Faster and cleaner dynamic kinetic resolution via mechanochemistry

Abstract

Application of the ball-milling techniques to dynamic kinetic resolution accelerates reactions while avoiding the use of toxic organic solvents and reactants commonly required in these processes. In this way, dynamic kinetic resolutions can be both faster and “cleaner” in the sense that mechanochemistry enables the reduction of their environmental impact.

---

Controlling the optical purity of molecules is essential in many areas, particularly in pharmaceutical research where biological activity highly depends on the chirality of active compounds. Numerous highly stereoselective methodologies have been developed to access enantiopure molecules and among them kinetic resolution provides an attractive approach. In this process, two enantiomers of a racemate are transformed into chiral products at different rates. When the resolution is efficient, one of the enantiomers of the racemic mixture is transformed into the desired product while the other is recovered unchanged. However, this procedure is limited to the maximum theoretical yield of 50%. To overcome this limitation, an in situ epimerization of the chirally labile substrate can be combined with the kinetic resolution to become a dynamic kinetic resolution (DKR).¹ However, the inherent necessity for the product formation to be slower than the substrate epimerization usually makes dynamic kinetic resolution quite a slow process.

α-Halo carbonyls (A, Scheme 1) can be taken as a good illustration of this process: the configurationally labile halogen atom in the α position of the carbonyl can be irreversibly substituted by a nucleophile. Stereo-differentiation in the S_N2 halogen displacement can be controlled by the chiral environment on R₂.

Scheme 1 Dynamic kinetic resolution of α-halo carbonyls by S_N2 halogen displacement.

Whereas DKR of such substrates allows access to products of type B with good yield and diastereoselectivity, long reaction times are needed. As many other examples of S_N2 reactions, this DKR requires polar and aprotic solvents such as toxic DCM or THF,² and is performed with organic, toxic and corrosive bases such as Et₃N to epimerise the substrate.³ Thus, reaction conditions allowing shorter reaction times and avoiding the use of problematic solvents and reagents would be highly preferable. Following our general objective to find greener alternatives to the otherwise environmentally troublesome chemical reactions,⁴ we envisioned performing DKR avoiding the use of undesirable solvents and bases. Due to its relevance for a laboratory scale study, the Ecoscale score⁵ was chosen as the green metric of choice to evaluate the environmental impact of the new reaction conditions. The Ecoscale is a score ranging from 0 (totally failed reaction) to 100 (ideal reaction) that is based on yield, cost, safety, technical set up, the temperature and time of reaction, work up and purification aspects. To each of these parameters are attributed penalty points that are subtracted from the ideal score of 100 to give the Ecoscale score of the studied reaction. The reaction conditions are ranked excellent if the Ecoscale score is >75, acceptable if >50 and inadequate if <50.

As the first example, we studied the DKR reaction of α-bromo-(R)-pantolactone ester 1 with Bn₂NH that was previously described by Durst and coworkers.⁶ Thus, we treated compound 1 with Et₃N, a catalytic amount of TBAI and Bn₂NH to obtain the corresponding α-dibenzylamino ester 2a in 59% yield with an excellent diastereoisomeric ratio (>98 : 2) (Table 1, entry 1). As described in the literature, this reaction was set up using an excess of Et₃N (2.0 eq.) in a toxic solvent (THF is suspected to be carcinogenic), accounting for a low Ecoscale score of 42.5, which corresponds to an inadequate synthesis.⁵ To reduce the environmental impact of this reaction, we envisioned replacing the problematic Et₃N and THF with innocuous NaHCO₃ and water.⁷ Under these conditions, the Ecoscale score was barely improved to 54, mainly due to a low yield of 34% (50% brsm; Table 1, entry 2). In addition, the reaction was much slower as 6 h were necessary to reach 50% conversion and 2a was obtained with a lower diastereomeric ratio (87 : 13). When a preferable solvent such as EtOAc⁸ was used in place of water, the reaction proceeded with a satisfying Ecoscale score of 62.5, and furnished 2a in a good yield of 73% with an excellent diastereomeric ratio (>98 : 2) (Table 1, entry 3). Even when careful attention was paid to use the minimum amount of solvent enabling proper agitation of this heterogeneous reaction media, 6 days were necessary to reach full conversion of the substrate. At this point of the study, we considered that using innocuous NaHCO₃ as the base would force us to use more polar solvents such as EtOH, THF or DMF. The use of THF or EtOH as the solvent was disappointing since 24 h and more than 6 days of reaction were respectively required to reach 95% of conversion (Table 1, entries 4 and 5). The use of DMF as the solvent resulted in a much shorter reaction time; only 2 h were required for a complete consumption of substrate 1 (Table 1, entry 6). Nevertheless, an unsatisfying Ecoscale score of 50.5 was obtained mainly due to the low yield (55%) and the fact that DMF presents high health-related risks, which hamper its environmental impact.⁹ Except for DMF, these relatively long reaction times could be attributed to the low solubility of either NaHCO₃ in organic solvents or α-bromo ester 1 in water, resulting in heterogeneous reaction mixtures that may lead to mass transfer limitations. As heterogeneity of the reaction mixture could be responsible for a low speed of reaction, the transformation was performed using the ball-milling technology.¹⁰ In this kind of apparatus, reagents (liquid or solid) are introduced into a jar with one or more balls. Rapid movements of the jar create repeated and violent contacts between reagents, balls and walls allowing for a very efficient mixing of solid-containing reaction mixtures.

Table 1 Influence of the solvent, base, and homogenization technique on the DKR of 1

Entry	Homogenization technique	Base (eq.)	Solvent	Time to reach >95% conversion^a	Yield^b (%)	dr^c (2a) : (2b)	Ecoscale score
a Determined by HPLC. b Isolated yield. c Determined using 300 MHz ^1H NMR. d Time to reach 50% conversion. e Obtained as a 50 : 50 molar mixture of 2a : 2b and 1. f Based on recovered starting material (brsm). g Not determined. h The η ratio is defined as the amount of added liquid to the sum of the mass of reactants. It is expressed in μL mg^−1.
1	Magnetic stirring	Et3N (2.0)	THF	5 h	59	>98 : 2	42.5
2	Magnetic stirring	NaHCO3 (1.2)	Water	6 h^d	34^e (50)^f	87 : 13	54
3	Magnetic stirring	NaHCO3 (1.2)	EtOAc	6 days	73	>98 : 2	62.5
4	Magnetic stirring	NaHCO3 (1.2)	THF	24 h	—^g	—^g	—^g
5	Magnetic stirring	NaHCO3 (1.2)	EtOH	>6 days	—^g	—^g	—^g
6	Magnetic stirring	NaHCO3 (1.2)	DMF	2 h	55	>98 : 2	50.5
7	Ball-milling	NaHCO3 (1.2)	Solvent-free	30 min	62	>98 : 2	63
8	Ball-milling	NaHCO3 (1.2)	Water (η = 2)^h	15 min	96	>98 : 2	80
9	Ball-milling	NaHCO3 (1.2)	Water (η = 1)^h	15 min	94	>98 : 2	79

---

Indeed, 30 min of vigorous agitation were enough for a complete conversion of α-bromo ester 1 into amino-ester 2a when 1 was placed in a 10 mL jar with one 10 mm diameter ball, NaHCO₃, dibenzylamine, and TBAI in the absence of any solvent (Table 1, entry 7). Under these conditions, dibenzylamino ester 2a was isolated in 62% yield with an excellent diastereomeric ratio (>98 : 2) and a satisfying Ecoscale score of 63. Obtainable benefits from adding a liquid in a grinded reaction mixture are now well established.^10b,11,12 Indeed, when adding small amounts of water in the jar (η ratio of 2 μL mg⁻¹)¹³ only 15 min were necessary to obtain complete conversion of the substrate while dibenzylamino ester 2a was isolated in excellent yield and diastereomeric ratio (96% yield, dr > 98 : 2; Table 1, entry 8). In these conditions, the highest Ecoscale score was obtained (80), corresponding to an excellent synthesis.⁵ Calculations leading to this high Ecoscale score include the solvent used during work-up and chromatographic purification (details on the Ecoscale calculations are available in the ESI†). Reducing the amount of water in the reaction media to 1 μL mg⁻¹ had little effect on the course of the reaction as 2a was obtained in 94% yield and >98 : 2 dr (Table 1, entry 9). We postulate that solving the suspected mass transfer limitations using the tremendous ability of the ball-mill technology to mix the solid-containing mixtures allowed:

– improvement of the yield without hampering the diastereoselectivity;

– reduction of reaction time;

– utilisation of the least problematic base (NaHCO₃) and solvent (H₂O) regardless of their solubility or solubilizing capacities.

To our knowledge, this is the first example of DKR to be performed through ball-milling technology.

At this stage of the study, we became interested in comparing the efficiency and the environmental impact of solution-based DKR reactions with the mechanochemistry-mediated approach on other known DKR reactions implying a S_N2 halogen substitution. When treated with (S)-(–)-α-methylbenzylamine in THF in the presence of Et₃N as the base, α-bromo ester 1 was transformed into 3 in 84% yield and >98 : 2 dr (Table 2, entry 1).⁶ The use of THF and Et₃N resulted in a low Ecoscale score of 52. Replacing these problematic chemicals with water and NaHCO₃ while mixing the reaction mixture with a vibrating ball-mill allowed for the production of 3 with much better Ecoscale score (72), high yield (80%) and diastereoselectivity (>98 : 2) (Table 2, entry 2). Treating α-bromo ester 1 with p-anisidine in a classical solution-based approach⁶ furnished amine 4 in 75% yield and >98 : 2 dr, though with a low Ecoscale score of 40.5. When applying our ball-mill mediated approach, the Ecoscale score could be improved up to 54, yet with a slightly lower yield (64%) and a drop in diastereoselectivity (88 : 12) (Table 2, entry 4). Oxygen-based nucleophiles such as p-methoxyphenol could also be used to perform DKR on α-bromo ester 1. Indeed, Durst and coworkers utilised NaH and THF to produce 5 with 70% yield and 95 : 5 dr, yet with a low Ecoscale score of 41 (Table 2, entry 5).¹⁴ Treatment of α-bromo ester 1 with p-methoxyphenol in a ball-mill with water and NaHCO₃ instead of THF and NaH resulted in the production of ester 5 with a better yield and an Ecoscale score of 86% and 62, respectively, albeit with a moderate 72 : 28 dr (Table 2, entry 6). When 2,6-dichlorophenol was used as a nucleophile, the reaction time, diastereoselectivity and the Ecoscale score were improved while the yield remained similar (Table 2, entries 7 and 8).¹⁵

Table 2 Comparison of classical solvent-based approaches with mechano-mediated DKR

Entry	Substrate	Product	Homogenization technique	Base	Solvent	Time	Yield^a (%)	dr	Ecoscale score
a Isolated yield. b Not indicated in the original publication. c Diastereomeric ratios were determined using 300 MHz ^1H NMR. d 2.16 eq. of Bn2NH were used. e 2.5 eq. of Bn2NH were used.
1			Magnetic stirring	Et3N	THF	—^b	84	>98 : 2	52
2			Ball-milling	NaHCO3	Water	30 min	80	>98 : 2^c	72
3	—		Magnetic stirring^6	Et3N	THF	—^b	75	>98 : 2	40.5
4			Ball-milling	NaHCO3	Water	6 h	64	88 : 12^c	54
5	—		Magnetic stirring^14	NaH	THF	6 h	70	95 : 5	41
6			Ball-milling	NaHCO3	Water	2 h	86	72 : 28^c	62
7	—		Magnetic stirring^15	NaH	THF	7 h	78	50 : 50	23.5
8			Ball-milling	NaHCO3	Water	1 h 30	77	56 : 44^c	47.5
9			Magnetic stirring^16	Et3N	CH2Cl2	24 h	81	87 : 13	51
10			Ball-milling	NaHCO3	Water	13 h	75^d	>98 : 2^c	68.5
11			Magnetic stirring^17	Bn2NH^e	THF	10 days	74	>95 : 5	58
12			Ball-milling	Bn2NH^e	Water	7 h	71	>98 : 2^c	66.5

---

After having changed the nature of nucleophiles, we focused our attention on other types of substrates such as α-bromo amide 7 and α-bromo ketone 9. When 1-(2-bromo-1-oxopropyl)-L-proline methyl ester 7 was reacted with Bn₂NH in CH₂Cl₂ in the presence of Et₃N for 24 h, N,N-dibenzyl-D-alanyl-L-proline methyl ester 8 was obtained in 81% yield with 87 : 13 dr, accounting for an Ecoscale score of 51 (Table 2, entry 9).¹⁶ Once again, utilising the ball-milling technology avoided the use of problematic CH₂Cl₂ and Et₃N. Thus, the reaction of α-bromo amide 7 with Bn₂NH in the presence of NaHCO₃ and water in a vibrating ball-mill furnished product 8 with a similar yield (75%) and with an improvement in the Ecoscale score (68.5 vs. 51 with the solution-based approach) (Table 2, entry 10). It is worth noting that utilisation of the ball-mill also resulted in improving the diastereomeric ratio from 87 : 13 to >98 : 2.

Finally, this approach was applied to the DKR reaction of γ-bromo-β-ketosulfoxide 9 with Bn₂NH. Salom-Roig and coworkers¹⁷ reported that treatment of 9 with 2.5 equivalents of Bn₂NH required 10 days in THF to reach reaction completion furnishing γ-dibenzylamino-β-ketosulfoxide 10 with 74% yield and >95 : 5 dr (Table 2, entry 11). Using Bn₂NH as the nucleophile and the base to deprotonate the HBr salts generated during the course of the reaction, the authors already avoided the use of the flammable, corrosive and toxic Et₃N, thus reducing the environmental impact of the reaction. This particularly resulted in a good Ecoscale score of 58. Yet, when the same reaction was performed in a ball-mill with water replacing THF, the time for reaction completion was dramatically reduced to only 7 h, while furnishing γ-dibenzylamino-β-ketosulfoxide 10 with a similar yield of 71% and diastereoselectivity (Table 2, entry 12). Replacing THF with water and reducing the required time to reach reaction completion resulted in an Ecoscale score improvement up to 66.5.

Conclusions

In conclusion, the great capacity of the ball-milling technology to mix solid-containing reaction mixtures allowed the design of reaction conditions that could mitigate the environmental impact of dynamic kinetic resolutions. This reduction was evaluated by calculating the Ecoscale score of every studied reaction conditions. Thus, we have shown that problematic chemicals such as THF or Et₃N could be replaced by innocuous water and NaHCO₃ without dramatically hampering the performance of the DKR reactions. In all cases, the time required to completion was severely reduced and in some cases yields and/or diastereomeric ratios were also improved. Exemplification of utilisation of the ball-milling technology to reduce the environmental impact of other problematic reactions is currently under progress.

We thank CNRS, Université Montpellier 1 and Université Montpellier 2 for financial support. François Métro is gratefully acknowledged for producing graphical abstract artwork.

References

1. (a) H. Pellissier, Tetrahedron, 2003, 59, 8291–8327; (b) H. Pellissier, Tetrahedron, 2008, 64, 1563–1601; (c) Y. S. Park, Tetrahedron: Asymmetry, 2009, 20, 2421–2427; (d) H. Pellissier, Tetrahedron, 2011, 67, 3769–3802.
2. DCM stands for dichloromethane and THF for tetrahydrofuran. DCM and THF are classified as toxic and carcinogenic (Category 2) according to Regulation (EC) No. 1272/2008.
3. Triethylamine is classified as highly flammable (R11), corrosive (R35) and harmful (R20/21/22) according to EU Directives 67/548/EEC or 1999/45/EC.
4. (a) V. Declerck, P. Nun, J. Martinez and F. Lamaty, Angew. Chem., Int. Ed., 2009, 48, 9318–9321; (b) T.-X. Métro, J. Bonnamour, T. Reidon, J. Sarpoulet, J. Martinez and F. Lamaty, Chem. Commun., 2012, 48, 11781–11783; (c) J. Bonnamour, T.-X. Métro, J. Martinez and F. Lamaty, Green Chem., 2013, 15, 1116–1120.
5. K. Van Aken, L. Strekowski and L. Patiny, Beilstein J. Org. Chem., 2006, 2, 3.
6. R. N. Ben and T. Durst, J. Org. Chem., 1999, 64, 7700–7706 . Reaction times were not indicated in the publication.
7. NaHCO₃ is considered to be a not hazardous substance as defined by the European regulation No. 1272/2008 and not classified as dangerous according to Directive 67/548/EEC.
8. (a) K. Alfonsi, J. Colberg, P. J. Dunn, T. Fevig, S. Jennings, T. A. Johnson, H. P. Kleine, C. Knight, M. A. Nagy, D. A. Perry and M. Stefaniak, Green Chem., 2008, 10, 31–36; (b) R. K. Henderson, C. Jiménez-González, D. J. C. Constable, S. R. Alston, G. G. A. Inglis, G. Fisher, J. Sherwood, S. P. Binks and A. D. Curzons, Green Chem., 2011, 13, 854–862; (c) D. Prat, O. Pardigon, H.-W. Flemming, S. Letestu, V. Ducandas, P. Isnard, E. Guntrum, T. Senac, S. Ruisseau, P. Cruciani and P. Hosek, Org. Process Res. Dev., 2013, 17, 1517–1525.
9. DMF is classified as presenting reproductive toxicity according to Regulation (EC) No. 1272/2008.
10. (a) B. Rodríguez, A. Bruckmann, T. Rantanen and C. Bolm, Adv. Synth. Catal., 2007, 349, 2213–2233; (b) S. L. James, C. J. Adams, C. Bolm, D. Braga, P. Collier, T. Friščić, F. Grepioni, K. D. M. Harris, G. Hyett, W. Jones, A. Krebs, J. Mack, L. Maini, A. G. Orpen, I. P. Parkin, W. C. Shearouse, J. W. Steed and D. C. Waddell, Chem. Soc. Rev., 2012, 41, 413–447; (c) G. W. Wang, Chem. Soc. Rev., 2013, 42, 7668–7700.
11. (a) E. Boldyreva, Chem. Soc. Rev., 2013, 42, 7719–7738; (b) G. A. Bowmaker, Chem. Commun., 2013, 49, 334–348.
12. T. Friščić, S. L. Childs, S. A. A. Rizvi and W. Jones, CrystEngComm, 2009, 11, 418–426.
13. The η ratio is defined as the amount of added liquid to the sum of the mass of reactants. It is expressed in μL mg⁻¹.
14. K. Koh and T. Durst, J. Org. Chem., 1994, 59, 4683–4686.
15. G. Massolini, G. Fracchiolla, E. Calleri, G. Carbonara, C. Temporini, A. Lavecchia, S. Cosconati, E. Novellino and F. Loiodice, Chirality, 2006, 18, 633–643.
16. J. Nam, J.-Y. Chang, K.-S. Hahm and Y. S. Park, Tetrahedron Lett., 2003, 44, 7727–7730.
17. P.-Y. Géant, J. Martinez and X. J. Salom-Roig, Eur. J. Org. Chem., 2011, 1300–1309.

---

Footnote

† Electronic supplementary information (ESI) available: General experimental procedures, characterisation data of all synthesized compounds and details on the calculations of Ecoscale scores. See DOI: 10.1039/c4gc01416b

---