Synthesis of Thiomorpholine via a Telescoped Photochemical Thiol–Ene/Cyclization Sequence in Continuous Flow

A procedure for the continuous flow generation of thiomorpholine in a two-step telescoped format was developed. The key step was the photochemical thiol–ene reaction of cysteamine hydrochloride and vinyl chloride as low-cost starting materials. This reaction could be conducted under highly concentrated (4 M) conditions using a low amount (0.1–0.5 mol %) of 9-fluorenone as the photocatalyst, leading to the corresponding half-mustard intermediate in quantitative yield. Thiomorpholine was subsequently obtained by base-mediated cyclization. The robustness of the process was demonstrated by performing the reaction for 7 h (40 min overall residence time), isolating the desired thiomorpholine via distillation.

Mass flow controller: calibrated for vinyl chloride from Bronkhorst with 0.5 bar inlet and 1 bar outlet pressure, max flow: 150 mLn/min Gas detector: Dräger Pac 8000 portable detector. Calibrated with a bumpstation using 10 ppm C 2 H 4 O/N 2 as calibration gas mixture.
Connectors and tubing: fittings from IDEX Health & Science Technologies were used. High purity PFA (perf luoroalkoxy) tubing with either 0.8 mm ID/1.6 mm OD or 1.6 mm ID/3.2 mm OD was used.
Photoreactor: a Corning Advanced-Flow Lab Photo Reactor was used. The reactor module (G1LF fluidic module) consisted of a compact glass fluidic module (155 × 125 × 8 mm size, 0.4 mm channel depth, 2.77 mL internal volume), encased within a high-capacity heat exchange channel (20 mL volume). LED panels were mounted on both sides of the fluidic module (40 mm from the center of the process stream). Each LED panel was equipped with 20 LEDs of 6 different wavelengths (120 LEDs in total) and a heat exchanger (T = 15 °C). The LED wavelength and intensity was controlled externally using a web-based interface, connected wirelessly to a router. Thermal regulation of the LED panels was carried out using a Huber Minichiller 280 filled with 30% ethylene glycol in water. Thermal regulation of the glass fluidic module was carried out using a Huber Ministat 230 filled with silicon oil (-20 °C to 195 °C). General batch procedure thermal: A 4 mL screw-cap vial was charged with cysteamine (77.2 mg, 1 mmol), vinyl acetate (1 equiv.), AIBN as initiator (5 mol%), and 1 mL of MeOH was added. The vial was closed using high p/T caps and heated in an Al-block to 60 °C.
General batch procedure photochemical: A 5 mL conical microwave vial was charged with cysteamine (77.2 mg, 1 mmol) and vinyl acetate (1 equiv.), and 1 mL of MeOH was added. The solution was sparged with Ar for ~1min to remove oxygen. The vial was closed using a crimp cap and irradiated using a 365 nm, 50 W LED lamp (7 cm distance to center of vial).

Batch Thiol-ene Reaction
The photochemical route not only provided full conversion, but also a very clean reaction profile compared to the thermal reaction ( Figure S6). For the optimization, a temperature and residence time screening was performed. A 5 M solution of cysteamine hydrochloride was prepared by dissolving 14.2 g in MeOH (25 mL). The thermostats were set to the desired temperatures (respective temperature for the reaction, 15 °C LED cooling) and the LEDs were turned on at the desired wavelength. The pumps were set to the desired flow rates. If not otherwise stated in Figure S8, 1.1 equiv of vinyl acetate are used. After allowing the reactor to reach steady state, the reactor output was collected for exactly 1 min into a flask containing biphenyl as external standard and then analyzed by GC-FID. However, clogging issues were encountered during the thiol-ene reaction at such high concentration.

Light Aborbance of Cysteamine Hydrochloride
UV-vis spectroscopy revealed an extremely low absorption coefficient at the irradiation wavelength of the LED ( = 0.025 Lmol −1 cm −1 at 363 nm). This would indicate poorer performance in flow, since at a concentration of 1 M of 2 only about 1% of the incident light is absorbed due to the shorter path length of 0.04 cm (plate´s channel size), compared with about 7% in batch with a path length of 1.3 cm (diameter of the vial). At higher concentrations, also the light absorbance increase s accordingly. Figure S9. Light absorbance as a function of path length utilizing the Beer−Lambert law of 2 at different concentrations. S11 6. Thiol-ene Reaction of Cysteamine Hydrochloride and Vinyl Chloride 6.1 Condensation of VC A 5 mL conical microwave vial equipped with a magnetic stir bar was charged with 2 mL of a solution containing cysteamine hydrochloride (1 M) and internal standard (methyl benzoate) in MeOH (HPLC grade). For the thermal reaction, AIBN (16 mg, 5 mol%, 0.1 mmol) was additionally added. The vial was closed (cap with septum), weighed (tara), immersed in an acetone bath and cooled with liquid N 2 below −50 °C (usually ~−60 °C). Under vigorous stirring of the solution, VC was charged (10 mLn/min) via a cannula that was immersed into the cold solution, whilst a second cannula prevents buildup of pressure (open to fume hood). After charging of the VC was finished (confirmed by weighing the vial), all cannulas were removed and the solution was allowed to warm to rt. Thermal: After condensation of VC as described in 6.1., the reaction mixture was then heated at 80 °C in the microwave reactor for the indicated time. After removal of remaining VC by sparging with Ar, 100 µL of the reaction mixture were mixed with 500 µL of MeOH-d4 and analyzed by 1 H NMR.
Photochemical: After condensation of VC as described in 6.1., the reaction mixture was then irradiated (50 W LED, 365 nm, 7 cm distance to center of vial, see Figure S14) for the desired time. After removal of remaining VC by sparging with Ar, 100 µL of the reaction mixture were mixed with 500 µL of MeOH-d4 and analyzed by 1 H NMR. The liquid feed solution was prepared by dissolving cysteamine -hydrochloride, 9-fluorenone, and methyl benzoate as internal standard in a volumetric flask (25 mL) in MeOH. Where indicated in Table S1, the solution was degassed by sparging with Ar using a balloon and needle. The thermostats were set to the desired temperature beforehand (respective temperature for the reaction, 15 °C LED-cooling). The liquid feed was directly pumped from the volumetric flask using a syringe pump (Syrris-Asia) at maximum flow rate (2.5 mL/min) until the reactor was filled with the substrate solution. Then, the flow rate was reduced to the desired value, the LEDs were turned on (100% intensity) and the MFC was set to deliver the desired amount of VC. After reaching steady-state (about 20 minutes), a sample was collected. 100 µL of this sample were diluted with 500 µL of MeOH-d4 and analyzed by 1 H NMR (300 MHz). S14

Isolation of Intermediate 4
The liquid feed solution was prepared by dissolving cysteamine -hydrochloride (11.36 g, 0.1 mol) and 9fluorenone (0.90 g, 5 mmol, 5 mol%) in a volumetric flask (25 mL) in MeOH. The solution was degassed by sparging with Ar using a balloon and needle. The thermostats were set to the desired temperature beforehand (20 °C reaction, 15 °C LED-cooling). The liquid feed was directly pumped from the volumetric flask using a syringe pump (Syrris-Asia) at maximum flow rate (2.5 mL/min) until the reactor was filled with the substrate solution. Then, the flow rate was was reduced to the desired value (0.139 mL/min), the LEDs were turned on (365 nm, 100% intensity) and the MFC was set to deliver the desired amount of VC (12.1 mLn/min, p = 0.8−0.9 bar). After reaching steady-state (about 20 minutes) the product stream was collected for 12 min. 1.30 g (105%) of product was isolated after evaporation of the solvent as a mixture of 95.6% of 4 and 4.4% of 9-fluorenone according to 1 H NMR.

Isolation of Thiomorpholine
The fractions of the long run were combined and extracted as described in the Experimental Section. The organic phase was then evaporated and distilled under vacuum at 20 mbar. Fractions 1 + 2 were combined and redistilled. An additional 4% of pure thiomorpholine was isolated, giving a total isolated yield of thiomorpholine of 54%. In the residue, peaks that could be assigned to thiomorpholine were detected in the NMR, but were not quantified. In addition, unidentified peaks were present. Possibly, thiomorpholine could either decompose under harsher conditions or further react with other unidentified side products under those conditions. The work-up and distillation procedures were performed only once for the long run experiment, and thus are not optimized. S18 8. Photographs of the Reactor Set-ups Figure S14. Set-up for batch photochemical thiol-ene reactions. A 365 nm, 50 W LED lamp was used, mounted at a distance of 7 cm to the vials. Figure S15. Set-up for the telescoped synthesis. The VC dosing unit is located in the adjacent fume hood. VC is fed to the reactor using PFA tubing (1.6 mm OD, 0.8 mm ID). A) Feed of cysteamine hydrochloride, 9fluorenone and methyl benzoate (IS) in MeOH. B) DIPEA feed. C) Feed of thiol-ene mixture from photoreaction. D) Hold vessel/gas separator after photoreactor. E) Temperature control for reaction plate. F) Fluidic module housing, with tinted plastic panels for light containment. G) Temperature control for LED panels and wireless reciever for LED control. H) Ultrasonic bath with the submerged reaction coil for the cyclization reaction.