Decatungstate‐Mediated C(sp3)–H Heteroarylation via Radical‐Polar Crossover in Batch and Flow

Abstract Photocatalytic hydrogen atom transfer is a very powerful strategy for the regioselective C(sp3)–H functionalization of organic molecules. Herein, we report on the unprecedented combination of decatungstate hydrogen atom transfer photocatalysis with the oxidative radical–polar crossover concept to access the direct net‐oxidative C(sp3)–H heteroarylation. The present methodology demonstrates a high functional group tolerance (40 examples) and is scalable when using continuous‐flow reactor technology. The developed protocol is also amenable to the late‐stage functionalization of biologically relevant molecules such as stanozolol, (−)‐ambroxide, podophyllotoxin, and dideoxyribose.


General information
All reagents and solvents were used as received without further purification. Reagents and solvents were bought from Sigma Aldrich, TCI and Flurochem. Technical solvents were bought form VWR International and Biosolve, and are used as received. The catalyst TBADT was prepared as illustrated below. According such procedure, the catalyst cost can be estimated around 0.7 €/g, employing 20 g tungstic acid (Sigma-Aldrich, BioUltra 72069) and 9.6 g of tetrabutylammonium bromide (Sigma-Aldrich, ReagentPlus, 193119), leading to 49 g of TBADT. For the preliminary experiment, LED strips (365 nm, 2.5 m, 300 SMD5050 LEDs, 36 W) were purchased form LedLightingHut. For scale-up, Vapourtec device was used, equipped with 60 W 365 nm LEDs. Disposable syringes were purchased from Laboratory Glass Specialist. Syringe pumps were purchased from Chemix Inc. model Fusion 200 Touch. Product isolation was performed manually, using silica (P60, SILICYCLE) or automatically by a Biotage® Isolation Four, with Biotage® SNAP KP-Sil 10 or 50 g flash chromatography cartridges. TLC analysis was performed using Silica on aluminum foils TLC plates (F254, SILICYCLE) with visualization under ultraviolet light (254 nm and 365 nm) or appropriate TLC staining (Cerium Ammonium Molybdate). 1 H (400 MHz) and 13 C (100 MHz) spectra were recorded at ambient temperature using a Bruker-Avance 400. 1 H NMR spectra are reported in parts per million (ppm) downfield relative to CDCl3 (7.26 ppm) and all 13 C NMR spectra are reported in ppm relative to CDCl3 (77.16 ppm) unless stated otherwise. The multiplicities of signals are designated by the following abbreviations: s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet), dd (doublet of doublets), dt (doublet of triplets), td (triplet of doublets), ddd (doublet of doublet of doublets). Coupling constants (J) are reported in hertz (Hz). NMR data was processed using the MestReNova 14 software package. Known products were characterized by comparing to the corresponding 1 H NMR and 13 C NMR with those available in the literature. The melting points were measured using a Büchi Melting Point M-565 apparatus. High resolution mass spectra (HRMS) were collected on an AccuTOF LC, JMS-T100LP Mass spectrometer (JEOL, Japen).

Homemade Setup
To assess the scope of the transformation reported, as well as for preliminary experiments, we adopted the homemade setup shown in Figure S1, left. It consists of a 3D-printed (PLA) reactor (inner diameter: 12.5 cm) that has been internally coated with LED strips (365 nm, 2.5 m, 300 SMD5050 LEDs, 36W). Cooling was applied via a strong compressed air flow to keep the temperature below 30 °C. When the reaction was run in batch-mode (see entries 15, 20 -23, 25, 27 in Scheme 2 or for preliminary experiments), the reactor was capped with a 3D-printed (PLA) lid with 8 holes serving as vials holder; in this way, up to 8 reactions could be run simultaneously ( Figure S1, center). When the reaction was run in flow-mode (referred to as "flow setup 1" in the main text) a different lid embedding a 3D-printed cylinder wrapped in PFA tubing was used ( Figure S1, right). The total volume of the reactor was 6.0 mL and the cylinder was covered with reflective tape to increase the efficiency of the setup. Dimensions of the cylinder: length 9 cm, diameter 7.5 cm. Dimensions of the lid: diameter 14.5 cm In this case, the reaction mixture was pumped via a syringe pump and irradiated for the indicated time.

Vapourtec system
For the final optimization experiments and the evaluation of the scope, a Vapourtec device with a UV-150 photochemical reactor was used, equipped with 60 W 365 nm LED. This system is referred to as "flow setup 2" in the main text.

Synthesis of the photocatalyst (TBADT)
In a round bottom flask, tungstic acid sodium dehydrate (20 g, 0.061 mol, 1 eq, 15.5 eur) was dissolved in 5 L of demineralized water. The resultant solution was heated at 90 °C and stirred for 2 h. The solution was cooled to 0 °C and pH was adjusted with conc. hydrochloric acid to 2. In another flask, tetrabutylammonium bromide (9.6 g, 0.030 mol, 0.5 eq, 18.3 eur) was dissolved in 5 L of demineralized water. The resultant solution was heated at 90 °C and stirred for 2 h. The solution was cooled to 0 °C and pH was adjusted with conc. hydrochloric acid to 2. Finally, the solution of tetrabutylammonium bromide was poured into that of tungstic acid sodium dehydrate at room temperature. Total reaction mixture was heated again at 90 °Cand stirred for 2 h. The reaction mixture was cooled to 0 °C and filtered. Obtained solid was diluted in CH2Cl2 (400 mL) and stirred at room temperature for 5 h, resulting in a white turbid mixture that was filtered to afford 16 g of crude TBADT in the residue and 5 g crude from the filtrate. The residue was treated with CH2Cl2 (200 mL) stirred at room temperature for 3 h, resulting again in a white turbid mixture that was filtered to afford 13 g of crude TBADT in the residue and 1.4 g crude from the filtrate. The above was repeated additional four additional 4 times to get 66.7 g of solid, which was treated with (CH3)2CO:CH3CN 1:1 mixture (530 mL) and stirred at room temperature for 3 h and filtered. The filtrate was evaporated to get a white solid, which was triturated with n-pentane (300 mL) to afford 40 g of TBADT as a white solid. Parallelly, the residue was diluted with (CH3)2CO:CH3CN 1:1 mixture (250 mL) and stirred at room temperature for 3 h. The solvent was evaporated and residue triturated with n-pentane (150 mL) to afford 9 g of TBADT as a white solid. Purity of both fractions was evaluated via UV-Vis spectroscopy and matched with that of authentic samples prepared via a previously reported procedure on a smaller scale. [1] Total yield (filtrate+residue): 49 g, 49% yield based on W.

Synthesis of 1m
Product 1m was synthesized by adapting a procedure reported in the literature. [2] A suspension of podofyllotoxin I (0.62 g, 1.5 mmol) in CH2C12 (15 mL) was treated with DMAP (73 mg, 0.6 mmol, 0.4 equiv), imidazole (122 mg, 1.8 mmol, 1.2 equiv) and acetic anhydride (1.5 mL). After 2 h the clear solution was washed once with saturated NaHCO3 solution and twice with water. The volume was reduced over a steam bath, and MeOH added; crystals of 1m formed upon standing (0.502 g, 73%). Spectroscopic data are in accordance with those reported in the literature (solvent: pyridine-d5). [

Synthesis of 1n
Product 1n was synthesized by adapting a procedure reported in the literature. [3][4][5] A mixture of 2′deoxythymidine (10 g, 41.3 mmol) and imidazole (11.8 g, 173 mmol, 4.2 equiv) in anhydrous DMF (60 mL) was stirred at room temperature for 5 min. Then tert-butyldimethylsilyl chloride (TBDMSCl; 13.1 g, 86.9 mmol, 2.1 equiv) was added, and the mixture was stirred for additional 12 h. After adding water (100 mL), the reaction mixture was extracted with hexane, dried with Na2SO4 and concentrated under vacuum to give II (15 g, 77%) as a white solid. [3] II (2.1 g, 4.4 mmol) and (NH4)2SO4 (1.2 g, 9.2 mmol, 2.1 equiv) were added into a oven-dried flask under N2. HMDS (20 mL) was added, and the solution was refluxed for 4 h. The HMDS was evaporated under reduced pressure and the residue was partitioned between water and cyclohexane. The organic layer was washed with saturated NaHCO3 solution and then distilled water. It was dried over Na2SO4 and evaporated under reduced pressure to give a yellow oil that was purified by flash chromatography (neutral Al2O3; cyclohexane:diethyl ether 2:1) (0.82 g, 54%). [4] Product III (0.81 g, 2.35 mmol) and 10% Pd/C (15 wt%, 184 mg) in i-PrOH (10 mL) was vigorously stirred at room temperature under ambient pressure of H2 for 24 h (balloon). Afterwards, the reaction mixture was filtered through a celite pad, and the filtrate was evaporated under reduced pressure. The residue was purified by column chromatography on silica gel (cyclohexane:Ethyl acetate 7:1) to give 1n as a light yellow oil (0.74 g, 91%). Spectroscopic data are in accordance with the literature. [

Synthesis of 2v
Product 2v was synthesized according to a procedure reported in the literature. [6] To a 100 mL N2-flushed round-bottomed flask equipped with a magnetic stir bar adenine (1.35 g, 10.0 mmol) and DMAP (0.122 g, 1.0 mmol) were added 50 mL of dry THF via gas-tight syringe. To the stirring suspension 8.7 g (39.8 mmol) of Boc2O were added under N2 atmosphere. The reaction mixture was stirred for 12 h at room temperature, after which solvent was removed by rotary evaporation to give I as a yellow oil. I (10 mmol) was taken in ethyl acetate (400 mL) and washed once with HCl 1 N (30 mL) followed by brine (3 × 100 mL). The ethyl acetate layer was dried over Na2SO4 and evaporated to give a colorless oil, which was used without any further purification. This oil was dissolved in MeOH (100 mL), to which 45 mL of saturated NaHCO3 solution was added. The so-obtained turbid mixture was stirred at 50 o C for 1 h to get product 2v. The reaction can be conveniently monitored via TLC (SiO2, cyclohexane:ethyl acetate 7:3). After evaporation of MeOH, water (100 mL) was added to the crude and the aqueous layer was extracted with CHCl3 (2 × 300 mL). The organic layers were gathered and dried over Na2SO4, filtered, and evaporated to give a white solid. The crude material was dissolved in ethyl acetate and filtered through silica gel and washed with ethyl acetate to give 2.1 g (63% yield) of pure 2v. Spectroscopic data are in accordance with the literature. [6]

Optimization of reaction conditions
The optimization of reaction conditions in batch was carried out by studying the cross-coupling between tetrahydrofuran (1a) and parent pyrazole (2a) in CH3CN (1 mL) on a 0.2 mmol scale (Tables S1). Reaction conditions: 2a (0.2 M), 1a (6 equiv, 1.2 M), TBADT (n mol%) in 1 mL of CH3CN were mixed in a 7 mL tube and irradiated with UV-A LEDs ( = 365 nm, 36 W) for the required time (see Figure S1). After irradiation, CD3CN (200 μl) was added to the crude along with pyrazine as external standard and the mixture was analyzed via 1 H-NMR to evaluate consumption and yield. The optimization of reaction conditions in flow was carried out by studying the cross-coupling between tetrahydrofuran (1a) and parent pyrazole (2a) in CH3CN (1 mL) on a 0.2 mmol scale (Tables S2). Reaction conditions: 2a (0.2 M), 1a (6 equiv, 1.2 M), TBADT (n mol%) in 1 mL of CH3CN were mixed in a 7 mL tube, the solution was withdrawn with a syringe and infused in flow setups described in Section S1 (see Figure S1 and S2). After irradiation, CD3CN (200 μl) was added to the crude along with pyrazine as external standard and the mixture was analyzed via 1 H-NMR to evaluate consumption and yield.

Radical Trapping Experiments
A 7 mL flame-dried tube was charged with TBADT (5 mol%), 2a (1 mmol), TEMPO (10 equiv), CH3CN (5 mL), TBHP (3 equiv), 1,3-dihydroisobenzofuran 1h (18 equiv). The mixture was swirled until homogenous, placed in a 10 mL disposable syringe and mounted on a syringe pump. The flow rate was set to 0.167 mL min -1 to a residence time of 1 h. When the syringe was fully empty, again acetonitrile was loaded into a syringe and injected to collect all product at the end of the reactor in a flask. Pyrazine (0.5 mmol) were added to the mixture and the yield was calculated by 1  To a 7 mL flame-dried tube was added TBADT (5 mol%), 2a (1 mmol), BHT (10 equiv), CH3CN (5 mL), TBHP (3 equiv), 1,3-dihydroisobenzofuran 1h (18 equiv). The mixture was swirled until homogenous, placed in a 10 mL disposable syringe and mounted on a syringe pump. The flow rate was set to 0.167 mL min -1 to a residence time of 1 h. When the syringe was fully empty, again acetonitrile was loaded into a syringe and injected to collect all product at the end of the reactor in a flask. Pyrazine (0.5 mmol) were added to the mixture and calculated the yield by 1 H-NMR. HRMS (FI) m/z calcd for C23H30O2: 338.2246; found: 338.2250.

Oxocarbenium Ion Trapping Experiments
To a 7 mL flame-dried tube was added TBADT (5 mol%), 1i (1 mmol), CH3CN (5 mL), TBHP (3 equiv), methanol (18 equiv). The mixture was swirled until homogenous, placed in a 10 mL disposable syringe and mounted on a syringe pump. The flow rate was set to 0.167 mL min -1 to a residence time of 1 h. When the syringe was fully empty, again acetonitrile was loaded into a syringe and injected to collect all product at the end of the reactor in a flask. Pyrazine (0.5 mmol) were added to the mixture and calculated the yield by 1  To a 7 mL flame-dried tube was added TBADT (5 mol%), 1i (1 mmol), CH3CN (5 mL), TBHP (3 equiv), tert-butanol (18 equiv). The mixture was swirled until homogenous, placed in a 10 mL disposable syringe and mounted on a syringe pump. The flow rate was set to 0.167 mL min -1 to a residence time of 1 h. When the syringe was fully empty, again acetonitrile was loaded into a syringe and injected to collect all product at the end of the reactor in a flask. The mixture was dried under reduced pressure and purified by column chromatography on silica gel to provide the product. The final product was weighed and characterized by HRMS, 1

Intermolecular Kinetic Isotopic Effect (KIE)
To a 7 mL flame-dried tube was added TBADT (5 mol%), 2a (1 mmol), CH3CN (5 mL), TBHP (3 equiv), 1a (9 equiv), 1a-d8 (9 equiv). The mixture was swirled until homogenous, placed in a 10 mL disposable syringe and mounted on a syringe pump. The flow rate was set to 0.167 mL min -1 to a residence time of 1 h. When the syringe was fully empty, again acetonitrile was loaded into a syringe and injected to collect all product at the end of the reactor in a flask. The mixture was dried under reduced pressure and purified by column chromatography on silica gel to provide the product. The final product was weighed and characterized by HRMS, 1

Laser Flash Photolysis
Laser flash photolysis (LFP) experiments were performed to study the decay of the reactive excited state of decatungstate (tagged W*) in the presence of increasing concentrations of quenchers 1a and 2a, which can be monitored at 780 nm. Thus, we measured the quenching constants for 1a and 2a through relevant Stern-Volmer plots and we found two comparable reaction rates. In detail, a bimolecular rate constant kQ = 2.3×10 8 M −1 ·s −1 was measured for 1a, while kQ = 1.6×10 8 M −1 ·s −1 was determined for 2a.

Experimental.
Nanosecond transient absorptions were recorded with an in-house assembled setup. An excitation wavelength of 324 nm was used. The excitation wavelength of 324 nm was generated using a tunable Nd:YAG-laser system (NT342B, Ekspla) comprising the pump laser (NL300) with harmonics generators (SHG, THG) producing 355 nm to pump an optical parametric oscillator (OPO) with SHG connected in a single device. The laser system was operated at a repetition rate of 5 Hz with a pulse length of 5ns. The probe light running at 10 Hz was generated by a high-stability short arc xenon flash lamp (FX-1160, Excelitas Technologies) using a modified PS302 controller (EG&G). Using a 50/50 beam splitter, the probe light was split equally into a signal beam and a reference beam and focused on the entrance slit of a spectrograph (SpectraPro-150, Princeton Instruments) with a grating of 150ln/mm blaze at 500nm. The probe beam (A = 1 mm2) was passed through the sample cell and orthogonally overlapped with the excitation beam on a 1 mm × 1 cm area. The excitation energy was recorded by measuring the excitation power at the back of an empty sample holder. In order to correct for fluctuations in the flash lamp spectral intensity, the reference was used to normalize the signal. Both beams were recorded simultaneously using a gated intensified CCD camera (PI-MAX3, Princeton Instruments) which has an adjustable gate of minimal 2.9 ns. Two delay generators (DG535 and DG645, Stanford Research Systems, Inc.) were used to time the excitation pulse, and to change the delay of the flash lamp and gate of the camera during the experiment. The setup was controlled by an in-house written Labview program. From decays reported in Figure S3, it is possible to derive the Stern-Volmer plots shown in Figure S4 following the equation:

Quantum yield measurement
The Quantum yield (QY) was determined via ferrioxalate actinometry according to a procedure reported in the literature. [7] Synthesis of the actinometer. Potassium ferrioxalate was prepared according to a procedure reported in the literature. [8] In particular, 3.2 g of FeCl3 was dissolved in 8 mL of distilled H2O and was added to a hot solution of 12 g of K2C2O4 in 20 mL of distilled H2O. After a couple of minutes at 100 °C, the mixture was cooled down at room temperature and crystallization was triggered with a glass stick. After crystallization was complete, the mother liquor was removed via a Pasteur pipette and the green crystals (K3[Fe(C2O4)3]) were dissolved again in 20 mL of distilled H2O. Potassium ferrioxalate was recrystallized two more times, washed with MeOH and dried at 45°C for 1 hour, light green crystals (5 g, 55%) were obtained. Then, a single-beam spectrophotometer was used to read the difference of absorbance (A) between Test 1 and 2 and the Blank.
for Test 1: ∆A = 0.15796 for Test 2: ∆A = 0.14166 We calculated the quantum yield after 2.5 hours of reaction, when 4.1×10 -5 moles (20% yield) of product 3 were formed ( Figure S5). QY was calculated from the ratio between the moles of product produced after 2.5 hours and the moles of photon reaching the reaction vial in 2.5 hours.  QY was determined to be 18% (as average of two experiments), indicating that the process is most likely not a radical-chain reaction.

General procedure
General Procedure 1 (GP1), batch conditions: A stock solution containing TBADT (5 mol%, 166 mg), azole (1 mmol), dry acetonitrile (5 mL), TBHP (545 µL, 5.5 M in decane or nonane, 3 equiv) and ether was prepared in a vial equipped with cap with septum and stirring bar (see each case for details). The solution was then split into 5 vials (1 mL each) and individual solutions were sparged with nitrogen. The reaction was stirred and irradiated with 36 W UV-A LEDs for 16 h. The solutions were collected, solvent was removed under reduced pressure and the crude was purified via column chromatography on silica gel to provide the product.