Rhoda‐Electrocatalyzed Bimetallic C−H Oxygenation by Weak O‐Coordination

Abstract

Rhodium‐electrocatalyzed arene C−H oxygenation by weakly O‐coordinating amides and ketones have been established by bimetallic electrocatalysis. Likewise, diverse dihydrooxazinones were selectively accessed by the judicious choice of current, enabling twofold C−H functionalization. Detailed mechanistic studies by experiment, mass spectroscopy and cyclovoltammetric analysis provided support for an unprecedented electrooxidation‐induced C−H activation by a bimetallic rhodium catalysis manifold.

Rhodium‐electrocatalyzed C−H oxygenations of amides and ketones proved viable under mild conditions via bimetallic electrocatalysis. Mechanistic studies revealed a high‐valent dimeric rhodium electrocatalysis.

Introduction

During the past decade, transition metal‐catalyzed C−H activation has been recognized as a transformative tool in molecular syntheses. ^[1] Phenols featuring an ortho‐substituted carbonyl group constitute important structural motifs of a diversity of bioactive molecules, ranging from natural products to drugs molecules. ^[2] Transition metal‐catalyzed C−H activation by weak chelation assistance provides a straightforward access to the assembly of phenols. ^[3] Pioneering studies with palladium catalysis were accomplished by Dong ^[4] and Rao. ^[5] In the same year, our group reported sustainable ruthenium‐catalyzed C−H oxygenations with a diverse range of weakly coordinating groups. ^[6] In addition, considerable efforts have been devoted to the development of different metal catalysts, along with various oxidants. ^[7] Despite indisputable advances, stoichiometric amounts of strong chemical oxidants, such as (diacetoxyiodo)benzene and K₂S₂O₈, are required, which leads to undesired byproducts.

Electrochemical synthesis has undergone a renaissance in recent years towards environmentally‐benign organic synthesis. ^[8] Significant recent impetus was gained by the merger of electrocatalysis with oxidative C−H activation, thus avoiding the use of toxic and expensive metal oxidants. ^[9] Although transition metal‐catalyzed electrochemically C−H oxygenation have been recently realized, ^[10] this mostly required strong N‐coordination, such as bidentate quinolinyl amides or pyridines, while the very recently devised ruthenium catalysis needed rather costly aryl iodides as additional redox mediators, jeopardizing the inherent atom‐economy. ^[10a] In contrast, the redox direct—mediator‐free—oxidation of homogeneous metal oxygenation catalysts at the anode surface typically called for a divided cell setup.[ ^10e , ^10f , ^10g ] Within our program on electrochemical C−H activation, ^[11] we have now devised mechanistically‐distinct rhoda‐electrocatalyzed C−H oxygenations of weakly‐O‐coordinating amides and ketones (Figure 1). Salient features of our findings include 1) undivided cell without redox mediator, 2) room temperature oxygenations, 3) electricity in lieu of strong oxidants, 4) high Faraday efficiency, 5) twofold electrochemical C−H functionalization towards dihydrooxazinones, ^[12] and 6) detailed mechanistic support for a bimetallic electrocatalysis manifold.

Figure 1.

Rhoda‐electrocatalyzed C−H oxygenation.

Results and Discussion

At the outset of our studies, various reaction conditions were explored for the envisioned redox‐mediator‐free, rhoda‐electrocatalyzed C−H oxygenation of 1 a in an operationally simple undivided cell equipped with a graphite felt (GF) anode and a platinum cathode (Table 1 and Table S1 in the Supporting Information). ^[13] Preliminary experimentation indicated that nBu₄NPF₆ was the optimal additive (Table 1, Entries 1–4). Further studies revealed that the reaction was viable at ambient temperature (Entries 5–6) in a solvent mixture of trifluoroacetic acid (TFA) and trifluoroacetic anhydride (TFAA) (1:1) (Entries 6–8). Control experiments showed that both the rhodium catalyst and electricity were essential (Entries 9–11). During the optimization, we found that NEt₃ could significantly improve the conductivity (Table S1). Thus we employed easily available TFA⋅NEt₃ salt (Entry 12), ^[14] enabling the use of a solvent mixture TFA/TFAA (1:20) without solubility or conductivity problems (Entries 12–13 vs. Entry 8). Here, the highest yield of 82 % was obtained with 2 mA (Entry 13). Interestingly, while [Rh(OAc)₂]₂ showed high catalytic efficacy, the commonly used [RhCp^*Cl₂]₂ or RhCl₃⋅3 H₂O provided unsatisfactory results (Entry 14).

Table 1.

Optimization of the rhoda‐electrocatalyzed C−H oxygenation.^[a]

Entry	Electrolyte	TFA/TFAA	I [mA]	t [h]	T [°C]	Yield [%]
1	CF3COONa	3:1	4	16	50	trace
2	LiClO4	3:1	4	16	50	24
3	nBu4NBF4	3:1	4	16	50	22
4	nBu4NPF6	3:1	4	16	50	28
5	nBu4NPF6	1:1	4	16	50	42
6	nBu4NPF6	1:1	4	16	RT	54
7	nBu4NPF6	1:2	4	16	RT	52
8	nBu4NPF6	1:3[b]	4	16	RT	trace
9	nBu4NPF6	1:1	0	16	RT	ND
10	nBu4NPF6	1:1	4	16	RT	trace[c]
11	nBu4NPF6	1:1	4	16	RT	57[d]
12	TFA⋅NEt3 [e]	1:20	4	8	RT	76[d]
13	TFA⋅NEt3 [e]	1:20	2	15	RT	82[d]
14	TFA⋅NEt3 [e]	1:20	2	15	RT	ND[f]
15	nBu4NPF6	3:1	4	16	50	28[g]
16	TFA⋅NEt3 [e]	1:20	2	15	RT	52[h]

[a] 1 a (0.5 mmol), [Rh(OAc)₂]₂ (2.5 mol %), electrolyte (0.10 M), TFA/TFAA, GF anode (10 mm × 10 mm × 6 mm), Pt cathode (10 mm × 15 mm × 0.125 mm). [b] Poor conductivity due to poor solubility of electrolyte. [c] Without [Rh(OAc)₂]₂. [d] Under N₂. [e] TFA⋅NEt₃ (0.33 M). [f] [RhCp^*Cl₂]₂ or RhCl₃.3 H₂O instead of [Rh(OAc)₂]₂ as the catalyst. [g] Ru(OAc)₂(p‐cymene) (5 mol %) as the catalyst. [h] [Rh(OPiv)₂]₂ as the catalyst. Piv=pivalate.

With the optimized reaction conditions in hand, we next examined the viable substrate scope of the rhodium‐catalyzed electrochemical C−H oxygenation with various weakly coordinating Weinreb amides 1 (Scheme 1 a). Electron‐rich as well as electron‐deficient Weinreb amides 1 a–1 p were amenable to the rhodium‐catalyzed electrochemical catalysis. Notably, a diverse array of valuable functional groups, including ester (2 f), halogen (2 g–2 j) and ketone (2 k) groups, were tolerated by the electrocatalysis, highlighting a notable potential for further late‐stage diversification. It is noteworthy that the rhoda‐electrocatalysis was not limited to Weinreb amides 1. Indeed, differently substituted amides 3 a–m were also efficiently converted into the corresponding oxygenated amides 4 a–m with remarkable catalytic efficiency (Scheme 1 b).

Scheme 1.

Rhoda‐electrocatalyzed C−H oxygenation of amides 1 and 3.

The outstanding robustness of the rhoda‐electrocatalyzed C−H oxygenation was further highlighted by its ability to transform more challenging ketones 5 (Scheme 2).

Scheme 2.

Rhoda‐electrocatalyzed C−H oxygenation of ketones 5.

During our optimization studies, small amounts of N‐demethylation product 7 a could often be isolated (Scheme 3 a). Hence, we rationalized that it was formed through a cascade C−H oxygenation, along with Shono‐type oxidation to generate Int2. Further experimentation at Q≥4 F mol⁻¹ revealed small amounts of an interesting dihydrooxazinone 8 a. Through rational design of the work up process with NEt₃ as non‐nucleophilic base, Int2 could be exclusively converted to the valuable dihydrooxazinone ^[12] 8 a (Scheme 3 a, blue path). With the optimized reaction conditions, we explored the substrate scope of the rhoda‐electrocatalyzed cascade reaction with diverse amides 1 and 3 to assemble various dihydrooxazinones 8 and 9 (Scheme 3 b). The tolerance of ester (8 f) and halogen (8 g, 8 i, 8 p, 9 d–e, 9 h) substitute provides an invaluable asset in terms of late‐stage modifications.

Scheme 3.

Rhoda‐electrocatalyzed cascade reaction for the synthesis of dihydrooxazinones.

Promysalin is a Pseudomonad secondary metabolite that exhibits narrow‐spectrum antibacterial activity, originally isolated from the rhizosphere (Scheme 4 a). ^[15] In 2016, Wuest reported the total synthesis of Promysalin analogues. ^[15b] The key proline‐salicylate fragment in Promysalin inspired us to apply our rhoda‐electrocatalyzed C−H oxygenation to the synthesis of various substituted proline‐salicylates (4 n–p) without any protection and deprotection of the phenol motif (Scheme 4 b). Next, we studied the efficiency of the catalysis with bromo analog 3 p through the gram‐scale synthesis with only 0.5 mol % of [Rh(OAc)₂]₂ with a turnover number of 166 based on the rhodium‐dimer, along with the solvent TFAA being recovered by simple distillation, highlighting the practical potential of this catalysis (Scheme 4 c).

Scheme 4.

Application to the synthesis of the analogous fragments of Promysalin.

Given the versatility of the redox mediator‐free electrochemical C−H oxygenation, we became attracted to probing its mode of action. Reactions conducted with isotopically labeled [D]₁‐TFA did not lead to any H/D scrambling. ^[13] Kinetic studies provided strong support for a fast C−H metalation with a minor kinetic isotope effect (KIE) of k _H/k _D≈1.1 (Scheme 5 a). Then, we explored the current dependence of the performance within a range from 2.0 to 5.0 mA, being indicative of a turnover‐limiting electron transfer step in the current region of 2.0–4.0 mA (Scheme 5 b). Beyond 4.0 mA the reaction rate did not increase significantly, being suggestive of a switch in the turnover‐limiting step.

Scheme 5.

Key mechanistic experiments.

Next, we turned our attention to investigate the ligand exchange effect at rhodium. Somewhat surprisingly, [Rh(OTFA)₂]₂ did not show any catalytic reactivity for this electrochemical catalysis (Scheme 6 a). Mass spectroscopy studies revealed a rapid ‐OAc/‐OTFA exchange in the solvent mixture TFA/TFAA (1:1), while in TFA/TFAA (1:20) this exchange was slow, ^[13] which may be caused by the low concentration of available TFA. These observations are matched with the low catalytic activity in TFA/TFAA (1:1) (Table 1), suggesting the ligand exchange to be harmful to the catalyst activity. While stepwise electrochemical oxygenation delineated an oxidation‐induced C−H activation and oxygenation, mass spectroscopic analysis of the electrolyzed mixture showed a plausible active intermediate [Rh(OAc)₂(OTFA)]₂, highlighting the dimeric form of the catalyst (Scheme 6 b). This was further experimentally supported by the inactivity of the monomeric Rh(OAc)₃ (Scheme 6 c).

Scheme 6.

Mechanistic studies on rhoda‐electrocatalyzed oxygenation.

Furthermore, we probed an electrochemical oxidation‐induced C−H activation by means of cyclovoltammetric analysis (Figure 2). First, the oxidation potential of amide 1 a was higher than [Rh(OAc)₂]₂ in TFA/TFAA (1:1), suggesting an oxidation‐induced C−H activation regime (Figure 2 a). In TFA/TFAA (1:20), [Rh(OAc)₂]₂ featured an onset potential of E _onset=0.8 V vs. Ag/AgCl with low current, due to the poor solubility of [Rh(OAc)₂]₂ (Figure 2 b, dark). In TFA/TFAA (1:1), the oxidation peak shifts to higher oxidation potentials, this shift is likely caused by gradually ‐OAc/‐OTFA ligand exchange, implying the more ‐OTFA substituents on rhodium, the higher oxidation potential exhibited (Figure 2 b, blue). In line with this observation, [Rh(OTFA)₂]₂ revealed a high onset potential of E _onset=1.3 V vs. Ag/AgCl, which further rationalized as to why [Rh(OTFA)₂]₂ was not catalytically competent (Figure 2 b, red).

Figure 2.

Cyclic voltammetry measurements in TFA/TFAA with 0.15 M TFA⋅NEt₃ and 5 mM substrate at RT with a scan rate of 100 mV s⁻¹: (a) in TFA/TFAA (1:1); (b) dark, [Rh(OAc)₂]₂ in TFA/TFAA (1:20), stirred 10 min, poor solubility; blue, [Rh(OAc)₂]₂ in TFA/TFAA (1:1), measured at stirring 5, 30 and 60 min; red, [Rh(OTFA)₂]₂ in TFA/TFAA (1:1).

Likewise, the catalyst's mode of action was investigated by means of DFT studies at the B3LYP‐D4/6‐311++G**,Rh/SDD+SMD(DCE)//B3LYP‐D3/6‐31G**,Rh/SDD level of theory. ^[13] Thus, several bimetallic rhodium complexes were probed, namely monocationic (Figure S9) or dicationic (Figure 3) Rh^III‐Rh^III as well as monocationic Rh^II‐Rh^III (Figure S11). According to our findings, a C−H rhodation via dicationic Rh^III‐Rh^III complex was identified as the most plausible pathway with an activation barrier of 25.1 kcal mol⁻¹.

Figure 3.

Computed relative Gibbs free energies (ΔG _298.15) in kcal mol⁻¹ for the C−H activation elementary step at the B3LYP‐D4/6‐311++G**,Rh/SDD+SMD(DCE)//B3LYP‐D3BJ/6‐31G**,Rh/SDD level of theory for a dicationic Rh^III‐Rh^III complex. In the computed transition state structure, non‐relevant hydrogens were omitted for clarity.

On the basis of our mechanistic studies, we propose a plausible catalytic cycle to be initiated by facile electrochemical oxidation of [Rh(OAc)₂]₂ to generate the rhodium(III)‐rhodium(II) species 10 (Figure 4). A subsequent anodic oxidation generates the catalytically competent bimetallic rhodium(III) species 11. Notably, the bimetallic nature of the electrocatalysis allows for effective direct oxidations in the absence of a redox mediator. Then, isomerization and substrate coordination occur to deliver intermediate 12. Thereafter, the high‐valent bimetallic rhodium‐induced C−H activation takes place, while, reductive elimination and decoordination deliver the trifluoroacetate ester 16. Finally, anodic oxidation regenerates the catalyst 10, with an alternative two‐electron oxidation pathway being depicted in Scheme S12. ^[13]

Figure 4.

Proposed catalytic cycle.

Conclusion

In conclusion, we have reported on mechanistically‐distinct rhoda‐electrocatalysis for C−H oxygenations of synthetically useful amides and ketones by challenging weak O‐coordination. In an undivided cell, an easily accessible ionic liquid and mild reaction conditions set the stage for a operationally‐friendly C−H oxygenation in the absence of redox mediators. By adjusting the current, valuable dihydrooxazinones could be selectively assembled by double C−H functionalization. The application to the late‐stage assembly of Promysalin analogue fragments with a low catalyst loading, as well as gram‐scale bimetallic electrocatalysis accounts for the considerable practical potential. Detailed mechanistic studies revealed an oxidation‐induced C−H activation by a bimetallic rhodium catalysis manifold.

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

X. Tan, L. Massignan, X. Hou, J. Frey, J. C. A. Oliveira, M. N. Hussain, L. Ackermann, Angew. Chem. Int. Ed. 2021, 60, 13264.