Maria
Talavera
*ab,
Soodeh
Mollasalehi
b and
Thomas
Braun
*b
aFacultad de Química, Universidade de Vigo, Campus Universitario, 36310 Vigo, Spain. E-mail: matalaveran@uvigo.gal
bDepartment of Chemistry, Humboldt Universität zu Berlin, Brook-Taylor Straße 2, 12489 Berlin, Germany. E-mail: thomas.braun@cms.hu-berlin.de
First published on 2nd May 2024
The reaction of [Rh{(E)-CFCHCF3}(PEt3)3] with Zn(CH3)2 results in the methylation of the alkenyl ligand to give [Rh{(E/Z)-C(CH3)CHCF3}(PEt3)3]. Variable temperature NMR studies allowed the identification of a heterobinuclear rhodium–zinc complex as an intermediate, for which the structure [Rh(CH3)(ZnCH3){(Z)-C(CH3)CHCF3}(PEt3)2] is proposed. Based on these stoichiometric reactions, unique Negishi-type catalytic cross-coupling reactions of fluorinated propenes by consecutive C–H and C–F bond activation steps at room temperature were developed. The C–H bond activation steps provide a fluorinated ligand at Rh and deliver the fluorinated product, whereas the C–F bond activation and C–C coupling occur via outer-sphere nucleophilic attack at the fluorinated alkenyl ligand.
A possible functionalization pathway would consist of the formation of C–C bonds by cross-coupling reactions. Such conversions have been investigated for chlorinated or brominated olefin derivatives.14,15 For fluoroalkenes, C–C couplings have been less studied and examples include group 10 catalysts and the use of gem-difluoroalkenes.16–28 The couplings proceed by an initial C–F bond oxidative addition – in some cases using lithium salts to promote the activation step – or insertion of a fluorinated olefin into a metal–carbon bond followed by β-fluorine elimination. Cao and Wu developed a Suzuki–Miyaura cross coupling arylation of gem-difluoroalkenes using a nickel catalyst,21 whereas Tsui and Liu applied a palladium catalyst.16 Xu et al. published the arylation of an in situ formed difluorovinyl ketone in the presence of a palladium catalyst and aryl boronic acids.19 Regarding rhodium catalyzed reactions, Xia and co-workers described the formation of monofluorinated dienes from gem-difluorocyclopropanes, which are employed as fluoroallyl surrogates, by C–C coupling with allylboronates.29 Negishi-type cross-coupling reactions based on vinyl C(sp2)–F bond oxidative addition steps are rare. Saeki et al. functionalized the olefin CF2CH(1-naphtyl) with arylzinc derivative MeC6H4ZnCl in the presence of a palladium catalyst.30 In addition, examples using Pd or Ni catalysts and lithium salts to promote the C–F bond activation have been described by Zhang et al. and Ogoshi and co-workers developing Negishi-type C(sp2)–F bond alkylations of 2,3,3,3-tetrafluoropropene or tetrafluoroethylene, respectively.31,32
Recently, the reactivity of Z-1,3,3,3-tetrafluoropropene (5a) towards rhodium(I) complexes has been studied and it was shown that rhodium alkenyl complexes with a low fluoroorganyl content such as [Rh{(E)-CHCHCF3}(PEt3)3] can activate higher fluorinated olefins (Scheme 1). For the latter, in a C–H activation step [Rh{(E)-CFCHCF3}(PEt3)3] (1) and 3,3,3-trifluoropropene were produced.33
Scheme 1 Reaction of [Rh{(E)-CHCHCF3}(PEt3)3] with Z-1,3,3,3-tetrafluoropropene.33 |
Herein studies on catalytic C–C coupling reactions at fluorinated alkenyl ligands are described. Stoichiometric model reactions give an insight into a possible reaction mechanism. Unprecedented rhodium catalyzed Negishi-type cross-coupling reactions of fluorinated olefins have been developed, based on both C(sp2)–F and C(sp2)–H bond activation and, remarkably, involving a C–C coupling step in the outer coordination sphere.
In an alternative approach in which complex 1 was treated with LiCH3 instead of Zn(CH3)2, a mixture of complexes (Z/E)-3 and 4 was also obtained. However, in this case a heterobinuclear intermediate similar to 2 was not observed. Note that Rh–Zn bimetallic complexes have been described before.35–39
The formation of 2 followed by the generation of E/Z-3 from complex 1 can be considered as a case for a rare outer-sphere reaction where a nucleophile attacks the fluorinated ligand bonded at rhodium. Note that the attack of a fluorosilicate at the carbon atom of a fluorinated pyridyl ligand has been proposed at rhodium.40 The attack of PEt3 at the β-carbon atom of a perfluorovinyl ligand at nickel was also described.41 In addition, outer-sphere electrophilic fluorination was reported by Lynam, Slattery and co-workers.42,43 The rearrangements of the alkenyl ligands for the conversion of 2 into E-3, 1 into 2 and E-3 into 6a comprise an isomerization at the double bond. This might involve an intermediate vinylidene complex that can be formed by a reversible migration of a methyl group from the alpha carbon to the metal center. Intermediate metallacyclopropene-like species have also been discussed for such rearrangements.44–46 In addition, a negative hyperconjugation of the π-electron density into antibonding orbitals at the CF3 group might weaken the CC double bond and allow for an isomerization of the alkenyl ligands. Note that Ojima et al. proposed a zwitterionic carbene–rhodium complex as an intermediate for such a cis/trans rearrangement.47 However, another alternative route for the isomerization to give complexes 2 or E-3 could imply a rearrangement during the nucleophilic substitution. After addition of CH3− to Cα, the β-carbanion formed could show a free rotation about Cα–Cβ. A Lewis-acidic cation of the type [ZnMe]+ would then abstract the fluoride forming the oligomeric [ZnFMe]n.
The suggested structure of complex 2 is supported by NMR spectroscopic data, in part based on the data for the 13C labeled derivative [Rh(13CH3)(Zn13CH3){(Z)-C(13CH3)CHCF3}(PEt3)2] (2′) (see ESI‡). Thus, in the 1H NMR spectrum of complex 2′ a resonance for the rhodium-bound methyl ligand is observed at −0.42 ppm as a doublet of triplets with coupling constants of 121.1 and 4.6 Hz due to the coupling to the carbon atom and the two phosphine ligands. In addition, a resonance at −0.21 ppm is assigned to the methyl group at the Zn atom. It appears as a doublet coupled with 1JH–C = 123.2 Hz. The 13C{1H} NMR displays two resonances at −27.5 and −9.2 ppm for the rhodium- and zinc-bonded methyl moieties, respectively. The former exhibits a doublet of pseudo quartets as a result of the coupling to rhodium (9.3 Hz), the two phosphorus atoms and the carbon atom bonded to zinc with a coupling constant of 5.2 Hz. The latter signal also shows the 5.2 Hz carbon–carbon coupling together with the extra coupling to rhodium resulting in a pseudo triplet. In addition, the geometry of complex 2 was optimized by DFT calculations using toluene as a solvent (BP86/def2-SVP, see ESI‡). The complex exhibits a tetragonal pyramidal structure at rhodium in which the methylzinc group is at the apical position. The distance between the Zn nucleus and one of the fluorine atoms of the CF3 moiety of 2.747 Å is shorter than the sum of van der Waals radii which might be the reason for the favored cis arrangement of the Zn center and the CF3 group at the moiety containing the double bond. In fact, this product is 7.4 kJ mol−1 more stable than the corresponding trans isomer.
Based on the stoichiometric functionalization of the fluorinated alkenyl ligand at complex 1 to yield 6a, a process for the catalytic methylation of fluoroolefins was developed. The reaction of Z-1,3,3,3-tetrafluoropropene (5a) with Zn(CH3)2 in THF-d8 in the presence of 10 mol% [Rh{(E)-CFCHCF3}(PEt3)3] (1) as a catalyst gave full conversion of the olefin into the C–C coupling product E-1,1,1-trifluorobut-2-ene (6a) (Scheme 3, Table 1, entry 1).
Entry | Catalyst | Methyl source | Conversiona (%) |
---|---|---|---|
a Conversion based on the consumption of Zn(CH3)2 and determined by 1H NMR spectroscopy. b 1 h reaction time, 5 mol% of complex 7. | |||
1 | Complex 1 | Zn(CH3)2 | 99 |
2b | Complex 7 | Zn(CH3)2 | 99 |
3 | [Rh(F)(PEt3)3] | Zn(CH3)2 | 70 |
4 | [Rh(H)(PEt3)3] | Zn(CH3)2 | 70 |
5 | Complex 7 | MgBrCH3 | — |
6 | Complex 7 | LiCH3 | — |
7 | Complex 7 | Zn(CH3)Cl | 15 |
Complex [Rh{(E)-CFCHCF3}(PEt3)3] (1) is synthesized by C–H bond activation of 5a using [Rh(CH3)(PEt3)3] (7).33 Therefore, complex 7 was also tested as a pre-catalyst and full conversion to compound 6a was obtained after 1 h using only 5 mol% of catalyst (Scheme 3, Table 1, entry 2). Remarkably the transformation occurs at room temperature within 1 h. Note that Zn(CH3)2 does not react with the olefin in the absence of a rhodium catalyst.
In general, there are few examples of rhodium-mediated Negishi-type cross-coupling reactions described in the literature. For these cases, C–I bond activation functionalization steps do not proceed through an outer-sphere reaction; instead they are proposed to occur via common oxidative addition/transmetalation cycles. Thus, Takagi et al. described alkyl-aryl C–C coupling reactions using [RhCl(cyclooctadiene)]2 with bidentate phosphines48 and work by Ozerov and co-workers using a PNP pincer rhodium complex as a precatalyst afforded a 7% yield in an aryl–aryl coupling.39 It might additionally be worth noting that examples on the functionalization of fluoroaromatics by cross-coupling reactions involving C–F bond cleavage have been reported, for instance Negishi-type conversions at Ni and Pd by the research groups of Love, Ogoshi and Radius.13,49–64
Other rhodium complexes as catalysts for the derivatization of 5a were then studied using Zn(CH3)2 as a nucleophile. Accordingly, Zn(CH3)2 methylates the Rh–F bond of [Rh(F)(PEt3)3] and in the presence of Z-1,3,3,3-tetrafluoropropene (5a), the generation of E-1,1,1-trifluorobut-2-ene (6a) with 70% conversion was achieved (Table 1, entry 3). A comparable outcome was obtained when [Rh(H)(PEt3)3] was tested as a catalyst (Table 1, entry 4). Note that the rhodium hydrido complex is known to activate preferentially vinylic C–F bonds followed by C–H bond activation of another equivalent of olefin.65–68
As an alternative to Zn(CH3)2, BrMgCH3 and LiCH3 were applied for the catalytic methylation of 5a, but the formation of the methylated derivative was not observed within one day (Table 1, entries 5 and 6). Note that an oligomerization of the nucleophilic reagents in solution can attenuate the nucleophilicity of their methyl anion. On the other hand, mono organylzinc chloride derivatives can also be used for cross-coupling reactions. Indeed, chloromethylzinc could be employed to yield 6a, but only with 15% conversion (Table 1, entry 7).
Further functionalization of Z-1,3,3,3-tetrafluoropropene (5a) was attempted using ZnEt2 and ZnPh2 as nucleophile sources. Indeed, reaction of 5a with ZnPh2 in the presence of complex 7 as a catalyst gave, within 1 d, E-3-phenyl-1,1,1-trifluoropropene (8a) with 90% conversion (Scheme 4), but after one week at room temperature, full conversion was achieved. Interestingly, ZnEt2 provided a different outcome. The reaction of 5a with ZnEt2 using complex 7 as a catalyst gave, after 1 day a mixture of ethylene, E-1,1,1-trifluoropent-2-ene (9a) and 3,3,3-trifluoropropene in a 5:1:2.3 ratio with 60% conversion (Scheme 4).
The formation of 9a should take place through a comparable mechanism as for the generation of 6a, leading to a Rh–Zn bimetallic intermediate which would bear an ethyl group at the olefin and an ethyl ligand bound to rhodium. This intermediate can produce ethene together with [Rh(H)(PEt3)3] by β-H elimination. The rhodium hydrido complex would then react with 5a to give 3,3,3-trifluoropropene and complex [Rh{(E)-CFCHCF3}(PEt3)3] (1) as described before in the literature.33
The scope of the rhodium defluorinative methylation was then investigated using other olefinic substrates and complex [Rh(CH3)(PEt3)3] (7) as a catalyst (Table 2, see also the ESI‡ for further optimizations). Thus, E-1,3,3,3-tetrafluoropropene (5b) was converted into compound 6a with 95% conversion. Higher fluorinated derivatives were also attempted and both Z-1,2,3,3,3-pentafluoropropene (5c) and perfluoropropene (5d) were used. In the first case Z-1,1,1,2-tetrafluorobut-2-ene (6c) was identified in 25% conversion as the main product, while in the second case a mixture of Z/E-1,2,3,3,3-pentafluorobut-2-ene isomers (6d) in a 1:0.9 ratio was formed (10% conversion). The latter stoichiometric conversion and decrease in reactivity for the perfluorinated substrate could be due to the preference of complex 7 for the activation of C–H over C–F bonds.67 Then, 2,3,3,3-tetrafluoropropene (HFO-1234yf, 5e) was used as a substrate, but only 11% conversion was achieved to form 6c and unidentified compounds. Finally, to investigate the influence of a CF3 group on the catalytic methylation of fluoroolefins, 1,1-difluoroethane (5f) and trifluoroethane (5g) were employed as reagents. While 5f yielded only traces of the methylated product 2-fluoropropene (6f) even at 60 °C, compound 5g provided a mixture of E-1,2-difluoropropene (6g) and an unknown compound in a 5:1 ratio with 17% conversion. These results also suggest a preference of the rhodium system for olefins with a geminal CHF group facilitating the initial C–H bond activation.
Footnotes |
† Dedicated to Prof. Dr Helmut Werner on the occasion of his 90th birthday. |
‡ Electronic supplementary information (ESI) available: Synthesis and analytics of all compounds and details of the DFT calculations. See DOI: https://doi.org/10.1039/d4sc00951g |
This journal is © The Royal Society of Chemistry 2024 |