Intermolecular Aryl C−H Amination through Sequential Iron and Copper Catalysis

Abstract A mild, efficient and regioselective method for para‐amination of activated arenes has been developed through a combination of iron and copper catalysis. A diverse range of products were obtained from an operationally simple one‐pot, two‐step procedure involving bromination of the aryl substrate with the powerful Lewis acid iron(III) triflimide, followed by a copper(I)‐catalysed N‐arylation reaction. This two‐step dehydrogenative process for the regioselective coupling of aromatic C−H bonds with non‐activated amines was applicable to anisole‐, phenol‐, aniline‐ and acetanilide‐type aryl compounds. Importantly, the arene substrates were used as the limiting reagent and required no protecting‐group manipulations during the transformation.

The ability to efficiently and selectively form aryl CÀNb onds is an important challenge in organic chemistry owing to the widespread occurrenceo fthis key linkage in natural products, pharmaceuticals, agrochemicalsa nd organic materials. [1] Historically,aryl amination was conducted by nitration of the aromatic ring, followed by reduction but the use of strongly acidic conditions has limited the general application of this method. [2] Another approach is the copper-or palladium-catalysed amination of aryl (pseudo)halides through Ullmann-Goldberg, Chan-Evans-Lam or Buchwald-Hartwig protocols. [3,4] Althought his methodt ypically produces as ingle regioisomer, prefunctionalisation of the aromatic ring is required.
To circumvent prefunctionalisation or preoxidation of either partner,n ew methods for directi ntermolecular dehydrogenative coupling of aryl CÀHb onds with non-activated amines have recently been reported. [5] For example, highly regioselective intermolecular ortho-amination through transition-metalcatalysed chelation-directed aryl CÀHa ctivation under oxidative conditions hasb een developed for efficient CÀNb ond for-mation( Figure 1a). [6,7] Procedures for directeda ryl CÀHa nd NÀ Hb ond coupling at more distal positions are less well established.O nly recently has ad irected meta-aminationp rocess been reported [8] and only afew examples are known for directed para-amination utilising transition-metal-, organo-o rp hotocatalysis, as well as non-catalytic oxidative methods. [9] Many of these tend to use the aryl compound as the solvent and/or generate mixtures of regioisomers. Exceptions that overcome these issues are known [10] and include work by Suna and coworkersw ho have used the electrophilicr eaction of arenes with hypervalent iodonium reagentst of orm unsymmetrical diaryl-l 3 -iodane intermediatesthat are then subjected to acopper(I)-catalyseda mination( Figure 1b). [11] We were interested in developing ac atalyticm ethod for amination of aryl CÀH bondst hat utilised only readily available, inexpensive, nontoxic first-row transition metals.H ere, we now report ao ne-pot, two-step method for the coupling of aryl CÀHb onds with NÀ Hb onds by using ac ombination of iron(III)-and copper(I)-catalysis ( Figure 1c). This para-directed process is complementary to ortho-directed methods and is applicable for the general coupling of aryl compounds with N-heterocycles, amides and sulfonamides.
Before developing the one-pot para-amination process, the scope,e fficiencya nd regioselectivity of the iron(III)-catalysed aryl bromination reaction was initially assessed. We recently demonstrated that metal triflimide catalysis could be used to activate N-iodosuccinimide (NIS) for regioselective iodination of arenes and so as imilarp rocedure was used to investigate bromination (Scheme1). [12,13] Am ixture of iron(III) chloride (5 mol %) with the inexpensive ionic liquid 1-butyl-3-methylimi-  dazolium bis(trifluoromethylsulfonyl)imide ([BMIM]NTf 2 ), which forms the powerful Lewis acid Fe(NTf 2 ) 3 in situ, [14] was found to activate N-bromosuccinimide (NBS) for the rapid and efficient bromination of aw ide range of anisoles,p henols, anilinesa nd acetanilides under mild conditions. [15] Arenesw ith strongly deactivating groups (e.g., 1l and 1m), naphthalenes (1q and 1r) and 2,3-benzodihydrofuran(1s)w ere all tolerated as substrates for effective bromination.A part from anisole (1a), which yielded a9 5:5r atio of p-a nd o-isomers, the transformation was found to generate the aryl bromides as single regioisomers. A dibromination reaction was also investigated with cyanophenol (1t). The use of two equivalents of NBSr esulted in clean dibromination and the efficient synthesis of bromoxynil (2t), ac ommerciallyu sed nitrile herbicide. [16] Having shown the generality of the Fe(NTf 2 ) 3 -catalysed aryl bromination,c onditions for the one-pot,t wo-step amination process were then investigated (Table 1). Initially, the standard methodf or the brominationr eaction[ with FeCl 3 (5 mol %) and the ionic liquid as the solvent] was then followed with ac opper(I) iodide-catalysed amination with indole as the nucleo-phile and caesium carbonatea st he base (Table1,e ntry 1). Althoughb romination was successful, no aminationw as observed.T oa llow better compatibility between the two steps, ab romination protocol with catalytic amountso fb oth FeCl 3 and [BMIM]NTf 2 in toluene was developed. Through this strategy and under typical conditions (40 8C, 4h), complete bromination was observeda nd, when combined with the copper(I)catalysed amination, conversion to the coupled product 3a was observed (14 %, entry 2). Af urther reduction of the catalyst loading in either DMF or toluene resulted in improved overall conversions (entries 3a nd 4). However,asignificant breakthrough was achieved when water was added as ac osolventt ot he second step. This allowed dissolution of the base andb etter mixingo fr eagents,w ith the result of nearly quantitative conversion to 3a (entry 5). Further optimisation studies showedt hat the catalystl oading of both FeCl 3 (2.5 mol %) and [BMIM]NTf 2 (7.5 mol %) could be lowered further without affecting the rate or overall conversion of the one-pot process (entry 6).
Following optimisation studies, the scope of the one-pot amination process of anisole (1a)w ith various N-nucleophiles was explored (Scheme 2). With ar ange of N-heterocycles such as indole, pyrazole, imidazole, pyrrole and pyrrolidin-2-onea s well as amides and sulfonamides, the one-pot process was found to generate the coupledp roducts cleanly andi ng ood yields as single regioisomers. The use of this procedure for multigram synthesis of aminateda ryl compounds was also demonstrated with the large-scale synthesis of pyrazole 3b in essentially quantitative yield.
With pyrazole and benzamide as nucleophiles, the scope of the aryl coupling partner was then explored (Scheme 3). The one-pot process was found to afford para-aminated products in good yields for various anisole-, phenol-, aniline-and acetanilide-based coupling partners, in the presence of either activating or deactivatings ubstituents. In addition, despite the use of aryl compounds containing nucleophilic functional groups (e.g.,p henols, anilines), the one-pott ransformation afforded the N-coupled products cleanly.
Scheme1.Scope of iron(III)-catalysed bromination. As part of this research program, ap reliminaryi nvestigation into ortho-N-arylation by using substrates with the para-position blockedw as also conducted. The one-pot bromination and aminationo fp-chloroanisole (1c)w ith pyrazoler equired more forcing conditions but afforded 4m cleanly in 53 %y ield (Scheme 3). However,alimitation of this approachwas demonstrated in as imilar one-pot process with p-nitroaniline (1l), which yieldeda ni nseparable mixture of 1l and the coupled product 4n. [17] In this case, slower copper-catalysed ortho-amination resultsi nacompeting pathway involving reduction of the organocopperi ntermediate, as evidencedb yt he regeneration of 1l from the bromide intermediate. [18] Methodsf or the one-step amination of aromatic halidesi nvolvingi ron and copperc atalysis have been previously reported. Correa and Bolm reported an iron-catalysed N-arylation reaction that was later found to be catalysed by coppercontaminants, [19] whereas Ta illefere ta l. described an iron-copper cocatalysed process. [20] To investigate the individual role of each metal catalyst during the amination step of this current process, as eries of experimentsw as performed to mimic the second step of the one-pot process ( Table 2). As expected, these resultsc onfirmed that copper iodide is essential for efficient conversion to the N-coupled product. Although this does not rule out an iron-copper co-catalysed amination, the Taillefer study does emphasise that pre-complexation of iron(III) with the acetylacetonate ligand was criticalf or effective co-catalysis (use of FeCl 3 as ac o-catalyst afforded low yields). [20] In conclusion, ag eneral and efficient approach for the regioselectiveb romination of aryl compounds has been demonstrated by using iron(III) triflimide activation of NBS. Combination of am odified version of this transformation thatu tilises catalytic amountso fb oth FeCl 3 and [BMIM]NTf 2 with ac opper(I)-catalysed amination allowed the development of ao nepot, two-step dehydrogenative process for the regioselective coupling of aromatic CÀHb onds with non-activated amines.
This processw as found to be general fort he para-amination of ar ange of aromatic compounds in combination with various nitrogen nucleophiless uch as N-heterocycles, amides and sulfonamides. This study provides ac onceptually new,o ne-pot Scheme2.Scope of the one-pot aminationp rocess with anisole (1a).  approachf or the preparation of valuable synthetic building blocks that is complementary to ortho-CÀHamination methods and overcomes the limitations of the multistep and harsh conditions of more traditional chemistries.