Mediator-Enabled Electrocatalysis with Ligandless Copper for Anaerobic Chan–Lam Coupling Reactions

Abstract

Simple copper salts serve as catalysts to effect C–X bond-forming reactions in some of the most utilized transformations in synthesis, including the oxidative coupling of aryl boronic acids and amines. However, these Chan–Lam coupling reactions have historically relied on chemical oxidants that limit their applicability beyond small-scale synthesis. Despite the success of replacing strong chemical oxidants with electrochemistry for a variety of metal-catalyzed processes, electrooxidative reactions with ligandless copper catalysts are plagued by slow electron-transfer kinetics, irreversible copper plating, and competitive substrate oxidation. Herein, we report the implementation of substoichiometric quantities of redox mediators to address limitations to Cu-catalyzed electrosynthesis. Mechanistic studies reveal that mediators serve multiple roles by (i) rapidly oxidizing low-valent Cu intermediates, (ii) stripping Cu metal from the cathode to regenerate the catalyst and reveal the active Pt surface for proton reduction, and (iii) providing anodic overcharge protection to prevent substrate oxidation. This strategy is applied to Chan–Lam coupling of aryl-, heteroaryl-, and alkylamines with arylboronic acids in the absence of chemical oxidants. Couplings under these electrochemical conditions occur with higher yields and shorter reaction times than conventional reactions in air and provide complementary substrate reactivity.

Graphical Abstract

INTRODUCTION

Reactions that form challenging C–X bonds, including C–N, C–F, C–O, and C–CF₃, are critically important for the discovery and design of drug-related molecules.^1,2 These coupling reactions have long relied on the judicious design of ligands that accelerate rates of reductive elimination from organometallic intermediates.^3–9 Recent approaches leverage the reactivity of high-valent organometallic complexes bearing inexpensive ligands to effect similar reactions.^10–14 However, accessing these high-valent organometallic complexes typically requires oxidation with superstoichiometric quantities of peroxides,^15,16 N-fluoropyridiniums,^12,17–19 xenon difluoride,^20,21 silver salts,²² hypervalent iodide,²³ or pressurized oxygen.^24,25 Efforts to mitigate the hazards and wastes associated with strong chemical oxidants have focused on new strategies for electron transfer (ET) and oxidation, including photoredox catalysis^26–30 and electrosynthesis.^31–40 Electrochemistry is particularly well-suited for net oxidative (or reductive) organic transformations because reactions can be designed for the anode while providing a benign reagent, like a proton, for consumption at the opposite electrode. In doing so, electrical energy and a proton can replace energetic chemical oxidants.⁴¹

These advantages have motivated the recent development of a wide range of catalytic electrooxidative methodologies that employ complexes of Co,^42,43 Fe,^{44, Ni,45,46} Ru,⁴⁷ Mn,^48–50 Rh,⁵¹ Ir,⁵² Cu,^53–58 and Pd.⁵⁹ However, catalysts for all of these reactions are metal complexes with strongly chelating ligands. While such ligands serve to modulate redox potentials, improve ET kinetics, and inhibit metal aggregation, some of the most common oxidative transformations such as Pd-catalyzed C–H oxygenation or Cu-catalyzed coupling require “ligandless” metal halides or acetates as catalysts.^60–72 Replacing prohibitive chemical oxidants in these transformations through electrooxidation remains an unmet goal because ligandless metal salts can undergo competing reduction and plating at the cathode (Figure 1a).

Figure 1.

(a) Challenges associated with electrooxidative, ligandless transformations. (b) Redox mediator-enabled ligandless copper Chan–Lam coupling (this work).

Separation of the anodic reaction from the cathodic reaction with a divided cell is one strategy for protecting the homogeneous catalyst. While the use of divided cells poses some limitations to the general synthetic community, the approach has been successfully applied to electrooxidative Wacker^73,74 or C–H acetoxylation^75–77 reactions catalyzed by simple Pd(OAc)₂ salts. Copper salts can similarly be confined to an anodic chamber to protect from reductive plating but have very poor kinetics for electrooxidation, allowing for competitive oxidation of substrates. Because of these combined challenges, there are no examples of electrochemical reactions catalyzed by simple Cu salts in the absence of added chemical oxidants.

This work details the development of electrooxidative C–N coupling reactions in an undivided cell with simple Cu salts as catalyst and a redox mediator that serves to (i) oxidize Cu^I to the active Cu^II form, (ii) strip plated Cu⁰ to regenerate the homogeneous Cu catalyst, and (iii) protect substrates from oxidation. This general approach for protecting ligandless metal catalysts with a mediator is successfully applied to the Chan–Lam cross-coupling of (hetero)aryl or alkyl amines with aryl boronic acids (Figure 1b). The electrochemical reactions are scalable and typically occur with higher rates than the conventional reactions under air. The battery-inspired mediator is critically important, as electrochemical reactions in its absence fail to generate product even in a divided cell.

Previous efforts to replace chemical oxidants with alternative redox mechanisms have addressed some but not all of the listed challenges. Chan–Lam coupling has been reported using photoredox cocatalysts⁷⁸ or electrochemical systems with Cu electrodes as both the anode and cathode to sustain the homogeneous catalyst by continuously stripping Cu metal.^55,56 However, all reported systems rely on molecular oxygen and are low-yielding in the absence of exogenous oxidants.

Application of a mediator-assisted strategy toward the development of an electrochemical Chan–Lam coupling reaction addresses the long-standing challenge of eliminating oxygen or peroxide additives from this important transformation.^{61,65–67,79,80} The requirement for hazardous oxidants has precluded the use of Chan–Lam couplings on a large scale, despite their ubiquitous application in drug discovery and small-scale synthesis.^24,81–84 The limited scale on which Chan–Lam reactions are currently performed is particularly frustrating because the reactions offer an easily accessible approach to forming important C–N bonds with an inexpensive copper salt as catalyst and abundant substrates. As such, this work represents not only a rare example of electrocatalysis with simple copper salts^85–87 in an undivided cell but also—to the best of our knowledge—the first example of Chan–Lam coupling to forge C–N bonds in the absence of a stoichiometric oxidant.

RESULTS AND DISCUSSION

The importance of the Chan–Lam coupling reaction in organic synthesis has motivated detailed mechanistic studies that guided our reaction design.^80,88–90 Previous reports implicate a bimolecular disproportionation of Cu^II as a critical step to forming a putative Cu^III intermediate that undergoes bond-forming C–N reductive elimination. As such, we sought to establish an electrochemical system that could maintain high concentrations of Cu^II without molecular oxygen or other chemical oxidants. We envisioned three challenges that would have to be addressed. First, Cu^II is easily reduced at an onset of −0.8 V (vs Fc/Fc⁺) and outcompetes the desired cathodic reaction of proton reduction in nonaqueous solvents (Figure 2a). The resulting Cu^I species can be further reduced to plate Cu⁰ at the cathode or can undergo disproportionation with a second Cu^I species to generate Cu⁰ metal and Cu^II, which can be reduced further. Second, electrooxidation of ligandless Cu^I to the desired Cu^II species is kinetically slow.⁵⁵ The cyclic voltammogram (CV) of Cu(OAc)₂ reproduced in Figure 2a reveals a reduction of Cu^II but no subsequent reoxidation on the return scan (black trace). The only observed anodic peak is the result of Cu⁰ stripping that had plated on the electrode during the reductive sweep. Third, the slow and poorly defined oxidation of Cu^I allows competitive oxidation and degradation of the amine substrates (red trace).

Figure 2.

(a) CV of Cu(OAc)₂ (black, 20 mM) and aniline (red, 100 mM). (b) Targeted synergy of redox mediators and copper electrocatalysts. CV conditions: 0.1 M KPF₆ in MeCN, 100 mV/s scan rate, room temperature, glassy carbon working electrode (WE) and Pt counter electrode (CE).

These limitations directed our evaluation of redox-active compounds that can assist electrochemically slow redox processes of simple metal salts. Recently, redox mediators have come to the forefront of electrocatalysis as a means to mediate homogeneous ET, most commonly with an organic substrate.^91–95 Indirect redox of organic substrates does not rely on ET directly at a heterogeneous surface, which often leads to degradation. In analogy to these strategies, we hypothesized that mediators could be employed to promote the redox processes of a Cu catalyst, which similarly suffers from problematic ET reactions at heterogeneous surfaces. With properly tuned redox potentials and ET kinetics, the mediator can undergo rapid oxidation at the anode and (i) homogeneously oxidize Cu^I to Cu^II, (ii) oxidatively strip plated Cu⁰ to regenerate catalyst, and (iii) protect amine substrates from oxidative degradation (Figure 2b).

Redox Mediator Design.

We first evaluated a wide range of conditions that are commonly employed for Cu-catalyzed Chan–Lam coupling reactions but in an inert atmosphere and without a redox mediator (Figure 3). As expected, electrolysis of these solutions in both undivided and divided cells resulted in trace yields (<5%). Our experience merging battery chemistries with electrosynthesis for overcharge protection and redox shuttling further guided the reaction design.⁴¹ Specifically, Chan–Lam reactions were conducted with redox shuttles that undergo oxidation between −0.6 and +0.2 V (vs Fc/Fc⁺). Mediators with potentials in this range can readily oxidize Cu^I and Cu⁰ (E_1/2 < −0.8 V) but are too mild to oxidize amine substrates (E_1/2 > +0.5 V).

Figure 3.

Evaluation of ferrocenyl mediators for the electrocatalytic reaction. Yields were determined by calibrated GC analysis with dodecane as an internal standard.

Mediator-assisted couplings were conducted with 20 mol % of the mediator. Of the tested additives (see Figure S1 in the Supporting Information for a complete list of evaluated mediators), only ferrocene-based mediators were compatible with the reaction conditions. Reactions with ferrocene (Fc) as a mediator dramatically improved product yields (47%) compared to any nonmediated electrolysis reaction. Similar or better yields were obtained from reactions conducted with less-oxidizing mediators (Fc-1 through Fc-3) with potentials between 0 and −0.15 V (vs Fc/Fc⁺). However, the benefit of the mediator was lost when the redox potential of the ferrocenyl derivative was lowered to more negative potentials (Fc-4). Low yields observed from reactions with the weak oxidant Fc-4 likely stem from the mediator’s inability to readily oxidize Cu^I. Reactions with mediators that are more oxidizing than Fc (Fc-5 through Fc-8) occurred with complete consumption of the aniline but low yields. These results suggest that mediators with potentials above +0.1 V (vs Fc/Fc⁺) allow for competitive oxidation of the aniline (E_onset = 0.4 V).

Despite these promising yields, reactions with ferrocene failed to reach full conversion regardless of the electron equivalents added, which suggests that degenerate redox shuttling occurs between the electrodes. Rather than relying on electrochemistry to generate significant quantities of the oxidized mediator, we instead employed the preoxidized mediator as the starting material. Reactions with 20 mol % ferrocenium hexafluorophosphate (Fc⁺) occurred with complete conversion, and the coupled product was isolated in 81% yield. Reactions with oxidized forms of mediators Fc-1 through Fc-3 similarly formed coupled products in high yields, but the unfunctionalized Fc⁺ was selected as the reaction mediator because of its commercial availability, ease of handling, and synthetic accessibility.

Reaction Development.

With the standard conditions summarized in Table 1, arylamine products could be prepared in high yields with an inexpensive mediator and catalyst in just 5 h under air-free conditions and constant current electrolysis in an undivided cell (entry 1). Control experiments reveal that Fc⁺, Et₃N, NaOAc, Cu(OAc)₂, and electrochemistry are all required for products to be formed under nitrogen (entries 2–6). While reactions can be performed with a variety of anodic materials (entries 7 and 8), reactions with Pt cathodes were the highest yielding. The observed improvement likely stems from decreased overpotentials for proton reduction at Pt compared to reduction at alternative cathodic materials.^96–100

Table 1.

Reaction Development^a,b

While the transformation is reported under inert conditions to highlight the use of electrochemistry as the means for oxidative turnover, reaction setup and electrolysis can be performed in air (entry 9). The slightly lower yield likely stems from degenerate redox events in the presence of air. An added benefit to the electrochemical reaction is the decreased reaction time compared to conventional reactions in air. Reactions in air without both electrochemistry and mediator formed products in just 8% yield (entry 10). Surprisingly, the mediator proved beneficial to even the nonelectrochemical reactions where yields were improved from 8% to 34% in the same timeframe (entry 11). Collectively, these results highlight the importance of both electrochemistry and redox mediators for Cu-catalyzed Chan–Lam reactions.

Substrate Scope.

The developed methodology was applied to couplings of aryl-, heteroaryl-, and alkylamines with arylboronic acids (Chart 1). Coupled products were isolated in good yields following electrooxidative reactions under inert conditions for arylamines with both electron-donating and electron-withdrawing substituents. Steric hindrance ortho to nitrogen had little effect on product yields (5–7). Electron-rich alkylamines that were susceptible to oxidative degradation underwent coupling but were more sensitive to the steric environment around nitrogen. While products of reactions with primary amines were isolated in good yields, yields from reactions with secondary amines were lower (14 and 15 vs 18 and 19). A similar effect was observed with hindered primary amines such as the valine derivative (17) that underwent coupling in reduced yields. No racemization was observed for this amino acid derivative as well as for the enantiopure benzylamine 16 under the mild reaction conditions. Coupling reactions of heteroarylamines occurred with lower conversion than with arylamines, but acceptable yields of products could be isolated (20–22). The lower reactivity likely stemmed from competing coordination to the Cu catalyst, but yields could be improved with longer electrolysis to pass additional Faradaic equivalents. Finally, the standard reaction could be easily scaled 100-fold using a media bottle as the undivided electrochemical cell. Under air-free conditions, compound 3 was formed as a single product in over 85% GC yield and was isolated in 72% yield (3.7 g).

Chart 1. Substrate Scope^h.

^aUnder air, no electrochemistry. ^bUnder air, without Fc⁺, no electrochemistry. ^c40 °C. ^dReaction was performed at double the concentration. ^eReaction was performed at triple the concentration. ^fWithout NEt₃. ^gFcMe₂ was used instead of Fc⁺. Minor variations to the applied current or electron equivalents for specific substrates are detailed in the Supporting Information. ^hThe reported yields are isolated yields from reactions of 0.30 mmol of amine and 0.45 mmol of boronic acid.

We next evaluated the scope in arylboronic acid under the anaerobic electrochemical conditions. Unlike reactions conducted with oxygen, no oxygenation or dimerization of the boronic acid was observed, and reactions generally form the target compound as a single product.^66,101 Classically, Chan–Lam reactions are the highest yielding with electron-rich arylboronic acids and are more challenging with electron-deficient substrates.^67,102 The opposite trend in reactivity was observed with this electrochemical methodology. Substrates with electron-withdrawing groups were coupled in the highest yields (31–33), while couplings of substrates with strongly donating groups para to the boronic acid were more challenging (26). Reactions reverted to the expected reactivity patterns when Chan–Lam coupling was performed under conventional conditions without electrochemistry. As an example, the low-yielding electrochemical reaction of electron-rich 26 (31%) improved to 72% when in air with Fc⁺. In contrast, electrochemistry was essential for coupling of the electron-deficient substrate to form 33 (88% yield vs 15% in air and Fc⁺).

The investigation of the reaction scope reveals that the developed protocol is not simply an air-free alternative to Chan–Lam coupling reactions. Rather, this methodology provides complementary reactivity for reactions of challenging substrates that have previously been low yielding.^67,102 Additionally, electrochemical reactions are higher yielding than those performed in air alone. Finally, these studies reveal that the mediator has a beneficial impact on yields of nonelectrochemical reactions performed in air (see examples 3, 14, 20, 26, and 28).

Role of the Mediator.

The dramatic impact of the redox mediator on the success of the electrochemical Chan–Lam reaction inspired mechanistic studies that could provide insights for the design and implementation of mediators in other challenging electrocatalytic methodologies. We first evaluated the role of Fc⁺ in homogeneous oxidation of low-valent Cu salts to regenerate the active Cu^II species. CVs of Cu^I(OAc)—the Cu species generated after C–N bond formation—scanned from 0 to −1.7 V revealed a single reduction event with a peak current at −1.5 V (Figure 4, red trace). This current response is consistent with Cu^I/0 reduction. The only oxidation event on the return scan to 0 V was the stripping of plated Cu⁰ indicated by the non-Nernstian peak centered at −0.45 V. Notably, there is no second redox event that can be attributed to oxidation of Cu^I. Additional evidence that Cu^I is not electrochemically oxidized was provided by analysis of the baseline current on the initial reductive sweep of the Cu^I-containing solution. Specifically, Cu^I should be unstable at the starting potential of the CV sweep (+0 V), and a positive baseline current would be observed for electrochemical oxidation of the bulk Cu^I species in solution. Additionally, any Cu^II that was generated at these high potentials would be reduced at the Cu^II/I couple with an onset of −0.8 V. However, no current was observed at either of these potential ranges, supporting a lack of electrochemical activity for the Cu^I salt under the reaction conditions.

Figure 4.

CVs probing Cu(I) oxidation by Fc⁺ with Cu(OAc) (red), Cu(OAc) with Fc⁺ (black), and Fc⁺ (blue). CV conditions: 20 mM in MeCN with 0.1 M KPF₆ as electrolyte and 100 mV/s scan rate at room temperature using glassy carbon WE and Pt CE. Potentials were calibrated with Fc as an internal reference.

CV scans after the addition of 1 equiv of Fc⁺ to the same solution of Cu^I(OAc) (Figure 4, black) matched those of Cu^II(OAc)₂ and revealed a Cu^II/I reduction that was not observed from CVs of Cu^I alone. Additionally, the baseline current above the potential of the Fc/Fc⁺ was positive, which indicates that the added Fc⁺ mediator was reduced to Fc and was undergoing oxidation at high potentials. For comparison, no oxidation was observed at high potentials in CVs of pure Fc⁺ (blue trace). Together, these data revealed that Fc⁺ rapidly oxidizes Cu^I to Cu^II under the reaction conditions.

We next analyzed deposits on the cathodes from reactions with and without added mediator. Visually, Pt cathodes from high-yielding reactions containing Fc⁺ appeared unchanged from before electrolysis (Figure 5a). In contrast, an even coating of metallic copper was visible on cathodes from reactions performed without Fc⁺ (Figure 5b). Elemental analysis by scanning electron microscopy with energy dispersive X-rays (EDX-SEM) of these cathodic surfaces confirmed the visual observations and underscored the dramatic impact of the mediator on maintaining a homogeneous catalyst. Images reproduced in Figure 5 reveal either a pristine Pt surface from reactions with added Fc⁺ or complete surface coverage by Cu⁰ in the absence of the mediator. Coating of the Pt surface only further diminishes the selectivity for the desired proton-reduction reaction over Cu plating. Finally, we demonstrated that plated cathodes were stripped of Cu deposits within seconds of submerging the electrodes in a solution of Fc⁺. CVs of the resulting solution were consistent with the conversion of Fc⁺ to Fc and the concomitant formation of Cu^II (Figure S2 in the Supporting Information). Catalyst recovery by oxidative stripping represents degenerate redox events: electrochemical Cu plating paired with Fc oxidation followed by back-electron transfer during the oxidative stripping of Cu⁰ with Fc⁺. It is for this reason that reactions require 4 electron equiv instead of the theoretical 2 F/mol for high yields. However, electrolysis beyond the theoretical capacity does not irreversibly waste electrons because the electrons are simply shuttled through the degenerate pathway for catalyst recovery.⁴¹ Together, these data implicate the mediator’s role in stripping Cu metal deposits from the cathode to maintain an active Pt surface for proton reduction and to regenerate catalytic Cu^II salts.

Figure 5.

(a) Photo and EDX-SEM image of Pt electrode from a reaction containing Fc⁺ showing elemental platinum in red. (b) Photo and EDX-SEM image of a Pt electrode from a reaction without Fc⁺ showing elemental copper in yellow.

Our final investigation into the function of the mediator targeted its role on preventing oxidation of amine substrates. Reactions without Fc⁺ form low yields of coupled product despite complete conversion of the amine substrate. In contrast, yields from reactions with the mediator closely mirror substrate conversion. To gain insight into the oxidative half-cell reactions, we performed mediated and unmediated Chan–Lam couplings with a Ag/Ag⁺ quasi-reference electrode to monitor the absolute operating potential of the anode (Figure 6a). Anodic reactions in solutions containing Fc⁺ occurred at +0.2 V (black trace), which is a potential consistent with oxidation of Fc (Figure 6b, black). In contrast, the anodic operating potential during unmediated reactions was +0.8 V (red trace). This high voltage is indicative of amine oxidation (Figure 6b, red) and provides a rational for the observed high conversions with low product yields under unmediated conditions.

Figure 6.

(a) Anodic voltage profile of a standard reaction with Fc⁺ (black) and a reaction without Fc⁺ (red). (b) CVs of anodic redox events of a standard reaction mixture (red) and of pure Fc⁺ (black). Conditions: 100 mM in MeCN with 0.1 M KPF₆ as electrolyte and 100 mV/s scan rate at 60 °C using glassy carbon WE and Pt CE. Potentials calibrated to Fc as an internal standard.

CONCLUSION

In summary, this work details a strategy for performing electrocatalytic synthetic transformations with catalytic quantities of ligandless copper salts and ferrocenium as an ET shuttle. Mechanistic studies reveal that the mediator serves multiple key roles during electrolysis, including as (i) an electrochemically generated oxidant to maintain high Cu^II concentrations, (ii) a stripping agent to regenerate active Cu catalyst and reveal the active Pt surface for proton reduction, and (iii) an overcharge protector to preclude undesirable anodic reactions. This strategy is applied to Chan–Lam coupling reactions of aryl-, heteroaryl-, and alkylamines with arylboronic acids in the absence of chemical oxidants to form coupled products in higher yields and with shorter reaction times than conventional reactions in air alone. We anticipate that this mediator-assisted electrocatalysis will be applicable to a wide range of oxidative transformations that utilize inexpensive catalysts of metal salts that have slow ET kinetics or are susceptible to reductive plating.