On Solute Recovery and Productivity in Chiral Resolution through Solid-State Deracemization by Temperature Cycling

Temperature cycling represents an effective means for the deracemization of chiral compounds that crystallize as conglomerates and racemize in solution. In such a process, a suspension enriched in the desired enantiomer is converted into an enantiopure one through periodic cycles of crystal dissolution and crystal growth. We show that performing temperature cycling at higher temperatures leads to faster deracemization and, consequently, higher productivity. However, this comes at the cost of lower recovery, as the solution contains potentially relevant amounts of solute due to the higher solubility at an elevated temperature. In this work, we introduce and compare two process variants that mitigate this issue. The first involves temperature cycling, followed by linear cooling, whereas the second is based on merging the temperature cycles and cooling crystallization. Experiments carried out with the chiral compound N-(2-methylbenzylidene)-phenylglycine amide show that the former variant is faster than the latter, and it is easier to design and implement. In this process, the choice of an appropriate cooling rate is essential to avoid nucleation of the undesired enantiomer.


S Supporting Information S.1 HPLC Protocol for Sample Analysis
To monitor the evolution of the enantiomeric excess during the process, a sample containing 60-100 µL of suspension was vacuum filtered and washed with anti-solvent at the end of each temperature cycle.In addition to these samples, a zero-th sample was taken from the suspension before the temperature cycles were started to determine the initial enantiomeric excess (ee 0 ).The samples were transferred to crimp top clear glass 1.5 mL HPLC vials and dissolved in acetonitrile for measurements.All measurements were carried out in a HPLC apparatus (DIONEX UltiMate 3000 series) that was equipped with a quaternary pump and DAD detector (Thermo Scientific, Reinach, Switzerland).Measurements were carried out at UV-VIS 213 nm on a CHIRALPAKL AY 20 µm stationary phase packed in a 250 mm × 4.6 mm column to separate enantiomers.From each HPLC vial, 2 µL was sampled and injected to the column kept at 27 • C. Pure acetonitrile was used as the mobile phase with a flow rate of 2 mL min −1 .Each measurement lasted 6 minutes and the retention times were found to be 2.07 min for (L)-NMPA and 3.26 min for (D)-NMPA as shown in Figure S1.The enantiomeric excess was calculated using the ratio of the HPLC peak areas for both enantiomers.

S.2 Stability of NMPA
Since deracemization experiments for NMPA have not been reported in the literature at temperatures of 50 • and above, we have analyzed products of some of the experiments by 1 H-NMR to verify the thermal stability of the compound.The ensuing spectra were compared with values reported in the supplementary information of the publication by Iggland et al. 1 , whereby we observed a quantitative agreement.This indicates that no degradation of NMPA took place under the given experimental conditions.This is further confirmed by the fact that the peaks measured using HPLC in all experiments appeared at the same retention times and exhibited regular shapes.

S.3 Additional Experiment to Improve Enantiomeric Excess
The enantiomeric excess in temperature cycling experiments reported both in this work, as well as in the literature, [2][3][4] in most cases does not reach a final value of ee = 1, but approaches a plateau at a slightly lower value on the order of 0.97-0.98.To understand the underlying reason behind this observation, we ran an experiment during which we decreased the cycle amplitude over time.
We conjectured that nucleation of the undesired enantiomer during the cooling step of a temperature cycle may affect the attainable enantiomeric excess.By decreasing the cycle amplitude, the difference in supersaturation between T max and T min decreases and thus, nucleation is expected to become less relevant.The results reported in Figure S2 show that decreasing the difference between the high and low temperature of the cycles does not lead to complete enantiopurity, which indicates that no relevant nucleation of the undesired enantiomer takes place, in contrast with our hypothesis.The reason behind this observation may thus instead be related to not yet understood shortcomings in the measurement methodology.

Figure S1 :
Figure S1: Example of a chromatogram of a sample analysed by above-mentioned HPLC protocol.

Figure S2 :
Figure S2: Experiment with decreasing cycle amplitude aimed at improving the final enantiomeric excess.