Method developmentInitial investigation of the deoxytrifluoromethylation/aromatization sequence exploited 4-phenylcyclohexanone as a model substrate, and each individual step (1,2-addition reaction, dehydration, and aromatization) was independently optimized. The 1,2-addition reaction using TMSCF3 as a reagent and catalytic tetra-n-butylammonium fluoride (TBAF) in tetrahydrofuran (THF) delivered a diastereomeric mixture of trifluoromethyl alcohol and trifluoromethyl silyl enol ether intermediate in quantitative yield (Table 1. Entry 1.1)37. From this mixture, a one-pot dehydration and aromatization could presumably deliver the desired Ar–CF3 substructure. Extensive screening established an initial hit using p-toluenesulfonic acid monohydrate (PTSA•H2O) as a dehydrating reagent and 2,3-dichloro-5,6-dicyano-p-benzoquinone (DDQ) as an oxidant (1 equiv. each) at 110 °C in o-xylene (0.1 M), which afforded the desired Ar–CF3 product in 5% yield (Entry 1.2)38. Optimization of the conditions (3 equiv. DDQ, 1 equiv. PTSA•H2O, 0.2 M toluene, reflux) improved the yield of this step to 70% (Entry 1.3). To develop a direct deoxytrifluoromethylation/aromatization sequence, we identified a single solvent that would facilitate all steps in the sequence. Since the 1,2-addition reaction proceeded in THF and the dehydration/aromatization proceeded in toluene, these two solvents were each explored in the one-pot sequence. However, no desired product was obtained in THF, and in toluene, only 4% product was obtained (Entry 1.4 and 1.5). We speculated that the trace amount of THF present in TBAF might react undesirably with DDQ and thus hinder the oxidation step of the sequence. Thus, other TMSCF3 activators were explored for the 1,2-addition step, and the combination of TMSCF3 with 10% CsF in toluene provided a quantitative of yield of a mixture of trifluoromethyl alcohol and trifluoromethyl silyl ether intermediate (Entry 1.6). With no workup, subjection of this crude reaction mixture directly to PTSA•H2O/DDQ increased the yield of the one-pot sequence to 14% (Entry 1.6). Switching the solvent from toluene to o-dichlorobenzene (o-DCB) increased the yield to 62% (Entry 1.7), and finally increasing the reaction temperature to 140 °C provided a quantitative yield for the one-pot sequence (Entry 1.8, Conditions A).Table 1 Deoxytrifluoromethylation/aromatization of activated substratesAs the model substrate, 4-phenyl cyclohexanone bears a benzylic position that presumably facilitates the final oxidation reaction with DDQ39, and we speculated that this oxidation might not translate to substrates that lack such stabilizing groups. Hence, a non-activated substrate, 4-t-butylcyclohexanone was used for further development of the reaction. Initially, the 1,2-addition reaction proceeded in quantitative yield using a combination of TMSCF3 and catalytic CsF in o-DCB. However, as expected, treatment of the intermediate mixture with PTSA•H2O/DDQ did not promote dehydration and aromatization to yield an Ar–CF3 derived product (Entry 1.9).Thus, we developed an alternate set of conditions using 4-t-butylcyclohexanone as a substrate. Specifically, 1,2-addition reaction using TMSCF3 and stoichiometric TBAF afforded trifluoromethyl alcohol intermediate in quantitative yield. Dehydration of the trifluoromethyl alcohol intermediate using thionyl chloride (SOCl2) and pyridine proceeded smoothy to deliver a vinyl–CF3 intermediate in quantitative yield (Table 2)40. However, oxidation of the vinyl–CF3 intermediate to the Ar–CF3 did not proceed using a variety of chemical oxidants {N-chlorosuccinimde, (2,2,6,6-tetramethylpiperidin-1-yl)oxyl, azobisisobutyronitrile (AIBN), benzoquinone-derived oxidants, etc.}, or using some Pd-catalyzed oxidation methods41. Ultimately, allylic functionalization and elimination successfully aromatized the vinyl–CF3 intermediate to the target Ar–CF3 product. As an initial hit, allylic bromination and elimination using 1 equiv. N-bromosuccinimde (NBS) and catalytic AIBN afforded 49% of the desired Ar–CF3 product (Entry 2.1)42, and increasing the loading of NBS to 3 equiv. afforded 96% yield of the desired product (Entry 2.2). Attempts to telescope this sequence to a single-pot operation were met with additional challenges, as the 1,2-addition and dehydration steps proceeded smoothly, though aromatization would not occur (Entry 2.3). After the formation of vinyl–CF3 intermediate, removal of THF and switching to o-DCB also did not afford any desired final product (Entry 2.4). To solve this issue, after the dehydration step, filtration of the reaction mixture through a silica plug and subsequent oxidation using NBS/AIBN afforded 82% overall yield (Entry 2.5, Conditions B). This reaction sequence also converted our original model substrate, 4-phenylcyclohexanone, to the anticipated product in 81% overall yield (Entry 2.6). Additionally, using the silica plug filtration step, DDQ could also be used as the oxidant for 4-phenylcyclohexanone affording 92% yield (Entry 2.7, Conditions C). In summary, we established three complimentary sets of conditions for transforming cyclohexanone precursors to Ar–CF3 products (Entries 1.8, 2.5, 2.7), which would become beneficial when encountering new substrates, functional groups, and/or substitution patterns that might not work for a single set of conditions.Table 2 Deoxytrifluoromethylation/aromatization of both activated/non-activated substratesDeoxytrifluoromethylation/aromatization reactions to access a diverse array of productsThe three optimized conditions (Fig. 2, Conditions A–C) enabled conversion of a variety of cyclohexan(en)ones to the corresponding Ar–CF3 products bearing many substitution patterns and useful functional groups. Typically, small-scale reactions (0.05–0.1 mmol) were run to assess the suitability of the conditions per substrate, and minor modifications in time, temperature, and reagent equivalents improved the yields for under-performing substrates. The best conditions were then repeated on a higher scale (0.5 mmol). Substrates bearing bulky aryl substituents at the 2-, 3-, and 4- positions of the ketone (1S–3S), as well as alkyl substituents (4S, 5S), were tolerated. A 2,6-disubstited substrate (6S) and one bearing fused saturated rings (7S) both reacted to provide good yields of products. Several important functional groups, cyano (8P), alkynes (9P), nitro (10P), methoxy (11P), and ethoxy (14P) groups, were also tolerated, as well as halogens (12P–13P, and 15P) that provide opportunities for further transformations via transition metal-catalyzed coupling reactions. Additionally, many aromatic and aliphatic heterocycles, were tolerated, including thiophene (16P), indole (17P), and morpholine (18P; for more see Fig. 3). Though methyl and phenyl esters were incompatible with the initial 1,2-addition reaction, the use of a t-butyl ester (19S) allowed the 1,2-addition reaction with CF3TMS to proceed in quantitative yield. Subsequently, subjection of the corresponding t-butyl ester intermediate to thermal/acidic aromatization (Conditions A) promoted aromatization and concurrent release of isobutylene to deliver benzoic acid product (19P1). In contrast, subjecting the intermediate to DDQ alone, without PTSA•H2O, delivered the t-butyl ester product (19P2) in moderate yield. Of note, though many substrates underwent the deoxytrifluoromethylation/aromatization sequence from cyclohexanone-based substrates (1S–5S, 8S, 10S–11S), cyclohexenone-based substrates (6S–7S, 9S, 14S–19S) were also converted to the corresponding Ar–CF3 compounds, which provides a complementary option for certain applications. In a related fashion, the reaction conditions converted tetralone derivatives (12S–13S), which also bear an additional level of unsaturation, to naphthyl trifluoromethane products in good yields (12P–13P). Generally, the aryl-substituted substrates proceeded under all three sets of Conditions A–C, whereas the alkyl-substituted cyclohexanone only proceeded under Conditions B.Fig. 2: The deoxytrifluoromethylation/aromatization strategy tolerates many useful functional groups and substitution patterns.All reactions were run in 0.5 mmol batch of substrate. Conditions A: i) TMSCF3 (1.1 equiv.), cat. CsF, o-DCB (0.5 M), rt–50 °C, 4–36 h. ii) PTSA•H2O (0–2 equiv.), DDQ (2–4 equiv.), o-DCB (0.2–0.1 M), 120–140 °C, 14–48 h. Conditions B: i) TMSCF3 (1.1 equiv.), TBAF (1 equiv.), THF (0.5 M), rt, 4–12 h. ii) SOCl2 (3 equiv.), Pyridine (3 equiv.), 10 mol% DMAP, THF (0.5 M), 50 °C, 18–24 h. Then silica filter. iii) NBS (2–4 equiv.), 10 mol% AIBN, o-DCB (0.2 M), 120 °C, 14 h. Conditions C: i) TMSCF3 (1.1 equiv.), TBAF (1 equiv.), THF (0.25 M), rt, 4 h. ii) SOCl2 (3 equiv.), Pyridine (3 equiv.), 10 mol% DMAP, THF (0.25 M), 50 °C, 18 h. Then silica filter. iii) DDQ (2 equiv.), o-DCB (0.2 M), 110 °C, 14 h. a For step ii, PhMe used as solvent at 110 °C. b 19F NMR yield using Conditions B. c For step iii, 1,2-DCE used as solvent at 90 °C. d No PTSA•H2O was used. e Both 19P1 and 19P2 were accessed from same substrate 19S bearing a t-butyl ester.Fig. 3: The deoxytrifluoromethylation/aromatization strategy pushes the limits of accessible Ar–CF3 compounds.The transformation enables (A) conversion of readily available ketone substrates to Ar–CF3 derivatives, (B) systematic elaboration of simple precursors to highly functionalized Ar–CF3 products, (C) conversion of natural products into Ar–CF3 derivatives, (D) annulation reactions to access highly substituted substrates that convert to Ar–CF3 products, and (E) conversion of heterocyclic ketones to (Het)Ar–CF3. All reactions were run in 0.5 mmol batch of substrate. Conditions A: i) TMSCF3 (1.1 equiv.), 10 mol% CsF, o-DCB (0.5 M), rt–50 °C, 4–48 h. ii) PTSA•H2O (2 equiv.), DDQ (3 equiv.), o-DCB (0.2 M), 120 °C–140 °C, 12–24 h. Conditions B: (i) TMSCF3 (1.1 equiv.), TBAF (1 equiv.), THF (0.5 M), rt, 4 h. (ii) SOCl2 (3 equiv.), Pyridine (3 equiv.), 10 mol% DMAP, THF (0.5 M), 50 °C, 18 h. Then silica filter. (iii) NBS (4 equiv.), 10 mol% AIBN, o-DCB (0.2 M), 120 °C, 18 h. Conditions C: (i) TMSCF3 (1.1 equiv.), TBAF (1 equiv.), THF (0.25–0.5 M), rt–35 °C, 4–12 h. (ii) SOCl2 (3 equiv.), Pyridine (3 equiv.), 10 mol% DMAP, THF (0.25–0.5 M), rt–60 °C, 8–24 h. Then silica filter. (iii) DDQ (2–3 equiv.), o-DCB (0.2 M), 90–120 °C, 12–48 h. a THF was used instead of o-DCB in step 1 followed by basic alumina plug filter and swapping of solvent with o-DCB. b 1,2–DCE was used as a solvent in step 3.Most significantly, the deoxyfluoroalkylation/aromatization strategy afforded Ar–CF3 compounds that either cannot be accessed by currently available reactions or that are impractical due to limited availability of starting substrates (Fig. 3). For instance, tetralone substrates 20S–22S are commercially available, and can be readily converted to previously unreported trifluoromethyl naphthalene products (20P–22P) in single-pot operations in good yields (Fig. 3A; crystal structure of 22P in SI, CCDC 2366007). However, the corresponding halides or carboxylate derivatives that would be needed for metal-catalyzed transformations are either (a) not commercially available, (b) available for unreasonably high costs (e.g. $1500/mg), or (c) have not previously been reported (except halide precursors for 20P). Additionally, highly substituted cyclohexenone 23S was readily converted to tetrasubstituted arene 23P in good yield (Fig. 3A); though the corresponding aryl halides and aryl carboxylic acid derivatives of 23P are not available commercially and the routes to access them are not known.Combined with a plethora of classical organic functionalization reactions of cyclohexan(en)ones, the deoxytrifluoromethylation/aromatization sequence provides countless opportunities for delivering highly functionalized Ar–CF3 products, most of which are not readily accessible by other methods due to (a) issues with the presence of functional groups that are prone to decompose under harsh conditions (e.g. SF4, Cl2/SBF4, HF), (b) limitations in commercial availability of aryl halide precursors and/or synthetic challenges associated with preparation of appropriate aryl halides, and/or (c) issues with regioselective preparation of certain isomers. For example, starting from 5-phenyl-1,3-cyclohexanedione, routine halogenation and cross-coupling reactions can systematically program the placement of substituents (23S–26S, Fig. 3A, B), and application of the deoxytrifluoromethylation/aromatization sequence subsequently delivers the corresponding Ar–CF3 derivatives in good yields (23P–26P). With synthetic creativity, the deoxytrifluoromethylation/aromatization sequence can be combined with routine organic transformations to convert cyclohexan(en)one-containing natural products derivatives, such as carvone into custom and highly substituted trifluoromethyl arenes (27P, Fig. 3C)43. In addition to functionalization reactions of cyclohexan(en)ones, complex substrates can also be accessed from natural product derivatives via Birch reduction of aromatic ethers. For instance, Birch reduction of dimethylated estradiol delivered cyclohexenone substrate (28S), which was then subjected to deoxyfluoroalkylation-aromatization sequence)44 – in this case using the Burgess reagent to promote the elimination – to yield valuable and previously inaccessible Ar–CF3 product (28P, Fig. 3C). This strategy demonstrated that aromatic ethers (even in complex natural products) could be reduced to cyclohexenones via a Birch Reduction and be subsequently subjected to deoxyfluoroalkylation/aromatization sequence to access Ar–CF3 products, providing a complimentary approach to cross-coupling chemistry. Considering the abundance of phenols and aromatic ethers in natural products, commercially available building blocks, and therapeutic candidates, many opportunities exist to access fluoroalkyl derivatives of known O-based compounds.An additional entry to accessing highly substituted Ar–CF3 products involves the use of regioselective annulation strategies to generate highly substituted cyclohexan(en)one substrates that can then undergo deoxyfluoroalkylation/aromatization to produce highly substituted, currently inaccessible products (Fig. 3D). For instance, highly substituted cyclohexan(en)ones were regioselectively prepared using Pd-catalyzed net [4 + 2] cycloaddition (29S)45, Co-catalyzed [5 + 1] annulation (30S)46, and Robinson annulation (31S–32S) reactions47, and subjection of the respective cyclohexan(en)ones to the deoxyfluoroalkylation/aromatization sequence delivered the corresponding highly substituted Ar–CF3 products in good yields (29P–32P), all of which are inaccessible by known routes. Notably, these highly substituted products (20P–23P, 25P–27P, 29P–32P) cannot be synthesized through previously reported methods, due to the aforementioned issues with harsh conditions, control of regiochemistry, and/or lack of functionalized substrates for the transformations (Fig. 1A). Further, in most cases, the corresponding aryl halide precursors required for cross-coupling reactions are not currently known, and accessing such aryl halides would require development of de novo routes that present notable synthetic challenges.The deoxyfluoroalkylation/aromatization strategy also translates to N-containing heterocycles that are found in a wide variety of biologically active compounds (Fig. 3E). For example, subjection of the FDA-approved drug, ondansetron (33S), which bears imidazole and carbazole rings, and quinolone substrate (34S) effectively afforded the corresponding trifluoromethyl heteroarene products in good yield (33P–34P). Though these examples involve the formation of benzannulated systems, the deoxyfluoroalkylation/aromatization sequence can also convert piperidinones (35S) to trifluoromethyl pyridines (35P), suggesting that a range of trifluoromethyl heterocycles can be generated from this strategy.
