Chiral auxiliary recycling in continuous flow: automated recovery and reuse of Oppolzer's sultam† †Electronic supplementary information (ESI) available: Experimental details, details of flow reactor and equipment, characterization data of compounds. See DOI: 10.1039/c7sc05192a

The telescoping of a three-stage, chiral auxiliary-mediated transformation in flow is described, including continuous separation of the product and auxiliary, enabling automated auxiliary reuse.


General experimental details
Unless otherwise noted reagents were used as received. Palladium on carbon was purchased from Strem, (10% palladium on activated carbon, reduced, dry powder, product 46-1900). All other chemicals were obtained from either Sigma Aldrich or Combi-Blocks. Aliquat 336 was absorbed onto silica gel, dry loaded onto a 40×100 mm silica column, eluted with DCM→95:5 DCM:MeOH, and evaporated prior to

Flow reactor setup:
A schematic of the reactor setup is shown in Figure S1 and a photo of the setup in the fumehood is shown in Figure S2 Figure S1. The series of mixers was constructed by making active mixer units as reported by Ley and co-workers. 2 Three 1.00 mL plastic HSW syringes with the plungers removed were heated at the open end until soft and then a thread was cut using 1/4-28 PEEK male nuts. Four 3×10 mm PTFE coated stir bars were inserted in the mixer and PTFE tape was used to improve the seal between the PEEK fitting and the cut thread. The mixers were connected together with 10 cm of PFA tubing (1/16" O.D., 0.5 mm I.D., 20 µL). 100 cm PFA tubing (1/16" O.D., 0.5 mm I.D. 100 μL volume) was connected after last chamber to allow the emulsion to settle into plugs before entering the gravity liquid-liquid separator. The agitators were operated by a magnetic stir plate set at maximum (1400 rpm). A ~7.5 mm air gap between the stir plate surface and the mixers was maintained through use of cardboard spacers to prevent heat transfer from the stir plate surface that became warm to the touch with extended hours of operation.

Gravity liquid-liquid separator with active withdrawal of both organic and aqueous phases
(type 1) S1 and S3 in Figure S1. A 2.5 mL glass Hamilton syringe was used for the body of the separator. The

Packed bed reactor (PBR)
R2 in Figure  The packed bed reactor was interfaced with the commercially available H-Cube Mini from ThalesNano to provide in situ generated hydrogen. The decision to use an in-house fabricated PBR rather than purchasing catcarts was solely financially driven. Our PBR was re-packed after each experiment with a mixture of commercially available Pd/C (110 mg) and 150-212 μm glass beads (3.0 g, product G9018 from Sigma Aldrich) to limit pressure drop. 3 Liquid volume of the packed bed was determined to be 1.0 mL by subtracting the mass once filled with water from the dry mass of the freshly backed PBR. Estimating equal occupancies of each of the 3 fluid phases (organic, aqueous and gaseous) gives a very rough estimate of residence time for the organic phase of 1.3 min. Measuring the time until breakthrough when introducing substrate through the clean packed bed was in rough agreement with this estimate.  Figure S1. A tube-in-tee mixer 4 was used for the acid wash after the methanolysis. The NaOMe in methanol formed small plugs in the tolueune stream, causing inefficient quenching when mixing with HCl in a simple tee. The tube-in-tee was fabricated by taking an ordinary PFA tee mixer with 1.0 mm I.D.
through-holes and enlarging the straight bore to 1/16" all the way through with a 1/16" drill bit, then further boring out one half of the straight bore to 2.0 mm I.D. as show in Figure S8. by addition of EtOH (100 mL), water (200 mL) and NaOH (65 g, 10 eq.) 8 and stirring the resulting biphasic mixture rapidly at room temperature for 5 h until a single phase was achieved. The solution was then washed with 50 mL of 1:1 hexanes:EtOAc followed by 25 mL of hexanes. The pH was adjusted to <1 with conc. HCl resulting in the formation of a separate organic phase that was collected. The aq. phase was extracted with 50 mL 1:1 hexanes:EtOAc then 25 mL hexanes and all organic phases were combined and dried over Na 2 SO 4 . The solvent was evaporated to yield an oil. EtOAc (100 mL) was added followed by cyclohexylamine (15.3 g, 154 mmol) with rapid stirring, resulting in a thick white slurry of 3-methyl-5-phenyl-2-pentenoic acid·cyclohexylamine salt that was recrystallized twice from EtOAc 9 to remove the z isomer. The resulting white solid was then added to 3 M HCl (100 mL) and the free acid was extracted with 2×25 mL DCM. The combined extracts were dried over Na 2 SO 4 and the solvent evaporated yielding a colourless oil. Boiling hexanes (80 mL) were added resulting in a fine suspension that was hot filtered and cooled to -20 °C yielding colourless crystals of the desired product in 97% purity. Recrystallizing a second time from hexanes yielded white needles of pure (E)-3-methyl-5-phenylpent-2-enoic acid.  Figure S3). Yields were ~quantitative as monitored by off-line GC analysis. 4% NaCl could be added to the NaOH feed to improve downstream separations without influencing the yield.

Hydrogenation in PBR from Equation 1
Start-up as per the description in section 5.1 general procedures for process start-up but with only the acylation and hydrogenation stages telescoped together. 97:3 (E):(Z)-3-methylnon-2-enoic acid chloride was used as the substrate. The effluent was collected for 1 h 50 min then evaporated and the residue chromatographed (silica gel, 3:1 hexanes:Et 2 O) to yield 790 mg; 97% hydrogenated product, d.r.: 90:10. Table 2 Various phase transfer catalysts were screened in batch. The combination of 18-crown-6 and 2,5dimethyl-2,5-hexanediol 13 with 55% (w/w) KOH (aq) and 0.1 M substrate in toluene was effective but when transitioning into flow precipitate formation and slow reaction rates were problematic.

Methanolysis optimization from
Delivering instead KOH in MeOH (30% w/w, 18 µL/min, 5 eq.) and extraction: 10 min, plus 10 min for each in-line phase separator ≈ 2 h total for start-up. Each gravity liquid/liquid phase separator was maintained with ~1 mL of organic phase and ~0.25 mL of aqueous phase. The flow rate of pumps internal to the flow path (i.e., withdrawing from phase separators) were adjusted to maintain a constant volume in the separator (i.e., flow in = flow out for each liquid phase).
Diastereoselectivity of the hydrogenation reaction was monitored off-line by withdrawing a ~25 μL aliquot from the post-hydrogenation organic phase and submitting to GC analysis every 30 min (see Section 9).

Telescoped process with auxiliary recovery
The acid chloride, auxiliary and PTC were all combined in the organic solution and 4% (w/w) NaOH, 4% (w/w) NaCl was used as the aqueous solution for the PTC acylation. No background sultam amide formation in the stock solution was observed over the timeframe of the experiment (~6 h).
After steady state was reached for the entire process, collection of the product and recovered auxiliary streams commenced. Effluents were collected for 3 h, then all pumps were stopped. The product phase effluent was evaporated and the methyl ester was purified by chromatography (20×150 mm silica gel stationary phase, pentane→5% Et 2 O in pentane mobile phase).
The recovered auxiliary phase effluent was acidified to pH<1 with conc. HCl and then extracted with 2×15 mL DCM. The combined organic extracts were dried over Na 2 SO 4 , and evaporated to yield the crude auxiliary. Recrystallization from hexanes gave white needles of pure camphorsultam.

Telescoped process with auxiliary recycling
For auxiliary recycle experiments the acid chloride, 35% auxiliary make up and PTC were combined in the organic solution and the other 65% of auxiliary was dissolved in a solution of 4% NaOH, 4% NaCl used for the acylation during start-up.
The auxiliary in 4% NaOH, 4% NaCl solution used for start-up was prepared by taking camphorsultam (0.46 g, 2.2 mmol) and making up to 25.00 mL with 4% NaOH, 4% NaCl.
After steady state was reached for the entire process as in the single pass experiments, collection of the product stream was commenced and the recovered auxiliary stream was connected to the acylation phase, completing the recycle loop. Product containing effluent was collected for 4.5 h, then all pumps were stopped. The effluent was evaporated and the methyl ester purified by chromatography (20×150 mm silica gel stationary phase, pentane→5% Et 2 O in pentane mobile phase).