Zachary P.
Sercel
and
Ilan
Marek
*
Schulich Faculty of Chemistry and the Resnick Sustainability Center for Catalysis, Technion – Israel Institute of Technology, Technion City, Haifa, 3200009, Israel. E-mail: chilanm@technion.ac.il
First published on 30th May 2025
We report a method for the direct palladium-catalyzed cross coupling reactions of cyclopropenyl esters bearing a variety of substitution patterns with Csp2 iodides. This reaction is largely insensitive to the electronic nature of the coupling partner. Tetramethylammonium acetate, a halide sequestrant, was exceptionally effective as an organic base. An observed KIE of 2.5 revealed C–H bond cleavage to be involved in the turnover-limiting step. This method enables the rapid assembly of cyclopropenes whose preparation previously required the use of toxic tin and arsenic reagents.
The cyclopropene ring, with ∼54 kcal mol−1 of strain energy,6 is primed to undergo numerous chemical reactions involving π bond breakage. Outside of synthetic methods developed by our group7 and others,8,9 cyclopropenes are commonly used as intermediates in total synthesis,10 monomers in polymer chemistry,11 and precursors to vinylcarbene intermediates.12 In nature, cyclopropenes occur in certain toxic fatty acids.13
A direct Pd-catalyzed cross coupling of cyclopropenyl esters was reported by Gevorgyan in 2005 (Scheme 1a).14 This excellent method enables the arylation of trisubstituted cyclopropenes bearing aryl substituents in moderate to high yields. However, an alkyl-substituted substrate underwent the reaction in only 46% yield, and the use of starting materials bearing fewer than three ring substituents was not reported. A series of mechanistic experiments indicated that the reaction likely proceeds via coordination of the cyclopropene to a palladium species to form complex 1, which bears carbocationic character, and a fast deprotonation leads to a cyclopropenylpalladium species that can go on to afford the cross-coupling product. Substrates bearing alkyl substituents (R2), which do not stabilize positive charge as effectively as aryl substituents, would therefore not undergo the reaction as efficiently.
In 2019, Hashmi reported a Ag/Au-cocatalyzed alkynylation of cyclopropenes which proceeds via Ag-mediated C–H activation (Scheme 1b).15 A similar catalytic system was recently used by Waser to synthesize cyclopropene heterodimers.16 These powerful methodologies were used to synthesize a wide array of privileged products, but gem-diester-containing cyclopropenes are required for high yields, and electron-withdrawing olefinic substituents are not tolerated. Cyclopropene alkynylation, as well as allylation, were also demonstrated by Xie using a dinuclear gold catalyst.17
If cyclopropenes are considered to be analogous to alkynes (based on hybridization18 and reactivity), utilizing Ag-mediated C–H activation to activate cyclopropenes for Pd-catalyzed cross coupling would be analogous to Ag-promoted formal Sonogashira couplings. Mori reported that reactions resembling the Sonogashira coupling can be promoted in high yield by either Ag2O or ionic bases like TBAF (Scheme 1c),19 and Pale reported similar reactions with cocatalytic silver halides.20
Based on this precedent, we have succeeded in developing a general, direct cross-coupling of di- and trisubstituted cyclopropenyl esters with aryl and styrenyl iodides (Scheme 1d).
Disubstituted cyclopropenyl ester 2a, prepared by Rh-catalyzed cyclopropenation,21 was identified as an ideal model substrate due to challenges previously encountered with the derivatization of this substrate by stannylation and Stille coupling. We first subjected 2a to Gevorgyan's direct cross coupling conditions with PhI14 but observed only a trace of the desired product 3a with nonspecific decomposition of the starting material (Scheme 2).
We then set out to identify reaction conditions that would effect the cross coupling of cyclopropene 2a with PhI (Table 1). Based on the hypothesis that silver-mediated C–H activation could enable a Sonogashira-type mechanism, the AgI-cocatalyzed conditions demonstrated by Pale for alkyne coupling were tested (entries 1 and 2), but desired product 3a was not detected.20 However, employing stoichiometric AgOAc as a base afforded 39% of the desired product (entry 3). Switching to Ag2O-mediated conditions based on those reported by Mori led to a high yield of 3a in just 2 hours at 70 °C (entry 4). The reaction proceeded slowly at ambient temperature (entry 5). Based on the precedent of Mori, we envisioned that a base-mediated reaction should be possible and would obviate the use of (super)stoichiometric silver. While several bases tested did promote formation of 3a (entries 6–8), these reactions were sluggish and proceeded with poor mass recovery. We therefore hypothesized that Ag is necessary to promote cross coupling by iodide abstraction from a Pd intermediate.22 An alternative halide sequestrant could obviate the use of stoichiometric silver. While most candidates (i.e. Pb and Tl salts) present a safety and environmental hazard, Larrosa demonstrated the use of tetramethylammonium acetate as an iodide sequestrant in Pd-catalyzed C–H activation,23 and a related application has since been explored on a process scale.24
Entry | Catalyst loading | Base | Additive | Solvent | Temperature | Time | % SM | % Yield |
---|---|---|---|---|---|---|---|---|
a Reactions were conducted on a 0.2–0.3 mmol scale and yields were determined by 1H NMR analysis using 1,3,5-trimethoxybenzene as an internal standard. | ||||||||
1 | 14.5 mol% | DIPEA | AgI (29 mol%) | DMF | 23 °C | 17 h | n.d. | — |
2 | 14.5 mol% | DIPEA | AgI (29 mol%) | THF | 70 °C | 17 h | n.d. | — |
3 | 7.2 mol% | AgOAc | — | THF | 70 °C | 18 h | — | 39% |
4 | 7.2 mol% | Ag2O | — | THF | 70 °C | 2 h | — | 84% |
5 | 7.2 mol% | Ag2O | — | THF | 23 °C | 18 h | 51% | 28% |
6 | 7.2 mol% | TBAF | — | THF | 70 °C | 21 h | 42% | 22% |
7 | 7.2 mol% | KOPh | — | THF | 70 °C | 16 h | 20% | 33% |
8 | 7.2 mol% | CsOAc | — | THF | 70 °C | 17 h | 23% | 20% |
9 | 7.2 mol% | Me4NOAc | AgI (29 mol%) | THF | 70 °C | 16 h | 19% | 68% |
10 | 5.7 mol% | Me4NOAc | AgI (29 mol%) | EtOAc | 90 °C | 17 h | — | 85% |
11 | 5.7 mol% | TBAOAc | AgI (29 mol%) | EtOAc | 90 °C | 17 h | — | 51% |
12 | 5.7 mol% | Me 4 NOAc | — | EtOAc | 90 °C | 19 h | — | 86% |
13 | 1.1 mol% | Me4NOAc | — | EtOAc | 90 °C | 15 h | — | 77% |
Conducting the arylation of 2a with Me4NOAc as the base in the presence of AgI led to a promising level of conversion and high mass recovery (entry 9). In the higher boiling and green solvent EtOAc, reaction at 90 °C afforded complete conversion of 2a and 85% yield in 17 h (entry 10). Me4NOAc was confirmed to be uniquely effective (entry 11), and furthermore, AgI does not play a catalytic role (entry 12). Indeed, a control experiment (see ESI†) revealed that Ag2O does not independently effect C–H activation of cyclopropene 2a. The reaction proceeded with a decreased catalyst loading, albeit in diminished yield (entry 13).
However, upon encountering reproducibility problems at lower catalyst loadings, we opted to explore the substrate scope using 5% Pd.
A variety of aryl iodides were suitable for cross-coupling with cyclopropene 2a in moderate to high yields (Scheme 3). Phenyl product 3a was prepared on a 1.9 g (10 mmol) scale without a significant drop in yield. Products bearing electron-donating (3b, 3i) and electron-withdrawing substituents (3c, 3h) could be furnished, and an aryl bromide (3d) did not interfere with the reaction, demonstrating high chemoselectivity toward aryl iodides. 2-Iodothiophene and an iodoindole were also suitable coupling partners (3e, 3f). Notably, ortho-iodoanisole provided product 3g in high yield despite potential steric clash.
Cyclopropene starting materials bearing a variety of substitution patterns were competent substrates in this reaction. An enantioenriched trisubstituted cyclopropene prepared by the method of Davies afforded optically active product 3j.25 A p-bromophenyl group was tolerated in a similar substrate (3k). Remarkably, cyclopropenes bearing gem-diester moieties (3l, 3m) underwent the cross-coupling smoothly without ring fragmentation.26 Similarly, fully substituted cyclopropenes with methyl-substituted quaternary centers (3n, 3o) were obtained in high yields. The reaction is also relatively insensitive to the nature of the olefinic substituent. A tethered silyloxy group (3p) and alkyl chloride (3q) were tolerated, albeit with some loss of yield to side reactions. Secondary and tertiary alkyl-substituted products were formed in moderate to high yields (3r, 3s).
This method was also applied to the alkenylation of cyclopropenes. However, in this case, milder reaction conditions were necessary due to product instability. Ag2O-mediated conditions derived from those evaluated during the initial reaction optimization (Table 1, entry 4) provided diene products 3t and 3u in 77% and 71% yields, respectively. For substrates and electrophiles that did not undergo the desired reaction, see the ESI.†
To interrogate the mechanism of this cross-coupling, the deuterium kinetic isotope effect for the reaction was measured by preparing deuterated cyclopropene 2a-D and conducting two reactions in parallel (Scheme 4). A primary KIE (kH/kD = 2.5) was observed, indicating that C–H bond cleavage occurs in the turnover-limiting step of the reaction.27 This is in contrast to the related reaction studied by Gevorgyan,14 which exhibited a kH/kD value of 1.0.
On the basis of this result and literature precedent, we propose a catalytic cycle for the reported reaction (Scheme 5). An anionic Pd(0) species 4 would engage in oxidative addition with aryl iodide to provide Pd(II) acetate complex 5, following iodide sequestration. The latter complex could coordinate to a cyclopropene 2, leading to complex 6. Then, turnover-limiting deprotonation would afford cyclopropenyl complex 7, which could undergo reductive elimination to release product 3 and complete the catalytic cycle (Scheme 6).
Solutions of Pd(PPh3)4 containing acetate have been shown to transiently form anionic complex 4, which is responsible for oxidative addition.28 Amatore found that ligand substitution of iodide with acetate is necessary for PhPdI(PPh3)2 to react with styrene via migratory insertion. While in the present case migratory insertion does not occur, as evidenced by the lack of formation of Heck-type methylenecyclopropane products when R1 = alkyl, iodide sequestration is likely necessary to favor formation of acetate complex 5 and enable coordination of 2 to the palladium center.
The subsequent step, deprotonation of the substrate and formation of cyclopropenyl palladium species 7, is turnover-limiting as evidenced by the observed KIE. While we cannot definitively rule out an intramolecular concerted metallation-deprotonation (CMD) mechanism involving a palladium-coordinated acetate ligand, the competence of fluoride and phenoxide bases (see Table 1, entries 6, 7) and the lack of a significant rate acceleration in the presence of acetate suggest an intermolecular deprotonation event akin to the step proposed by Gevorgyan14 or similar processes in Cu-free Sonogashira-type couplings.29
The products generated by the method reported herein possess synthetic utility. For example, hydrogenation of cyclopropene 3g over Pd/C afforded all-cis cyclopropane 8. Cu-catalyzed diastereoselective carbometallation2c of cyclopropene 3n yielded pentasubstituted cyclopropane 9 with excellent diastereoselectivity. Anisole derivative 3h was recently advanced to cyclobutane 10via the formation of a bicyclo[1.1.0]butane intermediate.2b
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5sc03096j |
This journal is © The Royal Society of Chemistry 2025 |