% mol/mol) in organic solvents are unsafe. To avoid dangerous situations as mentioned in points iii and iv above, Jones6 recommends adhering to the following checklist: (i) Avoid any contamination in the reaction vessel (which may induce an uncontrolled reaction). (ii) Avoid build-up of oxygen pressure (by venting and flushing with N2). (iii) Keep the concentration of the peroxo compound below 20% mol/mol (by presetting the reaction temperature, adding the peroxo compound last,7 stirring the reaction mixture, making sure that the peroxo compound reacts completely before adding more, and providing cooling if required). (iv) Destroy excess peroxo compound before work-up. (v) Never use acetone or other low boiling ketones as the solvent for cleaning or extraction. 2.2. Dioxygen Free radical autoxidation of an aldehyde is facile and affords the corresponding peracid. In the presence of a reactive substrate, e.g., an olefin or a ketone, Gerd-Jan ten Brink was born in Rijnsaterwoude, The Netherlands, in 1971. He received his M.Sc. degree from the Free University of Amsterdam (1995) under the supervision of Professor F. Bickelhaupt. In 2001, he received his Ph.D. degree (cum laude) for his research on “Green Catalytic Oxidations” under the supervision of Professor R. A. Sheldon. After a year of postdoctoral research spent jointly in the Sheldon group and at Avantium Technologies BV, he now works for ChemShop BV in Weert, The Netherlands. Isabel W. C. E. Arends (born 1966) studied chemistry at the University of Leiden (The Netherlands), where she received her Ph.D. in physical organic chemistry in 1993, under the supervision of Professor R. Louw and Dr. P. Mulder. Postdoctoral work followed with Professor K. U. Ingold at the National Research Council in Canada on liquid phase oxidations catalyzed by biomimetic iron complexes. She joined the group of R. A. Sheldon in 1995, where she was appointed Assistant Professor in 2001. Her research interests focus on enzyme- and metal-catalyzed redox reactions and green selective oxidations employing O2 and H2O2 in particular. Roger Sheldon (1942) received a Ph.D. in organic chemistry from the University of Leicester (United Kingdom) in 1967. This was followed by postdoctoral studies with Professor Jay Kochi in the United States. From 1969 to 1980, he was with Shell Research in Amsterdam, and from 1980 to 1990, he was R&D Director of DSM Andeno. In 1991, he moved to his present position as Professor of Organic Chemistry and Catalysis at the Delft University of Technology (The Netherlands). His primary research interests are in the application of catalytic methodologiesshomogeneous, heterogeneous, and enzymaticsin organic synthesis, particularly in relation to fine chemicals production. He developed the concept of E factors for assessing the environmental impact of chemical processes. Figure 1. Oxidation of menthone with Caro’s acid. 4106 Chemical Reviews, 2004, Vol. 104, No. 9 ten Brink et al

the peracid can transfer an oxygen atom to the substrate, resulting in the formation of one equivalent of epoxide or ester and acid. Oxidations involving the in situ formation of a peracid from an aldehyde and dioxygen are generally referred to as the Mukaiyama method.8 One industrial route to -caprolactone, for instance, involves the in situ formation of peracetic acid from acetaldehyde (Figure 2).9 The use of metal catalysts is optional,10 and the combination aldehyde/dioxygen is often not significantly different from peracids, although it should be noted that radical type chemistry may take place instead of the intended BV reactions. In alcoholic solvents, radical type side reactions are suppressed to a certain extent.11 Recently, Ishii12 reported on the “aerobic” BV oxidation of a cyclohexanol/cyclohexanone mixture (KA oil) to yield lactones (Figure 3). However, in this reaction, cyclohexanol is first oxidized to give cyclohexanone and hydrogen peroxide and the latter is used as the true oxidant in the BV reaction. The latter methodsusing O2 for the in situ formation of hydrogen peroxide as the actual oxidantshas been receiving much attention over the past years because it is cheaper than hydrogen peroxide itself.13 3. Theoretical Considerations 3.1. Mechanism of the Reaction For an in-depth discussion on the mechanism of the BV reaction, we refer to the excellent reviews of Krow2 and Meunier.3a A few key issues are briefly mentioned here. The generally accepted mechanism for the BV oxidation is a simple two-step reaction that involves the so-called Criegee intermediate or adduct. In the first step, a peroxide attacks the polarized CdO bond. The second step follows a concerted pathway (Figure 4). Only with acylperoxo type oxidants can the hydroxyl proton in this intermediate migrate intramolecularly. Hence, these oxidants are more effective than alkylperoxo type oxidants, which generally require a catalyst.14 It should be noted that in many reactions the two steps have activation energies that are in the same order of magnitude. Hence, catalysts may need to facilitate both steps of the reaction. With some exceptions,15 the rearrangement step is usually rate limiting.16 In the Criegee intermediate, a proper alignment is required for the rearrangement step: The migrating group RM needs to be antiperiplanar17 to the O-O bond of the leaving group (primary stereoelectronic effect) and antiperiplanar to a lone pair of the hydroxyl group (secondary stereoelectronic effect). In 1980, Noyori18 provided evidence for the existence of the secondary effect, but compelling evidence for the primary effect was not yet available.19 Criegee rearrangements in allyl hydroperoxides20 already hinted at such an effect, but in 2000, Crudden et al. showed its existence in a true BV oxidation of trans- and cis- 4-tert-butyl-2-fluorocyclohexanone (Figure 5).21 When the 2-fluoro substituent in 4-tert-butyl-2- fluorocyclohexanone is aligned in an axial position, differences in dipole effects in the various conformations are minimal and do not influence the migration of either the CH2 or the CHF group. In this case, a normal product distribution22 is observed and the electron-rich CH2 group migrates preferentially. However, when the 2-fluoro substituent is placed in an equatorial position, the conformation with CH2 antiperiplanar to the O-O bond creates an unfavorable dipole interaction of the perester with the CHF group. In this case, the electron-poor CHF group can achieve the required alignment more easily and is “forced” to migrate. Thus, at least in these cases, the primary stereoelectronic effect is more important than the migratory aptitude.23 Figure 2. Mukaiyama oxidation of cyclohexanone to caprolactone. Figure 3. “Aerobic” BV reaction of KA oil. Figure 4. Mechanism for BV reaction as proposed by Criegee. RM is the migrating group. Figure 5. BV oxidation of trans- and cis-4-tert-butyl-2- fluorocyclohexanone. The Baeyer−Villiger Reaction Chemical Reviews, 2004, Vol. 104, No. 9 4107

A more detailed mechanism (Figure 6) shows the possible mechanisms by which catalysts may improve BV reactions. Here, one can distinguish (1) electrophilic activation of the substrate, (2) electrophilic activation of the intermediate, (3) nucleophilic activation of the intermediate, (4) nucleophilic activation of (hydrogen) peroxide, and (5) electrophilic activation of (hydrogen) peroxide. 3.2. Electrophilic Activation of Substrate The action of acids (H+ or metal cations) is in part to activate the carbonyl functionality toward nucleophilic attack of peroxide or peracid via increasing the polarization of the CdO double bond (Figure 6, intermediate 1). Therefore, the combination CF3- CO3H/CF3CO2H gives one of the most reactive peracids, even though CF3CO3 - is a weak nucleophile, reluctant to attack the polarized carbonyl functionality. Indeed, in a buffered solution, the activity of CF3- CO3H is strongly diminished indicating that an improved leaving group effect of CF3COO- may not be important. Other work, however, indicated that electron-withdrawing substituents on the leaving group did actually facilitate rearrangement, an effect observed in oxidation both with peracids24 and with hydrogen peroxide.3b One example of transition metal-catalyzed electrophilic activation of substrates is the platinum-CF3 system, described below, which was developed in the group of Strukul (Figure 7).25 Activation of the ketone via coordination to Lewis acids seems to be the most general way to activate substrates for BV oxidation. In this case, the ketone coordinates to an electronpoor platinum center and becomes susceptible to attack of free hydrogen peroxide (intermediate I). This activation is somewhat reminiscent of activation of R,â-unsaturated ketones in Diels-Alder reactions. Not surprisingly, cationic platinum complexes of (chiral) diphosphines proved to be active in this reaction as well.26 To our knowledge, catalysts that are typically successful in Diels-Alder reactions, such as lanthanides, are rarely used to activate ketones for BV reactions with H2O2, 27 although these water stable Lewis acids seem to meet all of the requirements for a successful BV reaction. Other Lewis acids such as gallium(III) or tin(IV) chloride are too water sensitive and have mainly been successful under anhydrous conditions with, e.g., bis- (trimethylsilyl)peroxide as the oxidant (Figure 8).28,29 Clearly, the method is far from green. Corma et al.30 developed solid tin catalysts that are water stable and use hydrogen peroxide as the oxidant (see later under solid Lewis acids). 3.3. Electrophilic Activation of Intermediate In BV reactions with peracids as oxidants, strong acids, such as CF3CO2H, may also catalyze the rearrangement step via protonation of the carbonyl functionality of the leaving group (Figure 9). As this rearrangement step is usually rate limiting, the catalyst has a large effect here. Activation of the intermediate hydroperoxy adduct is similar to activation of the acylhydroperoxy intermediate. A Lewis acid may also facilitate the migration step, via coordination or protonation of the hydroxide (alkoxide), which is otherwise a very poor leaving group. This is again illustrated with the (dppe)Pt(CF3)]+ complex where the platinum center facilitates the rearrangement step via coordination to hydroxide (Figure 7, intermediate I). In most if not all cases, Lewis acid catalysts can facilitate both steps of the reaction. It is not always trivial to make a distinction between Brønsted and Lewis acid catalysis, as the pH of a solution may decrease when Lewis acidic metals are added to the reaction mixture. Therefore, carrying out BV reactions in buffered solutions may sometimes lead to surprising results. One important difference in BV reactions with hydrogen peroxide is Figure 6. Electrophilic and nucleophilic activation of the BV reaction. Figure 7. BV oxidation with (dppe)Pt(CF3)]+. Figure 8. Lewis acid-catalyzed oxidation of 2-(3-methyl- 2-butenyl)cyclopentanone. Figure 9. Acid-catalyzed BV oxidation with peracids as the oxidant. 4108 Chemical Reviews, 2004, Vol. 104, No. 9 ten Brink et al

that with Brønsted acid catalysts, dimeric, trimeric, or polymeric peroxides seem to be formed more easilyscompounds that are potentially explosive (Figure 10). Indeed, a BV reaction can sometimes proceed via such a dimeric peroxide intermediate as recently shown by Berkessel and co-workers (see also later Figure 17).31 3.4. Nucleophilic Activation of Intermediate On the basis of the mechanism depicted in Figure 9, it is difficult to imagine base catalysis to activate the intermediate. Base catalysis was observed when bicarbonate was added to a solution of m-CPBA and a bicyclic ketone in dichloromethane.32 The reaction rate nearly doubled, which was ascribed to an accelerated rearrangement step of an anionic Criegee adduct as compared to the neutral adduct (Figure 11). Renz and Meunier noted in their review3 that in the reaction mentioned above, bicarbonate also removed the coproduct, m-CBA, from the reaction mixture via deprotonation and precipitation. In this way, the m-CBA could not compete with m-CPBA for the substrate, resulting in an increase in rate. Although BV reactions are sometimes carried out under neutral to basic conditions to avoid acidcatalyzed side reactions, base catalysis is not commonly observed in BV reactions with hydrogen peroxide.33 3.5. Nucleophilic Activation of Hydrogen Peroxide Very few transition metals can catalyze BV reactions with hydrogen peroxide. The early transition metals (Ti, V, Mo, and W) may form peroxo complexes with hydrogen peroxide, but these are generally electrophilic in nature. Therefore, these complexes are active in, for example, epoxidation via electrophilic attack on preferably electron-rich olefins. A nucleophilic attack on the partially positively charged carbon of the CdO functionality is unlikely to occur with these complexes. Such a reaction seems to be the domain of the late transition metal peroxo complexes such as (ligand)Pt(O)2 or (ligand)Pd(O)2, which are partly nucleophilic in nature (see later). Indeed, the first example of transition metal catalysis, which involved a (dipicolinato)MoVI peroxo complex (Figure 12) in the oxidation of cyclic ketones with 90% hydrogen peroxide, later turned out to be a simple acid-catalyzed reaction, rather than the first example of nucleophilic reactivity of a group VI peroxo metal complex.34 With this in mind, the MTO-catalyzed BV oxidation of cyclobutanone with aqueous hydrogen peroxide becomes all the more suspicious.35 MTO is an extremely active catalyst for the epoxidation of olefins with aqueous hydrogen peroxide.36 The active bisperoxo intermediate gives an electrophilic attack on the double bond of the alkene. Therefore, a proposed nucleophilic attack of the same bisperoxo complex on the CdO double bond of, e.g., cyclobutanone (Figure 13), would seem unlikely. However, with the evidence available until now, it appears that MTO can exhibit electrophilic properties in epoxidation and nucleophilic properties in BV oxidation.37 The reason that MTO may change its nucleophilic/ electrophilic behavior depending on the substrate is not entirely clear. If the ketone coordinates to rhenium, then the metal plays a role as a Lewis acid and induces electrophilic activation of the substrate. The coordination of basic ligands (ketone) also increases the electron density on the metal center, which in turn increases the nucleophilic character of the peroxo groups.38 Reaction of the bisperoxo complex with thianthrene-5-oxide did reveal a partly nucleophilic character of the catalyst,39 and in the oxidation of 1,3-diketones, MTO acts as both an electrophilic and a nucleophilic catalyst.37 A similar intermediate has been proposed for the rhenium-catalyzed reaction as for the molybdenumcatalyzed reaction, but in this case, 17O NMR revealed polarization of the peroxo moiety, which might explain the nucleophilic character. Furthermore, contrary to the MoVI system, a stoichiometric reaction between the rhenium bisperoxo complex and cycloalkanones in the absence of hydrogen peroxide did lead to product formation, indicating that the rhenium bisperoxo complex is more than an expensive (BrønFigure 10. Formation of organic peroxides under acidic conditions. Figure 11. Rearrangement of anionic Criegee adduct. Figure 12. Molybdenum-catalyzed BV reaction of cyclopentanone. Figure 13. MTO-catalyzed oxidation of cyclobutanone. The Baeyer−Villiger Reaction Chemical Reviews, 2004, Vol. 104, No. 9 4109

sted) acid. It should be noted, however, that the active rhenium complex contains one aqua ligand, which is highly acidic (similar to the molybdenum complex). This acidity may still account for part of the activity of MTO under catalytic conditions. Examples of catalysis with late transition metal complexes are the Pt systems with bridging hydroxy ligands developed in the group of Strukul.40 The work is based on the premise that platinum-η2-peroxo complexes, which can be formed from (ligand)Pt(0) in a reaction with dioxygen, give a nucleophilic attack on ketones. However, such platinum-η2-peroxo species only react in stoichiometric reactions (Figure 14). When platinum salts are used in combination with hydrogen peroxide, a platinumhydroperoxo complex may be active. In the [(dppb)Pt(µ-OH)]2 2+-catalyzed oxidation of ketones (Figure 15), again, the platinum center gives an electrophilic activation of the ketone via coordination. In this case, hydrogen peroxide is also believed to coordinate to the platinum center and attack on the ketones proceeds intramolecularly. The difference with the previous platinum system (Figure 7) is that the platinum center may activate the hydrogen peroxide if (dppb)PtOOH]+ is indeed more nucleophilic than HOOH. It should be noted, however, that attack of coordinated hydrogen peroxide and attack of free hydrogen peroxide on the ketone are indistinguishable in kinetic investigations. Alternatively, coordination of platinum to hydrogen peroxide makes the latter more acidic, which might also promote BV oxidation. This would constitute an electrophilic activation of hydrogen peroxide. Again, platinum facilitates the rearrangement step by coordinating with the hydroxide leaving group () electrophilic activation).41 The platinum systems will be discussed further in asymmetric BV (see above and section 4.1.5). 3.6. Electrophilic Activation of Hydrogen Peroxide Interestingly, in a recent study, Brinck et al.42 showed that in the BF3-catalyzed reaction of acetone and hydrogen peroxide the Lewis acid facilitated the reaction via coordination to hydrogen peroxide, making the latter more acidic and increasing hydrogen bonding to the carbonyl functionality (Figure 16). The coordination of BF3 to acetone was calculated to lead to stabilization of the adduct, rendering it nearly unreactive! The same Lewis acid also facilitated the rearrangement step after migration to the outer peroxygen, creating a BF2OH leaving group rather than a hydroxide leaving group. As was pointed out above, many early transition metals can coordinate to hydrogen peroxide, making it more electrophilic and more eager to attack electronrich substrates such as olefins. By the same token, such electrophilic activation would decrease the tendency to attack already electron-poor ketones in a BV reaction. Neumann recently reported on the electrophilic activation of hydrogen peroxide by 1,1,1,3,3,3-hexafluoro-2-propanol in the oxidation of olefins and ketones.43 This solvent may form hydrogen bonds with hydrogen peroxide, but it is not able to receive hydrogen bonds back, due to the decreased electron density on CF3CHOHCF3 itself. This particular solvent, therefore, makes hydrogen peroxide more electrophilic, which should indeed promote attack of the peroxygens on olefins but not on ketones. This apparent anomaly was subsequently explained by Berkessel and co-workers31 who showed that Brønsted acid-catalyzed BV oxidations with hydrogen peroxide proceed by a nonclassical mechanism in CF3- CHOHCF3 (Figure 17). The intermediacy of spiro bisperoxide (1) was established by following the course of the reaction in (CF3)CDOD with 13C NMR. Figure 14. Nucleophilic reaction of platinum-η2-peroxo complex on ketone. Figure 15. BV oxidation with [(dppb)Pt(µ-OH)]2 2+. Figure 16. BF3-catalyzed oxidation of acetone with hydrogen peroxide. 4110 Chemical Reviews, 2004, Vol. 104, No. 9 ten Brink et al

4. Catalytic Reactions 4.1. Homogeneous Catalysts 4.1.1. Oxidation of Aldehydes One of the most underestimated oxidation reactions is undoubtedly the oxidation of (benz)aldehydes. A selective route to form benzoic acids is oxidation of the hydratesformed from the aldehyde and waters with strong inorganic oxidants such as KMnO4, CrO3, fuming HNO3, Jones reagent, etc. However, environmental considerations have shifted the attention to BV type reactions. In this case, the reaction can yield two products: the corresponding benzoic acid and the ester of the corresponding phenol and formic acid. Formation of the latter product from benzaldehydes is a useful alternative to direct hydroxylation of aromatics (Figure 18). With electron-donating hydroxy or amino substituents on the ortho or para position, the so-called Dakin reaction can be carried out without a catalyst under alkaline conditions. Recenty, a number of articles have appeared on the oxidation of aldehydes with aqueous hydrogen peroxide. The reaction is catalyzed by, e.g., Brønsted acids,44 MTO,45 arylseleninic acids, and SeO2. 46 Although the titles of some articles may imply that the catalysts involved are particularly effective to direct the reaction to either acid or phenol, the electron density on the phenyl ring largely determines this selectivity. Electron-donating substituents favor ring migration to yield phenols, whereas electron-withdrawing substituents favor hydrogen migration to yield acids.47 However, some differences in selectivity can be found depending on solvent type and pH. For instance, the oxidation of piperonal (Figure 19) gives the phenol under acidic conditions,48 whereas under alkaline conditions mainly the corresponding acid is formed.49 Under more or less neutral conditions, the selectivity is directed to the phenol when bis(2- nitrophenyl) diselenide is used as the catalyst (precursor).50 Generally speaking, seleninic acid catalysts show a high tendency to form phenols if electron-donating substituents are present on the aromatic ring of the substrate.51 With MTO,45 selectivity to the (electronrich) phenols is lower than with bis(2-nitrophenyl) diselenide. Table 1 gives an overview of the selectivity of several catalysts active in the oxidation of aldehydes. Several catalysts based on seleniumsnotably SeO2 and Ph2Se2 in THF52sshowed a high selectivity for the carboxylic acid. Noyori et al.44 used a simple lipophilic acid catalyst, [CH3(n-C8H17)3N]HSO4, to convert aldehydes to carboxylic acids under halide and metal-free conditions without the presence of any organic solvent (Figure 20). Although details were not given, it seems reasonable that substantial amounts of phenol were formed in those cases when only low yields of acid were reported. Possibly, the phenols are oxidized further. The system developed by Noyori is probably the easiest and greenest way to oxidize aldehydes to carboxylic acids to date, and the protocol is suitable to oxidize aldehydes on a mole-scale. The Figure 17. Nonclassical Brønsted acid-catalyzed BV oxidation in (CF3)2CHOH. Figure 18. BV reaction of aldehydes. Table 1. H2O2 Oxidation of Aldehydes to Acid/Phenol Mixtures substratea SeO2 b (ref 46) ArSe(O)OHb,c (ref 51) WO4 2- d (ref 53) [CH3(n-C8H17)3N]HSO4 d (ref 44) H+/MeOHb (ref 48) 4-NO2-æ 87/0 99/0 89 93 80/0 4-Cl-æ 83/0 50/50 86 76 87/0 4-CH3-æ 88/0 45/55 57 41 51/28 4-H-æ 97/0 NDe 84 85 NDe 4-MeO-æ 46/41 0/95 6 9 0/90 octanal 91/0 95/1 87 82 NDe a æ: -C6H4CHO. b Acid to phenol ratio. c ArSe(O)OH is 3,5-(CF3)2C6H3Se(O)OH. d Selectivity to acid; yield of phenol not given. e ND ) not determined. Figure 19. Oxidation of piperonal. Figure 20. Oxidation of 4-nitrobenzaldehyde. The Baeyer−Villiger Reaction Chemical Reviews, 2004, Vol. 104, No. 9 4111

results were not greatly different from those obtained by Venturello with a tungstate catalyst.53 SeO2 is especially good for the oxidation of linear aldehydes to acids but also for the oxidation of heteroaromatic aldehydes to acids.46 Both electronpoor pyridine-3-carbaldehyde (86%) and electron rich thiophene-2-carbaldehyde (88%) and furan-2-carbaldehyde (73%) were oxidized to the respective acids (Figure 21).54 The SeO2/H2O2 combination was not useful in the oxidation of 2-indolecarbaldehyde, as these types of compounds are too sensitive. However, a combination of m-CPBA and p-toluenesulfonic acid proved to be successful for a series of substituted indole derivatives.55 In this case, however, the 3-hydroxyindole derivative was formed, instead of the acid. To avoid contamination of the products with selenium catalysts, Knochel56 immobilized 2,4-bis(perfluorooctyl)phenyl butyl selenide in a fluorous phase.57 The system was improved slightly by using the 3,5- bis(perfluorooctyl)phenyl butyl selenide isomer of the catalyst and by using the more polar dichloroethane as a cosolvent system instead of benzene (Figure 22).58 The latter selenium isomer is slightly more difficult to synthesize, but both catalysts could be recycled without a serious loss of activity. Alternatively, aldehyde oxidation can be carried out with dioxygen as the sole oxidant. In this case, a peracid is formed from the aldehyde that attacks a second equivalent of aldehyde to yield two equivalents of acid. The nickel bis(triacontafluoroheptadeca- 8,10-dionate) catalyst59 (Figure 23)salso developed in the group of Knochelscan be recycled with a minimal loss of activity when operating under fluorous biphasic conditions, but it should be noted that solvent properties of perfluoroalkanes are not optimal for these (polar) oxidations reactions. Ionic liquids60 allow more fine tuning of the solvent properties, but so far, only one example of aerobic aldehyde oxidation in [bmim]PF6 has been reported and the results are suboptimal.61 4.1.2. Oxidation of Cycloalkanones From an industrial point of view, oxidation of cyclohexanone to -caprolactone is one of the more interesting BV reactions. The product lactone is polymerized and used in foams, biodegradable plastics, etc. The reaction is often carried out with oxygen in combination with a sacrificial aldehyde such as acetaldehyde or benzaldehyde.62 Alternatively, oxidation is carried out with a carboxylic acid, which is converted in situ to its corresponding peracid by the action of hydrogen peroxide plus an acid catalyst.63 A large number of catalysts have been shown to be active in the oxidation of cycloalkanones to lactones using only hydrogen peroxide as the oxidant. MTO is moderately active in the oxidation of linear ketones37 or higher cycloalkanones,64 but it is particularly active in the oxidation of cyclobutanone derivatives (Figure 24), which are oxidized faster with MTO than with other existing methods.65 Interestingly, with bicyclo[3.2.0]hept-2-en-6-one (Figure 24) competing, epoxidation of the olefin was negligible, even though MTO is an extremely good epoxidation catalyst. With <1 mol % MTO, cyclobutanones are fully converted within 1 h. Even R,Rdichlorocyclobutanones are oxidized, compounds that Figure 21. Oxidation of heteroaromatic aldehydes. Figure 22. Fluorous phase BV oxidation immobilized selenium catalysts. Figure 23. Nickel-catalyzed aerobic oxidation of substituted benzaldehydes. Figure 24. MTO-catalyzed oxidation of cycloalkanones. 4112 Chemical Reviews, 2004, Vol. 104, No. 9 ten Brink et al

are otherwise almost inert.65 Successful oxidation of a 4-chromanone derivative was achieved with a â-methoxy substituent present in the substrate.64 Other (higher) cycloalkanones react very slowly, and the use of strong acids in combination with (anhydrous) urea hydrogen peroxide (UHP; Figure 25) or preformed peracid catalysts is required66 (see also linear ketones). Another approach consists of the use of a fluorous Sn catalyst under biphasic conditions.67 A perfluorinated tin(IV) compound, Sn[NSO2C8F17]4, was recently shown to be a highly effective catalyst for BV oxidations of cyclic ketones with 35% hydrogen peroxide in a fluorous biphasic system (Figure 26). The catalyst, which resides in the fluorous phase, could be easily recycled without a loss of activity. Analogous compounds of Hf, Sc, and Yb were also shown to catalyze the BV oxidation, albeit less effectively than the Sn compound. 4.1.3. Oxidative Ring Contraction The oxidation of cycloalkanones to lactones is probably the standard BV reaction. For some time, however, ring contraction of cycloalkanones with SeO2/H2O2 has been known to be a synthetically useful side reaction.68 Recently, Mlochowski reinvestigated the reaction and developed effective catalysts for the conversion of cycloalkanones to cycloalkanecarboxylic acids (Figure 27).69 It has been proposed6 that a Se(VI) species is responsible for oxidative ring contraction, whereas a Se(IV) species catalyzes conventional BV reactions. This is consistent with the observation that only electron-rich selenium species seem to catalyze this side reaction. Products such as cyclopentanecarboxylic acid and cyclohexanecarboxylic acid are important in the synthesis of natural products and pharmaceuticals. They cannot always be obtained via the Favorski reaction. The present route is considerably more environmentally friendly than the use of stoichiometric amounts of thallium(III) salts.70 4.1.4. Oxidation of Linear Ketones Partly because of the lack of ring strain, linear ketones are generally reluctant to undergo the BV reaction. For the reaction of linear ketones with H2O2, strong protic acids are required to activate the ketone, but this also facilitates hydrolysis of the ester and the formation of stable peroxides or polymers containing peroxy groups.71 An alternative oxidant is HOF/CH3CN formed in situ from elemental fluorine.72 In this case, the reaction proceeds via an oxirane intermediate rather than the Criegee intermediate (Figure 28). The use of Lewis acid catalysts can largely circumvent the problem of peroxide formation. Lewis acids such as BF3 and SbF5 have already been used in BV reactions to activate (linear) ketones. Also, the platinum catalysts developed by Strukul26 and the tin zeolite â developed by Corma30 facilitate the reaction in a similar fashion. 4.1.5. Enantioselective Oxidations In 1994, Strukul et al. were the first to report an enantioselective BV reaction of a racemic mixture of 2-substituted cycloalkanones using hydrogen peroxide as an oxidant and chiral Pt complexes as catalysts.73 Shortly after, Bolm et al. published their system, using chiral Cu complexes and aldehyde/O2 as the oxidant for this reaction.74 Since then, a number of chiral catalysts for such reactions of racemic ketone mixtures (Figure 29 and Table 2) and Figure 25. BV oxidation of cyclotridecanone. Figure 26. Use of a fluorous Sn catalyst for BV oxidations in a fluorous biphasic system. Figure 27. Oxidative ring contraction of cyclohexanone. Figure 28. BV oxidation of 2-alkanones. Figure 29. Asymmetric BV oxidation of bicyclo[4.2.0]- octanone. The Baeyer−Villiger Reaction Chemical Reviews, 2004, Vol. 104, No. 9 4113