A Next‐Generation Air‐Stable Palladium(I) Dimer Enables Olefin Migration and Selective C−C Coupling in Air

Abstract We report a new air‐stable PdI dimer, [Pd(μ‐I)(PCy2 tBu)]2, which triggers E‐selective olefin migration to enamides and styrene derivatives in the presence of multiple functional groups and with complete tolerance of air. The same dimer also triggers extremely rapid C−C coupling (alkylation and arylation) at room temperature in a modular and triply selective fashion of aromatic C−Br, C−OTf/OFs, and C−Cl bonds in poly(pseudo)halogenated arenes, displaying superior activity over previous PdI dimer generations for substrates that bear substituents ortho to C−OTf.

Metal catalyzed cross-coupling reactions to forge CÀC bonds [1] as well as olefin migrations [2,3] for the construction of stereochemically defined double bonds are central strategies in the synthesis of key building blocks, pharmaceuticals, materials as well as of societal and industrial relevance, for example, in the production of commodity chemicals, such as gasoline, nylon, detergents, cosmetics, fragrances or food additives. [4] A central challenge in this context is the control of selectivity: whereas the site-selective C À C bond formation of poly(pseudo)halogenated arenes is a powerful strategy to densely functionalized arenes, [5] the positional as well as geometrical control (E vs. Z) of an olefin is key for its properties as well as follow up transformations. [2,6] Pd I dimers have emerged as especially powerful catalysts for these challenges. [7] Their catalytic role ranges from engaging in dinuclear Pd I catalysis [8] to being precursors for Pd 0 in CÀC and C-heteroatom bond formations [7b, 9] or for Pd II -H in olefin isomerizations. [10] Ultimately, the properties, catalytic role and efficiency are dictated by the nature of the ligand and the bridging unit in the dinuclear Pd I entity. For example, the bromide-bridged [Pd(m-Br)(P t Bu 3 )] 2 dimer D1 [9g,h, 11] (Figure 1) is highly labile and as such readily transformed to Pd 0 by weak nucleophiles. [12] It is also highly sensitive to oxygen or coordinating solvents, [7b, 13] and there-fore does not allow cross-coupling in air. A small modification, that is, swap of bromide to iodide in the bridge, makes the Pd I dimer highly robust: iodide-bridged dimer D2, that is, [Pd(m-I)(P t Bu 3 )] 2 , is a bench-stable species and no longer an efficient pre-catalyst to Pd 0 ; it needs strong nucleophiles for its conversion to Pd 0 . [12] However, nucleophiles which also function as an effective bridge in the dinuclear entity, can be incorporated into arenes via an exchange under direct dinuclear Pd I catalysis. [8] Moreover, D2 triggers extremely rapid and fully selective, sequential CÀC bond formations (both arylation and alkylation) of aromatic CÀBr, CÀOTf/ OFs and CÀCl bonds under open flask conditions ( Figure 1). [14] Conversely, owing to its robustness, dimer D2 is ineffective in olefin isomerization (see below), which requires an initial intramolecular palladation [10] with the ligand for Pd II -H generation. The more labile and air-sensitive bromide-bridged Pd I species D1 is therefore to date the most effective Pd I dimer precursor for olefin isomerizations, as demonstrated by Gooßen. [10] A less precious Ni I dimer was recently shown to also engage in selective olefin migration, albeit under Ni-radical rather than Ni-H reactivity. [6] Skrydstrup demonstrated that the in situ generated [(P t Bu 3 ) 2 Pd II (H)(Cl)] also triggers effective olefin isomerization. [15]  As part of our ongoing program with dinuclear metal catalysts, [7, 9g,h, 13, 14] we questioned whether a hybrid version of the dimers D1 and D2 could potentially be created, which would combine the practical air-stability feature of D2 with the reactivity modes of D1 and D2 to ultimately enable rapid and selective cross-coupling reactions as well as olefin migrations under complete tolerance of air.
We focused especially on the E-selective generation of enamides, which owing to their exquisite reactivity and stability features make them tuneable enamine equivalents for a wide variety of transformations. [16] Double bond migrations to make fully N-substituted and acyclic enamides are currently largely limited to precious metals however, that is, Ru, Rh and Ir catalysis. [17] Only isolated examples have been reported for other metals, [18] and identification of a Pdbased methodology would therefore be valuable, especially if paired with air-tolerance.
We envisioned that a new generation of air-stable Pd I dimers for olefin isomerization ideally also bears the iodide bridges, which appear to be key for air stability, but then should be combined with a slightly less stabilizing ligand to also allow for intramolecular cyclopalladation and [Pd II -H] generation. We wondered whether a partial replacement of tert-butyl by cyclohexyl [19] would be possible and potentially lead to the desired features. Such a Pd I dimer is unknown. We therefore initially set out to synthesize the corresponding Pd 0 complex, i.e. Pd(PCy 2 t Bu) 2 which was found to have a characteristic 31 P-NMR spectroscopic signal at 54.1 ppm. We subsequently attempted a comproportionation of the Pd 0 complex with Pd II I 2 , which resulted in a new species. 1 H, 13 C and 31 P NMR spectroscopic analyses (signal at 79.2 ppm in 31 P NMR) as well as X-ray diffraction (see Figure 2) unambiguously confirmed that the novel dimer [Pd(m-I)-(PCy 2 t Bu)] 2 (D3) was generated in high yield (89 %). [20] Our assessment of its air stability indicated that it is bench stable as a solid for at least three months (which constitutes the time of our testing).
With a new generation of air-stable Pd I dimer D3 in hand, we subsequently investigated its reactivity in the isomerization of substrate 14 a (see Figure 2 A). To our delight, quantitative isomerization of 14 a to E-enamide 14 occurred in MeOH under open-flask conditions in 2 h at 50 8C. [21] By contrast, [Pd(m-I)(P t Bu 3 )] 2 D2 gave only 10 % conversion in MeOH (and 3 % in toluene) even after 7 h at 50 8C. The more labile Br-bridged D1 gave no conversion whatsoever under open-flask conditions due to its oxygen-sensitivity, but effectively yielded 14 under argon. Similarly, Skrydstrups Pd II -H system C2 [15] that is generated in situ from a mixture of Pd 2 (dba) 3 , isobutyryl chloride, and tri-tert-butylphosphine (C1) was effective under argon, but not when the reaction was set-up and performed in air.
The open-flask isomerization process is also amenable to scale-up: when we performed the reaction on 1 gram scale we isolated enamide 14 in 96 % yield with E/Z > 99:1 after simple filtration and solvent removal as purification.
We next set out to test the wider potential of catalyst D3 in isomerizations other than amide to enamide conversions. Pleasingly, D3 was just as powerful for all-carbon-based olefin transpositions and we successfully converted a number of allyl benzenes. Electron-rich as well as electron-poor substituents were tolerated, including OH groups, even if positioned ortho to the isomerization site (40). [22,23] Our preliminary mechanistic data on the nature of the active species indicated a palladium hydride to be likely. While we found that dimer D3 is of low solubility in MeOH, in the presence of a substrate in MeOH, we were able to detect two species by 31 P-NMR analysis during the course of the reaction, which likely correspond to a [Pd II -H] (63.8 ppm) and a cyclopalladated Pd species (À9.9 ppm). Analogous species were previously seen in isomerizations with D1. [10] Interestingly, when we performed the one-pot simultaneous isomerization of 52, 44 a with D3 in MeOH (see Figure 2 B), we only saw very little deuterium scrambling in the products 53, 44 b. Roughly 8 % deuterium content was detected in 44 b, which indicates that the monophosphine Pd II -H appears to stay largely coordinated to the substrate during the course of the isomerization.
Moreover, when we performed the isomerization of allyl benzene 44 a in deuterated methanol (CD 3 OD), we observed only very little D incorporation in the substrate (< 5 %, Bu and P t Bu 3 ] [24] C2 and C3 and subjected these species to the isomerization of 44 a in CD 3 OD, substantial deuteration in the product was observed. For L = P t Bu 3 , 95 % of deuterium content was detected in 44 c, whereas 50 % deuteration was seen for L = PCy 2 t Bu. These data indicate that the ligation state of the palladium hydride substantially affects the deuterium scrambling and intermolecular cross-overs. Pd I dimers are in this respect privileged in forming monophosphine [Pd II -H] directly.
Overall, these data showcase that double bond migrations are feasible in open air for a range of substrates and substitution patterns, which was previously not possible with the first generation air-stable Pd I dimer D2.
We previously showed that dimer D2 gives rise to extremely rapid and fully selective arylation and alkylation of aromatic C À Br bonds in the presence of C À OTf and C À Cl. [14a,h,i] As opposed to typical Pd 0 based transformations, the selectivity was independent of any steric or electronic impacts imposed by the substrates. Even if challenged with an adamantyl group ortho to CÀBr, the coupling still occurs there in favour over a less hindered CÀOTf site. [14a] Our data indicated that while Pd I can in principle directly add to C À I or C À Br bonds, the barriers for addition to C À Cl or C À OTf are too high. However, we found that use of a polar solvent along with an organometallic reagent allows for CÀOTf functionalization with D2 in air in less than 10 min (at r.t.), [14d] which is most likely due to an ate complex having been generated Yields of isolated products after chromatography, E/Z ratios were determined by GC-MS or quantitative 1 H NMR. [25] Angewandte Chemie Communications 21932 www.angewandte.org which then triggers the transformation. While we found C-OTf functionalization to be quite broad, it occasionally proceeded inefficiently when there was ortho substitution to the coupling site ( Figure 4).
We were intrigued whether D3 can also trigger siteselective CÀC bond formations, especially if sterically challenged. Figure 4 shows the results. The data indicate that dimer D3 can also trigger selective C sp 2 -C sp 3 as well as C sp 2 -C sp 2 couplings of C-OTf in NMP under open-flask conditions in less than 10 min at r.t. Interestingly, while D2 yielded poor conversion for the hindered 45 and 47 in < 20 %, the new dimer D3 gave excellent conversion and exclusive C-OTf selectivity. For comparison, bromide-bridged D1 also gives poor (< 20 %) conversion when subjected to the same transformations under argon (in air the dimer is deactivated prior to reaction). For the functionalization of the analogous fluorosulfate C-OSO 2 F (OFs) in the presence of C À Cl, [14b] D1 gave no selectivity, whereas D2 and D3 gave fully selective couplings in excellent yields. Moreover, in analogy to D2, [14d] the new generation D3 appears to be similarly effective in CÀ Br and CÀCl couplings under the analogous reaction conditions, which allows for the triply selective and modular functionalization of arenes also.
In conclusion, disclosed herein was the preparation and reactivity studies of a new generation of air-stable Pd I dimer. The newly developed [Pd(m-I)(PCy 2 t Bu)] 2 D3 allowed for the first efficient E-selective olefin migrations in air in just 2 h reaction time to generate E-enamides or styrene derivatives in the presence of various functionalities, including sulfones, (phosph)amides, aldehydes, ester, stereocenters, and even tolerating the frequently reactive aromatic C À I bonds or ortho-OH groups. Our mechanistic data indicate that barely any H/D-crossover and exchange takes place, which likely is due to a monophosphine [Pd II -H] being generated as active species. Dimer D3 was shown to also trigger rapid and chemoselective C sp 2 -C sp 2 as well as C sp 2 -C sp 3 cross coupling reactions of poly(pseudo)halogenated arenes in air, showing superior reactivity compared to previous dimer generations for triflate substrates that contain ortho substituents.