Olefin Metathesis in Continuous Flow Reactor Employing Polar Ruthenium Catalyst and Soluble Metal Scavenger for Instant Purification of Products of Pharmaceutical Interest

In recent years, the development of continuous-flow reactors has attracted growing attention from the synthetic community. Moreover, findings in the precise control of the reaction parameters and improved mass/heat transfer have made the flow setup an attractive alternative to batch reactors, both in academia and industry, enabling safe and easy scaling-up of synthetic processes. Even though a majority of the pharmaceutical industry currently rely on batch reactors or semibatch reactors, many are integrating flow technology because of easier maintenance and lower risks. Herein, we demonstrate an operationally simple flow setup for homogeneous ring-closing metathesis, which is applicable to the synthesis of active pharmaceutical ingredients precursors or analogues with high efficiency, low residence time, and in a green solvent. Furthermore, through the addition of a soluble metal scavenger in the subsequent step within the flow system, the level of ruthenium contamination in the final product can be greatly reduced (to less than 5 ppm). To ensure that this method is applicable for industrial usage, an upscale process including a 24 h continuous-flow reaction for more than 60 g of a Sildenafil analogue was achieved in a continuous-flow fashion by adjusting the tubing size and flow rate accordingly.


General Procedure for RCM (Ring Closing Metathesis) Reactions in Batch
Scheme S1. General scheme for RCM reactions in batch.
Reactions were performed in dry solvents (DCM or EtOAc) under argon atmosphere. Stock solution of catalyst was prepared by dissolving 0.02 mmol of the appropriate catalyst (Ru1 or Ru2) in 1 mL of DCM or EtOAc (CM = 0.02 M). 0.1 M substrate solution (3 mL, 0.3 mmol, 1 equiv.) was placed in an oven-dried vial followed by the addition of 0.02 M solution of the catalyst (15/150 μL, 0.0003/0.0030 mmol, 0.1/1.0 mol%). The reaction mixture was stirred at the given temperature (30/60 °C) and then, the solution of SnatchCat (0.045 M in DCM) was added (0.44/4.4 mol% respectively for 0.1/1.0 mol% of the ruthenium catalyst). The resulting mixture was stirred for 30-120 min (200 rpm). The product was purified using column chromatography (stationary phase: SiO2, eluent: reaction solvent). Yields for all substrates but 15 were calculated based on GC results using durene (1,2,4,5-tetramethylbenzene) as an internal standard or NMR experiments (for substrate 15).
Note: Compound 16 precipitated from the reaction mixture as colorless crystals when the reaction is carried out in EtOAc. Table S1. Results of RCM reaction performed in batch.

Notes:
• after basic extraction the product does not contain TFA anion -it appears as free amine form (based on 19 F NMR) • E/Z ratio was found by calculations based on NMR signals. Two appropriate signals (one from E-isomer and second from Z-isomer) were found and integrated (7.81 ppm -Eform, 7.76 ppm -Z-form). The integral sum was set at the value of 100.
Scheme S3. General flow set-up diagram for testing substrates in small scale with varying concentration of the substrates depending on the solubility.
Substrate (2.4 mmol, 1 equiv.) and the catalyst Ru2 (6.1 mg, 0.0072 mmol, 0.3 mol%) were weighed separately, put into two different 20 mL vials (preheated in the oven for at least 30 min), and covered with rubber septum caps tightly. Atmosphere exchange was done employing argon (as an inert gas) using vacuum pumps (three cycles were performed before attaching balloons filled with argon to each of the vials). Then, 8 mL of dried and argon-bubbled EtOAc was transferred into the vials via a disposable syringe and sonication was done to ensure that all the catalyst and substrate were fully dissolved. The final concentration before mixing in the T-connector for the substrate was 0.4 M. Concentration of the substrates were varied based on the solubility in EtOAc.
Using an 8 mL KDS stainless syringe attached with a needle connector ( Figure S1), the substrate solution was extracted from the 20 mL vial and connected to a KDS syringe pump. Another 8 mL KDS stainless syringe was filled with the Ru2 solution mixture and attached to the same KDS syringe pump. Meanwhile, SnatchCat scavenger (6.1 mg, 0.030 mmol, 1.32 mol%, 4.4 equiv. vs the catalyst) was dissolved in 5 mL of EtOAc and transferred into an 8 mL KDS stainless syringe. Likewise, the syringe was attached to a KDS syringe pump. The flow set-up is as shown in Scheme S3 with the tubing soaked in an oil bath ( Figure S4) to provide heating for the reaction. After 2 residence timing (tR) of 20 minutes, the product was collected at the end of the silica packed-bed and concentrated under vacuum before further analysis using GC, NMR or ICP-MS were performed. Figure S4. General flow set-up for testing substrates in small scale. The pump on the left was used for pumping substrates and catalyst, while the pump on the right was used to pump SnatchCat for the workup. At the end of the reaction, the residues were filtered through the silica gel packed bed and collected in the 20 mL glass vial.

Procedure for RCM Reaction of Pacritinib Precursor in Flow
The precursor of Pacritinib 17 (30 mg, 0.064 mmol, 1 equiv.) and catalyst Ru2 (4.3 mg, 0.005 mmol, 8.0 mol%) were weighed separately, put into two different 20 mL vials (preheated in the oven for at least 30 min), and covered with rubber septum caps tightly. Atmosphere exchange was done employing argon (as inert gas) using vacuum pumps (three cycles were performed before attaching balloons filled with argon to each of the vials). Then, 8 mL of dried and argonbubbled EtOAc was transferred into the vials via a syringe and sonication was done to ensure that all the catalyst and substrate were fully dissolved. Afterwards, 10 µL of trifluoroacetic acid (15 mg, 0.135 mmol, 2.1 equiv.) was added to the Pacritinib precursor solution in order to acidify the free amine groups in the starting material via a micro-syringe.
Using an 8 mL KDS stainless syringe attached with a needle connector ( Figure S1), the acidified Pacritinib solution mixture was extracted from the 20 mL vial and connected to a KDS syringe pump. Another 8 mL KDS stainless syringe was filled with the Ru2 solution mixture and attached to the same KDS syringe pump. Meanwhile, solid SnatchCat (5.0 mg, 0.023 mmol, 35.2 mol%, 4.4 equiv. vs the catalyst) was dissolved in 8 mL of untreated EtOAc and extracted into an 8 mL KDS stainless syringe. Likewise, the syringe was attached to a KDS syringe pump. The flow setup was as shown in Scheme S3 with the tubing soaked in an oil bath at 90 °C to provide heating for the reaction. After 2 residence timing (tR) of 40 minutes, the product was collected at the end of the silica packed bed and concentrated under vacuum. Concentrated sodium carbonate S10 solution was used to neutralize the product via liquid-liquid extraction. The organic layer was concentrated and purified using flash column chromatography (stationary phase: SiO2, eluent: EtOAc/n-hexane 20:80 v/v) to obtain the desired RCM product 18 (72%, E/Z = 66:34). Further analysis was performed using NMR and ICP-OES (<15 ppm of Ru).

Procedure for RCM Reaction of Sildenafil Derivative in Flow in Larger Scale
Scheme S4. Flow set-up diagram for the synthesis of Sildenafil derivative in a 24-hour flow system.
Sildenafil derivative 15 (72.84 g, 150 mmol, 1 equiv.) and catalyst Ru2 (759.6 mg, 0.9 mmol, 0.3 mol%) were weighed separately, put into two different 2.5 L blue capped bottles (preheated in the oven for at least 30 min), and tightened with the blue cap shown in Figure S1. The starting materials were purged with argon gas for 10 minutes to remove any residual air in the bottle. Subsequently, 1.5 L of dried and argon-bubbled EtOAc was transferred into the bottles and sonicated until all the catalyst and substrate were fully dissolved into the solvent.
The two large bottles were connected to an Asia pump which was connected to the reaction tubing. Meanwhile, SnatchCat scavenger (871.2 mg, 3.96 mmol, 1.32 mol%, 4.4 equiv. vs the catalyst) was dissolved in 1.5 L of EtOAc and connected to another Asia pump (shown in Scheme   S4 and Figure S5). The flow set-up was illustrated in Scheme S4 with the tubing soaked in an oil bath at 90 °C to provide heating for the reaction. After 2 residence timing (tR) of 20 minutes, the product was collected at the end. The post-reaction mixture was being collected for 24 hours and then purified using flash column chromatography (stationary phase: 60 g of SiO2, eluent: EtOAc). The product solution was concentrated under vacuum to furnish product 16 (62 g, 94%). The product was analyzed using NMR and ICP-MS (0.51 ppm of Ru).
Measurements were carried out at National University of Singapore. Sample preparation was done by measuring 5 mg of the starting material (substrates) into GC vials. A standard stock solution of 56.0 mg of 1,3,5-trimethoxybenzene was measured and added with 10 mL of EtOAc as solvent. For each substrate, 1.5 mL of the stock solution was added to it and measured with a GC using temperature programming. The K value of the ratio between the internal standard and substrate was obtained. In addition, the retention time of the substrate was obtained. Subsequently, 5 mg of product sample was measured and added with 1.5 mL of standard stock solution. The amount of the starting material was calculated with reference to the internal standard and K value. Then, the mass of both starting material and product was calculated before converting to moles for the conversion.

ICP-OES Measurement Details
Approximately 100 mg of the product was extracted after reaction in flow system, introduced into PTFE vessels and topped up with 3 mL of Supra Pure nitric acid (65%). The vessels were capped and placed in a microwave digester to decompose the organic substances. The microwave was heated up to 180 °C at 500 W for 15-30 min. The temperature was then fixed at 180 °C and 500 W for 10 min. The digested substances were then left to cool to room temperature before transferring to a 25 mL or 50 mL volumetric flasks and topping up with ultra-pure water. Standard calibration curve was plotted with commercially available standards which are diluted to 0, 1, 10, 100, 250, 500 and 1000 ppb. Ruthenium concentrations of the samples were then determined by ICP-OES from this standard calibration curve. The concentration obtained, in parts per billion (ppb), from ICP-OES was multiplied by the dilution factor of 25 and divided by the mass (measured beforehand). Example results: 9.094 ppb signal in the ICP-OES concluded that in 101.66 mg sample, there was 2.2 parts per million (ppm) of ruthenium present.