Iodine–Iodine Cooperation Enables Metal-Free C–N Bond-Forming Electrocatalysis via Isolable Iodanyl Radicals

Abstract

Small molecule redox mediators convey interfacial electron transfer events into bulk solution and can enable diverse substrate activation mechanisms in synthetic electrocatalysis. Here we report that 1,2-diiodo-4,5-dimethoxybenzene is an efficient electrocatalyst for C–H/E–H coupling that operates at as low as 0.5 mol% catalyst loading. Spectroscopic, crystallographic, and computational results indicate a critical role for a three-electron I–I bonding interaction in stabilizing an iodanyl radical intermediate (i.e., formally I(II) species). As a result, the optimized catalyst operates at more than 100 mV lower potential than the related monoiodide catalyst 4-iodoanisole, which results in improved product yield, higher Faradaic efficiency, and expanded substrate scope. The isolated iodanyl radical is chemically competent in C–N bond formation. These results represent the first examples of substrate functionalization at a well-defined I(II) derivative and bona fide iodanyl radical catalysis and demonstrate one-electron pathways as a mechanistic alternative to canonical two-electron hypervalent iodine mechanisms. The observation establishes I–I redox cooperation as a new design concept for the development of metal-free redox mediators.

Graphical Abstract

INTRODUCTION

The development of indirect electrochemical mediators, which are redox-active small molecules that participate in well-defined interfacial electron transfer (ET) and convey the resulting electron or hole equivalents into the bulk phase, has powerfully enabled the development of organic electrocatalysis.^{1, 2} Identification of new mediators that engage in diverse modes of substrate activation and that can aggregate the multiple electron equivalents needed for the two-electron bond-making processes in organic synthesis, provide the opportunity to marry interfacial electron transfer with an array of synthetic transformations. Hypervalent iodine compounds are a broadly deployed class of chemoselective oxidants,^3–7 and the potential to utilize aryl iodides as indirect electrochemical mediators (i.e., electrocatalysts) has garnered significant attention (Figure 1).^8–12 Thus far, high catalyst loading, limited substrate scope, and the high onset potential for anodic oxidation has stymied development of hypervalent iodine electrocatalysis and largely limited application of these mediators to ex cell transformations.^{13, 14}

Figure 1.

Strategies for hypervalent iodine electrochemistry, either multiple electron oxidation to form I(III) species or single electron oxidation for iodanyl radical electrocatalysis.

Organic hypervalent I(III) and I(V) reagents are commonly encountered oxidants in fine-chemical synthesis that operate via selective two-electron oxidation-reduction processes. The potential role of iodanyl radicals (i.e. formally I(II) species) in substrate functionalization chemistry is far less explored.^15–17 Recently, transient iodanyl radicals have been proposed as intermediates in the photochemistry of I(III) compounds¹⁸ and as intermediates in aerobic and electrochemical syntheses of I(III) derivatives.^{11, 18–20} For example, in 2020 we reported an electrochemical C–H / N–H coupling catalyzed by iodoanisole that was proposed to proceed via carboxylate-stabilized iodanyl radicals.¹¹ While increasingly invoked, the complete lack of isolable organic I(II) compounds has prevented interrogation of potential reactions of these open-shell species towards substrates.

Inspired by 1) the reversible electrochemistry of veratrole derivatives (i.e., dimethoxybenzenes)²¹ and 2) the delocalized I–I interactions in σ-aromatic C₆I₆²⁺,²² we identified 1,2-diiodo-4,5-dimethoxybenzene (1a) as a highly efficient and robust catalyst for a suite of C-H functionalization reactions at catalyst loadings as low as 0.5 mol%. Detailed electrochemical and in situ spectroscopic experiments indicate that these reactions are mediated by the iodanyl radical generated by one-electron oxidation of 1a. Synthesis and isolation of the iodanyl radical (1a⁺) enabled complete spectroscopic and crystallographic characterization. The isolated iodanyl radical is chemically competent as an intermediate in C–H functionalization. These observations raise the specter of I(I/II) catalytic cycle involving direct substrate engagement by an iodanyl radical, which contrasts with traditional two-electron hypervalent iodine cycles.

EXPERIMENTAL

Representative procedures are presented below; for detailed descriptions of materials, methods, synthetic procedures for starting materials, and characterization data, see Supporting Information.

Synthesis of 1,2-diiodo-4,5-dimethoxybenzene radical cation (1a⁺).

A 100-mL round bottom flask charged with 1a (184 mg, 0.472 mmol, 1.00 equiv.), bis(trifluoroacetoxy)iodobenzene (PIFA) (104 mg, 0.242 mmol, 0.513 equiv.), and CH₂Cl₂ (10 mL). The reaction mixture was cooled −20 °C. BF3×OEt2 (0.06 mL, 0.486 mmol, 1.03 equiv.) was added dropwise resulting in an immediate blue color and further stirred at −20 °C for 3 h. The reaction was then filtered through a fine frit and washed with cold CH₂Cl₂ and dried under reduced pressure to afford 1a⁺ (121 mg, 55% yield) as a dark blue solid. UV-vis spectroscopy: λ_max = 613 nm in hfip. Single crystals were obtained from the blue concentrated filtrate upon standing at −20 °C for 16 h; data are summarized in Figure 4a and Tables S4, S6–S7.

Figure 4.

Oxidation of 1a affords iodanyl radical 1a⁺. Oxidation: 5 mM 1a in a 0.2 M TBAPF₆/hfip solution, CPE 1.22 V vs. Fc⁺/Fc; or, 0.5 equivalents PIFA, BF3·OEt2 in CH₂Cl₂. Reduction: 1.0 equivalent of N,N-dimethylaniline in hfip. a) In situ UV-vis spectra collected during CPE of 5 mM 1a (—). TD–DFT absorption spectrum of 1a⁺ (—), and electronic configurations of excited states for 1a⁺ (—). b) Most significant contributors to the computed transition at 645 nm. c) In situ EPR spectra collected during CPE of 1 mM 1a (—) and simulated spectrum of 1a⁺ (—).

Reaction of 1a⁺ with 2.

A 20-mL scintillation vial charged with 1a⁺ (95.4 mg, 0.201 mmol, 2.00 equiv.) and 2 (21.2 mg, 0.100 mmol, 1.00 equiv.) in hfip (5.0 mL). The reaction mixture was cooled to 0 °C. Tetramethylammonium acetate (26.6 mg, 0.187 mmol, 0.929 equiv.) was added batch wise at 0 °C before the reaction was allowed to warm to 23 °C and stirred for an additional 6 h. The reaction was then concentrated under reduced pressure, diluted with CDCl₃, and trimethoxybenzene (10.0 mg, 0.60 mmol) was added as an internal standard. The reaction outcome was analyzed by ¹H NMR spectroscopy to show product 3 (35% yield) and 1a (quant.). Compound 3: ¹H NMR (δ, 23 °C, 400 MHz, CDCl₃): 7.38 (d, J = 8.8 Hz, 2H), 6.91 (d, J = 8.8 Hz, 2H), 5.85 (s, 1H), 4.22 (td, J = 9.8, 4.5 Hz, 1H), 4.16−4.09 (m, 2H), 3.81 (s, 3H), 2.18 (s, 3H). The obtained spectral data were in good agreement with the literature.¹¹

Intramolecular C–N Coupling Catalyzed by 1a.

A 10-mL glass vial was charged with N-arylacetamide 2 (42.4 mg, 0.201 mmol, 1.00 equiv.), 1a (1.00 × 10⁻³ mmol, 0.5 mol%), tetramethylammonium acetate (57.0 mg, 0.401 mmol, 1.99 equiv.), tetrabutylammonium hexafluorophosphate (390 mg, 1.01 mmol, 5.02 equiv.), and hfip (5.0 mL) and was fitted with a glassy carbon anode, platinum cathode, and Ag⁺/Ag reference electrode. From a stock solution of 1a (7.84 mg, 0.0201 mmol, 1 mL hfip), 50 μL were transferred to the reaction flask with a micro syringe to obtain 0.5 mol% catalyst loading. Constant potential electrolysis was applied to the reaction mixture at 1.22 V vs. Fc⁺/Fc, until ~50 C charge (~2.6 F/mol) was passed. The reaction was then concentrated under reduced pressure, diluted with CDCl₃, and trimethoxybenzene (10.0 mg, 0.60 mmol) was added as an internal standard and the reaction outcome was analyzed by ¹H NMR spectroscopy to show product 3 (92% yield). Spectral data are consistent with those reported above.

Spirocyclization of 4 Catalyzed by 1a.

A 10-mL glass vial was charged with 4 (47.7 mg, 0.201 mmol, 1.00 equiv.), 1a (2.01 × 10⁻³ mmol, 1 mol%), tetramethylammonium acetate (57.0 mg, 0.401 mmol, 1.99 equiv.), tetrabutylammonium hexafluorophosphate (390 mg, 1.01 mmol, 5.02 equiv.), and hfip (5.0 mL) and was fitted with a glassy carbon anode, platinum cathode, and Ag⁺/Ag reference electrode. From a stock solution of 1a (7.84 mg, 0.0201 mmol, 1 mL hfip), 100 μL were transferred to the reaction flask with a micro syringe to obtain 1 mol% catalyst loading. Constant potential electrolysis was applied to the reaction mixture at 1.22 V vs. Fc⁺/Fc, until ~50 C charge (~2.6 F/mol) was passed. The reaction was then concentrated under reduced pressure, diluted with CDCl₃, and trimethoxybenzene (10.0 mg, 0.60 mmol) was added as an internal standard and the reaction outcome was analyzed by ¹H NMR spectroscopy to show product 5 (61% yield). ¹H NMR (δ, 23 °C, 400 MHz, CDCl₃): 6.94 (d, J = 10.2 Hz, 2H), 6.37 (d, J = 10.1 Hz, 2H), 6.14 (s, 1H), 2.70 (t, J = 7.9 Hz, 2H), 2.38 (t, J = 8.0 Hz, 2H), 2.14 (s, 2H). The obtained spectral data were in good agreement with the literature.²⁰

Lactonization of 6 Catalyzed by 1a.

A 10-mL glass vial was charged with 6 (39.8 mg, 0.201 mmol, 1.00 equiv.), 1a (0.101 mmol, 5 mol%), tetramethylammonium acetate (57.0 mg, 0.401 mmol, 1.99 equiv.), tetrabutylammonium hexafluorophosphate (390 mg, 1.01 mmol, 5.02 equiv.), and hfip (5.0 mL) and was fitted with glassy carbon anode, platinum cathode, and Ag⁺/Ag reference electrode. From a stock solution of 1a (7.84 mg, 0.0201 mmol, 1 mL hfip), 500 μL were transferred to the reaction flask with a micro syringe to obtain 5 mol% catalyst loading. Constant potential electrolysis was applied to the reaction mixture at 1.22 V vs. Fc⁺/Fc, until ~50 C charge (~2.6 F/mol). The reaction was then concentrated under reduced pressure, diluted with CDCl₃, and trimethoxybenzene (10.0 mg, 0.60 mmol) was added as an internal standard and the reaction outcome was analyzed by ¹H NMR spectroscopy to show product 7 (99% yield). ¹H NMR (δ, 23 °C, 400 MHz, CDCl₃): 8.38 (dd, J = 8.2, 1.2 Hz, 1H), 8.21 (d, J = 7.9 Hz, 1H), 8.15 (dd, J = 8.1, 1.4 Hz, 1H), 7.92 (td, J = 7.8, 1.3 Hz, 1H), 7.67−7.63 (m, 1H), 7.58−7.53 (m, 1H), 7.45−7.36 (m, 4H). The obtained spectral data were in good agreement with the literature.²³

Benzylic Acetoxylation of 8 Catalyzed by 1a.

A 10-mL glass vial was charged with 8 (39.0 mg, 0.201 mmol, 1.00 equiv.), 1a (19.5 mg, 0.0500 mmol, 25 mol%), tetramethylammonium acetate (57.0 mg, 0.401 mmol, 1.99 equiv.), tetrabutylammonium hexafluorophosphate (390 mg, 1.01 mmol, 5.02 equiv.), and hfip (5.0 mL) and was fitted with glassy carbon anode, platinum cathode, and Ag⁺/Ag reference electrode. Constant potential electrolysis was applied to the reaction mixture at 1.22 V vs. Fc⁺/Fc, until ~50 C charge (~2.6 F/mol). The reaction was then concentrated under reduced pressure, diluted with CDCl₃, and trimethoxybenzene (10.0 mg, 0.60 mmol) was added as an internal standard and the reaction outcome was analyzed by ¹H NMR spectroscopy to show product 9 (74% yield). ¹H NMR (δ, 23 °C, 400 MHz, CDCl₃): 7.38 (d, J = 8.8 Hz, 2H), 6.91 (d, J = 8.8 Hz, 2H), 5.85 (s, 1H), 4.22 (dq, J = 9.8, 4.5 Hz, 1H), 4.16−4.09 (dq, J = 9.8, 4.5 Hz, 1H), 3.81 (s, 3H), 2.18 (s, 3H), 1.24 (t, 7.13 Hz, 3H). ¹³C NMR (δ, 23 °C, 400 MHz, CDCl₃): 170.4, 169.1, 160.3, 129.1, 126.0, 114.2, 74.3, 61.6, 55.3, 20.8, 14.0. HRMS–ESI: calculated for [M+Na]= 175.0890, observed [M+Na]= 175.0884.

RESULTS AND DISCUSSION

We initiated our studies by examining the impact of aryliodide structure on the electrochemical C–N coupling of biarylamide 2 to afford the corresponding carbazole (3). Previous attempts to lower the catalyst loading from 25 mol% (optimized condition with 4-io-doanisole (1b) as catalyst) resulted in significantly decreased reaction efficiency: At 5 mol% 1b only 30% carbazole was produced and lowering further to 0.5 mol% catalyst loading resulted in 6% carbazole formation (Figure 2a, Table S1). 4-Iodo-1,2-dimethoxybenzene (1c), which features a second methoxy substituent, displays significantly increased electrochemical reversibility and decreased the onset potential for oxidation, but poor catalytic activity (8% yield in cyclization of 2). 1,2-Diiodo-4-methoxybenzene (1d), which features a second iodine substituent, displays poorer electrochemical reversibility and neither lowered the observed E_1/2 nor resulted in improved catalysis. In contrast, 1,2-diiodo-4,5-dimethoxybenzene (1a), which features two iodine and two methoxy substituents, displays a reversible wave at E_1/2=1.13 V vs. Fc⁺/Fc, a 200 mV lower E_pa than 1b (Figure 2b), and is a highly efficient catalyst. With 1a, the catalyst loading could be lowered to 0.5 mol% with no loss of yield or FE (i.e., 92% yield, 73% FE, Figure 2a).

Figure 2.

a) Impact of catalyst structure on cyclization efficiency. Conditions: CPE at the E_pa of 1 determined by CV in an undivided cell with a glassy carbon anode, a platinum-plated cathode, and a Ag⁺/Ag reference electrode. b) CVs of 1a (—), 1b (—), 1c (—), and 1d (—) measured of 1 (5 mM) in 0.2 M TBAPF₆/hfip at 0.1 V/s.

The relative position of the iodine substituents is crucial to catalytic efficiency: 1,4-Diiodo-2,5-dimethoxybenzene (1e) displays E_1/2 = 1.09 but is an inefficient C–N coupling catalyst (8% yield in cyclization of 2). Other 1,2-dimethoxybenzene derivatives were evaluated including 1,2-dibromo-4,5-dimethoxybenzene (1f), 1,2-dichloro-4,5-dimethoxybenzene (1g), and 1,2-dimethoxybenzene (1h). At 0.5 mol% catalyst loading, these aryl iodides afforded carbazole 3 in 40%, 20%, and 4% yield, respectively (Figure 2a). The mixed halogen catalyst 1-bromo-2-iodo-4,5-dimethoxybenzene (1i) performed worse compared to the dihalides 1a, 1f, or 1g affording 3 in 16% yield. Finally, we also examined 2-iodo-4,5-dimethoxybenzoic acid (1j) based on previous proposals that carboxylate ligands can stabilize iodanyl radicals generated from reduction of I(III) reagents.^{24, 25} At 0.5 mol% catalyst loading 1j afforded only 2% yield of 3. Additional electron rich aryl iodides were examined and were generally poorly efficient catalysts (Figure S1).

The nature of the electrochemical processes observed of 1a was analyzed by examining the relative cathodic (I_PC) and anodic currents (I_PA) as well as the separation of the oxidation and reduction peaks (ΔE_p) in CVs of 1a. In these experiments, CVs were obtained for hfip solutions of 1a with [TBA]PF₆ as electrolyte and a scan rate of 0.1 V/s. Under these conditions, the I_pc/I_pa ratio, which provides a measure of electrochemical reversibility and thus the stability of the electrochemically generated species, was 0.94 (c.f. under these conditions, I_pa/I_pc = 1.0 for ferrocene; Figure 3a). ΔE_p for 1a was measured to be 183 mV; for comparison, ΔE_p for ferrocene was 187 mV under these conditions (Figure 3a). The significant deviation from ideality (i.e., ΔE_p = 59 mV) likely results from slow electron transfer kinetics under the reaction conditions.^{26, 27} Analogous results were obtained at different scan rates (Figure S2, Table S2). Further support for a one-electron event was obtained by integration of the square-wave voltammetry data of 1a and Fc, which are similar (Figure 3b, Figure S3).²⁸ Analyses of other veratrole derivatives showed significantly less electrochemical reversibility, which we speculate may result in poor catalyst performance (Figure S4, Table S3).

Figure 3.

a) Cyclic voltammograms of 5.0 mM 1a in a 0.2 M TBAPF₆/hfip solution (—) and 5.0 mM ferrocene in a 0.2 M TBAPF₆/hfip solution (—) at 0.1 V/s with normalized potentials. b) Square wave voltammetry of a 5.0 mM 1a and 0.5 mM ferrocene in a 0.2 M TBAPF₆/hfip solution. c) CVs of 1a (5 mM) with 35.5 mM [TMA]OAc in 0.2 M TBAPF₆/hfip at 0.1 V/s varying biarylamide 2 loading 0.0 mM (—), 1.9, 6.1, and 12 mM (—).

To gain insight into the impact of substrate (2) and acetate, which is required for efficient C–N coupling, CVs were measured as a function of the concentration of these species. Addition of [TMA]OAc to CV solutions of 1a resulted in increased oxidative current and partial loss of reversibility (Figure S5, I_pc/I_pa = 0.52 with 35.5 mM [TMA]OAc). We previously interpreted similar observations in CVs of 4-iodoanisole 1b as consistent with acetate binding to an anodically generated iodanyl radical.¹¹ In the case of 1a, reversibility can be recaptured at higher scan rates, as the rate of reduction outcompetes the electron transfer with added acetate (Figure S6). Addition of substrate 2 to CVs of 1a does not affect the reversible wave observed for 1a (a new peak at E_pa=1.55 V vs. Fc⁺/Fc grows in (Figure S7), a result of direct substrate oxidation (Figure S8)). Addition of both 2 and [TMA]OAc to CVs of 1a resulted in the observation of catalytic current where reversibility could not be regained regardless of scan rate, indicative of rapid catalytic turn over (Figure 3c).

Given the single-electron inventory indicated by CV analysis, we were interested in the possibility that substrate activation may arise from one-electron oxidation of 1a without subsequent generation of an I(III) intermediate. Constant potential electrolysis (CPE) of an hfip solution of 1a (0.2 M [TBA]PF₆ added as electrolyte) in the absence of either acetate or 2 resulted in a drastic color change from colorless to dark blue (Figure 4). The UV-vis spectrum of this solution, acquired in situ via spectroelectrochemistry or ex situ following the cessation of electrolysis, displayed a broad low-energy absorbance centered at 613 nm (Figure 4a (—), Figure S9 and S10). Acquisition of spectra periodically during the electrolysis reveal the presence of an isosbestic point at 380 nm, which indicates the lack of an intermediate in the anodic oxidation of 1a. The observed UV-vis spectrum is consistent with reported veratrole-derived radical cations^{21, 29, 30} (red shifted due to the heavy atom effect of two iodides). Time-Dependent Density Functional Theory (TD-DFT) computations reproduce the low-energy feature in the spectrum of 1a⁺ (λ_max=645 nm, Figure 4a (—, —)).³¹ The largest orbital contribution to this transition is β-HOMO(−2) to β-LUMO, which is primarily π-to-π* with significant contribution from the in-phase combination of I–centered p-orbitals (Figure 4b).

Two experiments were carried out to confirm the blue solution represented single-electron oxidation of 1a. First, addition of one equivalent of N,N-dimethylaniline results in the consumption 1a⁺ (with concurrent regeneration of 1a) and the quantitative evolution of the N,N-dimethylaniline radical cation (Figure S11). Second, in situ EPR experiments³² are consistent with one-electron oxidation of 1a to generate an open-shell species. The resulting spectrum displays narrow spectral width (25 G) that is well-simulated with 2a_I = 1.67 G, 2a_H = 3.35 G, and 2a_Me = 0.343 G (Figure 4c). For comparison, under identical in situ conditions, neither 4-iodotoluene or 4-bromoanisole provide a spectral signature (Figure S12).

To simplify isolation and independent characterization of iodanyl radical 1a⁺, we pursued chemical oxidation of 1a to avoid the use of supporting electrolyte. Inspired by Kita’s seminal work on single-electron oxidation of anisole derivatives by bis(trifluoroacetoxy)iodobenzene (PIFA),^{33, 34} we found that treatment of 1a with 0.5 equivalents of bis(trifluoroacetoxy)iodobenzene (PIFA) and excess BF3·OEt2 in CH₂Cl₂ resulted in a dark-blue solution that was spectroscopically identical to the electrochemically generated solution (for comparison of UV-vis spectra and solvatochromism, see Figure S13).³⁵ Dark blue X-ray quality single-crystals of 1a⁺ were obtained on prolonged standing of reaction solutions at −22 °C (Figure 5a; solid-state and solution-phase UV-vis of redissolved crystals are collected in Figure S13). X-ray diffraction of the obtained single crystals provided the molecular structure pictured in Figure 4a. The metrical parameters of 1a are similar to those of 1a⁺, with the exception of a marked contraction of the I-I distance from 3.71157(2) in 1a to 3.6391(7) in 1a⁺ (for complete comparison of 1a and 1a⁺, see Figure S14, Table S4–S6). Charge balance in the crystal is maintained by 0.5 B₂F₆O²⁻ dianions per 1a⁺(See Table S7 and Figure S15 for refinement details).

Figure 5.

a) Displacement ellipsoid plot of 1a⁺ drawn at 50% probability, H-atoms are omitted for clarity. Selected metrical parameters: C1–I1 = 2.076(6) Å, I1–I2 = 3.6392(7) Å, B1–O1 = 1.513(9) Å, B2–O1 = 1.520(9) Å. b) Computed SOMO of 1a⁺. c) Atoms-in-molecules (AIM) analysis indicates an I–I bond critical point (BCP).

Oxidation of 1e-1j under the same conditions used for the preparation of 1a⁺ resulted in significant and distinct color changes. Oxidation of compounds 1f and 1g produced red and yellow solutions respectively (Figure S16a). Oxidation of 1e, 1i, and 1j afforded purple solutions; in the case of 1j the color dissipated quickly, which may be due to decarboxylative decomposition (Figure S16b). Treatment of veratrole (1h) under these conditions did not result in an observed color change. Unfortunately, attempts to crystalize these radical cations (i.e., 1e⁺-1j⁺) were unsuccessful.

DFT optimized structures of 1a and 1a⁺ provide metrics that are in close agreement with those determined by X-ray crystallography (Tables S8 and S9). In particular, the noted contraction of the I-I vector is well described in the computed structures: one-electron oxidation leads to a computed I–I contraction from 3.70 Å for 1a to 3.65 Å for 1a⁺. The singly occupied molecular orbital (SOMO) of 1a⁺ displays predominantly π-character with significant orbital contribution from the iodine atoms (~35% iodine-iodine AO contribution to the SOMO, Figure 5b). We envisioned that the observed I–I contraction may arise from oxidatively induced delocalized 3-electron bonding, which is well-precedented in proximal heavy main-group elements, such as diselenides and disulfides,^36–38 but unknown in hypervalent iodine chemistry. Atoms-in-molecules (AIM) analysis indicates an electron density of ρ(r)= 0.0123 e·bohr⁻³ and a Laplacian distribution of ∇²ρ(r)=0.0323 e·bohr⁻⁵ at the I–I bond critical point (BCP) for 1a⁺ (Figure 5c). For comparison, no BCP is observed between the I atoms in 1a (Figure S17). The ρ(r) observed at the BCP of 1a⁺ is within the range of values expected for I–I bonds.³⁹

AIM analysis was also carried out for 1f⁺, 1g⁺, 1i⁺, and 1j⁺. No BCPs were observed for radical cations derived from the ortho-di-bromo (i.e., 1f⁺), ortho-dichloro (i.e., 1g⁺) or the ortho-iodobromo compounds (i.e., 1i⁺), which suggests that in the context of organo-halides the large atomic radius and high polarizability of iodine may be critical to the observed X–X interactions; potential translation of X–X interaction to the smaller halogens such as in hypervalent bromine chemistry remains an area for development.^{40, 41} In contrast, a BCP is observed for the iodanyl radical cation derived from 1j (the BCP is observed both for 1j⁺ and the corresponding deprotonated analogue, see Figure S18). This result is consistent with previous reports that an ortho-carboxylate substituent can stabilize iodanyl radicals generated during the reduction of I(III) compounds.^{24, 25}

Isolation of 1a⁺ provided the first opportunity to directly evaluate the reactivity of organic I(II) derivatives in substrate functionalization. Treatment of 1a⁺ reactions with biarylamide 2 and [TMA]OAc in hfip resulted in immediate disappearance of the blue color and the evolution of carbazole 3 (35% yield) and quantitative recovery of 1a (Figure 6a). With demonstration of the chemical competence of iodanyl radical 1a⁺ in substrate functionalization and the significantly decreased oxidation potential of 1a as compared to 1b, we examined the potential to achieve other C–E bond-forming reactions (Figure 6b). Oxidative dearomatization of 4 was successfully demonstrated with catalyst loading as low as 1 mol% producing 61% product 5. C(sp²)–O bond formation can be accomplished from biarylcarboxylic acid 6 to afford 7 under electrocatalytic conditions (5 mol% 1a) loading.⁴² In the absence of 1a no background oxidation of either 4 or 6 was observed. Additionally, intermolecular benzylic acetoxylation of 8 to generate 9 was accomplished in 74% although higher catalyst loading (25 mol%) was required for high yield. Direct oxidation of 8 for extended reaction times in the absence of catalysts shows greater preference for the benzylic ketone product (40%) and only 20% acetoxylated product 9.

Figure 6.

a) Iodanyl radical promotes conversion of 2 to 3. b) Diiodide 1 catalyzes i) oxidative dearomatization, ii) lactone cyclization, and iii) benzylic acetoxylation. CPE at 1.22 V vs. Fc⁺/Fc. Reactions were optimized to the lowest catalyst loading without suffering current density (>1.0 mA).

CONCLUSION

In conclusion, here we introduce a new aryl iodide electrocatalyst — 1,2-diiodo-4,5-dimethoxybenzene (1a) — that operates at catalyst loadings down to 0.5 mol%. This catalyst activity is ascribed to a combination of reversible veratrole-based electrochemistry and co-operative I–I bonding that stabilizes the iodanyl radical intermediate that results from one-electron oxidation. Isolation of the incipient iodanyl radical enabled both complete characterization of the first formally I(II)-based organic molecule and direct interrogation of the substrate functionalization chemistry of this exotic species in substrate functionalization reactions. The results described here 1) suggest I(II) intermediates can be catalytically competent rather than simply intermediates en route to more traditionally invoked I(III) species and 2) expand the reactivity modes that are available to iodine-based catalysts in metal-free substrate oxidation chemistry.