Synlett 2016; 27(03): 331-336
DOI: 10.1055/s-0035-1561282
synpacts
© Georg Thieme Verlag Stuttgart · New York

Dehydrogenation Adjacent to Carbonyls Using Palladium–Allyl Intermediates

Aneta Turlik
Department of Chemistry, Yale University, 225 Prospect Street, New Haven, Connecticut 06520-8107, USA   Email: timothy.newhouse@yale.edu
,
Yifeng Chen
Department of Chemistry, Yale University, 225 Prospect Street, New Haven, Connecticut 06520-8107, USA   Email: timothy.newhouse@yale.edu
,
Timothy R. Newhouse*
Department of Chemistry, Yale University, 225 Prospect Street, New Haven, Connecticut 06520-8107, USA   Email: timothy.newhouse@yale.edu
› Author Affiliations
Further Information

Publication History

Received: 27 October 2015

Accepted after revision: 16 November 2015

Publication Date:
23 December 2015 (online)

 


Abstract

Palladium–allyl chemistry has historically been used as a means of allylation of nucleophiles, as developed by Tsuji and Trost in the 1970s. Also during this decade, the Saegusa oxidation, the most prominent palladium-catalyzed dehydrogenation reaction, was developed. This Synpacts article provides a historical overview of dehydrogenation adjacent to carbonyls and our recent contribution to this area: palladium–allyl catalyzed dehydrogenation of nitriles and esters.


#
Zoom Image
Aneta Turlik(right) was born in Białystok, Poland, and grew up in Brooklyn, New York. She received her BA in Biochemistry at Barnard College, where she worked under the direction of Professor Christian Rojas. In 2013, she began PhD studies at Yale University as a member of the Newhouse group.
Yifeng Chen (left) was born and grew up in Anhui, China. He received his BA in Chemistry from Soochow University in 2007 and completed his PhD under the supervision of Prof. Yuanhong Liu at Shanghai Institute of Organic Chemistry (SIOC) in 2012, focusing on homogenous gold catalysis. In 2013, he moved to the US as a postdoctoral fellow in the Buchwald group at MIT, working on continuous-flow chemistry. In 2014, he joined the Newhouse group as a postdoctoral fellow.
Timothy R. Newhouse (middle) was born in New Hampshire and grew up in northern New England. He received his BA in chemistry from Colby College (2005) in Waterville, ME, where he was mentored by Prof. Dasan M. Thamattoor. After moving to La Jolla, CA, he completed his PhD at The Scripps Research Institute with Prof. Phil S. Baran (2010). During his time at Scripps, he also worked in the laboratories of Prof. Donna G. Blackmond. He then returned to the east coast for postdoctoral studies with Prof. E. J. Corey at Harvard University. As of 2013, he is an Assistant Professor at Yale University in the Department of Chemistry.

While studies initiated in the mid-19th century used the oxidation of inert C–H bonds as a tool to understand the structure and electronic properties of organic molecules, modern efforts use C–H oxidation to understand reactivity and to access molecules by direct means. These oxidation methodologies often allow for reduction of the number of chemical operations in a synthetic sequence and eliminate the production of waste associated with classical processes. Palladium catalysis has accelerated advancements in the functionalization of C–H bonds with significant contributions coming from the research laboratories of Yu, Sanford, and others to convert C–H σ bonds to C–X σ bonds.[1] A less-studied reaction class is the palladium-catalyzed oxidation of alkanes to alkenes, which converts C–H σ bonds into C–C π bonds. Practical methods using this strategy to transform carbonyls to their α,β-unsaturated counterparts have recently been developed by Stahl and coworkers, as well as by our laboratory.[2]

1

Selenium- and Sulfur-Based Dehydrogenations

The direct dehydrogenation adjacent to electron-withdrawing groups dates to studies by Riley and coworkers using selenium dioxide during the interwar period.[3] Since then, several other dehydrogenation methods have been developed.[4] While investigating the introduction of the ketonic function α to esters (12, Scheme [1, a]), it was serendipitously discovered that the oxidation of ethyl succinate (3) produces the dehydrogenated product, ethyl fumarate (4) instead of the expected keto ester. These oxidation reactions may proceed via β-keto seleninic acid or selenium ester intermediates.[5]

Zoom Image
Scheme 1 Early studies on selenium- and sulfur-based dehydrogenation strategies

Similar to the earliest studies in carbonyl α,β-dehydrogenation, most methodologies involve initial α-functionalization of a carbonyl compound followed by an elimination reaction (Scheme [1]). Another traditional method, α-halogenation and base-mediated dehydrohalogenation, is still occasionally employed in multistep synthesis.[6] α-Phenyl selenide intermediates are more commonly utilized, owing to the mild conditions under which they can be produced: deprotonation with an amide base followed by treatment with PhSeX at low temperature (Scheme [1, b]).[7] In a separate step, the resulting phenyl selenides can be transformed into the α,β-unsaturated products 6 by oxidation of the selenide to a selenoxide 7, which spontaneously undergoes syn elimination by a [2,3]-sigmatropic rearrangement.[8] Unlike most other dehydrogenation methodologies, this transformation is capable of successfully dehydrogenating most types of carbonyl compounds, including those containing esters and amides. However, it rarely tolerates the presence of other oxidation-prone functionalities such as unprotected alcohols and amines.

In an analogous manner, sulfenylation and sulfoxide elimination also leads to pericyclic dehydrogenation, as shown by Trost in 1973 (89, Scheme [1, c]).[9] Although it avoids the toxicity of selenium, this technology has been less frequently employed in part due to the higher temperatures under which the elimination occurs (>50 °C). A conceptually similar dehydrogenation was developed by Mukaiyama and coworkers,[10] using N-tert-butyl phenylsulfinimidoyl chloride (12), which can effect an in situ dehydrogenation via the intermediacy of an α-sulfinimidoyl group (13, Scheme [1, d]). A similar cyclic transition state leads to the dehydrogenated compound. While ketone-containing compounds are efficiently dehydrogenated, other carbonyl compounds, such as esters, are more challenging and lead to incomplete conversion of starting materials.[11] Due to their similar polarities, isolating starting materials from products is difficult by chromatographic purification. In addition, the method suffers from limited stability of the reagent, as well as difficulty in removal of the byproducts.


# 2

DDQ and IBX

In the early 20th century, quinones were identified as oxidants capable of dehydrogenating activated carbonyl compounds, and in the mid-20th century 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) was identified as the most generally useful quinone oxidant.[12] This reaction represents another class of dehydrogenations, which do not proceed via an α-functionalized intermediate. Historically, DDQ was developed for the dehydrogenation of steroids. While its use for dehydrogenation of unactivated nonsteroidal ketones is limited, studies have attempted to remedy this issue by activation via a silyl enol ether.[13] Despite some advancements, the limited scope of compatible substrates for the DDQ oxidation and the difficulty in removing the stoichiometric byproduct produced has limited its impact on organic synthesis.

Within the area of carbonyl α,β-dehydrogenation, this reagent has been used more modernly for compounds in which aromaticity is a driving force for oxidation,[14] such as chromanones (1415, Scheme [2, a]).[15] α,β-Dehydrogenations of groups less acidic than aldehydes and ketones using DDQ are limited to isolated examples which result in compounds with high degrees of conjugation or aromaticity.[16]

Zoom Image
Scheme 2 One-step dehydrogenation methodologies using stoichiometric oxidants

A major advancement was made towards the end of the 20th century by the identification of an alternative stoichiometric oxidant – a hypervalent iodine reagent, 2-iodoxybenzoic acid (IBX) – that effects the conversion of ketones and aldehydes into the corresponding enones and enals (Scheme [2, b]).[17] This dehydrogenation is thought to proceed via a single-electron-transfer mechanism.[18] Taking advantage of this mechanistic insight, ligands such as 4-methylmorpholine-N-oxide (NMO) and 4-methoxypyridine-N-oxide (MPO) have been used to effect mono- or didehydrogenation in a controlled manner.[19] As is the case with other oxidants, IBX has the ability to oxidize alcohols in the presence of carbonyl compounds, as well as to concomitantly perform sequential oxidations (e.g., alcohol to enone). Although limited to ketones and aldehydes, IBX-mediated α,β-dehydrogenation remains one of the premier synthetic methodologies.


# 3

Palladium Catalysis

In the late 20th century, the first examples of palladium-mediated dehydrogenation reactions were discovered. The Saegusa oxidation, now another one of the most widely employed methodologies, proceeds via an alternative mode of carbonyl activation – formation of an enoxysilane (18) – and subsequent dehydrogenation mediated by palladium (Scheme [3, a]).[20] The reaction is believed to proceed via a palladium enolate 19 that undergoes a β-hydride elimination to deliver the α,β-unsaturated product 11, although mechanistic investigations have been limited.[21] Despite the instability of the intermediate enoxysilanes and the cost associated with the high catalyst loadings of palladium required for efficient transformations, it is one of the most frequently employed methods for dehydrogenation adjacent to unactivated ketones.

Zoom Image
Scheme 3 Dehydrogenation using palladium catalysis

A mechanistically distinct palladium-catalyzed conversion of ketones and aldehydes into enones and enals was recently reported by Stahl and coworkers (Scheme [3, b]).[22] In these reactions, unsaturated carbonyl compounds are directly accessed by treatment of carbonyl-containing substrates with a palladium(II) source and oxygen as an oxidant under acidic conditions. Choice of ligand allows for the selective formation of the α,β-unsaturated product instead of the higher-oxidation-state phenol product.[23] This process is thought to proceed via palladium-mediated C–H activation (2021), rather than by dehydrosilylation, as in the Saegusa oxidation. Importantly, this methodology allows for the direct conversion to the unsaturated derivatives, avoiding the synthesis of unstable enoxysilane compounds.

Although useful one-step dehydrogenation methodologies have been developed, these techniques have been limited to the oxidation of ketones and aldehydes, while a one-step methodology for α,β-dehydrogenation of less-acidic functionalities, like esters, nitriles, and amides,[24] has remained elusive (Scheme [4, a]). We focused our studies on the synthetic problem of developing a method that would allow for the α,β-dehydrogenation of a wide variety of electron-withdrawing groups that have typically required multiple steps to efficiently dehydrogenate.

Zoom Image
Scheme 4 Development of a one-step palladium-catalyzed reaction for the α,β-dehydrogenation of a variety of electron-withdrawing groups

We envisioned a dehydrogenation that would initiate with an in situ deprotonation and transmetalation to form a palladium enolate (23, Scheme [4, b]). The reaction pathway would proceed with a β-hydride elimination, in a similar fashion to the Stahl and Saegusa methods, to form the unsaturated product 22. The catalytic cycle would be completed after the resulting allyl–Pd hydride species 24 reductively eliminated to form propene and palladium(0), which would be oxidized by an allyl electrophile to reform the active catalyst 26.[25]

The selection of an allyl electrophile as an oxidant under these reaction conditions meant that the catalytically formed Pd–allyl species could also form the typical allylation product 28 (Scheme [4, c]).[26] Historically, Pd–allyl species have been used primarily for allylation of nucleophiles in the Tsuji–Trost reaction and its variants. This reaction can be traced to 1965, a few years after the discovery of the Wacker process,[27] when Tsuji and coworkers reported the nucleophilic attack of [Pd(allyl)Cl]2 by carbanions to form allylated products, using stoichiometric amounts of palladium. Beginning in the 1970s, Trost and coworkers developed a catalytic and enantioselective variant, culminating in the identification of a ligand that reliably produces allylation products with high enantioselectivity.[28] This reaction proceeds by attack of a nucleophile, such as an enolate, at the carbon of an electrophilic Pd–η3-allyl species to produce a new carbon–allyl bond.[29] Palladium(0) is thus formed, which can undergo oxidative addition to form a new Pd–allyl species.

If the allylation pathway could be avoided, using allyl acetate as a mild oxidant would allow for compatibility with a wide range of functional groups, as allyl acetate is considerably less electrophilic than most typical oxidants. Early investigations of our dehydrogenation reaction were complicated by this well-precedented allylation byproduct 33, as well as condensation byproducts such as 34 that arose from direct attack of the enolate onto the allylic ester oxidant (Scheme [5]). The allylation byproduct may arise via direct attack of the enolate at the carbon of the Pd–allyl species.[30] Alternatively, an inner-sphere mechanism could be envisioned wherein initial attack at palladium of the Pd–allyl complex is followed by C–C bond-forming reductive elimination.[31] We sought to prevent byproduct formation by altering the identity of the enolate.

Zoom Image
Scheme 5 Deprotonation with LiTMP and transmetalation to a zinc enolate allows for full conversion into α,β-unsaturated product

Transmetalation from a lithium enolate to a less reactive zinc enolate by addition of ZnCl2 eliminated the formation of the condensation byproduct 34. Additionally, this allowed for selectivity for dehydrogenation (32, Scheme [5]), without detection of the allylation byproduct 33. This reversal in selectivity from the usually observed allylation product is likely a function of the ligand environment on palladium and not solely a function of the identity of the enolate,[32] as Negishi has observed that zinc enolates also react at the carbon of a Pd–π-allyl complex with Ph3P as a ligand.[33] The efficiency of this transformation was also a function of the amide base – LiTMP was found to be optimal relative to other readily available amide bases. While allyl pivalate was the most successful oxidant for ester dehydrogenation, allyl acetate was superior in the case of nitriles.

With these optimized reaction conditions in hand, we set out to evaluate the scope of this dehydrogenation reaction. Substrates containing protected alcohols (36a), protected amines (36c,j, 36k), furans (36e), and indoles (36d) all produced high yields of dehydrogenation product (Scheme [6]). It is noteworthy that tertiary amines 36k underwent smooth dehydrogenation, despite their known propensity to undergo oxidation under palladium catalysis.[34] In addition, dehydrogenation adjacent to nitriles and esters both tolerated α-substitution (36a,c,g,h,j,k) as well as β,β-disubstitution (36b,l). Lactones as well as linear and cyclic esters were successfully dehydrogenated. Unlike with other one-step dehydrogenation methods, overoxidation was not observed for any substrate. Dehydrogenation was unsuccessful with some functionalities, including α-amino esters and nitriles, lactones without β-substitution, and α-aryl nitriles.

Zoom Image
Scheme 6 Select scope of dehydrogenation depicts a wide variety of nitrile and ester substrates

In order to gain insight into the mechanistic pathway by which this reaction proceeds, we performed kinetic isotope studies. Measurements examining substitution of the β-hydrogen atom with deuterium revealed a parallel KIE value of 2.3 ± 0.1 (Scheme [7, a]), which indicates that C–H bond cleavage or formation is involved in a turnover-limiting step.

Zoom Image
Scheme 7 Parallel and intramolecular kinetic isotope effect studies suggest that reductive elimination is the turnover-limiting step

To distinguish whether β-hydride elimination or reductive elimination is turnover limiting, the intramolecular KIE value was measured and found to be 1.0 ± 0.1 (Scheme [7, b]). These studies suggested that β-hydride elimination is a reversible process that is followed by a turnover-limiting reductive elimination.

In addition to providing a historical overview of dehydrogenation adjacent to carbonyl compounds, this Synpacts has highlighted the synthetic potential of Pd–allyl intermediates generated from allyl electrophiles. In the area of α,β-dehydrogenation, future work will focus on more challenging substrate classes and on determining by what means site selectivity may be obtained when multiple easily oxidized functionalities are present.


#
#

Acknowledgment

We acknowledge Yale University, the NSF (GRF to A.T., DGE-1122492) and the Anderson Foundation (postdoctoral fellowship to Y.C.) for funding. Mr. Justin P. Romaire is acknowledged for experimental contributions.



Zoom Image
Aneta Turlik(right) was born in Białystok, Poland, and grew up in Brooklyn, New York. She received her BA in Biochemistry at Barnard College, where she worked under the direction of Professor Christian Rojas. In 2013, she began PhD studies at Yale University as a member of the Newhouse group.
Yifeng Chen (left) was born and grew up in Anhui, China. He received his BA in Chemistry from Soochow University in 2007 and completed his PhD under the supervision of Prof. Yuanhong Liu at Shanghai Institute of Organic Chemistry (SIOC) in 2012, focusing on homogenous gold catalysis. In 2013, he moved to the US as a postdoctoral fellow in the Buchwald group at MIT, working on continuous-flow chemistry. In 2014, he joined the Newhouse group as a postdoctoral fellow.
Timothy R. Newhouse (middle) was born in New Hampshire and grew up in northern New England. He received his BA in chemistry from Colby College (2005) in Waterville, ME, where he was mentored by Prof. Dasan M. Thamattoor. After moving to La Jolla, CA, he completed his PhD at The Scripps Research Institute with Prof. Phil S. Baran (2010). During his time at Scripps, he also worked in the laboratories of Prof. Donna G. Blackmond. He then returned to the east coast for postdoctoral studies with Prof. E. J. Corey at Harvard University. As of 2013, he is an Assistant Professor at Yale University in the Department of Chemistry.
Zoom Image
Scheme 1 Early studies on selenium- and sulfur-based dehydrogenation strategies
Zoom Image
Scheme 2 One-step dehydrogenation methodologies using stoichiometric oxidants
Zoom Image
Scheme 3 Dehydrogenation using palladium catalysis
Zoom Image
Scheme 4 Development of a one-step palladium-catalyzed reaction for the α,β-dehydrogenation of a variety of electron-withdrawing groups
Zoom Image
Scheme 5 Deprotonation with LiTMP and transmetalation to a zinc enolate allows for full conversion into α,β-unsaturated product
Zoom Image
Scheme 6 Select scope of dehydrogenation depicts a wide variety of nitrile and ester substrates
Zoom Image
Scheme 7 Parallel and intramolecular kinetic isotope effect studies suggest that reductive elimination is the turnover-limiting step