N-Hydroxybenzimidazole as a structurally modifiable platform for N-oxyl radicals for direct C–H functionalization reactions

A novel class of N-oxy radicals based on flexibly modifiable N-hydroxybenzimidazole skeleton was designed and applied to C–H functionalization reactions.


Introduction
The direct functionalization of C-H bonds has been recognized as an innovative approach to realize step-and atomeconomical synthesis of various functionalized molecules. Considerable research efforts have been devoted to the development of a variety of C-H functionalization reactions based on transition-metal-catalyzed or photoredox-promoted approaches. 1,2 On the other hand, transition-metal-free and non-photolytic direct C-H functionalization reactions still remain a signicant challenge, 3 although they will provide new synthetic strategies with complementary reactivity and selectivity. 4 In this context, N-oxyl radicals have attracted much attention as promising organoradical catalysts for these transformations. 5 As represented by N-hydroxyphthalimide (NHPI), several N-hydroxy compounds have been used to generate active N-oxyl radicals in situ, which selectively abstract hydrogen atoms from C-H bonds of organic molecules. The resulting alkyl radicals react with various radical acceptors to afford the corresponding functionalized products. 6 In order to improve the catalytic efficiency and stability of the key N-oxyl radicals, several N-hydroxy compounds as organoradical precursors have been designed based on the structure of NHPI and utilized in the same type of transformations (Fig. 1a). 7 Most of the reported N-hydroxy compounds, however, have similar structures featuring carbonyl groups adjacent to hydroxylamine moiety as a reactive site. While these carbonyl groups signicantly contribute to enhancing the catalytic performance of N-oxyl radicals, 8 only a few sites are available for structural modication due to their existence. As a result, such a lack of structural diversity for these N-hydroxy compounds renders it difficult to alter the specic properties of the corresponding N-oxyl radicals such as bond dissociation energies (BDEs), a fundamental parameter to estimate their reactivities in the hydrogen atom transfer process. Accordingly, further progress in this eld demands a completely novel design of N-oxyl radicals with structural diversity. Herein, we report the design and synthesis of novel organoradical species based on N-hydroxybenzimidazoles (NHBIs), which contain multiple modication sites (Fig. 1b), and demonstrate their distinct reactivities as organoradical catalysts and efficient radical initiators in the direct C-H functionalization reactions.
N-Hydroxybenzimidazoles, which are planar and stable heterocycles bearing an N-hydroxy moiety, have recently attracted interest in the eld of biological and pharmaceutical sciences as anti-virulence or anti-cancer agents. 9,10 However, they have rarely been used in synthetic organic chemistry and, to the best of our knowledge, studies on the potential of NHBIs to generate the corresponding N-oxyl radicals have not yet been conducted. 11 In this context, we became interested in the potential of NHBIs as novel organoradical precursors based on the following features: (1) NHBIs can be readily prepared from 2-nitroaniline derivatives in a few steps (see the ESI †); (2) substituents can be easily introduced at both the aromatic ring and the 2-position of the benzimidazole moiety; (3) NHBIs contain additional modication sites such as the nitrogen atom at the 3-position of the benzimidazole moiety and the counteranion of the resulting benzimidazolium species, which may potentially be exploited for further functionalization.

Results and discussion
To test our hypothesis, we initially carried out density functional theory (DFT) calculations in order to estimate the BDE values of the O-H bonds in our designed NHBI derivatives (1a-f) (Fig. 2). 12 As expected, the results of the DFT calculations revealed that the BDE values for 1 can be tuned within a wide range ($10 kcal mol À1 ) by facile modications such as the introduction of substituents at the aromatic ring or the 2position of the imidazole moiety, or by N-alkylation at the 3position. Moreover, some of synthesized NHBI derivatives have the similar or even higher BDEs compared to that of NHPI, which has been widely used as an efficient catalyst for several hydrogen atom abstraction reactions. 6 Thus, these results indicate that N-oxyl radicals derived from NHBI derivatives potentially work as organoradical catalysts for direct C-H functionalization via hydrogen atom abstraction.
In order to evaluate the catalytic activities of NHBI derivatives, we attempted to apply them to the reported C-H functionalization reactions. Among several transformations catalyzed by N-oxyl radicals, we selected the benzylic C-H amination of ethylbenzene (2) as a model reaction, which was originally reported as the NHPI-catalyzed reaction (Scheme 1). 6d Aer the reaction was run using diethyl azodicarboxylate (3) in the presence of 10 mol% of NHPI at 80 C for 24 h, the desired aminated product 4 was obtained in moderate yield. On the other hand, yields of 4 for reactions using NHBI derivatives varied signicantly, depending on their substituents. Notably, when 1d was used under the same reaction conditions, the yield of 4 was much higher than the use of NHPI. On the contrary, cationic NHBI derivative 1e, which has a highest value of BDE among examined N-hydroxy compounds, did not show remarkable catalytic activity for this transformation, probably due to difficult regeneration of active N-oxy radicals. These experimental results clearly indicate that our designed NHBI derivatives can be applied to direct C-H functionalization reactions.
Having conrmed the potential of NHBI derivatives as active organoradical precursors, we set about exploring their distinct reactivities for the development of direct C-H functionalization reactions. Our continuing concern about methods for generation and application of acyl radical species 13,14 led us to investigate the direct functionalization of aldehydes via C(sp 2 )-H bond activation by NHBI derivatives. To our delight, we found that in the presence of catalytic amount of NHBI, the aldehydic C-H bond was directly converted into C-F bond by using Selectuor as a uorine atom transfer reagent, affording the corresponding acyl uorides (Scheme 2). Acyl uorides have recently attracted much attention in synthetic and biological chemistry owing to their unique reactivity, which is not observable in commonly employed acyl chlorides or acid anhydrides. 15,16 While several synthetic methods for acyl uorides have been reported, they are typically prepared by using toxic uorination reagents and/or precious metal catalysts. 17,18 Therefore, the development of their different synthetic approach with high practicability is in great demand.  Scheme 1 Evaluation of 1 as organoradical catalyst for benzylic C-H amination. Yields were determined by 1 H NMR spectroscopy using 1,1,2,2-tetrachloroethane as an internal standard.
Initially, the effect of NHBI derivatives was assessed in the reaction of 3-phenylpropanal (5a) with Selectuor (Table 1). Although the reaction using 10 mol% of 1a did not afford the desired acyl uoride 6a (entry 1), uorine-substituted NHBIs 1b-d afforded 6a in moderate to good yields (entries [2][3][4]. It is noteworthy that cationic NHBI derivative 1e, which was not active for the aforementioned benzylic C-H amination reaction (Scheme 1), turned out to be effective for this C-H uorination reaction (entry 5). During the optimization, we observed the formation of 3-phenylpropionic acid (6a 0 ) as a side product, which could potentially be formed during the reaction of an acyl radical with oxygen. 19 Thus, we carried out the reaction under an inert atmosphere of argon, which further improved the yield of 6a with the generation of 6a 0 sufficiently suppressed (entry 6). On the other hand, the introduction of triuoromethyl group at the 2-position of benzimidazolium structure gave only a small amount of the product (entry 7). While the effect of counteranion for cationic NHBI derivative is almost negligible (entry 8), the importance of free hydroxylamine moiety of NHBI was conrmed by the reaction with N-benzyloxybenzimidazole 1h (entry 9). Neither 6a or 6a 0 was obtained in the absence of NHBI derivative (entry 10). As a comparative study, NHPI and its derivatives were used for this transformation (entries [11][12][13]. Although the reactions proceeded moderately in the presence of these N-hydroxy compounds, the yields of 6a were much lower than the use of 1e. With the optimal NHBI derivative in hand, the scope of aldehyde was then investigated ( Table 2). While isolated yields of products were determined aer one-pot conversion into benzyl amides 7 to avoid loss of acyl uorides 6 due to their volatility and/or instability on silica-gel, several reactions could be scaled up to 2.0 mmol and the obtained acyl uorides were successfully isolated aer ash column chromatography in acceptable yields. The reactions of aliphatic aldehydes proceeded even at room temperature, affording the desired products in moderate to high yields. Desired products bearing benzylic C-H bond (6a-7a), acyclic or cyclic ether moieties (7f and 7i) and nitrogen-containing functional groups (6d-7d, 6h-7h and 7k) were successfully obtained, highlighting the advantage of mild reaction conditions. Although pivalaldehyde gave the corresponding product 7j in low yield due to competitive decarbonylation, 18d functionalized aldehyde with two stereogenic centers provided the desired benzylamide 7k without a Unless otherwise specied, reactions were carried out in MeCN for 2 h in the presence of 5a (0.20 mmol), Selectuor (2.0 equiv.) and N-hydroxy compound (10 mol%). b Yields of 6a were determined by 1 H NMR spectroscopy using benzotriuoride as the internal standard. c The reaction was conducted under an atmosphere of argon. N.D.: not detected.
Scheme 2 Aldehydic C-H fluorination reaction using NHBI. epimerization. Moreover, products from aryl aldehydes (7l, 6m-7m and 6n-7n) and trans-cinnamaldehyde (6o-7o) could also be obtained at increased reaction temperature. The direct conversion of aldehydes into acyl uorides and the subsequent reaction with various nucleophiles allow for the divergent synthesis of carbonyl compounds (Scheme 3). Other than benzylamine, secondary amines (8a-b), amino acid derivatives (8c-d) and even less nucleophilic oxazolidinone (8e) were also applicable to the one-pot procedure. The use of other nucleophiles such as alcohol (8f-h) and thiol (8i) provided the corresponding esters and thioester in good to high yields. This protocol was also applied to the conversion of b-hydroxy aldehyde 9 into b-lactone 10 in one pot.
Moreover, starting from isolated acyl uoride 6a, several unsymmetrical ketones (11a-e) could be synthesized via the reaction with various carbon nucleophiles (Scheme 4). 20 Both the arylation of 6a by a palladium-catalyzed cross-coupling with phenylboronic acid and the Friedel-Cras reaction with 1,3dimethoxybenzene (A) afforded aryl ketones 11a and 11b, respectively, in good yields. Alternatively, the reaction with a Wittig reagent furnished b-keto ester 11c, while the use of allyltrimethylsilane afforded b,g-unsaturated ketone 11d in high yield. Finally, the reaction with silyl enol ether B in the presence of tributyltin uoride afforded b-keto ester 11e quantitatively. These transformations demonstrate the utility of acyl uorides as versatile synthetic intermediates.
To better understand the reaction mechanism, several experiments were carried out (Scheme 5). The addition of 2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPO) as a radical scavenger under standard conditions completely inhibited the formation of acyl uoride (Scheme 5a). In addition, when a stoichiometric reaction of 1e with styrene and TEMPO was run in the presence of Selectuor, a 1 : 1 : 1 adduct of 1e, styrene and TEMPO was detected by high-resolution mass spectrometry (Scheme 5b). On the other hand, Selectuor is known to be activated under photoirradiation condition, generating the corresponding radical cation 12. 21 This species can work as a strong hydrogen atom abstraction reagent for C-H activation of various hydrocarbons. 21a-c,22 Based on recent reports, we conducted the C-H oxidative arylation of cyclooctane with isoquinoline in the presence of a catalytic amount of 1e (Scheme 5c). 21b,c As a result, arylated product 13 could be obtained in good yield even without photoirradiation. We also conrmed 13 was not detected in the absence of 1e and photoirradiation. Considering a high BDE value for C-H bond of cyclooctane (95.8 kcal mol À1 , see the ESI †), a reaction pathway where the N-oxyl radical derived from 1e abstracts a hydrogen atom from C-H bond of cyclooctane would not be feasible. Therefore, these results indicate that our NHBI/Selectuor system allows for the efficient generation of highly reactive radical cation 12, which is a key species for the activation of aldehydic C-H bond.
Based on these studies, we propose a reaction mechanism as shown in Fig. 3. In the presence of NHBI 1, radical cation 12 is generated from Selectuor with concomitant formation of Noxyl radical 14. Then 12 abstracts a hydrogen atom from aldehyde 5 to form the corresponding acyl radical 15. Subsequently, a uorine atom of another Selectuor is trapped by 15 to give the desired acyl uoride 6 and radical cation 12, and the reaction proceeds via a chain process until completion.
With respect to the initiation process generating two different radicals from 1 and Selectuor, further theoretical studies by DFT calculation provided some insights into the remarkable reactivity difference among NHBI derivatives (Fig. 4). Namely, when this process is divided into two steps as shown in Fig. 4a, the energy gap between LUMO of putative oxoammonium 16 and HOMO of monocationic amine 17 is widely changed depending on the substitution pattern of NHBI structure. In addition, these values of HOMO-LUMO energy gap are in proportion to the Gibbs free energy difference (DG) for step II (Fig. 4b). It should be noted this correlation is not the only determinant for the efficiency of initiation process; the exceptionally lower reactivity of 1f (Table 1, entry 7) would be attributed to an unfavorable uphill in free energy for the step I, which makes it difficult to be oxidized by Selectuor to form the corresponding oxoammonium species leading to step II (see the ESI †). Thus, these results demonstrated that the highly designable structures of 1 allow for the tuning of their electronic state, which can alter the energy prole of the process for generation of active radical species.

Conclusions
In conclusion, we have synthesized a novel class of N-oxyl radicals based on NHBI, which is a exibly modiable structural platform. The evaluation of these NHBI derivatives as organoradical catalysts revealed that the substituents on these structures signicantly alter their catalytic performance in benzylic C-H amination reaction. Moreover, we have developed a novel metal-free and non-photolytic method for the synthesis of acyl uorides by using a catalytic amount of NHBI derivatives. Mechanistic studies indicated a distinct character of NHBI derivatives as an efficient radical initiator to generate a more active radical species for hydrogen atom abstraction. Further investigations into the applications of NHBI derivatives to other C-H activation strategies, as well as the development of novel NHBI catalysts, are currently underway in our laboratory. We believe that these studies indicate a new direction for the chemistry of N-oxyl radicals, which will spur further research on organoradical catalysis for direct C-H activation reactions.

Conflicts of interest
There are no conicts to declare.