Feriel
Rekhroukh
ab,
Charlie
Blons
ab,
Laura
Estévez‡
c,
Sonia
Mallet-Ladeira
d,
Karinne
Miqueu
*c,
Abderrahmane
Amgoune
*ab and
Didier
Bourissou
*ab
aUniversité de Toulouse, UPS, 118 route de Narbonne, 31062 Toulouse, France. E-mail: amgoune@chimie.ups-tlse.fr; dbouriss@chimie.ups-tlse.fr
bCNRS, LHFA, UMR 5069, F-31062 Toulouse, France
cCNRS, Univ Pau & Pays Adour, Institut des Sciences Analytiques et de Physico-Chimie pour l'environnement et les Matériaux, UMR 5254, Pau, 64000, France. E-mail: karinne.miqueu@univ-pau.fr
dUniversité Paul Sabatier, Institut de Chimie de Toulouse (FR 2599), 118 route de Narbonne, 31062 Toulouse Cedex 9, France
First published on 19th April 2017
The synthesis and characterization of the first gold(III)–arene complexes are described. Well-defined (P,C)-cyclometalated gold(III)–aryl complexes were prepared and characterized by NMR spectroscopy. These complexes swiftly and cleanly reacted with norbornene and ethylene to provide cationic gold(III)–alkyl complexes, in which the remote phenyl ring was η2-coordinated to gold. The interaction between the aromatic ring and the gold(III) center was thoroughly analyzed by NMR spectroscopy, X-ray diffraction, and DFT calculations. The π–arene coordination was found to significantly influence the stability and reactivity of low coordinated gold(III) alkyl species.
In particular, no gold(III)–arene complex has been characterized to date. This is all the more striking and unfortunate as gold(III) complexes have the remarkable ability to activate aromatic C–H bonds, and π–arene complexes likely represent the initial stage of such processes.12–14 This view is supported by a recent computational study by Swang and co-workers.15 Accordingly, benzene was predicted to coordinate in an η1-fashion to a dicationic gold(III) fragment [(bpy)AuIII(C6H5)]2+, and low activation barriers were computed for benzene rotation as well as proton jumping between the coordinated phenyl and benzene ligands. However, attempts to synthesize relevant gold(III) complexes were unsuccessful.15 While gold(III)–arene complexes remain unprecedented, it must be noted that a few gold(I) complexes featuring intra and intermolecular Au(I)–arene interactions have been structurally characterized.16–19 In these complexes, the arene acts as a weakly coordinating ligand and it can be easily displaced by substrates such as alkynes.
The key bottlenecks in the quest for gold(III)–arene complexes lie in the design and access of low coordinated cationic gold(III) complexes. We have recently developed a simple synthetic route to cyclometalated [(P,C)Au(III)–Me]+ complexes and have shown their ability to readily insert olefins into the gold(III)–carbon bond.8,20 The resulting low coordinated gold(III) alkyl complexes showed a high tendency for β-hydride elimination21 and enabled the first authentication of a C–H agostic interaction with gold.22 These results motivated us to explore the synthesis of gold(III)–arene complexes by olefin insertion into a gold(III)–aryl bond (Chart 1).
Chart 1 Schematic representation of the targeted gold(III)–arene complexes and their generation by olefin insertion into a Au–aryl bond. |
As reported hereafter, this strategy proved fruitful and empowered us to characterize gold(III)–arene complexes. Migratory insertion of norbornene and ethylene into the Au(III)–Ph bonds is shown to be a very facile process that leads to gold(III) alkyl complexes stabilized by intramolecular coordination of the phenyl ring. The mechanism of formation of the gold(III)–arene complexes and their bonding situation have been thoroughly investigated spectroscopically, crystallographically, and computationally. The influence of arene coordination on the fate of low coordinated gold(III) alkyl species is also discussed.
Scheme 1 Preparation of the cationic gold(III)–phenyl complex 3 and molecular view of its neutral precursor 2 with thermal ellipsoids drawn at the 50% probability level. |
The iodide at the gold was then abstracted by adding complex 2 to AgSbF6 in dichloromethane. According to 31P NMR monitoring, the reaction was instantaneous at −80 °C and afforded a mixture of two cationic species in a 70/30 ratio (δ31P: 77.4 and 80.8 ppm, respectively). The 1H and 13C NMR spectra also display two sets of similar signals, suggesting that the two species correspond to closely related forms of complex 3.25 Variable-temperature 31P NMR studies indicated that the ratio between the two forms does not vary significantly from −80 to −20 °C (decomposition into several unidentified species occurs above −20 °C). However, the existence of a dynamic exchange between the two species was clearly apparent from the cross-peaks observed in the 2D 31P{1H} EXSY NMR spectrum. The 13C NMR signal for the phenyl ipso carbon atom Cipso bound to gold is deshielded by about 10 ppm upon cationization (δ from 168.4 ppm for 2 to 177.2 ppm for the major form of 3). The large 2JPC coupling constant is retained (134 Hz for 2, 120 Hz for 3), indicating that the phenyl ring still occupies the position trans to the phosphine moiety, in line with the dissymmetric trans influence of the (P,C) chelate. The reaction of complex 3 with olefins was then investigated by carrying out the iodide abstraction of 2 in the presence of norbornene or ethylene.
The CH of the norbornyl moiety bound to gold resonates in 13C NMR as a doublet at 57.0 ppm with a large 2JPC coupling constant of 72.5 Hz, indicating a trans arrangement of the phosphorus atom and norbornyl group (Fig. 1a). In addition, six 13C NMR signals are found for the phenyl ring, suggesting hindered rotation around the Csp3–Csp2 bond and possibly some interaction between the phenyl ring and the gold center. The chemical shift of the ipso carbon is shifted to high field compared to that of 2-phenylbicycloheptane (PBH)26 (δ 111.8 ppm for 4vs. 147.6 ppm for PBH) and appears as a doublet due to coupling with phosphorus (JPC = 9.1 Hz). The presence of the gold(III) center also affects the ortho-CH signals, one of them being shifted to high field by about 10 ppm (δ 124.1 and 135.6 ppm). These spectroscopic data are strongly reminiscent of those observed by Cheng and Catellani for related neutral Pd(II)–arene complexes27–29 and suggest η2-coordination of the phenyl ring to gold(III). Some of these Pd(II)–arene complexes have been crystallographically characterized and their bonding situation has been thoroughly investigated.27–29
With the aim of obtaining structural confirmation of the gold(III)–arene interaction, we prepared another gold(III)–aryl complex 2′ with an OMe group in the para position of the phenyl ring (Scheme 2). This electron-rich arene ring was envisioned to impart greater stability to the gold(III) π-complex and hopefully higher crystallinity. Starting from 2′, the insertion of norbornene also works well. The ensuing gold(III)–arene complex 4′ was isolated and fully characterized. To our delight, crystals suitable for X-ray diffraction analysis were obtained in this case. Accordingly, complex 4′ (Fig. 1b) adopts a 3-coordinated T-shaped structure (PAuCNB and PAuCnaphthyl bond angles of 177.29(7)° and 84.15(7)°, respectively). As expected, the insertion of norbornene into the Au–Ph bond places the gold atom and the phenyl ring cis and exo to the norbornyl group. The arene sits trans to the naphthyl backbone with short contacts between gold and two carbon atoms (Cipso and one Cortho). The Au–Cipso distance (2.416(2) Å) is shorter than the Au–Cortho distance (2.593(3) Å), indicating dissymmetric η2-coordination of the arene to gold(III). The interaction of the (p-OMe)-phenyl ring with the gold(III) center is also apparent from NMR (Fig. 1a). As for 4, the 13C signals for Cipso (δ 90.3 ppm) and one of the Cortho atoms (δ 115.7 ppm) of 4′ are shifted to high field compared to those of 2-(p-OMe)-phenylbicycloheptane26 (Δδ13C = 49 ppm for Cipso and 13 ppm for Cortho).30 Interestingly, the NMR signals for the (p-OMe)-phenyl ring are desymmetrized even at room temperature, suggesting tighter arene coordination in 4′vs.4.
Scheme 2 Formation of the gold(III) norbornyl complexes 4 and 4′ by insertion of norbornene into the Au(III)–Ar bond of 2 and 2′. |
The reaction profile accounting for the formation of complex 4 from 3 was also computed. This involves the coordination and migratory insertion of norbornene into the Au–Ph bond, followed by cis/trans isomerization (Fig. 2). The insertion of NB proceeds via an in-plane 4-center transition state and gives the gold(III)–norbornyl complex trans-4. The associated activation barrier is low (ΔG≠ = 10.7 kcal mol−1), much lower than that computed for the reaction of NB with the corresponding Au(III)–Me complex (ΔG≠ = 18.7 kcal mol−1).31 Complex trans-4 is not observed experimentally. It spontaneously isomerizes (with a low barrier ΔG≠ of 14.3 kcal mol−1) into the more thermodynamically stable complex cis-4. The Gibbs free energy ΔG for the overall transformation cis-3 → cis-4 is −19.4 kcal mol−1.
Fig. 2 Reaction profile (ΔG in kcal mol−1) computed at the B3PW91/SDD+f(Au)/6-31G**(other atoms) level of theory for the formation of the gold(III)–arene complex 4 upon reaction of cis-3 with NB. |
The insertion of NB into the Au–Ph bond of 3 is a very facile process. It affords a straightforward entry to the cationic gold(III) norbornyl complex 4 and to the neutral cyclometalated complex 5 following C–H activation. The η2-coordination of the phenyl ring to gold has been unambiguously authenticated in 4. This intramolecular gold(III)–arene interaction provides significant thermal stability to complex 4. Indeed, cationic gold(III) alkyl complexes tend to rapidly undergo reductive elimination or β-hydride elimination even at low temperature and only a few such species have been characterized so far.21,22 These observations prompted us to explore the reactivity of the cationic phenyl complex 3 towards ethylene and to thereby probe the impact of π–arene interactions on the reactivity of gold(III) alkyl species.
Scheme 4 Formation of the gold(III) alkyl complex 6 upon reaction of the cationic (P,C)gold(III) phenyl complex with ethylene. |
The methine group at the gold resonates in 13C NMR at δ 68.6 ppm with a small 2JPC coupling constant of 7.7 Hz, indicating that the sec-alkyl moiety sits in this case in a trans position to the Cnaphthyl atom. The 13C NMR spectrum also revealed a significant shift to high field of the Cipso and one of the Cortho carbon atoms of the phenyl ring compared to those of ethylbenzene (Δδ13C = 20 ppm for Cipso and 16 ppm for Cortho, Fig. 3a).32 All the carbon atoms of the phenyl ring except for one of the Cmeta couple with phosphorus (4.4 < JPC < 10.3 Hz), and one of the Hortho resonance signal also display coupling to phosphorus (JPH = 3.4 Hz). All these data support η2-coordination of the phenyl ring to gold(III), as in the case of the gold(III) norbornyl complex 4.
Complex 6 shows unusual thermal stability (no degradation is observed after 24 h at 0 °C) given the presence of hydrogen atoms in a β position and the proclivity of cationic 3-coordinated gold(III) alkyl complexes to undergo β-H elimination.21 Indeed, the corresponding n-propyl and n-butyl complexes were previously found to rapidly release propene and butenes at 0 °C. A similar process with the formation of styrene is observed with 6 but only after 1 h at room temperature. The higher stability of complex 6 is most likely due to the π-interaction of the phenyl ring with the gold(III) center.
The structure of complex 6 was further assessed by DFT and NBO calculations. Accordingly, complex 6 adopts a slightly distorted T-shaped geometry (Fig. 3b and Table S3†) with PAuC(H) and CnaphthylAuC(H) bond angles of 109.4 and 166.3°, respectively. The pendant phenyl ring is significantly bent towards the cationic [(P,C)Au]+ fragment [AuC(H)Cipso bond angle = 74.2°] and interacts with the gold center via the Cipso and Cortho carbons in an asymmetrical manner (Au–Cipso: 2.28 Å, Au–Cortho: 2.55 Å). The η2-coordination of the phenyl ring to gold(III) is corroborated by the 13C NMR data computed for the corresponding Cipso and Cortho carbon atoms (Table S3†). The respective resonance signals are shifted to high field (δ 124.6 and 120.6 ppm for Cipso and Cortho, respectively) and display JPC coupling constants of 5 and 10.7 Hz, respectively. NBO analysis (Fig. S31†) identified a donation from the π(CipsoCortho) orbital to the Au σ*(AuCnaphthyl) orbital with a delocalization energy ΔE(2) of 38.6 kcal mol−1, a value very close to that found in the gold(III) norbornyl complex 4. It is interesting to note that complex 6 is characterized experimentally in its trans form (the Cnaphthyl and Calkyl atoms are in a trans arrangement) and does not isomerize into the more thermodynamically stable cis-isomer (ΔG = −5.2 kcal mol−1 and ΔG≠ = 25 kcal mol−1, Fig. S30†).
To gain a deeper insight into the mechanism of formation of complex 6, the reaction of the cationic complex 3 with ethylene was thoroughly investigated by DFT calculations (Fig. 4 and S32†). Ethylene first coordinates to gold to give complex cis-3π (ΔG = 1.5 kcal mol−1) and then inserts into the Au–Ph bond to afford the Au(CH2CH2Ph) complex trans-A. The activation barrier for the migratory insertion is low (ΔG≠ = 9.5 kcal mol−1 from the initial reactants). In line with experimental observations, insertion into the Au–Ph bond is kinetically more facile than into the Au–Me bond (the corresponding activation barrier reaches 16.5 kcal mol−1). Complex trans-A displays some π–arene interaction [according to NBO, ΔE(2) = 10.5 kcal mol−1] (Fig. S33†) but coordination of the pendant phenyl ring to gold(III) is less stabilizing with the flexible CH2CH2 spacer than that encountered in complex 6. In line with the fact that trans-A is not observed experimentally, it easily undergoes β-hydride elimination (ΔG≠ = 4.8 kcal mol−1).33 The reaction is slightly exergonic and leads to intermediate cis-H. Finally, the Au(III)–H styrene complex readily undergoes (2,1) re-insertion of styrene into the Au–H bond to give complex trans-6 with an activation barrier of ΔG≠ = 8.0 kcal mol−1 from cis-H to trans-6 (styrene dissociation is comparatively disfavored with ΔG≠ = 14.5 kcal mol−1).34,35 As mentioned above, η2-coordination of the phenyl ring substantially stabilizes complex trans-6 (ΔG = −22.9 kcal mol−1), and the latter species is actually the thermodynamic product of ethylene insertion into the Au–Ph bond, enabling its experimental characterization. Note that in accordance with experimental observations, the arene coordination disfavors β-hydride elimination from complex trans-6 (Fig. S34†).24 The associated energy barrier ΔG≠ amounts to 16.5 kcal mol−1vs. 7.1 kcal mol−1 for the corresponding [(P,C)Au(III)(n-Bu)]+ complex.21 The arene coordination also disfavors isomerization of trans-6 into its more thermodynamically stable cis isomer (Fig. S30†). The latter process requires a high activation energy (ΔG≠ = 25 kcal mol−1) and is even less favored than β-hydride elimination.
This study also shows that migratory insertion of olefins such as norbornene and ethylene into the Au(III)–Ph bond is a very facile process. This transformation and the ensuing intermediates, which are stabilized intramolecularly by π–arene coordination, are relevant to a number of key catalytic processes (such as Mizoroki–Heck coupling, the Catellani reaction, and the hydroarylation of olefins).
Future work from our group will seek to take advantage of the increasing scope of gold(III) species and reactivities to develop valuable catalytic transformations.
Footnotes |
† Electronic supplementary information (ESI) available: Experimental details (syntheses, analytical data and copies of 1H, 13C and 31P NMR spectra), crystallographic data of 2 and computational details. CCDC 1526613 and 1538015. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c7sc00145b |
‡ Current address: Departamento de Química Física, Universidade de Vigo, Facultade de Química Lagoas-Marcosende s/n, 36310 Vigo, Galicia, Spain. |
This journal is © The Royal Society of Chemistry 2017 |