Catalytic Synthesis of Cyclopropenium Cations with Rh-Carbynoids.

Herein, we report the first catalytic one-step synthesis of cyclopropenium cations (CPCs) with readily available alkynes and hypervalent iodine reagents as carbyne sources. Key to the process is the catalytic generation of a novel Rh-carbynoid that formally transfers monovalent cationic carbynes (:+C-R) to alkynes via an oxidative [2+1] cycloaddition. Our process is able to synthesize a new type of CPC substituted with an ester group that underpins the regioselective attack of a broad range of carbon and heteroatomic nucleophiles, thus providing a new platform for the synthesis of valuable cyclopropenes difficult or not possible to make by current methodologies.

Cyclopropenium cations (CPCs), discovered by Prof. Ronald Breslow in the late 1950s,[1−3] are the smallest member of the Hückel aromatic systems. These aromatic cations with two π-electrons delocalized over three 2p orbitals are known to have considerable thermodynamic stability and molecular strain (Figure A). The highly stable tris(dialkylamino)CPC derivatives have recently found applications as electrophotocatalysts,[4] gene delivery agents,[5] catholytes for nonaqueous redox batteries,[6] or liquid crystals.[7] However, CPCs have been largely underappreciated in organic synthesis, despite their potential as three-carbon building blocks in the construction of complex architectures.

Figure 1

Catalytic synthesis of cyclopropenium cations.

One of the main reasons is that the majority of synthetic strategies toward CPCs, developed between the 1950s and 1980s, rely on multistep sequences and show limitations in efficiency. Those methods require the synthesis of cyclopropene derivatives that lead to the CPC upon (pseudo)halide, nitrile, carbonyl, or hydride abstraction with strong Lewis/Brønsted acids.[1,3,8] Alternative approaches based on functionalizations of alkynes with chlorocarbenes generated from chlorodiazirines under UV light irradiation[9] or with a cationic metal-carbyne [(η5-C5H5)(CO)2Mn≡CPh]+ show very limited scope (only two examples are described in the latter case).[10] In addition to this, nonsymmetrical CPCs are likely to react with poor regioselectivity in nucleophilic events. In fact, pioneering work by Padwa showed that reactions between CPCs and Grignard reagents provided mixtures of regioisomers.[11]

Herein, we would like to disclose the invention of a one-step Rh-catalyzed process for the preparing of CPCs that combines readily available alkynes and hypervalent iodine reagents as formal cationic carbyne sources (Figure B).[12] The process showed a broad scope of a new class of CPCs substituted with an ester group that, upon treatment with a diverse range of nucleophiles, provided access to valuable and elusive classes of cyclopropene derivatives.[13]

Our research group is focused on the development of a carbyne transfer platform in organic synthesis using tailored hypervalent iodine reagents[14] as carbyne synthons.[15] The catalytic activation with dirhodium carboxylate catalysts[16] provides access to Rh-carbynoids () (Figure ),[15b,15d] a novel class of Rh-carbene species substituted with an ester group and a hypervalent iodine moiety as outstanding nucleofuge.[17] Such species have the ability to emulate the carbene/carbocation behavior of a monovalent cationic carbyne (:+C–R) and provide a transient cyclopropyl-I(III) intermediate upon alkene cyclopropanation. Recently, we questioned whether we could exploit our Rh-catalyzed carbyne transfer platform for the discovery and development of the first one-step and catalytic synthesis of cyclopropenium cations (Figure ). We hypothesized that the catalytically generated Rh-carbynoid (1) could intercept alkynes and provide cyclopropenyl-I(III) intermediates () that, upon an ionization process that occurs with the departure of the I(III) leaving group, would lead to a new class of CPC substituted with an ester group.[18]

Figure 2

Mechanistic hypothesis.

With this mechanistic proposal in mind, we evaluated the feasibility of this idea using 1-phenyl-1-propyne, a broad range of commercial Rh2 catalysts, and diazomethyl-based hypervalent iodine reagents 2 (see Supporting Information for full optimization studies). Taking into account the possible instability and reactivity of the desired CPC, we employed 1,3,5-trimethoxybenzene as external nucleophile to quantify the efficiency of the transformation. To our delight, after extensive optimization studies, we were able to find that pseudocyclic reagent 2a and the Du Bois catalyst[19] Rh2(esp)2 (1 mol %) led to cyclopropene 3a* with excellent efficiency (94% yield, >20:1 r.r.) (Table ). The only regioisomer observed comes from the selective attack of the nucleophile to the cyclopropenium carbon atom substituted with the ester group.

Table 1

Optimization Studiesa,b

entry	deviation of reaction conditions	yieldb3a*, %
1	Rh2(OPiv)4 instead of Rh2(esp)2	93
2	Rh2(OAc)4 instead of Rh2(esp)2	4
3	Rh2(TFA)4 instead of Rh2(esp)2	0
4	2b instead of 2a	50
5	2c,d instead of 2a	0
6	2d used with Zn(OTf)2 instead of 2a	60c
7	reaction carried out at –50 °C	81

Reactions performed with alkyne 1a (0.2 mmol, 2.0 equiv), reagents 2 (0.1 mmol, 1 equiv), Rh catalyst (1 mol %), and 1,3,5-trimethoxybenzene (4 equiv) in CH2Cl2 (0.067 M).

Yields are reported on the basis of 1H NMR analysis using CH2Br2 as the internal standard.

0.5 equiv of Zn(OTf)2 was used.

During the optimization of the reaction, we observed that other sterically demanding catalysts worked well (Table , entry 1); however, Rh2(OAc)4 or the more electrophilic Rh2(TFA)4 was not competent (entries 2, 3). The nature of the hypervalent iodine reagent played also a key role in the efficiency of the transformation. The use of triflate as counterion provided a moderate yield of 3a* (entry 4). However, although full consumption of linear reagent 2c was observed, no conversion to the desired product was detected (entry 5). On the other hand, cyclic reagent 2d was inert to catalytic diazo activation with Rh2(esp)2 (entry 5). Only upon addition of Zn(OTf)2 as the additive, a well-known activator of cyclic I(III) reagents,[14] was a moderate yield of 3a* observed (entry 6). We also found that higher reaction temperatures could give product 3a* but with lower efficiency (entry 7).

During the optimization process, we observed that the addition of 2a to 1a and Rh2(esp)2 at −60 °C provided the formation of a slurry, which quickly turned into a clear solution after the addition of 1,3,5-trimethoxybenzene. We believed that this observation suggested the potential formation of either intermediate or CPC 3a. Initial experiments performed to isolate the intermediate at room temperature were unsuccessful; however, slow addition of dry hexane at −60 °C and quick filtration provided a white solid, which was subjected to spectroscopic analysis using mono- and bidimensional nuclear magnetic resonance (1H, 13C, 19F, 31P, 1H–13C NMR, CD3NO2 as solvent), IR, and HRMS. The three deshielded signals at 171.5, 166.7, and 156.0 ppm observed in the 13C NMR spectra clearly referred to the cyclopropenium carbon atoms, thus confirming the formation of 3a (86% yield) (Table ). 3a could be stored at −20 °C under argon for at least 1 month without detectable decomposition and can be handled at room temperature without the need of a glovebox.

Table 2

Scope of the Catalytic Synthesis of Cyclopropenium Cations 3a

Performed with alkyne 1 (0.26 mmol, 1.3 equiv), hypervalent iodine reagents 2 (0.2 mmol, 1.0 equiv), CH2Cl2, −60 °C, 50 min, 1,3,5-trimethoxybenzene (0.8 mmol, 4.0 equiv) was added for the synthesis of 3*.

Yield of the reaction to give 1.08 g of 3a and 1.10 g of 3ad.

1.5 mol % catalyst.

2 mol % catalyst.

Reaction carried out at −63 °C.

1.5 equiv of alkyne.

2.0 equiv of alkyne.

Encouraged by the results, we next aimed to develop the scope of our catalytic protocol for the synthesis of CPCs 3 by examining a wide range of alkynes (Table ). In case the CPC was difficult to isolate, it was transformed to the corresponding cyclopropene 3* with 1,3,5-trimethoxybenzene. We were delighted to observe that our process worked well for alkynes with aryl rings substituted with synthetically useful functionalities such as halogens (3b–e, 3l–n), acetoxy (3f), phenyl (3g), CF3 (3h), ester (3i), alkene (3j), and methyl (3k) in para, meta, and ortho positions as well as naphthalene (3o,p) and heterocycle cores (3q*).

Alternative primary alkyl groups such as ethyl (3r), sec-butyl (3s), chloroalkyls (3t*, 3u), benzyl (3v*), and alkyl groups functionalized with protected amines (3w*) or alcohols (3x*, 3y) were well tolerated. Secondary alkyl substituents like isopropyl required an increment on the catalyst loading (2 mol %) and alkyne (1.5 equiv) (3z); however, cyclic derivatives such as cyclopropyl (3aa), cyclobutyl (3ab), and cyclohexyl (3ac*) worked well under the standard conditions. Alkynes substituted with tertiary groups like tert-butyl or trimethylsilyl provided traces of the desired CPCs. On the other hand, while alkynes substituted with two alkyl groups were well tolerated (3ad–ai*), diphenylacetylene was unreactive and terminal alkynes such as phenylacetylene provided mixtures of products difficult to identify.[20] We also demonstrated that alternative ester substituents at the hypervalent iodine reagents were possible (3aj–al, 3am*). Finally, it is worth highlighting that CPC 3a and 3ad were prepared in >1 g without compromising the efficiency of the process.

We next turned our attention to evaluate the reactivity of our cyclopropenium cations with a broad range of carbon and heteroatomic nucleophiles aiming to develop a novel synthetic route to cyclopropenes (Table ). These highly strained, three-membered unsaturated carbocycles are well known for their unique potential as versatile building blocks in organic synthesis that can undergo nucleophilic or electrophilic additions, substitutions, rearrangements, cycloadditions, and ring-opening reactions, delivering pharmaceutical-relevant scaffolds like cyclopropanes or complex structures,[21] but also as biorthogonal reagents for chemical biology applications.[22]

Table 3

Synthesis of Tri- and Tetrasubstituted Cyclopropenes by Nucleophilic Attacka

Performed with 3 (0.1 mmol), carbon or heteroatomic nucleophiles (0.15–0.2 mmol), CH2Cl2, 0 °C, 2–15 min.

Cs2CO3 (0.1 mmol) was added as base. Yields are reported on the basis of isolated pure product.

Considering that the nucleophilic attack of 1,3,5-trimethoxybenzene to 3a proceeds with outstanding regioselectivity, we hoped that alternative nucleophiles could behave analogously. We were delighted to observe that a variety of commercial or readily available aryl, alkyl, vinyl, and alkynyl Grignard reagents provided instantaneous access to cyclopropenes 4–19 as single regioisomers and in high efficiency with simple reaction conditions (CH2Cl2, 0 °C, 2–15 min). Moreover, alternative nucleophiles such as boronic acid (20), organozinc (21), organosilicon (22, 23), carbonyl (24, 25) and isocyanide nucleophiles (26) performed well, and in many cases, no chromatographic column was needed to obtain the corresponding cyclopropene product.

It is worth highlighting that our methodology with cyclopropenium cations provides a new approach to the synthesis of complex cyclopropenes[23] and solutions to challenges observed in metal-catalyzed carbene transfer with diazo acetates to internal alkynes. Catalytically generated metal-carbenes substituted with alkyl or allyl groups undergo faster β-hydride migration[24] or intramolecular cyclopropanation,[25] respectively, before the intermolecular cyclopropenation of the internal alkyne takes place.

The remarkable promiscuity observed of our CPCs to react with carbon nucleophiles encouraged us to question whether heteroatomic nucleophiles could work. If successful, such reactions would provide access to a type of tetrasubstituted cyclopropenes not possible to make by any reaction currently available, because of the lack of heteroatom-substituted diazoacetates or alternative carbene sources.[26] We were delighted to observe that a selection of commercial nitrogen, oxygen, phosphorus, and sulfur nucleophiles provided cyclopropenes 27–37 with high efficiency. In addition, a reaction carried out with tetrabutylammonium borohydride provided trisubstituted cyclopropene derivative 38 by the regioselective hydride attack to 3a.[27,28]

Notably, all kinds of heteroatomic nucleophiles underwent regioselective attack to the cyclopropenium carbon atom substituted with the ester group. The reactions provide access to a plethora of novel cyclopropene derivatives with unknown reactivity, which promise applications in reaction discovery and in the construction of complex skeletons.[21]

Finally, in order to provide an explanation of the outstanding and intriguing regiocontrol observed in the nucleophilic addition to CPCs 3, we calculated the geometry optimization and LUMO map using SPARTAN 20 at the ωb97xd/6-31G(d) level. In Figure , it can clearly be appreciated that the carbon atom substituted with the ester group has the highest LUMO coefficient among the cyclopropenium carbon sites, thus suggesting that nucleophilic attack occurs under orbital control.

Figure 3

LUMO map of 3a (method: ωB97X-D, basis set: 6-31G(D)).

In summary, we have discovered and developed the first catalytic synthesis of cyclopropenium cations based on the formal transfer of monovalent cationic carbynes (:+C–R) to readily available alkynes from a catalytically generated Rh-carbynoid. This type of group transfer process is uncommon for metal-carbyne complexes, which mainly evolve via [2+2] cycloadditions with alkynes.[29,10] Our process accesses previously unknown ester-substituted CPCs, which can be handled outside of a glovebox, from a broad range of internal aryl- and alkyl-substituted alkynes. The synthetic utility of our CPCs has been demonstrated by the regioselective attack of a broad range of carbon and heteroatomic nucleophiles that provided valuable cyclopropenes. Several of those compounds cannot be made by current approaches due to the lack of appropriate diazoacetate reagents as carbene sources or limitations in current methodologies. Current studies are focused on exploiting novel cyclopropenium reactivity for asymmetric synthesis and other applications.[30]