Rapid Enantiomeric Excess Measurements of Enantioisotopomers by Molecular Rotational Resonance Spectroscopy

Abstract

Recent work in drug discovery has shown that selectively deuterated small molecules can improve the safety and efficacy for active pharmaceutical ingredients. The advantages derive from changes in metabolism resulting from the kinetic isotope effect when deuterium is substituted for a hydrogen atom at a structural position where rate limiting C−H bond breaking occurs. This application has pushed the development of precision deuteration strategies in synthetic chemistry that can install deuterium atoms with high regioselectivity and with stereocontrol. Copper-catalyzed alkene transfer hydrodeuteration chemistry has recently been shown to have high stereoselectivity for deuteration at the metabolically important benzyl C−H position. In this case, stereocontrol results in the creation of enantioisotopomers—molecules that are chiral solely by virtue of the deuterium substitution—and chiral analysis techniques are needed to assess the reaction selectivity. It was recently shown that chiral tag molecular rotational resonance (MRR) spectroscopy provides a routine way to measure the enantiomeric excess and establish the absolute configuration of enantioisotopomers. High-throughput implementations of chiral tag MRR spectroscopy are needed to support optimization of the chemical synthesis. A measurement methodology for high-throughput chiral analysis is demonstrated in this work. The high-throughput ee measurements are performed using cavity-enhanced MRR spectroscopy, which reduces measurement times and sample consumption by more than an order-of-magnitude compared to the previous enantioisotopomer analysis using a broadband MRR spectrometer. It is also shown that transitions for monitoring the enantiomers can be selected from a broadband rotational spectrum without the need for spectroscopic analysis. The general applicability of chiral tag MRR spectroscopy is illustrated by performing chiral analysis on six enantioisotopomer reaction products using a single molecule as the tag for chiral discrimination.

INTRODUCTION

The synthesis and characterization of deuterated small molecules continues to evolve and gain importance, especially in pharmaceutical process chemistry. Selective incorporation of deuterium into positions prone to metabolic oxidation may favorably alter the absorption, distribution, metabolism, and excretion (ADME) properties of a given drug molecule.¹ This concept, known as the deuterium switch,² was successfully deployed in the development of the first FDA approved deuterated drug, deutetrabenazine,³ in 2017. While applying deuterium switch tactics to improve current medicines continues to yield potential drug candidates, deuterium is also being incorporated in the early stages of the drug discovery process.² This approach led to the second FDA approved deuterated drug, deucravacitinib, in 2022.

Beyond applications in the synthesis of novel drug candidates, deuterated small molecules have other critical roles in process chemistry, especially as molecular probes. With the high-performance capability of mass spectrometers, deuterated drug analogues containing at least three deuterium atoms, or deuterium in combination with other stable isotopes such as ¹³C or ¹⁵N to achieve M ≥ 3, can be utilized as internal standards in measurements involving human samples.^4–6 Deuterated small molecules also play a crucial role in the development of reactions in process chemistry. They are utilized to elucidate reaction mechanisms in certain synthetic steps.^4,7 Importantly, this has led to improved syntheses of active pharmaceutical ingredients.⁸

The increasing applications of selectively deuterated small molecules as novel drug candidates and tools for studying chemical processes has driven demand for highly selective deuteration reaction methodologies. A key challenge in deuteration reactions is controlling the exact quantity and placement of deuterium in a small molecule. While hydrogen isotope exchange reactions provide a powerful strategy to incorporate deuterium into small molecules, especially for late-stage deuteration,^9–16 it remains difficult to control the quantity and placement of deuterium in the reaction products.^17,18 The inherent challenges with nonselective deuteration reactions are that isotopic product mixtures are typically inseparable on chromatographic columns, and traditional spectroscopic techniques (e.g., NMR, IR, and MS) are insufficient for characterizing and quantifying all isotopic species in complex product mixtures. These challenges may limit the utility of deuterated reaction products for use in ADME studies, kinetic studies, or drug development, as misleading or compromised data can result if unknown quantities of underdeuterated, misdeuterated, or overdeuterated species persist as a mixture with the desired isotopically labeled target molecule under study.^19,20

Recently, precision deuteration reactions have emerged that incorporate deuterium or hydrogen in a highly controlled fashion across alkene or alkyne functionality in a small molecule (Scheme 1a).^18,21–33 Using transfer hydrodeuteration techniques, our research groups have developed an enantioselective reaction to make enantioisotopomers that are chiral by virtue of deuterium substitution (Scheme 1b).³⁴ Now, enantioisotopomers can be readily accessed in one step from achiral and readily accessible alkene precursors. To enable this chemistry, we also developed new molecular rotational resonance (MRR) spectroscopy analysis methods to quantify all isotopic species in reaction mixtures.^23,35 Furthermore, we reported “chiral tagging” in MRR to provide the first general technique for enantiomeric excess (ee) and absolute configuration (ac) determination for enantioisotopmers.³⁴

Scheme 1. Overview of the Synthesis and Characterization of Isotopic Products from Transfer Hydrodeuteration Reactions^A.

A (a) Transfer hydrodeuteration involves the addition of HD to an organic molecule from a non-H₂, HD, or D₂ source or sources. (b) Prior work involving enantioselective copper-catalyzed alkene transfer hydrodeuteration for the synthesis of high-enantiopurity enantioisotopomers. (c) N-heterocycle-containing substrates were found to lead to reduced enantioselectivities for the copper-catalyzed enantioselective alkene transfer hydrodeuteration reaction. (d) This work describes the development of a high-throughput chiral tag MRR spectroscopy technique that is used to support reaction optimization for the enantioselective copper-catalyzed alkene transfer hydrodeuteration reaction.

The separation of enantiomers presents a general challenge to pharmaceutical chemistry.³⁶ In particular, there is recognition for the need of high-throughput techniques for ee determination.^37–42 High-speed techniques are needed, for example, to screen catalysts for asymmetric reactions⁴³ and to develop chiral building block toolboxes^44,45 in drug discovery. Emerging continuous manufacturing methods based on flow chemistry concepts^46–50 will also require rapid ee determination techniques for real-time monitoring of chiral species in crude reaction mixtures.⁵¹ MRR spectroscopy is an emerging spectroscopic technique for analytical chemistry. It offers high chemical specificity so that challenging problems in structural chemistry, such as the presence of regioisomers or diastereomers, can be tackled directly from the crude reaction mixture.⁵² High-speed measurements can be performed using cavity-enhanced spectrometer designs that can be coupled to flow chemistry systems for online analysis.⁵³ High-speed measurements also facilitate rapid impurity screening of raw materials in pharmaceutical production.⁵⁴ MRR spectroscopy has recently been extended to chiral analysis using the chiral tag rotational spectroscopy approach,^34,55–58 and this measurement methodology is extended in the present analysis of enantioisotopomers.

Following the initial studies of highly enantioselective deuteration reactions, our laboratories discovered several reactions that resulted in diminished ee of the product. Specifically, substrates containing N-heterocycles were problematic (Scheme 1c). Given the ubiquity of quinoline heterocycles in drug molecules,^59,60 we undertook further reaction optimization studies with a substrate containing a quinoline in an effort to improve reaction selectivities. To support reaction optimization, a high-throughput technique for the rapid measurement of ee for enantioisotopomers was required. Importantly, a technique that consumed a minimal amount of sample, had short analysis times, and provided reproducible and reliable data was necessary. Given that no techniques currently exist for the rapid measurement of ee for enantioisotopomers, we were motivated to develop a high-throughput chiral tag MRR measurement technique. In addition to reporting the first high-throughput technique for ee determination of enantioisotopomers, we also present a method by which the ee of a new enantioisotopomeric compound can be ascertained without the need for any spectroscopic analysis—a time-consuming step that requires some expertise in rotational spectroscopy (Scheme 1d). These two developments can allow MRR to be effectively utilized for this unique and analytically challenging class of molecules.

RESULTS AND DISCUSSION

Chiral Analysis Using MRR Chiral Tag Spectroscopy.

The principles of chiral tag rotational spectroscopy are briefly described. MRR spectroscopy detects transitions between the quantized energy levels associated with the rotational kinetic energy of a freely rotating molecule.⁶¹ One advantage of MRR spectroscopy is that the Hamiltonian for the quantized rotational kinetic energy of a distortable rotor, known as the Watson Hamiltonian,⁶² predicts all transition frequencies to experimental accuracy for most molecules. The key parameters in the Hamiltonian are the three rotational constants, A, B, and C, that are inversely related to the moments-of-inertia in the principal axis system of molecular rotation. Therefore, the spectroscopy parameters are directly tied to the geometry of the molecule through the distribution of the mass about the molecule’s center-of-mass. Each distinct molecular geometry—including each possible isomer of a single chemical species—has its own set of rotational constants that can be used for identification. One major strength of MRR spectroscopy for chemical analysis is that equilibrium geometries from quantum chemistry, which are used to calculate the rotational constants, are sufficiently accurate to make a high-confidence identification of a molecule without the need for certified reference standards.⁶³ Although the measurement itself is nondestructive, there is no attempt in these experiments to recapture the analyte, and so, the sample is consumed as it is evacuated from the measurement chamber by the vacuum pumps.

MRR instruments measure the spectrum in a method analogous to Fourier transform NMR spectroscopy.^64,65 The molecular sample is polarized using a short pulse of resonant light as the dipole moment of the molecule is aligned with the polarized electric field of the excitation pulse (as a result, MRR spectroscopy is only applicable to polar molecules). The excitation pulse creates time-dependent quantum states that subsequently emit light at the frequencies of allowed rotational transitions. The coherent emission, often called the free induction decay, is collected by the receiver and digitized. The time-domain signal is converted to a frequency-domain spectrum using fast Fourier transform processing. MRR spectrometer designs operate in either broadband^65,66 or narrowband modes.^64,67,68 Broadband spectrometers, using chirped-pulse excitation to achieve large excitation bandwidths, are used to record a large portion of the MRR spectrum for spectroscopic analysis, from which the spectral patterns for each species present can be fit to obtain the rotational constants. Narrowband instruments perform the measurement in a cavity resonator—a design pioneered by Balle and Flygare.⁶⁴ The resonator restricts the measurement to a single transition of the full rotational spectrum. However, the use of a high-quality factor (Q) cavity significantly enhances the detection sensitivity, which reduces measurement time and sample consumption. This design is well-suited to chemical monitoring applications^53,67 and is employed in this work for high-throughput ee measurements. Both instrument designs offer exceptionally high spectral resolution with the practical advantage that there is negligible overlap in spectral patterns of different species, even for complex chemical mixtures, so that the method does not require chromatography to separate the mixture prior to analysis.

Chiral molecules exist in two forms, called enantiomers, that are nonsuperimposable mirror images. A common way to designate the enantiomers is the Cahn-Ingold-Prelog method that labels the chiral centers as either (S) or (R) defined by ranking the priority of the groups attached to the chiral center. The physical properties of isolated enantiomers are identical. However, the enantiomers can behave differently when placed in a chiral environment—this chiral recognition of the environment can alter the behavior of the enantiomers of an active pharmaceutical ingredient in the body, for example. Since MRR spectroscopy performs measurements on isolated, freely rotating molecules, the rotational spectra of the enantiomers are identical. Chiral analysis can thus be performed in MRR spectroscopy using chiral derivatization—an approach also used for NMR spectroscopy.^69,70 In chiral derivatization, a single enantiomer of a new chiral molecule is attached to the analyte to provide a local chiral environment that can differentiate the analyte enantiomers. The addition of the new chiral center creates distinguishable molecules that are known as diastereomers. Using the Cahn-Ingold-Prelog labels, the tag-analyte diastereomers can be designated as homochiral when the two chiral centers have (S) or (R) chirality and heterochiral when the labels differ. In chiral tag MRR spectroscopy, the derivatization uses noncovalent interactions such as hydrogen bonds to attach the derivatizing agent.^{54–58,71–73} This is achieved by adding a small amount of the “tag” to the neon carrier gas used to introduce the sample into vacuum. Clusters between the analyte and tag are automatically generated during the pulsed jet expansion. This derivatization method requires no additional chemical preparation steps, and the noncovalent interactions will not cause racemization in the analyte that would compromise the measurement accuracy.

The chiral tag methodology involves two measurements. The spectrum of the complexes formed between the analyte and a racemic sample of the tag is acquired first. This spectrum is used to calibrate the instrument response as well as spectroscopic factors that determine the measured signal intensities. The second measurement uses a high-enantiopurity tag sample, and the relative signal intensities predominantly reflect the enantiomer composition of the analyte. The derivation of the formula used to determine the enantiomeric excess has one fundamental assumption: the number densities of the homochiral and heterochiral complexes are linearly proportional to the number densities of the tag and analyte in the pulsed jet expansion.⁵⁵ Under this assumption, the intensities for rotational transitions in the spectra of the homochiral and heterochiral tag and analyte complexes can be written:

$$I_{\mathit{\text{Homo }}}=C_{\mathit{\text{Homo }}}([(+)‐\text{Analyte}][(+)‐\text{Tag}]+[(-)‐\text{Analyte}][(-)‐\text{Tag}])$$ (1)

$$I_{\mathit{\text{Hetero }}}=C_{\mathit{\text{Hetero }}}([(-)‐\text{Analyte}][(+)‐\text{Tag}]+[(+)‐\text{Analyte}][(-)‐\text{Tag}])$$ (2)

The constants, C_Homo and C_Hetero, include the instrument response functions, spectroscopic factors that determine the transition intensity, and the populations of the homochiral and heterochiral complexes in the pulsed jet expansion. The ee determination uses the normalized transition intensities defined using the intensities in the spectra acquired using the racemic and enantiopure tags as

$$N_{\mathit{\text{Homo }}}=\frac{I_{\mathit{\text{Homo }}}^{\mathit{\text{Pure }}}}{I_{\mathit{\text{Homo }}}^{\mathit{\text{Rac }}}}$$ (3)

and similarly for the heterochiral spectrum intensities. Using the normalized intensities, the ratio R is defined

$$R=\frac{N_{\mathit{\text{Homo }}}}{N_{\mathit{\text{Hetero }}}}$$ (4)

This ratio is the equal to the enantiomer ratio of the analyte when the tag is exactly a single enantiomer. The ee of the analyte can be derived using

$$\frac{R-1}{R+1}=\left(e{e}_{\mathit{\text{analyte }}}\right)\left(e{e}_{\mathit{\text{tag }}}\right)$$ (5)

The tag ee is known in advance of the measurement from a separate analysis.

Application to the Enantiomers Generated by Precision Deuteration Reactions.

The chiral tag method is illustrated in Figures 1 and 2 for a set of deuterated compounds produced using the enantioselective copper-catalyzed alkene transfer hydrodeuteration reaction chemistry. There are three sets of analytes that are structurally similar, with one analyte in each set containing a nitrogen atom. The quantum chemistry structures of the chiral tag complexes formed between the deuterated reaction product and the enantiopure chiral tag, (S)-1,1,1-trifluoropropan-2-ol (trifluoroisopropanol, or TFIP), that give rise to the highest-intensity spectra in a broadband MRR measurement are shown in Figure 1. The quantum chemistry calculations are detailed in the Supporting Information and use the dispersion-corrected density functional method, B3LYP-GD3BJ/def2-TZVP, for the geometry optimization in Gaussian 16.⁷⁴ The spectroscopic analysis that supports the identification of these chiral tag complexes is also reported in the Supporting Information. The reaction chemistry adds a single deuterium at the benzylic carbon. Deuterium substitution at this position creates either the (S)-enantiomer (blue position) or (R)-enantiomer (red postion) of the analyte. When the chiral tag complex with (S)-TFIP is formed, the mass change associated with deuteration will make different and predictable changes to the moments-of-inertia. Spectroscopic analysis, therefore, makes it possible to associate a specific enantiomer of the reaction product with each observed rotational spectrum. That is, spectroscopic analysis supported by quantum chemistry calculations is required to assign the absolute configuration.⁷¹

Figure 1.

Structures for the chiral tag complexes formed between (S)-1,1,1-trifluoro-propan-2-ol and six reaction products of the copper-catalyzed alkene transfer hydrodeuteration reaction are shown. TFIP attaches to the nitrogen heterocycle species by a classical hydrogen bond formed between the electronegative nitrogen atom and the hydrogen atom of the alcohol in the tag molecule. For the hydrocarbon species where there is no classical hydrogen bond acceptor, the noncovalent attachment occurs to the aromatic ring. Dispersion-corrected density functional theory calculations correctly identify the lowest-energy geometries of the chiral tag complexes in both cases. In each case, the structure is the lowest-energy isomer of the chiral tag complex obtained from quantum chemistry calculations using the B3LYP GD3BJ def2TZVP method. MRR spectra corresponding to these geometries are observed in broadband MRR measurements. The main reaction product incorporates a single deuterium nucleus at one of the benzylic hydrogen positions. The prochiral positions shown in blue produce the (S)-enantiomer of the analyte, while the red positions give the (R)-enantiomer.

Figure 2.

Ee measurements of the deuterated reaction products of Figure 1 are illustrated. In each case, a single transition from the homochiral and heterochiral chiral tag complex MRR spectrum is shown. The measurement in black uses a racemic tag sample. For the transitions shown, this racemic tag gives signal intensities that are approximately equal for the diastereomeric complexes. In the red spectrum, high-enantiopurity (S)-TFIP is used. In this case, the transition intensities reflect the enantiomer composition. For the hydrocarbons on the left, the reaction product has a high enantiomeric excess of the (S)-enantiomer. A significantly lower ee is observed for nitrogen-containing reaction products. For the two reaction products at the bottom, a second sample preparation was performed under cold room conditions and increased the ee as shown by the blue spectra.

The enantiomeric excess of the deuterated reaction product is determined from the transition intensities in the rotational spectra of the homochiral and heterochiral chiral tag complexes. This determination can be made using a single transition from the full spectral signature of each complex. The enantiomeric excess determinations for the six compounds in Figure 1 are shown in Figure 2. For each reaction product, the intensity of a single transition from the broadband MRR spectrum of the homochiral and the heterochiral complex is shown. The spectrum in black uses the racemic TFIP tag sample to calibrate the instrument response. Because the homochiral and heterochiral complexes are isotopomers of each other, the response factors (C_Homo and C_Hetero from eqs 1 and 2) are nearly identical. The red spectrum shows the transition intensities when the measurement is performed with high-enantiopurity (S)-TFIP (ee = 0.993). Since the tag is nearly enantiopure, the ratio of the transitions in the (S)-TFIP tag measurement is a good approximation to the enantiomer ratio of the analyte. The analyte chirality, (S) or (R), is denoted in the figure. The sample ee, calculated using eq 5, is also given in the figure. This value is determined from the average using several transition pairs from the homochiral and heterochiral MRR spectra as described in the previous work.³⁴ These results illustrate the generality of the chiral tag MRR method. A single tag molecule can be used to analyze a wide range of molecules, and any attachment position of the tag generates distinguishable spectra because deuterium substitution at the two possible positions produces different mass distributions.

These results from broadband MRR spectrum measurements show that the high ee achieved for the hydrocarbon reaction products is not maintained when a nitrogen atom is present. The spectra for the structurally similar 4-ethylbiphenyl and 2-(4-ethyphenyl)pyridine show a second measurement in blue. In this case, the reaction was performed at 0 instead of 25 °C. In both cases, the ee increases for lower-temperature reaction conditions. This observation motivated a study of reaction conditions, involving changes to both temperature and solvent, to optimize the ee of 8-ethylquinoline-d1—the nitrogen-containing reaction product with the highest ee under room-temperature reaction conditions.

High-Throughput Enantiomeric Excess MRR Measurement Methodology.

The MRR analysis method using broadband rotational spectroscopy requires about 2 h and consumes about 100 mg of the enantioisotopomer sample. These parameters are not favorable for reaction condition screening studies or other applications where rapid ee monitoring is required. Also, the full MRR spectroscopy analysis, where spectra are assigned to specific complex geometries based on agreement of experimental and theoretical rotational constants, is time-consuming. In many applications, ee determinations have the highest importance, while assignment of the absolute configuration can be undertaken at a later time. Under conditions where only the ee is required and the sample is known to have high chemical purity, chiral tag MRR spectroscopy can be implemented in an efficient manner that bypasses the need for any spectroscopic analysis. This measurement process is illustrated for the analysis of 8-ethylquinoline-d1 and used to analyze a set of samples prepared under different reaction conditions.

The first step of the measurement requires identifying transitions in the MRR spectra of the homochiral and heterochiral complexes that can be used for the ee determination. This identification can be made using the variation in transition intensity in a broadband MRR measurement between the racemic and enantiopure tag measurements. The potential complication is that there are generally several isomers of the chiral tag complexes formed in the pulsed jet expansion. For the 8-ethylquinoline/TFIP system, two isomers dominate the isomer population. The spectra for the two complexes are shown in Figure 3, while the spectroscopy analysis details are presented in the Supporting Information. Unlike the case where full spectroscopic analysis is performed first, so that transitions for the homochiral and heterochiral complexes of the same parent structure are chosen for the ee analysis, a spectroscopy-free analysis approach has the potential to select transitions from different isomers.

Figure 3.

Broadband MRR spectrum of 8-ethylquinoline-d1 with (S)-TFIP is shown. The spectrum in the left panel shows the transition in the measurement that requires the presence of both the tag and analyte. The spectrum, in black, was the average of 65 000 signal acquisitions—a 40 min measurement consuming approximately 20 mg of sample. Below the experiment, calculated spectra for the homochiral (blue) and heterochiral (red) tag complexes shown in the right panel that use the fit rotational constants provided in the Supporting Information are shown. The observation of a higher-intensity spectrum for the heterochiral tag complex indicates that the (R)-enantiomer is in excess. A small frequency range of the spectrum is shown in the middle panel. Note that several transitions in the spectrum are not part of the spectral signature of the complexes of the lowest-energy complex. The same experimental spectrum is shown in (B); however, in this case, the calculated spectra correspond to the second highest-energy isomer of the chiral tag complex identified in quantum chemistry calculations. This second isomer accounts for the additional strong transitions as can be seen by a comparison of the middle panels of (A) and (B).

The ee analysis based on eqs 1 and 2 will still be valid in this case as long as the isomer populations are stable from measurement-to-measurement. This assumption is tested in Figure 4 where the histogram of ee determinations from pairs of homochiral and heterochiral transitions of the two isomers of the 8-ethylquinoline/TFIP complexes are compared. Within the measurement precision, the ee determination does not depend on the isomer chosen for either the homochiral or heterochiral transition used in eq 5. This behavior has been noted in another recent chiral tag measurement.⁵⁷

Figure 4.

Enantiomeric excess determinations using eq 5 and the broadband spectra shown in Figure 4 are shown. The analysis uses the 36 highest-intensity MRR transitions in each isomer spectrum. The ee is obtained from the mean value of the 36² set of transition pairs. The error estimate reported for the determinations is obtained from the standard error of the distribution. Panels (A) through (D) give very similar ee values, indicating that the isomer ratios are stable in the MRR spectrometer.

The implication of the result of Figure 4 is that the normalized transition intensity, eq 3, is the same for all transitions in the rotational spectra of isomers of the homochiral and heterochiral complexes. This result is validated in Figure 5. This measurement shows a histogram of normalized transition intensities for the 100 most intense spectroscopic transitions of the chiral tag complexes. This analysis first removes all observed MRR transitions for TFIP (available from a laboratory reference spectrum) and 8-ethylquinoline-d1 (available from an initial measurement performed when optimizing the sample conditions for MRR spectroscopy) so that only transitions related to the interaction of the tag and analyte are analyzed. The exceptionally high spectral resolution of MRR spectroscopy ensures that the process of cutting the transitions from the TFIP and 8-ethylquinoline monomer species does not simultaneously remove a significant number of transitions associated with the tag complexes. The normalized signal histogram in Figure 5 has two clear peaks corresponding to transitions of the homochiral and heterochiral complexes, regardless of their isomer geometry. When these transitions are used to determine the ee using eq 5, an analysis also shown in Figure 5, the result is the same as Figure 4. Pairs of homochiral and heterochiral transitions for ee determinations can be selected without the need for any spectroscopic analysis. However, it should be emphasized that without the spectral analysis, these transitions cannot be designated as homochiral or heterochiral, that is, the absolute configuration of the higher-abudnance enantiomer cannot be established.

Figure 5.

(A) The distribution of normalized transition intensities, eq 3, for the broadband MRR spectrum of Figure 3 is shown for the 100 highest-intensity transitions of the chiral tag complexes. (B) Using the clear separation of the normalized intensities observed in (A), an ee determination is made from all transition pairs between the two sets. Note that in this case it is not possible to determine which enantiomer is in excess, since no spectroscopic analysis is performed.

Once a homochiral-heterochiral transition pair is selected based on the normalized transition intensity in the broadband spectrum, ee determinations can be performed using a narrowband MRR spectrometer to reduce measurement time and sample consumption. The samples prepared under different reaction conditions were analyzed using the Bright-Spec isoMRR instrument. An example measurement is shown in Figure 6. Details of the measurement protocol are presented in the Supporting Information. The measurement requires measuring the intensities of two transitions corresponding to the different analyte enantiomers (at frequencies 6336.21 and 6341.35 MHz). The resonance frequency of the cavity is determined by the cavity length, so the mirror position is automatically changed in the instrument software between the two measurements.

Figure 6.

IsoMRR cavity-enhanced measurements are shown for one of the samples of 8-ethylquinoline-d1 prepared for the reaction conditions study (Table 1, Entry 1). The transition intensities for racemic tag sample are shown in (A) and are used to calibrate the instrument response as described in the Supporting Information. The transitions for the heterochiral and homochiral tag complexes using enantiopure tag (ee = 0.993) are shown in (B). The heterochiral measurement time is 40 s, while the homochiral measurement time is 160 s.

The calibration process measures the transition intensities using a racemic TFIP tag sample. These are used to calculate the normalized transition intensities, eq 3, which in turn are used to calculate the ee in eq 5. This calibration step makes it possible to determine the number of averages required to measure the two transitions to sufficient sensitivity to make the ee determination. In a Fourier transform MRR measurement, the noise is reduced by the square root of the number of signal acquisitions, so the ee sensitivity limit is adaptable by choosing the number of signal averages. After the calibration process is complete, the automated ee analysis proceeds using the enantiopure TFIP tag sample. In the 8-ethylquinoline-d1 measurements, three separate ee determinations are performed for each sample to characterize the measurement precision. In all cases, the precision is about ±1.5% ee—performance comparable to chiral gas chromatography mass spectrometry measurements we have performed on other chiral samples. In addition, the signal intensity of a MRR transition of the 8-ethylquinoline-d1 monomer is checked at the beginning and end of the measurement sequence to verify that the sample was not depleted during the chiral analysis. This measurement sequence was performed using 3 mg of the 8-ethylquinoline-d1 sample. The measurement cycle time, including system purging to prepare for the subsequent analysis, is 10 min—both significant reductions compared to the broadband MRR measurement methodology previously used for enantioisotopomer analysis.

The reaction conditions and ee determinations for 8-ethylquinoline-d1 samples are reported in Table 1. The high-throughput ee determination on the 40 °C sample in THF solvent, ee = 47.8%(1.5), is in good agreement with determination using broadband MRR spectroscopy, ee = 50%(1) (note that these two samples were prepared in different synthetic runs). In this case, varying the reaction conditions does not increase the ee, so other paths to improved enantiopurity need to be explored.

Table 1.

Enantiomeric Excess Determinations Using the IsoMRR Spectrometer^a

Entry	Solvent	Temperature	Yield (%)b	ee (%)c
1	THF	40 °C	82	47.8(1.5)
2	THF	rt	89	53.8(1.4)
3	THF	3 °C	77	48.0(1.5)
4	THF	−20 °	87	49.5(1.5)
5d	THF	−45 °C	41	48.2(1.5)
6	THF	−45 °C	45	49.6(1.5)
7	1,4-Dioxane	3 °C	86	40.2(1.7)

^aReactions were conducted on a 0.6 mmol scale.

^bYields are of the isolated product.

^cMeasurement uncertainties are the standard deviation from three replicate measurements.

^dProduct contained trace recovered starting material.

The enantioisotopomer analysis in this work is a particularly challenging problem in chiral analysis. A few techniques have been reported for analysis of isotopically chiral molecules but have limitations for high-speed analysis. Partial enantiomer separation by chiral chromatography has been reported but requires long measurement times (greater than 1 h), and it is unclear whether the approach is general.⁷⁵ Chiral derivatization for nuclear magnetic resonance (NMR) spectroscopy has been successful for the analysis of ethylbenzene-d1,⁷⁶ but chiral derivatization by covalent chemistry is a cumbersome process and often requires careful selection of the derivatization agent to achieve separation of the NMR resonances.⁷⁰ The measurement of the ²H NMR spectrum of the analyte in a chiral polymer has been reported for ee determinations and has been shown to be applicable to a range of enantioisotopomers containing polar heteroatom functionality.^77–80 However, the reported method requires a relatively large amount of sample (40 mg) and has long spectrum acquisition times. MRR spectroscopy also has a second implementation for chiral analysis called three-wave mixing rotational spectroscopy.^81–83 This method uses two excitation pulses to create a coherent emission signal that differentiates the enantiomers based on the signal phase (or sign) and, as a result, has similarities to other chiroptical methods like circular dichroism. Three-wave mixing rotational spectroscopy has been applied to enantioisotopomers to demonstrate that the enantiomers give enantiomer-dependent signals.⁸⁴ One limitation of this measurement approach is that a sample of known ee is required to calibrate the instrument response, and these are not typically available for enantioisotopomers. However, this second MRR measurement method may offer additional paths to developing high-throughput ee determinations that maintain the MRR spectroscopy advantage of high chemical selectivity that makes it possible to perform measurements in a complex reaction mixture.

EXPERIMENTAL SECTION

MRR Experiments.

Broadband MRR spectra were measured on the 2–8 GHz chirped-pulse Fourier transform microwave spectrometer at the University of Virginia. The sample injection uses the reservoir nozzle design from NIST with a 1 cm × 1 mm cylindrical channel nozzle. The reservoir nozzle assembly has temperature control, and the reservoir temperature is optimized using the monomer sample. The vapor pressure of the analyte in the reservoir is entrained in neon gas. The analyte and tag concentrations are typically 0.1% (by mole) in the carrier gas. The racemic and enantiopure trifluoroisopropanol (TFIP) samples were obtained from SynQuest Laboratories and used without additional purification.

The high-speed measurement of the enantiomeric excess of the 8-ethylquinoline-d1 prepared under different reaction conditions was made in a BrightSpec isoMRR instrument—a cavity-enhanced Fourier transform microwave spectrometer based on the Balle-Flygare instrument design and its NIST implementation. This instrument uses the NSIT reservoir nozzle design. The TFIP tag samples were obtained from SynQuest Laboratories. Additional MRR spectroscopy measurement and data analysis details are reported in the Supporting Information.

General Information.

The following chemicals were purchased from commercial vendors and were used as received: Cu(OAc)₂ (99.999% from Alfa Aesar); dimethoxy-(methyl)silane (TCI); ethanol-OD (Millipore Sigma); (R)-(−)-DTBM-SEGPHOS (TCI). Anhydrous tetrahydrofuran (THF) was purified by an MBRAUN solvent purification system (MB-SPS). Chloroform-d (CDCl₃) was stored over 3 Å molecular sieves. Thin-layer chromatography (TLC) was conducted with Silicycle silica gel 60 Å F254 precoated plates (0.25 mm) and visualized with UV and a KMnO₄ stain. Flash chromatography was performed using SiliaFlash P60, 40–60 mm (230–400 mesh), purchased from Silicycle. For reactions that required heating (transfer hydrodeuteration), a PolyBlock for 2-dram vials was used on top of a Heidolph heating/stir plate. ¹H NMR spectra were recorded on a Varian 300, 400, or 600 MHz spectrometer and are reported in ppm using deuterated solvent as an internal standard (CDCl₃ at 7.26 ppm). Data is reported as follows: s = singlet, d = doublet, t = triplet, q = quartet, p = pentet, m = multiplet, br = broad; coupling constant(s) are reported in Hz; integration. ¹³C NMR spectra were recorded on a Varian 76 or 101 MHz spectrometer and are reported in ppm using deuterated solvent as an internal standard (CDCl₃ at 77.16 ppm). ²H NMR spectra were recorded on a Varian 61 MHz spectrometer. Data for (S)-ethylbenzene-d1 and (S)-ethyl-naphthalene-d1 in Figure 2 was previously published.³⁴

General Procedure.

In a N₂ filled glovebox, (R)-(−)-DTBM-SEGPHOS (12.9 mg, 0.011 mmol, 0.022 equiv), Cu(OAc)₂ (50 μL of a 0.2 M solution in THF, 0.010 mmol, 0.02 equiv), and THF (0.200 mL) were added to an oven-dried 2-dram vial followed by dropwise addition of dimethoxy-(methyl)silane (246 μL, 2.00 mmol, 4.0 equiv). A color change from green/blue to orange was observed while stirring for 15 min at room temperature. In a separate oven-dried 1-dram vial was added the alkene substrate (0.50 mmol, 1.0 equiv), THF (0.250 mL), and ethanol-OD (76 μL, 1.30 mmol, 2.6 equiv). The solution in the 1-dram vial was added dropwise over 20 s to the 2-dram vial. The total volume of THF was calculated based on having a final reaction concentration of 1 M based on the alkene substrate. The 2-dram vial was capped with a red pressure relief cap, taken out of the glovebox, and stirred for 16–20 h at the desired temperature in a PolyBlock for 2-dram vials. Upon completion, diethyl ether (10 mL × 3) was added to the reaction vial and transferred to a 200 mL round-bottom flask to dry load and was purified by flash column chromatography. Deuterium incorporation for all reaction products was >97% at the desired benzylic position.

(S)-4-(Ethyl-1-d)-1,1′-biphenyl. Reactions were performed according to the general procedure, with reaction one occurring at room temperature and reaction two occurring at 5 °C. Flash column chromatography using 100% hexanes was used to yield the pure product as a white crystalline solid. The room-temperature reaction gave the final product in 95% yield with 93% ee, and the 3 °C reaction gave the final product in 93% yield with 97% ee. ¹H NMR (400 MHz, CDCl₃): δ 7.70 (d, J = 8.1 Hz, 2H), 7.63 (d, J = 8.2 Hz, 2H), 7.52 (t, J = 7.9 Hz, 2H), 7.46–7.40 (m, 1H), 7.38 (d, J = 8.1 Hz, 2H), 2.83–2.74 (m, 1.01H), 1.38 (d, J = 7.7 Hz, 3H). ²H NMR (61 MHz, CHCl₃): δ 2.79, (s, 0.99D). ¹³C NMR (101 MHz, CDCl₃): δ 143.45, 141.30, 138.73, 128.83, 128.41, 127.20, 127.13, 127.08, 28.30 (t, J = 19.3 Hz), 15.66. HRMS (EI⁺) m/z: [M]⁺ Calculated for C₁₄H₁₃D 183.1158; Found 183.1150.

(R)-8-(Ethyl-1-d)-quinoline. Reactions were performed according to the general procedure. See Table 1 for specific changes to the reaction conditions for each entry. Flash column chromatography using gradient elution with 98% hexanes and 2% ethyl acetate was performed to yield the pure product as a yellow oil.¹H NMR (400 MHz, CDCl₃): δ 8.96–8.93 (m, 1H), 8.15–8.11 (m, 1H), 7.66 (d, J = 8.0 Hz, 1H), 7.57 (d, J = 7.2 Hz, 1H), 7.51–7.45 (m, 1H), 7.41–7.36 (m, 1H), 3.35–3.25 (m, 1.02H), 1.39 (d, J = 7.7 Hz, 3H). ²H NMR (61 MHz, CHCl₃): δ 3.33 (s, 0.98D). ¹³C NMR (75 MHz, CDCl₃): δ 149.31, 146.83, 142.95, 136.43, 128.45, 128.00, 126.50, 125.89, 120.86, 24.37 (t, J = 19.5 Hz), 15.06. HRMS (EI⁺) m/z: [M]⁺ Calculated for C₁₁H₁₀DN 158.0954; Found 158.0945.

(S)-(4-(Ethyl-1-d)-phenyl)pyridine. Reactions were performed according to the general procedure, with reaction one occurring at room temperature and reaction two occurring at 5 °C. Flash column chromatography using gradient elution with 95% hexanes and 5% ethyl acetate was performed to yield the pure product as a clear colorless oil. The room-temperature reaction gave the final product in 67% yield with 11% ee, and the 3 °C reaction gave the final product in 89% yield with 35% ee. ¹H NMR (400 MHz, CDCl₃): δ 8.69 (dd, J = 4.7, 1.8 Hz, 1H), 7.97–7.91 (m, 2H), 7.75–7.67 (m, 2H), 7.35–7.29 (m, 2H), 7.23–7.15 (m, 1H), 2.74–2.65 (m, 1.01H), 1.29 (d, J =7.6 Hz, 3H). ²H NMR (61 MHz, CHCl₃): δ 2.71 (s, 0.99D). ¹³C NMR (75 MHz, CDCl₃): δ 157.55, 149.66, 145.32, 136.93, 136.73, 128.36, 126.93, 121.86, 120.35, 28.38 (t, J = 19.3 Hz), 15.54. HRMS (EI⁺) m/z: [M]⁺ Calculated for C₁₃H₁₂DN 184.1111; Found 184.1104.

CONCLUSIONS

Rapid determination of ee for enantioisotopomers, specifically compounds that are chiral by virtue of deuterium substitution, was achieved by performing chiral tagging experiments on an isoMRR spectrometer. This represents the first spectroscopic technique for high-throughput ee determination of enantioisotopomers. Typically, between 1 and 3 mg of sample is required for each measurement. Analyses generally take between 10 and 15 min and can be performed successively with minimal time required between sample injections. Additionally, we have shown that it is possible to perform the enantiomeric excess determination without assigning molecular spectra, enabling the ability to perform high-throughput reaction optimization without performing time-intensive spectroscopic analysis. We envision this spectroscopic technique being useful for supporting enantioselective reaction optimization in process chemistry for processes involving the synthesis of both isotopically labeled or nonisotopically labeled compounds. Ongoing studies into reaction design and chiral ligand design are being performed in our laboratories to achieve high ee in the enantioselective transfer hydrodeuteration reactions for alkenes containing N-heterocycles and other Lewis basic heteroatom functionality.