Radical cascade synthesis of azoles via tandem hydrogen atom transfer

A radical cascade strategy for the modular synthesis of five-membered heteroarenes (e.g. oxazoles, imidazoles) from feedstock reagents (e.g. alcohols, amines, nitriles) has been developed. This double C–H oxidation is enabled by in situ generated imidate and acyloxy radicals, which afford regio- and chemo-selective β C–H bis-functionalization. The broad synthetic utility of this tandem hydrogen atom transfer (HAT) approach to access azoles is included, along with experiments and computations that provide insight into the selectivity and mechanism of both HAT events.


I. General Information
All chemicals and reagents were purchased from Sigma-Aldrich, Alfa Aesar, Acros, TCI, or ChemImpex. DCE, MeCN and Et3N were distilled over CaH2 before use. PhMe, CH2Cl2, and Et2O were dried using an Innovative Technology solvent system. Silicycle F60 (230-400 mesh) silica gel was used for flash column chromatography. Thin layer chromatography (TLC) analyses were performed using Merck silica gel 60 F254 plates and visualized under UV (254 nm) or KMnO4 stain. 1 H, 19 F, 13 C NMR spectra were recorded using a Bruker AVIII 400 or AVIII 600 MHz NMR spectrometer. 1 H NMR and 13 C NMR chemical shifts are reported in parts per million and referenced with respect to CDCl3 ( 1 H: residual CHCl3 at δ 7.26, 13 C: CDCl3 triplet at δ 77.16) or DMSO-d6 ( 1 H: residual DMSO quintet at 2.50, 13 C: DMSO-d6 sept at 39.51). 1 H NMR data are reported as chemical shifts (δ ppm), multiplicity (s = singlet, bs = broad singlet, d = doublet, t = triplet, q = quartet, quint = quintet, sept = septet, m = multiplet, app t = apparent triplet, app q = apparent quartet, app qd = apparent quartet of doublets), coupling constant (Hz), relative integral. 19 F NMR data are reported as chemical shifts (δ ppm). High resolution mass spectra were obtained using Bruker MicrOTOF (ESI). IR spectra were recorded using a Thermo Fisher Nicolet iS10 FT-IR and are reported in terms of frequency of absorption (cm -1 ). Melting points were determined using a Laboratory Devices MEL-TEMP II.
Photochemical reactions were performed by placing reaction vessels approximately 5 cm away from two 23W TCP Model EDXO-23 compact fluorescent lightbulbs (CFL, 0.380 A, 1450 lumens). Reaction temperature was maintained at approximately 23 °C by with two fans (Figure S1). To address challenges with benzimidate synthesis, a new method was developed (see Section III for further discussion) To a 4-dram vial containing a stir bar was added alcohol (1 equiv.), nitrile (1.1 equiv.), PhMe (0.5 M), and triflic acid (1.2 equiv.). The solution was heated to 110 °C and stirred. After 24 h the solution was cooled to room temperature, and then stored at -15 °C until crystallization of the hydrotriflate salt was observed. The salt was then isolated via filtration and washed with cold hexanes and Et2O; residual solvent was removed under vacuum. The salt was suspended in Et2O (0.1 M), and NaHCO3 (sat. aqueous) was added dropwise until the dissolution of the salt observed (typically ~5 minutes). The aqueous phase was extracted with CH2Cl2, and the combined organic phases were dried over Na2SO4, concentrated under vacuum, and then used as is, or purified via column chromatography (silica gel treated with 1% Et3N/hexanes to avoid hydrolysis).

General Procedure for Transimidation (GP2)
Step 1: To a pressure tube equipped with a stir bar was added nitrile (1 equiv.), trifluoroethanol (12 equiv.), and acetyl chloride (8 equiv.). The solution was heated to 80 °C and stirred. After 48 h the reaction was cooled to room temperature and carefully vented (Note: HCl gas is formed as a by-product, see below for additional instructions on safe handling), which immediately induced precipitation of the benzimidate hydrochloride salt. The benzimidate salt was collected via filtration with cold hexanes.
Caution: The pressure tube reactions were vented by bubbling through a solution of saturated NaHCO3 to neutralize the by-product HCl gas. Even at room temperature, venting may be violent.
Step 2: To a 2-dram vial equipped with a stir bar was added trifluoroethyl benzimidate hydrochloride salt (1 equiv.), alcohol (1 equiv.), and MeCN (0.16 M). The reaction was heated to 50 °C and stirred. Reaction progress was monitored by consumption of starting trifluoroethyl benzimidate via crude 1 H NMR. Upon completion, the solution was concentrated and the resulting crude solid was suspended in dry Et2O and subjected to the free-base protocol from GP1. The crude reaction mixture was then purified via column chromatography (silica gel treated with 1% Et3N in hexanes to avoid hydrolysis).
Note: Transimidation is time-sensitive. Prolonged reaction times lead to decomposition.

III. Imidate Synthesis Optimization
To improve the efficiency, handling, and safety of our benzimidate synthesis, we developed a triflic acid variant of the Pinner reaction. 1 To activate benzonitrile (pKa of conjugate acid = -10), 2 we used triflic acid (pKa = -14) as opposed to HCl (pKa = -8). Triflic acid also has the significant advantage of being a liquid, preventing the need to use a sealed pressure tube or venting of gaseous species upon work-up.

S7
Decomposition of the free-based imidate to esters and amides was observed after a few months, while the salt was stable indefinitely.
In the course of substrate exploration, we observed that secondary or benzylic alcohols afforded amides via cationic rearrangement, but primary alcohols (including phenol) and thiols formed imidate salts with varying efficiency (Figure S2). 2-(o-tolyl)ethan-1-ol (0.14 g, 1.0 mmol) was subjected to GP2, with acidification after transimidation to facilitate removal of unconsumed alcohol (see below). Upon completion, the free-based crude was concentrated, dissolved in dry Et2O, and acidified with 2M HCl in Et2O. Upon precipitation, the hydrochloride salt was isolated via filtration, and washed with cold hexanes and Et2O, yielding the benzimidate salt (0.17 g, 63%) as a white solid. The benzimidate salt was free-based according to GP1; no purification was needed, yielding benzimidate S4 (0.14 g, 98%) as a clear oil.
Note: The benzimidate salt was found to rapidly tautomerize to the corresponding amide under free-base conditions.
Note: The phenethyl pivalimidate hydrotriflate salt took over a week to precipitate out at -15 °C. Alkyl imidates are generally more hydrolytically unstable than their corresponding benzimidates. As such, column chromatography for this substrate was done quickly, in less than 3 minutes. The crude free-based material can be used without any observable difference in reactivity.
Note: the transimidation is temperature sensitive; at 50 °C the alcohol is not consumed, and at 80 °C, the imidate decomposes into benzamide. Multiple columns may be needed to obtain pure benzimidate.

Solvent Effects with CsI
Benzimidate 3 (0.2 mmol) was subjected to GP3, with different solvents listed below. Upon completion, the crude mixture was quenched, concentrated, and quantified via 1 H NMR using 1 equiv of 1,2-dichloroethane as an internal standard.

Counter-Ion Effects
Benzimidate 3 (0.2 mmol) was subjected to GP3, with different iodide salts listed below. Upon completion, the crude mixture was quenched, concentrated, and quantified via 1 H NMR using 1 equiv of 1,2-dichloroethane as an internal standard. Note: NaI affords a variable range of yields, while CsI affords both more consistent and higher yields Figure S3. Summary of improved yields with CsI compared to NaI, a 1 H NMR yields with 1 equiv 1,2-dichloromethane as an internal standard

Light, Thermal, and Oxygen Controls
Benzimidate 3 was subjected to GP3, with different conditions listed below. Upon completion, the crude mixture was quenched, concentrated, and quantified via 1 H NMR using 1 equiv of 1,2dichloroethane as an internal standard.

Oxidant Loading and Catalytic I2 Studies
Benzimidate 3 was subjected to GP3, with different loadings of CsI and PhI(OAc)2 listed below. Upon completion, the resultant crude mixture was concentrated and quantified via 1 H NMR using 1 equiv of 1,2-dichloroethane as an internal standard. Intermediate oxazoline 2 was subjected to oxidation conditions which have been reported in the literature for oxazoline to oxazole conversion.
Benzimidate 3 was subjected to conditions reported in the literature for PhI(OAc)2 oxidation with catalytic amounts of I2 listed below. 7 Upon completion, the resultant crude mixture was concentrated and quantified via 1 H NMR using 1 equiv of 1,2-dichloroethane as an internal standard.

Time Studies
Benzimidate 3 was subjected to GP3, and quenched at different times listed below. Upon completion, the resultant crude mixture was concentrated and quantified via 1 H NMR using 1 equiv of 1,2-dichloroethane as an internal standard.  Longer runtimes did not result in a decreased yield of oxazole 3, indicating the stability of the product under reaction conditions.
The reaction was heated to 50°C (by placing vial in an aluminum heating block) and stirred for 2 hours. Upon completion, the reaction was quenched with 10% aq. Na2S2O3, extracted with EtOAc, and washed with H2O. The crude was then purified via column chromatography (silica gel treated with 1% Et3N in hexanes), yielding oxazoline 2 (0.34 g, 85%) as a light yellow oil. Characterization data is consistent with reported literature data.
Characterization data is consistent with reported literature data.

5-methyl-2,4-diphenyloxazole (40)
The two-step protocol outlined in IX. Mechanistic Studies: Steric Effects of Benzimidates was used to obtain oxazole 40 in 58% via 1 H NMR with 1,2-dichloromethane as internal standard. The crude material was worked up according to GP3. Purification via column chromatography yielded oxazole 40 as a white solid.

Tandem Oxidation from Benzimidate Hydrotriflate Salt
In the development of a one-pot procedure, we explored the efficiency of the tandem oxidation from the benzimidate hydrotriflate salt. Benzimidate hydrotriflate salt S2 was subjected to GP3 with the addition of base described in Table S9. H NMR with 1 equiv of 1,2-dichloroethane as internal standard, parentheses denotes isolated yield.

One-pot Conversion of Alcohol to 2-Phenyl Oxazole
2-phenylethan-1-ol (24.0 μL, 0.2 mmol) was subjected to GP1 for 24 hours. The crude was reduced under vacuum and immediately subjected to GP3 with the following modification: addition of NaOAc (33 mg, 0.4 mmol) to effect the freebase in situ. After the tandem oxidation was complete, analysis of the crude by 1 H NMR with 1 equiv of 1,2-dichloroethane as an internal standard gave 45% yield. Purification via column chromatography yielded oxazole 3 (17.6 mg, 40%) as a white solid.
Modified Procedure 1: The crude imidate hydrotriflate salt from (i) was first filtered and residual solvent removed under vacuum. The washed salt was then subjected to GP3 with the following modification: addition of NaOAc (33 mg, 0.4 mmol) to effect the freebase in situ. After the tandem oxidation was complete, purification via column chromatography yielded oxazole 3 (27.3 mg, 62%) as a white solid.
Modified Procedure 2: The crude imidate hydrotriflate salt from (i) was first filtered and residual solvent removed under vacuum. The washed salt was then free-based according to GP1. The resulting imidate was then subjected to GP3. After the tandem oxidation was complete, purification via column chromatography yielded oxazole 3 (35 mg, 78%) as a white solid.

General Procedure for Derivatization of Trichloromethyl Oxazoles (GP4)
To a 2-dram vial equipped with a stir bar was added oxazole 28 (0.2 mmol, 1 equiv.), NaH2PO4•H2O (0.6 mmol, 3 equiv.), amine (0.8 mmol, 4 equiv.), and MeCN (0.1 M). The reaction was stirred at 80 °C and progress was monitored by TLC. Upon completion, the reaction was quenched with water and extracted with ethyl acetate. The combined organic layers were dried over Na2SO4, concentrated, and then purified via column chromatography (silica gel with ethyl acetate and hexanes).
Additionally, the appearance and disappearance of oxazoline 2 under short time points (Table  S7) and quantitative yield of oxazole 3 from 2 under our normal tandem reaction conditions strongly support oxazolines as an intermediate product en route to oxazoles.
Analysis by 1 H NMR showed consumption of starting material but did not show a clear oxazole major product, and TLC confirms a complex mixture. Coupled with the absence of observed diiodide in previous reactions, a mechanism invoking di-iodination followed by substitution and elimination to the oxazole is unlikely.

S44
To probe if the second HAT event was mediated by residual imidate starting material, the TFE benzimidate was added in sub-stoichiometric quantities to the reaction mixture to observe if it had an effect on the efficiency of the oxazoline to oxazole transformation. The yields obtained are comparable to reactions without additional TFE benzimidate and thus does not support the hypothesis that the second HAT event is mediated by intermolecular imidate HAT.
This substrate afforded only oxazoline 35, with no additional functionalization at the α-oxygen position, indicating that HAT is not efficient at this position.

Deuterated Solvent Experiments
Oxazoline 2 (23 mg, 0.1 mmol) was subjected to GP3 with the following modifications, PhMe-d8 was used instead of PhMe. Upon completion, the crude mixture was quenched, concentrated, and quantified via 1 H NMR using 1 equiv of 1,2-dichloroethane as an internal standard.
Oxazoline 2 (23 mg, 0.1 mmol) was subjected to GP3 with the following modifications, MeCN-d3 was used instead of PhMe. Upon completion, the crude mixture was quenched, concentrated, and quantified via 1 H NMR using 1 equiv of 1,2-dichloroethane as an internal standard.
To explore the role of solvent as a hydrogen atom source, and therefore quenching any radical intermediates generated under our reaction conditions and suppressing the second HAT, we employed PhMe-d8 and MeCN-d3 to observe any possible deuterium trapping. No deuterium incorporation was observed in either reaction suggesting that solvent C-H bonds are not serving as a hydrogen atom source.

Benzylic Hydrogen Atom Transfer via Other Oxidants
Oxazoline 2 (23 mg, 0.1 mmol) and chloranil (49 mg, 0.2 mmol) were added to an 8 mL vial with stir bar. The vial was degassed and DCE (1 mL) was added. The reaction was then stirred for 15 h at 80 o C. The crude mixture was quenched with 1M NaOH and extracted with CH2Cl2, dried over MgSO4 and concentrated. The reaction mixture was quantified via 1 H NMR using 1 equiv of 1,2dichloroethane as an internal standard.
To probe whether the conversion of the intermediate oxazolines to oxazoles occurs via an intermolecular HAT, we subjected oxazoline 2 to oxidants known to perform HAT at benzylic positions. Chloranil has been shown to oxidize benzylic C-H bonds, and it performs well in the aromatization of the intermediate oxazoline.
Additionally, DDQ at 50 °C in CH2Cl2 was found to decompose the oxazoline.

Steric Effects of Secondary Benzimidate
Benzimidate 38 (48 mg, 0.2 mmol) was added to an 8 mL vial and heated at 50 °C for 3 h with PhI(OAc)2 (77 mg, 0.24 mmol) in 1 mL of an I2 solution (0.01 M in DMF). The solvent was removed and residue was purified by column chromatography to yield a diastereomeric mixture of syn/anti oxazoline 39 (34 mg, 72%, dr = 1.4:1, syn:anti) as a clear oil. The mixture of syn/anti oxazoline 39 (31 mg, 0.13 mmol) was subjected to GP3. The reaction crude was quantified via 1 H NMR using 1,2-dichloroethane as an internal standard, yielding 58% oxazole 40 and 30% of anti oxazoline 39. Figure S4. Top spectra is of the mixture of syn/anti 39 prior to subjection to GP3. Middle spectra is of the crude reaction after subjection to GP3. Bottom spectra is of independently synthesized oxazole 40.

X. Substrate Limitations
In the course of substrate exploration, the following substrates afforded oxazolines, but not oxazoles. The above data were collected from Reaxys for the indicated structural motifs on 14 Mar 2019 with the following options: As Drawn, Stereo, Additional Ring Closures, Related Markush, Salts, Charges. The results were filtered by "product for purchase" and "SigmaAldrich".

Computational Methods
All calculations were performed using the Gaussian 16 (revision A.03) suite of programs. 9 Geometry optimizations and frequency calculations were performed using the ωB97X-D functional 10 with the 6-311++G(d,p) basis set 11 (Def2-TZVPP for iodine), 12 and the default PCM solvation model for PhMe, 13 and an integration grid of "UltraFine" to help minimize uncertainty in computed free energies. All stationary points were confirmed to have no imaginary frequencies.
Transition states were confirmed as first order saddle points with the presence of a single imaginary frequency, and scanned in both directions along their intrinsic reaction coordinate. Reported Gibbs free energies and enthalpies in solution include thermal corrections computed at 298.15 K and 1 atm. Visualization carried out with CYLview. 14

Oxidation Potentials
The oxidation potentials of oxazole and imidazole were calculated according to the method described in the literature 15

Reaction Modeling
Activation energies (ΔG ‡ ) were calculated as the difference in free energies between the optimized transition state and its preceding ground state. Reaction energies (ΔG°) were calculated as the difference between the free energies of the product(s) and reactant(s). In the case of the bimolecular 2 nd HAT reactions, the products and reactants were treated separately to properly account for entropy. These are summarized in Table S11 and all relevant energies for these calculations are contained in Table S12. Bond dissociation free energies were calculated as the difference in free energies between the sum of radical products resulting from hemolytic cleavage of the indicated bond and the noncleaved starting material. Multiple combinations of common functionals and solvent models were used to provide a range of bond dissociation free energies (Table S13).

Global and Local Radical Philicities
The global and local radical electrophilicities were calculated according to the method described in the literature. 18 Geometries for each of the radical species were first optimized using the ωB97X-D functional, 6-311++g(d,p) basis set, and the PCM solvation model for PhMe. From these optimized geometries, electronic energies were then determined at N, N+1, and N-1 electron counts without additional geometry optimization, which are summarized in Table S15. The global and local radical electrophilicities, along with the electronic properties from which they are derived, are summarized in