Abstract

Differences between the solid-state 13C CP-MAS NMR spectra of holemic samples of the two enantiomers of 2,2′-dihydroxy-1,1′-binaphthyl (binol) were not sufficiently emphatic to reliably distinguish them, though they are readily distinguishable from the spectrum of the bimate of the compound crystallized from an equimatic sample. Inducing an additional chiral environment by cocondensation with sucrose as a chiral selector (CS) provided a method to yield differential spectra for the two enantiomers and thus effect enantiodifferentiation by way of solid-state NMR using weak interactions from a CS.

1. Introduction

One of the most important and challenging determinations in chemistry is the state of chirality (“enantiodifferentiation,” i.e., the identification of an enantiomer, the quantification of enantiomeric composition, and/or the determination of absolute configuration [1, 2]), ever more so with the increasingly stringent requirements placed upon the pharmaceutical industry for enantiomerically pure drugs. As part of our investigations [3] into chiral processes involving weak interactions, in particular the ESDAC [37] (enantiomer self-disproportionation on achiral chromatography) phenomenon stemming from our original observations [4, 8] of 9-hydroxy cineole (1) and its related metabolites (Figure 1) such as the dihydroxylated compounds 2,9- (2) and 2,10-dihydroxy-1,8-cineole (3) [9], we were greatly interested in new methods of enantiodifferentiation. 9-Hydroxy cineole (1) is an interesting compound because it readily undergoes SDE (self-disproportionation of enantiomers) [10], for example, by ESDAC [4] and also by sublimation [8], since it, like 2 and 3, is obtained as a scalemate from its natural source. A scalemate is a sample containing an unequal mixture of enantiomers—the preferred terminology and descriptors set out recently [11, 12] with respect to chiral systems are used herein. Thus, not only were we interested in absolute configuration determination but also quantification of the enantiomeric composition (ec), and for which we had to take care not to elicit perturbation of the ec from its original state.

Fortunately, however, many methods are available to accomplish enantiodifferentiation and even within NMR there are a number of broad approaches [1319] including chemical derivatization, chiral solvating agents (CSAs), and chiral complexation. In addition, chiral cavitands, either under isotropic conditions [17, 2024] or otherwise [25], have also been explored with respect to evaluation of the ec and/or identification of the predominant enantiomer. The most formidable advances within NMR of recent times in this context have involved the application of chiral liquid crystals (CLCs) [2638]—originating from the pioneering work [39, 40] of Snyder et al.—for determination of the ec or even for absolute configuration determination. Such are the dramatic differences in favorable cases that even (fluoro)hydrocarbon enantiomers [3133] and enantiomers chiral by virtue of isotopic substitution [36] have been successfully analyzed using this methodology. The great success of CLCs lies in the strong response by various order-dependent parameters to the small differences between the enantiomers in the chiral environment based on their orientation to the field; though differential orientation of enantiomers by other, field-independent media is also possible [41, 42]. However, despite all efforts, a general solution [13, 19, 27, 28, 43] to the central questions has not been found. A general solution is not only desirable in its own right but also to limit the shortcomings of the various present methodologies [1315, 28, 41, 42] which include technical and interpretative challenges. All of the previous work, however, suggests that if a general solution is to be found, then the solution will likely involve a compact state and/or aligned media. In this work, the aim is to examine a new approach to chiral analysis in the solid state by use of a chiral selector (CS), namely, sucrose, to effect the enantiodifferentiation of 2,2′-dihydroxy-1,1′-binaphthyl (4, binol) by means of 13C CP-MAS NMR (cross-polarization-magic angle spinning NMR).

2. Results and Discussion

One intriguing notion is whether enantiomers are able to display observable spectroscopic differences under isotropic conditions (in addition to parity violation effects [44]): in dilute solution (considering the typical concentrations used for solution-state NMR), the presumption is that the analyte molecules are well separated and effectively prevented from intimate interaction with each other by solvation. This is obviously not the case in the solid state but, thus far at least, there are no reports of such observations. Applying the same rationalization for an isolated pair of chiral molecules (considering them equivalent in energy) to a chiral molecule in a matrix of identical molecules, then the presumption would be that this system is equivalent to its mirror-image molecule also surrounded by identical molecules. Distinctions between holemates (i.e., enantiopure samples) and their equimates (i.e., samples consisting of 1 : 1 ratios of the two enantiomers) have been reported in nonisotropic phases [13, 18, 4547], and distinctions have even been made on the order-dependent parameters for cases where the holmate and the bimate (i.e., the “racemate”) exhibit identical isotropic chemical shifts and anisotropic values [48, 49]. These differences can even permit quantitation of the ec for scalemates in favorable cases in either of the aforementioned situations. However, these observations are attributed to the different space groups that these particular enantiomers and their corresponding equimates belong to [50, 51] in the former case, or in the latter case, due to the differing molecular symmetries and packings for the molecules in the holematic and bimatic states [48, 49]. These results have been suitably reviewed [50, 51]. Given the lack of differential observations between the spectra of two enantiomers in the confines of the solid state leads either to the conclusion that there is no discernible effect or that the same net effect is resultant for both enantiomers—and also the equimate in cases where the space groups are the same or the compound behaves as a homomate (conglomerate) or unimate (pseudoracemate). Alternatively, perturbation of the chiral environment in the solid state by the introduction of a CS could also lead to enantiodifferentiation. This approach does not seem to have been considered previously, though NMR examination of the enclathration in the solid phase of an enantiomer by a chiral medium has been reported [52] and reviewed [51]. Thus, whilst noncovalent interactions for the creation of a chiral environment have been useful for the analysis of enantiomers [1319, 28], for example, CSAs have been in use for some time now and have been subsequently followed by CLCs [51], the next logical step is chiral solid solutions (CSSs). Of note, in a weak chiral environment (induced by itself nonetheless), 1 readily displays chiral-state dependent differential behavior and this led us to the premise that perhaps not just in solution-state NMR, but also in solid-state NMR similar phenomena could be observed based on the analogous phenomena of SIDA (self-induced diastereomeric anisochronism) [3, 53] in the former method. Herein, enantiodifferentiation by an extremely weak external CS, namely, sucrose, as the provider of a chiral environment, is examined.

To explore the aforementioned premise, binol (4) was examined by solid-state 13C CP-MAS NMR. The two enantiomers of binol (R-4 and S -4, see Figure 2) each crystallizes in the P32 space group whilst the bimate (i.e., a “racemic compound,” bim-4) crystallizes in the Iba2 space group [54]. Four samples consisting of holemic samples of each of the two enantiomers (R-4 and S-4), an equimate sample as a mechanical mixture of the two (R-4 + S-4) prepared by the addition and thorough solid-phase mixing of equal amounts of R-4 and S-4 using a vortex mixer for 10–15 minutes, and an equimate sample prepared as the bimate (bim-4) by crystallization of an equimatic sample from chloroform were examined. Although solid-phase admixture of the enantiomers R-4 and S-4 has been reported [54] to rapidly yield the bimate, morphing was not apparent in this study upon mixing and the bimate was only formed after dissolution and removal of the solvent by evaporation. The results are portrayed in Figure 3 where the spectrum, as expected, of the bimate (bim-4) is clearly distinct from the spectra of the two enantiomers R-4 and S-4 and the mechanical mixture of the two (R-4 + S-4). However, although there are slight discernible differences between each of the enantiomers and the mechanical mixture in the figure shown, for example, the intensity of the pairs of signals at ca. 112 and 154 ppm and the signal at ca. 134 ppm, these apparent distinctions are blurred by the run-to-run inconsistencies of the acquired spectra and thus cannot by unequivocally attributed to the inherent differences between the two enantiomers R-4 and S-4 and the mechanical mixture R-4 + S-4. To confirm that the differences were not due to sample inconsistencies, the samples were also examined by solution-state NMR, whereby the spectra of R-4 and S-4 were found to be, to all intents and purposes, identical by 1H and 13C NMR.

To perturb the intrinsic chiral environment and thereby create discernible differences, an enhanced chiral environment was induced in the samples by the admixture of a CS, and for this purpose sucrose was used. Physical intimacy between the hydrophobic and hydrophilic compounds was ensured by the use of finely powdered samples; in the case of sucrose, consumer-grade crystals were ground in a mortar and pestle prior to admixture whilst the binol (4) was used as received. Following weighing and admixture, thorough dispersion and intimate contact between the crystals was ensured by several minutes of vortex mixing. Several samples were again prepared: the two enantiomers (R-4 and S-4, each ca. 1 : 1 with the CS and two further samples of R-4 in a 1.19 : 1 ratio and a 0.55 : 1 ratio), the equimate as a mechanical mixture (R-4 + S-4, ca. 1 : 1 with the CS), and the bimate (bim-4, ca. 1 : 1 with the CS). Again, although in particular comparisons discernible differences were apparent in the 13C solid-state NMR, the run-to-run inconsistencies of the acquired spectra again prevented any firm conviction of a distinction between either of the enantiomers (R-4 and S-4) and the mechanical mixture of the two (R-4 + S-4). Indeed, for all directly comparable spectra, reliable distinctions between the samples consisting solely of 4 (R-4, S-4, R-4 + S-4, and bim-4) and the admixture of each sample with sucrose were not forthcoming. Needless to say, variation of the enantiomer to sugar ratio failed to realize any significant perturbation.

To accentuate the perturbations of the chiral environment, each of the binol (4) admixtures with sucrose was taken up in DMSO and the resulting solutions then reduced to dryness under high vacuum. These cocondensed samples were then subsequently again examined by 13C CP-MAS NMR. For the enantiomer (R-4) with sucrose, new distinctions were not apparent above that of the run-to-run inconsistencies. However, most intriguingly, upon close examination of the S-enantiomer (S-4) with sucrose, the spectrum of the dried sample was distinct, but this was only evident after the course of several weeks (Figure 4). The main distinctions were for the pairs of signals at ca. 112 ppm and ca. 154 ppm wherein both pairs were substantially increased in intensity, similarly also for the signal at 134 ppm. The two peaks at 127 ppm were altered such that a reversal of signal intensity relative between the two occurred. The freshly dried sample also showed these distinctions, but the perturbations were only very slight and were intermediate between the physically mixed-only sample of S-enantiomer (S-4) with sucrose and the cocondensed aged sample. It is of considerable note that the differences displayed by the aged S-enantiomer (S-4) with sucrose sample appear to exceed the run-to-run inconsistencies and that these distinctions are thus attributable to the induced chiral environment provided by the CS. In principle, therefore, this method could provide a means for enantiomer identification or even determination of ec with appropriate data processing (e.g., deconvolution of signals). It must be noted though that cocondensation will form the bimate if the R-enantiomer (R-4) is present, so cocondensation is limited to homomates and unimates and not applicable to systems that display heteromatic behavior such as bimates—as is the case here. Moreover, this time-dependent change in the spectra was not apparent for the physically mixed-only sample of the S-enantiomer (S-4) with sucrose. Also, clearly distinct changes were not evident over the same course of time for any of the R-enantiomer (R-4) with sucrose samples. And finally, differences in the spectra of the CS sucrose were not in evidence.

It should be noted that although CP is not conducive to quantitative studies, in this work, we are not comparing atoms in different molecules nor are we comparing different atoms within the one molecule, but rather the same atoms under different chiral conditions. This has been shown to be a valid assumption [46]. Thus, by maintaining constant CP conditions across the samples, valid comparisons were ensured.

3. Conclusions

Thus, this as yet unexplored avenue of examination could potentially yield a general and easy-to-implement solution for chiral analysis. The obvious appeals of this approach are its simplicity, the ease of sample recovery (in this particular case, only liquid-liquid extraction), the circumvention of specific interactions, and the dispensing of any special technique for sample preparation or spectral interpretation. Thus, whilst the power of CLCs at present is unparalleled, the technical and interpretative challenges limit the chance of a general solution being easily found. With further work it is hoped that development of this approach could be advanced to the stage where enantiomeric identification and the quantification of enantiomeric excess can be accomplished, not only in specific cases, but as a more general solution.

4. Experimental

13C CP-MAS NMR spectra were acquired without lock at ambient temperature in a field of 9.4 T (100 MHz) using 4 mm rotors spun at 9 kHz with a contact time of 16 ms and postacquisition delay of 300 s. A long contact time was utilized to facilitate observation of the quaternary carbons via long-range couplings; thus, a compromise value was applied between the one-bond and multiple-bond couplings with this compromise value being determined in practice from a short series of measurements using optimized conditions to locate an appropriately sufficient value. The 90° pulse for 1H in the irradiation channel (3.90 μs) was found by acquisition of a series of 13C CP-MAS NMR spectra on a standard adamantane sample and locating the optimal value at a set power level. This same power level was used for the Hartman-Hahn condition where CP was effected by continuous wave irradiation of both the 1H and 13C nuclei. The power level for the Hartman-Hahn condition on 13C was found by varying the power level applied to the 13C nuclei for a standard adamantane sample at a set contact time (5 ms) and locating the optimal value for maximal signal response. Power levels amounted to −3.5 and 0.32 dB for the 1H and 13C nuclei, respectively, utilizing 100 and 300 Watt amplifiers, respectively. Typically, a spectrum was acquired for 12 h consisting of 146 scans. 1 k data points were collected for each FID (acquisition time = 6.4 ms) which were processed by zero-filling to 2 k and applying a line broadening of 100 Hz. Selected samples were also rerun to gauge run-to-run aberrations, either sequentially for the same sample insertion, at different sessions for the same sample, or by repacking the sample rotor with the same material. For comparative purposes, spectra were vertically scaled and shifted horizontally. Solution-state spectra were acquired as previously described [55, 56].