University of Groningen Palladium catalyzed carbon-carbon bond formation under reductive, oxidative and redox neutral conditions Gottumukkala, Aditya Lakshmi Narasimha Raju IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Publication date: 2012 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Gottumukkala, A. L. N. R. (2012). Palladium catalyzed carbon-carbon bond formation under reductive, oxidative and redox neutral conditions. Groningen: s.n. Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date: 21-07-2020
Transcript
University of Groningen
Palladium catalyzed carbon-carbon bond formation under reductive, oxidative and redoxneutral conditionsGottumukkala, Aditya Lakshmi Narasimha Raju
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.
Document VersionPublisher's PDF, also known as Version of record
Publication date:2012
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):Gottumukkala, A. L. N. R. (2012). Palladium catalyzed carbon-carbon bond formation under reductive,oxidative and redox neutral conditions. Groningen: s.n.
CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.
The work described in this thesis was performed at the Stratingh Institute for Chemistry, University of Groningen, The Netherlands.
The research work described in this thesis was performed within the framework of the CatchBio Program (Project No. 053.70.206). The authors gratefully acknowledge the support of the SmartMix Program of the Netherlands Ministry of Economic Affairs and the Netherlands Ministry of Education, Culture and Science.
Cover designed by Jeffrey Bos and Aditya Gottumukkala. Made real by Jeffrey Bos.
Chapter 2: Conjugate addition versus Mizoroki-Heck reaction of aryl iodides 21
2.1 Introduction 22
2.2 Goal 25
2.3 Results and discussion 25
2.4 Conclusions and future perspectives 41
2.5 Experimental 43
2.6 References 54
Chapter 3: Pd-BIAN: A highly selective catalyst for the base-free oxidative Heck reaction 57
3.1 Introduction 58
3.2 Goal 66
3.3 Results and discussion 66
3.4 Mechanistic insights 74
3.5 Conclusion and future perspectives 75
3.6 Experimental 77
3.7 References 88
Chapter 4: Palladium-catalyzed benzylic quaternary stereocenter formation via conjugate addition to cyclic enones 91
4.1 Introduction 92
4.2 Goal 95
4.3 Results and discussion 95
4.4 Synthesis of (–)- -cuparenone 106
4.5 Conclusion and future perspectives 106
4.6 Experimental 108
4.7 References 122
Chapter 5: Facile construction of benzylic quaternary centers via palladium catalysis 125
5.1 Introduction 126
5.2 Goal 129
5.3 Results and discussion 129
5.4 Summary and conclusions 138
5.5 Experimental 140
5.6 References 152
Chapter 6: Palladium-catalyzed enantioselective conjugate addition of arylboronic acids to acyclic enones 155
6.1 Introduction 156
6.2 Goal 157
6.3 Results and discussion 157
6.4 Summary and conclusions 171
6.5 Future perspectives 171
6.6 Experimental 173
6.7 References 181
Chapter 7: Studies toward an enantioselective Matsuda-
Heck reaction 183
7.1 Introduction 184
7.2 Goal 187
7.3 Results and discussion 187
7.4 Mechanistic insights and the Meerwein arylation 197
7.5 Recent developments 199
7.6 Summary and conclusions 199
7.7 Experimental 200
7.8 References 211
Summary 215
Samenvatting 221
Résumé 227
Acknowledgements 233
Chapter 1
Palladium-catalyzed Heck and conjugate addition reactions
In this chapter, a concise overview of the different types of palladium catalyzed Heck reactions, conjugate additions and their interplay is presented. Recent developments in the Pd-catalyzed synthesis of benzylic quaternary centers are discussed, along with an outlook over the remainder of the thesis.
Chapter 1
2
“You would be forgiven if you thought the most important element in an organic transformation was carbon. …….. for just over half a century in many of chemistry's most renowned organic reactions, it has actually been palladium.” 1
1.1 Introduction
The formation of carbon-carbon bonds is central to the study of organic chemistry. In what may be regarded as an overly-simplified view, the linkage of a carbon nucleophile with a carbon electrophile, is, in fact, an encompassing multitude of reactions which allow the linking of molecules with a variety of conditions, positions and functions. One might argue that Synthetic Organic Chemistry, in the terms we understand it today, rests squarely on this set of reactions, and in particular, the ability to engage at will, a reaction to afford the desired transformation. Thus, the endeavor to develop a new carbon-carbon bond forming reaction remains an evergreen enterprise in organic chemistry.
In recent decades, the formation of carbon-carbon bonds by transition metal catalysis2,3 has gained much importance. Among the various transition metals applicable for C-C bond formation, Pd holds a particularly prominent place.1 Its significance in synthesis is not limited to carbon-carbon bond formation, and is documented in several well renowned treatises.4-6
1.2 The Heck reaction
While learning about the mechanistic details of the, then recently-developed, Wacker oxidation7,8 from colleague Patrick Henry9 at Hercules Inc., Richard F. Heck wondered what would happen if an organo Pd-species lacking a -hydrogen atom, would react with an olefin. The very first reaction (Scheme 1) Heck set up, “to see what would happen”, was a thumping success.10 This sparked a detailed investigation into the reaction, which resulted in seven consecutive, single authored communications11-17 in 1968.
Palladium-catalyzed Heck and conjugate addition reactions
3
Scheme 1: The first olefination experiment performed by Heck.
The success and importance of the reaction was quickly recognized, and the laboratory of Tsutomu Mizoroki reported18,19 that toxic and hazardous arylmercury salts could easily be replaced by aryl iodides, making the reaction much more convenient to handle and operate. In the decades that followed, several important developments took place, such as the development of a catalytic version (what is now known as the Mizoroki-Heck reaction), and more favorable conditions for application in organic synthesis (Scheme 2). The developments thenceforth are numerous and outside the scope of this chapter. A number of excellent reviews20-22 and a recent monograph23 are dedicated to this.
Scheme 2: General representation of a Heck reaction.
1.3 The oxidative Heck reaction
The oxidative variant of the Heck reaction involves PdII, formed by oxidation of Pd0 with an external oxidant, in the catalytic cycle. Thus, instead of oxidative addition to an aryl halide (as in the case of the Mizoroki-Heck reaction), there is a transmetallation (no change in the oxidation state of Pd) of an organometallic reagent such as a boronic acid, a siloxane, etc. This seemingly simple mechanistic variation results in a host of properties that set the oxidative Heck reaction apart, in its own right. By avoiding oxidative addition to halides, an endergonic process, the oxidative Heck reaction can be performed even at room temperature. Heck is credited for disclosing the first example of an oxidative Heck reaction (Scheme 3, but in fact also Scheme 1). Pivotal contributions by Larhed24-32 and Jung,33-36 among others,37-39 have made this reaction the highly versatile transformation it is today. Recent studies by Sigman,40 have shown that the reaction can be tuned to yield (E)-styrenyl products in high yields and selectivities, for a series of linear, electronically unbiased olefins.
Chapter 1
4
Scheme 3: The first example of an oxidative Heck reaction.
The choice of the oxidant is crucial for the success of the reaction. Initial studies focused on the use of CuII salts. However, compelling advances in the area came with the use of oxygen as the oxidant41 (aerobic oxidative Heck), avoiding stoichiometric waste at the end of the reaction. An important consideration in this connection is that electron-rich phosphines usually cannot be applied as ligands in the reaction, as these are oxidized. Instead, nitrogen ligands such as 2,2’-bipyridine; 1,10-phenanthroline; and 2,9-dimethylphenanthroline (neocuproine); are employed. The first examples of the reaction in which a ligand was used to stabilize the Pd were reported by Larhed.30
The application of the oxidative Heck reaction in organic synthesis has been the topic of several reviews.42-47 The reaction is particularly successful in the alkenylation and arylation of terminal olefins.35 A variety of boron-based organometallics such as boronic acids,48 borate esters, and trifluoroborates49 of aromatics, heteroaromatics and alkenes have been successfully employed.
Elegant studies by Larhed and coworkers on the regioselective arylation of electron rich olefins29 have established the charged nature of the metal during the reaction (Scheme 4.a). When the metal is cationic (as drawn), the arylation is selective at the -position. DFT calculations of the intermediate (Scheme 4.b) are in agreement with the observed experimental results. Recently, Stahl and coworkers have shown that this selectivity is not only based on electronic parameters, but also influenced by the steric parameters of the ligand.50
Scheme 4: a) Regioselectivity of oxidative Heck reaction with cationic Pd species b) DFT
calculated structure prior to olefin insertion.
Palladium-catalyzed Heck and conjugate addition reactions
5
While bearing in mind the success of the reaction, it is important to point out the drawbacks while employing it for internal alkenes. In addition to the expected Heck product, the conjugate addition product, phenol and biphenyl are regularly observed.35 These issues are addressed further in Chapter 3.
1.4 The Matsuda-Heck reaction
Following the seminal discoveries of Heck11-16 and Mizoroki,18,19 the group of Tsutomu Matsuda reported that aryldiazonium salts are highly effective for the same coupling reaction (Scheme 5).51 In a series of publications,52-66 the group laid the foundation for what is known today as the Matsuda-Heck reaction,67,68 exploring its scope, reaction parameters and application.
Scheme 5: A Matsuda-Heck reaction.
The application of aryldiazonium salts in Pd-catalyzed reactions has several advantages:
i) Dinitrogen, being an excellent leaving group, makes the oxidative addition of aryldiazonium species to Pd0 facile, even at room temperature.
ii) Loss of dinitrogen renders, the aryl species cationic which, in turn, results in a cationic metal complex
iii) The cationic species formed possesses a vacant coordination site that also facilitates the insertion of the alkene
iv) Aryldiazonium salts can be readily synthesized from the corresponding anilines.
All these attractive features allowed the application of the Matsuda-Heck reaction on an industrial scale. The first industrial process was the synthesis of the herbicide Prosulfuron (Scheme 6) by Syngenta AG (formerly Ciba-Geigy AG).69
Scheme 6: Industrial Synthesis of Prosulfuron.
Chapter 1
6
The choice of the appropriate counterion of the aryldiazonium salt is essential for the reaction, as it has an important influence on its stability. Over the years, aryldiazonium tetrafluoroborates70 have become quite popular and are most commonly employed due to their stability and ease of synthesis,71 while disulfonimides,72 tosylates73 and carboxylates74 have also been reported to be stable for application.
Despite being known for several decades, the enantioselective version of the reaction was lacking. The chief cause for this shortcoming, is that phosphines readily react with the aryldiazonium salts.75 As most enantioselective catalysts use chiral phosphines to make the catalyst asymmetric, this route is ineffective. Very recently, the first example of an enantioselective Matsuda-Heck reaction has been reported. It employs nitrogen-based chiral bisoxazoline ligands, and has been reported for only one substrate (Scheme 7).76
Scheme 7: The first example of an enantioselective Matsuda-Heck reaction.
1.5 A comparison of the different types of Heck reactions
The following is a qualitative comparison of the three different types of Heck reactions presented thus far. The characteristics presented are only those typically reported, and intended as a simplified guide for choosing a reaction to effectuate the transformation.
All three variants of the Heck reaction usually work best for terminal olefins possessing an electronic bias. The reaction temperature for a Mizoroki-Heck reaction varies, typically, between 60 – 140 oC, depending on the nature of the aryl halide, electronics of its substitution and electron-richness of the phosphine employed. The mechanism of the Mizoroki-Heck reaction is the most studied of all the three, and has been found to proceed via different pathways, depending on the substrate and additives.77-81 The aerobic oxidative Heck26 and the Matsuda-Heck82 reaction have been demonstrated to proceed via cationic pathways.
Mizoroki-Heck reactions typically require a base,83 and the choice of the base has a significant influence on the outcome of the reaction.77 The bases most commonly employed are trialkylamines, alkali carbonates and acetates. The addition of base is usually not necessary for oxidative Heck or Matsuda-Heck reactions, particularly,
Palladium-catalyzed Heck and conjugate addition reactions
7
for terminal alkenes. Internal alkenes however often react faster in the presence of a base.
Table 1: A qualitative comparison of the different Heck reactions
As a result of the varying reaction temperatures needed for the reaction, the choice of solvent for the reaction also varies, though in all cases polar solvents are preferred. High boiling solvents such as DMF, NMP, and MeCN are most commonly reported for the Mizoroki-Heck reaction, while THF, MeOH and MeCN are usually applied in oxidative Heck reactions. Alcohols are widely used for the Matsuda-Heck reaction. While external oxidants are not needed for the Mizoroki-Heck and Matsuda-Heck reaction, and can even be detrimental, they are imperative for the oxidative Heck reaction. Consequently, compounds bearing oxidation-sensitive functional groups are not well suited for this reaction. The compatibility of phosphines in a reaction is often an important consideration. While most phosphines are readily compatible with the Mizoroki-Heck reaction, there are only a limited number of reports84 that use phosphines in the oxidative Heck reaction, due to their ease of oxidation under the reaction conditions. Phosphines
reaction temp. 60 -140 oC room temperature room temperature
Solvents polar, high boiling
(DMF, MeCN)
polar, moderate BP
(Alcohols, ethers)
polar, moderate BP
(Alcohols, ethers)
External
oxidant no required no
Phosphines compatible incompatible, with few
exceptions incompatible
Chapter 1
8
are not suitable for the Matsuda-Heck reaction, as they react readily with diazonium salts.75
1.6 Conjugate addition reactions
Transition metal catalyzed conjugate addition reactions of organometallics form an essential tool in modern organic synthesis. Copper and rhodium catalyzed reactions have been well studied over the past decade, and have been reviewed extensively.85-90 While copper is generally most applicable for reactive organometallic nucleophiles91 such as organozinc,92 -magnesium or -aluminum reagents93, rhodium is best suited for less reactive organometallics like aryl and alkenyl borates, boronic acids48 or trifluoroborates.49,94
The field of palladium catalyzed conjugate addition reactions89,90,95,96 is relatively underdeveloped. One of the limiting factors is the competing Heck reaction via -hydride elimination. The desired conjugate addition may be favored in three ways:
i) by the absence or unavailability of a -hydride
ii) by the fast protonolysis of the alkyl-Pd intermediate, avoiding -hydride elimination
iii) by reductive cleavage of the alkyl-Pd intermediate, avoiding -hydride elimination.
Criterion (i) is dependent on the substrate. For criterion (ii) to operate, the combination of a cationic Pd species and water is most suited, while for criterion (iii) a good reducing agent present in excess is beneficial.
Initial reports by Uemura and coworkers on the 1,4-addition of arylantimony97,98 and aryltin99,100 reagents laid the foundation for this area. The first report of an enantioselective conjugate addition reaction with Pd catalysis used triarylbismuth reagents, and Chiraphos-Pd complex 25 (Scheme 8)101. The addition of a stoichiometric amount of Cu(BF4)2·6H2O was found to be necessary as a co-oxidant for the reaction. 0.6 Equivalents of the triarylbismuth reagent was required to achieve full conversion.
Scheme 8: Asymmetric conjugate addition of triarylbismuth reagents.
Palladium-catalyzed Heck and conjugate addition reactions
9
Currently, arylboronic acids have become the reagents of choice for the introduction of an aryl group via transition metal catalysis, on account of their stability, ease of operation, storage and commercial availability.48 The first examples of the enantioselective Pd-catalyzed conjugate addition of arylboronic acids were described by Minnaard et al.102 using Me-Duphos (27) as the ligand. Enantioselectivities up to 99% were obtained for a variety of cyclic enones (Scheme 9).
Scheme 9: First example of the Pd-catalyzed asymmetric conjugate addition of arylboronic acids
As versatile as these reactions are, it must be noted that essentially all the organometallic reagents described in the above transformations are synthesized from their corresponding halides. Thus, it would represent a considerable advance if the aryl halides themselves could be employed for conjugate addition reactions, as it avoids the additional steps for the synthesis of the organometallics.
Surprisingly little interest has been paid to this argument. Only the group of Cacchi103-111 investigated this approach in detail, whereby they observed that the -alkyl-Pd complex (or the palladium enolate) formed during a Heck reaction, could be reduced to afford the formal conjugate addition product (Scheme 10). For this purpose, they found that a combination of a trialkylamine and an acid was most effective.
Scheme 10: Conjugate addition of aryl halides as described by Cacchi et al.
Studies detailed in Chapter 2 of this thesis bring this concept further and point out that the base itself is capable to act as a reducing agent.
Chapter 1
10
1.7 Conjugate addition versus the Heck reaction
The formation of the conjugate addition product as side-product during a Heck reaction112 and, inversely, the formation of a Heck product as side-product during the conjugate addition110,113-115 has been observed in several Pd and Rh catalyzed reactions.116-119 This may be understood as a result of the competition between reductive cleavage and -hydride elimination of the formed -alkyl metal species (Scheme 11).
Scheme 11: Competition between conjugate addition and Heck reaction.
Obtaining either of the products selectively is an important chemoselectivity challenge, for this class of reactions. The selectivity is attributed to different factors in different reactions, such as solvent,120,121 substrate,122 ligands,101,102 and pH.79
Recently, Wu and coworkers probed this competition by DFT calculations123 for isoelectronic RhI and PdII species. They conclude that competition is strongly influenced by the nature of the ligand. Furthermore, the authors conclude that, in case of cyclic substrates, the rate-limiting steps for the Heck type reaction are isomerization of the C-bound metal to align itself for -hydride elimination, and protonolysis for the conjugate addition reaction. It must be noted that the conclusions drawn in the above study are based on the use of arylboronic aids, and the ligands studied were limited to acetic acid, trimethylphosphine and bipyridines. While this provides insight, the results need to be interpreted carefully, especially when considering conjugate addition with aryl halides (Chapter 2) in which there is a change in oxidation state of the metal (Pd0 to PdII) during the reaction.
The selective construction of a compound in which a carbon atom is surrounded by four different carbon substituents (quaternary stereocenters, sometimes referred to as “all carbon” quaternary stereocenters) is an important challenge in organic synthesis.124,125 Though, at first sight, one might be easily misled to believe that this is a simple extension of the conjugate addition reaction to substrates bearing two
-substituents, it is far from it. , -Disubstitution increases the steric hindrance significantly, making the insertion of the alkene into the aryl-metal species more
Palladium-catalyzed Heck and conjugate addition reactions
11
challenging. Furthermore, for the asymmetric version of the reaction, the catalyst must distinguish between two alkyl or aryl substituents (compared to an alkyl/aryl and a H) making enantioselectivity more challenging, as well.
Quaternary stereocenters are a common motif in a variety of natural products (Figure 1), and may be grouped into several classes. For the sake of the following discussion and to put emphasis on the studies presented in this thesis, we limit ourselves to benzylic quaternary stereocenters (a quaternary center at the benzylic position of an aromatic substituent). Until the advent of transition metal catalysis in organic synthesis, there were very few methods that allowed the construction of such benzylic quaternary stereocenters in a selective fashion.
Transition metal catalysis offers the synthetic chemist several reactions to build this motive. Examples of such reactions include -arylation of substituted carbonyls,126-
128 allylic alkylation of aryl bearing substrates,86,129 asymmetric Heck reactions46,130-
132 and conjugate addition to substituted enones.133
Figure 1: Examples of natural products containing quaternary centers. b) Natural products
bearing benzylic quaternary centers.
Chapter 1
12
Of the reactions listed above, conjugate addition to , -unsaturated ketones is a particularly convenient protocol, especially in view of subsequent synthesis (Scheme 12). Upon conjugate addition, one is left with three convenient handles for further synthesis: two regiochemically distinct -positions and the carbonyl moiety. All these entities can be selectively targeted for further functionalization.
Scheme 12: Conjugate addition provides several handles for further modification.
Palladium catalyzed conjugate addition reactions forming quaternary centers are a very recent development. In 2010, Lu reported134 that cationic μ-hydroxo-Pd complex C1 (Scheme 13) was capable of catalyzing the conjugate addition of arylboronic acids to enones.
Scheme 13: Lu’s pioneering example of Pd-catalyzed benzylic quaternary center formation.
Subsequent DFT calculations by Houk135 on the catalyst system presented above, revealed that the insertion of the alkene into the Pd-aryl is the rate-limiting step of the reaction. In addition, calculations revealed that when there is an alkoxy group in the -position, the substitution product (instead of the conjugate addition product) is obtained, due to energetically favored -alkoxy elimination.
During the course of our own studies on the development of an asymmetric Pd-catalyzed quaternary center forming reaction, Stoltz and coworkers reported136 a similar study, using the combination of Pd(O2CCF3)2 and 36 as the catalyst (Scheme 14). While the reaction demonstrated good substrate scope and impressive enantioselectivity, it was limited to cyclic enones.
Palladium-catalyzed Heck and conjugate addition reactions
13
Scheme 14: Conjugate addition reported by Stoltz et al.136
1.9 Outline of this thesis
The studies presented in this thesis aim to understand the different types of palladium-catalyzed Heck reactions, conjugate addition reactions and their relationship.
Chapter 2 deals with the relationship between the Mizoroki-Heck reaction and the conjugate addition of aryl halides (formally, a reductive Heck arylation reaction). Our study indicates that under optimized conditions, the selectivity towards either of the products could be governed by the choice of the base. We found that a reductive base such as tributylamine is capable of steering the reaction to the conjugate product selectively, whereas use of a non-reductive base such as a carboxylate affords the Heck product.
Chapter 3 is dedicated to the study of the oxidative Heck reaction, in which we found that a BIAN ligand is superior to 2,9-dimethylphenanthroline in the Heck reaction of cyclic enones. The reaction proceeds under mild, base-free conditions, with molecular oxygen as the sole oxidant.
In Chapter 4, a highly enantioselective Pd-catalyzed conjugate addition reaction of arylboronic acids for the formation of benzylic quaternary centers is presented. The reaction is carried out under mild conditions, employing a catalyst system comprising of readily available PdCl2, PhBOX and AgSbF6.
The formation of benzylic quaternary centers in a racemic fashion is the topic of study in the chapter 5. Our studies show that a simplified catalyst system comprising Pd(O2CCF3)2 and bipyridine is capable of affording benzylic quaternary centers in excellent yields. For cyclic substrates, 1 mol% of the catalyst was sufficient, while 5 mol% was necessary for acyclic substrates.
In Chapter 6, the low enantioselectivities encountered in the addition to acyclic substrates in chapter 4 is addressed further. It was found that a quinolinyl-oxazoline ligand affords improved ee’s in the reaction. A higher temperature was found necessary.
Chapter 1
14
Chapter 7 details our attempts to develop an enantioselective Matsuda-Heck reaction. Our efforts in that direction did not result in a favorable outcome. In addition, studies on diastereoselective Matsuda-Heck reactions were performed.
Palladium-catalyzed Heck and conjugate addition reactions
15
1.10 References
(1) Hartings, M. Nat. Chem. 2012, 4, 764.
(2) Hartwig, J. F. Organotransition Metal Chemistry: From Bonding to Catalysis; University
Science Book: Sausalito, 2009.
(3) Transition Metals for Organic Synthesis: Building Blocks and Fine Chemicals; Beller, M.;
(125) Christoffers, J.; Baro, A. Quaternary Stereocenters: Challenges and Solutions for Organic
Synthesis; Wiley-VCH: Weinheim, 2005.
(126) Johansson, C. C. C.; Colacot, T. J. Angew. Chem. Int. Ed. 2010, 49, 676.
(127) Burtoloso, A. C. B. Synlett 2009, 320.
(128) Bellina, F.; Rossi, R. Chem. Rev. 2010, 110, 1082.
(129) Falciola, C. A.; Alexakis, A. Eur. J. Org. Chem. 2008, 3765.
Chapter 1
20
(130) Shibasaki, M.; Vogl, E. M.; Ohshima, T. Adv. Synth. Catal. 2004, 346, 1533.
(131) Dounay, A. B.; Overman, L. E. In The Mizoroki-Heck Reaction; Oestreich, M., Ed.; John
Wiley & Sons: 2009, p 533.
(132) Dounay, A. B.; Overman, L. E. Chem. Rev. 2003, 103, 2945.
(133) Hawner, C.; Alexakis, A. Chem. Commun. 2010, 46, 7295.
(134) Lin, S.; Lu, X. Org. Lett. 2010, 12, 2536.
(135) Lan, Y.; Houk, K. N. J. Org. Chem. 2011, 76, 4905.
(136) Kikushima, K.; Holder, J. C.; Gatti, M.; Stoltz, B. M. J. Am. Chem. Soc. 2011, 133, 6902.
Chapter 2
Conjugate addition versus Mizoroki-Heck reaction of aryl iodides
In this chapter, a study is presented that contributes to the understanding of the factors that differentiate between palladium-catalyzed conjugate addition and Mizoroki-Heck reaction of aryl halides to Michael acceptors. Our findings demonstrate that the reaction can be shaped in such a way that only the nature of the base determines the outcome of the reaction. This selectivity was found to be general for cyclic and linear enones but not for various other Michael acceptors. In addition, this study reports the first examples of the conjugate addition of aryl halides to nitro styrenes.
Parts of this chapter have been published: Gottumukkala, A. L.; de Vries, J. G.; Minnaard, A. J. Chem. Eur. J. 2011, 17, 3091.
Chapter 2
22
2.1 Introduction
Conjugate addition reactions form an important class of reliable carbon-carbon bond forming reactions. Among these, conjugate addition reactions of organometallics catalyzed by transition metals1,2 have great significance as these offer a wide scope of nucleophiles and pro-nucleophiles that can be employed.
Of the conjugate addition reactions of organometallics, those catalyzed by copper,3-5 rhodium,6,7 and palladium8-10 are well described in literature. Copper has been particularly successful for the conjugate addition of reactive organometallic reagents such as Grignard reagents11, organozincs12 and organoaluminiums.13 Rhodium catalyzed conjugate addition14 of soft organometallic reagents like boronic acids,15 borate esters, tetrafluoroborates,16 organosilicon reagents17,18 is also well established. There are also a handful of examples for the rhodium catalyzed conjugate addition of organozinc19 and organotitanium6,7 reagents. Compared to the former, palladium catalyzed conjugate addition14 of soft organometallics such as boronic acids20 is a recent development, becoming increasingly popular due to their broad functional group tolerance, mild reaction conditions and wide scope.
Irrespective of the catalysis, the required organometallic reagents; Grignard reagents, organozincs, and boronic acids or their derivatives, are nearly invariably synthesized from the corresponding halides. Therefore it would be a significant development if the aryl halides themselves could be used in conjugate addition reactions, as it would avoid the synthesis of the organometallic reagent.
It is remarkable that little attention has been paid, to date, to the direct use of aryl halides in metal catalyzed reductive conjugate addition reactions; all the more so because the closely related Mizoroki-Heck21-24 reaction has been studied and applied extensively. For the result to be a conjugate addition, a reductive cleavage of the intermediate alkyl-Pd complex, following olefin insertion, would be necessary as opposed to the -hydride elimination that takes place during the Mizoroki-Heck reaction (Scheme 1). It must be noted therefore, that this mechanism leading to the formation of the conjugate addition product, is distinct from copper catalysis, wherein, the nucleophilicity of the alkylcopper results in a Michael-type reaction. One might, therefore, view the conjugate addition reaction with aryl iodides, as involving a “catalytic umpolung”, which avoids a stoichiometric umpolung, that is involved in the synthesis of organometallic reagents.
Conjugate addition vs. Mizorioki-Heck reaction
23
Scheme 1: Reductive cleavage affords the conjugate addition product while -hydride
elimination leads to the Mizoroki-Heck product.
In fact, the conjugate addition product in palladium-catalyzed Mizoroki-Heck reactions is regularly observed as a side product25,26(Scheme 2) and this derailing is sometimes referred to as the reductive Heck reaction or reductive arylation. The extent of the reductive Heck product formed in the reaction varies greatly with base, temperature, substrate and solvent.
Scheme 2: Influence of reaction conditions on product distribution.26
This derailing has been only partially appreciated in literature in inter25,26 - and intramolecular27 cases (Scheme 3). However, it mostly remains unclear whether this involves a bona fide reductive conjugate addition or instead a Mizoroki-Heck reaction followed by reduction of the double bond, the latter being commonly observed also.28,29 A base dependent selectivity was observed by Buchwald and coworkers,27 who found that an exocyclic double bond could be formed (5b, Scheme 3b) when the insipient Pd-alkyl species reacted with the N-methyl of 1,2,2,6,6-pentamethylpiperidine (PMP).
Chapter 2
24
Scheme 3: a) Intramolecular reductive Heck. b) Exocyclic double bond obtained with PMP.
Surprisingly, only the group of Cacchi has studied the reductive conjugate addition of aryl halides in some detail25,30-36 for the addition of iodoarenes to benzalacetone (Scheme 4). The authors reasoned that the formation of the conjugate addition product could be favored by protonation of the alkylpalladium intermediate B (Scheme 5). It was therefore concluded that addition of an acid was imperative.25 Further, it was proposed that formic acid reduced the PdII to Pd0. Hence the addition of formic acid to the reaction, would serve both to protonate, and reduce the PdII to Pd0. The overall reaction proposed by the authors is depicted in Scheme 6.
Scheme 4: Conjugate addition to benzalacetone with aryl iodides.
Scheme 5: Role of the acid in conjugate addition.
Conjugate addition vs. Mizorioki-Heck reaction
25
Scheme 6: Proposed mechanism for conjugate addition, according to Cacchi et al.25
2.2 Goal
The goal of the present study was to develop a convenient and selective procedure for the conjugate addition of aryl halides to enones catalyzed by palladium. In order to attain high selectivity, it is imperative to understand the factors that lead to conjugate addition and Mizoroki-Heck reaction. Furthermore, an extension of the conjugate addition strategy, using aryl halides, to other classes of substrates such as nitrostyrenes would be of interest.
2.3 Results and discussion
Our initial study of the reaction parameters that influence the reaction pathway was based on the studies of Cacchi et al. Though initial studies37 from the authors used the combination of Et3N and HCO2H, subsequent studies25 reported improved selectivity using the combination of Bu3N and CF3CO2H. Reproducing the reaction as described (Table 1, entry 1), we scrutinized both the product composition and the reaction parameters. Next to the expected conjugate addition
Chapter 2
26
product 8a (62% yield), the Mizoroki-Heck product 8b, and 4,4’-dimethoxybiphenyl (8c) were obtained (Scheme 7). In addition, careful analysis revealed that varying, yet significant, amounts of 4-methoxybutyrophenone 8d (up to 20%) and 1-butenyl-4-methoxybenzene 8e (traces) were also formed, resulting from the oxidation of tributylamine (See section 2.3.6).
Scheme 7: Product distribution using conditions described by Cacchi et al.
Lowering the catalyst loading led to an increased reaction time (18 h, entry 2 vs. 12 h, entry 1), along with a similar distribution of products. Attempts to lower the amount of added iodoanisole lead to an incomplete conversion even after 18 h and thus lower isolated yields (not shown). Further we looked at various solvents in order to ascertain their influence on the reaction. Of the solvents tried†, the reaction proceeded to full conversion in the case of DMF and DMA. DMF quickly became the solvent of choice, as it is readily available in a dry form.
Next we looked at the influence of the various additives in the reaction as described by Cacchi.25 Essentially, all tetralkylammonium salts tested‡ led to a similar result in terms of conversion and selectivity. Interestingly, we observed that eliminating the additive from the reaction altogether also had no influence, on the scale at which we were running the reaction (1 mmol). Thus it was avoided in subsequent runs. Surprisingly, it turned out that this reaction gives the same outcome even in the absence of trifluoroacetic acid, which strongly simplifies the catalytic system (entries 5-10). Interestingly this led to an improvement in selectivity, as the Mizoroki-Heck product was no longer observed.
Table 1: Optimization of reaction parameters for the conjugate addition of 4-iodoanisole.
Reaction conditions: 6 (1.14 mmol), 7 (2.7 mmol), Bu3N (5.1 mmol), CF3COOH (3 mmol), nBu4NI (0.11 mmol), 80 oC, DMF, N2, 18 h. a Isolated yields. b Reaction completed in 12 h. c 1.4 mmol of 7 used. d CF3COOH and nBu4NI omitted. e Incomplete conversion after 18 h. Only trace amounts of product observed. f 10 mol% KOtBu added, isopropanol as solvent.[15] g Reaction carried out at 100 oC. h Reaction carried out at 120 oC. BINAP = 2,2'-bis(diphenylphosphino)-1,1'-binaphthyl. NHC = N-Heterocyclic Carbene. nd = not determined.
Experiments performed in the absence of Pd or Bu3N resulted in full recovery of the starting materials. This confirms that the reaction is catalyzed by Pd and Bu3N is essential for the reaction. Subsequently, we examined the use of a Pd0
Entry Pd mol% Ligand mol% Yielda (%)
1b Pd(OAc)2 10 PPh3 24 62
2 Pd(OAc)2 5 PPh3 11 60
3c Pd(OAc)2 5 PPh3 11 54
4 Pd2(dba)3.CHCl3 2.5 PPh3 11 62
5d Pd2(dba)3.CHCl3 2.5 PPh3 11 60
6 d Pd2(dba)3.CHCl3 2.5 Tol-BINAP 10 58
7 d Pd2(dba)3.CHCl3 2.5 rac-BINAP 10 64
8 d PEPPSI-IPr® 3 - nde
9 d,f PdII-NHC (9a) 3 - 74
10 d Pd0-NHC (10) 1.5 - 82
11 d,g Pd0-NHC(10) 0.2 - 58
12 d,h Pd0-NHC(10) 0.02 - 54
Chapter 2
28
precursor along with several bisphosphines♣ instead of PPh3 (entries 6, 7). This did not lead to a considerable improvement of the reaction. Taken together, the requirement of Pd and NBu3 and the independence of the outcome of the reaction on the nature of the added phosphine, strongly points at colloidal Pd particles24 being the catalytically active species. In order to have a well-defined catalyst that would also be directly relevant for enantioselective conjugate addition, we turned our attention to N-heterocyclic carbenes38 (NHCs) as ligands (Figure 1).
Figure 1: Pd-NHCs used for this study.
Our first attempts with these catalysts were with PEPPSITM-iPr39 (entry 8), which resulted in only trace amounts of product. We then focused our attention on PdII-NHC 9a, a catalyst that has been reported40-43 to be highly active for Mizoroki-Heck reactions. While the catalyst showed a respectable increase in isolated yield, the reaction was not always reproducible (See section 2.3.4 for details, and a plausible explanation).
We observed an increased activity and good reproducibility (Section 2.3.4) using commercially available Pd0–NHC 10, a catalyst found to be successful for the telomerization of dienes with alcohols.44,45 This enabled us to lower the catalyst loading considerably to 1.5 mol% (3 mol% Pd). Further attempts to lower the catalyst loading by 10-fold (0.2 mol%, entry 11) or 100-fold (0.02 mol%, entry 12) resulted in lower but still significant yields, albeit at higher temperature (100 oC). Taking the similarity in the lowered isolated yields (58% & 54% respectively) in these cases into consideration, though GC showed complete consumption of the starting material, this suggests decomposition of the substrate at higher temperature under these conditions. ♣ Other bisphosphines assayed included: di(hexafluorophenyl)phosphinoethane, dppb, dppp dppf, Xylyl-BINAP, R,Sp-Josiphos, R,Rp- Walphos and Taniaphos ligands. A handful of phosphoramidites were also assayed. All ligands tested gave similar yields as entries 6,7 and no change in product distribution was noticed.
Conjugate addition vs. Mizorioki-Heck reaction
29
Table 2: Substituted aryl iodides in the conjugate addition reaction.
aIsolated yield
Varying the electron density on the aryl iodide (Table 2) showed that also electron-poor aryl iodides perform reasonably well in the reaction. Replacement of aryl iodides with aryl bromides or chlorides under the same reaction conditions, only led to recovery of the starting materials. Although from a viewpoint of cost this is a disappointing observation, the selectivity of the reaction for aryl iodides allows other halogens to be present in the substrates, which in turn may be exploited for further functionalization.
After having achieved selective conjugate addition, we aimed for selective formation of the Mizoroki-Heck product as well, choosing bases that are incapable of reductively cleaving Pd. We studied several inorganic bases for application in this reaction, and many were quite successful, albeit at higher temperatures (> 120 oC). We were fortunate to find that cesium pivalate (CsOPiv) was very effective even at lower temperature (80 oC), perhaps due to its solubility in organic solvents. In short, complete reversal of the selectivity to the Mizoroki-Heck product was achieved, establishing that the base is key in tuning the selectivity.
As is invariably found with Heck reactions of -substituted enones, a mixture of double bond isomers was obtained which indicates a fast palladium C-bound-to-O-bound equilibrium8 (Scheme 8). It is, at present, unclear whether in the conjugate addition, reduction takes place from the C- or O-bound palladium.
Entry Substituent Yield (%)a Product
1 4-MeO 82 8a
2 H 83 11a
3 3-Cl 52 12a
4 3-Br 56 13a
5 4-Cl 58 14a
6 4-Br 63 15a
Chapter 2
30
Scheme 8: Mixture of E / Z isomers obtained in the reaction.
An alternative explanation could be that the Pd-hydride formed upon -hydride elimination could re-insert into the newly formed double bond without any facial preference and upon a second -elimination, results in E / Z isomers.46 This process could occur multiple times, thereby etching away selectivity obtained in the reaction.
2.3.2 Scope of the reaction.
Following our studies on the optimization of the reaction parameters, the scope of the reaction was investigated, with different enones.
Benzalacetone 6, afforded the conjugate addition product (8a) in 83% yield, while the corresponding Mizoroki-Heck product (8b) was obtained in 82% yield. Other products were only observed in trace amounts by GC. Chalcone gave 16a in 73% yield and 16b in a somewhat higher yield of 80%.
(E)-Hept-3-en-2-one gave 17a in 63% yield and the corresponding Mizoroki-Heck product (17b) in 84% yield. Cyclohexenone afforded 18a and 18b in 56% and 64% yield, respectively, but cycloheptenone gave only the conjugate addition product 19a in 69% yield. No Heck product was observed, despite several attempts. In fact, the Heck reaction with cycloheptenone has never been described!
Conjugate addition vs. Mizorioki-Heck reaction
31
Table 3: Selectivity governed by the base.
a Isolated yield. b Obtained as a mixture of E/Z isomers. nd = not determined.
The fact that the conjugate addition reaction on this substrate works readily, while the Heck reaction does not, suggests that the -hydride elimination is unfavorable
due to ring strain. Nitrochalcone furnished 20a in an excellent 92% yield while 20b was obtained in 71% yield (entry 6). trans,trans-Dibenzylideneacetone (dba) afforded 21a in 74% yield and the Mizoroki-Heck product (21b) in 92% yield. 4% Monoarylated product was also observed. Impressively, the reactions could also be easily carried out in a microwave, without any change in reaction parameters, and shortening reaction times considerably. The yields and selectivity of the product formation remained consistent with reactions carried out in an oil bath. The reaction of 6 to 8a could be completed in 7 min, while all the reported substrates in Table 3 went to complete conversion in less than 30 min, while these reactions took 18 h to reach full conversion in an oil bath (See experimental section for details). When D-mannitol derived 22 was employed, a diastereoselective reductive conjugate addition took place affording a 5:1 mixture in favor of the anti product47 (Scheme 9).
Scheme 9: Diastereoselective conjugate addition of aryl iodides.
In contrast to enones, , -unsaturated esters (including acrylates), amides and nitriles yielded only their corresponding Mizoroki-Heck products, regardless of the base used (Scheme 10a). Nonetheless, when the ester contained a strongly electron-withdrawing moiety, such as a tetrafluorophenyl group, we were able to observe the formation of the conjugate addition product 37, though only in low yields. Due to the ease of hydrolysis of the substrate and the formed product, the product was isolated as its methylester (Scheme 10b). Unsurprisingly, “unactivated” esters are poor substrates, also for the conjugate addition of boronic acids catalyzed by palladium. Miyaura et al. explain this as the result of a slow equilibration between C-bound and O-bound enolates8 (see section 2.3.5).
Conjugate addition vs. Mizorioki-Heck reaction
33
Scheme 10: a) Substrates that afford only the Mizoroki-Heck product. b) Conjugate addition in
case of an ester bearing a highly electron withdrawing group.
Chapter 2
34
2.3.3. -Nitrostyrenes
Table 4. Conjugate addition to -nitrostyrenes.
Entry Substrate Yield Producta
1
28%
38
2
54%
39
3
64%
40
4
21%
41
5
38%
42
6
-- -- --
a Isolated yield
Conjugate addition vs. Mizorioki-Heck reaction
35
In order to stretch the utility of the approach, we studied -nitrostyrenes (Table 4), a substrate class that until now failed to undergo catalytic Mizoroki-Heck reactions.48 Intriguingly, we observed that although attempted Mizoroki-Heck reaction lead to recovered starting materials, conjugate addition readily took place using Bu3N as the base (Table 4). This questions the current opinion that substrate coordination or sequestration of Pd by the nitro-group, is the reason for failure48 of these substrates in the Mizoroki-Heck reaction.
The isolated yields of the product seem to strongly mirror the electron density of the substrate. 2,3-Dimethoxy- -nitrostyrene provided 40 in 64% isolated yield (entry 3). 1-Nitro-1-cyclohexene was recovered unreacted (entry 6).
2.3.4. Synthesis of N-heterocyclic carbenes
Realizing the mechanistic similarity of the desired conjugate addition of aryl halides to the Mizoroki-Heck reaction, we surmised that highly active catalysts for the latter would also perform well in the former. Consequently, we went on to synthesize and test catalysts that were reported to be the most active for Mizoroki-Heck reaction at the time of the study.
The success of palladium N-heterocyclic carbene catalysts38 (NHCs) in promoting various cross coupling reactions,49 and in particular the Mizoroki-Heck reaction40,42 is well documented. We therefore proceeded to synthesize 9a, a highly active precatalyst for the Mizoroki-Heck Reaction.40
Scheme 11: Synthesis of PdII –NHC 9a.
Realizing the influence of steric parameters of the carbene ligand on the catalytic activity of the complex,50 we also synthesized 9b and 9c by a similar procedure, albeit in a lower yield (Scheme 12). The X-ray structure of 9b was obtained, which confirms the square planar PdII center (Figure 2).
It must be noted that these are PdII complexes, and for the catalyst to be active in the reaction, it must be reduced to Pd0. This implies that there needs to be an “activation” of the catalyst via reduction of the Pd center (to Pd0). Such activations have been previously documented.51 Thus, we chose to add 10 mol% KOtBu in
Chapter 2
36
100 vol% of isopropanol as an activator, to a solution of the catalyst, prior to the addition of the reactants. The activation is thought to work via the mechanism shown in Scheme 13. A solution of 9a in DMF (pale yellow) turns black within minutes after adding KOtBu and isopropanol, at 80 oC.
Scheme 12: PdII-NHC complexes with varying steric bulk.
Figure 2: X-ray structure of 9b.
Conjugate addition vs. Mizorioki-Heck reaction
37
Scheme 13: Proposed activation of 9a with KOtBu and isopropanol.
The catalytic activity of precatalysts 9a, 9b and 9c was examined in the conjugate addition of iodoanisole to benzalacetone, under the optimized conditions (Table 5). Although in general the yields obtained were comparable or marginally higher than those obtained with phosphines (except 9c), the yields were not always reproducible. We attributed the cause of this to difficulty in reproducing the activation.
Table 5. Conjugate addition with complexes 9a-c
a Isolated yield.
In order to overcome this difficulty, we scouted for Pd0–NHCs that would not require this activation. We were delighted to find that 10, a commercially available catalyst, satisfied all the required criteria for the reaction, and was thus chosen for all subsequent experimentation.
Entry Conditions Yielda
1 9a (3 mol%) + NBu3 + IPA + KOtBu (10 mol%) 74%
2 9b (3 mol%) + NBu3 + IPA + KOtBu (10 mol%) 62%
3 9c (3 mol%) + NBu3 + IPA + KOtBu (10 mol%) 43%
Chapter 2
38
2.3.5. Studies with chiral phosphines
Prior to our effort on the use of N-heterocyclic carbenes, we studied the use of chiral phosphines in the reaction with the desire to render the reaction enantioselective. Though almost all the assayed phosphines gave good conversions, we did not observe any enantioselectivity. We reasoned this to be a consequence of extensive competing coordination of the added trialkylamine to the metal, which could displace the phosphine and hence lead to a loss of enantioselectivity. This observation re-iterated our interest in the use of N-heterocyclic carbene bound Pd for our reaction.
2.3.6. Mechanistic insights
We propose the following mechanism (Scheme 14), which delineates the similarity between conjugate addition and Mizoroki-Heck reaction. The conjugate addition pathway is depicted with solid arrows, and the Mizoroki-Heck with dashed arrows.
Oxidative addition of iodoanisole to Pd0–NHC results in species B, which undergoes migratory insertion into benzalacetone to form species C. The Pd center can remain bound to the -carbon of the ketone (C-bound) as in species C, or may exist in an equilibrium with its O-bound form (C ) via keto-enol tautomerism In our case, we believe that the equilibrium is largely in the C-bound form (species C). In the presence of a reductive base such as tributylamine, there is a coordination of the tributylamine to the Pd, forming D. The tributylamine possesses six -hydrogens which are available to Pd for -hydride elimination, forming species E. The elimination product (the imine) can disintegrate further to give side-products. Species E reductively eliminates to form the conjugate addition product and regenerates the Pd0 catalyst.
In the absence of a reductive base, however, species C can transform to species F via rotation about the single bond, thereby aligning the -hydrogen and Pd syn to one another. This facilitates a -hydride elimination, possibly assisted by the base, to afford 8b and regenerate the Pd0 catalyst. The Mizoroki-Heck product is obtained as a mixture of E / Z isomers (Scheme 8).
Conjugate addition vs. Mizorioki-Heck reaction
39
Scheme 14: Proposed reaction pathway.
Interestingly, in mechanistic studies performed by Friestad and Branchaud52 on the intramolecular conjugate addition of aryl iodides, it was observed that there was a buildup of the Mizoroki-Heck product, which began to disappear towards the end of the reaction and afforded the conjugate addition product. This led the authors to propose that there was an initial Mizoroki-Heck reaction taking place forming product 45b. As the reaction proceeded, a stoichiometric amount of triethylammonium salt was formed, which the authors suppose is in equilibrium
Chapter 2
40
with a Pd-H species, and therefore involved in the reduction of the Mizoroki-Heck product to the conjugate addition product (45a).
Scheme 15: Studies from Branchaud proposing the formation of the mixture of 45a and 45b
Taking these observations into consideration, we followed the progress of our reaction in time, by GC. We only observed a time dependent increase of 8a and no formation of 8b was observed, during the entire reaction. This makes it unlikely that our reaction works via a Heck-followed-by-reduction sequence.
Further we studied our reaction in presence of added 16b (The Heck product, Scheme 16). Should the Branchaud mechanism hold, one would expect, using aryl iodides, 16b being converted to 16a competitively, towards the end of the reaction. However, 16b was recovered fully, unreacted (Scheme 16).
Scheme 16: Recovery of 16b in the reaction indicates the absence of degenerate reduction via Pd-hydride species.
Further we observed, that oxidized tributylamine underwent further reactions, as described in Scheme 17. Although it was not possible to isolate dibutylamine and p-anisylbutanone (8d) in a quantitative way from the reaction due to the workup involved, these products were unambiguously identified by NMR.
Conjugate addition vs. Mizorioki-Heck reaction
41
Scheme 17: Side products observed by GCMS and rationale for their formation.
Following our study, a report appeared in literature53 wherein trialkylamines (51) are converted to the corresponding arylketones (53) using aryl iodides via Pd catalysis. The authors find it necessary to add ZnO for the success of the reaction, which presumably assists in the formation of the iminium species (Scheme 18).
Scheme 18: Pd-catalyzed synthesis of aryl ketones from trialkylamines.
Based on these collective observations and reports, we surmise that the reaction works as depicted in Scheme 14.
2.4 Conclusion and future perspectives.
In summary, a catalytic system has been developed which, by choice of the base, selectively switches between reductive conjugate addition and Mizoroki-Heck reaction of aryl iodides to , -unsaturated enones. For conjugate addition reactions, this avoids the preparation and use of organometallics, rendering it atom economic and cost-effective. Reductive cleavage of Pd, instead of protonolyis, is proposed to release the product.
For the first time, this reaction has been extended to -nitrostyrenes, a class of substrates unamenable to the corresponding Mizoroki-Heck reaction. The reaction is completed in 30 min under microwave irradiation, and can be performed with
Chapter 2
42
good diastereoselectivity. This conjugate addition reaction has considerable scope, also in cases were the Mizoroki-Heck reaction fails.
These studies are an excellent prelude to the development of the enantioselective version of this reaction. Our studies using chiral phosphines were not fruitful, most probably because the chiral phosphine ligand does not stay on the palladium during the reaction. Future research should focus on the application of chiral N-heterocyclic carbene complexes of palladium, as NHC ligands are known to coordinate to Pd strongly. It is relevant to note that most of the Pd-NHC complexes reported to date are PdII complexes, while the most effective catalyst for our reaction is a Pd0 complex. Thus significant attention to the design and synthesis of these complexes would be profitable, unless a reliable activation (reduction) protocol is developed. Though amines are usually considered to be excellent reductants for Pd, however, they do not seem to be able to activate the precatalysts 9a-c.
In terms of mechanistic understanding of the reaction, our results strongly suggest that the conjugate addition product is not formed by a sequential Heck-reduction pathway, but instead via bona fide conjugate addition. The scope of the reaction remains to be extended to , -unsaturated esters, nitriles, amides and aldehydes.
Conjugate addition vs. Mizorioki-Heck reaction
43
2.5 Experimental
2.5.1 General
All experiments were carried out in flame dried or oven dried (150 oC) glassware,
in an atmosphere of nitrogen, unless specified otherwise, by standard Schlenk
techniques. Schlenk reaction tubes with screw caps, and equipped with a Teflon-
coated magnetic stir bar were flame dried under vacuum and allowed to return to
room temperature prior to being charged with reactants. A manifold permitting
switching between nitrogen atmosphere and vacuum was used to control the
atmosphere in the reaction vessel. Once charged with all the reactants, the
reaction vessel was cycled through at least 3 cycles of nitrogen-vacuum-nitrogen
to ensure the atmosphere was inert. Reaction temperature refers to the
temperature of the oil bath. Flash chromatography was performed using Merck
silica gel type 9385 (230-400 mesh), using the indicated solvents.
Experiments were performed on a CEM Discover Microwave apparatus, in crimp-
sealable reaction tubes. The tubes were prepared similar to reactions in oil bath.
The reactions were performed under Constant-Temperature Mode (Temperature
Max 80 oC), 110 W, with air-cooling, for 30 min. Work up and analysis was similar
to reactions performed using an oil bath.
All solvents used for extraction, filtration and chromatography were of commercial
grade, and used without further purification, except for pentane, which was
distilled prior to use. Peptide Synthesis grade DMF was purchased from
BIOSOLVE BV (The Netherlands) and degassed by a freeze-pump-thaw
procedure (repeated cycles of freezing under vacuum with liquid nitrogen and
thawing to room temperature) prior to use. The degassing was performed
regularly in small batches, and once degassed, the DMF was used within a few
days.
Reagents were purchased from Sigma-Aldrich, Strem or Acros and used without
further purification. Analytical grade tri-n-butylamine was purchased from Sigma-
Aldrich and stored in an atmosphere of nitrogen. Palladium complex PdII-NHCs
(9a-c) was synthesized as described in literature40,41 and stored under ambient
Chapter 2
44
conditions. Palladium complex Pd0-NHC (10) was purchased from Sigma-Aldrich
and stored in a nitrogen-filled glovebox, from which the requisite amounts were
weighed out.
TLC was performed on Merck silica gel 60, 0.25 mm plates and visualization was
done by UV and staining with Seebach’s reagent (a mixture of phosphomolybdic
acid (25 g), cerium (IV) sulfate (7.5 g), H2O (500 ml) and H2SO4 (25 ml)). 1H- and 13C-NMR were recorded on a Varian AMX400 (400, 100.59 MHz, respectively)
using CDCl3 as solvent, unless specified otherwise. Chemical shift values are
reported in ppm with the solvent resonance as the internal standard (CHCl3:
7.26 for 1H, 77.0 for 13C). Data are reported as follows: chemical shifts ( ),
multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, br = broad, m =
multiplet), coupling constants J (Hz), and integration.
GC-MS measurements were made using a HP 6890 Series Gas Chromatograph
system equipped with a HP 5973 Mass Sensitive Detector. GC measurements
were made using a Shimadzu GC 2014 gas chromatograph system bearing a AT5
column (Grace Alltech) and FID detection. Whenever GC yield is reported, the
quantification was done using cyclo-octane as internal standard. High Resolution
Mass measurements were performed using a ThermoScientific LTQ Oribitrap XL
spectrometer.
2.5.2 General experimental procedure for conjugate addition:
A Schlenk tube equipped with screwcap and egg-shaped Teflon coated stirbar
was flame dried under vacuum and allowed to cool down to room temperature
under vacuum. The tube was backfilled with nitrogen prior to opening the cap. 4-
Iodoanisole (2.72 mmol, 637 mg), and benzalacetone (1.14 mmol, 166 mg) were
charged under a stream of nitrogen. The tube was capped with a rubber septum
and subjected to 3 cycles of vacuum and nitrogen. 10 (1.5 mol%, 0.017 mmol)
was weighed under nitrogen, dissolved in 1 ml of degassed DMF and injected into
the reaction tube. Tri-n-butylamine (5.1 mmol, 1.2 ml) was added via syringe, and
the septum was replaced by a screw cap. The reaction mixture is biphasic with the
DMF layer being colored and the tri-n-butylamine being nearly colorless or faint
Conjugate addition vs. Mizorioki-Heck reaction
45
yellow. The Schlenk tube was then alternated through 3 cycles of vacuum and
nitrogen, and placed into a pre-heated oil bath at 80 oC.
Upon completion as judged by GC-MS and/or TLC, (sampled from the DMF layer)
the reaction was cooled to room temperature and poured into 10% HCl (v/v)
solution (10 ml) and extracted with ether (3 X 25 ml). The organic extracts were
combined, dried over anhydrous MgSO4 and concentrated in vacuo. The
concentrate was loaded directly or adsorbed onto silica prior to loading onto a
silica gel column and eluted.
2.5.3 General experimental procedure for the Mizoroki-Heck reaction:
A Schlenk tube equipped with screw cap and egg-shaped Teflon coated stirbar
was flame dried under vacuum and allowed to cool down to room temperature
under vacuum. The tube was backfilled with nitrogen prior to opening the cap. 4-
Pd-BIAN: A highly selective catalyst for the base-free oxidative Heck reaction.
In this chapter, a systematic study is presented on the development of a selective catalyst for the palladium-catalyzed base-free oxidative Heck reaction. We have found that N,N’-bis-2,6-xylyl-acenaphthenequinonediimine (BIAN) is an excellent ligand for this reaction, inducing high rates, being stable under the oxidative conditions of the reaction. In addition, its use led to minimal formation of biphenyl and phenol. Our study also indicates that a cationic Pd complex is needed as addition of excess halide retarded the reaction
Parts of this chapter have been published: Gottumukkala, A. L.; Teichert,J.F.; Ferrer, C.; Eisink, S.; Heijnen, D.; van Dijk, S.; van den Hoogenband, A.; Minnaard, A. J.; J. Org. Chem. 2011, 76 , 3498.
Chapter 3
58
3.1 Introduction
3.1.1 The Oxidative Heck and Mizoroki-Heck reactions
Development of methodologies that allow facile construction of carbon-carbon bonds under mild reaction conditions, with minimal use of additives and low waste production remains a highly active field of research. The oxidative variant of the Mizoroki-Heck reaction (or simply the “oxidative-Heck” reaction), may be viewed as an example of such an effort. While sharing attributes with the Mizoroki-Heck reaction, the oxidative Heck reaction often requires milder conditions and does not require a base for the reaction, though a stoichiometric oxidant is essential. A simplified comparative scheme depicting the two reactions is presented in Figure 1.
Figure 1: A simplified comparison of the oxidative Heck and the Mizoroki-Heck reaction.
The oxidant (in this case, oxygen) serves to regenerate the PdII species A, that transmetalates with an organometallic reagent (in this case, phenylboronic acid) to form product B. This then undergoes migratory insertion to C, followed by isomerization via an oxa-allyl species D to E, in which the Pd is oriented syn to the
Oxidative Heck reactions of cyclic enones
59
-hydrogen, facilitating -hydride elimination, expelling the formed product and palladium (0).
3.1.2 Background
The first example of an oxidative Heck reaction was demonstrated by Richard Heck1 with the use of arylmercuric chlorides. The reaction of phenylmercuric chloride with methylacrylate, catalyzed by Li2PdCl4, using CuCl2 as the oxidant, yielded 3 in 57% (Scheme 1). It is interesting to note that the reaction was run at room temperature and under base-free conditions.
Scheme 1: The first example of an oxidative Heck reaction.1
However, this reaction failed to catch the attention of synthetic chemists for decades, most probably due to the toxic organomercury compounds employed. However, this changed with the report of Uemura et al. in 1994,2 wherein the authors demonstrated that the reaction could be carried out equally well with arylboronic acids. The reaction on iso-butene resulted in a mixture of isomers with varying yields (Scheme 2), but this report set the stage for several developments. In addition, the authors proposed the oxidative addition of the arylboronic acid to a Pd0 center, much like an aryl halide, to generate the catalytically active PdII species. Later reports have questioned this proposal,3,4 and it is now accepted that the Pd0 center is oxidized to PdII, before transmetalation with the arylboronic acid.
Scheme 2: The first example of arylboronic acids in the oxidative Heck reaction.2
While important contributions have since been made using various other oxidants, (most notably CuII salts) the following section focuses on aerobic oxidations (oxidation with dioxygen), the topic of the current study. The reader is directed to a recent review5 for detailed information.
Chapter 3
60
3.1.3 Aerobic oxidative Heck reactions
Jung and coworkers reported the first example4 of the aerobic oxidative Heck reaction. The reaction was carried out under ligand-free conditions, with Na2CO3 as a base and at 50 oC (Scheme 3). Despite being limited in substrate scope to terminal olefins and with a high catalyst loading (10 mol%), the study paved the way for a flurry of investigations into the use of molecular oxygen as an effective oxidant for the reaction. Subsequent studies6 by the same group resulted in lower catalyst loadings and an extension of the scope to alkenyl boronic acids.
Scheme 3: The first example of an aerobic oxidative Heck reaction.4
Larhed and coworkers studied the influence of an added ligand for this reaction.7 They found 2,9-dimethyl-1,10-phenanthroline (dmphen) to be a highly effective ligand for this reaction, when carried out in the presence of N-methylmorpholine (as base). Phosphines were found to be ineffective for this reaction, as they were prone to oxidation under these conditions. The use of dmphen allowed them to reduce the catalyst loading to 1 mol% (Scheme 4). However, the reaction remained limited to terminal olefins and required elevated temperatures (50 oC).
Scheme 4: The use of dmphen as a ligand, pioneered by Larhed.7
The same group further investigated the regioselectivity of the arylation of enamides as a probe for the influence of electronic properties of the boronic acid (Scheme 5).8 They observed a higher selectivity for α−arylation with electron rich boronic acids, whereas use of electron poor boronic acids led to lower selectivity. This selectivity was completely lost upon addition of LiCl to the reaction. This observation, along with DFT calculations led them to propose a cationic Pd intermediate being active in the reaction. The difference in regioselectivity was attributed to variation in the energy barriers for the complexation of the enamide to the cationic arylpalladium species (formed by transmetallation of the boronic acid to Pd) with the varying electron-density of the aromatic group of the boronic acid.
Oxidative Heck reactions of cyclic enones
61
They further observed much lower yields and the requirement for a higher catalyst loading (40 mol%) with electron-poor boronic acids (Scheme 5).
Scheme 5: Regioselectivity as a function of electron density of the arylating agent.8
Independent studies by Larhed9 and Jung10, further demonstrated that an improved oxidative Heck reaction with dmphen could be carried out in the absence of base. While the latter study focused on the coupling of alkenylboronic acids to enones, the former study concentrated solely on arylboronic acids. Studies by Larhed showed that the reaction was easily scalable, and could even be performed in a microwave.9,11 The study also reported an expanded scope of alkenes that may be employed for the reaction. In addition to commonly employed acrylates, various enones, enals, enamides, vinylsulfones, styrenes, enol ethers and vinylboronate esters were found amenable to the reaction (Figure 2), and were found to react in synthetically applicable yields in the microwave within 10 minutes.
Figure 2: Substrate scope for the oxidative Heck reaction with arylboronic acids in microwave.
Jung et al. were the first to report examples of the oxidative Heck reaction on cyclic enones (Scheme 6).10 Unlike terminal enones, cyclic enones were found to react poorly under the base-free and ligandless conditions. Their reactivity improved in the presence of base. In addition to Heck product, varying amounts of conjugate addition product were also observed. A further improvement in reactivity was observed when dmphen was added as a ligand. They reported 81% isolated yield for an aerobic, base-free oxidative Heck reaction carried out on cyclohexenone with phenylboronic acid. It must be noted however, that in our hands we were unable to reproduce this claim on three separate attempts. In addition to the
Chapter 3
62
expected Heck product (46%), the conjugate addition product, biphenyl and phenol were also observed.
Scheme 6: The first example of an oxidative Heck reaction with cylcohexenone, as reported by
Jung.10
3.1.4 Ligand stability
Dmphen as a ligand has been used extensively in Pd catalyzed aerobic oxidation12,13 of alcohols to aldehydes and ketones. The catalyst was found to be highly active and using polar groups, such as sulfonates, on the ligand, these reactions could also be performed in water.14 The authors found that the presence of methyl groups at the 2 and 9 positions prevents dimerization, and thus increases the lifetime of the active catalyst (Scheme 7).14
Scheme 7: a) Pd catalyzed aerobic oxidation of alcohols. b) role of the methyl groups in
preventing dimerization.
Studies by Waymouth et al,15 into the decomposition pathways of this catalyst during the reaction have shown that the inactivation of the catalyst is likely via the oxidation of the proximal methyl group resulting in the carboxylate complex 30. The authors found that the formed complex is inactive as a catalyst (Figure 3).
Oxidative Heck reactions of cyclic enones
63
Figure 3: X-ray structure of the inactivated catalyst upon oxidation of the ligand. The triflate
counterion is omitted for clarity.
3.1.5 BIAN ligands
Bis-imines of acenaphthenequinone, (commonly abbreviated as BIAN) are an important class of ligands in organometallic chemistry and transition metal catalysis. They have been studied extensively by Elsevier and co-workers16-18 and they have been documented in the literature since the 1990s.
Their ability to form stable complexes with s, p19 and d block elements has made them widely applied in polymerization,20-25 hydrogenation26-29 and other carbon-carbon18 and carbon-hydrogen28 bond forming reactions. The Pd complexes of BIAN ligands have been particularly successful for Z-selective alkyne semi-hydrogenations,18,19 and also exhibit interesting redox properties.30 Their role in facilitating Pd catalyzed carbon-carbon bond formation has been reviewed in detail.31
Their facile synthesis32,33 (Scheme 8, see experimental section for details) allows a large degree of variation in the aniline building block to tune their properties.33,34
Scheme 8: a) General synthesis of Bis-imine-acenapthaquinone ligands. b) o,o’- Dimethyl-BIAN
(34) used in this study.
Chapter 3
64
3.1.6 Mechanistic attributes and side products
In line with empirical observations that cationic Pd intermediates were involved in the reaction, Larhed and coworkers studied the aerobic oxidative Heck reaction by ESI-MS/MS.35 They were able to identify several stable reaction intermediates with this method, confirming a cationic reaction pathway for the reaction. The observed intermediates were largely consistent with the commonly accepted mechanism (Scheme 9).
Scheme 9: Intermediates in the oxidative Heck reaction observed by ESI-MS/MS.
The role of an added base was also studied. While it was found to have a positive influence on reactions carried out under ligandless conditions, it was found to have a detrimental role in reactions carried out in the presence of ligands. The base assists in the formation of reactive arylboronate species, which can readily undergo homocoupling, forming biaryl. In the absence of base, this side reaction was found to be considerably diminished. The conjugate addition product results from hydrolysis of the alkylpalladium species following migratory insertion as depicted in Scheme 10.
Oxidative Heck reactions of cyclic enones
65
Scheme 10: Formation of the conjugate addition product.
The Pd catalyzed formation of phenol from arylboronic acids may be envisioned to proceed via the catalytic cycle depicted in Scheme 11.36 To date, this mechanism remains experimentally unverified.
Scheme 11: Pd-catalyzed formation of Phenol from Phenylboronic acid.
Furthermore, the conversion of arylboronic acids to phenols, under a variety of oxidative conditions, is a well-described reaction.37-41 However, recently, Stahl and coworkers42 have demonstrated that cyclohexenone can be converted to phenol very efficiently via oxidative palladium catalysis. The reaction is assumed to work via a C-H activation of the -position of the cyclohexenone, followed by -hydride elimination and subsequent tautomerization (Scheme 12).
Chapter 3
66
Scheme 12: Formation of phenol from cyclohexenone via oxidative Pd catalysis.
3.2 Goal
The goal of this project was to develop a highly selective protocol for the oxidative Heck reaction of cyclic enones, under base-free conditions and with molecular oxygen as the sole oxidant. Previous studies of the oxidative Heck reaction often resulted in product mixtures for this class of substrates, consisting of the corresponding conjugate addition product, biphenyl and phenol in addition to the desired Heck product.
Furthermore, a ligand that remains stable under the oxidative conditions of the reaction should provide a higher turnover number and possibly a higher turnover frequency. As noted earlier, phosphines, the most commonly employed class of ligands in transition metal catalysis can not be used for this reaction, as they are prone to oxidation. In addition, a ligand that provides stabilization to high valent metal species, while remaining inert to side reactions (such as C-H activation of proximal alkyl units) is an important factor in choosing the right ligand.
3.3 Results and discussion
3.3.1 Development of the reaction protocol
Cyclohex-2-en-1-one (23) was chosen as benchmark substrate in the oxidative Heck reaction with phenylboronic acid (9). The ability to carry out the reactions at room temperature and with oxygen (balloon connected to flask) as the sole oxidant was set as an important criterion (Scheme 13).
Scheme 13: Oxidative Heck reaction of cyclohexenone and phenylboronic acid.
Oxidative Heck reactions of cyclic enones
67
During the initial study of the reaction parameters, 3 equiv of phenylboronic acid were employed, as is common in literature. Performing the reaction in DMF, in the absence of any ligand led to only 50% consumption of the starting material (Table 1, entry 1). The addition of 5 mol% of BIAN to this reaction led to an improved conversion, along with the formation of biphenyl (10%). Lower conversions were observed when the reaction was performed in THF, ethanol (entries 2, 3) or THF : water (2:1). Dry methanol afforded 80% conversion, but a 9:1 ratio of MeOH : H2O led to full conversion towards the product, though 5% of biphenyl was observed (Table 1, entry 5). The improved selectivity to product in the presence of water may be rationalized on two accounts. First, water assists the transmetalation of the boronic acid to palladium via the formation of either the reactive phenyltrihydroxyborate species, or alternatively a hydroxopalladium species.43 Secondly, it prevents the formation of boroxine (40), via the equilibrium depicted in Scheme 14.
Scheme 14: The formation of boroxine from phenylboronic acid.
Under these conditions, it was sufficient to use 1.5 equiv of phenylboronic acid to attain full conversion. Increasing the proportion of water further (MeOH : H2O = 4:1) led to poor solubility of the substrates and catalyst in the reaction medium. The formation of biphenyl could be avoided by the addition of a slight excess (2 mol%) of BIAN (Table 1, entry 6). A three-fold excess of the ligand compared to the metal led to no change in conversion or product distribution (Table 1, entry 7). Taking this together, in addition to the observed increase of biphenyl formation in the absence of ligand (Table 1, entry 1), it is likely that the presence of unligated Pd under these reaction conditions results in the formation of biphenyl. The use of dmphen under these conditions led to only 40% consumption of starting material, in addition to 30% biphenyl (Table 1, entry 8). Using 15 mol% of dmphen in the reaction led to a complete arrest of the reaction (Table 1, entry 9).
To ascertain our observations of increased reactivity and selectivity of Pd(OAc)2/BIAN versus Pd(OAc)2/dmphen, we synthesized these complexes (Figure 4) independently and studied them in our reaction. Use of 42a resulted in complete consumption of the starting material, with minimal formation of biphenyl
Chapter 3
68
(< 5%), whilst use of 41a resulted in incomplete conversion, along with 10% biphenyl.
From the first set of optimizations, we conclude that BIAN is superior to dmphen as a ligand for the oxidative Heck reaction performed under base-free conditions at room temperature, and the presence of water in the reaction medium increases the conversions.
Table 1: Comparison of dmphen and BIAN as ligands in the oxidative Heck reaction .a
Entry Solvent Ligand (mol%) Conv. (%)b Side Product (%)b
1 DMF -- 50 biphenyl (50%)
2 THF BIAN (5) 40
3 EtOH BIAN (5) 10 biphenyl (trace) +
phenol
4 MeOH (dry) BIAN (5) 80 -
5 MeOH:H2O (9:1) BIAN (5) full biphenyl (5%)
6 MeOH:H2O (9:1) BIAN (7) full -
7 MeOH:H2O (9:1) BIAN (15) full -
8 MeOH:H2O (9:1) dmphen (5) 40 biphenyl (30%)
9 MeOH:H2O (9:1) dmphen (15) -- --
10c MeOH:H2O (9:1) 42a full biphenyl (5%)
11c MeOH:H2O (9:1) 41a 80 biphenyl (10%)
a Cyclohex-2-en-1-one (0.1 mmol), phenylboronic acid (0.3 mmol), O2 balloon, rt, MeOH : water (9:1) 0.2 ml, 30 h. dmphen = 2,9-dimethyl-1,10-phenanthroline. b Determined by GC with dodecane as internal standard.
c Complex used instead of ligand.
For further optimization (Table 2), complex 42a was employed as the catalyst. The influence of reducing the amount of phenylboronic acid on the reaction was studied. We found that full conversion could be achieved with as little as 1.5 equiv, though reaction times were slightly longer (30 h vs. 24 h). Lowering the catalyst
Oxidative Heck reactions of cyclic enones
69
loading to 2.5 mol% (entry 4) increased the reaction time to 45 h, whilst the reaction was incomplete even after 72 h with 1 mol% of the catalyst (entry 5).
Figure 4: Palladium complexes synthesized for this study.
To determine whether air could be used instead of pure dioxygen, the reaction vessel was left open to air. The reaction was found to proceed (entry 6) albeit with a lower yield. Air at balloon pressure (entry 7), or at 2 atm (entry 8) did not result in a significant improvement.
Table 2: Optimization of reaction parameters for the oxidative Heck reaction on cyclohexenone.a
Entry 42a
(mol%)
Ratio
(23:9) Solvent Atm. t (h) Yield
(%)a
1 5 1:3.0 MeOH (dry) O2c 24 75
2 5 1:2.5 MeOH/H2O (9:1) O2c 28 86
3 5 1:1.5 MeOH/H2O (9:1) O2c 30 85
4 2.5 1:1.5 MeOH/H2O (9:1) O2c 45 77
5 1 1:1.5 MeOH/H2O (9:1) O2c 72 ndb
6 5 1:1.5 MeOH/H2O (9:1) Air 30 45
7 5 1:1.5 MeOH/H2O (9:1) Airc 30 49
8 5 1:1.5 MeOH/H2O (9:1) Aird 30 50
a Isolated yield.b Not determined. Incomplete conversion even after 72 h.c Balloon pressure. d 2 atm.
Thus, the optimal conditions for the reaction are a 9:1 MeOH : water mixture as solvent at room temperature and using dioxygen (from a balloon appended to the
Chapter 3
70
reaction vessel). In situ catalyst formation with a slight excess of ligand (7 mol% BIAN to 5 mol% Pd(OAc)2 gives the highest yield of product (86%) with good selectivity.
3.3.2 Reaction progress monitoring of Pd complexes
To validate the improved conversion and selectivity afforded by BIAN, studies were performed to monitor the progress of the reaction with palladium complexes of dmphen and BIAN. Four complexes; 41a, 41b, 42a, and 42b (Figure 4) were prepared to study the conversion to the Heck product and biphenyl formation in time (Figure 5a) under the optimized conditions. Trifluoroacetate complexes, 41b and 42b, were chosen to investigate the influence of the counterion on the reaction, as it has been proposed that they participate in the regeneration of the active catalyst from the peroxoboronate species in the mechanism put forward by Jung et al.10 The formation of products over time as determined by GC in the reactions catalyzed by complexes 41a, 41b, 42a, and 42b was plotted (Figure 5)
Comparing acetate complexes 41a and 42a, BIAN-based catalyst (42a) turned out to be the enhanced initial rate of the reaction. In the case of the trifluoroacetate complexes 41b and 42b, both complexes displayed similar initial rates. The reaction progress shows a fast initiation of the oxidative Heck reaction, and minimal biphenyl formation with BIAN as the ligand. No induction periods were observed for all the reactions. GC-detectable quantities of biphenyl were formed at the beginning of the reaction, presumably en route to the formation of the active species.
However, the biphenyl concentration remained unchanged in time for BIAN complexes 42a and 42b, while an increase was observed with 41a and 41b. Correlating this result with entries 1 and 2 from Table 1, it can be concluded that very little Pd remains unligated in the case of Pd-BIAN. With complexes 41a and 41b, the rate of biphenyl formation was found to exceed product formation, resulting in a poor overall selectivity for the reaction. 42a And 42b showed comparable activity, but a compelling difference was observed between 41a and 41b. Apparently, either the active catalyst formation is different for dmphen and BIAN, or the reactions have a different rate determining step.
Thus, the use of Pd(O2CCF3)2 in combination with dmphen (41b) leads to a considerably improved catalyst compared to the reported Pd(OAc)2/dmphen combination (41a). However, the BIAN-based catalysts 42a and 42b outperform the dmphen-based catalysts.
Oxidative Heck reactions of cyclic enones
71
a Conversion measured by GC.
Figure 5: Progress of reaction in time with complexes 41a, 41b, 42a, 42b. The graphs indicate
the consumption of the starting material (green), the formation of product (red) and biphenyl
(blue).
42b
0
1
2
3
4
5
6
7
8
0 200 400 600 800 1000 1200 1400
Time (min)
Co
nce
ntra
tion
(mg
/ml)
42a
-1
0
1
2
3
4
5
6
7
8
0 200 400 600 800 1000 1200 1400
Time (min)
Con
cen
trat
ion
(mg/
ml)
41b
0
1
2
3
4
5
6
7
8
0 200 400 600 800 1000 1200 1400
Time (min)
Con
cen
tra
tion
(mg/
ml)
41a
0
1
2
3
4
5
6
7
8
0 200 400 600 800 1000 1200 1400
Time (min)
Co
nce
ntra
tion
(mg
/ml)
cyclohex-2-en-1-one biphenyl 24
Chapter 3
72
3.3.3 Scope of the reaction
Equipped with the knowledge that complexes 42a and 42b are efficient catalysts, and the fact that in situ formed Pd(OAc)2/BIAN is equally active as complex 41a, we proceeded to explore the scope and limitations of our system. To understand the influence of steric and electronic factors on the reaction, a variety of substituted boronic acids was chosen.
In general, use of meta- and para- substituted phenylboronic acids leads to high to excellent yields (Figure 6). The highest isolated yield (43i, 95%) was obtained with m-nitro phenylboronic acid. m-Tolyl, m-anisyl, p-chloro, and p-fluoro phenylboronic acid gave the corresponding Heck products 43f, 43h, 43e, 43b in yields exceeding 90%. Good yields ( 75%) of 43a, 43c and 43j were obtained with p-tolyl, p-trifluoromethyl, and m-chlorophenylboronic acid, respectively. m-Fluoro-phenylboronic acid gave 43g in 52% yield.
As may be reasoned on the account of steric congestion, use of ortho-substituted phenylboronic acids led to lower yields compared to the corresponding meta or para substituted phenylboronic acids bearing the same functional group (except 43l), irrespective of the electronic factors. This trend was not observed in case of 43l, mostly likely due to the only slight increase in steric hindrance, due to the small difference in the atomic radius of fluorine (0.73 Å) vs. hydrogen (0.53 Å), though the best yield was obtained upon reflux. o-Tolyl-, and the sterically demanding o,o’-xylyl-boronic acid were found to be unreactive at room temperature. At reflux temperature, however, the reactions afford the corresponding Heck products, 43k and 43n in moderate yields. o-Nitrophenylboronic acid did not react even upon reflux.
The scope of the reaction with regard to Michael acceptors was also investigated (Figure 7). Acrylates afforded the corresponding Heck products 44a, 44b in a high yield (85%). Due to its volatility, a two-fold excess of methylvinyl ketone with respect to phenylboronic acid was employed to furnish 44c in high yield (87%). Boc-protected 2,3-dihydropyridin-4(1H)-one, an important building block for alkaloid synthesis, gave the corresponding Heck product 44d in good yield (74%).
Oxidative Heck reactions of cyclic enones
73
Figure 6: Boronic acids used in the oxidative Heck reaction of cyclohexenone.
aPd(OAc)2 (5 mol%), BIAN (7 mol%), cyclohex-2-en-1-one (1 mmol), boronic acid (1.5 mmol), O2, 2 mL MeOH : water (9:1), rt, 30 h. Isolated yields b Reaction carried out at reflux.
Chapter 3
74
Figure 7: Scope of Micheal acceptorsa
a Pd(OAc)2 (5 mol%), BIAN (7 mol%), enone (1 mmol), phenylboronic acid (1.5 mmol), O2, 2 ml MeOH : H2O (9:1), rt, 30 h. Isolated yields. b Methylvinyl ketone (2 mmol), phenylboronic acid (1 mmol) used.
3.4 Mechanistic insights
It was intriguing to observe that Pd(OAc)2/BIAN gives high yields in the oxidative Heck reaction also when electron poor phenylboronic acids are employed. Based on experimental observations and DFT calculations, Larhed and coworkers8 proposed that the oxidative Heck reaction catalyzed by bis-nitrogen ligated Pd species proceed via a cationic mechanism. A consequence would be that electron-poor boronic acids react slower compared to electron-rich boronic acids. It was therefore expected that electron-poor boronic acids would perform poorly in this reaction, and this was indeed found to be the case with dmphen ligand.
A convenient test for the presence of cationic species in a catalytic reaction is to study the influence of added halide salts. Cationic metal species react with halides to form neutral complexes, which should arrest the progress of the reaction. Larhed and coworkers concluded from these experiments that the oxidative Heck reaction proceeds via a cationic pathway, and explained the poor conversions with electron poor boronic acids, as a consequence.8
In our hands as well, the addition of halide salts resulted in a dramatic drop in reactivity of the catalyst (Table 3). This was found to be the case irrespective of the counterion of the halide (Table 3, entries 2, 5, 6). A large excess of LiCl inhibited the formation of 24 completely, even at reflux (entry 1). When 1 equiv of LiCl with respect to the substrate was added, the reaction was retarded at rt (entry 2), but proceeded with moderate conversion at reflux. 5 mol% (1 equiv with respect to Pd) of LiCl did not hinder the reaction significantly (entry 3). When salts without halides were added, the catalytic activity was unaffected (entries 4, 5 vs 7, 8).
Oxidative Heck reactions of cyclic enones
75
Table 3: Influence of added halides
Entry Additive Quantity Conv. (%)a
1 LiCl 10 equiv < 2
2b LiCl 1 equiv < 2
3 LiCl 5 mol% 80
4 LiBr 1 equiv < 2
5 ZnCl2 1 equiv < 2
6 Bu4NCl 1 equiv < 2
7 Li(OTf) 1 equiv 78
8 Zn(OTf)2 1 equiv 74
a Determined by GC, with dodecane as internal standard. b Proceeds to moderate conversion upon reflux.
It remains, therefore, inconclusive why on the one hand electron poor arylboronic acids react well in the Pd-BIAN catalyzed oxidative Heck reaction whereas on the other hand a cationic intermediate is most probably involved. One might speculate that in case of BIAN, the charge is better delocalized on to the ligand, leaving the metal center more electron rich than in case of dmphen, though both complexes are cationic per se. This might, therefore, enable the BIAN complex to undergo transmetalation conveniently with electron poor boronic acids as well.
An alternative explanation might come from the fact that in the absence of competitive side reactions (for instance C-H activation of the proximal methyl group in case of dmphen), the catalyst has a longer life-time, enabling it also to react with electron poor boronic acids.
3.5 Conclusion and future perspectives
This study demonstrates that Pd(OAc)2/34 is an excellent catalyst for the base-free oxidative Heck reaction of arylboronic acids with cyclic enones at room temperature. The mild reaction conditions permit the use of a diverse range of electron-rich and electron-poor arylboronic acids and Michael acceptors. Time-variant reaction data show the high selectivity and reaction rate of BIAN ligand 34
Chapter 3
76
compared to commonly used catalysts like 2,9-dimethylphenanthroline (dmphen). This study represents a novel and selective route to many of the described compounds, as most of them have not been synthesized by Heck procedures previously, and only observed as side products in conjugate addition reactions. From a mechanistic perspective, evidence suggests that there is a cationic intermediate involved during the reaction.
Despite the success of this reaction, a few aspects remain to be investigated. While arylboronic acids are very successful in this reaction, alkenylboronic acids are hardly reactive under the current reaction conditions. It also remains to be seen how this reaction performs with other classes of Michael acceptors.
Oxidative Heck reactions of cyclic enones
77
3.6 Experimental
3.6.1 General: All reactions were performed in flame dried Schlenk tubes,
equipped with a Teflon-coated magnetic stir bar and screw cap. A manifold
permitting switching between dinitrogen atmosphere and vacuum was used to
control the atmosphere in the reaction vessel, while oxygen was made available via
a balloon connected to the side-arm of the Schlenk. The Schlenk tube was flushed
thrice with oxygen before capping the vessel.
Flash chromatography was performed using Merck silica gel type 9385 (230-400
mesh). Where ever an isolated yield was obtained, the reactions were performed
on at least 1 mmol scale. TLC was performed using pre-coated silica gel plates
(0.25 mm thick, 60 F254). The compounds were visualized under UV light (254 nm)
or by staining with Seebach Stain (25 g phosphomolybdic acid, 7.5 g cerium (IV)
sulfate, 500 ml H2O, 25 ml H2SO4) or a KMnO4 dip. All solvents used for extraction,
filtration and chromatography were of commercial grade, and used without further
purification.
All the chemicals, except BIAN (see synthesis below) were sourced from
commercial sources, and were used without further purification. 2,9-dimethyl-1,10-
phenanthroline (dmphen) was recrystallized before use. Methanol was distilled
prior to use and mixed in 9:1 (v/v) ratio with water.
1H- and 13C-NMR were recorded on a Varian AMX400 (400, 100.59 MHz,
respectively) using CDCl3 as solvent, unless specified otherwise. Chemical shift
values are reported in ppm with the solvent resonance as the internal standard
(CHCl3: 7.27 for 1H, 77.1 for 13C). Data are reported as follows: chemical shifts
( ), multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, br = broad, m =
multiplet), coupling constants J (Hz), and integration. GC-MS measurements were
made using a HP 6890 Series Gas Chromatograph system equipped with a HP
5973 Mass Sensitive Detector.
GC measurements were made using a Shimadzu GC 2014 gas chromatograph
system bearing a HP-5 column (30 m × 0.25 mm ID) and FID detection.
Chapter 3
78
Conversions were calculated from multi-point calibration curves, with dodecane as
internal standard. The values for the plotted data points were the average of two
reproducible runs. Time- variant measurements were made studied by adding the
internal standard prior to the onset of the reaction, and conversions were
determined by normalizing the integrated value of the internal standard against
other runs in the set. After sampling the reaction mixture (via syringe), the sample
was filtered over a plug of silica and diluted with MeOH, prior to analysis by GC.
3.6.2 General procedure oxidative Heck reaction
General Procedure A: To a Schlenk tube equipped with a magnetic stirring bar
and a septum was added palladium acetate (11 mg, 5 mol%, 0.05 equiv), and
BIAN (28 mg, 7 mol%, 0.07 equiv), under and atmosphere of dinitrogen. The
mixture was dissolved in 2 ml of a solution of MeOH : H2O (9:1) and stirred for 30
minutes at room temperature, followed by the addition of the olefin (1.0 equiv) and
the boronic acid (1.5 equiv). The Schlenk tube was then connected to an oxygen
balloon on the side arm, and oxygen was flushed thrice through the flask, via
cycles of evacuation and oxygen backfilling, and remained connected to the
oxygen balloon. Upon addition of the oxygen the reaction mixture turned dark red-
black. The reaction mixture was allowed to stir at room temperature. When the
reaction was judged complete (by TLC, GC), the reaction mixture was diluted with
diethyl ether and filtered though a pad of silica. The reaction mixture was then
concentrated in vacuo and loaded directly on the column, and the desired product
was isolated.
General procedure B: Identical to General Procedure A, except that the Schlenk
tube was placed in an oil bath at 80 °C, after being backfilled with oxygen.
3.6.3 Synthesis of Bis(aryl)acenaphthequinonediimine (BIAN, 34)
Procedure adapted from literature.31,32 To an
ovendried 100 ml doublenecked round-bottom
flask, under nitrogen, containing a magnetic stir-
bar was added acenaphthoquinone (500 mg,
2.74 mmol, 1.0 equiv) and dry zinc chloride (1 g,
N N
Oxidative Heck reactions of cyclic enones
79
7.34 mmol, 2.7 equiv). These compounds were suspended in glacial acetic acid
(20 ml) turning the reaction yellow. After 5 minutes of stirring at 60 °C, 2,6-
dimethylaniline (0.75 ml, 6.30 mmol, 2.3 equiv) was added and the mixture was
stirred at reflux for 1 hour. During this time the reaction changed from yellow to red
and back again. The mixture was then vacuum filtered while still hot and the
residue was washed with ether (2 x 10 ml) to provide the zinc chloride complex
(1.30 g, 2.47 mmol, 90%) as a yellow solid. The zinc chloride complex was
suspended in CH2Cl2 (40 ml) in a 100 ml separatory funnel. A solution of sodium
oxalate (453 mg, 3.38 mmol, 1.25 equiv) in water (20 ml) was added. After washing
for 5 min, a white precipitate was formed in the water layer. The layers were
separated and the organic layer was washed with water (2 x 10 ml), dried over
MgSO4 and concentrated in vacuo. The obtained red oil was further dried under
reduced pressure. The desired product 34 (0.95 g, 2.44 mmol, 99%) was obtained
as a red foamy solid, with an overall yield of 90%. 1H-NMR (400 MHz, CDCl3)
(38) Prakash, G. K. S.; Chacko, S.; Panja, C.; Thomas, T. E.; Gurung, L.; Rasul, G.; Mathew, T.;
Olah, G. A. Adv. Synth. Catal. 2009, 351, 1567.
(39) Gogoi, A.; Bora, U. Synlett 2012, 23, 1079.
(40) Xu, J.; Wang, X.; Shao, C.; Su, D.; Cheng, G.; Hu, Y. Org. Lett. 2010, 12, 1964.
(41) Zou, Y.-Q.; Chen, J.-R.; Liu, X.-P.; Lu, L.-Q.; Davis, R. L.; Jørgensen, K. A.; Xiao, W.-J.
Angew. Chem. Int. Ed. 2012, 51, 784.
(42) Izawa, Y.; Pun, D.; Stahl, S. S. Science 2011, 333, 209.
(43) Carrow, B. P.; Hartwig, J. F. J. Am. Chem. Soc. 2011, 133, 2116.
(44) Chen, I. H.; Yin, L.; Itano, W.; Kanai, M.; Shibasaki, M. J. Am. Chem. Soc. 2009, 131, 11664.
(45) Palais, L.; Alexakis, A. Chem. Eur. J. 2009, 15, 10473.
(46) Kehrli, S.; Martin, D.; Rix, D.; Mauduit, M.; Alexakis, A. Chem. Eur. J. 2010, 16, 9890.
(47) Ege, S. N.; Yates, P. Can. J. Chem. 1967, 45, 2933.
(48) Kantchev, E. A. B.; Peh, G. R.; Zhang, C.; Ying, J. Y. Org. Lett. 2008, 10, 3949.
(49) Fraley, M. E.; Rubino, R. S. Tetrahedron Lett. 1997, 38, 3365.
Chapter 3
90
(50) Beaulieu, E. D.; Voss, L.; Trauner, D. Org. Lett. 2008, 10, 869.
(51) Ye, Z.; Chen, F.; Luo, F.; Wang, W.; Lin, B.; Jia, X.; Cheng, J. Synlett 2009, 2198.
Chapter 4
Palladium-catalyzed benzylic quaternary stereocenter formation via conjugate addition to cyclic enones
In this chapter, the study of a novel catalytic system is presented, that delivers benzylic all-carbon stereocenters by conjugate addition of arylboronic acids to -substituted cyclic enones in good yields and excellent enantioselectivities. The catalyst is synthesized in a single step from PdCl2 and PhBOX, and activated in situ by AgSbF6.
Parts of this chapter have been published:
Gottumukkala, A. L.; Matcha, K.; Lutz, M.; de Vries, J. G.; Minnaard, A. J. Chem. Eur. J., 2012, 18, 6097.
Chapter 4
92
4.1 Introduction
The selective synthesis of compounds with all-carbon quaternary stereocenters is an important challenge in organic chemistry.1,2 Transition metal-catalyzed reactions offer convenient alternatives for the synthesis of these motifs in high yields and selectivities.
While conserving focus on the construction of benzylic quaternary stereocenters by transition metal catalysis, four reactions have gained much acclaim. These include, though are not limited to:
a) -Arylation of substituted ketones3-5 delivers a quaternary stereocenter in the position of a carbonyl moiety, most commonly a ketone.
b) Allylic alkylation6-8 forms a quaternary stereocenter by positioning an alkyl substituent at the benzylic position of an aryl substrate, along with the disposition of the double bond.
c) Asymmetric Heck9-12 reaction affords a quaternary stereocenter via a newly formed benzylic carbon-carbon bond, combined with the shift of the double bond away from the newly formed stereocenter.
d) Conjugate Addition13,14 installs a quaternary stereocenter at the -position to a carbonyl moiety, along with the loss of unsaturation.
Of the above mentioned, conjugate addition to enones presents considerable advantage over other methods for further molecular modification, namely the generation of two stereochemically distinct -positions, and an unconjugated ketone function.
4.1.1 Conjugate Addition to -substituted enones
Transition metal-catalyzed conjugate addition of organometallics to -substituted enones has been exploited considerably in recent years to afford quaternary stereocenters.13,15,16 Of particular mention are the reactions catalyzed by copper17-
27 and rhodium.28-33 Application of these methodologies in synthesis is often considered to be complimentary, as copper is most effective for the addition of alkyl substituents, whilst rhodium works most effectively for the addition of aryl or alkenyl substituents. While organometallics like alkylzinc reagents, Grignard reagents are amenable to Cu catalysis; aryl or alkenyl boronic acids, esters, borates or their derivatives are better suited for rhodium catalysis.
Conjugate additions to cyclic enones
93
Each of these reactions presents its own set of advantages and disadvantages, which are summed up in Table 1.
Table 1. A qualitative comparison of Cu and Rh catalyzed conjugate addition reactions.
Palladium catalysis offers several opportunities to overcome the disadvantages of either the copper or the rhodium systems. Most significantly, Pd catalyzed reactions can be performed at ambient temperatures (40 – 80 oC), without a rigorous exclusion of air or moisture from the reaction. Further, Pd is considerably cheaper than Rh, facilitating large-scale application. In addition, Pd has been demonstrated to be successful for substrates with increased steric hindrance close to the -position of the enone, which still remains a challenge for Rh catalysis.34
Arylboronic acids35 have proven to be the most convenient organometallics for the introduction of an aryl moiety via transition metal catalysis. Their ease of handling, long shelf-life, stability to air and moisture under ambient conditions; combined with their large commercial availability of derivatives has compounded their wide scale application. Further, the reactivity of arylboronic acids can be tuned by the
Copper Rhodium
Advantages Disadvantages Advantages Disadvantages
Inexpensive Requires reactive
organometallics
Ambient reaction
conditions Very expensive
Highly successful
for the introduction
of alkyl groups
Has limited functional
group tolerance
Highly successful for the
introduction of aryl and
alkenyl groups, along
with a broad functional
group tolerance
Rigorous exclusion of air
and moisture is required
Does not require
rigorous exclusion of air
and moisture
Low temperatures (-50 oC
and lower) prohibit large-
scale application
Chapter 4
94
formation of the corresponding boroxines,35 trihydroxyborates 36, trifluoroborates37 or MIDA boronates.38,39 It is therefore that these reagents have been greatly exploited in transition metal catalysis.35
In 2010, the group of Lu40 disclosed the first example of Pd-catalyzed conjugate addition of arylboronic acids to -substituted enones. Using only 0.5 mol% of 4 (1 mol% of Pd), at room temperature, isolated yields up to 96% were obtained (Scheme 1).
Scheme 1: The first example of Pd-catalyzed quaternary center formation via conjugate addition.
During the course of our own investigation into the development of an enantioselective variant of this reaction, the group of Stoltz41 reported a competent procedure (Scheme 2) using the easily synthesized phenyloxazoline ligand 5. While the catalyst displays a good substrate scope, both in terms of enones and arylboronic acids, it was reported to be ineffective for linear substrates and produced biphenyl as a significant byproduct.
Following our investigation, the group of Lee34 reported a mild, ligand-free procedure for conjugate addition reactions using aryl boroxines, in the presence of sodium nitrate and triflic acid (Scheme 3). The reactions could be performed in air, and proceeded with high diastereoselectivity.
Conjugate additions to cyclic enones
95
Scheme 3: Highly diastereoselective conjugate addition using arylboroxines.
The groups of Li and Duan42 reported a Pd catalyzed procedure for the desulfitative coupling of arylsulfinic acids with enones, leading to conjugate addition products in moderate to good yields (Scheme 4). The extrusion of SO2
during the reaction required higher temperatures, whereas in cases in which the enone was not -substituted, the corresponding Heck product was observed as a side-product.
4.2 Goal The goal of the study was to develop a Pd-catalyzed reaction for the enantioselective formation of benzylic quaternary stereocenters, via conjugate addition of arylboronic acids to , -disubstituted enones. This aims to combine the convenience of using a relatively inexpensive metal (Pd), along with readily available arylboronic acids, under conditions which do not require the strict exclusion of air or moisture, for the formation of benzylic quaternary centers with high stereoselectivity. To facilitate easier application, we aimed to do this with commercially available compounds, as much as possible.
One could readily see the importance of this strategy in organic synthesis, as was readily seen in the shortest known synthesis of (–) - α-cuparenone was achieved.
4.3 Results and discussion
Given the importance of the reaction in the construction of all-carbon benzylic quaternary centers, the new insights provided by the work of Lu40 and our own insights into the enantioselective conjugate addition of arylboronic acids to enones forming tertiary centers catalyzed by cationic Pd species,43 we pursued a study to
Chapter 4
96
develop a novel Pd-catalyzed route to achieve all-carbon benzylic quaternary centers.
4.3.1 Optimization of reaction parameters
As a benchmark, the addition of phenylboronic acid (2) to 3-methyl-cyclohexenone (1) was investigated (Table 2) in the presence of various Pd precursors and chiral ligands (Figure 1). Reproducing the reaction of Lu, using catalyst 440, gave the desired racemic compound, 3-methyl-3-phenylcyclohexanone (3a), in an isolated yield of 92% (entry 1).
Figure 1. Ligands and complexes studied.
Conjugate additions to cyclic enones
97
Our quest for an optimal catalyst system for the reaction began with MeDuPhos (11)/Pd(O2CCF3)2 (entry 3), which we previously disclosed to be highly successful in the conjugate addition of arylboronic acids to cyclohexenone.43 However, the conversion obtained was very poor, despite an excellent ee (96%). Variations of temperature and solvent in this system did not prove beneficial.
Table 2: Optimization of reaction parameters.
Reaction conditions: methylcyclohexenone (1 equiv), phenylboronic acid (1.5 equiv), Pd precursor, ligand, additive, 40 oC, solvent, 24 h. a MeOH as solvent. b THF:water (10:1) as solvent c Acetone as solvent d MeOH:water (4:1) as solvent. e Conversion determined by GC / NMR analysis of the crude
reaction mixture. f Determined by chiral GC analysis of the crude reaction mixture. g -chloro- -phenyl-acetophenone.44
Analysis of other chiral phosphines, phosphoramidites, and cationic Pd complexes thereof, also met with disappointing results. This was, therefore, followed by a screen of various chiral nitrogen ligands.
Interestingly, bidentate amine BINAM (13) provided a significant level of enantioinduction (72% ee), albeit with poor conversion (16%). Cinchonine (14) or diaminocyclohexane-based 15 were found to be ineffective for the reaction (entries 7,8). Compelling improvements began to appear only when bisoxazoline (BOX) ligands were studied. Using 16 as ligand, a jump in conversion from 22% to 51% whilst maintaining a high ee (92%) was observed when methanol was substituted with acetone as the solvent (entries 9, 10). This result, together with the observation of Pd black when the reaction was performed in methanol, led us to believe that the reduction of PdII to Pd0 by MeOH was the cause of the poor conversions.45
Since Pd0 is catalytically inactive for the transformation, we sought methods to re-oxidize Pd0 to PdII. Amongst the potential oxidants, notable improvement in conversion was observed when 20 mol% of Cu(BF4)2·6H2O was added. However, this came at the cost of complete loss of enantioselectivity, probably because 16 was scavenged from the palladium by CuII (entry 11). This warranted the addition of 27 mol% of 16 in subsequent runs, to maintain the excellent enantioselectivity (97% ee) with 67% conversion (entry 12). Organic oxidants such as benzoquinone or desyl chloride46 ( -chloro- -phenyl-acetophenone) completely retarded the reaction (entries 16, 17).
Other BOX ligands showed no prominent improvement in conversion or ee (entries 13–15), and the reaction was found to be completely blocked by 18 (entry 14). At this stage, we scouted for alternative approaches to generate cationic Pd species in situ, with the goal of maintaining the high enantioinduction by 16 and to improve the conversion. A reliable protocol to this end is the dehalogenation of metal halides with silver salts, in situ.
Scheme 5: Synthesis of PdCl2 – PhBOX complex (21).
Conjugate additions to cyclic enones
99
We synthesized complex 21 (Scheme 5) by refluxing PdCl2 and 16 in acetonitrile for 3 h. Switching to MeOH : water (4:1) as the solvent and in situ dehalogenation with AgBF4 resulted in a significant increase in conversion to 74% (entry 18). A decisive improvement leading to full conversion however; with 96% ee in 24 h (entry 19), was achieved by employing AgSbF6 as the silver salt! It is very fortunate that this superior catalyst is composed of commercially available and affordable PhBOX, PdCl2 and AgSbF6. The role of SbF6
- as the counterion seems to be critical for the increased activity of the catalyst. Such improvements in activity of the catalyst by changing the counterions have been observed previously.47
Replacing PhB(OH)2 with PhBF3K (entry 20) gave poor conversion (<10%) although the ee remained excellent (97%). Substituting phenylboronic acid with potassium trihydroxyboronate36 (entry 21) or phenylboronic acid MIDA ester38 gave no reaction (entry 22).
The X-ray crystal structure of catalyst 21 is presented in Figure 2. The structure confirms the square planar geometry of PdII. The stereodirecting phenyl groups point above and below the plane, providing the chiral environment around the metal center.
Figure 2: X-Ray structure of the PdCl2 – PhBOX complex (21).48
Chapter 4
100
4.3.2 Substrate scope of enones.
Equipped with the (21)/AgSbF6 system, we advanced to investigate the scope of the new reaction (Table 3). We were delighted that the catalyst system was highly successful in the addition of phenylboronic acid to five-, six- and seven-membered enones, delivering excellent enantioselectivities uniformly along with high isolated yields. The reaction tolerates alkyl substituents at the position, but not aryl substituents (entries 2, 3 vs entry 4). Overall, this coincides with the results of Stoltz et al., but the ees with the current (21)/AgSbF6 catalyst are substantially higher.
Next to carbocyclic enones, heterocycles possessing quaternary stereocenters are key building blocks in the synthesis of natural products and pharmaceuticals. In general, however, their synthesis is challenging and procedures using catalytic asymmetric conjugate additions are limited. Lactone 25 turned out to be a suitable substrate for the catalyst system. Although the desired product 25a was isolated with a reduced 57% yield, it had a respectable 88% ee (entry 6). Cyclic ether 26a was obtained in 28% yield and 69% ee, starting from 6-methyl-2H-pyran-4(3H)-one 26 (entry 7). The lower activity may be a consequence of the poor electrophilicity of 26, formally a vinylogous ester, arising from electron donation by the lone pair of the ring oxygen. Next we also studied two substrates (27 and 28) bearing a substituted nitrogen atom instead of an oxygen atom. However, both the substrates were recovered unreacted.
Compared to cyclic enones, acyclic enones are well-known to be considerably more challenging substrates for conjugate addition, and the current reaction is no exception.6 Substrate 29 underwent rapid E / Z isomerization, combined with a poor conversion, resulting in low yield and ee (entry 10). Enone 30 met a similar fate (entry 11). Realizing that success of the asymmetric addition to acyclic substrates rested in controlling this isomerization, several substrate classes were designed and studied. We were fortunate to discover that an allylic ether function in the substrate arrested the isomerization and led to considerably higher yields of the desired products. Compound 31a was obtained in a high yield of 84% from 31 though with only 23% ee (entry 12). Interestingly, benzyl ethers E-32 and Z-32 provided the same enantiomer of the product 32a, suggesting dominant catalyst control over the reaction, although a higher ee was obtained starting from Z-32 (entries 13, 14). This stands in contrast to observations in the corresponding Rh-catalyzed conjugate additions, wherein E and Z isomers of the substrate provide the opposite enantiomers of the product.28 Following the reaction by GC over time, showed no isomerization of E and Z isomers during the reaction.
Conjugate additions to cyclic enones
101
Table 3. Substrate scope of enones for the reaction.
Entry Substrate Product Yield (%)a ee (%)b
1
21
21a 93 93
2
2
3 95 96
3
22
22a 91 99
4
23 -- -- -- --
5
24
24a 80 94
6c
25
25a 57 88
Chapter 4
102
Entry Substrate Product Yield (%)a ee (%)b
7
26
26a 28 69
8
27 -- -- -- --
9
28 -- -- -- --
10d
29
29a 14 8
11 d
30 -- <10 nd
12 d
31
31a 84 23
13 d
E-32 81 25
14 d
Z-32 32a
78 36
15 d
33 53 51e
16 d
34
33a
68 27e
Conjugate additions to cyclic enones
103
Reaction conditions: enone (0.5 mmol), phenylboronic acid (0.75 mmol), 21 (8 mol%), AgSbF6 (20 mol%), MeOH : water (4:1) 2.5 ml, 40 oC, 24 – 40 h. a Isolated yields. b Determined by chiral GC / HPLC analysis of the isolated products.c 3 Equiv of boronic acid added. d Reaction performed at 60 oC. e ee determined by ring opening of the acetal, see supporting information. TBDPS = tert-butyldiphenylsilyl; Bn = benzyl, TIPS = triisopropylsilyl.
The use of trityl ether 33 (entry 15) resulted in the formation of acetal 33a, due to cleavage of the trityl moiety following conjugate addition. This is probably due to the acidic environment of the reaction. An improved ee of 51% was observed, however. The ee determination of 33a could be performed by ring opening the acetal (See experimental section). Silyl ethers 33 and 34 also suffered cleavage of the protecting group. Reaction of 34 resulted in the formation of 33a, but TBDPS bearing 35a could be isolated from 35 in a reduced yield of 38%, next to acetal 33a (9%), and with a respectable ee of 60% (entry 17). The cleavage of the silyl ethers is probably due to the presence of fluoride ions, produced by the decomposition of SbF6
–.
Realizing the comparatively poor levels of enantioinduction in the linear substrates, we devised a complementary route to these products via the ring opening of lactone 25a. To this end, Weinreb amide 36 was obtained in 91% yield. 36 offers several convenient handles for further modification (Scheme 6).
Scheme 6. Access to acyclic compounds with quaternary stereocenters via ring-opening of a lactone.
4.3.3 Arylboronic acids
Further, we examined the scope of arylboronic acids that could be employed in the reaction. This scope turned out to be broad, with arylboronic acids bearing electron
Entry Substrate Product Yield (%)a
ee (%)b
17 d
35
35a 38 60
Chapter 4
104
Table 4. The conjugate addition of arylboronic acids to 3-methyl-cyclohexenone.
Entry Arylboronic acid Product Yield (%)a
ee (%)b
1
3b 89 97
2
3c 96 97
3
-- -- --
4
3d 44 93
5
3e 30 98
6
3f 98 >99
7
3g 85 98
Conjugate additions to cyclic enones
105
Reaction conditions: 3-methylcyclohexenone (0.5 mmol), arylboronic acid (0.75 mmol), 21 (8 mol%), AgSbF6 (20 mol%), MeOH :water (4:1) 2.5 ml, 40 oC, 24 – 40 h. a Isolated yields. b Determined by chiral GC / HPLC analysis of isolated products. c Did not proceed to completion even after 72 h. d Reaction performed at 60 oC.
donating groups performing the best. Enantiomeric excesses higher than 95% are regularly observed with alkoxy, alkyl or halide substituted phenylboronic acids (Table 4). The steric encumbrance of ortho substituents was found to block the reaction (entry 3). Once again, the enantioselectivities delivered herein are repeatedly higher than those reported earlier, for the same substrates.41 Of particular mention is 3g for which a 29% improvement in enantioselectivity was observed. Ferrocenylboronic acid, however, was found to be unreactive, even at 60 oC (entry 10). Alkenylboronic acids were found to be ineffective for the reaction, under these reaction conditions (entries 11,12).
Entry Arylboronic acid Product Yield (%)a ee (%)b
8
3h 98 96
9
3i 88 98
10d
-- -- -- --
11
-- -- -- --
12 -- -- -- --
Chapter 4
106
4.4 Synthesis of (–)- -cuparenone
As an example of the impact of this new reaction in synthesis, we completed the synthesis of sesquiterpene (–)- -cuparenone, isolated from Thuja orientalis. 38 Was previously synthesized49-53 by considerably longer routes. The shortest route prior to our synthesis, was 5 steps long,54 starting from non-commercially available compounds, while the longest route was 17 (!) steps.53
Our strategy involved only 2 steps, starting from materials that are readily available commercially. Starting from 3-methylcylopentenone, 37a was obtained in 68% yield and 90% ee. Subsequent regioselective dimethylation could be performed in a single step with excess MeI, by slow addition of KOtBu in THF, to give 38 in a respectable 65% yield, a considerable improvement compared to literature (see Experimental section). The selectivity obtained for the gem-dimethylation is remarkable, and may be explained on the account of coordinative assistance by the aryl group in directing the deprotonation with KOtBu.55
Scheme 7. A two-step synthesis of (–)- -cuparenone.
4.5 Conclusion and future perspectives In summary, we describe herein (21)/AgSbF6 as an efficient catalytic system for the conjugate addition of arylboronic acids to , -disubstituted enones. Benzylic quaternary stereocenters are formed in excellent yields (up to 98%) and enantioselectivities (up to 99%) for carbocyclic enones, and with appreciable success for heterocyclic (up to 88% ee) and acyclic enones (up to 60% ee). As this is the first example of a single catalytic system, able to address all these classes of substrates, this breaks the ground to further developments in this direction.
An allylic ether moiety was found to be a key factor in the addition to acyclic substrates. However, the exact role of the allylic ether in the reaction is not completely understood.
The catalyst system has also been employed in the shortest synthesis to date of (–)- -cuparenone. In addition to the convenient quaternary center synthesis, the selective gem-dimethylation in a single step is practical for synthesis.
Conjugate additions to cyclic enones
107
The remarkable influence of SbF6– as a counterion remains to be understood.
Further studies could focus on the developing conditions for the application of alkenyl and heteroaromatic boronic acids.
Chapter 4
108
4.6 Experimental Section
4.6.1 General:
All experiments were carried out in flame dried or oven dried (150 oC) glassware, in
an atmosphere of dinitrogen, unless specified otherwise, by standard Schlenk
techniques. Schlenk reaction tubes with screw caps, and equipped with a teflon-
coated magnetic stir bar were flame dried under vacuum and allowed to return to
room temperature prior to being charged with reactants. A manifold permitting
switching between dinitrogen atmosphere and vacuum was used to control the
atmosphere in the reaction vessel. Reaction temperature refers to the temperature
of the oil bath.
Flash chromatography was performed using Merck silica gel type 9385 (230-400
mesh), using the indicated solvents, or using Grace Revelaris® automatic column
machine, using silica gel column cartridges. Detection from the machine was
performed using UV. All solvents used for extraction, filtration and chromatography
were of commercial grade, and used without further purification. Anhydrous
methanol, and acetonitrile were sourced from Sigma-Aldrich or Acros and stored
under dinitrogen. Reagents, ligands and complex 11 were purchased from Sigma-
Aldrich, Strem or Acros and used without further purification. Silver
hexafluoroantimonate was stored in a nitrogen dry-box. Bisoxazoline ligands 16-19
were stored at -20 oC.
Complex 21 was prepared as described below and stored under ambient
conditions. TLC was performed on Merck silica gel 60, 0.25 mm plates and
visualization was done by UV (210 nm) and staining with Seebach’s reagent (a
(47) For examples when the hexafluoroantimonate ion has a remarkableinfluence on catalysis,
see: a) Johnson, J. S.; Evans, D. A. Acc. Chem. Res. 2000, 33, 325. b) O'Mahony, D. J. R.;
Belanger, D. B.; Livinghouse, T. Synlett 1998, 443.
(48) CCDC-855151 contains the supplementary crystallographic data forthis paper. These data
can be obtained free of charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
(49) Shivakumar, I.; Salunke, G. B.; Kumar, S. Synth. Commun. 2011, 41, 1952.
(50) Natarajan, A.; Ng, D.; Yang, Z.; Garcia-Garibay, M. A. Angew. Chem., Int. Ed. 2007, 46,
6485.
(51) Chavan, S. P.; Dhawane, A. N.; Kalkote, U. R. Tetrahedron Lett. 2007, 48, 965.
(52) Chavan, S. P.; Dhawane, A. N.; Kalkote, U. R. Synthesis 2007, 3827.
(53) Asaoka, M.; Takenouchi, K.; Takei, H. Tetrahedron Lett. 1988, 29, 325.
(54) Zhang, P.; Le, H.; Kyne, R. E.; Morken, J. P. J. Am. Chem. Soc. 2011, 133, 9716.
(55) Posner, G. H.; Lentz, C. M. J. Am. Chem. Soc. 1979, 101, 934.
(56) Palais, L.; Alexakis, A. Chem. Eur. J. 2009, 15, 10473.
Chapter 5
Facile construction of benzylic quaternary centers via palladium catalysis
This chapter details our study to develop a straightforward, practical and scalable protocol for the palladium catalyzed conjugate addition of arylboronic acids to , -disubstituted enones, leading to the formation of benzylic quaternary centers. Compared to existing literature, this procedure avoids the synthesis of preformed catalysts, or forcing conditions. Further, the role of KSbF6 as an additive to obtain increased yields in case of acyclic enones is studied.
Parts of this chapter will be submitted for publication: Gottumukkala, A.L., Suljagic, J., Matcha, K., de Vries, J. G., Minnaard, A. J., Manuscript in preparation.
126
Chapter 5
5.1 Introduction
5.1.1 Occurrence and non-catalytic synthesis
The benzylic quaternary center1-5 is a widely prevalent motif in a variety of natural products,5,6 drug candidates7 and fragrances. A few examples of molecules containing this motif are presented in Figure 1.
Figure 1: Examples of natural products bearing a benzylic quaternary center
The synthesis of this architectural element has remained a challenge and continues to be an actively studied area over the past decade.1,2 The convenience rendered by a selective and mild approach to install these units is so significant that it can shorten synthesis routes considerably and greatly increase yields. A clear example of this is the synthesis of (-)-α-cuparenone, in which this seemingly simple target, by non-catalytic approach, required a linear sequence of 16 steps to synthesize the molecule.8 A straightforward conjugate addition, on the other hand, allowed the synthesis of the same compound in merely 2 steps (Section 4.5, Chapter 4).
Conjugate addition to enones is a particularly facile approach for building molecular diversity as it offers several convenient handles for functionalization. Upon conjugate addition, the molecule is left with two α positions that are regiochemically distinct, in addition to the ketone moiety; all of which can be targeted selectively for further chemical transformations (Scheme 1).
Scheme 1: Handles for selective functionalization following conjugate addition.
127
Facile Construction of Benzylic Quaternary Centers via Pd catalysis
Commonly, the synthesis of these motifs via conjugate addition, under uncatalyzed conditions is achieved by Gilman reagents9 (Scheme 2).
Scheme 2: Gilman reagents applied in uncatalyzed conjugate addition reactions.
Transition metal catalysis offers a convenient alternative to the above. Conjugate addition reactions catalyzed by copper10-21 and rhodium22-25 have been well studied in recent years, and are the subject of a recent review.26 The application of these metals in catalysis is often considered to be complementary: whilst copper catalysis is particularly successful for the conjugate addition of alkyl substituents, rhodium catalysis allows facile addition of alkenyl and aryl groups.26 A qualitative comparison of the two approaches has been presented in Section 4.1 (Chapter 4).
5.1.2 Palladium in conjugate addition
For many years, while it was possible to obtain tertiary and quaternary centers via conjugate addition with Cu27,28 and Rh29,30 catalysis, it was thought that Pd catalysis29,31 was limited to the formation of tertiary centers. The first report to dispel this notion came in 2010, when the group of Lu disclosed32 that cationic Pd complex C1 was able to catalyze the formation of 8 (Scheme 3) via the conjugate addition of phenylboronic acid to 3-methylcyclohexenone (1).
Scheme 3: The first Pd-catalyzed benzylic quaternary center formation via conjugate addition.
The reported synthesis of C1 was via the dehalogenation of 9 using AgBF4. It was however, found to proceed with a rather low yield (27 %).
Scheme 4: Synthesis of catalyst C1.
128
Chapter 5
Recently, Lee and coworkers reported33 good diastereoselectivities for the Pd-catalyzed conjugate addition of arylboroxines, using Pd(MeCN)4(OTf)2 or alternatively a combination of Pd(OAc)2 and TfOH, as catalysts (Scheme 5). The authors found that the addition of 1 equiv of NaNO3 was helpful to minimize the formation of biphenyl, though they failed to comment whether NaNO3 functioned as an oxidant of Pd. While the conditions of the reactions are indeed mild, and the diastereoselectivities impressive; the reaction requires 5 mol% of Pd, in addition to prior synthesis of the arylboroxines (via dehydration of the corresponding boronic acids), and the addition of 1 equiv of NaNO3. Further, the catalyst, Pd(MeCN)4(OTf)2 is not readily available and moisture sensitive (it decomposes within 2 min upon exposure to ambient atmospheric conditions).
Scheme 5: Diastereoselective conjugate addition of arylboroxines.
The groups of Li and Duan recently reported34 a desulfitative conjugate addition of arylsulfinic acids with enones (Scheme 6). The authors probed the reaction mechanism via ESI-MS/MS. The intermediates identified were largely consistent with the mechanism put forth by Lu.32 Despite the relevance of this procedure to synthesis, a catalyst loading of 5 mol% was necessary for success of the reaction, in addition to high temperatures (90 oC). Further, the acidic conditions of the reaction limit the functional groups that could be used in this reaction.
Scheme 6: Desulfitative conjugate addition of arylsulfinic acids to enones.
Computational studies disclosed by Houk et al.35, verified the mechanism advanced by Lu.32 DFT calculations explain several empirical observations such as the catalytic activity of hydroxo-palladium species (Scheme 4), the regioselectivity of the addition and the necessity for a cationic Pd species. In addition, the authors rationalized why conjugate addition does not take place when there is an aryl substituent in the -position, as this gives rise to a steep energy barrier of 37 kcal/mol, going to the transition state. Further, the formation of the substitution
129
Facile Construction of Benzylic Quaternary Centers via Pd catalysis
product observed when an alkoxy function is present in the -position, was rationalized on the account of energetically favorable -alkoxy elimination.
5.2 Goal
The goal of this study was to develop a straightforward and easily scalable procedure for Pd catalyzed conjugate addition reactions leading to the formation of benzylic quaternary centers, using low catalytic loadings. We chose to develop a procedure based on arylboronic acids, as these are easily accessible commercially, available with a variety of functional groups, and can be stored and manipulated at ambient conditions. Such a procedure would preferably avoid the dehydration of the arylboronic acids to the corresponding boroxines, and also the combination of acid and high temperature that is required to extrude SO2 in the reported desulfitation reaction.
5.3 Results and discussion
Our initial focus was to study the role of BIAN as a ligand for Pd in the formation of quaternary centers, following its success in the oxidative Heck reaction (Chapter 3). We chose the reaction of 3-methylcyclohexenone with phenylboronic acid to optimize the reaction.
5.3.1 Optimization of reaction parameters for conjugate addition
With the optimized catalyst system of the oxidative Heck reaction but without the oxygen balloon; ie Pd(OAc)2 (5 mol%), BIAN (7 mol%), MeOH : H2O (9:1), rt, 12 h (Table 1, entry 1) no product formation was observed. Performing the same reaction at 40 oC did not have any beneficial influence (entry 2). Subsequently, we assayed cationic sources of Pd, along with the use of dry methanol, as described by Lu (Scheme 3).32
Using Pd(CH3CN)4(BF4)2, 14% conversion was obtained (entry 3). Switching to Pd(O2CCF3)2 gave an improved conversion of 33%. Performing the reaction in a solution of MeOH : H2O (9:1) improved the conversion, though it was still below 50% (entry 5). Interestingly, performing the reaction at higher temperature (60 oC) gave full conversion (entry 6). However, when the catalyst loading was lowered to 1 mol%, conversion was found to drop significantly (entry 7).
130
Chapter 5
Table 1: Reaction of 3- methylcyclohexenone with phenylboronic acida
Entry Pd (mol%) Ligand
(mol%) Solvent
Temp.
(oC)
Conv.
(%)b
1 Pd(OAc)2 (5 ) BIAN (7) MeOH:H2O (9:1) rt 0
2 Pd(OAc)2 (5 ) BIAN (7) MeOH:H2O (9:1) 40 0
3 Pd(CH3CN)4(BF4)2 (5) BIAN (7) MeOH rt 14
4 Pd(O2CCF3)2 (5) BIAN (7) MeOH rt 33
5 Pd(O2CCF3)2 (5) BIAN (7) MeOH:H2O (9:1) rt 43
6 Pd(O2CCF3)2 (5) BIAN (7) MeOH:H2O (9:1) 60 full
7 Pd(O2CCF3)2 (1) BIAN (1.5) MeOH:H2O (9:1) 60 35
8 Pd(O2CCF3)2 (5) Phen (7) MeOH:H2O (9:1) 60 --
9 Pd(O2CCF3)2 (5) Bipy (7) MeOH:H2O (9:1) 60 full
10 Pd(O2CCF3)2 (1) Bipy (1.5) MeOH:H2O (9:1) 60 full
15e Pd(O2CCF3)2 (1) Bipy (1.5) MeOH:H2O (9:1) 60 0
a 3-methylcyclohexenone (0.5 mmol), phenylboronic acid (1mmol), Pd precursor, ligand, solvent (1 ml), 12 h.b Conversion determined by GC analysis. c 9 used instead of 7. d 10 used instead of 7. e 11 used instead of 7.
131
Facile Construction of Benzylic Quaternary Centers via Pd catalysis
At this stage, we assayed other nitrogen based ligands. 2,9-Dimethylphenanthroline proved to be a poor ligand (entry 8). On the other hand, 2,2’-bipyridine gave full conversion even with 5 mol% Pd(OAc)2 (entry 9), and the conversion remained the same when the Pd loading was lowered to 1 mol% (entry 10). Trying the reaction at room temperature resulted in poor conversion (entry 11), and increasing the proportion of water in the solvent was not beneficial (entry 12). Substituting phenylboronic acid by potassium trihydroxyphenylborate36 (entry 13), potassium phenyltetrafluoroborate37 (entry 14) or phenyl MIDA boronate38 (entry 15) did not lead to product formation.
Comparing to the system developed by Lu,32 our protocol avoids the need to synthesize the catalyst, C1, a synthesis which is low yielding. Both the protocols require only 1 mol% Pd, while a higher reaction temperature (60 oC) was found to be necessary in our case.
5.3.2 Conjugate addition to cyclic enones
With the optimized conditions [1 mol% Pd(O2CCF3)2, 1.5 mol% 2,2-bipyridine, 60 oC, MeOH:H2O (9:1)] at hand, we went on to explore the scope and limitations of the reaction (Figure 2). In general, the reaction proceeded well for a series of 3-substituted 5, 6 and 7 membered cyclic enones with variety of arylboronic acids, affording yields higher than 90% in many cases. The reactions proceeded cleanly affording only the product.
Reactions of 3-methylcyclohexenone (1) afforded 3b and 3c in yields over 90% with 4- and 3-tolylboronic acid. 4-Fluorophenylboronic acid gave 3d in a reduced yield of 70%. Alkoxy substituted arylboronic acids afforded 3e and 3f in 72% and 85%, respectively. Whilst 3-chlorophenyl boronic acid afforded 3h in 90% yield, 3-chloro-4-methoxyphenyl boronic acid gave 3g in 95% yield. As observed earlier, (Chapter 4, Table 3, entry 4), 3-phenylcyclohexenone 12 was found to be unreactive, and 3-cyclopentyl-cyclohexenone (13) formed the product only in trace amounts. The failure of these substrates (12,13) in the reaction has been explained more recently by the work of Houk et al. as described earlier.35
3-Methyl-cyclopentenone proved to be an excellent substrate for the conjugate addition of arylboronic acids. Yields for the conjugate addition products of 4- and 3-tolylboronic acids were found to be 96% (14b) and 93% (14c) respectively. 4-fluorophenylboronic gave 14d in 88% yield. Alkoxy substituted phenylboronic acids afforded 14e and 14f in 90% and 93% yield respectively. 14g Could be isolated in 90% yield, indicating that the benzyl protecting group remained intact during the reaction. 3-Methylcycloheptenone afforded 15a, in a respectable 80% yield
132
Chapter 5
Figure 2: Conjugate addition of arylboronic acids to cyclic enones.a,b
a Enone (1 mmol), arylboronic acid (2 mmol) Pd(O2CCF3)2 (1 mol%), 2,2’-Bipyridine (1.5 mol%), MeOH : H2O (9:1) 2 ml, 12 h. b Isolated yields.
Reactions with 3-methylcyclopentenone gave higher yields and often exhibited shorter reaction times (6-8 h), as compared to 3-methylcyclohexenone (12-18 h) with the same arylboronic acid. This may be explained by a release in ring strain of the enone, following conjugate addition. Both electron rich and electron deficient arylboronic acids were amenable to the reaction, though best yields were usually
133
Facile Construction of Benzylic Quaternary Centers via Pd catalysis
obtained with electron rich boronic acids. The applicability of halogenated arylboronic acids for the reaction allows further functionalization via Pd-catalyzed cross coupling reactions.
Despite the impressive yields obtained with a variety of arylboronic acids, certain classes of boronic acids did not result in any product formation. These include heteroaromatic boronic acids, alkenylboronic acids, phenylboronic acids bearing an ortho-substituent or phenylboronic acids substituted with a nitro- or trifluoromethyl substituent and are listed in Figure 3.
Figure 3: Boronic acids that failed to undergo conjugate addition reactions.
Further we were interested to test the scope of our reaction on acyclic systems.
5.3.3 Conjugate addition to acyclic enones
Acyclic substrates have been particularly challenging substrates for conjugate addition reactions. This is often the case because the lack of structural rigidity, which, permits an easy s-cis to s-trans isomerization of the substrate. As a result, most often, the conditions that are successful for cyclic enones are not the same for acyclic enones,39 which was also found to be true in our case.
For our study, we chose the reaction of 26 with phenylboronic acid to optimize the reaction conditions (Table 2). In general, the reactivity of linear substrates was lower than that of cyclic substrates, (reaction times ranging between 18-36 h to attain full conversion, while cyclic substrates required 8-12 h). A higher temperature (80 oC) was also found to be necessary. The use of the optimized conditions for cyclic substrates (Table 2, entry 1) only led to 20% conversion after 36 h. Increasing the catalyst loading to 5 mol% led to an improved conversion, though it remained incomplete, even at higher temperature. At this point, we considered the use of additives to improve the conversion. We found that adding 20 mol% of potassium hexafluoroantimonate (KSbF6) to the reaction afforded full
134
Chapter 5
conversion (entry 4). The choice of KSbF6 was based on our previous studies on the enantioselective conjugate addition to cyclic enones (Chapter 4, Section 4.3.1), in which the use of SbF6
– as a counter ion for the cationic Pd complex resulted in improved yield and selectivity. Therefore, we surmised that adding KSbF6 to the current reaction system could have a similar influence, and this was indeed found to be the case. The improved activity upon the addition of SbF6
– may be attributed to a salt metathesis between the trifluoroacetate (CF3CO2
–) ion and SbF6 – leading
to catalytically more active species. However, lowering the catalyst loading to 1 mol% in the presence of the additive, gave diminished conversions (entry 5), though still higher than in the case where no additive was present (entry 5 vs. entry 1). Hence it was decided to use 5 mol% of Pd for subsequent experimentation.
Table 2: Conjugate addition of cyclic enones.
Entry Pd (mol%) Ligand
(mol%) Additive
Temp.
(oC)
Conv.
(%)
1 Pd(O2CCF3)2 (1) BIAN (1.5) - 60 20
2 Pd(O2CCF3)2 (3) dmphen (5) - 80 38
3 Pd(O2CCF3)2 (5) Bipy (7) - 80 63
4 Pd(O2CCF3)2 (5) Bipy (7) KSbF6 (20 mol%) 80 full
To test the scope and limitations of the reaction on acyclic enones, we assayed several classes of linear substrates for the reaction. The results are summarized in Table 3. We started our study with 28, a substrate found to be suitable for the conjugate addition of boronates by rhodium catalysis. However, the product was formed only in trace quantities. 29, bearing a bulkier tert-butyl group was found to
135
Facile Construction of Benzylic Quaternary Centers via Pd catalysis
Table 3: Substrate scope for conjugate addition of acyclic enones.a
Facile Construction of Benzylic Quaternary Centers via Pd catalysis
be completely inactive towards conjugate addition. Substrate 30, a derivative of Meldrum’s acid could be recovered unreacted as well. 26 Was found to be a successful substrate for conjugate addition. A variety of boronic acids were found to be amenable to the reaction. Phenylboronic acid afforded 27 in 72% yield, while m-tolylboronic acid gave 27b in 75% yield. p-Fluorophenylboronic acid gave 27c in 70% yield, and 27d could be obtained from p-methoxyphenylboronic acid in 72% yield. In general, boronic acids bearing alkoxy functionalities performed well in the reaction. Compounds 27e, 27f, 27g and 27h were all obtained in yields between 72-78% (entries 8-11).
Substrate E-31, bearing a benzyl moiety was also found to be applicable in conjugate addition under the optimized reaction conditions. Phenylboronic acid and m-tolylboronic acid afforded 32 and 32b in 74% and 90% yield respectively. Substrate Z-31 also afforded 32, albeit in a slightly reduced yield of 68%. In order to ascertain whether the observed success of substrates 26 and 31 was due to the coordination of the allylic oxygen atom to the metal center during the catalysis, we designed substrate 33. Substrate 34 was designed additionally to test the amenability of a nitrogen-containing substituent in the reaction. However, both these substrates were found to be unreactive under the reaction conditions. Thus it remains unclear whether the observed influence of allylic oxygen atom for the reaction is due to a coordination of the substrate to the metal center or due to an electronic influence.
It is noted that in most of the successful cases, the isolated yields were in the range of 70-80%, though, in all cases complete consumption of the substrate in the reaction had been indicated by GC. Further, TLC of the worked up reaction mixture, showed only a single spot. The cause of the lower yield was not apparent at the time of the investigation. It was found later (See section 6.3.5 for details), that a side reaction leading to the decomposition of the starting material was occurring. The decomposition products could be identified, when using a substrate bearing a larger substituent on the -position of the carbonyl, such as a n-butyl (35) or phenyl (38). Ketoaldehydes, 37 and 40 were obtained from 35 and 38 respectively (Scheme 7), and their formation accounts for the gap between the conversion of the starting material and isolated yield of the product.
The boronic acids described in Figure 3 were also found to be inapplicable for the reaction with linear substrates.
138
Chapter 5
Scheme 7: Side products due to decomposition of starting material.
5.4 Summary and conclusions
This study has resulted in the development of a simple catalyst system for the Pd-catalyzed conjugate addition of arylboronic acids to β,β-disubstituted enones resulting in the formation of quaternary centers. This procedure avoids the synthesis of preformed catalysts, arylboroxines or forcing conditions necessary for the extrusion of SO2. Only 1 mol% of Pd(O2CCF3)2 along with 1.5 mol% of inexpensive 2,2’-bipyridine is necessary to achieve full conversion and good yields in the case of cyclic enones. 5, 6 and 7-membered rings were found to be amenable to the reaction, though 5-membered rings usually exhibited higher yields and shorter reaction times. Linear substrates bearing an allylic oxygen function were found to be good substrates for the reaction. A higher catalyst loading (5 mol%) of Pd(O2CCF3)2, along with 20 mol% of KSbF6 as an additive was needed in that case to obtain full conversion. Isolated yields for the products were mostly in the range of 70-80%, and the “missing yield” could be attributed to the formation of
139
Facile Construction of Benzylic Quaternary Centers via Pd catalysis
a ketoaldehyde side product. The study also established limitations on the scope of arylboronic acids that currently can be employed in the reaction.
Future research in this direction could focus on expanding the scope of linear substrates for conjugate addition, and suppression of side product formation.
140
Chapter 5
5.5 Experimental
5.5.1 General
All experiments were carried out in flame dried or oven dried (150 oC) glassware, in
an atmosphere of dinitrogen, unless specified otherwise, by standard Schlenk
techniques. Schlenk tubes with screw caps, and equipped with a teflon-coated
magnetic stir bar were flame dried under vacuum and allowed to return to rt prior to
being charged with reactants. A manifold permitting switching between dinitrogen
atmosphere and vacuum was used to control the atmosphere in the reaction
vessel. Reaction temperature refers to the temperature of the oil bath.
Flash chromatography was performed using Merck silica gel type 9385 (230-400
mesh), using the indicated solvents. All solvents used for filtration and
chromatography were of commercial grade, and used without further purification.
Anhydrous methanol, and acetonitrile were sourced from Sigma-Aldrich or Acros
and stored under dinitrogen.
TLC was performed on Merck silica gel 60, 0.25 mm plates and visualization was
done by UV and staining with Seebach’s reagent (a mixture of phosphomolybdic
acid (25 g), cerium (IV) sulfate (7.5 g), H2O (500 mL) and H2SO4 (25 mL)) or
Vanillin stain (a mixture of vanillin (6g), conc. sulphuric acid (1.5 mL) and ethanol
(95 mL)) or KMnO4 stain.
1H- and 13C-NMR were recorded on a Varian AMX400 (400, 100.59 MHz,
respectively) using CDCl3 as solvent, unless specified otherwise. Chemical shift
values are reported in ppm with the solvent resonance as the internal standard
(CHCl3: 7.27 for 1H, 77.1 for 13C). Data are reported as follows: chemical shifts
( ), multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, br = broad, m =
multiplet), coupling constants J (Hz), and integration.
GC-MS measurements were made using a HP 6890 Series Gas Chromatograph
system equipped with a HP 5973 Mass Sensitive Detector. GC measurements
were made using a Shimadzu GC 2014 gas chromatograph system bearing a AT5
column (Grace Alltech) and FID detection. Whenever GC conversion is reported,
the quantification was done using cyclo-octane as internal standard. High
141
Facile Construction of Benzylic Quaternary Centers via Pd catalysis
Resolution Mass Spectrometry was performed using a ThermoScientific LTQ
Oribitrap XL spectrometer.
Synthesis of starting materials: The synthesis of 26 and 31 has been described
in Chapter 3 (Section 3.5). The substrates 12,40 28,41 and 3042 were synthesized
according to literature procedures.
5.5.2 General procedure for the conjugate addition of arylboronic acids to
ββββ,ββββ-disubstituted enones
Method A: Conjugate addition to cyclic enones
To a Schlenk tube equipped with a magnetic stirring bar and a septum was added
(42) Fillion, E.; Wilsily, A. J. Am. Chem. Soc. 2006, 128, 2774.
(43) Kikushima, K.; Holder, J. C.; Gatti, M.; Stoltz, B. M. J. Am. Chem. Soc. 2011, 133, 6902.
154
Chapter 5
Chapter 6
Palladium-catalyzed enantioselective conjugate addition of arylboronic acids to acyclic enones
In this chapter, our study towards higher enantioselectivities in the palladium catalyzed conjugate addition of arylboronic acids to acyclic enones is discussed. The highest ee obtained was 77%. Yields of the isolated products were diminished, despite complete consumption of the starting material, due to a competing decomposition reaction.
Parts of this chapter will be submitted for publication: Gottumukkala, A. L., Matcha, K., de Vries, J. G., Minnaard, A. J., Manuscript in preparation.
156
Chapter 6
6.1 Introduction
The transition metal catalyzed conjugate addition of organometallics to enones is an important tool for enantioselective carbon – carbon bond formation. The importance of this reaction in synthesis is clear from the extensive literature dedicated to this transformation.1-4
Acyclic enones are considerably more challenging substrates for conjugate addition reactions than cyclic enones, partly due to the existence of s-cis and s-trans conformers (Scheme 1).1,5
Scheme 1: s-Cis and s-trans conformations of enones.
As a result, very often, a different chiral ligand is necessary when compared to cyclic enones. For example, in the conjugate addition of organozinc reagents to cyclic and acyclic enones using a phosphoramidite ligand derived from TADDOL under similar reaction conditions,5 a clear difference in enantioselectivity was observed (Scheme 2).
Scheme 2: Difference between cyclic and acyclic substrates in the Cu-catalyzed addition of diethylzinc.
Similarly, in an example of rhodium catalyzed conjugate addition of arylboroxines, acyclic substrates gave lower yields (82%) and enantioselectivities (86% ee) compared to cyclic enones (99% yield, and 99% ee), under identical conditions (Scheme 3).6
157
Conjugate Additions to Acyclic Enones
Scheme 3: Variation between cyclic and acyclic substrates in the Rh catalyzed addition of
phenylboroxine.
Following the same trend, the conjugate addition of phenylboronic acid to 12 using PdCl2-PhBOX gave only 23% ee (Scheme 4) as described in Chapter 4. The use of the bulkier TBDPS protecting group permitted an increase in ee (60%) but this was accompanied by the decomposition of the starting material and the formed product during the reaction.
Scheme 4: Conjugate addition of phenylboronic acid to 12 catalyzed by PdCl2-PhBOX.
6.2 Goal
While the reaction above clearly demonstrates the applicability of Pd-catalyzed conjugate addition reactions to form quaternary centers in acyclic substrates, the selectivity obtained was much lower than in case of cyclic enones. Thus a detailed and separate study was necessary, to develop a set of reaction conditions necessary to obtain improved selectivity. We were interested in the development of a Pd-catalyzed conjugate addition of arylboronic acids to acyclic , -disubstituted enones, leading to the formation of benzylic quaternary centers with good enantioselectivity. Unlike cyclic enones, acyclic enones possess lesser structural rigidity.
6.3. Results and discussion
Based on our previous studies detailed in chapters 4 and 5, we surmised the following aspects.
158
Chapter 6
1) Substrates bearing an allylic oxygen were necessary for the success of the reaction
2) New classes of chiral nitrogen ligands had to be assayed for the reaction.
Due to the limited commercial availability of chiral nitrogen ligands, it was decided to spend considerable effort in synthesizing several chiral nitrogen ligands for the reaction.
6.3.1 Synthesis of chiral nitrogen ligands
6.3.1.1 C2–Symmetric chiral bipyridine ligands
Inspired by the success of bipyridine in conjugate addition reactions (Chapter 5) and the success of C2 symmetric ligands in affording excellent enantioselectivities for cyclic substrates (Chapter 4), it was a natural choice to investigate the influence of chiral bipyridines (Figure 1). Chiral bipyridine ligands have been investigated in transition metal catalysis, and were found to be successful for a variety of reactions. Copper catalyzed allylic oxidation using ligand 15 provided the product in 82% ee (Scheme 5).7
Figure 1: Chiral bipyridine ligands studied in transition metal catalysis.
Scheme 5: Chiral bipyridine ligand 15 applied in Cu catalyzed allylic oxidation.7
6.3.1.2 Synthesis of iso-PINDY ligand
The synthesis of ligand 14 has been described by Kocovský7 starting from (+)-α-pinene (Scheme 6). The first step of the reported synthesis is a singlet oxygen ene reaction, forming pinocarvone (–)-24. In our hands, this reaction remained difficult
159
Conjugate Additions to Acyclic Enones
to reproduce. This was perhaps due to the large size of the singlet oxygen apparatus (1 l) that was at our disposal, compared to the scale of the reaction we were working on (2 g). The stream of oxygen gas needed for the reaction, also led to the accelerated evaporation of the reaction mixture. Fortunately, we could overcome this shortcoming by an alternative route,8 involving the epoxidation of 21, followed by ring-opening with LiNEt2 to form 23 and subsequent oxidation with PCC.
In our hands, the subsequent Kröhnke annulation9,10 with pyridinium salt 25 proved to be particularly low yielding (8%). Using a recrystallized portion of 25 did not help. The subsequent triflation of 26 proceeded in 92% yield, the product of which was subjected to a nickel mediated Negishi-type coupling. The reaction remained incomplete even after 24 h (the reported yield after reaction overnight was 51%), and the majority of the product obtained was the hydrodetriflated product 28. The expected homocoupled product 29 was obtained only trace quantities. Due to the difficulties experienced with the synthesis of 29 and the poor yields, it was decided to discontinue further efforts in this direction.
Scheme 6: Attempted synthesis of iso-PINDY ligand.
6.3.1.3 Synthesis of MINDY ligand
Next, we attempted the synthesis of MINDY (Scheme 7), a ligand derived from (–)-menthone (30). Claisen condensation of menthone with HCO2Et afforded 31 in 74% yield, and subsequent Knoevenagel condensation with -cyanoacetamide,
160
Chapter 6
followed by spontaneous cyclization led to the formation of pyridine 32. At this stage, however, a 1:1 mixture of epimers was obtained, which was carried forward in the synthesis. Acidic hydrolysis gave the hydroxy acid, which was subjected to pyrolytic decarboxylation, affording 34 in 70% yield. Next, chlorination of 34 with POCl3 gave 35 in 77% yield, which was subjected to a nickel mediated dimerization. This led as expected to a mixture of diastereomers, which could be separated by column chromatography. The desired ligand 15 was isolated in only 11% yield.
Scheme 7: Synthesis of MINDY ligand.
15 Was tested in a conjugate addition reaction of phenylboronic acid to 12, using the reaction conditions described in Chapter 5, for acyclic substrates (Scheme 8a). Disappointingly, the product was formed in only trace amounts, and found to be racemic. After several failed reactions, it was decided to verify the complexation of 15 with Pd. For this purpose, we attempted the complexation of 15 with PdCl2
(Scheme 8b) as we had observed that PdCl2 gave the corresponding Pd-complexes in good yields (Chapter 4). The reaction only resulted in a black residue which did not correspond to the product. Unfortunately, none of the added ligand could be recovered, or identified by 1H-NMR, suggesting disintegration. Further, we tested if the residue obtained from the reaction was catalytically active, by adding
161
Conjugate Additions to Acyclic Enones
starting materials (12 and phenylboronic acid) to the reaction. No reaction was observed.
Scheme 8: Attempted reactions with ligand 15.
The failure of complexation can perhaps be explained by the fact that the ligand is too sterically hindered for Pd. Spacefill models of 15, generated using Chem3D Pro®, indeed indicate that the pyridine nitrogen is quite shielded by the bulky isopropyl group. In addition, in its energetically most stable conformation the molecule probably adopts a conformation wherein the two nitrogens are “trans” to one another (Figure 2), thereby making complexation difficult.
Figure 2: Conformations of ligand 15.
Taking these failures with C2 symmetric ligands into account, it was decided to study C1-symmetric ligands. Our decision in this direction was further leveraged by the report of Stoltz,11 wherein excellent enantioselectivities for cyclic substrates were obtained using a C1 symmetric ligand.
6.3.2 Synthesis of C1–symmetric ligands
Learning from the good selectivities obtained using oxazoline based ligands12 (Chapter 4), we decided to continue with this class of ligands. Candidates that were not available commercially had to be synthesized, however.
162
Chapter 6
6.3.2.1 C1–symmetric chiral pyridyl oxazolines
Learning from the work of Stoltz,11 we went on to prepare a set of pyridyl-oxazolines (L4-L6), using various amino acids. Their two step synthesis is described in Scheme 9. These ligands were then used in catalysis.
Scheme 9: Synthesis of pyridyl-oxazoline ligands.
6.2.2.2 C1–symmetric chiral quinoline-oxazolines.
PdII-complexes of chiral quinoline-oxazolines have been found to be excellent catalysts for enantioselective oxidative cascade cyclizations13 and aza-Wacker-type cyclizations of olefinic tosylamides.14 The synthesis of these ligands proceeds similar to ligands L4-L6 (Scheme 10), when starting from quinoline-2-nitrile (40)
Scheme 10: Synthesis of quinolone-oxazoline ligands.
6.3.2.3 C1–symmetric chiral quinolinyl-oxazolines
Furthermore, we proceeded to synthesize a set of C1 symmetric quinolinyl-oxazolines. This class of ligands has been found to be effective in the Pd catalyzed asymmetric hydroarylation of norbornenes.15 The synthesis proceeded starting from 8-quinolinecarboxylic acid 42 as presented in scheme 11.
Scheme 11: Synthesis of quinolinyl-oxazoline ligands.
163
Conjugate Additions to Acyclic Enones
6.3.3 Optimization of reaction parameters.
With the synthesized ligands and complexes in hand (Figure 3), the Pd-catalyzed conjugate addition reaction of phenylboronic acid to 12 was tested. The results are summarized in Table 1. As noted previously (Chapter 5, section 5.3.3), acyclic enones, in general, required longer reaction times and higher reaction temperatures (80 oC) to reach full conversion. Hereinafter, 3 equiv of the phenylboronic acid was necessary to afford full conversion, while only 1.5 equiv was necessary in the case of cyclic substrates (Chapter 4). This could be due to increased protodeboronation of the arylboronic acids at elevated temperatures.16
Picking up from the success of bisoxazoline ligands in the conjugate addition reaction to cyclic enones, we started our screening program with the optimized conditions (Table 1, entry 1) reported in Chapter 4. The starting material was fully consumed (as measured by GC) during the reaction, although the product was formed in only 24% ee. Retaining the ligand and using Pd(O2CCF3)2 as the Pd precursor, resulted in a diminished conversion, while maintaining the same ee (entry 2). Bulkier tert-butyl substituents on the ligand led to poor performance (entry 3) in the reaction. A similar fate was shared by ligand L3 (entry 4) and its corresponding complex C2 (entry 5). In the latter case, upon dehalogenation it is likely that a C-H activation of the proximal iso-propyl group by the dicationic palladium results in the deactivation of the catalyst.17 Ligand L4 was tested under the conditions described by Stoltz,11 which led to an improved 34% ee, though conversion was incomplete (entry 6). Learning from the work of Stoltz, we opted for DCE as a solvent, in view of the higher reaction temperature. The addition of 20 vol% water along with KSbF6 resulted in full conversion (entry 7). The beneficial influence of the water could be due to improved transmetalation.18 The role of SbF6
– in improving conversions has been discussed before (Chapter 4, section 4.3.1). The use of other ligands from the same class (L5, L6) only resulted in poorer results (entries 8, 9). Complex C3, synthesized from ligand L4, however demonstrated an improved ee (entry 10) and full conversion. Use of quinoline-oxazolines L7 and L8 led to full conversion, although lower ee’s were obtained (entries 11,12). Use of L7 derived C4 also resulted in full conversion, although the ee obtained was only 42%. Application of tert-butyl bearing quinolinyl-oxazoline L9 induced 33% ee, while ligand L11 afforded an impressive 69% ee. Complex C5 derived from L11 also afforded full conversion with 71% ee.
P,N-ligands L12 and L13 induced very high ee’s (entries 18, 19), but the conversions remained low. Finally, we tried the Spirobox class of ligands L14 and L15, which only resulted in trace amounts of product being formed.
164
Chapter 6
Figure 3: Ligands and complexes used for the optimization studies
165
Conjugate Additions to Acyclic Enones
Table 1: Optimization of reaction parameters for the conjugate addition of phenylboronic acid.a
Entry Pd Ligand Solvent Additive Convb eec
1 C1 - MeOH:H2O AgSbF6 full 24
2 Pd(O2CCF3)2 L1 MeOH:H2O - 65 25
3 Pd(O2CCF3)2 L2 MeOH:H2O - < 2 nd
4 Pd(O2CCF3)2 L3 MeOH:H2O - nd
5 C2 - MeOH:H2O AgSbF6 - nd
6 Pd(O2CCF3)2 L4 DCE - 87 34
7 Pd(O2CCF3)2 L4 DCE:H2O KSbF6 full 35
8 Pd(O2CCF3)2 L5 DCE:H2O KSbF6 full 11
9 Pd(O2CCF3)2 L6 DCE:H2O KSbF6 full 6
10 C3 - MeOH:H2O AgSbF6 full 54
11 Pd(O2CCF3)2 L7 DCE:H2O KSbF6 full 33
12 Pd(O2CCF3)2 L8 DCE:H2O KSbF6 full 38
13 C4 - MeOH:H2O AgSbF6 full 42
14 Pd(O2CCF3)2 L9 DCE:H2O KSbF6 78 33
15 Pd(O2CCF3)2 L10 DCE:H2O KSbF6 full 56
16 Pd(O2CCF3)2 L11 DCE:H2O KSbF6 full 69
17 C5 - MeOH:H2O AgSbF6 full 71
18 Pd(O2CCF3)2 L12 DCE:H2O KSbF6 31 89
19 Pd(O2CCF3)2 L13 DCE:H2O KSbF6 26 91
20 Pd(O2CCF3)2 L14 DCE:H2O KSbF6 <10 nd
21 Pd(O2CCF3)2 L15 DCE:H2O KSbF6 <10 nd
a 12 (0.2 mmol, 1 equiv), phenylboronic acid (0.6 mmol, 3 equiv) Pd precursor or Pd complex (5 mol%), ligand (7 mol%), additive (20 mol%), solvent : H2O (4:1) 0.5 ml, 18 h, 80 oC b Conversion determined by
166
Chapter 6
GC analysis of reaction mixture. c ee determined by chiral HPLC analysis of the isolated product [chiralpak OJ-H, n-heptane : i-PrOH (99:1), 210 nm]. DCE = 1,2-dichloroethane, nd = Not determined.
From the above set of experiments, we learned that ligand L11 afforded the best ee’s, along with full conversion. Realizing that the difference in ee when using Pd(O2CCF3)2 / L11 and complex C5 was marginal, we decided to use the former for further experimentation, in view of convenience.
6.3.4 Substrate scope of the reaction
With the optimized conditions in hand, we proceeded to study the scope of the reaction. The results are summarized in Table 2. The conjugate addition of phenylboronic acid to 12, gave 13 in 71% yield and 69% ee. With p-tolylboronic acid, 13b was obtained in 74% yield, with a minor drop in ee to 65%. Use of m-tolylboronic acid gave 13c gave the product in 70% yield and 64% ee. Reaction with alkoxy-substituted boronic acids, gave the corresponding products 13d and 13e in a reduced yield of 42% and 53%, respectively. The ee in these cases was 59% and 57%, respectively. Interestingly, when a chloro-substituted alkoxyboronic acid was used, the yield and ee improved, affording 13f in 83% yield and 77% ee. Substrate 46, bearing a butyl chain, was arylated to give the corresponding product 47 in 71% ee. Phenyl containing substrate 48, afforded 49 with similar selectivity (69%). Comparing entries 1, 7, 8 and 9, it is concluded that the substituent on the α- position of the carbonyl, does not have an significant influence on the ee of the formed product. Substrate 52 bearing a benzyl substituent also afforded the expected conjugate addition product (53) in 63% yield and 71% ee.
167
Conjugate Additions to Acyclic Enones
Table 2: Substrate scope with a variety of enones.a
Entry Substrate Boronic Acid Product (yieldb) eec
(%)
1
13 (71) 69
2
13b (74) 65
3
13c (70) 64
4
13d (42) 59
5
13e (53) 57
6
12
13f (83) 77
7
46
47 (58) 71
168
Chapter 6
a enone (0.5 mmol, 1.equiv), phenylboronic acid (1.5 mmol, 3 equiv) Pd(O2CCF3)2 (8.3 mg, 5 mol%), L11 (9.6 mg, 7 mol%), KSbF6 (27 mg, 20 mol%), DCE : H2O (4:1) 1 ml, 18h. 80 oC. b Isolated yield. c ee determined by chiral HPLC analysis of the isolated product.
6.3.5 Substrate decomposition
Examining Table 2, it is apparent that the isolated yields of the products are essentially in the range of 65-75%, despite the starting material being consumed completely in the reaction. This remained a puzzle until substrate 46 and 48 were tested in the conjugate addition reaction. In addition to the expected conjugate addition product, we also observed the formation of a keto-aldehyde (54, 55) as side product (Scheme 12). 54 and 55 could be isolated and characterized.
Scheme 12: Side products obtained from the reaction.
Entry Substrate Boronic Acid Product (yield)
ee
(%)C
8
48
49 (48) 69
9
50
51 (32) 71
10 52
53 (63) 71
169
Conjugate Additions to Acyclic Enones
By analogy, one might expect that such a side product (56) might even be formed in the case of substrate 12. However, 56 is of low molecular weight, and expected to be quite volatile. Consequently, it escapes detection by GC/MS or crude 1H-NMR of the reaction mixture. The formation of 56 might account for the “missing yield” of 13 (Scheme 13).
Scheme 13: Side product could not be detected.
This hypothesis is further confirmed by the formation of ketoaldehyde 57, formed in the reaction of 50, which was observed by GC-MS, but could not be isolated (Scheme 14).
Scheme 14: Side product detected by GC/MS.
Two plausible mechanisms are proposed for the formation of the ketoaldehyde side product (Scheme 15, 16). Both mechanisms invoke a Pd allyl species (B and J) that is formed by an oxidative addition to Pd0 species. It must be emphasized that at this time, both mechanisms are speculative, and due to time constraints no experimental data has been collected to prove or disprove these mechanisms.
Considering the mechanism presented in Scheme 15, a sequence involving oxidative addition of Pd0, -allyl formation (B), and reductive elimination liberating Pd0 is proposed to explain the isomerization of the double bond, leading to the formation of D. Since phenylboronic acid is a weak acid (pKa of 8.9 in water, comparable to phenol),16 one might expect that it protonates the enol ether, forming species E and F. These can expel isobutene and form the observed product 54. Alternatively, F can undergo hydrolysis via hemi-acetal G, to give 54 and tert-butanol. A weak point of this hypothesis is that D has not been observed in the reaction.
170
Chapter 6
Scheme 15: Plausible mechanism for the formation of 54.
An alternative pathway involves a “Tsuji-Trost type” mechanism, in which an allyl species is formed by expelling (protonated) tert-butanol (J), followed by nucleophillic attack of water to form K. This then isomerizes to L, as shown in Scheme 16. Tautomerization of L gives 54. Although K was not observed as such in the reactions, we did observe such deprotection (especially incases of substrates bearing silyl protecting groups) and hemiacetal formation, in an earlier phase of the research (Chapter 4, section 4.3.2).
Scheme 16: Plausible formation of side-product 54 via a Tsuji-Trost type pathway.
Control experiments performed in the absence of Pd did not afford the ketoaldehydes. Since the catalyst employed for the reaction is chiral, it is possible that the formed ketoaldehyde is enantiomerically enriched. However, 54 and 55 did not show any optical activity in a polarimeter. This could also be explained by the rapid racemization of the stereocenter next to the aldehyde via keto-enol
171
Conjugate Additions to Acyclic Enones
tautomerism, under the acidic conditions of the reaction or alternatively, that they are indeed enantioenriched but their specific rotation is extremely low. Further studies, could focus on the synthesis of the racemate for comparison.
6.4 Summary and conclusions
In this chapter, we describe the first examples of Pd catalyzed enantioselective conjugate addition of arylboronic acids to β,β-disubstituted linear enones. Enantioselectivities up to 77% could be achieved for a series of substrates bearing an allylic oxygen. Substituting the methyl group of substrate 12 with an n-butyl or phenyl did not seem to have an influence on the enantioselectivity of the formed product.
Though the ee’s obtained are considerably lower than those obtained with cyclic enones, these represent the first examples. The isolated yield of the products was mostly in the range of 60-70% despite the substrates being completely consumed in the reaction. The “missing yield” could be accounted for by competing Pd-catalyzed decomposition of the substrate.
6.5 Future perspectives
Future studies should include a mechanistic elucidation of the formation of the side product, leading to changes in the catalyst system to arrest its formation. In addition, design and study of new ligands is desirable, to increase the enantioselectivity.
While examining the most successful ligands for conjugate addition to acyclic substrates from the current chapter (L11) and cyclic substrates described in chapter 4 (L1), it might be interesting to note that both these ligands form a 6- membered palladacycle (Figure 3). On the other hand, ligand L4, reported by Stoltz,11 forms a 5-membered cycle, and led to lower enantioselectivities compared to L1 for cyclic substrates and L11 for acyclic substrates The 6-membered cycle might allow the stereodirecting element in oxazole to be closer to the metal center, thus affording higher enantioselectivities. Thus, future development could focus on the design of more ligands bearing 6-membered palladacycles, and with increased steric bulk of the stereodirecting element of the ligand. Plausible structures are presented in Figure 4.
172
Chapter 6
Figure 3: Palladacycles formed from the ligands L1, L4 and L11.
Figure 4: Plausible ligand designs for improved enantioselectivity.
173
Conjugate Additions to Acyclic Enones
6.6 Experimental
6.6.1 General
All experiments were carried out in flame dried or oven dried (150 oC) glassware, in
an atmosphere of dinitrogen, unless specified otherwise, by standard Schlenk
techniques. Schlenk reaction tubes with screw caps, and equipped with a teflon-
coated magnetic stir bar were flame dried under vacuum and allowed to return to
room temperature prior to being charged with reactants. A manifold permitting
switching between dinitrogen atmosphere and vacuum was used to control the
atmosphere in the reaction vessel. Reaction temperature refers to the temperature
of the oil bath.
Flash chromatography was performed using Merck silica gel type 9385 (230-400
mesh), using the indicated solvents. All solvents used for filtration and
chromatography were of commercial grade, and used without further purification.
Anhydrous methanol, and acetonitrile were sourced from Sigma-Aldrich or Acros
and stored under dinitrogen.
TLC was performed on Merck silica gel 60, 0.25 mm plates and visualization was
done by UV and staining with Seebach’s reagent (a mixture of phosphomolybdic
acid (25 g), cerium (IV) sulfate (7.5 g), H2O (500 ml) and H2SO4 (25 ml)) or Vanillin
Stain (a mixture of vanillin (6g), conc. sulphuric acid (1.5 ml) and ethanol (95 ml))
or KMnO4 stain.
1H- and 13C-NMR were recorded on a Varian AMX400 (400, 100.59 MHz,
respectively) using CDCl3 as solvent, unless specified otherwise. Chemical shift
values are reported in ppm with the solvent resonance as the internal standard
(CHCl3: 7.27 for 1H, 77.1 for 13C). Data are reported as follows: chemical shifts
( ), multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, br = broad, m =
multiplet), coupling constants J (Hz), and integration.
GC-MS measurements were made using a HP 6890 Series Gas Chromatograph
system equipped with a HP 5973 Mass Sensitive Detector. GC measurements
were made using a Shimadzu GC 2014 gas chromatograph system bearing a AT5
column (Grace Alltech) and FID detection. Whenever GC conversion is reported,
174
Chapter 6
the quantification was done using cyclo-octane as internal standard. High
Resolution Mass Spectrometry was performed using a ThermoScientific LTQ
Oribitrap XL spectrometer.
The absolute configuration of the products described herein is not known.
Synthesis of starting materials: Synthesis of substrates 12 and 52 has been
Studies toward an enantioselective Matsuda-Heck reaction
This chapter details our studies towards an enantioselective version of the Matsuda-Heck reaction. While our efforts to this end did not bear fruit, we studied potential substrates for a diastereoselective version of the reaction. Our results led us to speculate that the reaction also involves a radical pathway leading to unselective product formation.
This work was performed in collaboration with Edwin Kroon, MSc, as a part of his master research project.
184
Chapter 7
7.1 Introduction
Within a few years following the initial reports of Heck1-6 and Mizoroki7,8 on the use of aryl iodides for palladium-catalyzed coupling with alkenes, the group of Matsuda9 reported use of aryldiazonium salts in place of aryl halides for this highly effective coupling reaction. In a series of publications,10-24 the group laid the foundation for what would become the Matsuda-Heck reaction,25,26 exploring its scope, reaction parameters and application. Important contributions to this field of study, in recent years, have been made by the groups of Correia27-43 and Felpin,44-
50 among others.
Disappointingly, however, the reaction escaped the interest of most of the synthetic chemists, most probably due to the supposed instability of diazonium salts. Nonetheless, the reaction regained popularity in the nineties, with the emergence of stable and dry diazonium salts in organic synthesis and catalysis.51 These developments of the Matsuda-Heck reaction have been the focus of several excellent reviews.25,26,51
7.1.1 Aryldiazonium salts
Aryldiazonium salts have been known for over 150 years, since their discovery by Griefs.52 However, their application in organic chemistry has been limited, owing to their reputation of being unstable and explosive. This reputation arises partly from the fact that the first reports of aryldiazonium salts were concerned with their chloride salts, which are unstable above 0 oC, and even explosive when subjected to shock or stress.26 In fact, they can rarely be prepared in a dry from.
However, the stability of these species varies significantly with the nature of the counterion. Aryldiazonium tetrafluoroborates53 are generally considered stable and are most commonly employed, while the corresponding disulfonimides,54 tosylates55 and carboxylates56 have also been reported to be stable. Despite their stability, only few aryldiazonium salts are available commerically and therefore most of them are required to be synthesized in-house.
An alternative and practical approach to address this issue of instability is to form them in situ. Aryldiazonium salts with BF4
–, ClO4–, CF3CO2
–, F– and CH3SO3– as
counterions have been prepared in situ and found to be very effective for Matsuda-Heck reactions.26 However, the yields obtained in those cases are often inferior to reactions in which crystalline aryldiazonium salts have been employed.25 For this reason, we chose to perform our studies with crystalline aryldiazonium tetrafluoroborates.
185
Matsuda-Heck Reaction
7.1.2 Diastereoselective Matsuda-Heck reactions
Following the revival of interest in the Matsuda-Heck reaction, and parallel developments in stereoselective Mizoroki-Heck reactions, reports of diastereoselective Matsuda-Heck reactions have been published.
A particularly successful example of a diastereoselective Matsuda-Heck reaction is the one using 1-tert-butylsulfinylcyclopentene (1) as the substrate.57 High diastereoselectivities (dr’s up to 96:4) were obtained, though the products were obtained in moderate yields (72-79%) and a high catalytic loading (20 mol %) was necessary. The authors demonstrated that the products could be reductively desulfonylated to the corresponding alkene (Scheme 1). Interestingly, the Mizoroki-Heck reaction on the same substrate with 4-iodoanisole did not work, and a complex mixture of products was obtained, most probably due to thermal instability of the alkyl sulfoxide. In spite of the attractiveness of this approach, the high catalytic loading combined with the moderate yields limit the application of this strategy.
Scheme 1: Matsuda-Heck on 1-tert-butylsulfinylcyclopentene.
The arylation of substituted dihydropyrroles was also found to proceed in a highly stereoselective manner (Scheme 2).40 The best selectivities in favor of the trans product were obtained when both the pyrrole nitrogen and the substituent on the 2’ position bore sterically bulky entities. The arylation of dihydrofurans also followed a similar trend.
Scheme 2: Dihydropyrroles as substrates for the Matsuda-Heck reaction.
The synthesis of C-aryl glycosides via Mizoroki-Heck reaction often results in a mixture of isomers due to undesired double-bond isomerizations at higher temperature.58 However, employing the mild conditions of the Matsuda-Heck
186
Chapter 7
arylation on glycal 7 gave the desired product in good yield and as a single diastereomer (Scheme 3).59
Scheme 3: Synthesis of aryl-glycosides via Matsuda Heck reaction.
The high facial selectivity obtained towards the trans isomer (“trans selectivity”) in the Matsuda-Heck reaction was exploited in an efficient total synthesis of (–)- isoaltholactone (11).34 The good yield obtained in this key step allowed the completion of the synthesis of the natural product in only 7 steps, with an overall yield of 25% (Scheme 4).
Scheme 4: Diastereoselective Matsuda-Heck reaction on a 2’-substituted dihydrofuran.
Recently, the group of Correia studied the role of directing groups embedded in the substrate for Matsuda-Heck reactions60 on cyclopentene 12 (Scheme 5). The selectivity obtained was rationalized as a result of chelation of the aryl-Pd species by the Boc group of the substrate, directing the arylation (15). This strategy was exploited in a concise total synthesis of the sphingosine 1-phosphate receptor (S1P1) agonist VPC01091.
The goal of this study was to develop and demonstrate the first examples of an enantioselective Matsuda-Heck reaction. Despite being known for more than 30 years, at the time of our study, there were no examples of an enantioselective Matsuda-Heck reaction in literature. Nonetheless, there were several examples of enantioselective reactions known for the closely related Mizoroki-Heck reaction.
This may, in part, might be due to the fact that the most common route to render palladium catalyzed reactions enantioselective is to replace achiral phosphines with an appropriate chiral phosphine. However, with the Matsuda-Heck reaction, this poses a difficulty as the diazonium compounds readily react with the phosphines,61 as depicted in Scheme 6, making them poor ligands.
Scheme 6: Reaction of phosphines with aryldiazonium salts.
Our strategy to address this issue was to assay chiral nitrogen based ligands (such as bisoxazolines) or -accepting ligands such as phosphoramidites, and chiral N-heterocyclic carbene (NHC) ligands. Further, we chose substrates that were highly successful for the enantioselective Mizoroki-Heck reaction.
7.3 Results and Discussion
7.3.1 Synthesis of aryldiazonium salts
The limited commercial availability of aryldiazonium salts led us to search for a convenient method to obtain them in multigram quantities in a pure and dry form. Realizing the importance of the appropriate counterion in synthesis and stability of the salt,26 we chose to work with aryldiazonium salts with tetrafluoroborate counterions. The synthesis was adapted from a report of Doyle et al.62 (Scheme 7).
Scheme 7: Synthesis of aryldiazonium tetrafluoroborates in a dry form.
While the method is highly effective for the synthesis of the aryldiazonium salts, we consistently observed colored impurities along with the formed products, which are otherwise usually described as off-white solids. The side products were also
188
Chapter 7
observed by 1H-NMR, and estimated to be present in 5 - 10%. The color persisted despite the use of freshly distilled anilines The colored impurities could be a result of side reactions such as the azo-coupling, which are known to produce brightly colored dyes.63
In order to decolorize and purify these diazonium salts, various methods were attempted. Trituration with pentane, ether or dichloromethane did not decolorize the salt, and the salts did not survive column chromatography. After careful observation, it was found that decolorized diazonium salts could be obtained by dissolving the colored diazonium salts in acetone and precipitating them by dropwise addition of ice-cold diethyl ether. The precipitates were filtered, washed with additional diethyl ether and dried under vacuum. Once dry, the salts could be stored up to 2 months at –18 oC (See experimental section of details). Four diazonium salts were prepared and purified with this procedure (Figure 1). The purity of these salts was assayed by 1H-NMR to be > 95%.
Figure 1: Diazonium tetrafluoroborates synthesized for this study.
7.3.2 Reaction parameters
To ascertain the quality of the synthesized diazonium salts, we performed a benchmark Matsuda-Heck reaction on butyl acrylate (Scheme 8), under standard reaction conditions.
Scheme 8: Matsuda-Heck reaction with butyl acrylate.
Once we were convinced that the reaction could be performed as reported in the literature, we proceeded to test pro-chiral substrates.
189
Matsuda-Heck Reaction
7.3.3 2,3-Dihydrofuran
We investigated different substrates that could be employed for the enantioselective Matsuda-Heck reaction. Our initial study involved 2,3-dihydrofuran, a substrate employed extensively in the enantioselective Mizoroki-Heck reaction.64,65 The results are summarized in Table 1.
In the majority of the cases (except entry 3), the de-diazotized product was observed. The formation of this formal reduction of the aryldiazonium salt was found to be independent of the electronic nature of the substituents on the aryldiazonium salt and the nature of the solvent, suggesting that it proceeded via a radical pathway.66,67 However, when Pd2(dba)3 was employed as the catalyst for the reaction between 4-methoxybenzenediazonium tetrafluoroborate (2b) and 2,3-dihydrofuran (18) in the presence of NaOAc as base, 19 was obtained in 66% yield (entry 3). Disappointingly however, other diazonium salts assayed under the same conditions only gave the de-diazotized product (entries 5, 7).
Table 1: Matsuda-Heck reaction of 2,3-dihyrofuran.a
Subsequently, we attempted to make this reaction enantioselective with the addition of a chiral ligand (Table 2). Due to the inapplicability of chiral phosphines in this reaction (see introduction) we employed chiral nitrogen based ligands, L1
190
Chapter 7
and L2. As observed previously, anisole was the only product obtained, implying that the reduction took place much faster than the Matsuda-Heck reaction (entries 1, 2). To ascertain that the formation of anisole was not limited to L1 and L2, we also studied the reaction with 2,2’-bipyridine (L3) as the ligand (entry 3). Additionally, Pd-NHC (C1) was also tested as a catalyst (entries 5 - 8). Further, different bases were also tested to validate their role in the reaction (entries 7, 8). To our dismay, in all cases the expected Heck product was not observed.
Table 2: Attempted asymmetric Matsuda-Heck reaction of 2,3-dihydrofuran.a
Table 3: Attempted asymmetric Matsuda-Heck on cyclic ester.
[Pd] (mol%) Ligand (mol%) Base Yieldb eec
1 C1 (1) - - 60% -
2 Pd(OAc)2 (2) - - 61% -
3 Pd2(dba)3 (1) - - 9% -
4 Pd(OAc)2 (2) - CaCO3 38% -
5 Pd2(dba)3 (1) - CaCO3 27% -
6d Pd(OAc)2 (2) - - 57% -
7d Pd2(dba)3 (1) - - 14% -
8 Pd(OAc)2 (2) L1 (5) - - -
9 Pd(OAc)2 (2) L2 (5) - -
10 Pd(OAc)2 (2) L4 (5) - 70% 0
11 Pd(OAc)2 (2) L5 (5) - 47% 0
12 Pd(OAc)2 (2) L6 (5) - 50% 0
13 Pd(OAc)2 (2) L7 (5) - 54% 0
a Reaction conditions: 20 (0.5 mmol, 63 mg, 1 equiv), Pd precursor ligand, base (2 equiv), 2b (0.6 mmol, 1.2 equiv), DMF (1 ml), rt, 14 h. b Isolated yield. c Determined by chiral HPLC. d Reaction performed at 40 oC.
192
Chapter 7
Next it was chosen to study methyl 1-cyclopentene-1-carboxylate (20) as a substrate for the Matsuda-Heck reaction. Jung and coworkers had previously studied this substrate, for application in the enantioselective oxidative-Heck reaction.68 20 Proved to be a good substrate for the Matsuda-Heck reaction as well (Table 3). Pd-NHC (C1) and Pd(OAc)2 gave the highest yields. Surprisingly, Pd2(dba)3 gave a poor yield (9%, entry 3), while it gave the best yield with 2,3-dihydrofuran. The addition of a base to the reaction led to lower yields (entries 4, 5).
Performing the reaction at a higher temperature (40 oC) did not affect the reaction significantly (entries 2 vs. 6; 3 vs. 7). Encouraged by these results, we tested various chiral ligands in this reaction. Ligands L1 and L2 proved to be detrimental to the reaction (entries 8 and 9). Interestingly, ligand L4, a chiral secondary phosphine oxide, gave an improved yield of 70%, though without any enantioselectivity. Next we assayed phosphoramidites (L5-L7), taking the risk that these could be oxidized under the reaction conditions. Our rationale in assaying these was that since phosphoramidites are less electron rich ligands than phosphines in general, they might withstand oxidation. However, the yields from these reactions (entries 11 - 13) were only lower compared to reactions where no ligand was added (entry 2). In addition, no enantioinduction from these ligands was observed.
In view of the lack of enantioselectivity observed, along with the frequent observation of de-diazotization to the corresponding arenes, it was decided to concentrate our efforts on studying diastereoselective versions of the Matsuda-Heck reaction instead of the enantioselective version. This takes away the necessity to use chiral ligands.
7.3.5 Diastereoselective Matsuda Heck reactions
Despite being known in literature for nearly a decade, most examples of diastereoselective Matsuda-Heck reactions are limited to cases in which the stereodirecting groups were attached irreversibly (or reversibly with difficulty) to the substrate. Therefore, we attempted to develop a convenient route in which a chiral auxiliary is easily attached to the substrate, and easily removed as well after the diastereoselective reaction (Scheme 9).
193
Matsuda-Heck Reaction
Scheme 9: Proposed route for diastereoselective Matsuda-Heck reaction.
To this end, we chose to install chiral auxiliaries based on (1R,2S,5R)-menthol (A1) and (1R,2S)-camphorsultam (A2), as these can be easily appended and cleaved. The substrates 23, 25 were synthesized from commercially available 20, which upon hydrolysis gave 22, and subsequently were coupled to the corresponding auxiliary (Scheme 10).
Scheme 10: Synthesis of substrates 23 and 25.
The results of the attempted Matsuda-Heck reaction of 23 and 25 are presented in Table 4. The reactions did not proceed to completion even after 24 h and yields remained below 50%. The use of Pd-NHC C1 resulted in product formation in trace amounts (entries 3, 6). Unfortunately, no diastereoselectivity was observed in the
194
Chapter 7
reactions with either of the substrates. We reasoned here that the chirality was located remote from the double bond, plausibly resulting in poor chirality transfer.
Table 4: Diastereoselective Matsuda-Heck reaction with substrates 23 and 25.
Pd (mol%) Substrate Product Yield (%)b drc
1 Pd(OAc)2 (2.5) 23 24 38 1:1
2 Pd2(dba)3 (1.25) 23 24 47 1:1
3 C1 (1.25) 23 24 traces 1:1
4 Pd(OAc)2 (2.5) 25 26 30 1:1
5 Pd2(dba)3 (1.25) 25 26 36 1:1
6 C1 (1.25) 25 26 traces 1:1
a Reaction conditions: 23 / 25 (0.5 mmol, equiv), Pd precursor, 2b (0.6 mmol, 1.2 equiv), DMF (2 ml), rt, 14 h.b Isolated yield.c Determined by 1H-NMR and GC analysis of the crude reaction mixture.
As we observed no diastereoselectivity in the above approach, it was surmised to design novel substrates wherein the stereodirecting group was present in closer proximity to the Michael acceptor, such as chiral bicyclic enones. Substrate 35 was designed this objective.
Synthesis of 35 has been previously reported69 from 2-carene (30). However, 2-carene is considerably more expensive than its readily available isomer, 3-carene (29). In order to have sufficient starting material for our study, at an affordable price, we chose to synthesize 2-carene. For this, we elected a strategy based on the base-mediated isomerisation70 of 3-carene into a thermodynamic mixture
195
Matsuda-Heck Reaction
(55:45) of 3-carene and 2-carene (Scheme 11). As 29 and 30 are extremely difficult to separate by column chromatography or other means, we chose to continue the synthesis with this mixture and then isolate the products. 29 And 30 were oxidized into the corresponding diols (31 and 32), and subsequently oxidatively cleaved by NaIO4 into ketoaldehydes 33 and 34. The ketoaldehydes were found to be unstable to column chromatography and therefore were carried forward to a subsequent aldol condensation. Finally, 35 and 36 could be separated from one another by column chromatography. On occasion, a side product (37) was also formed during the aldol condensation step (Scheme 12), making isolation by column chromatography difficult. In this case, 37 could be converted to 36 by treating the reaction mixture with DBU, simplifying the separation (Scheme 12).
Scheme 11: Synthesis of 35.
Scheme 12: Formation of 37, and conversion to 36.
Subsequently, we proceeded to investigate 35 as a substrate for the Matsuda-Heck reaction (Table 5). Discouragingly, the reaction did not proceed to full conversion, even at a higher temperature (40 oC), and an inseparable mixture of the Heck
196
Chapter 7
product and the conjugate addition product were obtained. An improved yield of the mixture was obtained using K2CO3 as the base in acetonitrile (entry 2). The ratio of Heck product (38) to the conjugate addition product (39) seemed rather constant (3:1) for most of the bases tried (determined by GC), except for tributylamine, that led to a 1:1 ratio of the Heck product and conjugate addition product. This was interpreted as a consequence of the reductive nature of tributylamine (Chapter 2).
Table 5: Matsuda-Heck reaction of 35.a
Base Solvent Yield (%)b Ratio (38:39)c
1 K2CO3 DMF 20 3:1
2 K2CO3 CH3CN 60 3:1
3 i-Pr2NEt DMF 6 3:1
4 i-Pr2NEt CH3CN 18 3:1
5 KOAc CH3CN 44 3:1
6 NBu3 CH3CN 43 1:1
7d K2CO3 CH3CN 68 3:1
8e K2CO3 CH3CN 70 3:1
9f K2CO3 CH3CN 80 3:1
a Reaction conditions: 35 (0.5 mmol, 75 mg, 1 equiv), Pd2(dba)3 (1 mol%, 4.5 mg), base (1 mmol, 2 equiv), 4-methoxybenzenediazonium tetrafluoroborate (1 mmol, 2 equiv), solvent (2 ml), 40 oC, 14 h.b combined yield of 38 and 39. c Determined by GC analysis of the crude reaction mixture. d 3 Portions of Pd2(dba)3 (0.3 mol% each) were dosed every 3 h. e 3 Portions of 2b, (0.3 mmol each) were dosed every 3 h. f 8 Portions of a mixture of Pd2(dba)3 (0.13 mol% each) and 2b (0.13 mmol each) were added every 1 h.
Dosing additional quantities of catalyst and 4-methoxybenzenediazonium tetrafluoroborate (2b), individually and together, led to considerable improvements but full conversion could not be attained (entries 7 – 9). The best combined yield of 38 and 39 (80%, entry 9) was obtained when 8 portions of a mixture of Pd2(dba)3 (0.13 mol%) and 4-methoxybenzenediazonium tetrafluoroborate (2b, 0.13 mmol)
197
Matsuda-Heck Reaction
dissolved in CH3CN were added every 1 h, for the first 8 hours of the reaction and the reaction was then allowed to stir overnight (entry 9).
Due to the inseparability of the formed products, it was difficult to determine the diastereoselectivity of the formed Heck product, though only a single peak was observed on GC. Further studies with this substrate and catalyst system became impractical due to the long dosing procedures, and thereby, reproducibility of this procedure was not satisfactory.
The above considerations, combined with the fact that each substrate presented herein required its own optimization, with difficulty in reproducing yields, lead us to discontinue our study.
7.3 Mechanistic insights and the Meerwein arylation.
In spite of the growing importance of this reaction to modern synthesis and chemical industry, to date, only one study of its mechanism of the Matsuda-Heck reaction has been reported. In 2004, Eberlin and Correia disclosed their investigation39 on the mechanism by ESI-MS/MS (Scheme 13). Using 4-methoxyphenyldiazonium tetrafluoroborate and 2,3-dihydorfuran in the reaction, they were able to identify species B, C D and F. The authors found that a series of ligand exchanges took place on the metal during the reaction.
.
Scheme 13: Proposed catalytic cycle for Matsuda-Heck reaction by ESI-MS/MS.
198
Chapter 7
They propose that the dibenzylideneacetone (dba) ligand in Pd2(dba)3 is replaced by acetonitrile, forming species A, which undergoes oxidative nitrogen extrusion to give cationic species B. A ligand exchange on species B affords species C, into which, the alkene inserts, forming species D. Subsequent syn -hydride elimination (involving a disposition of the double bond) expels the product and affords palladium hydride E. The base deprotonates E to give F, which undergoes ligand exchange to form species A, a species that re-enters the catalytic cycle. This mechanism is largely consistent with the well-studied mechanism of the Mizoroki-Heck reaction.
While trying to rationalize the poor yields and selectivities observed for the Matsuda-Heck reaction in our study, especially when compared to the Mizoroki-Heck reaction, we looked for alternative routes that could lead to the same product. A reaction that is very similar to the Matsuda-Heck reaction in terms of the starting materials and the product is the Meerwein arylation71 (Scheme 14), which is mediated by radicals,72 and is promoted by CuII salts. Though Matsuda, in his initial work ruled out such a possibility based on the high yields of the product and substrate scope exhibited by the reaction,9 one is left to wonder if a radical pathway also contributes to the product formation.
Scheme 14: A Meerwein arylation reaction of aryldiazonium salts.
In fact, it has been observed, on occasion, that the “Matsuda-Heck” reaction is enhanced in presence of Cu salts.73 However, control experiments performed in the absence of Pd resulted in poor yields or no reaction, making it difficult to conclude if the reaction proceeded via Meerwein arylation or Matsuda-Heck reaction.
It might be of interest to note that the currently accepted experimental evidence for the mechanism of Matsuda-Heck reaction comes from ESI-MS/MS analysis of intermediates and disintegration products.39 As the method is not quantitative and dependent on charged species (or species that can be ionized under the conditions of the measurement), it is questionable if it may be held as unequivocal evidence for a two-electron transfer process. In fact, Pd and other group 10 metals are also known to promote one-electron radical processes.74 This might also account for the difficulty of making the reaction enantioselective. This is further supported by the
199
Matsuda-Heck Reaction
fact that there is a significant solvent effect on the reactivity (Table 5). DMF is known to form hydrogen radicals in the presence of a radical source, while CH3CN does not. This could perhaps explain the improved results observed when CH3CN is used as a solvent. Studies designed to ascertain the presence of radicals might shed more light into this.
7.5 Recent Developments
Following the discontinuation of our study, a report appeared in literature which marked the first example of an enantioselective Matsuda-Heck reaction.75 The combination of Pd/bisoxazoline ligand L8 was found to be the best ligand for the system, delivering ee’s up to 84% (Scheme 15). Though the catalyst loading is quite high (10% Pd), and the substrate scope demonstrated was limited, this report sets the path for further investigation in the area.
Scheme 15: The first example of an enantioselective Matsuda-Heck reaction.
7.6 Summary
During the course of this study, a straightforward protocol to obtain pure aryldiazonium salts via crystallization was developed. Butyl acrylate was found to be a very good substrate for the Matsuda-Heck reaction. 2,3-dihydrofuran (18) and methyl 1-cyclopentene-1-carboxylate (20) were also moderately successful for the reaction with yields of 66% and 60% respectively, though no enantioinduction was observed. Attempts towards a diastereoselective Matsuda-Heck via easily affixable chiral auxiliaries such as menthol and camphorsultam proved to be unsuccessful, and the corresponding products were isolated only in low yields. Substrate 35 resulted in an inseparable mixture of the Heck product and conjugate addition product, making the accurate determination of diastereoselectivity rather difficult.
Overall, it may be concluded that the Matsuda-Heck reaction for internal alkenes is very sensitive to minor changes in the reaction conditions. Yields vary drastically, and it seems plausible that competing radical reactions contribute the poor yields.
200
Chapter 7
7.7. Experimental
7.7.1 General
All experiments were carried out in flame dried or oven dried (150 oC) glassware, in
an atmosphere of nitrogen, unless specified otherwise, by standard Schlenk
techniques. Schlenk reaction tubes with screw caps, and equipped with a Teflon-
coated magnetic stir bar were flame dried under vacuum and allowed to return to
room temperature prior to being charged with reactants. A manifold permitting
switching between nitrogen atmosphere and vacuum was used to control the
atmosphere in the reaction vessel. Once charged with all the reactants, the
reaction vessel was cycled through at least 3 cycles of nitrogen-vacuum-nitrogen to
ensure the atmosphere was inert. Reaction temperature refers to the temperature
of the oil bath. Flash Chromatography was performed using Merck silica gel type
9385 (230-400 mesh), using the indicated solvents.
All solvents used for extraction, filtration and chromatography were of commercial
grade, and used without further purification. Reagents were purchased from
Sigma-Aldrich, Strem or Acros and used without further purification. Pd2(dba)3,C1
was purchased from Sigma-Aldrich and stored in a nitrogen filled glove-box.
TLC was performed on Merck silica gel 60, 0.25 mm plates and visualization was
done by UV and staining with Seebach’s reagent (a mixture of phosphomolybdic
acid (25 g), cerium (IV) sulfate (7.5 g), H2O (500 ml) and H2SO4 (25 ml)). 1H- and 13C-NMR were recorded on a Varian AMX400 (400, 100.59 MHz, respectively)
using CDCl3 as solvent, unless specified otherwise. Chemical shift values are
reported in ppm with the solvent resonance as the internal standard (CHCl3: 7.26
for 1H, 77.0 for 13C). Data are reported as follows: chemical shifts ( ), multiplicity
(s = singlet, d = doublet, t = triplet, q = quartet, br = broad, m = multiplet), coupling
constants J (Hz), and integration.
GCMS analysis was performed on a Agilent HP-6890 series with HP-5 ((5%-
phenyl)-methylpolysiloxane) or HP-1 (dimethylpolysiloxane) column (25 m x 0.25
mm x 0.25 μm). HRMS (EI) were obtained with a Thermo Scientific LTQ Orbitrap
201
Matsuda-Heck Reaction
XL spectrometer. Enantiomeric excess determination was performed by HPLC
(Shimadzu LC-20 AD) analysis using UV detection.
7.7.2 General procedure for the synthesis of aryldiazonium tetrafluoroborate
salts.
CAUTION! Diazonium compounds are potentially explosive and should be handled
with care. It is advisable to prepare them in small quantities.
Procedure adapted from literature62. A three-necked round bottom flask with
condenser and addition funnel was charged with BF3•OEt2 (22.5 mmol, 2.78 ml)
and cooled to -10 °C. Aromatic amine (15.0 mmol) in dry CH2Cl2 (30 ml) was
added, additional solvent (usually 5 ml) was added when stirring was hampered.
To the cooled solution was added dropwise a solution of t-BuONO (18.0 mmol,
2.14 ml) in dry CH2Cl2 (15 ml). After the addition was complete the reaction was
stirred for 30 min at the same temperature. Then the reaction was allowed to warm
to rt and stirred for 30 min. During this, the diazonium compound precipitated from
the reaction. Dry pentane (40 ml) was added and the precipitate was collected,
washed with ice cold dry Et2O (3 x 15 ml), and dried under vacuum.
Impure diazonium salts were dissolved in as little acetone (or acetonitrile) as
possible and then carefully precipitated with ice cold dry Et2O. The diazonium salt
was collected by filtration and washed with ice cold dry Et2O, and dried under
vacuum.
Note: All diazonium salts prepared were stored under nitrogen, at -18 °C.
Benzenediazonium tetrafluoroborate (2a): Prepared according
to the general procedure above and purified by crystallization.
Obtained in 87% yield, as a yellow solild. Product is commercially
available (CAS # 369-57-3).
N2BF4
202
Chapter 7
4-Methoxybenzenediazonium tetrafluoroborate (2b):
Prepared according to the general procedure above and
purified by crystallization. Obtained in 73% yield as an off-white
(74) Jahn, U. In Radicals in Synthesis III; Heinrich, M., Gansäuer, A., Eds.; Springer:
Heidelberg: 2012; Vol. 320, p 323.
(75) Correia, C. R. D.; Oliveira, C. C.; Salles Jr, A. G.; Santos, E. A. F. Tetrahedron Lett. 2012,
53, 3325.
(76) Iwasaki, T.; Maegawa, Y.; Hayashi, Y.; Ohshima, T.; Mashima, K. J. Org. Chem. 2008, 73,
5147.
(77) Fukuyama, T.; Arai, M.; Matsubara, H.; Ryu, I. J. Org. Chem. 2004, 69, 8105.
214
Chapter 7
(78) Vitnik, V. D.; Ivanovi , M. D.; Vitnik, Ž. J.; or evi , J. B.; Žižak, Ž. S.; Jurani , Z. D.;
Jurani , I. O. Synth. Commun. 2009, 39, 1457.
Summary
215
Summary
The construction of carbon-carbon bonds is central to organic chemistry. The advent of transition metal catalysis in organic synthesis has greatly contributed to the development of new reactions that allow the linkage of organic molecules with a variety of conditions, positions and functions. In recent years, palladium (Pd) has occupied a particularly prominent role, in terms of the variety of reactions that it can catalyze under reductive, oxidative and redox neutral conditions.
The Mizoroki-Heck reaction, a Pd-catalyzed reaction of aryl halides with alkenes, is a well studied synthetic transformation. The Mizoroki-Heck reaction with , -unsaturated ketones is regularly observed to result in a mixture of products, of which the conjugate addition product (or formally, the reductive arylation product) forms an important component. Interestingly, to date little attention had been paid to this, in order to exploit it as a successful synthetic protocol to arrive at conjugate addition products. Alternative procedures employ organometallics, which in turn are often obtained from the corresponding halides. Thus, it would be of considerable benefit if the halides themselves could be used to arrive at the conjugate addition product as the sole product of the reaction.
Chapter 2 of this thesis addresses this issue. For a series of aryl iodides, under the optimized reaction conditions described, the base plays the paramount role in discerning the selectivity between the Mizoroki-Heck reaction and conjugate addition. Our studies show that a conjugate addition product is favored when reductive bases (such as trialkylamines) are used, while non-reductive bases, (such as carbonates, carboxylates etc.) result in the Mizoroki-Heck product. Thus, just by changing the base, the selectivity between conjugate addition and Mizoroki-Heck reaction can be shifted completely (Scheme 1). In addition, we found that only 1.5 mol% of a commercially available Pd0-NHC is sufficient for this reaction. Furthermore, it was found that the conjugate addition of aryl halides could be extended to -nitrostyrenes, a previously unknown class of substrates for this reaction. Interestingly, the corresponding Mizoroki-Heck reaction of these substrates is unknown.
216
Scheme 1: Base dependent selectivity in the Pd-catalyzed reaction of aryl halides with enones.
Chapter 3 addresses the issue of selectivity and reactivity for cyclic enones as substrates in the Pd-catalyzed base-free, aerobic oxidative Heck reaction. While the oxidative Heck reaction is known to be high yielding and selective for many terminal enones, the selectivity and activity drop considerably when cyclic enones are employed. Furthermore, 2,9-dimethyl-1,10-phenanthroline (dmphen), a very commonly used ligand in oxidative Pd catalysis, has shown to be oxidized during the reaction via C-H activation of the proximal methyl groups. We therefore desired to develop a novel catalyst system that would result in a good selectivity and activity for the base-free aerobic oxidative Heck reaction of cyclic enones.
Our studies revealed that aryl bisimines of acenapthaquinone (commonly abbreviated as BIAN) are successful for this purpose. Using 5 mol% of readily available Pd(OAc)2 and 7 mol% of o,o’-dimethyl BIAN (Scheme 2), a number of arylboronic acids was found to undergo reaction with cyclohexenone. Both electron rich and electron poor boronic acids were found to be successful in the reaction, though arylboronic acids bearing an ortho-substituent were found to react slowly. Studies carried out in the presence of added halide salt were retarded, indicating that a cationic intermediate was pertinent in the reaction pathway.
Scheme 2: BIAN in the Pd-catalyzed base-free oxidative Heck reaction.
Chapters 4,5 and 6 of this thesis are dedicated to the study of Pd-catalyzed quaternary stereocenter formation via conjugate addition to , -substituted enones. While such a reaction catalyzed by copper or rhodium has been known for several years, both these strategies suffer from shortcomings that impede their application on a large scale. These include the rigorous exemption of air and moisture from the reaction vessel, combined with low temperatures in case of the former; and the
Summary
217
high price of rhodium in the case of the latter. A Pd-catalyzed procedure would circumvent both these shortfalls.
Chapter 4 addresses this issue for cyclic enones. A screening of ligands and Pd precursors taught us that cationic Pd complexes bearing chiral nitrogen ligands are essential for the success of the reaction. We were fortunate to find that the combination of readily available PdCl2, PhBOX and AgSbF6 was an excellent catalyst system for the reaction. Hexafluoroantimonate (SbF6
-) counterions were
found to play an important role in improving the conversion. Excellent enantioselectivities (up to 99%) were obtained for a series of 5, 6 and 7 membered carbocyclic enones (Scheme 3), while lower conversion and enantioselectivity was observed with heterocyclic enones.
As a consequence of the above strategy, we were able to perform the shortest known synthesis of (–)- -cuparenone (Scheme 4), in only 2 steps, starting from commercially available starting materials.
Scheme 4: A 2-step synthesis of (–)- -cuparenone.
In chapter 5, our efforts to develop a simple, and easily scalable catalytic system for the racemic conjugate addition for arylboronic acids to , -disubstituted enones are described. Using only 1 mol% of Pd(O2CCF3)2 and 1.5 mol% of 2,2’-bipyridine, good to excellent yields were obtained for a variety of cyclic enones. In case of acyclic enones, 5 mol% of Pd(O2CCF3)2 and 7 mol% of 2,2’-bipyridine was necessary, in addition to 20 mol% of KSbF6 as an additive and higher temperature. The presence of an allylic ether linkage was found to be essential for the success of these substrates. Replacing an allylic oxygen with a nitrogen was found to be detrimental.
218
Scheme 5: Construction of racemic benzylic quaternary centers.
Chapter 6 of this thesis is dedicated to the study of Pd-catalyzed conjugate addition of arylboronic acids to acyclic substrates. Acyclic substrates are usually more challenging for enantioselective catalysis, due to their lack of structural rigidity and ease of s-cis and s-trans isomerization. Building on studies described in chapters 4 and 5, we found that C1 symmetric chiral oxazoline ligands afforded improved selectivity for linear substrates bearing allylic ether linkages (Scheme 6). The optimal ligand, L1, could be synthesized in four steps. Yields and enantioselectivities obtained for the catalytic reaction were moderate. The highest ee obtained was 77%, using 3-chloro-4-isopropoxyphenylboronic acid. The cause of the lowered yield of the product, despite complete consumption of the starting material, was found to be a competing decomposition of the starting material.
Scheme 6: Pd-catalyzed conjugate addition of arylboronic acids to acyclic substrates.
Despite being known for more than three decades, an enantioselective version of the well known Matsuda-Heck reaction has remained an unsurmounted challenge. Our efforts to develop such a reaction are described in chapter 7. A few substrates that had been successful for the closely related, enantioselective Mizoroki-Heck reaction were studied, but these were found to be unsuccessful. Next, we focused our efforts on a diastereoselective version of the reaction, using easily appendable chiral auxiliaries. However, these efforts also did not bear fruit. Finally, we studied substrate 14, derived from 2-carene. This substrate resulted in a mixture of the expected Heck and the formal conjugate addition product (Scheme 7). Due to
Summary
219
impracticable addition times for the reaction and difficulty with reproducing the results, further studies in the topic were discontinued.
Scheme 7: A diastereoselective Matsuda-Heck reaction.
220
Samenvatting
221
Samenvatting
De constructie van koolstof-koolstof bindingen is cruciaal voor de organische chemie. De komst van overgangs metaal katalyse in organische synthese heeft een grote bijdrage gehad aan de ontwikkeling van nieuwe reacties, die de combinatie van organische moleculen met een variatie aan condities, posities en functies toestaan. In de afgelopen jaren heeft palladium (Pd) daar een bijzonder prominente rol in genomen, met name in betrekking tot de reacties die het kan katalyseren onder reducerence, oxiderende en redox neutrale condities.
De Mizoroki-Heck reactie, een Pd gekatalyseerde reactie van aryl halides met alkenen, is een goed bestudeerde synthetische transformatie. De Mizoroki-Heck reactie met , -onverzadigde ketonen resulteert meestal in een mengsel van producten, waarbij het geconjugeerde product (of formeel het gereduceerde arylatie product) een belangrijk bestanddeel vormt. Het is interessant dat hier tot op heden weinig aandacht aan is besteed, met de bedoeling dit tot een succesvol synthetisch protocol te gebruiken voor geconjugeerde addities. Alternatieve procedures gebruiken organometaal verbindingen, die op hun beurt verkregen worden van de overeenkomstige halides. Het zou dus zeer aantrekkelijk zijn om halides zelf te kunnen gebruiken om het geconjugeerde additie product als het enige product van de reactie te verkrijgen.
Hoofdstuk 2 van dit proefschrift gaat in op dit punt. Voor een serie van aryl jodides, onder de geoptimaliseerde beschreven condities, speelt de base een beslissende rol in het onderscheid van de selectiviteit tussen de Mizoroki-Heck reactie en de geconjugeerde additie. Ons onderzoek laat zien dat het geconjugeerde additie product het belangrijkste product is wanneer reductieve basen (zoals trialkylamines) worden gebruikt. Wanneer niet-reductieve basen (zoals carbonaten, carboxylaten, enz.) worden gebruikt resulteert dit in het Mizoroki-Heck product. Door alleen de base te veranderen, kan de selectiviteit tussen de geconjugeerde additie en de Mizoroki-Heck reactie dus compleet worden omgewisseld (Schema 1). Slechts 1.5 mol% van een commercieel verkrijgbare Pd0-NHC katalysator bleek voldoende te zijn voor deze reactie. Daarnaast vonden we dat de geconjugeerde additie van aryl halides kan worden uitgebreid tot -nitrostyrenen, een voorheen onbekende klasse van substraten voor deze reactie. Opmerkelijk genoeg is de corresponderende Mizoroki-Heck reactie niet bekend voor deze substraten.
222
Schema 1: Base afhankelijke selectiviteit in de Pd-gekatalyseerde reactie van aryl halides met
enonen.
Hoofdstuk 3 adresseert de kwestie met betrekking tot de selectiviteit en reactiviteit van cyclische enonen als substraten in de Pd-gekatalyseerde base-vrije, aërobe oxidatieve Heck reactie. Ondanks het feit dat de oxidatieve Heck reactie hoge opbrengsten en selectiviteiten voor veel eindstandige enonen geeft, vermindert de selectiviteit en activiteit beduidend wanneer cyclische enonen worden gebruikt. Bovendien wordt 2,9-dimethyl-1,10-phenanthroline (dmphen), een vaak gebruikt ligand in oxidatieve Pd katalyse, geoxideerd gedurende de reactie via C-H activatie op de proximale methyl groepen. Wij stelden ons daarom als doel om een nieuw katalytisch systeem te ontwikkelen dat zou resulteren in een goede selectiviteit en activiteit voor de base-vrije aërobe oxidatieve Heck reactie van cyclische enonen.
Ons onderzoek heeft aan het licht gebracht dat aryl bisiminen van acenapthaquinone (meestal afgekort als BIAN) succesvol zijn voor dit doeleinde. Gebruikmakende van 5 mol% Pd(OAc)2 en 7 mol% van o,o’-dimethyl BIAN (Schema 2) vonden we dat een redelijke hoeveelheid arylboorzuren een reactie kunnen ondergaan met cyclohexenon. Elektronenrijke en elektronenzuigende boorzuren bleken beide succesvol te zijn in de reactie. Arylboorzuren met een ortho-substituent bleken echter langzaam te reageren. Studies uitgevoerd met toegevoegd halide zout werden vertraagd, hetgeen er op wijst dat een kationisch intermediair passend is in het reactie mechanisme.
Scheme 2: BIAN in de Pd-gekatalyseerde base-vrije oxidatieve Heck reactie.
Hoofdstuk 4, 5 en 6 van dit proefschrift zijn gewijd aan het onderzoek naar de Pd-gekatalyseerde formatie van quaternaire stereocentra via geconjugeerde additie op
, -gesubstitueerde enonen. Ondanks het feit dat dit soort reacties al jaren bekend
Samenvatting
223
zijn met koper en rhodium hebben deze strategiën nog tekortkomingen die hun applicatie op grote schaal verhinderen. Dit zijn het benodigde rigoreuze verwijderen van lucht en vocht van het reactievat, gecombineerd met de noodzaak voor lage temperaturen in het geval van koper en de hoge prijs van rhodium in het geval van de laatste. Een Pd-gekatalyseerde procedure zou deze tekortkomingen omzeilen.
Hoofdstuk 4 adresseert dit probleem voor cyclische enonen. Een screening van liganden en Pd precursors leerde ons dat kationische Pd complexen met een chirale stikstof liganden essentieel zijn voor het succes van de reactie. Wij waren content om te vinden dat de combinatie van gemakkelijk verkrijgbaar PdCl2, PhBOX en AgSbF6 een perfect katalytisch systeem is voor deze reactie. Hexafluoroantimonaat (SbF6
-) tegenionen bleken een belangrijke rol te spelen in
het verbeteren van de conversie. Uitstekende enantioselectiviteiten (tot en met 99%) werden vekregen voor een serie van 5, 6 en 7 ledige koolstofcyclische enonen (Schema 3). Daarentegen werden lagere conversies en enantioselectiviteiten geobserveerd voor heterocyclische enonen.
Schema 3: Pd gekatalyseerde geconjugeerde additie voor de formatie van quaternaire
stereocentra.
Als gevolg van de hiervoor genoemde strategie konden we de kortste bekende synthese van (–)- -Cuparenone uitvoeren, in slechts 2 stappen, beginnend met commercieel verkrijgbare uitgangsmaterialen.
Schema 4: Een 2-staps synthese van (–)- -cuparenone.
In hoofdstuk 5 wordt onze bijdrage om een simpele en eenvoudig op te schalen katalytisch systeem voor de racemische geconjugeerde additie van arylboorzuren op , -digesubstitueerde enonen beschreven. Gebruikmakend van slechts 1 mol%
224
Pd(O2CCF3)2 en 1.5 mol% 2,2’-bipyridine werden goede tot uitstekende opbrengsten verkregen voor een diversiteit van cyclische enonen. In het geval van acyclische enonen was 5 mol% Pd(O2CCF3)2 en 7 mol% 2,2’-bipyridine nodig, met daarnaast een toegevoegde 20 mol% KSbF6 als additief en een hogere temperatuur. De aanwezigheid van een allylische ether binding bleek essentieel voor het succes van deze substraten. Het vervangen van een allylisch zuurstof voor een stikstof bleek nadelig.
Schema 5: Constructie van racemische benzylische quaternaire centra.
Hoofdstuk 6 van dit proefschrift is toegewijd aan het bestuderen van de Pd-gekatalyseerde geconjugeerde additie van arylboorzuren op acyclische substraten. Acyclische substraten zijn over het algemeen meer uitdagend voor enantioselectieve katalyse door de grotere flexibiliteit in hun structuur en het gemak van s-cis en s-trans isomerizatie. Verderbouwend op onderzoek uit hoofdstuk 4 en 5, vonden we dat C1 symmetrische chirale oxazoline liganden een verbeterde selectiviteit voor lineaire substraten met allylische ether bindingen gaven (Schema 6). Het optimale ligand, L1, kon worden gesynthetiseerd in vier stappen. Opbrengsten en enantioselectiviteiten vekregen voor de katalytische reacties waren gemiddeld. De hoogst verkregen ee was 77%, gebruikmakende van 3-chloor-4-isopropoxyfenylboorzuur. De oorzaak van de lagere opbrengst van het product, ondanks complete consumptie van het uitgansmateriaal, is een concurrerende decompositie van het uitgangsmateriaal.
Schema 6: Pd gekatalyseerde geconjugeerde additie van arylboorzuren op acyclische
substraten.
Samenvatting
225
Ondanks zijn bekendheid voor meer dan drie decennia is de enantioselectieve versie van de welbekende Matsuda-Heck reactie een onoverwonnen uitdaging. Onze inspanningen om deze reactie te ontwikkelen worden beschreven in hoofdstuk 7. Een aantal substraten die succesvol waren voor de verwante, enantioselectieve Mizoroki-Heck reactie werden onderzocht, maar bleken niet succesvol. Hierna verlegden we onze inspanningen op een diastereoselectieve versie van de reactie, gebruikmakende van eenvoudig aangebrachte chirale hulpstoffen. Deze inspanningen worpen echter ook hun vruchten niet af. Als laatste bestudeerden we substraat 14, een derivaat van 2-careen. Dit substraat gaf een mengsel van het verwachte Heck en het formele geconjugeerde additie product (Schema 7). Door onpraktische additietijden voor de reactie en het moeilijk kunnen reproduceren van de resultaten werd hier geen verder onderzoek naar gedaan.
Schema 7: Een diastereoselectieve Matsuda-Heck reactie.
Translated from English by Danny Geerdink. Edited by Derk Jan van Dijken.
226
Résumé
227
Résumé
La création de liaisons carbone-carbone est essentielle en chimie organique. L’apparition de la catalyse basée sur les métaux de transitions a largement contribué au développement de nouvelles réactions, permettant dès lors la formation de liaisons entre molécules organiques avec une grande variété de conditions, positions et fonctions. Ces dernières années, le palladium (Pd) a joué un rôle particulièrement important au niveau de la variété de réactions qu’il permet de catalyser, notamment en conditions réductives, oxydantes et neutres.
La réaction de Mizoroki-Heck, une réaction entre alcènes et halogènes d’aryles catalysée par du Pd, est une transformation synthétique qui a été largement étudiée. Cette réaction, si elle est employée avec des cétones - insaturées, résulte le plus souvent en un mélange de produits, parmi lesquels le produit d’addition conjuguée (ou formellement le produit d’arylation réductive) est particulièrement important. Il est intéressant de noter que jusqu’à présent, peu d’attention a été portée à ce phénomène dans le but d’exploiter cette réaction comme un protocole efficace pour arriver à ces produits d’addition conjuguée. Les procédures alternatives utilisent des composés organométalliques, qui sont eux-mêmes dérivés des halogènes. Dès lors, il serait très intéressant d’utiliser les halogènes eux-mêmes pour synthétiser sélectivement ces produits.
Le chapitre 2 de cette thèse aborde ce problème. Pour une série d’iodure d’aryles, avec les conditions optimales décrites, la base joue un rôle primordial dans la détermination de la sélectivité entre la réaction de Mizoroki-Heck et l’addition conjuguée. Nos études montrent que l’addition conjuguée est favorisée lorsque des bases réductives (comme les trialkylamines) sont utilisées, alors que l’utilisation de bases non-réductives (telles les carbonates et carboxylates) favorise le produit de la réaction de Mizoroki-Heck. Ainsi, juste en variant la base, la sélectivité entre l’addition conjuguée et la réaction de Mizoroki-Heck peut être complètement changée (Schéma 1). En outre, nous avons démontré que l’emploi de seulement 1.5 mol% d’un complexe au palladium (Pd0-NHC) disponible dans le commerce était suffisant, et que l’addition conjuguée d’iodures d’aryles était également possible avec des -nitrostyrènes, substrats qui étaient jusque-là
228
inconnus pour cette réaction. Etonnamment ces substrats ne sont toujours pas décrits pour la réaction de Mizoroki-Heck.
Schéma 1: Sélectivité en fonction de la base utilisée dans la réaction entre iodures d’aryles et
énones.
Le chapitre 3 traite le problème de sélectivité et de réactivité pour les énones cycliques comme substrats dans la réaction d’oxydation aérobique de Heck, sans base et catalysée au Pd. Alors que la réaction de Heck est connue pour ses rendements élevés et sa bonne sélectivité pour beaucoup d’énones terminales, l’utilisation d’énones cycliques conduit à des baisses notables en termes de rendements et de sélectivités. De plus, il a été démontré que 2,9-diméthyl-1,10-phénantroline (dmphen), un ligand très largement utilisé en catalyse oxydante au Pd, est oxydé pendant la réaction par activation de la liaison C-H des groupements méthyles proximaux. Par conséquent, nous avons voulu développer un système catalytique qui garantirait de bonnes sélectivités ainsi que de bons rendements dans la réaction d’oxydation aérobique de Heck, pour les énones cycliques.
Nos études ont révélé que les aryles bisimines d’acénaphtaquinone (communément abrégés BIAN), permettent d’atteindre avec succès cet objectif. En utilisant 5mol% de Pd(OAc)2 et 7 mol% de o,o’-diméthyl BIAN (Schéma 2), un nombre important d’aryles d’acide boronique réagissent avec la cyclohexénone. Les acides boroniques, électro-riches comme électro-déficients, se sont montrés efficaces pour cette réaction, bien que ceux étant substitués en position ortho s’avèrent avoir une vitesse de réaction plus lente. Les réactions effectuées en présence d’un sel d’halogène sont ralenties, mettant en évidence l’existence probable d’un intermédiaire cationique au cours de la réaction.
Schéma 2: BIAN dans la réaction d’oxydation aérobique de Heck, sans base et catalysée au Pd.
Résumé
229
Les chapitres 4, 5 et 6 sont consacrés à l’étude de la formation de stéréocentres quaternaires via addition conjuguée à des énones , -substitutées catalysée au Pd. Alors qu’une telle réaction catalysée par du cuivre ou du rhodium est connue depuis plusieurs années, ces deux stratégies souffrent de lacunes qui entravent leur application à grande échelle. Cela inclus l’absence totale d’air et d’humidité pendant la réaction, combinée avec des températures basses pour le cas du cuivre, et un prix très élevé dans le cas du rhodium. Une procédure catalysée au Pd permettrait alors d’outrepasser ces limitations.
Le chapitre 4 aborde ce problème dans le cas des énones cycliques. L’étude de différents ligands et précurseurs au Pd nous a appris que les complexes cationiques possédant des ligands azotés chiraux sont essentiels pour cette réaction. Nous sommes par la suite parvenus à démontrer que la combinaison de PdCl2, PhBOX et AgSbF6 représente un excellent système catalytique pour cette réaction, nos études montrant que les contre-ions SbF6
- jouent un rôle important dans l’amélioration de la conversion. D’excellentes énantiosélectivités (jusqu’à 99%) ont été obtenues pour une série d’énones carbocycliques à 5, 6 et 7 atomes (Schéma 3), alors que des conversions et énantiosélectivités inférieures sont observées dans le cas des énones hétérocycliques.
Schéma 3: Stéréocentres quaternaires formés par addition conjuguée.
Grâce à cette stratégie nous avons pu accomplir la synthèse la plus courte connue à ce jour de la (–)- -Cuparénone (Schéma 4) en seulement deux étapes, et ce à partir de substrats disponibles dans le commerce.
Schéma 4: Synthèse en 2 étapes de la (–)- -Cuparénone.
230
Dans le chapitre 5, nos efforts se sont portés sur le développement d’un système catalytique simple et facilement transposable à grande échelle pour l’addition conjuguée racémique d’aryles d’acide boronique à des , -énones disubstituées. En utilisant seulement 1 mol% de Pd(O2CCF3)2 et 7 mol% de 2,2’-bipyridine, des rendements allant de bons à excellents ont été obtenus pour une variété d’énones cycliques. Dans le cas des énones acycliques, 5mol% de Pd(O2CCF3)2 et 7 mol% de 2,2’-bipyridine sont nécessaires, en plus de l’ajout de 20 mol% de KSbF6 en tant qu’additif et d’une température plus élevée. La présence d’une fonction éther allylique est essentielle pour le succès de ces substrats, le remplacement de l’atome d’oxygène allylique par un atome d’azote étant préjudiciable.
Schéma 5: Construction de centres racémiques benzyliques quaternaires.
Le chapitre 6 de cette thèse est dédié à l’étude de l’addition conjuguée d’aryles d’acide boronique catalysée au Pd à des substrats acycliques. Ces derniers sont souvent plus délicats en termes de catalyse énantiosélective, dû à leur manque de rigidité structurelle ainsi qu’à la facilité d’une isomérisation entre s-cis et s-trans. En se basant sur les études décrites dans les chapitres 4 et 5, nous avons pu démontrer que l’utilisation d’oxazolines de symétrie C1 comme ligands résultait en une amélioration de la sélectivité pour les substrats linéaires pourvus d’un éther allylique (Schéma 6). Le ligand optimal, L1, a été synthétisé en 4 étapes. Les rendements et énantiosélectivités obtenus pour cette réaction catalytique sont modestes, la meilleure énantiosélectivité obtenue étant de 77%, en utilisant l’acide 3-chloro-4-isopropoxyphénylboronique. La cause de ce faible rendement, malgré une consommation complète du substrat, s’est avérée être une réaction concurrente de décomposition du substrat.
Résumé
231
Schéma 6: Addition conjuguée d’aryles d’acide boronique à des substrats acycliques.
Bien que connue depuis plus de trois décennies, la version énantiosélective de la réaction de Matsuda-Heck reste à ce jour un défi non surmonté. Nos efforts pour développer une telle réaction sont décrits dans le chapitre 7. Certains des substrats qui se sont montrés efficaces lors de la réaction énantiosélective de Mizoroki-Heck ont été étudiés, mais n’ont montré aucune réaction. Nous avons alors recentré nos efforts sur une version diastéréosélective de cette réaction, en utilisant des auxiliaires chiraux facilement incorporables. Cependant, ces efforts n’ont pas porté leur fruit. Enfin, le substrat 14 dérivé du 2-carène a été étudié, résultant en un mélange de produits d’addition conjuguée et de réaction de Heck (Schéma 7). Face aux difficultés à reproduire les résultats, et à des temps d’addition fastidieux, la poursuite d’études sur ce sujet a été abandonnée.
Schéma 7: Réaction diastéréosélective de Matsuda-Heck.
Translated from English by Céline Nicklaus and Mathieu Colomb-Delsuc.
232
233
Standing on the shoulders of giants…
As four years spent in Groningen come to an end, its time to look back and thank all the people who made this wonderful sojourn possible. Many have contributed, in ways more than one; to this book and my personal development. The following is my humble effort to express my gratitude to them.
I would like to start off by thanking the person who contributed the most for the success of the work described in this book: my promoter, Adri Minnaard. It has been an absolute pleasure to be a part of your group, and to have you as my mentor. I have learned a great deal from you, both personally and in chemistry. Your never-ending patience, and acute interest in scientific discussions, often led to several long and exciting discussions. The emails and conversations that followed our project on the quaternary centers being “scooped”, I shall remember for ever. Thanks to your support and confidence in my choices, I could continue on the project, and it has formed 3 chapters of my thesis. And perhaps, I could not finish my thesis in time, had it not been for you being proactive in getting a date for my defense, well in advance!
Hans, it has been wonderful to work with you. Your insights in Pd catalysis have been very helpful, in shaping the projects described in this book. I particularly remember our frequent discussions about mechanisms in your office, during your visits to Groningen. You hold my dream job, being at the interface of industry and academia, and I hope to step on several of the stones you have stepped on, during my career.
Next, I would to thank the members of the reading committee, Prof. Mats Larhed, Prof. Romano Orru and Prof. Sijbren Otto, for carefully going through my thesis,, providing valuable suggestions for its improvement, and most importantly, approving it for defense.
I would not have known a lot of the chemistry described in this book, myself, had it not been for my mentor, Prof. Pierre Dixneuf. I can never forget those amazing lectures on the fundamentals of organometallics and catalysis, during my days as a master’s student in Rennes. Your enthusiasm, good humor, hospitality and willingness to go out of your way to assist, shall always be remembered. Many thanks for the continued help and encouragement, over the years.
Ben, your energy and enthusiasm in chemistry, have been a great source of inspiration. Though, your involvement in regular sub-group discussions became
234
limited, over time, due to your extremely-busy schedule, your comments during our short discussions have been beneficial. Many thanks for stepping in to write a letter of support, whenever necessary. Gerard, Syuzi, Wes, Edwin and Anna: thanks so much for all the discussions and many useful suggestions, over the last four years.
I also take this opportunity to thank members from the Catchbio team, and particularly, our industrial-contact persons: Dr.Lavinia Panella, Dr. Bernard Kapteijn and Dr. Paul Alsters, for patiently listening to my updates and being supportive, even at times when progress seemed slow.
Over the last four years, two friends have shared every crest, trough and curve with me. Stepping into various roles, ranging from philosophers to cooks; you have become an integral part of my life. I would have not learnt so much about myself, had it not been for your care, affection and constructive criticism. Ashoka and Shiva: thanks-a-million for being around at all times and being the wonderful people you are. Though, I still have my own opinions of who is the best cook amongst us!
Kiran and Derk-jan (DJ), thanks for agreeing to be my paranymphs. Kiran, thanks for being the “synthesis muscle”, in the hour of need. It has been amazing working next to you. DJ, thanks for being my “translator-on-demand”, assisting me with all the paperwork and being a super-fun guy to hang out with. (Psst…I am not sure if I can make our “sausage-parties”, public knowledge!). Thanks to both of you, for proof-reading my thesis. Hari, of course, it would have been wonderful to see you standing beside me, dressed in a penguin-suit, on the day of my defense, but I absolutely understand, life’s callings are far more significant. You are with me in spirit, and that means a lot to me.
The two and half years spent in the Nijenborgh 4 are full of good memories. It was nothing short of a wonderful melting pot of exciting science, music and good times. Sharing the erstwhile Koffie Klub Island (14.245) with Danny, Santi, Miriam, Jeffrey, Bart, Maxime and Yange was a nice experience. It was fun to go through the daily playlist ranging from Santi’s classical music to Jeffrey’s Iron Maiden albums, spaced by my own requests for Roxette. Miriam, my neighbor for all the four years: It has been nice working next to you, and having hearty conversations while working late in the evenings. I wish you the best for finishing up your PhD! Santi, sharing the lab with you was fun, and I cannot think of anyone but you when I see a orange colored-microphone! Yange, it was fun to stand next to you during your days as a masters student. All the best with finishing-up your PhD.
Danny, it has been wonderful sharing the lab and the office, over all the four years with you. Thanks so much for the countless favors you have done for me, be it in
235
translating, making phone calls or even filling up the gaps in my dutch homework! And of course, travelling with you was a lot of fun. Our US trip would not have been the same without you and Chantal. And it was sure fun to go to the Roxette concert with you! Best wishes for your stay in Mulheim.
Being next to the Vyscom lab could only mean fun! Thomas, thanks so much for all the help over the period you were in Groningen; particularly for helping grow my first crystal for X-ray. I am very grateful for all the help and hospitality during my visit to the Firmenich facility at La Plaine. Felix, fellow labmate, officemate and Mac-mate: It was great fun to hangout with you, both in the lab and outside. Your lab space has been a treasure-trove of all sorts of glassware, perfectly organized. (Hey…but I wasn’t always to blame when things went missing!) Thanks so much for inviting me over to your wedding! Chris, you were great fun in the lab and it was memorable to hangout with you at Lowlands! Cati, thanks so much for your assistance with the BIAN project! I wish we had more time to collaborate. Peter, my personal gym guru: thanks so much for encouraging me in my efforts to get fit! Your plan for a CKD worked like a charm for me, but I only couldn’t stop eating like crazy after getting off it . I cannot think of anyone else but you when I hear Gunther! Manuel, my fellow Pd-crusader and Belly-Brother, it was nice to have many fruitful discussions with you. We ended up sharing rooms at most of the conferences we attended together. It’s good that you don’t snore louder than me, so I always had a good night’s sleep. By the way, we should take our Belly-Brothers act on tour, someday!
Bas and the India crew: Groningen, for me, would not have been the same without you! Bas and Bjorn (“B & B”), it was so much fun to hangout with you, be it at the coffee-breaks, in the bars - dancing the night away or sitting across the table and being involved in a “wise” discussion. Thanks so much for taking the effort to travel north, once in a while to reminisce the old times. Suzzane and Suzzane (!), thanks so much for being great friends. Judit and Johannes, congratulations once again on becoming proud parents! I am grateful for the wonderful hospitality during my visit to Zurich. Your wedding was a memorable occasion. I wish you both, the very best, in your future endeavors. Johannes, I eagerly look forward to the day, when I will hear of you as Prof. Teichert. Lachlan and Pieter, I was very hopeful of joining you guys in Boston, but things turned out otherwise! I hope that we can meet once again in the coolest city on the west-coast, Berkeley! ;) Alena, I consider myself luckily that my PhD period overlapped with your stay in Groningen. I wish you and Manolo, lots of success and happiness in Canada.
Martin, it was nice to have so many interesting discussions with you, be it in the lab or at the borrel! I wish Almudena and you, a bright future. Bin, I wish you the best
236
for finishing up your PhD. Celine et Mathieu, merci beaucoup pour votre aide avec la traduction. As both of you enter the final months of your PhD program, I keep my fingers crossed for you, to finish everything the way you planned. Anne, I still cannot knock the image, out of my mind, of you being a fresh PhD student, who just graduated with a master’s degree. I wish you much success for the (long!) PhD ahead. Pat and Massimo (“Is there a problem?”): all the very best for you both. Natasha, thanks-a-million, for hosting me in Zagreb.
Pub-quizzing occasionally with “Doctor A and the Minions” was an important activity on the social calendar. I, being irregular, had the advantage of having several different team mates over the period. Annick (“Dr. A”), Bas, Julia, Felix, Jack, Celine, Mathieu, Petra, Gerald, Francesco: all helped at making “logical guesses” for the answers. (Gerald, thanks so much for lending me your tent for the weekend at Lowlands!) And of course, there were other teams who tried to be as cool as ours’ ;). Jos, Wiktor, Jochem, Johannes, Pieter, Lachlan: It was fun competing with you. Wiktor, thanks so much for all the help, particularly, for willing to be a part of all my practice talks, even on short notice. Jos, thanks for giving me the first lesson in wall climbing. Jochem, thanks for the many fun-filled evenings we have had. It was great fun to travel with you on our tour of the US west-coast. “What happens in Vegas, stays in Vegas”, so they say, and hence I say no further. Megan and Jack, thanks so much for hosting me during my visit to the west-coast.
The Dancing Queens, Matea and Francesca, thanks for the many Salsa outings. Matea, thanks so much for inviting me over to your wedding, and convincing me to take dancing lessons. It has been one of the best experiences I have had during my time in Groningen.
Moving to the Linnaeusborg, brought a new mix of labmates, and particularly closer to the members of the Hirsch group. Tiziana (Ehi!), it has always been great fun to hangout with you and Ugo. Blijke, it was fun sharing the office with you, and to have someone around to complain to, when I was in the mood to do so. ;). Best wishes for your future endeavors. Milon, thanks so much for all the wonderful dinners. Simon, Steven, Stephen, Vasu, Leticia..best wishes for the days ahead. Bea, Jeffrey, Zhongtao, Selma, Wendy, Claudia, Wim, Maria, Micheal, Mathieu Denis, Peter ( X 3!), it was nice sharing the lab with you all.
Jeffrey Bos, Stratingh’s Photographer-Laureate, thanks so much for your help with designing the cover. I look forward to see Stratingh’s first ever 3D movie.
My friends from the Inorganic world, Julia, Seb, Johanne, Douwe.. it was fun to hang out with you guys at the borrels.
237
Over the years, I was fortunate to work with several motivated students. Dorus, it has been wonderful sharing the lab and the office with you. My stay in the Netherlands would have been half the fun, if it hadn’t been for you. I will fondly remember the many conversations we had: over coffee, lunch, in bars, over the internet and even in sultry Mumbai. Thanks so much for giving me my first lessons in skating, and taking me out on bike rides. Edwin, it was great working with you on the Matsuda-Heck project. I appreciate the bold face you maintained despite all the difficulties the project seemed to throw at us. I wish you much success in your PhD. Niek, I have had a lot of fun working with you when you were bachelor’s student and then having you as my neighbor during your master’s. It has been fantastic to see you transition from a bachelors student, to a masters student and now, as a PhD-student-to-be. I wish Melanie and you, the very best! Jasmin, it was very nice working with you during your 18 month stay in Groningen. Thank you very much for your contributions on the quaternary centers project. I wish I had learnt more basketball from you, during your time here! I wish you success in your career. Maxime, it was nice to be involved in the synthesis of chiral NHC ligands with you. I learned a lot from the effort. Merci beaucoup! Sven, it was nice to see that our work on the oxidative Heck reaction could find application in a biochemical environment. I look forward to hear more exciting news about it, in the future!
In addition to above, I was fortunate to have several students over short periods. Though the time was short, the fun was not. Pim, Wendy and Anouk: Many thanks!
Being part of an institute as large as the Stratingh, means there are several people whom we meet, share good times with and engage in interesting conversations. In view of practicality, I make a list of all the people who contributed, in their own unique way to make my stay memorable.
Many thanks are due to Monique and Theodora for their help with instrumental analysis. Assistance from Hilda, Ebe, Alphons, Hans van der Welde and Pieter van den Meulen is gratefully acknowledged.
Life outside the institute is equally important. I was fortunate to serve a year on the board of the Gronignen Indian Students Association (GISA), which brought me in contact with many people in different spheres. Harsh, Bhushan, Prachi, Yamini,
238
Gopi (X 2), Sasanka, Swagath, Lakshmi, Amol, Sneha, Praveen, Amol, Shivashankar, Rama, Swapnil, Milind, Tushar, Siddesh: Thanks for all the good times. Participating occasionally in the weekend cricket matches at the parking lot in Zernike ( better known as Zernike International Cricket Stadium or ZICS), has been a good way to meet and socialize. Afzal, Irfan, Gaurav, Gautham (muscle-man), Eswar (dialogue king!), Appu (our star batsman!), Arun (a multifaceted sportsman), Pranov, Praneeth, Soumen, Milon and Suresh. Special thanks to you all, and for hosting excellent parties. Suresh, thanks so much for offering to read through my thesis and weed-out minor mistakes, and making me feel at home with your excellent culinary skills.
Anna, special thanks to you for the wonderful times. Thanks to you, I have become a lot more organized and patient. I wish you the very best, and I hope to take you around India, sometime. And now that I have been taking Salsa lessons, we have to give it a try ;).
Finally, to my family back home, in India. Amma and Nanna, your unwavering support and love, have been my biggest strengths. Thanks so much for having the courage to take several difficult decisions and being unhesitant to sacrifice, in order to see me where I am today. This is neither the beginning nor the end, but I am grateful for affection and care, at each instance. To my grandparents, who do not understand much of what I do all day, but have been wishing and praying for my success and well-being at all times, thank you. My uncles, aunts and cousins who have celebrated each of my success and shared each of my difficulty, as if it were their own: many more thanks.
Anyone, whom I missed to thank in the lines above, was only due to over-sight. I beg your pardon and express my heartfelt gratitude, for all the help.
As this phase in my life as “scientific nomad” comes to end, I thank all the giants, on whose shoulders; I have had the opportunity to stand. With a heavy heart, it remains to set course to a new destination, and look forward to what’s next.
– Aditya
About the cover: The cover is a representation of the times I spent in Groningen. The front cover depicts the projects described in this thesis: as articles for published results, and data for unpublished work; along with an X-ray structure of the most successful catalyst from my work. The challenges that remain after my research are marked on the post-it. The back cover, offers a glimpse of the times spent outside the lab and things special to Groningen and the Netherlands.
