Optical resolution of H/D isotopic chirality via asymmetric C–H amination

Abstract

Isotopic chirality arising from hydrogen/deuterium substitution has long resisted optical resolution because isotopic substitution introduces only minimal stereochemical differences, rendering such molecules structurally pseudo-achiral. Recognizing that the small yet distinct difference in C–H/C–D bond dissociation energies (BDEs) creates an energetic asymmetry, we envisioned that an enantioselective C–H functionalization exploiting this disparity could achieve the kinetic resolution of isotopically chiral molecules. Here, we report a catalytic method for the optical resolution of benzylic and allylic H/D isotopic chirality. This method achieved a remarkably high separation factor (S = 83.1), enabling the isolation of highly enantioenriched isotopically chiral substrates from racemic mixtures and revealing retention of stereochemical memory during radical recombination in asymmetric C–H functionalization.

INTRODUCTION

Replacement of hydrogen by deuterium represents the smallest possible structural modification of an organic molecule, yet it has profound consequences in chemistry owing to the kinetic isotope effect (KIE). H/D substitution has become an indispensable strategy for mechanistic investigations, isotope labeling, and the development of deuterated pharmaceuticals with improved metabolic stability. Beyond these established applications, H/D substitution can also generate isotopic chirality, in which enantiomers differ solely in isotopic composition. Notably, isotopic substitution represents the smallest structural difference capable of generating molecular chirality. Accordingly, isotopically chiral molecules constitute one of the most subtle forms of molecular chirality. Access to highly enantioenriched isotopically chiral CHDXY compounds therefore provides a valuable opportunity to rigorously investigate the stereochemical consequences of isotopic substitution. Their general availability is expected to stimulate advances in chiral recognition, precision chiral spectroscopy, and broader studies of isotope-induced stereochemistry (1, 2).

This type of chirality remains one of the few classes of chirality that cannot be resolved into enantiomerically pure forms from their racemic mixtures by established optical resolution methodologies. Although recent advances in spectroscopic techniques, including anisotropic NMR in chiral oriented media, have enabled the enantiodiscrimination of certain isotopically chiral molecules, these methods are primarily analytical and do not offer a practical solution for the preparative resolution of such compounds (1–13). As a result, access to such molecules has largely depended on derivation from pre-existing chiral frameworks or asymmetric synthesis (14–18). Representative strategies such as recrystallization of diastereomeric salts and chiral stationary phase-supported HPLC have proven ineffective (19–25). This is because isotopic substitution does not impart substantial stereochemical differences, such as steric hindrance or bond length, causing these molecules to behave as if they were achiral. Several attempts have been reported, yet recrystallization methods were irreproducible, while chiral HPLC typically afforded separation factors (S) close to unity, as exemplified with phenyl-d₅-phenylmethanol and isotopically substituted diphenylmethylamine derivatives (Fig. 1A). To date, one of the few successful examples has been reported by Kimata and co-workers, who resolved isotopically chiral diphenylmethanol, albeit requiring 65 cycles of recycling chiral HPLC (24). Notably, however, there have been no successful examples of resolving isotopic chirality arising from the simplest case, monodeuterated compounds. These limitations underscore the absence of a general and practical strategy for the optical resolution of isotopic chirality, highlighting the urgent need for effective methodologies.

Fig. 1. Optical resolution of isotopic chiral molecules.

(A) Previous examples: Diastereomeric interaction with chiral resolving reagents. (B) This work: kinetic resolution of isotopic chirality through C–H amination.

To address these long-standing challenges, we turned to the subtle but fundamental difference in BDE between hydrogen and deuterium, the only energetic handle available for stereochemical discrimination. The difference in C–H/C–D BDEs (≈5 kJ/mol) provides the only measurable energetic asymmetry that can be exploited kinetically. Indeed, although isotopic chiral molecules appear stereochemically pseudo-achiral, the difference in C–H/C–D BDEs renders them energetically chiral. By directly engaging the isotopic stereogenic center as the reaction site and enforcing an irreversible hydrogen atom abstraction (HAT) process as the rate-determining step, we sought to transform this elusive energetic gap into a decisive stereochemical outcome. Such a system requires a catalytic process in which C–H cleavage constitutes the rate-determining step and occurs with precise stereochemical control. Guided by this concept, we chose the chiral Ru(CO)salen complex as the catalytic system because its nitrene-transfer pathway is known to involve a rate-determining HAT step (Fig. 1B).

In such asymmetric C–H functionalizations, the relative rate difference between isotopic enantiomers can be conceptualized as the product of two contributions: the intrinsic enantioselective rate difference of a prochiral substrate undergoing asymmetric transformation, and the kinetic isotope effect (KIE) arising from the C–H/C–D BDE gap (Fig. 2) (26–30). Based on this understanding, it is strongly suggested that an effective system must not only precisely control the substrate’s approach, orientation, and conformation, but also preferentially abstract the hydrogen atom over deuterium during the functionalization step.

Fig. 2. Kinetic resolution of isotopic chirality through KIE-governed C–H functionalization.

Here we demonstrate that isotopic chirality, once considered practically unresolvable, can be addressed through optical resolution. Chiral ruthenium–salen complexes catalyze asymmetric benzylic and allylic C–H amination with azides as nitrene sources, enabling the kinetic resolution of H/D isotopic chirality. This transformation not only recovers isotopically chiral molecules in highly enantioenriched form but also achieves a remarkably high S value of 83.1.

These findings establish a general strategy for resolving isotopic chirality and, more broadly, reveal that bond energetics, rather than geometry alone, can serve as a powerful axis for stereochemical discrimination, expanding the scope of asymmetric synthesis.

RESULTS

Previously, we found that the Ru(CO)salen complex 1 served as an efficient catalyst for C–H amination (Fig. 3) (31). The ruthenium-catalyzed C–H amination occurred preferentially at less hindered benzylic and allylic positions, with good to excellent enantio- and site-selectivity (Fig. 3A). Moreover, in mechanistic studies using cis-β-ethylstyrene derivatives as the radical clocks, cis/trans isomerization was observed only in the reaction using the electron-donating methoxy-substituted derivative (Fig. 3B). This observation indicated that the very short-lived radical intermediates were generated through a HAT process involving the putative ruthenium-nitrene intermediate (Fig. 3C). These findings strongly suggested that this amination process would be effective for the kinetic resolution of isotopic chirality.

Fig. 3. Ru(CO)salen-catalyzed asymmetric C–H amination.

(A) Overview of the ruthenium-catalyzed asymmetric C–H amination. (B) Mechanistic studies of the ruthenium-catalyzed C–H amination using cis-β-ethylstyrene derivatives as radical clocks. (C) Plausible reaction mechanism of the ruthenium-catalyzed C–H amination. (D) Competitive and independent measurements of the KIE in the ruthenium-catalyzed C–H amination.

Therefore, we measured the deuterium KIE of the ruthenium-catalyzed C–H amination using 2-ethylnaphthalene as the substrate. For this purpose, we employed the previously reported complex 1 and the modified complex 2, designed to improve catalyst robustness (31–33). Large KIE values were observed: 11 and 10 in competitive conditions, and 6 and 10 under independent measurements, respectively (Fig. 3D). Collectively, these results establish beyond doubt that ruthenium-catalyzed asymmetric C–H amination fulfills the critical mechanistic requirement for the kinetic resolution of isotopic chirality: a rate-limiting HAT step.

Encouraged by these observations, we tested whether ruthenium-salen-catalyzed asymmetric C–H amination could resolve isotopic chirality (Fig. 4A). Using racemic mono-deuterated 2-(1-d₁-ethyl)naphthalene, the reaction catalyzed by complex 1 at −10°C delivered the corresponding amination product with an enantiomeric ratio of 96:4 at 51.5% conversion. Although the enantiomeric ratio of the recovered starting material could not be determined directly, the recovered isotopically chiral substrate was also examined by anisotropic ²H NMR spectroscopy using a chiral gel matrix; however, no detectable enantiomeric discrimination was observed under the conditions examined. Chiral LC-MS analysis of the deuterium content of each product enantiomer allowed us, based on the premise that the stereochemistry of the amination product reflects that of the starting material, to deconvolute the overall conversion into contributions from each isotopic enantiomer, thereby enabling estimation of the enantiomeric ratio of the recovered substrate (Fig. 4B). As a result, the analysis revealed that the recovered starting material was enriched to an enantiomeric ratio of 83:17, corresponding to an S value of 9, which was calculated by Kagan’s first-order equation (34, 35). This result demonstrates that isotopic chirality can be successfully resolved by kinetic resolution through ruthenium-catalyzed asymmetric C–H amination. Furthermore, a higher S value of 10 was observed in the reaction catalyzed by complex 2, affording the recovered substrate with an enantiomeric ratio of 90:10 at 56.5% conversion.

Fig. 4. Kinetic resolution of rac-2-(1-d₁-ethyl)naphthalene via ruthenium-catalyzed asymmetric amination.

(A) Ru(CO)salen-catalyzed asymmetric C–H amination of rac-2-(1-d₁-ethyl)naphthalene. (B) Estimation of the enantiomeric ratio of the recovered substrate and calculation of the separation factor (S) using Kagan’s first-order equation.

The kinetic resolution with complex 2 as the catalyst was also examined in various solvents, including haloalkanes, halobenzenes, and ethyl acetate (table S1). With the exception of the strongly coordinating solvent THF, these conditions afforded good S values (entries 1–8). Among them, CH₂Cl₂ enabled recovery of the starting material with a good enantiomeric ratio at an appropriate level of conversion (entry 1). Furthermore, lowering the reaction temperature enhanced the selectivity. At −40°C, the kinetic resolution reached an S value of 15, allowing recovery of the isotopically chiral substrate in highly enantioenriched form at 61.7% conversion (entry 12). Lowering the substrate concentration to 1.0 M further improved both the enantiomeric ratio and the S value (entry 13; S = 19, estimated er = 0.4:99.6).

With these preliminary results in hand, we expanded the study to the kinetic resolution of benzylic and allylic H/D isotopic chiral substrates (Fig. 5A). A series of mono-deuterated 1-d₁-ethylbenzene derivatives underwent ruthenium-catalyzed asymmetric C–H amination to give enantiomerically enriched starting materials with good to excellent separation factors. Although the precise electronic effects of substituents remain unclear, a general trend was observed: para-substituted derivatives tended to show higher conversions than others. In particular, substrates bearing moderately bulky substituents such as bromine, chlorine, methoxy, and siloxy groups at the para-position consistently afforded highly enantioenriched recovered materials with excellent selectivity. Notably, para-bromoethylbenzene achieved the highest separation factor (S = 83.1), recovering the highly enantioenriched isotopically chiral compound at 55.0% conversion. This scaffold is a known synthetic intermediate for the preparation of the SGLT2 inhibitor tofogliflozin, highlighting the potential of the present methodology for providing enantioenriched isotopically chiral building blocks for the synthesis of pharmaceutically relevant compounds (36). These observations suggest that substrates facilitating hydrogen atom abstraction at the benzylic position tend to provide better outcomes. Substituents capable of stabilizing the benzylic radical intermediate while introducing minimal steric hindrance near the reaction site appear to favor both higher conversion and selectivity.

Fig. 5. Kinetic resolution of isotopic chiral molecules and the determination of the actual enantiomeric ratio of the recovered substrate.

(A) Substrate scope of the kinetic resolution of isotopic chirality. ^a Reaction conditions: substrate (0.2 mmol), Ru(CO)salen 2 (4.0 μmol), SESN₃ (0.15 mmol), MS4Å (20 mg) in CH₂Cl₂ (0.1 mL) at −40°C under N₂. ^b Determined by GC using bicyclohexyl as the internal standard. ^c Estimated based on enantiomeric ratio of amination products, conversion of the starting materials, and D-contents of each enantiomer of the amination products. ^d Calculated by using Kagan’s first-order equation. ^e At −50°C. ^f Using 5 mol% of 2 at −15°C. (B) The recovered substrate was converted into a diastereomeric mixture of tert-alcohols via a four-step sequence. (C) Determination of the actual enantiomeric ratio of the recovered trans-3-d₁-1-phenylbutene. ^g Each diastereomer was independently analyzed by ¹H NMR spectroscopy (600 MHz, CDCl₃).

Beyond simple ethylbenzene derivatives, the method was also applicable to functionalized substrates. For example, the Weinreb amide, (E)-4-d₁-N,O-dimethyl-N-hydroxy-2-pentenamide, proceeded sluggishly under standard conditions and required a slightly higher reaction temperature (−15°C) to achieve measurable conversion. Even so, the reaction afforded kinetic resolution with an S value of 8, recovering the starting material with an enantiomeric ratio of 80.5:19.5, thus demonstrating the broader applicability of this strategy. However, these results are based on estimates, and to ascertain whether genuine kinetic resolution has taken place, it is essential to determine the actual enantiomeric ratio of the recovered isotopic chiral compounds.

To experimentally determine the actual enantiomeric ratio of an isotopic chiral molecule, we further investigated the reaction of rac-trans-3-d₁-1-phenylbutene. The recovered starting material exhibited an estimated enantiomeric ratio of 94:6 at 61.9% conversion, corresponding to a calculated S value of 9. To directly verify this estimate, the recovered substrate was transformed into a diastereomeric mixture of tertiary alcohols through a four-step sequence: olefin epoxidation, reductive epoxy-ring opening, alcohol oxidation, and nucleophilic addition with a phenyl Grignard reagent (Fig. 5B). The resulting tertiary alcohols were separated into two fractions according to the configuration of the newly introduced hydroxy-bearing stereocenter using chiral HPLC. Comparison with authentic samples prepared from (S)-ethyl lactate confirmed that the recovered isotopic chiral substrate was R-enriched in a 93:7 enantiomeric ratio, as determined by ¹H NMR analysis (Fig. 5C) (37). This result closely matched the estimated value, thereby experimentally validating that true kinetic resolution had been achieved. Furthermore, the excellent agreement between the estimated and experimentally determined enantiomeric ratios of the recovered isotopically chiral substrate strongly suggests that the radical intermediate, generated via hydrogen atom abstraction by the putative ruthenium–nitrene species, retains its stereochemical information to a large extent during the radical–radical recombination step leading to functionalization.

DISCUSSION

In summary, we have established a preparation method for highly enantioenriched H/D isotopically chiral compounds from racemic mixtures, representing a substantial advance in chemical synthesis. The lack of substantial stereochemical differences between isotopes has long hindered the synthesis of highly enantioenriched compounds from racemic mixtures. Moreover, no general and practical method for the preparative optical resolution of such compounds has been established. To overcome this challenge, we developed an innovative kinetic resolution method using the chiral Ru(CO)salen 2-catalyzed C–H amination reaction with azide as the nitrogen source. This method enabled the efficient synthesis of H/D isotopically chiral molecules, particularly ethylbenzene derivatives and specific alkenes bearing H/D stereogenic centers, achieving S values ranging from 6.9 to 83.1.

The key to our success lies in the strategic utilization of the subtle yet decisive difference in the BDEs between C–H and C–D bonds, combining high enantioselectivity with a KIE to generate relative differences in reaction rates between H/D isotopically chiral enantiomers. This process preferentially converted one enantiomer into the product, thereby increasing the enantiomeric purity of the remaining substrate. Furthermore, the analysis provides mechanistic insight into the ability of C–H functionalization to preserve H/D isotopic stereochemical information, as evidenced by the consistency between the estimated and actual enantiomeric ratios.

We believe that the concept of kinetically resolving H/D isotopically chiral molecules represents an important strategy for accessing highly enantioenriched H/D isotopically chiral compounds, for which no general and practical method for the preparative optical resolution has been established. To further enhance the applicability of our methodology, it will be important to explore various catalytic C–H functionalization strategies, including transition-metal catalysis and enzymatic approaches, enabling access to a broader range of H/D isotopically chiral compounds. This work thus provides a practical platform for accessing highly enantioenriched isotopically chiral molecules, thereby expanding opportunities for systematic studies of isotopic chirality and its future applications in chemistry, biology, and functional molecular systems. More broadly, it establishes a framework for asymmetric catalysis in which bond energetics, rather than structural differences alone, dictate stereochemical outcomes.

MATERIALS AND METHODS

¹H NMR and ¹³C NMR spectra were recorded on a JEOL JNM-AL-400 spectrometer at 400 and 100 MHz, respectively. ²H NMR spectra were recorded on a Bruker AVANCE III 600 spectrometer at 92 MHz. Chemical shifts are expressed in parts per million (ppm) with respect to tetramethylsilane (TMS) as the internal standard (0 ppm) for ¹H NMR, and CDCl₃ as the internal standard (77.0 ppm) for ¹³C NMR. Coupling constants were reported as hertz (Hz), signal shapes and splitting patterns are indicated as follows: s, singlet; br, broad; d, doublet; t, triplet; q, quartet; quin, quintet; sext, sextet; dd, doublet of doublets; m, multiplet. Mass spectra were recorded on a BRUKER DALTONICS MICRO TOF-KS1 focus (HR-ESI); peaks are given in m/z (% of base peak). Infrared spectra were obtained with SHIMADZU IRAffinity-1 FTIR spectrophotometer, and only diagnostic signals are listed below. High-performance liquid chromatography was carried out using SHIMADZU HPLC Prominence equipped with a column packed with an appropriate optically active material. Optical rotations were recorded on an ANTON PAAR MCP-300. Gas chromatography was performed on a SHIMADZU GC-2025 using a COATING CP CHIRASIL-DEX CB column (30 m, 0.25 mm, 1.4 μm) or a HP-5 column. Thin-layer chromatography was performed on Merck silica gel 60 F254-coated glass plates and visualized by fluorescence quenching under UV light or staining with the standard solutions of 12MoO₃·H₃PO₄, p-anisaldehyde and ceric ammonium molybdate. Column chromatography was conducted on Cica (Kanto Chemical Co., INC.) silica gel 60 N (spherical, neutral, 63–210 μm), or Fuji Silysia Chemical Ltd. Chromatorex NH silica gel (spherical, basic, 100–200 μm). Ru(CO)salen 1 and 2-(trimethylsilyl)ethane-1-sulfonyl azide were prepared according to the reported procedures (31–33). Other starting materials were synthesized according to the following procedures. All substrates were distilled before use.

Typical Ruthenium-Catalyzed C–H Amination

Molecular sieves 4 Å (20 mg), 2-ethylnaphthalene (31.5 μL, 0.20 mmol), and bicyclohexyl (10 μL) as an internal standard for GC analysis were placed in a 5.0 mL Schlenk tube under N₂. A CH₂Cl₂ (200 μL) solution of the Ru(CO)-salen complex (0.20 μmol, 0.10 mol%) was added, and the mixture was cooled to −40°C. After stirring for 30 min, 2-(trimethylsilyl)ethane-1-sulfonyl azide (SESN₃) (48.9 μL, 0.26 mmol) was slowly added to the reaction mixture.

An aliquot (10 μL) of the reaction mixture was withdrawn at time zero (and at designated time points thereafter), diluted with EtOH, and analyzed by GC using a CP CHIRASIL-DEX CB chiral column to determine the conversion. After completion of the reaction, the reaction mixture was directly subjected to silica gel chromatography to afford the aminated product. The enantiomeric ratio (er_pro) of the corresponding amide and the D-content of each stereoisomer were determined by LC–MS analysis using a chiral stationary phase.

Kinetic Isotope Effect (KIE) Measurement under Independent Conditions

2-Ethylnaphthalene (A, 31.5 μL, 0.20 mmol) or 2-(1,1-d₂-ethyl)naphthalene (B, 31.6 μL, 0.20 mmol), Ru(CO)-salen complex 1 or 2 (4.0 μmol, 2 mol%), molecular sieves 4 Å (20 mg), and bicyclohexyl (10 μL) as an internal standard for GC analysis were placed in a 5.0 mL Schlenk tube under N₂. CH₂Cl₂ (200 μL) was added, and the reaction mixture was cooled to −10 or − 40°C. After stirring for 30 min, SESN₃ (37.6 μL, 0.20 mmol) was slowly added to the mixture.

An aliquot (10 μL) of the reaction mixture was withdrawn at time zero (and at designated time points thereafter), diluted with EtOH, and analyzed by GC using a CP CHIRASIL-DEX CB chiral column to determine conversion.

KIE Measurement under Intermolecular Competition Conditions

2-Ethylnaphthalene (15.8 μL, 0.10 mmol) and 2-(1,1-d₂-ethyl)naphthalene (15.8 μL, 0.10 mmol), Ru(CO)-salen complex 1 or 2 (4.0 μmol, 2 mol%), molecular sieves 4 Å (20 mg), and bicyclohexyl (10 μL) as an internal standard for GC analysis were placed in a 5.0 mL Schlenk tube under N₂. CH₂Cl₂ (200 μL) was added, and the reaction mixture was cooled to −10 or − 40°C. After stirring for 30 min, SESN₃ (37.6 μL, 0.20 mmol) was slowly added to the reaction mixture.

The reaction was stirred for 24 h, and the crude mixture was directly subjected to silica gel chromatography (hexane/EtOAc = 4:1) to afford the corresponding amide. The D-content of the product was determined by ¹H NMR analysis.

Optimized Reaction Conditions: Kinetic Resolution of H/D Isotopic Chiral Compounds via Ru(CO)-salen Complex 2-Catalyzed C–H Amination

rac-2-(1-d₁-Ethyl)naphthalene (0.20 mmol), Ru(CO)-salen complex 2 (4.54 mg, 4.0 μmol, 2 mol%), molecular sieves 4 Å (20 mg), and bicyclohexyl (10 μL) as an internal standard for GC analysis were placed in a 5.0 mL Schlenk tube under N₂. Solvent (100 μL) was added, and the reaction mixture was cooled to −10 to −40°C. After stirring for 30 min, SESN₃ (28.2 μL, 0.15 mmol) was slowly added, and the reaction mixture was stirred at that temperature.

At appropriate time intervals, an aliquot (10 μL) was withdrawn from the reaction mixture as the time zero (or time-point) sample, diluted with EtOH, and analyzed by GC using a CP CHIRASIL-DEX CB chiral column to determine conversion. The crude mixture was then directly subjected to silica gel chromatography to isolate the aminated product.

The enantiomeric ratio of the aminated product (er_amide) and the deuterium content of each isomer were determined by LC–MS analysis using a chiral stationary phase column. The enantiomeric ratio of the recovered substrate (er_sub.) was estimated based on these measured values according to the reported method. The separation factors (S) were calculated using Kagan’s equation (eq. 3) from the conversion and estimated er_sub. values (34, 35).

General Procedure for Kinetic Resolution of H/D Isotopic Chiral Compounds

rac-Monodeuterated substrate (0.20 mmol), Ru(CO)-salen complex 2 (4.54 mg, 4.0 μmol, 2 mol%), molecular sieves 4 Å (20 mg), and bicyclohexyl (10 μL) as an internal standard for GC analysis were placed in a 5.0 mL Schlenk tube under N₂. CH₂Cl₂ (100 μL) was added, and the reaction mixture was cooled to −40°C. After stirring for 30 min, SESN₃ (28.2 μL, 0.15 mmol) was slowly added to the reaction mixture.

At appropriate time intervals, an aliquot (10 μL) was withdrawn from the reaction mixture as the time zero (or time-point) sample, diluted with EtOH, and analyzed by GC using either a CP CHIRASIL-DEX CB or an HP-5 column to determine the conversion. The reaction mixture was then directly subjected to silica gel chromatography to isolate the aminated product.

The enantiomeric ratio of the aminated product (er_amide) and the deuterium content of each isomer were determined by LC–MS analysis using a chiral stationary phase column. The enantiomeric ratio of the recovered substrate (er_sub) was estimated from these measured values. The separation factors (S) were calculated using Kagan’s equation based on the conversion and estimated er_sub values.

Supplementary Materials

This PDF file includes:

Supplementary Text

Supplementary Notes S1 to S4

Figs. S1 to S45

Table S1