Silver-catalysed trifluoromethylation of arenes at room temperature†

The trifluoromethyl group is valued for its ability to modulate the properties of diverse materials such as pharmaceuticals, agrochemicals and polymers. Aryl CF3 groups are electronwithdrawing, hydrophobic and generally very stable, all properties that can be harnessed in the design of biologically active molecules and functional materials. Synthetic methods for aryl and heteroaryl trifluoromethylation are thus critical to the discovery and production of new molecules of high value to society. Recent developments in metal-mediated trifluoromethylation have produced significant advances in this area, with common functional groups such as aryl boronic acids and halides undergoing efficient trifluoromethylation under palladium and copper catalysis. Metal-catalysed trifluoromethylation of unactivated C–H positions, by contrast, is significantly less developed and has great potential for accelerating medicinal and agrochemical syntheses. Despite some recent groundbreaking developments in this area, there is still great demand for the development of catalytic C–H trifluoromethylation methods that function under mild and simple conditions. We were interested in developing a catalytic trifluoromethylation based on silver; in contrast to its Group 11 neighbour copper there have been few reports on silver-mediated trifluoromethylation and none that we are aware of using silver catalysis. The redox catalysis of silver, comprising one electron steps between 0, +1, +2 and +3 oxidation states, has been scarcely exploited in synthesis relative to other late TMs and could offer productive catalytic pathways for trifluoromethylation. We have recently developed silver-catalysed decarboxylative C–H cross-coupling under oxidative radical conditions, and were keen to see if a similar approach was viable for C–H trifluoromethylation. We started with a screen of reaction conditions based around TMSCF3 as the trifluoromethylating agent. 9 The groups of Sanford, Bräse and Wang have recently demonstrated the compatibility of this reagent with stoichiometric silver salts, encouraging us that it could form the basis of a catalytic system. Using 1,4-dimethoxybenzene (1a) as the substrate, we conducted an initial solvent screen using combinations of AgF, TMSCF3 and PhI(OAc)2 (Table 1). We worked at room temperature under air throughout, with the aim of developing a mild reaction with as broad a functional group tolerance as possible. The reaction proved sensitive to solvent choice with initially only MeCN from a selection of common organic solvents producing any reaction (entries 1 and 2). DMSO proved more effective still, affording the trifluoromethylated compound 2a in 51% conversion (entry 3). Fluoride was not a requirement, with Ag2CO3 being similarly effective at promoting reaction (entry 4). Alternative oxidants did not improve on PhI(OAc)2 (entries 5 and 6), and the use of a

The trifluoromethyl group is valued for its ability to modulate the properties of diverse materials such as pharmaceuticals, agrochemicals and polymers. Aryl CF 3 groups are electronwithdrawing, hydrophobic and generally very stable, all properties that can be harnessed in the design of biologically active molecules and functional materials. Synthetic methods for aryl and heteroaryl trifluoromethylation are thus critical to the discovery and production of new molecules of high value to society. 1 Recent developments in metal-mediated trifluoromethylation have produced significant advances in this area, 2 with common functional groups such as aryl boronic acids and halides undergoing efficient trifluoromethylation under palladium and copper catalysis. 3 Metal-catalysed trifluoromethylation of unactivated C-H positions, by contrast, is significantly less developed and has great potential for accelerating medicinal and agrochemical syntheses. Despite some recent groundbreaking developments in this area, 4 there is still great demand for the development of catalytic C-H trifluoromethylation methods that function under mild and simple conditions. We were interested in developing a catalytic trifluoromethylation based on silver; in contrast to its Group 11 neighbour copper there have been few reports on silver-mediated trifluoromethylation 4g,k,5 and none that we are aware of using silver catalysis. The redox catalysis of silver, comprising one electron steps between 0, +1, +2 and +3 oxidation states, has been scarcely exploited in synthesis relative to other late TMs 6,7 and could offer productive catalytic pathways for trifluoromethylation. We have recently developed silver-catalysed decarboxylative C-H cross-coupling under oxidative radical conditions, 8 and were keen to see if a similar approach was viable for C-H trifluoromethylation.
We started with a screen of reaction conditions based around TMSCF 3 as the trifluoromethylating agent. 9 The groups of Sanford, Bräse and Wang have recently demonstrated the compatibility of this reagent with stoichiometric silver salts, 4g,k,5c encouraging us that it could form the basis of a catalytic system. Using 1, 4-dimethoxybenzene (1a) as the substrate, we conducted an initial solvent screen using combinations of AgF, TMSCF 3 and PhI(OAc) 2 (Table 1). We worked at room temperature under air throughout, with the aim of developing a mild reaction with as broad a functional group tolerance as possible. The reaction proved sensitive to solvent choice with initially only MeCN from a selection of common organic solvents producing any reaction (entries 1 and 2). DMSO proved more effective still, affording the trifluoromethylated compound 2a in 51% conversion (entry 3). Fluoride was not a requirement, with Ag 2 CO 3 being similarly effective at promoting reaction (entry 4). Alternative oxidants did not improve on PhI(OAc) 2 (entries 5 and 6), and the use of a nitrogen atmosphere led to a reduction in yield (entry 7). Crucially, sub-stoichiometric amounts of silver salts proved equally effective (entries 8-10), indicating that a catalytic reaction was feasible. We settled on conditions of AgF (25 mol%) with TMSCF 3 (2 equiv.) and PhI(OAc) 2 (2 equiv.), at room temperature (entry 9) to take forward. The use of the more stable (and expensive) TESCF 3 reagent gave only marginal improvement (entry 10), so we continued with the cheaper TMSCF 3 reagent.
Substrate scope investigations established that the procedure was effective for a variety of electron rich arenes with broad substrate scope tolerance ( Table 2). For unsymmetrical substrates isomeric mixtures were generally observed, with regioselectivities consistent with radical S Ar H addition (vide infra). Importantly, the reaction was compatible with halogen groups, illustrating an orthogonal reactivity to conventional C-X trifluoromethylations whereby neighbouring C-H bonds undergo preferential reaction. The useful building blocks 2f, 2g, 2h and 2i were prepared in this fashion. Electron-withdrawing groups such as aldehyde (2j), ketone (2k) and ester (2l) were likewise tolerated without problem. Importantly, dialkylanilines could be trifluoromethylated, a key class of building block that has rarely featured in C-H trifluoromethylation reports. 4i,10 A slight preference for ortho over para selectivity was observed for simple dimethylamine (2m), with bromo substitution also tolerated (2n) along with N-acylation (2o). We were pleased to observe that the reaction was also effective for un-activated arenes (2p, 2q and 2r), although these substrates did require an excess of the arene and the reaction temperature raised to 70 1C.
The reaction could be extended to heteroarenes with N-Me pyrroles in particular being excellent substrates (2s, 2t). Electronwithdrawing groups on the heteroarene nucleus were welltolerated (2t), but on nitrogen less so (N-Boc, 2u). Furans (2v), thiophenes (2w, 2x) and indoles (2y) were all productive, indicating that the method is viable for the major classes of p-excessive heterocycle. p-Deficient heteroarenes, by contrast, were not generally effective in the reaction but could be efficiently captured by masking the azine nucleus with electron-donating groups (2aa).
We next turned to the trifluoromethylation of more complex, biologically active molecules -a major driver for the development of new methods in this area. Introduction of the CF 3 group at unactivated C-H positions represents a very versatile approach to fluorine incorporation for modulation of biological activity, 4c,d,j demanding mild reaction conditions that are tolerant of functional groups and reasonable stoichiometries with respect to the (often valuable) C-H substrate. Accordingly, we extended the reaction to trifluoromethylate some more complex molecules in the agrochemistry field, an area where the CF 3 group is particularly prevalent. We could successfully incorporate the CF 3 group into the commercial herbicides pyriftalid 11 and napropamide 12 (Scheme 1). The functional group tolerance of the reaction was illustrated by sulfide, lactone and a-hydroxyamide functionality all being stable to the reaction conditions (Scheme 1, 3 and 4).
A radical mechanism is implicated for the trifluoromethylation reaction, 13 as radical quenching reactions using TEMPO and galvinoxyl radical both shut down the reaction, with the TEMPO-CF 3 adduct being clearly observed in the crude 19 F NMR. The electrophilic CF 3 radical usually (but with some exceptions) 4c,j displays a marked preference for electron rich substrates, as seen here, underlining the likelihood of a radical pathway. A possible Table 2 Ag-catalysed trifluoromethylation: substrate scope a,b a 1 (0.3 mmol), TMSCF 3 (0.6 mmol), PhI(OAc) 2 (0.6 mmol), AgF (0.075 mmol), DMSO (1.0 mL), room temperature, 20 h. b Isolated yields. For isomer mixtures, the minor regioisomeric position is labeled with *. c Yields determined by 19 F NMR using 4-fluoroanisole as the internal standard. d Reaction conducted at 70 1C, 5-10 equiv. of arene.
Scheme 1 Agrochemical trifluoromethylation. mechanism is shown in Scheme 2 whereby TMSCF 3 is oxidised to the CF 3 radical, followed by S Ar H addition, then a second one electron oxidation and proton loss to give the product 2. Control experiments to investigate the role of silver in the first step of the proposed mechanism indicated that AgF alone was insufficiently oxidizing to generate CF 3 (mixing AgF with TMSCF 3 in the presence of TEMPO in DMSO at room temperature gave only trace quantities of TEMPO-CF 3 ). PhI(OAc) 2 alone was moderately effective (44% NMR yield of TEMPO-CF 3 ) and the combination of PhI(OAc) 2 and AgF highly effective (91% NMR yield). 14 The background oxidizing activity of the hypervalent iodine reagent could be quantified in the trifluoromethylation of 1,4-dimethoxybenzene 1a in the absence of any silver salt, producing a low conversion to the trifluoromethylated product 2a (26% NMR yield). Alternative mechanisms were investigated by treating dimethoxyanisole 1a with in situ prepared AgCF 3 5b in both MeCN and DMSO as solvents. No reaction could be observed in each case, suggesting organometallic AgCF 3 intermediates are not participating under our reaction conditions. A further control experiment with Togni's reagent 4a in DMSO at room temperature gave no reaction, ruling out simple S E Ar attack on an electrophilic CF 3 source. Finally, we considered the possibility of initial arene oxidation by PhI(OAc) 2 , followed by CF 3 anion addition to a cationic arene intermediate.
Extensive work by Kita has demonstrated the C-H functionalization of electron rich arenes using PhI(TFA) 2 in the presence of stoichiometric BF 3 ÁOEt 2 and nucleophiles. 15 It seems the present conditions are not sufficiently oxidizing to enable an analogous pathway, as a control reaction in the absence of TMSCF 3 gave no reaction, where some degree of homocoupling would be expected if this mechanism was in operation.
In conclusion, we have developed a silver-catalysed trifluoromethylation system for electron rich aromatic and heteroaromatic substrates. The reaction works at room temperature under air, does not require excessive stoichiometries of substrate or reagent, and is operationally simple to carry out. The application of this chemistry to new trifluoromethylation substrates will be the subject of future work in our laboratory.
We thank Syngenta, the University of Manchester and the EPSRC for funding (Leadership Fellowship to M.F.G.), and the EPSRC mass spectrometry service at the University of Swansea.