C–H Amination via Electrophotocatalytic Ritter-Type Reaction

Abstract

A method for C–H bond amination via an electrophotocatalytic Ritter-type reaction is described. The reaction is catalyzed by a trisaminocyclopropenium (TAC) ion in an electrochemical cell under irradiation. These conditions convert benzylic C–H bonds to acetamides without the use of a stoichiometric chemical oxidant. A range of functionality is shown to be compatible with this transformation, and several complex substrates are demonstrated.

Graphical Abstract

The invention of methods to convert unactivated C–H bonds to C–N bonds has long been established as a desirable goal.^1–2,3,4,5 In this regard, the classic Hoffmann-Löfler-Freitag reaction has been a mainstay for complex molecule synthesis,^6–7,8,9 while modern incarnations of this strategy have recently been described using photoredox catalysis.^{10 – 11 12} Alternatively, a variety of reactions involving nitrene intermediates that can undergo C–H insertion or H-atom abstraction have proven exceptionally useful.^{13–14 15 16 17 18} Moreover, the use of designer transition metal complexes that can effect C–N couplings of C–H and N–H bonds have become increasingly popular, and a wide range of impressive applications of this strategy have been disclosed.^{1,19–20,21,22} An orthogonal strategy entails the oxidation of a C–H bond to the corresponding carbocation, which can then be trapped by a nitrile, usually as solvent, leading to amide products (Figure 1a). This process is in essence a Ritter reaction,^{23–24,25,26} with the difference being the means by which the carbocation is generated. Several catalytic methods for Ritter-type C–H amination have been reported, including work by Ishii, Baran, Kiyokawa and Minakata, and Liu and Chen (Figure 1b).^{27–28 29 30} While these are important advances, the oxidants employed are incompatible with many functional groups and in the case of Selectfluor are quite expensive. A potentially attractive alternative would be to use electrochemistry,^{31–32,33,34,35,36,37,38,39,40,41,42,43} which would obviate the need for the chemical oxidant altogether. While there have been examples of electrochemical Ritter-type C–H aminations,^{44–45,46,47} these processes tend to be low-yielding and incompatible with much native functionality due to the high anodic potentials required. However, very recently an electrochemical method for benzylic C–H amination with sulfonamides was recently reported that proceeded with good efficiency,⁴⁸ although it required the use of HFIP solvent and the range of demonstrated functional group compatibility was narrow. Thus, there remains the need for an electrochemically-coupled version of this chemistry that has the generality needed to operate in the context of greater molecular complexity.

Figure 1.

(a) Ritter reaction of carbocation intermediates. (b) Selected examples of Ritter-type chemistry. (c) Electrophotocatalytic Ritter-type C–H amination.

In this regard, we recently reported⁴⁹ an electrophotocatalytic^{50,51,52,53,54,55,56,57,58,59 – 60 61 62} vicinal C–H diamination reaction that achieved the conversion of alkylarene substrates, particularly those that were α-branched, to 3,4-dihydroimidazole products (Figure 1c). Although these products represent a valuable increase in molecular complexity, it would also be desirable to access the classic Ritter-type monoamination products as well. In this Communication, we report that unbranched substrates undergo efficient single site C–H amination under modified electrophotocatalytic conditions.

With the goal of developing a method that would be applicable to a wide variety of substrates, we chose the complex substrate 2, which is an analogue of the pharmaceutical agent celecoxib, for our optimization studies (Table 1). Under the optimal conditions we identified, TAC 1 catalyzed the conversion of this substrate to acetamide 3 by reaction in a divided cell (E_cell = 2.2 V, E_anode = 1.6 V vs SCE) under irradiation from a compact fluorescent light (CFL), in the presence of TFA, H₂O, and n-Bu₄NPF₆ in acetonitrile solvent (entry 1). A small amount (4%) of aldehyde 4 was also observed. Table 1 illustrates the impact of variation from these conditions. We found that the identity of the electrolyte was important. Thus, the use of LiClO₄ (entry 2) or n-Bu₄NBF₄ (entry 3) resulted in a significant decrease in yield of 3 and a small increase in the formation of the side product 4. Not surprisingly, without the application of the cell potential (entry 4) or in the absence of the catalyst (entry 5), no reaction occurred. When only the light was omitted, a small amount (9%) of background reaction was observed (entry 6). Changing the acid additive from TFA to AcOH was detrimental, with the yield dropping to only 10% (entry 7). Although product was observed without the addition of water (entry 8), the yield was significantly diminished. Notably, the use of a divided cell was critical to the success of this process, as the yield of product in an undivided cell was very low (entry 9). Finally, we also attempted to conduct this reaction via direct electrolysis, with potentials up to 3.5 V (entries 10 and 11). While some product was generated under these conditions, the yields were low, the reaction mixtures were complicated, and a large amount of aldehyde side product 4 was formed. The contrast between the direct electrolysis and electrophotocatalytic approach underscores the ability of the latter to effect potent yet selective oxidative chemistry.

Table 1.

Optimization study for electrophotocatalytic Ritter-type C–H amination.^a,b

entry	change from standard	yield 3 (%)	yield 4 (%)
1	none	77	4
2	LiCIO4 instead of n-Bu4NPF6	50	8
3	n-Bu4NBF4 instead of n-Bu4NPF6	27	5
4	no cell potential	0	0
5	no catalyst	0	trace
6	no light	9	trace
7	AcOH instead of TFA	10	trace
8	no H2O	66	trace
9	undivided cell	14	4
10	direct electrolysis (3.0 V)	36	32
11	direct electrolysis (3.5 V)	33	20
12	Pb as cathode	28	trace

^aReaction conditions: 2 (0.5 mmol), H₂O (1.0 equiv), 1 (8 mol %), nBu₄NPF₆ (1.0 equiv), TFA (0.2 mL), MeCN (6.0 mL) [anode], and nBu₄NPF₆ (1.0 equiv), TFA (0.2 mL), MeCN (6.0 mL) [cathode] in H-type divided cell with carbon felt anode, Pt cathode. Reactions performed under constant cell potential conditions with irradiation for 48 h at rt under N₂.

^bYields are of isolated and purified products 3 and 4.

With these optimized conditions we explored the scope of this method (Table 2). Toluene and 4-halogenated toluenes gave rise to the corresponding benzylacetamides 5-9 in good yields. Notably, the position of halogenation had a significant impact on efficiency. Thus, whereas the 4-bromo product 8 was accessed in reasonable yield, the 2-bromo isomer 10 was formed less efficiently, and the 3-bromo isomer 11 was only produced in 31% yield. Also, in the latter two cases, a change of the electrolyte to LiClO₄ proved necessary for maximal yield.⁶³ On the other hand, 1-bromo-3,5-dimethylbenzene reacted with good efficiency to furnish acetamide 12 in 61% yield.

Table 2.

Electrophotocatalytic Ritter-type amination of benzylic C–H bonds.^a

^aReaction conditions: substrate (0.5 mmol), H₂O (1.0 equiv), 1 (8 mol %), nBu₄NPF₆ (1.0 equiv), TFA (0.2 mL), MeCN (6.0 mL) [anode], and nBu₄NPF₆ (1.0 equiv), TFA (0.2 mL), MeCN (6.0 mL) [cathode] in H-type divided cell with carbon felt anode, Pt cathode. Reactions performed under constant voltage (CV) (2.2 V) conditions with light irradiation for 48 h at rt under N₂, isolated yield.

^bLiClO₄ instead of nBu₄NPF₆.

^c24 h.

^d2.6 V.

^eWorked up with 1M NaOH (aq).

Products resulting from functionalization of alkyl substituents other than methyl (e.g. 13 and 14), as long as the benzylic position was not branched. In the case of a branched substrate like cumene, the reaction gave rise to no acetamide product, but rather the dihydroimidazole product 15 as we previously described.⁴⁹ In examining a substrate designed to probe the competition between methyl and methylene sites, we found that 16 was formed with a 6.2:1 preference over the alternative, methyl-functionalized site. Meanwhile, diphenylmethane was functionalized in good yield to furnish the α-diarylamine 17. Cyclic substrates such as tetralin, indane, and dibenzosuberone participated in the Ritter-type reaction as well, giving rise to acetamides 18–20 respectively.

In terms of functional group compatibility, we found that an alcohol (21), carboxylic acid (22), ester (23 and 24), alkyl chloride (25), and tosylate groups (26) survived the amination procedure. While a terminal alkyne containing substrate furnished no product 27, the internal alkyne product 28 was obtained in 70% yield. Notably, products possessing nitrogen functionality of various types could be accessed, including carbazole (29) trifluoroacetamide (30 and 31), pyridine (32), and even free amine-containing structures (33 and 34).

One of the most useful applications of C–H amination chemistry is the late-stage functionalization of complex molecules. For example, White has recently demonstrated efficient amination reactions on a number of bioactive molecules using a trichloroethylsulfoniminoiodinane using manganese catalysis.¹³ We decided to see whether some of these readily available compounds could be aminated using the electrophotocatalytic method, which has the benefits of using acetonitrile as the nitrogen source, requires no transition metal catalyst, and generates more desirable acetamide products (Table 3). Hence, the sertraline analogue 35 could be generated in 44% yield, while the retinoic acid receptor agonist derivative 36⁶⁴ was produced in 63% yield. Meanwhile, an isatin derivative was selectively converted to compound 37 in 58% yield. Similarly, compound 38 derived from an inhibitor of the CYP11B1⁶⁵ was generated in good yield and with complete selectivity for the position shown. Other derivatives bearing either imide (39)⁶⁶ or ketone functionalities (40 and 41) could also be accessed. Notably, products 42–44^{67 – 68 69} were generated with high efficiency despite the potential for undesired oxidation of the free amine functionality. To demonstrate the applicability of this method to preparative scale synthesis, we found that 41 could be generated on a 1.5 g scale in 50% yield by extending the reaction time to 72 h.

Table 3.

Electrophotocatalytic Ritter-type C–H amination of complex molecules.^a

^aReaction conditions: substrate (0.3–0.5 mmol, 1.0 equiv), H₂O (1.0 equiv), 1 (8 mol %), nBu₄NPF₆ (1.0 equiv), TFA (0.2 mL), MeCN (6.0 mL) [anode], and nBu₄NPF₆ (1.0 equiv), TFA (0.2 mL), MeCN (6.0 mL) [cathode] in H-type divided cell with carbon felt anode, Pt cathode. Reactions performed under constant voltage (CV) conditions with light irradiation for 48 h at rt under N₂, isolated yield.

^bLiClO₄ instead of nBu₄NPF₆.

^c72 h reaction time.

^dRun at 2.6 V.

A mechanistic rationale for this electrophotocatalytic Ritter-type reaction is shown in Figure 2. The TAC 1 can be oxidized to generate the stable radical dication 45 (E_p/2 = 1.12 V in 30:1 MeCN:TFA vs SCE). Irradiation of 45 then leads to the photoexcited intermediate 46, which engages in single-electron oxidation of the arene substrate 47 to generate radical cation 48. Deprotonation of this intermediate results in benzylic radical 49, which then undergoes a second oxidation event, either by the TAC radical dication 45 or directly at the anode, leading to carbocation 50. The carbocation then proceeds through the classic Ritter steps by reaction with acetonitrile solvent to form nitrilium 51 followed by hydrolysis (either by reaction with adventitious water or with trifluoroacetic acid) to furnish the amide product 52. Meanwhile, the redox reaction is balanced by cathodic reduction of protons to produce dihydrogen. A key aspect of this approach is that the cell potential is insufficient to oxidize the substrate directly, leading to selective single electron oxidations by the electrophotocatalyst, which minimizes the risk of side reactions.

Figure 2.

Mechanistic rationale for electrophotocatalytic Ritter-type reaction.

An important distinction should be made between the monoaminations described in this work and the diaminations we recently reported under similar conditions.⁴⁹ Our hypothesis for the diamination reaction is that initially formed benzylic acetamides 52 undergo an acid-catalyzed E₁ elimination reaction to furnish styrenes 53, which then undergo further oxidation and Ritter-type trapping reactions to eventually form one or both of the regiosiomeric dihydroimidazoles 54a,b. In the case of benzylic-branched substrates, the elimination reaction is facile with TFA, and thus monoamination products 52 are not observed. With unbranched substrates, the elimination reaction requires a stronger acid like triflic acid. Thus, under the current conditions using TFA, the unbranched substrates furnish only the monoamination products 52.

It might be stressed that the risk of overoxidation presents a significant challenge for C–H amination chemistry, particularly if the installed nitrogen is not sufficiently deactivated (e.g. by a tosyl protecting group). We have measured the redox potentials of several of the products shown in Table 2 using cyclic voltammetry and have found that they are very close to those of the corresponding starting materials (see Supporting Information for details). It seems that the introduction of an acetamide moiety slows the kinetics of single-electron oxidation by TAC 1, which may be why the electrophotocatalytic process is more efficient than the direct electrochemical one. However, we have found that prolonged exposure to the standard reaction conditions (48 h) does result in some acetamide decomposition (7–30% in the compounds tested).

The direct amination of C–H bonds via Ritter-type reactions offers a potentially simple and scalable approach to molecular upgrading. The current work demonstrates that electrophotocatalysis can help achieve these transformations without the need for stoichiometric chemical oxidants and with enough selectivity to operate with a reasonable range of native functionality. As such, this method should be a useful addition to C–H functionalization toolbox.