Complimentary and personal copy www.thieme.com SYNTHESIS Reviews and Full Papers in Chemical Synthesis This electronic reprint is provided for non- commercial and personal use only: this reprint may be forwarded to individual colleagues or may be used on the author’s homepage. This reprint is not provided for distribution in repositories, including social and scientific networks and platforms. Publishing House and Copyright: © 2015 by Georg Thieme Verlag KG Rüdigerstraße 14 70469 Stuttgart ISSN0039-7881 Any further use only by permission of the Publishing House SYNTHESIS0039-78811437-210X © Georg Thieme Verlag Stuttgart · New York 2015, 47, 3079–3117 3079 review Syn thesis S. E. Mann et al. Review Palladium(II)-Catalysed Oxidation of Alkenes Sam E. Manna R L. Benhamoub R R Tom D. Sheppard*b PdX2 p-allyl comPpdlXex Nu or Nu allylic oxidation aArgenta, Discovery Services, Charles River, 7-9 Spire Green Centre, Nu Nu Flex Meadow, Harlow, Essex, CM19 5TR, UK R PdX2 R – HPdX R bDInegpoaldrt Lmaebnotr aotfo Crihees,m 2i0st Gryo, rUdnoinv eSrts, iLtyo nCdoollne,g We LCo1nHd o0nA,J ,C UhKristopher H PdX Wacker oxidation [email protected] s-alkyl complex Nu R Nu = OR, NR2 1,2-difunctionalisation FG 1 Introduction Received: 01.05.2015 Accepted after revison: 25.06.2015 Published online: 26.08.2015 1.1 Background DOI: 10.1055/s-0035-1560465; Art ID: ss-2015-e0286-r Abstract This review provides a summary of recent developments in Alkenes are extremely abundant chemical feedstocks the palladium(II)-catalysed oxidation of alkenes, focusing largely on re- which are produced in large quantities from petrochemical actions which lead to the formation of new carbon–oxygen or carbon– sources. As a consequence, they have been exploited as nitrogen bonds. Three classes of reaction are covered: i) oxidations pro- ceeding via allylic C–H bond cleavage and formation of a π-allyl com- starting materials for the synthesis of a wide range of or- plex; ii) Wacker-type oxidations proceeding via nucleopalladation fol- ganic chemical building blocks. The oxidation of alkenes to lowed by β-hydride elimination; and iii) 1,2-difunctionalisation of introduce carbon–oxygen and other carbon–heteroatom alkenes proceeding via nucleopalladation followed by functionalisation bonds is a highly important process which enables higher of the resulting σ-alkylpalladium(II) intermediate. The mechanisms are discussed alongside the scope and limitations of each reaction. polarity molecules to be prepared from these abundant hy- 1 Introduction drocarbons. Metal-catalysed oxidation reactions are partic- 1.1 Background ularly notable in this respect and a palladium-catalysed 1.2 Oxidation Pathways alkene oxidation – the Wacker oxidation – was one of the 1.3 Observation of Reaction Intermediates first transition-metal-catalysed industrial processes, en- 2 Allylic Oxidation 2.1 Background abling the efficient generation of acetaldehyde from eth- 2.2 Allylic Oxygenation ylene.1 2.3 Allylic Amination Since this pioneering process was developed, the field 2.4 Allylic Functionalisation with Other Nucleophiles has grown considerably and a range of palladium-catalysed 3 The Wacker Oxidation oxidation reactions are routinely used in synthetic chemis- 3.1 Background 3.2 Variation of the Co-Oxidant try laboratories all over the world. These reactions general- 3.3 Direct Oxygen-Coupled Wacker Oxidations ly rely on the strong interaction of palladium(II) salts with 3.4 Aldehyde-Selective Wacker Oxidations the π-orbitals of an alkene, which opens up several differ- 3.5 Wacker Oxidation of Internal Alkenes ent reaction pathways via which new carbon–heteroatom 3.6 Aza-Wacker Oxidations bonds can be formed (Scheme 1). Importantly, these path- 4 Intermolecular 1,2-Difunctionalisation of Alkenes 4.1 Introduction ways are often finely balanced, with similar substrates 4.2 Oxyhalogenation Reactions sometimes undergoing oxidation via different mechanisms. 4.3 Dioxygenation Reactions The choice of catalyst and/or ligand can provide high levels 4.4 Oxycarbonylation Reactions of control over the reaction pathway, however, enabling dif- 4.5 Aminohalogenation Reactions 4.6 Diamination Reactions ferent products to be accessed from the same substrate. 4.7 Aminooxygenation Reactions This review covers the diverse methods available for the 4.8 Aminocarbonylation Reactions palladium-catalysed oxidation of alkenes, discussing the 5 Summary and Conclusions likely mechanisms of the reactions in each case and outlin- ing potential new directions for research in this area. Key words alkenes, oxidation, palladium, Wacker oxidation, catalysis © Georg Thieme Verlag Stuttgart · New York — Synthesis 2015, 47, 3079–3117 3080 Syn thesis S. E. Mann et al. Review 1.2 Oxidation Pathways R2 R1 R2 R1 1 β-hydride Coordination of an alkene 1 (Scheme 1) to a palladi- H elimination Nu 7 Pd(0) um(II) salt renders a typically electron-rich alkene consid- R2 R1 path 2 erably more electrophilic, allowing 2 to be attacked by oxy- R2 R1 Nu 4 2 gOenne ncoumclemoopnh ipleast hswucahy iansv owlvaetesr i, naitlciaolh aoblss,t roarc taiocent aotfe a ino nals-. X Pd XH Nu R2 PdX R1 R2 Nu R1 or lylic proton by the nucleophile (or a base) to generate π-al- path 1 Nu 6a PdX 6b lyl complex 3 (path 1). This relatively stable complex can then undergo a nucleophilic substitution reaction, general- R2 R1 path 3 FG ly via an outer sphere process, to give allylic oxidation prod- 3 'FG' R2 R1 ucts 4 and 5, together with a palladium(0) species, which PdX [O] Nu 8 Pd(II) R2 R1 must be re-oxidised to palladium(II) in order for the cata- 5 Nu [O] lytic cycle to be completed. Alternative reaction pathways Pd(0) Pd(II) R2 R1 involve direct attack of the nucleophile onto the activated alkene 2 to give a σ-alkylpalladium intermediate 6a (and/or Nu 4 its regioisomer 6b) via a concerted nucleopalladation.2 In- Scheme 1 Pathways for palladium-catalysed alkene oxidation termediate 6a can then undergo β-hydride elimination (path 2) to give either vinylic oxidation product 7 or allylic cleophile is typically water and the resulting enol 7 oxidation product 4, depending on the regioselectivity of (Nu = OH) undergoes tautomerisation to the ketone or alde- the β-hydride elimination. In the Wacker oxidation, the nu- hyde. It should be noted, however, that even this tautomeri- Biographical sketches Sam Mann graduated from University College London. His synthesis. In 2011, he joined Ar- the University of York in 2007, doctoral research focused on genta, which became part of where he obtained an MChem, the development of new syn- Charles River’s Discovery Ser- having spent a year on ex- thetic methodologies involving vices in 2014. He currently change at the Université Joseph cyclic oxygen, nitrogen and sul- holds the position of Senior Sci- Fourier in Grenoble, France. He fur acetal derivatives and their entist, working in the medicinal then moved to London to study applications in palladium(II)-ca- chemistry group. for a Ph.D. under the supervi- talysed oxidations, multicompo- sion of Dr. Tom Sheppard at nent reactions and medium ring Laure Benhamou obtained Davit Zargarian (Université de carbon–carbon bond-forming her Ph.D. in 2009 from Toulouse Montréal, Canada) for a year as reactions. Since September University (France). She worked a postdoctoral researcher to 2012, she has been a senior under the supervision of Dr Guy study the chemistry of Ni(III)- postdoctoral researcher with Dr Lavigne and Dr Vincent César on pincer complexes. She then Tom Sheppard (UCL, London) the elaboration of N-heterocy- moved to Geneva (Switzerland), where she is developing new clic carbene ligands and their as a postdoctoral fellow with catalytic methodology for the subsequent coordination to Prof. Peter Kündig in 2011, synthesis of boron–oxygen het- late-transition metals. In 2010, where she developed chiral Pd- erocycles. she joined the group of Prof. NHC catalysts for asymmetric Tom Sheppard obtained his Motherwell at University Col- tureship in organic chemistry at Ph.D. under the supervision of lege London, working on zinc- University College London, and Professor Steve Ley at the Uni- carbenoid cyclopropanation re- in 2013 he was promoted to versity of Cambridge in 2004, actions and multicomponent Reader. His current research in- working on the application of reactions. In 2007, he was terests include novel organobo- butane-2,3-diacetals in asym- awarded an Engineering and ron chemistry, transition-metal- metric synthesis. He then Physical Sciences Research catalysed reactions and sustain- undertook postdoctoral re- Council Advanced Research Fel- able organic synthesis. search with Professor William lowship and appointed to a Lec- © Georg Thieme Verlag Stuttgart · New York — Synthesis 2015, 47, 3079–3117 3081 Syn thesis S. E. Mann et al. Review sation is thought to be mediated by palladium.3 In combina- tion of reaction products which apparently derived from tion with the formation of regioisomer 6b in the initial nuc- different mechanistic pathways (Scheme 2). Product 11 ap- leopalladation process, path 2 can lead to the formation of pears to be the result of either a linear-selective allylic oxi- up to four different possible oxidation products. β-Hydride dation (path 1) or a Wacker-type oxypalladation (path 2) elimination generates an HPdX species which can readily with regioselective β-hydride elimination away from the undergo reductive elimination to extrude HX, and re-oxida- acetate. In contrast, product 12, derived from sterically tion to an active palladium(II) species is again necessary in crowded substrate 10, appears to result from a Wacker-type order to complete the cycle. oxypalladation reaction followed by regioselective β-hy- A third oxidation pathway involves further functional- dride elimination towards the acetate group (path 2). The isation of the σ-palladium intermediates 6a,b with an oxi- oxypalladation process seems to occur with opposite regi- dizing agent or trapping reagent to generate a new σ-bond oselectivity to a typical Wacker oxidation (which usually at the carbon where the palladium is bound (path 3), giving gives methyl ketones from terminal alkenes), presumably 1,2-difunctionalisation product 8. Often, this final pathway due to the presence of the sulfur ligand in the substrate leads to the regeneration of a palladium(II) species directly, which directs the palladium towards the more hindered in- so no further oxidation of the catalyst is required. Again, a ternal carbon atom. In the case of other substrates, mix- number of isomeric products can be obtained, depending tures of the two classes of products were obtained and on both the regioselectivity and stereoselectivity of the ini- these were demonstrated not to isomerise under the reac- tial nucleopalladation reaction and the mechanism of the tion conditions. subsequent oxidative functionalisation (retention or inver- sion of the stereochemistry of the carbon–palladium bond). AcOH Pd(II) From the three different pathways discussed above and S S S S OAc BQ depicted in Scheme 1, it can be seen that the mechanism of R MnO2 R a particular reaction can often be far from obvious. For ex- 9a R = Me DMSO 11a R = Me, 68% (E/Z 10:1) 9b R = H 11b R = H, 63% (E/Z 4.3:1) ample, the product of a formal allylic oxidation with alkene transposition 4 can be obtained from two different reaction AcOH OAc pathways (path 1 and path 2), and the situation can be com- Pd(II) S S S S plicated further by the fact that alkenes in both the prod- R BQ R ucts and the starting materials can undergo isomerisation MnO2 10a R = H DMSO 12a R = H, 58% (E/Z 1:1.2) under the reaction conditions. 10b R = Me 12b R = Me, 30% (E/Z 1:1.4) This review covers recent developments in the palladi- Scheme 2 Formation of different products from the palladium(II)- um(II)-catalysed oxidation of alkenes via the three path- mediated oxidation of closely related unsaturated thioacetals ways shown in Scheme 1, with a focus on intermolecular reactions that involve the formation of a new carbon– The dithiane group is a good bidentate ligand for palla- oxygen or carbon–nitrogen bond. dium(II) and readily coordinates to palladium(II) acetate. This leads to the formation of a well-defined complex in 1.3 Observation of Reaction Intermediates which there is a remarkable downfield shift of the nearby axial proton on the dithiane ring (Scheme 3). In computa- The direct observation of the intermediates 2, 3 and 6 is tional calculations, this effect was attributed to the close possible in many cases, and it is clear that the different re- proximity of this proton to the acetate ligands on palladi- action pathways by which a particular alkene undergoes um. Reaction of thioacetal 9a with palladium(II) acetate led oxidation are often finely balanced. Notably, the incorpora- first to the formation of an isomeric mixture of complexes tion of ligands for palladium into the substrate can lead to 13a and 13b, which gradually converted into the single π- stabilization of both π-allyl–palladium complexes and σ- allyl complex 14. This in turn was gradually converted into palladium complexes. π-Allyl–palladium complexes can the allylic oxidation product 11a, showing that this com- also be isolated in the absence of a suitable trapping nu- pound is formed exclusively via path 1. With substrate 10a, cleophile.4 Sheppard and co-workers have studied the pal- initial formation of the thioacetal complex 15 was stereose- ladium-catalysed oxidation of unsaturated thioacetals,5 lective, but the probable σ-organopalladium intermediate which provides a useful framework for direct observation 16 was not observed, with direct formation of the oxidation of the various organopalladium intermediates involved in product 12a being seen by NMR. However, it should be not- alkene oxidation. This work also gives a good illustration of ed that a Wacker-type oxidation (path 2) is the most plausi- the fine balance between the different reaction pathways. ble mechanism for the formation of this compound. It was, Oxidation of closely related substrates 9 and 10 using a pal- however, possible to obtain unusually stable σ-alkylpalladi- ladium(II) catalyst in the presence of p-benzoquinone (BQ) um complexes 18a,b by reaction of truncated substrates and manganese(IV) oxide as co-oxidants led to the forma- 17a,b with palladium(II) acetate. Interestingly, in complexes © Georg Thieme Verlag Stuttgart · New York — Synthesis 2015, 47, 3079–3117 3082 Syn thesis S. E. Mann et al. Review 18, the new carbon–palladium bond was formed with com- 2 Allylic Oxidation plete diastereoselectivity. These compounds are remarkably resistant to β-hydride elimination, but can undergo further With an abundance of alkenes available as starting ma- reactions. The carbon–palladium bond can be function- terials, methodologies that enable the synthetic chemist to alised with concomitant dithiane removal by treatment transform simple hydrocarbons into more complex hetero- with a strong oxidizing agent, providing a stepwise demon- atom-containing compounds are highly desirable. Oxida- stration of a 1,2-alkene difunctionalisation reaction (path tion at the activated allylic position is one approach that 3). A potential reason for the different reactivity observed has been heavily exploited and provides a reliable method with these substrates is the conformation of the six-mem- for incorporating heteroatoms into aliphatic building bered dithiane ring, and the orientation of the alkene-con- blocks. The resultant allylic alcohols and amines have a va- taining chain. In the case of 10a,b, the bulky gem-dimethyl riety of applications in synthesis, ranging from simple group forces the alkene-containing chain to occupy an enone precursors to complex substrates for the Tsuji–Trost equatorial position on the dithiane ring, which appears to reaction6 and Sharpless asymmetric epoxidation.7 As well promote direct oxypalladation. However, an axial alkene as being useful synthetic intermediates, allylic alcohols and group (complex 13b) appears to promote π-allyl formation amines are common structural features found in a number (14) which is able to out-compete direct oxypalladation of of biologically active natural products and medicines, mak- 13a in a Curtin–Hammett-type equilibrium for substrate ing them attractive targets in their own right. There are 9a. many methods available for the allylic oxidation of alkenes,8 The above observations serve to illustrate the fine bal- including the use of stoichiometric quantities of toxic chro- ance between the competing oxidation pathways: reaction mium(VI) reagents8a,d or malodourous organoselenium of a simple alkene with palladium(II) acetate can proceed compounds,8d,e causing environmental and safety concerns. either via direct formation of a π-allyl complex or through The Kharasch–Sosnovsky reaction,8f,g which offers a non- oxypalladation of the alkene, depending on the fine struc- toxic copper-mediated alternative, has also been widely ture of the substrate. As will be seen in the discussion that used to achieve this useful transformation.8b,c Whilst reli- follows, this can complicate the development of new reac- able, these traditional approaches tend to be incompatible tion manifolds, as subtle factors can affect which pathway with many functional groups, causing undesired oxidation is followed. of alcohols, at benzylic positions and adjacent to heteroat- oms and carbonyl groups. As a result, the use of suitable protecting groups is often required, which involves addi- S S 13a tional steps and increases the amount of waste generated in Pd HδH 4.58 S the synthesis. AcO OAc S The use of palladium(II) salts for the allylic oxidation of 9a AcO OAc H Pd OAc 11a alkenes is a comparatively new process that provides a Pd H δH 4.52 H much milder, more general approach to traditional meth- S 14 ods and is compatible with a wider range of functional S groups. Although it was first documented as early as the 13b 1960s, the power of this transformation was not fully real- ized until relatively recently. As a result, the field has re- H ceived much attention in recent years. However, to the best H S of our knowledge, the palladium-mediated allylic oxidation S S reaction has yet to be reviewed in any detail. This review 10a S 12a Pd OAc article therefore covers some of the key early break- Pd AcO 16 AcO OAc not observed throughs as a prelude to the contemporary research in this field. 15 Ar Ar S S S 2.1 Background S Pd 10 min AcO Pd AcO The palladium-mediated allylic oxidation of alkenes OAc AcO 2 17a Ar = Ph 18a Ar = Ph dates back to the 1960s (Scheme 4), when Mosieev first ob- 17b Ar = 2-naphthyl 18b Ar = 2-naphthyl served the formation of trace amounts of allyl acetate in the Scheme 3 Observation of divergent reaction mechanisms in the palla- reaction between propene and stoichiometric palladium dium(II)-mediated oxidation of closely related unsaturated thioacetals chloride in the presence of sodium acetate and acetic acid (Scheme 4, a).9 At around the same time, Winstein demon- strated that cyclic alkenes were converted largely into the corresponding allylic esters (Scheme 4, b) and higher termi- © Georg Thieme Verlag Stuttgart · New York — Synthesis 2015, 47, 3079–3117 3083 Syn thesis S. E. Mann et al. Review nal alkenes gave a larger proportion of the corresponding Pd(OAc)2 (0.5 mol%) OAc allylic acetates than propene, using palladium acetate as ox- 95% idant (Scheme 4, c).10 Internal linear alkenes, however, gave BQ, MnO2 AcOH, 60 °C almost exclusive conversion into branched allylic esters (Scheme 4, d), constituting the first examples of selective OAc OAc allylic oxidation with palladium. ( )n PdCl2, NaOAc (a) OAc n = 1, 3, 4, 8; 60–90% 87% (mixture of regioisomers) AcOH OAc OAc M1o9s6ie0ev Scheme 5 First example of the catalytic allylic oxidation of alkenes us- ing a palladium(II) salt trace Pd(OAc)2 OAc Winstein (b) 1963 idation of alkenes could also be extended to the preparation AcOH of higher allyl carboxylates, as demonstrated by Åkermark and co-workers.12d Replacement of acetic acid in the reac- (c) Pd(OAc)2 OAc tion mixture with two equivalents of a higher carboxylic OAc 9% acid gave rise to the corresponding cyclohexene allyl esters AcOH 80% OAc in good yields (Scheme 6). Benzoates, cinnamates and even 9% Winstein a bulky pivalate could be obtained under these conditions. 1966 One example of a chiral carboxylate was also successfully (d) Pd(OAc)2 OAc R = Me, 97% employed in this reaction, although the diastereoselectivity R R R = Et, 92% was very modest. Nevertheless, this opened up the possibil- AcOH ity for asymmetric induction in future developments. This Scheme 4 Early examples of palladium-mediated allylic oxidation of reaction was also able to employ tert-butyl hydroperoxide simple alkenes as a low-cost stoichiometric oxidant. Whilst at the time it was not clear what the mecha- Pd(OAc)2 (5 mol%) BQ (10 mol%) O R nisms for these particular transformations were, the obser- vation of allyl acetates hinted at the possibility that there tBuOOH, RCO2H O was more than one pathway involved in the palladium- CH2Cl2, 40 °C mediated oxidation of alkenes. As well as the known X Wacker-type 1,2-nucleopalladation (Scheme 1, path 2), the X = H; 77% O O Ph observed allyl acetates could also feasibly have formed via X = NO2; 82% X = OMe; 85% O O nucleophilic attack by acetate on a π-allyl palladium com- 89% plex (Scheme 1, path 1). As mentioned above, the precise OAc mechanism of these transformations is often unclear, as the O O same alkene can sometimes undergo oxidation via more O O than one different pathway and alkene isomerisation can 65% 82% (~60:40 dr) also take place. Scheme 6 Formation of higher allyl esters in the palladium-mediated Despite these early advances, it was not until much later oxidation of cyclohexene that a synthetically useful catalytic version of this reaction was developed. Heumann and Åkermark first reported the catalytic allylic oxidation of cyclohexene using an intricate Larock took this idea one step further and developed an three-component oxidation system comprising palladium elegant intramolecular oxidative allylic cyclisation of alke- acetate, p-benzoquinone and manganese(IV) oxide (Scheme noic acids that gave rise to a range of aliphatic lactones 5).11 A range of simple cyclic and internal linear alkenes (Scheme 7) using molecular oxygen as the stoichiometric were also efficiently oxidized to their corresponding allyl oxidant.12e Both five- and six-membered mono-, bi- and acetates under these conditions. even spirocyclic lactones were formed in excellent yields by A number of reports followed this breakthrough, ex- this procedure. As these species had previously only been tending the scope of the reaction to include alternative co- accessible via two-step halolactonisation–dehydrohaloge- oxidant systems, different palladium(II) sources and ever nation processes, there were clearly advantages in develop- more complex substrates.12 The palladium(II)-mediated ox- ing this chemistry. © Georg Thieme Verlag Stuttgart · New York — Synthesis 2015, 47, 3079–3117 3084 Syn thesis S. E. Mann et al. Review O R3 Pd(OAc)2 (5 mol%) OAc R3 R1 Pd(OAc)2 (5 mol%) BQ (10 mol%) ( )n O NaOAc O ( )m R1 R4 R1 R4 R2 R2 H2O2, AcOH R2 ( )m OH DMSO, O2 R2 ( )n 13 examples R1 OAc n, m = 1, 2; 15 examples (a) nBu nBu nBu O O O O O O H AcO H 71% 81% 86% 90% O O H H O O O O ( )3 79% ( )3 (b) iPr iPr 91% 87% 82% H H Scheme 7 Intramolecular allylic oxidation to form lactones H H H H AcO AcO OAc However, some substrates still behaved unpredictably, 20% (mixture of stereoisomers) giving rise to a variety of different oxidized products, and a (c) OAc general procedure for allylic oxidation remained elusive. 67% For instance, Åkermark obtained several different products AcO in the palladium acetate/hydrogen peroxide mediated oxi- (d) nBu nBu dation of a range of alkenes, depending on the nature of the 46% O starting material (Scheme 8).12d As demonstrated previous- OAc ly, internal linear and cyclic alkenes gave rise to the expect- ed allyl acetates in good yield (Scheme 8, a). Even a complex (e) tetracyclic cholesterol derivative underwent the desired transformation, albeit in much lower yield than the simpler 60% alkenes (Scheme 8, b). Oxidation of cyclohexa-1,3-diene Scheme 8 Different products obtained in the palladium-mediated oxi- gave cyclohexene-1,4-diacetate, which has been postulated dation of a variety of alkene substrates to proceed via 1,2-acetoxypalladation to give a π-allyl inter- mediate, followed by subsequent reaction with acetate (Scheme 8, c).13 Terminal alkenes, on the other hand, under- with high E/Z selectivity. Palladium acetate bis(sulfoxide) went exclusive 1,2-oxypalladation to give Wacker products complex 19, on the other hand, furnished branched allyl (Scheme 8, d). With cis-1,2-divinylcyclohexane (a 1,5-di- acetates in a similarly regioselective fashion. ene) the reaction proceeded via 1,2-acetoxypalladation as The bis(sulfoxide) ligand was shown to partially decom- with other terminal alkenes, although in this case the palla- pose under the reaction conditions to generate phenylvinyl dium–carbon σ-intermediate then underwent an intramo- sulfoxide.15 This commercially available compound could lecular Heck reaction to give a bicyclic compound (Scheme also be employed successfully in place of the bis(sulfoxide) 8, e). and it is likely to be the active ligand in this process. This Whilst the diversity of products available from this was the first example of a general, synthetically useful chemistry was intriguing, the lack of a more general meth- methodology for the selective formation of allyl acetates odology for palladium-mediated allylic oxidation still rep- from terminal alkenes. Only very small quantities of the resented a significant challenge. In order for the potential of corresponding methyl ketones were observed under these this reaction to be fully realized, a generic procedure that conditions, making this a valuable complimentary ap- could predictably furnish allyl acetates from alkenes was proach to the Wacker oxidation. necessary. This breakthrough came in the form of seminal As a consequence of this and other recent developments work by the White research group, who developed a sulfox- described herein, the palladium(II)-mediated allylic oxida- ide-promoted method for the regioselective synthesis of al- tion of alkenes has become a powerful synthetic tool, re- lyl acetates from monosubstituted alkenes.14 Depending on ceiving much attention in recent years. the conditions employed, terminal alkenes could be selec- tively oxidized to give either linear or branched allyl ace- 2.2 Allylic Oxygenation tates (Scheme 9). Palladium acetate in a mixture of dimeth- ylsulfoxide and acetic acid (1:1) regioselectively gave the In the decade or so since the publication of their origi- corresponding linear allyl acetates in reasonable yield and nal paper, the White group has elaborated upon the scope of their sulfoxide-promoted reaction with some elegant ap- © Georg Thieme Verlag Stuttgart · New York — Synthesis 2015, 47, 3079–3117 3085 Syn thesis S. E. Mann et al. Review O O the White group have described successful alternative Ph S S Ph(10 mol%) routes to a variety of synthetic targets, which circumvent Pd(OAc)2 (10 mol%) 19 P(OdAc)2 OAc traditional olefination reactions. A representative selection R OAc R of examples is outlined in Scheme 11. In the first case BQ, 4 Å MS R DMSO–AcOH (1:1) BQ, AcOH (Scheme 11, a), a linear allylic oxidation was carried out on dioxane a differentially protected diol, which is readily prepared 10 examples 13 examples 50–65% 56–83% from protected glycidol in two steps.16e The resulting linear >10:1 E/Z >16:1 branched/linear allyl benzoate is a key intermediate in the synthesis of (–)- >11:1 linear/branched laulimalide, which was previously synthesized by olefina- OAc selected tion in six steps.18 A second example using a similar strate- TBDPSO OAc maptacihrsed TBDPSO gy is given by the three-step synthesis of a linear allyl ben- zoate, a precursor to trans-fused polycyclic ethers present 50%; 31:1 linear/branched 72%; 1:16 linear/branched E/Z = 11:1 in the brevetoxin scaffold (Scheme 11, b).19 The efficient OAc three-step synthesis from commercially available cyclohex- H H N O N O ene oxide using allylic oxidation chemistry represents a sig- Ph OAc Ph nificant improvement over the previously reported six-step O O olefination route. Finally, anti-1,4-dioxan-2-ones were 64%; 13:1 linear/branched 64%; 1:26 linear/branched shown to be highly versatile intermediates that can be E/Z = 12:1 readily elaborated into syn-1,2-diols, stereo-defined ami- Scheme 9 Tunable regioselectivity in the sulfoxide-promoted allylic no-polyols and syn-pyrans (Scheme 11, c).16c By tethering oxidation of terminal alkenes the reacting carboxylic acid to a homoallylic alcohol, a car- boxylate-chelated inner-sphere allylic functionalisation re- plications of the chemistry.16 In 2006, they demonstrated action was enabled, giving access to anti-1,4-dioxan-2- the ability to rapidly access complex polyoxygenated scaf- ones. The example shown illustrates the utility of this pro- folds from bulk chemical starting materials with a short, de cess, shortening the previously reported route to the syn- novo synthesis of differentially protected L-galactose pyran intermediate used for preparing goniodomin A.20 The (Scheme 10).16i A homoallyl acetonide was subjected to pal- 1,4-dioxan-2-one, prepared by allylic oxidation, undergoes ladium(II)-mediated allylic oxidation using p-methoxyben- an Ireland–Claisen rearrangement to give the desired syn- zoic acid as nucleophile to give the corresponding linear al- pyran intermediate, which is isolated as its methyl ester in lylic ester in good yield and excellent regio- and stereose- excellent yield. This improved sequence is three steps lectivity. shorter than the previously reported approach based on traditional olefination chemistry. OBn Perhaps the most impressive achievement of this chem- istry is its successful employment in the total synthesis of O O [Pd(MeCN)4](BF4)2 (10 mol%) OBn 6-deoxyerythronolide B (Scheme 12).16f A late-stage intra- PhBQ, 4 Å MS molecular functionalisation of intermediate 20 gave access O O to advanced macrolactone 21a with high diastereoselectivi- DIPEA O OCOPMP PMP OH DMSO–CH2Cl2 (3:1) 75% yield ty for the natural epimer. This intermediate was then readi- PMP = 4-MeOC6H4 5 steps >300:1 9lin7e:3a rE/b/Zranched ly elaborated to give the natural product in three additional steps. Addition of fluoride ion to π-allyl palladium com- OTBS plexes is known to promote π–σ–π interconversion through BnO OTBS occupation of a coordination site on the metal centre.21 It dfifferentially protected was postulated therefore that addition of fluoride to the re- L-galactose HO O OH action mixture could alter the stereochemical outcome by Scheme 10 De novo synthesis of differentially protected L-galactose by disrupting the palladium–carboxylate chelate through highly regioselective linear allylic oxidation which 21 is presumably formed. Pleasingly, the addition of tetra-n-butylammonium fluoride (TBAF) produced the de- Taking advantage of the unique ability of p-methoxy- sired effect, resulting in a 1:1.3 mixture of the natural ma- benzoates of (E)-but-2-ene-1,4-diols to undergo highly terial 21a and its epimer 21b. diastereoselective asymmetric dihydroxylation,17 the re- The significance of this ability to selectively furnish ei- searchers rapidly completed the enantioselective synthesis ther one of the two epimers was highlighted when a syn- of the hexose framework. thesis of the same intermediates was attempted using more The use of allylic oxidation can offer a more efficient al- traditional chemistry. Yamaguchi macrolactonisation of al- ternative to traditional carbon–carbon bond-forming lylic alcohol 22a (prepared via a modification of the allylic chemistry. For example, several recent publications from © Georg Thieme Verlag Stuttgart · New York — Synthesis 2015, 47, 3079–3117 3086 Syn thesis S. E. Mann et al. Review 1) AllylMgCl, THF PMP (a) O –20 °C to r.t. OPMB TBSO TBSO O O O O O 2) PMB-X, TfOH 19 Et2O, r.t. (P3h eCqOu2ivH) [Pd(4DM ÅMe CSMNOS),,4 D]C(BIHPF2EC4A)l22, ,(P 14h01B m°QCol%) (B01Q.3:4,1 4eT.3%qB udAirvF) HO 20 1>94 (5006:1.B%3 Qd e.rq.uiv) OPMB O PMP PMP (–)-laulimalide TBSO O Ph O a) 19 (0.3 equiv) O oxidation route: 3 steps, 57% ArCO2H, BQ b) LiOOH olefination route: 6 steps, 53% c) K2CO3, MeOH 1) BrMg ( )2 O 21a O O (b) Et2O, 0 °C to r.t. OR O O O O 2) KHMDS, TBAI, R-Br ( )2 O O O O THF, 0 °C 21b PMP (A4r CeqOu2iHv) [Pd(4DM ÅMe CSMNOS),,4 D]C(BIHPF2EC4A)l22, ,(P 14h01B m°QCol%) HO O O O O O OH 3 steps Yamaguchi O O OR Yamaguchi 87% O 22a PMP O Ar 6-deoxyerythro- I ( )2 O Ar X O O O O O OH 2n2o lsidteep Bs precursor to trans-fused polycyclic oxidation route: 3 steps, 47% HO 7.8% overall ethers in brevetoxins olefination route: 6 steps, 19% 22b 1:1 (separated by chromatography) O O CO2H Ph S S Ph (10 mol%) O Scheme 12 Total synthesis of 6-deoxyerythronolide B by late-stage (c) Pd O C–H oxidation O 19 (OAc)2 O O O Cr(salen)Cl, BQ O O dioxane, 65 °C approach to the total synthesis of unnatural iminosugar (–)- 73%, 3:1 dr castanospermine (Scheme 13).22 In this instance, the Cbz 1) LiHMDS, TMSCl 2) MeI, K2CO3 nitrogen protecting group provides the nucleophilic oxygen Et3N, THF, –78°C DMF, r.t. then reflux, PhMe 2 steps; 82% source, which cyclises onto the allylic position in the pres- ence of the White catalyst and a Lewis acid. The resulting oxazolidinone was isolated as a single diastereomer and goniodomin A O H O H CO2Me was readily elaborated to give (–)-castanospermine. The O use of simple, readily available carbamates as nucleophiles oxidation route: 7 steps in this reaction is unprecedented and represents an exciting olefination route: 10 steps new prospect for future developments. The fact that this re- Scheme 11 Allylic oxidation as a more efficient alternative to olefina- action does not proceed in the presence of molecular sieves tion in natural product synthesis suggests that hydrolysis of the intermediate cationic cyclic carbamate is a key step in this process that drives the reac- oxidation chemistry) furnished the expected natural epi- tion to completion. mer 21a in excellent yield. However, exposure of the oppo- site diastereomer 22b to the same conditions yielded only O O Ph S S Ph O oligomers, demonstrating that macrolactone 21b is not ac- O OBn Pd O cessible via the common traditional approach. By using a N 19 (OAc)2 (10 mol%) N tunable late-stage allylic oxidation strategy rather than a H macrolactonisation approach, the researchers were able to BnO OBn dBioQx,a Yneb,( O75T f°)C3 BnO OBn circumvent this issue and access both the naturally occur- OBn OBn 71% ring compound and its unnatural epimer. A similar strategy was later employed to complete the total synthesis of N OH H erythromycin.16d HO OH In more recent years, a number of other research groups OH have begun to take advantage of the high chemo- and regio- (–)-castanospermine selectivity that can be achieved using palladium acetate bis(sulfoxide) complex 19 (generally referred to as the Scheme 13 Total synthesis of (–)-castanospermine using intramolecu- lar allylic oxidation as the key step White catalyst). For example, Malik recently applied this © Georg Thieme Verlag Stuttgart · New York — Synthesis 2015, 47, 3079–3117 3087 Syn thesis S. E. Mann et al. Review Similarly, the use of the bis(sulfoxide) palladium cata- (a) lyst 19 to promote an intramolecular oxidative allylic cycli- Pd(OAc)2 (5 mol%) PhS sation to form biologically relevant tetracyclic flavonoids R L1 (5 mol%) R OAc L1 = O was recently reported by Belani (Scheme 14).23 Until rela- BQ, AcOH 12 examples 23 40 °C 32–78% tively recently, oxygen nucleophiles employed in the palla- (b) Stambuli 2010 dium bis(sulfoxide)-catalyzed process have all been weakly Pd(OAc)2 (5 mol%) alecoidpihci l(etys ptiecnadll yt coa rpbeorxfoyrlmate lse wssi twh eplKl au <n 6d)e.r M tohrees eb arseiacc ntuiocn- R L2 (5 mol%) R OAc L2 = S conditions as they are not readily deprotonated and can BQ, AcOH 10 examples 40 °C 48–87% Stambuli 2013 also deactivate the catalyst. The successful use of phenols (pK ~10) and even aliphatic alcohols (pK ~16)16a as nucleo- philaes represents an important improveament in the scope (c) Pd(OL3A (c5)2 m (5o lm%o)l%) L3 = O R R OAc of this reaction. AcOH, NaOAc 11 examples O O O2, dioxane, 60 °C 52–84% N 24 N Ph S S Ph (10 mol%) Stahl 2010 R1 O R2 19 P(OdAc)2 R1 O R2 R(d) NPadOCAl2c ,( 16 matoml% O)2 R OAc Kaneda 2006 OH BQ, AcOH O DMA–AcOH, 80 °C 8 examples 56–90% O CH2Cl2, 40 °C O Scheme 15 Catalyst systems for linear-selective allylic oxidation 7 examples 60–90% Scheme 14 Tetracyclic flavonoid synthesis by intramolecular oxidative allylic cyclisation sults in the formation of branched products. This develop- ment opens up the possibility for ligand-derived stereo- Several other catalytic systems have been developed for chemical induction if the Pd/sox catalyst system could be the selective formation of either linear or branched allylic suitably modified. Notwithstanding, at the time of writing, oxidation products (Scheme 15). In 2010, Stambuli report- this catalyst system is the only alternative to the White cat- ed the linear-selective allylic oxidation of terminal alkenes alyst for selective formation of branched allylic esters. using thioether ligand 23 (Scheme 15, a).24 The scope and functional group tolerance of this reaction were compara- Pd(OAc)2 (10 mol%) O L (5 mol%) ble to the procedure first published by White, although the O R2 R1 reaction times tended to be slightly shorter in this case. In a BQ, R2CO2H R1 more comprehensive survey of palladium–sulfide catalysts, dioxane 45 °C 11 examples Stambuli later demonstrated that the simple and inexpen- 51–86% yield F sive tetrahydrothiophene was a highly active and linear-se- lective ligand (Scheme 15, b).25 In an effort to move away O L = from the use of stoichiometric quantities of oxidants such S+ N as p-benzoquinone or copper additives, Stahl developed the –O 25 tBu first catalytic system to achieve aerobic turnover with 4,5- diazafluorenone ligand 24 (Scheme 15, c).26 This system Scheme 16 A Pd/sox catalyst system for the branch-selective synthesis of allylic esters proved to be highly selective for linear allylic acetates with good yields and functional group tolerance. An earlier ex- ample of a direct oxygen-coupled allylic oxidation was re- As mentioned above, allylic alcohols and esters are com- ported by Kaneda (Scheme 15, d) alongside their procedure mon starting materials in a variety of different reactions. A for the Wacker oxidation (vide infra).27 This method uses methodology that enables the one-pot conversion of simple only 1 mol% of palladium catalyst but a high pressure of ox- alkenes into intermediate substrates that can then undergo ygen is necessary for efficient catalyst turnover. With mo- further manipulations would therefore be a powerful syn- lecular oxygen as the sole re-oxidant, these systems clearly thetic tool for rapidly accessing complex molecular archi- offer an environmental advantage over other approaches to tecture. Hartwig demonstrated the potential of this ap- allylic oxidation. proach by employing a one-pot allylic oxidation–enantiose- More recently, Liu has reported the use of a Pd/sox cata- lective functionalisation reaction (Scheme 17).29 Taking lyst system that is highly selective for branched allylic es- inspiration from the success of the one-pot iridium-cata- ters.28 The bulky oxazoline of sox ligand 25 (Scheme 16) lysed borylation–Suzuki coupling of aromatic carbon– hinders functionalisation at the terminal carbon atom, fa- hydrogen bonds,30 the research group sought to develop a vouring oxidation at the internal allylic position, which re- method for iridium-catalysed allylic functionalisation in © Georg Thieme Verlag Stuttgart · New York — Synthesis 2015, 47, 3079–3117
Description: