Limitations in Determining Enantiomeric Excess of Alcohols by 31 P-NMR of the Phosphonate Derivatives

O uso de derivados fosfonatos de dialquila diastereoisoméricos na determinação de excessos enantioméricos através da razão entre os sinais de P-NMR anisócronos dos isômeros meso e treo, foi bem sucedido para álcoois secundários (método de Feringa), porém revelou-se ineficiente para álcoois primários com dois ou mais grupos metilênicos entre as hidroxilas e os centros estereogênicos. Observou-se adicicionalmente que era importante obter os espectros acoplados e desacoplados para se certificar dos picos correspondentes aos fosfonatos de dialquila. Entretanto, esta metodologia merece ser aplicada sempre que a quantidade de álcool disponível e sua estrutura não forem fatores limitantes.


Introduction
Asymmetric processes and chemical conversion of chiral compounds require chromatographic or NMR methods for the determination of the enantiomeric excesses.The usual approach requires the availability of GC or HPLC chiral columns (chromatography).Chiral derivatizing or complexing agents like chiral lanthanides derivatives (NMR) are alternatively applied.Both NMR and chromatographic methods rely on the formation of diastereomeric derivatives or complexes to achieve the enantiomeric discrimination.The concept of using achiral derivatives to determine enantiomeric excess by chromatography and NMR was first introduced by Horeau 1 .Horeau's rationale was very simple, by derivatizing racemic alcohols with achiral derivatizing agents, he could detect either by chromatography or by 1 H-NMR the meso and threo diastereomers.He also observed that the ratio between them was proportional to the enantiomeric excess (Scheme 1).
This concept was adapted by Feringa 2 to 31 P-NMR experiments, using PCl 3 as the achiral derivatizing reagent.Therefore, racemic alcohols were converted into phosphonates in a fast and quantitative reaction, yielding stereoisomeric mixtures: an enantiomeric pair (R,R and S,S) and two meso compounds (R,S and S,R), in a 2:1:1 ratio (Scheme 2).These diastereomers were detected by 31 P-NMR, as three signals (0-10 ppm).Application of Horeau's equation 1 (p 2 = (K-1)/(K+1), where K = (peak area of threo / peak areas of meso 1 + meso 2 and p is the optical purity) allowed the calculation of the optical purity of enantiomerically enriched mixtures.This method provided an easy way to determine the enantiomeric excess making use of 31 P-NMR which has a lower detection sensitivity relative to 1 H (6.63 x 10 -2 ) but fewer signals to observe, thus simplifying the analyses.The method was extended to amines 3 and thiols 4 .
The main purpose of this work was to determine optical purity of alcohols 1-4 (Fig. 1) which were not discriminated with chiral GC and HPLC chiral columns.

Experimental
1 H, 13 C and 31 P-NMR spectra were obtained on a Varian Inova-500 spectrometer with standard pulse sequences operating at 499.885 MHz, 125.695MHz and 202.135MHz for 1 H, 13 C and 31 P, respectively.The chemical shifts are reported in ppm using TMS (0 ppm) as internal reference for 1 H and 13 C-NMR spectra and pulses of 45° for 1 H and 13 C and delay times of 1s and 2s respectively.For 31 P-NMR spectra, H 3 PO 4 (0 ppm) was used as external reference, 0.03 mol L -1 of PCl 3 per samples and 1 mL of CDCl 3 solution, spectral width 5250 Hz , acquisition time 0.5 s, delay of 2 s, 512 scans per spectrum, temperature 25 °C, pulse width 45°, WALTZ decoupling mode, The coupling constants (J) are in Hz.
HPLC chromatograms were obtained with a Hewlett Packard equipment, model 1090, serie II/M HPLC equipped with a photodiode array detector and a Waters Nova Pack ODS (4 µm, 150 mm X 3.9 mm i.d) column eluted with chiral mobile phase (β-cyclodextrin 15 mmol L -1 ) or a LiChroCART 250-4 ChiraDex GAMMA (5 µm, 250 mm X 4 mm i.d) chiral column eluted with H 2 O, at a flow rate of 0.4 mL min -1 and a column temperature of 30 °C.
Diastereomers obtained by derivatizing an alcohol enantiomeric mixture with phosphorus trichloride.
To a round bottom flask (3 L) equipped with a mechanical stirrer, thermometer and containing saccharose (75 g), zinc sulphate (0.2 g) and water (500 mL), fresh baker's yeast (135 g) was added.The mixture was stirred for two hours at room temperature before adding 2,2-dimethyl-1,3cyclohexanedione (2.4 g, 17.1 mmol).The reaction was further stirred for 3 days at 30 °C.Sodium chloride was added to the reaction mixture and continuously extracted with dichloromethane for 12 h.Solvent evaporation furnished the crude alcohol (4.1 g) and column chromatography eluted with hexane-diethyl acetate 12%, yielded compound (+)-1 (

General procedure for the preparation phosphonate derivatives of alcohols 1-4
To a stirred solution of alcohol (0.18 mmol) and pyridine (0.5 mmol) in 0.5 mL CDCl 3 was added 0.06 mmol of freshly distilled PCl 3 dissolved in 0.5 mL of CDCl 3 .After stirring for 10 min, the reaction mixture was transferred into a 5 mm NMR tube and then the spectra were recorded.31 P-NMR of the phosphonates are depicted in Table 1.

Results and Discussion
Following the global trend of working with enantiomerically enriched mixtures or pure enantiomers has brought a new problem to our research group, namely, enantiomeric discrimination.Chiral GC and HPLC columns and chiral eluents are not always efficient thus other methods of analysis to assess enantiomeric excess had to be investigated.We were particularly attracted by Feringa's method 2 , mainly due to its low cost.
Compound (+)-1 was obtained via enantioselective reduction using Saccharomyces cerevisiae but the enantiomeric excess could only be inferred by specific optical rotation compared to the literature 5 .The dialkylphosponate derivatives of the racemate showed three 31 P signals (Figs. 2b and 2c) in the proton decoupled NMR spectrum with K close to 1 (Table 1) indicating that chiral recognition was negligible during the coupling reaction, therefore the enantiomeric excess could be obtained via the ratio of the threo and meso diastereomers (deconvolution was applied to the Table 1. 31 P-NMR a data of phosphonates derivatives obtained from the reaction between the alcohols and PCl3.spectrum depicted in 2c for integrating purposes).The dialkylphosphonate derivative signals were further confirmed by proton coupled spectrum (Fig. 2a).The enantiomeric excess obtained by Feringa's method was 97.0%, assuming that meso 1/meso 2 ratio follows the same value observed for the racemic mixture (Table 1).This strategy was necessary because the meso 2 signal, in this particular example, is overlapped by the monoalkyl phosphonated derivative (Table 1, Figs. 2d and 2e).This result was compatible with the ee obtained by optical rotation comparison 5 .Therefore though less precise than a HPLC/UV or GC/FID method the ee was estimated without chiral reagents or columns.In an attempt to improve the precision of the ee determination and inspired in Horeau's pioneer work 1 , the crude phosponate derivatives of (±)-1 were analyzed by GC/FID and GC/MS.The analysis revealed that the mixture was too complex and the ee difficult to determine.
Application of the same methodology to alcohol (±)-2 furnished several 31 P signals in the NMR spectrum due to the presence of monoalkylphosphonate and dialkylphosphonate derivatives in the reaction mixture.The use of the expected phosphorus multiplicity for the dialkylphosphonate derivatives in a 1 H coupled spectrum allowed the recognition of the correct three anisochronous 31 P signals, which were integrated in the decoupled spectrum furnish-ing K close to 1 (Table 1).And we concluded that the method was valid for general application of the ee determination of (±)-2.
The versatility of the method had no apparent limit but primary alcohols have always been recognized as difficult substrates to be discriminated by chiral chromatography, but Feringa reported the successful application of his method to 2-phenyl-1-butanol 2 .Thus, primary alcohols (±)-3 and (±)-4 (Figs.3a and 3b), showing no separation in the available chiral columns, and possessing two and three methylene groups tethering the hydroxyl group and the stereogenic center, respectively, were selected for the final test.In both cases the phosphorus signals of the diastereomeric dialkylphosphonate derivatives were isochronous (Figs.3a and 3b).This ineffective discrimination pointed out that Feringa's method is mostly limited to compounds possessing hydroxyl groups at the chiral center.

Conclusions
The use of Feringa's method is recommended when: chiral chromatography fails; the alcohol is available in amounts ranging from 20 to 30 mg (otherwise more spectrometer time will be used) and finally the hydroxyl groups are close to the stereogenic center.Other stereogenic elements were not tested by Feringa neither by other group.

d
the enantiomeric excess (ee) was calculated from the integrated peak area of threo isomer (d,l pair) and meso isomers, respectively using the modified Horeau's formula 1 ee =[ (K-1)/(K+1)] 1/2 x 100 where K = (peak area of threo / peak areas of meso 1 + meso 2); e not available due to isochronous 31 P chemical shifts in the diastereomers.