L-Shaped Heterobidentate Imidazo[1,5-α]pyridin-3-ylidene (N,C)-Ligands for Oxidant-Free Au(I)/Au(III) Catalysis

Abstract

In the last decade, major advances have been made in homogenous gold catalysis. However, Au(I)/Au(III) catalytic cycle remains much less explored due to the reluctance of Au(I) to undergo oxidative addition and the stability of the Au(III) intermediate. Herein, we report activation of aryl halides at gold(I) enabled by NHC (NHC = N-heterocyclic carbene) ligands through the development of a new class of L-shaped heterobidentate ImPy (ImPy = imidazo[1,5-α]pyridin-3-ylidene) N,C ligands that feature hemilabile character of the amino group in combination with strong σ-donation of the carbene center in a rigid conformation, imposed by the ligand architecture. Detailed characterization and control studies reveal key ligand features for Au(I)/Au(III) redox cycle, wherein the hemilabile nitrogen is placed at the coordinating position of a rigid framework. Given the tremendous significance of homogenous gold catalysis, we anticipate that this ligand platform will find widespread application.

Graphical Abstract

We report activation of aryl halides at gold(I) enabled by NHC (NHC = N-heterocyclic carbene) ligands through the design of a new class of L-shaped heterobidentate ImPy (ImPy = imidazo[1,5-α]pyridin-3-ylidene) N,C ligands that feature hemilabile character of the amino group in a locked conformation. Experimental and DFT studies demonstrate pivotal stabilization of Au(III) species by the hemilabile nitrogen for successful Au(I)/Au(III) catalysis.

Introduction

Synthetic chemistry of gold has experienced explosive development over the last two decades.^[1] At present, in addition to indispensable applications of gold in electronics, medicines, functional materials, and aerospace engineering as well as its common presence in jewelry, gold catalysis is considered a fundamental method for the synthesis of chemicals in both academic and industrial research. In particular, gold complexes are among the most powerful catalysts for activation of unsaturated C─C bonds in alkynes, alkenes and allenes due to excellent carbophilic π-acidity of gold(I). The nucleophile addition to the gold activated π-systems provides diverse new C─C and C─heteroatom bonds without the change in oxidation state of the gold center during the catalytic cycle.^[2]

In parallel, much less progress has been made in Au(I)/Au(III) cycle for cross-coupling reactions owing to high redox potential of Au(I)/Au(III) (E⁰ = +1.41 V), compared with Pd(0)/Pd(II) (E⁰ = +0.92 V), and relative instability of the resulting Au(III) intermediates, which make Au(I) oxidative addition and Au(I)/Au(III) cycle a major challenge.^[3] In order to overcome the high redox potential of Au(I)/Au(III), sacrificial stoichiometric internal or external oxidants, such as hypervalent iodine reagents or electrophilic fluorinating oxidants, have been implemented to exploit Au(I)/Au(III) cycle (Figure 1A).^[4] However, the use of oxidants presents a major limitation, including low functional group tolerance, low atom economy or the necessity for pre-functionalized substrates, thereby reducing practical significance of these reactions. In addition, although a significantly milder approach in combining gold catalysis with photoredox catalysis has been achieved to address the high redox barrier of the Au(I)/Au(III) cycle, only limited electrophiles, such as aryldiazonium and aryl iodonium salts, are compatible with these protocols (Figure 1A).^[5]

Figure 1.

Au(I)/Au(III) redox cycle under oxidant-free conditions: context of this work.

In terms of oxidative addition of Au(I), the direct intramolecular oxidative addition of C(sp²)─I/Br and C(sp³)─Br bonds under oxidant-free conditions was demonstrated independently by Bourissou^[6] and Toste^[7] in 2014 (Figure 1B). In 2015, Ribas and co-workers reported oxidative addition of C(sp²)─I/C(sp²)─Br to Au(I) with the assistance of nitrogen chelation.^[8] However, at that point, C─X oxidative addition to Au(I) was still limited to intramolecular and chelation assisted models. In 2014, Bourissou group reported ligand-enabled intermolecular C─X oxidative addition to Au(I) by assistance of bidentate phosphine ligand under oxidant-free conditions.^[9] It was proposed that preorganized bidentate bent bonding of Au(I) is the key to promote the oxidative addition step by facilitating the formation of square-planar geometry of the Au(III) oxidative addition product. After this finding, two types of bidentate phosphine and pyridine ligands were applied to achieve ligand-enabled C─X oxidative addition to Au(I) under oxidant-free conditions (Figure 1C).^[10-13] In 2017, Bourissou group demonstrated Au(I) oxidative addition with P,N-bidentate ligand featuring hemilabile nitrogen atom to provide extra stabilization to Au(III).^[10] In 2018, Russell and co-workers demonstrated bipyridine-type N,N-ligands to promote C─X oxidative addition to Au(I) (Figure 1C).^[11] Furthermore, in 2019, Hashmi group reported that tri- and tetra-coordination of Au(I) facilitates oxidative addition of C(sp) centers to Au(I) using N,N ligands (Figure 1C).^[12,13] Recently, C,N hemilabile mesoionic MIC ligands with pendant pyridine moieties for Au(I)/Au(III) catalysis were developed by the Bourissou and Ribas groups to stabilize Au(III) complexes and catalyze the arylation-lactonization of γ-alkenoic acids, albeit these complexes feature low catalytic activity.^[14] Thus, although studies in ligand design demonstrated that the challenging Au(I) oxidative addition is feasible,^[10-15] to date, very few ligands promote Au(I)/Au(III) catalysis under oxidant-free conditions using ubiquitous halides as electrophiles and only bidentate Me-DalPhos has been demonstrated as a general ligand for this catalysis manifold.^[16] At present, the rational ligand design and mechanism elucidation represent an unmet challenge in ligand-controlled, oxidant-free Au(I)/Au(III) catalysis.

As part of our program in NHC catalysis,^[17] we proposed that a hemilabile group with locked conformation on NHC scaffold would facilitate oxidative addition of Au(I) and maximize the stabilization of Au(III) intermediates (Figure 1D). Herein, we report activation of aryl halides at gold(I) enabled by NHC ligands through the development of a new class of L-shaped heterobidentate ImPy (ImPy = imidazo[1,5-α]pyridin-3-ylidene) N,C ligands that feature hemilabile character of the amino group in combination with strong σ-donation of the carbene center in a rigid conformation, imposed by the ligand architecture (Figure 1E).^[17] These ligands inherit the features of easily-modifiable, strongly σ-donating NHC ligands merged with hard nitrogen donor for metal stabilization. The unique hemilabile N,C ImPy NHC framework facilitates C─X oxidative addition to Au(I), and stabilizes Au(III) intermediate, enabling mild, oxidant-free Au(I)/Au(III) C─H arylation. This system and its bench-stable Au(I) precatalysts have been successfully applied in mild cross-coupling reactions between electron-rich arenes and readily available aryl halides under Au(I)/Au(III) oxidant-free conditions. Extensive experimental and DFT studies elucidate the mechanism and key features of the catalyst framework, demonstrating pivotal stabilization of Au(III) species by the hemilabile nitrogen atom. We further present rapid and chromatography-free large scale ligand synthesis, enabling broad access to this family of heterobidentate N,C ligands.^[18] Considering the broad utility of N-heterocyclic carbene ligands in gold catalysis^[1,2] and given the general scarcity of ligands for Au(I)/Au(III) cycle, we anticipate that new ligand platform will find wide application for oxidant-free Au(I)/Au(III) catalysis and enable advanced ligand designs for Au(I)/Au(III) cycle.

Crucially, the (P,N) platform offers more possibilities for steric and electronic tuning cf. phosphines owing to the electronic and steric properties of NHC ligands, which is the key aspect for catalytic development.^[16-18]

Results and Discussion

Ligand Design and Reaction Optimization.

Inspired by previous studies on oxidative addition of C─X bonds to Au(I) (Figure 1A-C), our initial trials for Au(I)/Au(III) catalysis began with the model C─C coupling reaction between 1,3,5-trimethoxybenzene 1a (1.0 equiv) and iodobenzene 2a (1.0 equiv) (Table 1), a system established by Bourissou in 2017.^[10] Our initial reaction conditions employed NHC─Au─Cl (5 mol%) as precatalyst and AgOTf (1.1 equiv) as halide scavenger in MeOH. AgOTf was selected as halide scavenger to abstract the iodide from the oxidative addition intermediate and form the more accessible and electrophilic Au(III) species for the C─H activation step.

Table 1.

Summary of optimization and control reaction conditions.^a

entry	[Au─NHC]	AgX	yield (%)
1	IPrAuCl (L1)	AgOTf	0
2	IMesAuCl (L2)	AgOTf	0
3	6-tBu-2-PyDippAuCl (L3)	AgOTf	0
4	ImPyNMe2DippAuCl (L4)	AgOTf	55
5	ImPyNMe2MesAuCl (L5)	AgOTf	40
6	ImPyNEt2DippAuCl (L6)	AgOTf	<2
7	ImPyNEt2MesAuCl (L7)	AgOTf	<2
8	ImPyNPipDippAuCl (L8)	AgOTf	13
9	ImPyNPipMesAuCl (L9)	AgOTf	<2
10	ImPyTrippDippAuCl (L10)	AgOTf	0
11	ImPyTrippIMesAuCl (L11)	AgOTf	0
12	ImPyNMe2DippAuCl (L4)	AgNTf2	97
13	ImPyNMe2DippAuCl (L4)	AgSbF6	79
14	ImPyNMe2DippAuCl (L4)	AgNTf2	96b

^a1,3,5-trimethoxybenzene (1.0 equiv), iodobenzene (1.0 equiv), AgX (1.1 equiv), NHC-Au-Cl (5 mol%), methanol (0.125 M), 80 °C, 16 h.

^bNHC-Au-Cl (2.5 mol%).

As we expected, the classical imidazol-2-ylidene complexes, such as IPrAuCl (L1) and IMesAuCl (L2), were completely inactive in this Au(I)/Au(III) redox process (Table 1, entries 1-2). We hypothesized that a hemilabile or bidentate NHC framework is required to achieve the reluctant Au(I) oxidative addition. We thus tested NHC─Au─Cl complex 6-tBu-2-PyDippAuCl (L3) bearing hemilabile N-pyridine wingtip fixed in the syn conformation by the 6-t-Bu group;^[19] however, this complex was also completely unreactive and resulted in no reaction (Table 1, entry 3). Thus, we hypothesized that a more rigid NHC framework with a conformationally-locked hemilabile donor are required for coordination to Au(I) to achieve oxidative addition and stabilization of Au(III) species.

We were attracted to sterically demanding ImPy (ImPy = imidazo[1,5-α]pyridin-3-ylidene) architecture.^[20-22] Sterically, this class of ligands is well-positioned to satisfy the spacial requirements of the Au(I)/Au(III) cycle with variable N1/C5 substitution, while, electronically, the presence of a coordinating donor atom at the C5 position offers the hemilabile handle for stabilization of Au(I)/Au(III) intermediates.^[18] Thus, we designed and synthesized amino-decorated N,C hemilabile ligands on imidazo[1,5-α]pyridine architecture (L4–L9) (vide infra). The amino substituent is placed at the 5-position of the scaffold, which in combination with the sterically-flexible N2-aryl wingtip and imidazo[1,5-α]pyridin-3-ylidene bicyclic framework, results in unique features inherent to NHC ligands, such as stronger σ-donation, better π-acceptance and umbrella-type L-shape steric arrangement (cf. phosphines and N^N ligands). The rigid NHC template enforces conformationally-fixed hemilabile N-substitution and 1,4-relationship between the hard and soft donors, N and carbene. This arrangement enables coplanarity of Au coordinated to the NHC center and hemilabile N atom in the imidazo[1,5-α]pyridine scaffold, which offers transient stabilization of both Au(I) and Au(III) for Au(I)/Au(III) cycle. Finally, the inherent nature of NHC ligands offers the potential for facile structure diversification and reactivity tuning by N-wingtip substitution and amine modification.

To our delight, the parent complex prepared in this class, ImPyNMe₂DippAuCl (L4), bearing C5-dimethylamino substituent and N2-Dipp wingtip promoted the Au(I)/Au(III) C─H arylation in 55% yield (Table 1, entry 4). Interestingly, its N2-IMes analogue (L5) was also reactive (Table 1, entry 5), while congeners bearing more sterically-demanding C5-diethylamino (L6–L7) and C5-piperidinyl (L8–L9) substituents were less reactive (Table 1, entries 6-9). It should be noted that product formation was observed in all cases (Table 1, entries 6-9). These results suggest that bulkier NR₂ groups might decrease the facility of the C─H activation step (vide infra). Importantly, control reactions with ImPy C5-aryl ligands,^[20] ImPyTrippDipp (L10) and ImPyTrippMes (L11) (Table 1, entries 10-11) resulted in no reaction, indicating that the replacement of the amine substituent with an aryl at the C5-position is unproductive for Au(I)/Au(III) catalytic cycle (vide infra). After additional optimization, the yield of desired cross-coupling product 3a could be increased to 97% in the presence of AgNTf₂ in MeOH at 80 °C (Table 1, entries 12-13). Furthermore, decreasing catalyst loading to 2.5% had little effect on the reaction efficiency under the oxidant-free conditions in this model C─I/C─H cross-coupling (Table 1, entry 14).

Substrate Scope.

With this novel heterobidentate ImPy N,C Au(I) catalyst system in hand, we set out to explore the scope of various iodobenzenes in this reaction. As shown in Scheme 1, the arylation of 1,3,5-trimethoxybenzene proceeded smoothly with diverse iodobenzenes with excellent functional group tolerance, expected for Au(I)/Au(III) cycle under mild oxidant-free conditions. For example, various functional groups, such as electron-deficient CO₂Me, CF₃ and NO₂ at para- and meta-positions were well tolerated to afford 3b, 3j and 3k in 99%, 83% and 95% yields, respectively. Furthermore, electron-neutral and electron-rich iodobenzenes, such as 4-t-Bu, 4-Me, 4-OMe, 2-OMe as well as iodonaphthalene provided 3c (99%), 3i (99%), 3g (99%), 3h (83%) and 3f (83%) in excellent yields. Notably, complete chemoselectivity was observed using bifunctional handles, such as 1-bromo-4-iodobenzene 3d (99%), 1,4-diiodobenzene 3n (45%), pseudohalide 4-OTf 3e (83%) and nucleophile 4-Bpin 3m (70%). This chemoselectivity indicates excellent capacity of Au(I)/Au(III) catalysis to complement the traditional transition-metal-catalyzed cross-coupling reactions and highlights its potential in iterative processes. Moreover, complete arylation regioselectivity was obtained in the arylation of 1,3-dimethoxybenzene (3l, 60%). Of note, the synthetic utility of this mild C─H arylation protocol was demonstrated in late-stage functionalization by cross-coupling between menthol and cholesterol-derived arenes, which afforded the functionalized products 3o and 3p in 95% and 72% yields.

Scheme 1.

Substrate scope.^{a a}conditions: arene (1.0 equiv), iodobenzene (1.0 equiv), [Au(I)-NHC] = ImPyNMe₂DippAuCl (5 mol%), AgNTf₂ (1.1 equiv), methanol (0.125 M), 80 °C, 16 h. See SI for details.

Several additional points should be noted: (1) the reactivity of aryl iodides compares favorably with the previous system (see SI); (2) based on solvent screening results, coordinating solvents are not required but have beneficial effect for the yield; (3) the coupling takes place at 40 °C (44% yield), with reactivity observed at room temperature (9%) (see SI).

Control Studies.

The success of the present class of heterobidentate ImPy N,C ligands hinges upon the hemilabile N,carbene architecture in a 1,4-relationship between the hard and soft donors. In order to further elucidate the key factors responsible for the successful heterobidentate ImPy N,C ligand design for Au(I)/Au(III) catalysis, control ligands bearing C5-iPr and C7-NMe₂ groups and their gold complexes [ImPyiPr^C5DippAuCl] (L12) and [ImPyNMe₂^C7DippAuCl] (L13) were synthesized (Table 2). Importantly, the prepared analogue L12 (C5-i-Pr) features similar steric bulk to the most reactive heterobidentate catalyst ImPyNMe₂DippAuCl (L4), while the complex L13 (C7-NMe₂) features similar electronics with the most reactive ImPy catalyst ImPyNMe₂DippAuCl (L4). As shown in Table 2, both complexes L12 and L13 were found completely unactive in the model reaction, indicating that (1) C5-donor substituent is required for the successful catalysis, and (2) C5 substituent must be a hemilabile group instead of a group with a similar steric bulk on the NHC scaffold. The unproductive coupling with complex L13 (C7-NMe₂ vs. C5-NMe₂) highlights that the key factor in ligand design for Au(I)/Au(III) redox cycle is hemilabile positioning of the NHC ligand (cf. electronics, vide infra).

Table 2.

Control studies.^a

entry	[Au─NHC]	yield (%)
1	ImPyiPrC5DippAuCl (L12)	0
2	ImPyNMe2C7DippAuCl (L13)	0

^aConditions: 1,3,5-trimethoxybenzene (1.0 equiv), iodobenzene (1.0 equiv), AgNTf₂ (1.1 equiv), NHC-Au-Cl (5 mol%), methanol (0.125 M), 80 °C, 16 h.

Crystallographic Characterization.

The structures of the most reactive complex ImPyNMe₂DippAuCl (L4) and its structural C7-NMe₂ analogue (L13) were unambiguously determined by crystallographic analysis (Figure 2A). Interestingly, complex L4 showed slight deviation from the linear Au(I) geometry (C1─Au─Cl, 174.6°; C1-Au, 1.984Å) cf. complex L13 (C1─Au─Cl, 178.7°; C1-Au, 1.983 Å). Moreover, the distance of C5─N_(NMe2) (1.390 Å) in complex L4 is longer than the analogous C7─N_(NMe2) distance (1.368 Å) in complex L13. The distance between N─Au in complex L4 is 3.141 Å, indicating a weak interaction between Au and hemilabile N atom (vide infra).

Figure 2.

(A) X-ray crystal structure of ImPyNMe₂DippAuCl (L4) and control NHC-Au-Cl (L13). Hydrogen atoms have been omitted for clarity. Selected bond lengths [Å] and angles [°]. ImPyNMe₂DippAuCl, L4: Au─C1, 1.984(8); Au─Cl, 2.279(2); C1–N1, 1.35(1), C1–N2, 1.37(1); Au─N3, 3.141, C7-N3, 1.39(1); Cl─Au─C1, 174.6(2); N1–C1–N2, 103.6(6); Au-C1-N2, 130.3(6), Au-C1-N1, 124.6(6). ImPyNMe₂^C7DippAuCl, L13: Au─C1, 1.983(3); Au─Cl, 2.2816(8); C1–N1, 1.345(5); C1–N2, 1.363(6), C5-N3, 1.368(6); Cl─Au─C1, 178.7(1); N1–C1–N2, 104.4(3); Au-C1-N2, 127.1(3), Au-C1-N1, 128.3(3). CCDC 2213634 (L4); CCDC 2213788 (L13). (B) Topographical steric maps of L4 and L13 showing % V_bur per quadrant.

We used the steric map analysis developed by Cavallo and co-workers^[21] to determine the % buried volume (%V_bur) and define steric impact of ImPy NHC-Au-Cl complexes L4 and L13 (Figure 2B). As shown, the steric map analysis of complex L4 revealed the (%V_bur) of 43.6% with 47.2%, 48.4%, 27.8%, 50.9% for each quadrant. These values can be compared with the (%V_bur) of 35.6% for complex L13 with 43.7%, 47.2%, 25.3%, 26.2% for each quadrant, and the (%V_bur) of 46.3% with 46.2%, 46.5%, 46.2%, 46.5% for each quadrant for the analogous imidazol-2-ylidene IPr-Au-Cl complex. The steric map of ImPyNMe₂DippAuCl (L4) indicates a unique spatial impact of the C5-amino ImPy framework with one unencumbered quadrant and three sterically-hindered quadrants. This steric arrangement permits facile access of substrates to the ligand catalytic pocket for oxidative addition without the loss of necessary bulkiness of ligands important for reductive elimination, in addition to the hemilabile nitrogen donor atom in the rigid arrangement.

Mechanistic Studies.

To gain further insight into the successful reactivity of heterobidentate ImPy N,C ligand (L4) in Au(I)/Au(III) catalysis, we prepared the oxidative addition Au(III) complex 6 by reacting biphenylene with L4 (Figure 3A). The reaction occurred at room temperature, indicating facile reactivity of complex L4 to undergo oxidative addition.

Figure 3.

(A) Oxidative addition of biphenylene. Conditions: (a) NHC-Au-Cl (1.0 equiv), AgSbF₆ (1.1 equiv), CH₂Cl₂, 23 °C. (b) biphenylene (2.25 equiv), [CH₃(CH₂)₃]₄NCl (1.25 equiv), 23 °C. (B) X-ray crystal structure of Au(III) complex 6. Hydrogen atoms have been omitted for clarity. Selected bond lengths [Å] and angles [°]: Au─C1, 2.071(4); Au─Cl, 2.383(1); C1–N1, 1.368(5), C1–N2, 1.351(5); Au─N3, 2.934(4), C4-N3, 1.406(6); C1─Au─Cl, 90.6(1); N1–C1–N2, 104.0(3); N2-C1-Au, 130.4(3), N1-C1-Au, 124.8(3). CCDC 2213789 (6). (C) Topographical steric map of 6 showing % V_bur per quadrant.

In accordance with reports of Toste,^[23] Bourissou^[10] and Bertrand,^[24] the hemilabile nitrogen in L4 provides stabilization for gold center in the +3 oxidation state (Figure 3B), albeit there is no coordination with Au(I) center (Figure 2A, vide infra). Indeed, the X-ray structure of the Au(III) oxidative addition complex 6 confirms that the Au─N distance between Au(III) center and NMe₂ is 2.934 Å, which is a significantly lower than the sum of van der Waals radii [r(Au)+r(N) = 3.21 Å]. This indicates an intramolecular non-orbital-based electrostatic attractive interaction between gold(III) center and NMe₂ which is crucial for stabilization of the Au(III) species during Au(I)/Au(III) redox cycle.^[25,26] At this stage, studies to synthesize the cationic complex from 6 were not successful.

We performed a Hammett correlation study to elucidate the electronic impact of electrophiles on the Au(I)/Au(III) cycle (Scheme 2). The Hammett plot was constructed using intermolecular competition between para-substituted iodoarenes (p-CF₃, p-Br, p-Me, p-OMe) and iodobenzene under standard competition conditions. The Hammett plot showed negative slope (R² = 0.97, ρ = −0.66). The negative ρ indicates that oxidative addition of electron-rich aryl iodides is favored by electrophilic Au(I)─NHC (vide infra), which is consistent with the effect of heterobidentate N,P Me-DalPhos ligand.^[16a] This result further indicates unusual preference of Au(I)/Au(III) cycle for oxidative addition of electron-rich electrophiles vs. Pd(0)/Pd(II) catalysis.

Scheme 2.

Hammett correlation in the C─H arylation of 1,3,5-trimethoxybenzene with para-substituted iodoarenes.^{a a}Conditions: 1,3,5-trimethoxybenzene (2.0 equiv), iodobenzene (1.0 equiv), AgNTf₂ (1.0 equiv), NHC-Au-Cl (5 mol%), methanol (0.125 M), 80 °C, 16 h.

Furthermore, although we were unable to isolate the oxidative addition complex with iodobenzene due to its low stability, the oxidative addition Ar─Au(III)─I species was detected by HRMS under standard conditions (see SI).

Large Scale Ligand Synthesis.

To enable broad access to these L-shaped heterobidentate ImPy N,C ligands, we developed large-scale chromatography-free ligand synthesis (Scheme 3). Starting from the commercially-available 2,6-dibromopyridine 7, 6-bromo-2-N,N-dimethylaminopyridine 8 is selectively obtained in quantitative yield by S_NAr displacement with amine in the presence of K₂CO₃. Straightforward lithium-halogen exchange and formylation of 8 readily affords 6-substituted-amino-pinacolaldehydes 9. Condensation with 2,6-diisopropylaniline in EtOH affords imino-pyridine 10, which is purified by recrystallization. Finally, ImPy salt is formed by the reaction with formaldehyde under acidic conditions. It is worth noting that (1) the entire sequence is readily performed on 10 g scale, (2) the synthesis is highly modular enabling incorporation of different amines at the C5 position and wingtips at the N2 position, (3) all steps are performed without chromatographic purification. The rapid access to this novel class of heterobidentate ImPy N,carbene ligands should significantly expand the potential of this ligand platform in Au(I)Au(III) catalysis and organometallic catalysis.

Scheme 3.

Chromatography-free large-scale ligand synthesis.^{a a}Conditions: (a) 2,6-dibromopyridine (1.0 equiv), HNMe₂ (5.0 equiv), K₂CO₃ (1.1 equiv), CH₃CN, 100 °C, 72 h, 98%. (b) n-BuLi (1.5 equiv), DMF (1.2 equiv), THF, −78 °C, 2 h, 75%. (c) 2,6-diisopropylaniline (1.0 equiv), EtOH, 90 °C, 24 h, 76%. (d) (CHO)_n (1.5 equiv), HCl (2.0 equiv), EtOH, 70 °C, 36 h, 80%. See SI for full details.

DFT Studies.

To gain insight into the reactivity of ImPy ligands in Au(I)/(Au(III) catalysis and elucidate key factors for ligand design, extensive DFT studies were conducted. Notably, this study discloses the full catalytic cycle of Au(I)/Au(III)-catalyzed, ligand-enabled coupling for the first time and provides the full insight into DFT mechanism, including oxidative addition, C─H activation, deprotonation, and reductive elimination with the catalytic system ImPyNMe₂AuCl/AgNTf₂. Furthermore, the parallel DFT mechanism computation of control ligands was elucidated and compared with ImPyNMe₂ to explain the success of ligand design for ligand-enabled Au(I)/Au(III) catalysis.

The calculated results of the ImPyNMe₂AuCl catalyzed C─C coupling reaction of 1,3,5-trimethoxybenzene with iodobenzene are in Figure 4. The initial structure of the ImPyNMe₂ ligand is based on its X-ray crystal structure. Initially, the pre-catalyst ImPyNMe₂AuCl undergoes counter-anion exchange with AgNTf₂ to form a more stable ImPyNMe₂AuNTf₂ (IM1A), which is considered as the catalytically active species. Then, NTf₂⁻ in IM1A is substituted by iodobenzene to afford the iodine-coordinated Au^I σ-complex IM2A, from which the oxidative addition of aryl iodide to Au(I) occurs via TS1A with a relative energy of 16.6 kcal/mol, leading to the formation of Au(III)I─aryl intermediate IM3A, which features tetracoordinate Au(III) coordinated with hemilabile NMe₂. Subsequently, Au(III)NTf₂─aryl intermediate IM4A is obtained via the halide abstraction process between IM3A and AgNTf₂ with the release of AgI. The coordination of 1,3,5-trimethoxybenzene to IM4A affords IM5A, which then undergoes C─H activation and deprotonation, assisted by NTf₂⁻ via TS2A (21.7 kcal/mol relative to the reaction coordinate), to form Au(III)─aryl intermediate IM6A.

Figure 4.

Calculated Gibbs energy profile for the gold(I)-catalyzed C─C coupling reaction of 1,3,5-trimethoxybenzene and iodobenzene with ImPyNMe₂Dipp as the ligand of Au(I) complex.

Finally, by carrying out the C─C reductive elimination, IM6A evolves into the final coupling product via TS3A and at the same time the active catalyst is regenerated. TS3A with a relative energy of 26.0 kcal/mol is located as the highest point on the potential energy surface; thus, it is the rate-determining transition state, and the overall barrier of the reaction is calculated to be 29.0 kcal/mol (IM1A to TS3A). Such a moderately high energy barrier can be overcome at the experimental temperature (80 °C).

The calculated results with several other NHC ligands, including ImPyiPr^C5Dipp, ImPyNMe₂^C7Dipp, and IPr, are shown in Figures S1, S2 and S3 in the SI (pages S72–S73). In comparison, it is found that the energy profiles of all the three potential energy surfaces are similar to that shown in Figure 4. Compared with other NHC ligands, ImPyNMe₂AuCl showed much lower energy of key Au(III) intermediates (IM3, IM6) and Au(III) transition state (TS3) than other NHC ligands, which clearly proves that C5-NMe₂ group is essential to stabilize the Au(III) intermediates and the corresponding transition state. Among these Au(III) transition states, the C5-NMe₂ group in ImPyNMe₂ plays a substantial role for stabilizing the rate-determining C─C reductive elimination transition state. As indicated in TS3A, the −NMe₂ group coordinates to the Au(III) center, forming a 16-electron tetra-coordinated complex. In contrast, in TS3B, TS3C, and TS3D, the Au(III) centers present 14-electron tri-coordinated structures due to absence of the −NMe₂ group, and thus destabilizing these transition states.^[27] The overall barrier is 32.7 kcal/mol with ImPyiPr^C5Dipp, 31.1 kcal/mol with ImPyNMe₂^C7Dipp, and 32.4 kcal/mol with IPr. All these barriers are higher in comparison with that for the reaction with ImPyNMe₂, 29.0 kcal/mol (Table 3), accounting for the fact that no product was observed in the experiments when ImPyNMe₂ is replaced. It is expected that these barrier differences originated from the intrinsically different reactivities of NHC ligands, rather than from the DFT method used. In the present work, the chosen functional is B97D, which shows good performance for the transition metal systems in terms of the error spread.^[28]

Table 3.

Optimized transition state geometries for the rate-determining C─C reductive elimination and calculated overall energy barriers using different NHC ligands.

	ImPyNMe2	ImPyiPrC5Dipp	ImPyNMe2C7Dipp	IPr
Transition State
Barrier (kcal/mol)	29.0	32.7	31.1	32.4

From the calculated results, it is clear that the incorporation of −NMe₂ into the C5 position of the ImPy skeleton is of critical importance, which stabilizes the Au(III) complex and thus reduces the overall energy barrier in Au(I)/Au(III) catalysis.

In addition, we calculated the catalytic cycle with P,N ligand Me-DalPhos, which is the single most popular ligand thus far in Au(I)/Au(III) catalysis (Figure S4 in the SI, page S74). The overall barrier of the reaction with Me-DalPhos is found to be 28.2 kcal/mol, which is in accordance with the reported catalytic activity.^[10] Interestingly, compared with mechanism pathway of Me-DalPhos, although slightly higher overall energy barrier is required for ImPyNMe₂, this ImPy NHC ligand showed easier C─H activation and deprotonation (IM5 to TS2) based on DFT calculations, which is consistent with the fact that experimentally no base was required to reach high yields under ImPyNMe₂AuCl-catalyzed arylation conditions. This fact bodes especially well for expanding the scope of the reactions medicated by oxidant-free Au(I)/(Au(III) catalysis.

Overall, the results of DFT computations highlight the importance of C5─N coordination on stabilizing the Au(III) intermediate and outline the key benefits of heterobidentate ImPy N,C ligands for Au(I)/Au(III) catalysis.

In a broader sense, the complementarity and differences of this ImPy N,C ligand platform vs. the previously reported P,C ligands should be noted.^[16-18] The present class of ligands (1) enables steric and electronic tuning that is not possible with phosphine ligands due to the versatility of N-heterocyclic carbenes; (2) base is not required for arylation, which simplifies the reaction conditions and enables to use sensitive substrates (e.g., esters of bioactive products, such as 3o, 3p); (3) an important aspect is complementarity of the ligand systems in that bromides are recovered unchanged from the present ligand system, which enables to tolerate sensitive halides in the system (3d, 3n); (4) NHC ligands are intrinsically more stable to oxidation than phosphines, which offers new opportunities in oxidative gold catalysis; (5) the strong σ-donation of NHCs stabilizes various metals at high oxidation states, which provides unique properties for oxidative catalysis.

Conclusion

In summary, although gold catalysis represents a powerful and one of the most burgeoning fields of molecule activation, the development of Au(I)/Au(III) cycle in the absence of oxidants has had a very limited success due to general lack of ligands that promote oxidative addition to Au(I) and stabilization of Au(III) center.

In this study, we reported a novel class of L-shaped heterobidentate ImPy N,C ligands that enable activation of aryl halides at gold(I). These ligands merge the inherent properties of versatile N-heterocyclic carbene platform with donor stabilization of Au(I) and Au(III) intermediates. This class of heterobidentate ImPy N,C ligands merges hemilabile character of the hard amino group in combination with strong σ-donation of the soft carbene center. Ligand-enabled redox process in arylation of electron-rich arenes with readily available aryl halides under mild, oxidant-free Au(I)/Au(III) conditions has been demonstrated. Detailed characterization and extensive control studies revealed key features for Au(I)/Au(III) redox cycle, wherein the hemilabile nitrogen is placed at the coordinating position of a rigid imidazo[1,5-α]pyridine framework. DFT studies demonstrated crucial stabilization of Au(III) species by the hemilabile nitrogen atom. Large scale, chromatography-free ligand synthesis has been developed, providing operationally-simple and modular access to this class of ligands. The conformational lock prevents nitrogen from rotation and renders the pendant N at the closest position to gold, thus maximizing stabilization. Studies are currently underway to develop even more reactive classes of hemilabile ligands based on this design. Research on expanding the substrate scope to other processes by oxidative gold catalysis, including C─C coupling of electron-rich heterocycles, C─O heteroarylation and C─O carboxyarylation are currently underway to facilitate further reaction development. Preliminary results indicate high reactivity and unique properties of the present system. These processes are currently being optimized and will be reported in due course. Considering the importance of gold catalysis in modern synthesis, we anticipate that this approach will find widespread applications in oxidant-free Au(I)/Au(III) catalysis and enable future advances in ligand design for organometallic reactions.