Enantiomeric Discrimination by Double Quantum Excited Selective

Oct 11, 2007 - The experiment is robust, the pulse sequence is easy to implement, and the methodology has been demonstrated on different chiral molecu...
0 downloads 0 Views 135KB Size
Download PDF
J. Phys. Chem. B 2007, 111, 12403-12410

12403

Enantiomeric Discrimination by Double Quantum Excited Selective Refocusing (DQ-SERF) Experiment Bikash Baishya,†,‡ Uday Ramesh Prabhu,†,‡ and N. Suryaprakash*,‡ Solid State and Structural Chemistry Unit and NMR Research Centre, Indian Institute of Science, Bangalore 560 012, India ReceiVed: June 22, 2007; In Final Form: August 29, 2007

The differences in chemical shift anisotropies, dipolar couplings, and quadrupolar couplings of two enantiomers in the chiral liquid crystalline media are employed to visualize enantiomers. In spite of the fact that proton has high magnetic moment and is abundantly present in all the chiral molecules, 1H NMR is not exploited to its full potential because of severe overlap of unresolved transitions arising from long- and short-distance couplings. Furthermore, the two spectra from R and S enantiomers result in doubling of the number of observable transitions. The present study demonstrates the application of the selectively excited homonuclear double quantum (DQ) coherence correlated to its single quantum coherence of an isolated methyl group in a chiral molecule. The DQ dimension retains only the passive couplings within the protons of the methyl group while the long-distance passive couplings are refocused, removing the overlap of central transitions, and each enantiomer displays a doublet instead of a triplet unlike in regular selective refocusing experiment. The doublet separation being different for each enantiomer results in their discrimination. The cross section taken along the single quantum dimension pertaining to each transition in the DQ dimension provides the one-dimensional spectra for each individual enantiomer with the complete removal of the overlapped transitions from the other enantiomer. The experiment is robust, the pulse sequence is easy to implement, and the methodology has been demonstrated on different chiral molecules.

Introduction NMR spectroscopic visualization of enantiomers is extensively practiced using the weakly aligned chiral lyotropic liquid crystalline media.1,2 The difference in the orientational property of the enantiomers has been exploited for their visualization and to determine their excess. The difference in the elements of the Saupe order matrix between the enantiomers is small and is about 10-3 to 10-5.3 Its effect on the anisotropic NMR spectral parameters like chemical shift anisotropies (∆σi), dipolar couplings (Dij), and quadrupolar couplings (Qi) is significant as far as the enantiomeric discrimination is concerned. In the case of proton detection, the chemical shift difference between the enantiomers is not significant and the dipolar couplings have to be exploited. The strength of the dipolar couplings are not large unlike in the strongly orienting thermotropic liquid crystals, and the spin systems are weakly coupled. Thus, the first-order analyses of the spectra similar to that of the liquid-state spectra are generally possible. However, the unresolved transitions from several short- and long-distance dipolar couplings and the doubling of the spectra arising out of two enantiomers limit the routine use of 1H detection. When the number of interacting spins increases, there is excessive broadening which precludes the unraveling of the transitions for each enantiomer and their unambiguous discrimination. Therefore, the majority of NMR investigations of enantiomeric differentiation have focused on 2H NMR,4-10 taking advantage of the relatively large strength of the quadrupole interaction compared to chemical shift * To whom correspondence should be addressed. E-mail: nsp@ sif.iisc.ernet.in. † Solid State and Structural Chemistry Unit. ‡ NMR Research Centre.

anisotropies or dipolar couplings. There are also studies using the detection of 13C11,12 and 19F.13,14 In spite of severe overlap of the transitions, proton detection is always advantageous because of its (a) high magnetic moment, (b) high natural abundance, and (c) abundant presence in all the chiral organic molecules. Thus, there are continuous efforts to simplify the proton spectrum, namely, the selective excitation15 using the well-known selective refocusing (SERF) experiment16 used in liquid-state studies to determine the couplings, combining variable angle spinning with selective refocusing,17 twodimensional correlation,18 heteronuclear selective refocusing,19 and J-resolved experiment with bilinear rotation decoupling (BIRD) sequence.20 Each of these methods has its own advantages, aids the analysis of the spectrum, and results in the discrimination of the enantiomers. However, the lack of resolution and the severe overlap of transitions, more so in complex molecules, have restricted the routine use of proton NMR to its full potential. Thus, there is a dire need to develop experimental methodologies for the routine use of 1H NMR. In our recent study on heteronuclear spin systems, we demonstrated the selective detection of single quantum (SQ) coherence on the basis of the spin states of the heteronuclei using the highest order of homonuclear multiple quantum coherence.21 In the present study, we are demonstrating the application of a double quantum excited selective refocusing experiment (DQSERF) on chosen chiral molecules, each showing an isolated group of transitions arising from methyl protons. The DQ coherence reduces the triplet of the methyl group to a doublet in the DQ dimension and completely removes the central transition. The cross section taken along the SQ dimension pertaining to each transition in the DQ dimension provides the

10.1021/jp074873s CCC: $37.00 © 2007 American Chemical Society Published on Web 10/11/2007

12404 J. Phys. Chem. B, Vol. 111, No. 43, 2007

Baishya et al.

one-dimensional spectrum for each individual enantiomer with complete removal of the overlapped transitions from the other enantiomer. The selective refocusing 180° pulse in the middle of the DQ evolution retains only the short-distance couplings between the two active protons involved in the DQ coherence and the passive proton of the methyl group not involved in the double quantum coherence. The DQ dimension provides information on short-distance coupling among methyl protons in each enantiomer. Furthermore, the selective refocusing removes all the long-distance remote passive couplings. When the long-distance couplings are not of comparable magnitude to the short-distance couplings, the couplings to the remote protons can also be determined from the SQ dimension using this technique. In the SERF experiment on the isolated methyl protons, the refocusing of the long-distance couplings in the indirect dimension provides two triplets, one for each enantiomer when the enantiodiscrimination occurs. The central peaks of the triplets do not evolve during the refocusing period and are always overlapped. Thus, the chiral discrimination is not possible for the overlapped central transitions. The present DQ-SERF experiment which results in the removal of central transitions and the complete separation of enantiomer peaks serves as a better alternative to SERF experiment for the discrimination of the enantiomers. The experiment is robust and easy to implement. Multiple Quantum NMR. The multiple quantum (MQ) NMR and its applications have been discussed in detail in the reviews.22,23 Multiple quantum coherence is a coherent superposition of states for which the total spin magnetic quantum number is *1. Spins which undergo flipping are said to be active in the coherence, and all the remaining spins are said to be passive. The information content of the MQ coherences depends both on its order and on the active spins and their interactions with the passive spins. The group of active spins can be regarded as the super spin and the remaining passive spins can be regarded as the spectator spins. As the size of the super spin increases, the number of spectator spins decreases and consequently the multiplet pattern of the coherence becomes simpler. The number of allowed transitions for the zeroth and the m order quantum coherence of N nonequivalent interacting half integer spins is given by

1 Z0 ) [2NCN - 2N] 2 Zm )

2N! (N - m)!(N + m)!

(1) (2)

where Z0 refers to zero quantum (ZQ) and Zm refers to m quanta for m > 1. Thus, the highest quantum is for N ) m, and it pertains to a situation wherein all the spins are participating spins and there are no spectator spins. This corresponds to a situation where all the N spins flip simultaneously from the state |R〉 to the state |β〉 or vice versa. This implies that the scalar and the dipolar fields do not influence the spectrum. The highest quantum spectrum, however, provides information on the sum of all the chemical shifts. The different MQ orders of N interacting spins can be selectively excited24 or detected.25 N - 1 quantum is a situation in which N - 1 spins flip in the presence of one leftover spin. The super spin is then split by the scalar or the dipolar field of a spectator spin providing a doublet centered at the sum of the chemical shift positions of the active spins. The separation of the doublet provides the sum of the couplings of the active spins to the passive spin. For N

Figure 1. The pulse sequence used for the selective excitation of DQ coherence of methyl group in (R/S)-2-chloropropanoic acid, (R/S)-3butyn-2-ol, and (R/S)-propylene oxide. The pulse phase of Φ2 was x, y, -x, and -y with the receiver phase being x all the time. Phases of all the remaining pulses were set to x. Double quantum to single quantum coherence selection is achieved by setting the gradient ratio G1:G2 as 1:2. Seduce shaped pulse was used for all the experiments. The details of the pulse widths, delays, and phases are given in Table 1.

nonequivalent spins, there are N doublets in the N - 1 quantum spectrum. The N - 2 quantum spectrum is a situation where N - 2 spins flip at a time in the presence of the remaining two spins. This will have a greater number of transitions than N 1 quantum but significantly less compared to SQ transitions. The number of transitions and thereby the complexity of the spectrum increases by going to lower quantum. The N - 1 and N - 2 quantum spectra have sufficient number of transitions to provide all the spectral parameters. Furthermore, the sensitivity of the precessional frequency of any higher quantum coherence is proportional to its order. Thus, the field inhomogeneity contributes to the m quantum order by m times. However, the ZQ coherence is insensitive to the field inhomogeneity. Multiple Quantum-Single Quantum Correlation. The information on the MQ spectrum and the derivable information for different types of spin systems is available in the literature.26 In the schematic representation of the correlation of DQ coherence to its SQ coherence, with the nonselective detection pulse on a homonuclear scalar coupled AMX spin system,26 the DQ dimension provides three doublets with each doublet separation corresponding to the sum of the scalar couplings between each passive spin and the active spins centered at the sum of the chemical shift positions of the two active spins. Each doublet corresponds to the states |R〉 and |β〉 of the passive spin. The cross sections taken along the SQ dimension for each spin state of the passive spin give all 12 transitions expected for an AMX spin system, whose intensities depend on the flip angle. Thus, a nonselective detection pulse results in each |R〉 and |β〉 state of A, M, and X spins correlating to all the allowed transitions in the SQ dimension. The present study exploits the selective excitation of the double quantum coherence of an isolated methyl group not only to determine the spectral parameters but also for the separation of enantiomer peaks of the chiral molecules aligned in the chiral liquid crystalline media. Intensities of the Spectrum and the Determination of Enantiomeric Excess (ee) Using the MQ Sequence. The selective excitation of the double quantum coherence of the methyl protons is the essence of this work. All the twodimensional experiments were carried out using the DQ-SERF pulse sequence shown in Figure 1. It is a well-known pulse sequence employed for the multiple quantum studies.27 In the indirect dimension (t1 dimension), the DQ coherence is allowed to evolve, and at the end of the t1 period, the magnetization is converted back to SQ coherence for detection (t2 dimension).

Enantiomeric Discrimination

J. Phys. Chem. B, Vol. 111, No. 43, 2007 12405

All the pulses are selectively applied only on the isolated methyl proton resonances of the chosen molecules. It is well-known that MQ excitation is nonuniform.22 The amplitude of an excited coherence is dependent on the coupling constants. In the present case, though the nomenclature of the excited spin system is A3 for both the enantiomers, each enantiomer has different coupling constants. To excite DQ coherence in an A3 spin system, the delay, τ, between the first and the second 90° pulses should be less than 1/2(2THH), where 2T 2 HH (which corresponds to 3( DHH)) refers to the separation between the adjacent transitions of the triplet arising from proton that is two bonds away. However, the DQ excitation is maximum for 1/4(3 × 2DHH). This can be understood using the product operator formalism28 from the terms present just before the second 90° pulse of the pulse sequence shown in Figure 1, for an A3 spin system, as discussed below

Figure 2. The structure and numbering of the molecules (1) (R/S)-2chloropropanoic acid, (2) (R/S)-3-butyn-2-ol, and (3) (R/S)-propylene oxide.

-IY cos2 π(2THH)τ + 2IXSZ sin 2π(2THH)τ + 2IYSZRZ sin2 π(2THH)τ (3) However, the experimentally found optimum values generally do not agree with the theoretical one. One of the reasons for this is the long duration of 90° and 180° pulses. By suitably manipulating the power of radio frequency (rf) pulses, both 90° and 180° pulses were calibrated to 1.66 and 6.25 ms for the molecules investigated, and one has to consider the evolution of the magnetization during the pulses in the spin echo part of the selective DQ-SQ sequence. Furthermore, to derive optimum selectivity, only the central region of the spectrum was considered, implying there is also a definite offset dependence of the selective pulses used, particularly for (R/S)-3-butyn-2-ol (2) where the methyl region spread was 280 Hz and the excitation width was restricted to 40 Hz. Also, the racemic mixture was made of unequal amount of R and S enantiomers, and the delay, τ, had to be optimized to get the maximum signal intensity for both, which may be different from the maximum intensity of individual enantiomer. This is more so in (R/S)propylene oxide (3) where there is a severe overlap of transitions and the maximum intensity for each enantiomer cannot be monitored. Therefore, the intensities of the two-dimensional (2D) projections in our experiments are not comparable to that of normal 1D spectrum. This is due to several factors, such as (a) the DQ excitation is maximum for τ ) 1/4(3 × 2DHH) and there is a leakage of magnetization because of the first and third terms in eq 3 for this τ value, (b) the use of average τ delay, (c) the offset dependence of the shaped pulses used, and (d) the loss of magnetization because of zero quantum coherence. Therefore, the measurement of enantiomeric excess using the present method is tedious. However, there are other possible experiments to measure enantiomeric excess (ee) using uniform excitation of DQ coherence over a range of coupling constants, namely, repeating the experiment with several values of τ and co-adding the data;29-33 incrementing the delay, τ, with the evolution time t126 and thereby achieving τ averaging while 2D data set is being acquired; and using composite excitation and mixing periods designed to have a uniform effect over a range of coupling constants.34,35 Experimental Confirmation For the demonstration of the new experimental methodology, three different molecules (R/S)-2-chloropropanoic acid 1, 2, and 3, each having a chiral center, were chosen. The samples were purchased from Sigma and were used without further purification. The structures of these molecules are given in Figure 2.

Figure 3. Top trace: One-dimensional 1H spectrum of 1; middle trace: one-dimensional 1H spectrum of 2; and bottom trace: onedimensional 1H spectrum of 3. The expansions of the particular regions are plotted above the corresponding spectrum. The assignments of the protons have also been marked. The origin of the unmarked peaks is unknown. It could be either some impurity peaks or artifacts. Other details are given in the text.

The samples were prepared by using the method described in the literature.10,20 For the oriented sample 3, 78 mg of poly-γbenzyl-L-glutamate (PBLG) with DP 782 procured from Sigma, 42.5 mg of the racemic mixture, and 580 mg of CDCl3 were taken. For the oriented sample 2, 85 mg of PBLG, 59 mg of 2, and 450 mg of CDCl3 were taken. For the oriented sample 1, 50 mg of the racemic mixture, 80 mg of poly-γ-benzyl-Lglutamate, and 300 mg of CDCl3 were taken. The samples were sealed in a 5 mm NMR tube to avoid the evaporation of the solvent and then were centrifuged back and forth for several hours till the visually homogeneous phase was observed. The orientation of the sample was investigated by monitoring the 2H doublet separation of CDCl . The temperature was main3 tained at 303, 300.7, and 300 K for 3, 2, and 1, respectively, using Bruker BVT 3000 temperature controller unit of the DRX500 NMR spectrometer. The one-dimensional proton spectra of all the molecules, along with the assignment of peaks to different groups of transitions, are given in Figure 3. The assignment of peaks for each enantiomer R and S is normally carried out by first recording the spectrum of enantio pure sample and then by comparing it with the spectrum of a racemic mixture. The spectral parameters determined from the present study are in close agreement with the reported values.3,17,19,20 Thus, the assignments for R and S enantiomers were taken from the literature.

12406 J. Phys. Chem. B, Vol. 111, No. 43, 2007

Baishya et al.

TABLE 1: Acquisition and Processing Parameters Used for the Two-Dimensional DQ-SERF Experiments in 1, 2, and 3 Aligned in Chiral Lyotropic Liquid Crystal PBLGa molecules investigated and the experiments performed (R/S)-2-chloropropanoic acid DQ-SERF

(R/S)-3-butyn -2-ol DQ-SERF

(R/S)-propylene oxide DQ-SERF

parameter

F1 dimension

F2 dimension

F1 dimension

F2 dimension

F1 dimension

spectral width (Hz) number of points digital resolution (Hz) zero filling of data window function shaped pulse length (ms) optimized τ delay (ms) relaxation delay (s) number of accumulations

300 304 0.98 1k nil 1.66 for both π/2 and π 20 4 4

400 1228 0.32 2k nil

500 320 1.56 1k nil 6.25 for both π/2 and π 15.6 7 4

580 1228 0.38 2k nil

100 96 1.04 512 sine 6.25 for both π/2 and π 15.6 17 4

F2 dimension 100 400 0.25 2k sine

a All the SERF experiments for each molecule were carried out under identical experimental conditions of DQ-SERF. In 2 (where the methyl region spread is 280 Hz), the excitation width was restricted to 40 Hz in central region to avoid excitation of nearby peaks. To avoid the long duration of pulses, the 180° pulses were separately optimized with same duration as 90° pulses but with high power. All the spectra are displayed in magnitude mode without any linear prediction.

In 1, the methyl and the methine proton resonances are separated by more than 2 ppm. The methyl region shows a bandwidth of 140 Hz. Thus, the methyl protons of both the enantiomers could be simultaneously and selectively excited. However, the spectrum is a mixture of transitions from the R and S enantiomers. A series of two-dimensional experiments were, therefore, carried out to optimize the τ delay for maximizing the signal intensity of the methyl protons for both enantiomers. The intensity of the spectrum for the first t1 incremental delay corresponding to each enantiomer was plotted as a function of τ delay (graph not shown but given as the Supporting Information). The average delay of 20 ms was chosen as optimum to obtain maximum and comparable signal intensities for both the enantiomers and for all the lines. In 2, all the resonances are well isolated and the remote couplings of all the protons to other protons, except from the OH proton, are seen. The spectrum is an overlap of 12 transitions from each enantiomer, and the resulting 24 transitions are difficult to assign. The optimized τ delay that gave maximum intensities for all the lines was 15.6 ms. In 3, the methyl peaks for the two enantiomers are not resolved because of remote coupling assisted broadening and negligible difference in chemical shift anisotropy. Therefore, the peaks from both the enantiomers were indistinguishably overlapped. However, this did not preclude us from optimizing the τ delay. The delay of 15.6 ms gave maximum and comparable signal intensities for both the enantiomers. For all the molecules, the two-dimensional DQ-SERF and SERF experiments were carried out. The acquisition and processing parameters for all the molecules and for both types of experiments are summarized in Table 1. Results and Discussion In the molecules having a methyl group giving an isolated peak in 1H NMR, the selective DQ excitation of methyl protons has a distinct advantage. The DQ excitation corresponds to flipping of any two protons of the methyl group (active spins) in the presence of the third (passive spin) proton. Each cross section along SQ dimension displays a normal one-dimensional spectrum. An application of a selective refocusing 180° pulse on the methyl protons in the middle of the DQ dimension as shown in Figure 1 removes the effect of couplings from longdistance protons and retains only short-distance passive coupling to methyl proton not involved in DQ coherence resulting in a doublet. Because the sum of the two passive couplings from

TABLE 2: Basic Symmetry Functions for an A3 Spin System Showing the FZ Component for Each and Their Designation in the C3V Symmetry

number

spin functions

1 2 3 4 5 6 7 8

RRR (RRβ + RβR + βRR)/x 3 (RRβ + RβR - 2βRR)/ x6 (RRβ - RβR)/x2 (ββR + βRβ + Rββ)/x 3 (ββR + βRβ - 2Rββ)/x 6 (ββR - βRβ)/x2 βββ

designation of FZ component FZ component in C3V symmetry 3/2 1/2 1/2 1/2 -1/2 -1/2 -1/2 -3/2

(A1)3/2 (A1)1/2 E1/2 E1/2 (A1)-1/2 E-1/2 E-1/2 (A1)-3/2

short-distance proton is different in the two enantiomers, the doublets get separated from each other. Thus, there is a separation of all the peaks for each enantiomer with no single peak overlapped from the other enantiomer which aids in unambiguous discrimination. Total Separation via Reduced Multiplicity in DQ-SERF. The DQ dimension in a DQ-SERF experiment provides reduced multiplicity of transitions compared to SERF experiment. Flipping of the two spins of the methyl protons (A3 spin system) then mimic the spin system of the type AX, where A is the super spin and X is the spectator spin, in the DQ dimension. Exciting the DQ coherence and allowing it to evolve only under the sum of passive couplings to X spin (no chemical shift evolution) gives a doublet at frequencies +T and -T (where T ) 3 × 2DHH) as discussed below. The basic symmetry functions for the three equivalent nuclei in A3 spin system36 are given in Table 2. The dipolar coupling Hamiltonian that influence the spectrum of an A3 spin system can be written as37 3DHHI1ZI2Z + 3DHHI1ZI3Z + 3DHHI2ZI3Z. The DQ coherence evolves only under the sum of passive couplings and not under the sum of active coupling. The effective dipolar coupling Hamiltonian in the DQ dipolar resolved dimension for an A3 spin system for the DQ coherence of spins 1 and 2 can be written as 3DHHI1ZI3Z + 3DHHI2ZI3Z. The frequency for the two DQ transitions (calculated from Table 2) RRR f (ββR + βRβ + Rββ)/x3 and (RRβ + RβR + βRR)/x3 f βββ, that is, (A1)3/2 f (A1)-1/2 and (A1)1/2 f (A1)-3/2, are at - T and + T, where T ) 3(2DHH). Thus, the DQ dimension is a doublet with separation corresponding to 2T.

Enantiomeric Discrimination Because the value of T is different for the R and S enantiomers, the DQ transitions of R and S have different frequencies in DQ dimension spectra; schematically, one doublet is at +TR, -TR and the other doublet is at +TS, -TS enabling unambiguous chiral discrimination. As the last 90° pulse affects the spin state of the passive spin, there are three SQ transitions possible from each of the above DQ transitions. They are (A1)3/2 f (A1)1/2, (A1)1/2 f (A1)-1/2, and (A1)-1/2 f (A1)-3/2 with corresponding frequencies -T, 0, +T. Thus, the two triplets in the SQ dimension arise at the already separated DQ transitions. Since the separation of the enantiomer transitions have been achieved in the DQ dimension, the SQ transitions of the R and S forms appear in different cross sections in the SQ dimension. Thus, complete separation is achieved for the methyl group. As far as the SERF experiment is concerned, the three transitions for the dipolar coupled A3 spin system in the indirect dipolar resolved dimension can be calculated to be -T, 0, and +T. Since the central transition of the methyl peaks at the middle of the spectra does not evolve, it is always at zero frequency. The frequencies in the direct dimension appear at ′ΩA, ′ΩA + T, and ′ΩA -T. The absence of the 90° pulse at the end of the t1 dimension in SERF experiment unlike in DQSERF experiment indicates that the spin states involved in the SQ transitions in the indirect dimension remain unaffected in the direct dimension also. Hence, there is a diagonal tilt of the peaks (analogous to 2D J-resolved spectrum) in such a spin state selected spectra as shown in Figure 4B, Figure 5B, and Figure 6B. The different values of T for R and S enantiomers provide chiral discrimination in the indirect dimension for the outer transitions at +TR, -TR and +TS, -TS. However, the central transitions which have a zero frequency for both R and S forms are overlapped. In systems containing remote protons, the direct dimension is a multiplet because of the remote couplings for cross sections taken at frequencies +TR and -TR for the R enantiomer and +TS and -TS for the S enantiomer, and it is possible to pick up the transitions corresponding to each enantiomer at these cross sections. However, the multiplets arising out of the remote couplings for both the enantiomers are merged for the cross section along the SQ dimension taken at central transition frequency of the indirect dimension, and complete chiral discrimination is not possible in such a situation. The DQ-SQ correlation spectra reduces the triplet to a doublet and overcomes the problem of overlap. DQ-SERF Experiments. The DQ-SERF experiments of all the three molecules are discussed in this section, and the results are compared with SERF experiments in the subsequent section. The protons of 1 form a weakly coupled spin system of the type A3X. The first-order analysis of the spectrum is straightforward. It is evident from the top trace of Figure 3 that the resonance of the methyl protons is split into a 1:2:1 triplet because of residual proton dipolar couplings among the magnetically equivalent protons. The separation between two adjacent transitions provides 3(2DHH). Each of these transitions is further split into doublets of equal intensity because of coupling with methine proton. Similarly, the methine proton is a quartet because of its coupling with methyl protons. The spectrum is a mixture of transitions from the two enantiomers. This is a favorable system for the first-order analysis as the transitions are well-resolved. However, the prior requirement for the analysis is the identification or separation of the transitions belonging to each enantiomer. This is a challenging task.

J. Phys. Chem. B, Vol. 111, No. 43, 2007 12407

Figure 4. (A) 500 MHz 1H 2D DQ-SERF spectrum of 1 in the chiral liquid crystal PBLG correlating the DQ coherence of methyl protons to its SQ coherence along with the corresponding projections. All the details of the experiment are given in Table 1. Assignments of peaks for R and S enantiomers are taken from the literature. Peak separations which provide the values of 2THH (3 × 2DHH) and 3THH for R and S are marked. All the peaks of 2D projection in the direct dimension could be correlated to the peaks in R and S enantiomer cross sections. The discrete vertical lines marked identify the closely resonating signals for each enantiomer in the SQ dimension. The small internal doublet in the F1 projection centered at 0 Hz is possibly due to pulse imperfection or to some other experimental artifact, which is not clear at the moment. (B) 500 MHz 1H 2D SERF spectrum of 1 in the chiral liquid crystal PBLG. The pulse sequence was taken from the literature.17 Peaks corresponding to R and S enantiomers and the peak separations which provide the respective values of 2THH (3 × 2DHH) and 3THH are marked. The peak marked with an asterisk (*) in the F1 dimension is the overlap of two central transitions from both the enantiomers. The peaks in the central region of 2D projection in the direct dimension could not be correlated to the R and S as the central transition from R and S overlap in the indirect dimension shown in the enlarged plot in the box.

The 2D spectrum correlating the DQ coherence of the methyl protons to its SQ coherence is shown in Figure 4A along with the corresponding projections. The selective double quantum excitation of the methyl protons and the application of a selective 180° refocusing pulse in the middle of t1 dimension, which removes the passive coupling to methine proton, result in an AX spin system in the DQ dimension, where A corresponds to super spin (with two active methyl protons) and X is the passive methyl proton (nonparticipating spin in the double quantum). The spectrum observed in the DQ dimension is the A part of an AX spin system which shows a doublet because of the coupling with spin X. The separation of the doublet, which is 2(3 × 2DHH), for each enantiomer in the DQ dimension, therefore, corresponds to the separation of the outer lines of

12408 J. Phys. Chem. B, Vol. 111, No. 43, 2007

Figure 5. (A) 500 MHz 2D DQ-SERF spectrum of 3-butyn-2-ol in PBLG correlating the DQ coherence of methyl protons to its SQ coherence along with the corresponding projections. Peaks corresponding to S and R enantiomers and the peak separations which provide the values of 2THH (3 × 2DHH), 3THH, and 5THH are marked. All the peaks of 2D projection in the direct dimension could be correlated to the peaks of R and S enantiomer cross sections. The peak marked with an * in the F2 dimension corresponds to overlapped transitions from the two enantiomers. (B) 500 MHz SERF spectrum of 2 in PBLG. Peak separations which provide the values of 2THH (3 × 2DHH), 3THH, and 5THH for R and S are marked. The peaks in the central region of 2D projection in the direct dimension could not be correlated to the R and S as the central transition from R and S enantiomer overlap in the indirect dimension. The discrete vertical lines marked identify the positions of the transitions for S enantiomer in the SQ dimension. The peak marked with an * in the F1 dimension is the overlap of two central transitions from both the enantiomers. Other experimental details are discussed in the text.

the triplet of the SQ experiment. There are two doublets because of the enantiomers R and S. The cross section of the spectrum taken along the SQ dimension for each transition in the DQ dimension of an enantiomer corresponds to normal one-dimensional spectrum of the methyl group. It is evident from Figure 4A that there is a complete separation of the overlapped spectra. Each cross section provides a doublet of a triplet from which the dipolar couplings between the methyl protons (2DHH) and between methyl and methine protons (3THH, where the superscript 3 refers to the proton that is three bonds away) can be determined for both the enantiomers. The derived spectral parameters are given in Table 3. The one-dimensional 1H spectrum of 2 given in Figure 3 (middle trace) has well-resolved peaks for the CH3 and two CH groups. The protons of this molecule form a spin system of the

Baishya et al.

Figure 6. (A) 500 MHz 2D DQ-SERF spectrum of 3 in PBLG correlating the DQ coherence of methyl protons to its SQ coherence along with the corresponding projections. Peak separations which provide the values of 2THH (3 × 2DHH) and 3THH for R and S enantiomers are marked. All the peaks of 2D projection in the direct dimension could be correlated to the peaks in the R and S enantiomer cross sections. The discrete vertical lines marked identify the positions of the R enantiomer peaks in the SQ dimension. Other experimental details are given in the text. (B) The 500 MHz SERF spectrum of 3 in PBLG. Peak separations which provide the values of 2THH (3 × 2DHH) and 3T HH for the R and S enantiomers are marked. The peaks in the central region of 2D projection in the direct dimension could not be correlated to the R and S enantiomer as the central transition from R and S overlap in the indirect dimension. The discrete vertical lines marked identify the position of the R and S enantiomer peaks in the SQ dimension. Peaks marked with an * in the F1 dimension are the overlapped R and S transitions that cannot be discriminated. Other experimental details are given in the text.

type A3MX. The 2D spectrum correlating the double quantum coherence of the methyl protons to its SQ coherence is shown in Figure 5 A. As discussed earlier, the spectrum in the DQ dimension always provides 2(3 × 2DHH), and the cross sections of the spectrum along SQ dimension can be utilized to extract parameters, namely, 3 × 2DHH, 3THH (coupling between equivalent protons of methyl group and methine (-CH proton), and 5THH (where the superscript 5 refers to the proton that is five bonds away, i.e., between equivalent protons of the methyl group and the acetylinic proton). All the coupling parameters derived are given in Table 3. For the R enantiomer, the SQ dimension is well-resolved, and the magnitudes of 2DHH, 3THH, and 5THH can be determined from the analysis of the cross sections. However, 3THH could not be measured for the S enantiomer as the peaks are not resolved. The one-dimensional 1H spectrum of 3 in the chiral liquid crystal PBLG shown in Figure 3 (bottom trace) has wellseparated peaks for all the three groups of protons. The spin

Enantiomeric Discrimination

J. Phys. Chem. B, Vol. 111, No. 43, 2007 12409

TABLE 3: The Derived Coupling Parameters (Only Magnitudes Have Been Determined and Not the Signs of the Couplings) from the DQ-SERF and SERF Experiments in 1, 2, and 3 Aligned in Chiral Lyotropic Liquid Crystal PBLG molecules investigated and their parameter values in Hz (R/S)-2-chloropropanoic acid

(R/S)-3-butyn-2-ol

(R/S)-propylene oxide

parametera

DQ-SERF

SERF

DQ-SERF

SERF

DQ-SERF

SERF

(3 × 2DHH)R (3 × 2DHH)S (3THH)R (3THH)S (5THH)R (5THH)S

59.7 (b)b 110.3 (b) 15.6 (d) 25.1 (d)

59.5 (b) 109.3 (b) 15.6 (d) 25.3 (d)

103.4 (b) 44.2 (i) 49.7 (d)

103.6 (b) 44.35 (b) 49.4 (d) 32.8 (d) 9.1 (d) 9.3 (d)

12.3 (b) 4.9 (i)

12.3 (b) 4.9 (b) 6.0 (d) 7.0 (d)

9.1 (d) 9.3 (d)

5.6 (d)

a The errors on the parameters are of the order of digital resolution mentioned in Table 1. b b, d, and i refer to determinable parameters in both the dimensions, direct dimension, and indirect dimension, respectively.

system of the protons in 3 is of the type A3MNX where A, M, N, and X refer to methyl protons, two nonequivalent protons of methylene, and methine proton, respectively. The first-order analysis of this spectrum is straightforward. The methyl resonance is split into a 1:2:1 triplet because of dipolar couplings among them. The separation between two adjacent transitions provides 3 × 2DHH. Each of these transitions is further split into doublets of equal intensity because of coupling with methine proton (3THH). The coupling to long-distance methylene protons, 4T , results in further splitting of each of these transitions. HH This 4THH is very small and unresolved and contributes only to line broadening. Similarly, the methine proton is a quartet because of its coupling with methyl protons and is further split by methylene protons. The proton chemical shift difference between the two enantiomers is very small and is within the line width. Because of the existence of a large number of shortand long-distance dipolar couplings and also two sets of peaks arising from both R and S enantiomers, the transitions are indistinguishably overlapped. Thus, it is difficult to identify the peaks for each enantiomer from such a complicated onedimensional spectrum. The 2D spectrum correlating DQ coherence to its SQ coherence is given in Figure 6A. The spectrum, as discussed for the previous molecules, exhibits two doublets in the DQ dimension, from which 3 × 2DHH could be determined for both the enantiomers. The analyses of the cross sections of each enantiomer taken along the SQ dimension give all the couplings provided that the peaks are well-resolved. Thus, 2DHH and 3THH are available for the R enantiomer in the SQ cross sections. For S enantiomer because of poor resolution, this information could not be extracted in the SQ cross sections. Because of excessive broadening, 4THH could not be obtained for both the enantiomers. SERF Experiments. The significance of the DQ-SERF experiment is obvious when compared to the SERF experiment. The SERF spectrum of 1 is given in Figure 4B and is plotted under identical scale of DQ-SERF in both dimensions. The spectrum in the F1 dimension is the X part of the AXN type where N is the number of equivalent spins centered at zero chemical shift. As observed from the spectrum, the central transitions from both the enantiomers are always overlapped. The diagonal tilt of the doublets is visible from the spectrum as discussed earlier. This point has been ignored in earlier works.15,17 This situation appears in the weakly coupled spin systems in both the strongly and weakly oriented media. However, the advantage of this can be exploited in the present study as far as the determination of the couplings to longdistance protons is concerned. The cross section taken along the SQ dimension for each outer transition is a doublet with the separation 3THH. The derived parameters from the SERF experiment is also given in Table 3. The cross section of the central transition cannot, however, be utilized because of severe

overlap. This is obvious from the expansion of the cross section of the central peak given in Figure 4B. The SERF spectrum of 2 is given in Figure 5B. The cross section taken along the SQ dimension for each component of the triplet provides the doublet of a doublet from which (3THH)R/Sand (5THH)R/S could be obtained for each enantiomer. Again, the expansion of the central part of the spectrum clearly points out the overlap of the transitions from which the assignment of the peaks is difficult. The derived parameters are given in Table 3. The SERF experiment is useful in this case, especially for the determination of 3THH for the S enantiomer which could not be obtained from DQ-SERF experiment. The SERF spectrum of 3 is shown in Figure 6B. The analyses was carried out similar to the molecules discussed earlier. The derived coupling parameters are given in Table 3. In a study on this molecule using the variable angle spinning for visualization of the enantiomers,17 the authors claim that the resolution is enhanced by spinning the sample orienting at an angle 40° away from the magic angle, which also helps in selective excitation. Both SERF and phSERF experiments gave spurious peaks which authors claim arise from pulse imperfections. However, the resolution of the spectrum obtained by them by sample spinning is not better than our work which is carried out under static conditions. Unlike in spinning experiments, there are no spurious peaks in our experiments except in Figure 4A. Furthermore, with the observed well-isolated methyl peak with a spread of around 200 Hz, we could easily selectively excite methyl group without any demand on the resolution requirement. The similar experiment for the double quantum selective excitation of the heteronuclear coupling can also be implemented and exploited for the enhanced spectral dispersion. This will have a significant advantage compared to the heteronuclear selective refocusing reported in the literature.19 Progress has already been made in this direction, and the results will be published elsewhere. The work reported in the present study is on the systems wherein the protons of the methyl group are weakly coupled to remaining protons. The work on strongly coupled spin systems and the development of a phase-sensitive DQ-SERF experiment are other directions which require further investigations. Conclusions The use of chiral liquid crystal for enantiodiscrimination using selectively excited double quantum refocusing (DQ-SERF) experiment of methyl protons is demonstrated. We have also demonstrated that in a single SERF experiment both short- and long-distance couplings could be measured. There is no need of a biselective pulse to determine the remote couplings as reported in the earlier SERF experiment. There are distinct advantages of DQ-SERF experiment: (a) there is a complete

12410 J. Phys. Chem. B, Vol. 111, No. 43, 2007 separation of the overlapped spectra of methyl peaks for each enantiomer in the DQ dimension as well as SQ dimension in different cross sections, (b) the reduction of the triplet to doublet in the DQ dimension enhanced the resolution of the spectra of enantiomers and enabled unambiguous visualization, (c) the DQ dimension provided the direct measure of coupling between the selectively excited protons and the passive spin, that is, coupling among the methyl group protons, (d) the spectral resolution achieved in the DQ dimension is twice that of the earlier single quantum (SERF) experiment, (e) the cross section taken along SQ dimension at each transition in the DQ dimension provided all the coupling information, and (f) the experimental pulse sequence is simple, robust, and easy to implement and applicable to any molecule with isolated peaks from the methyl protons. As far as the disadvantage of the DQ-SERF is concerned, the quantitative measurement of enantiomeric excess is tedious. Acknowledgment. NS gratefully acknowledges the financial support by Department of Science and Technology, New Delhi for the grant no. SR/S1/PC-13/2004. Supporting Information Available: The intensity of the spectrum for the first t1 incremental delay corresponding to each enantiomer plotted as a function of τ delay. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Parker, D. Chem. ReV. 1991, 91, 1441-1457. (2) Rothchild, R. Enantiomer 2000, 5, 457-471. (3) Lesot, P.; Merlet, D.; Courtieu, J.; Emsley, J. W.; Rantala, T. P.; Jokisaari, J. J. Phys. Chem. A 1997, 101, 5719-5724. (4) Canet, I.; Courtieu, J.; Loewenstein, A.; Meddour, A.; Pe´chine´, J. M. J. Am. Chem. Soc. 1995, 117, 6520-6526. (5) Lafon, O.; Lesot, P.; Merlet, D.; Courtieu, J. J. Magn. Reson. 2004, 171, 135-142. (6) Solgadi, A.; Meddour, A.; Courtieu, J. Tetrahedron: Asymmetry 2004, 15, 1315-1318. (7) Merlet, D.; Sarfati, M.; Ancian, B.; Courtieu, J.; Lesot, P. Phys. Chem. Chem. Phys. 2000, 2, 2283-2290. (8) Lesot, P.; Merlet, D.; Loewenstein, A.; Courtieu, J. Tetrahedron: Asymmetry 1998, 9, 1871-1881. (9) Merlet, D.; Ancian, B.; Smadja, W.; Courtieu, J.; Lesot, P. Chem. Commun. 1998, 2301-2302. (10) Merlet, D.; Ancian, B.; Courtieu, J.; Lesot, P. J. Am. Chem. Soc. 1999, 121, 5249-5258.

Baishya et al. (11) Lesot, P.; Lafon, O.; Courtieu, J.; Berdague´, P. Chem. Eur. J. 2004, 10, 3741-3746. (12) Meddour, A.; Berdague, P.; Hedli, A.; Courtieu, J.; Lesot, P. J. Am. Chem. Soc. 1997, 119, 4502-4508. (13) Madiot, V.; Lesot, P.; Gre´e, D.; Courtieu, J.; Gree, R. Chem. Commun. 2000, 169-170. (14) Jakubcova, M.; Meddour, A.; Pe´chine´, J. M.; Baklouti, A.; Courtieu, J. J. Fluorine Chem. 1997, 86, 149-153. (15) Farjon, J.; Merlet, D.; Lesot, P.; Courtieu, J. J. Magn. Reson. 2002, 158, 169-172. (16) Fa¨cke, T.; Berger, S. J. Magn. Reson. A 1995, 113, 114-116. (17) Beguin, L.; Courtieu, J.; Ziani, L.; Merlet, D. Magn. Reson. Chem. 2006, 44, 1096-1101. (18) Lesot, P.; Merlet, D.; Courtieu, J.; Emsley, J. W. Liq. Cryst. 1996, 21, 427-435. (19) Farjon, J.; Baltaze, J. P.; Lesot, P.; Merlet, D.; Courtieu, J. Magn. Reson. Chem. 2004, 42, 594-599. (20) Ziani, L.; Courtieu, J.; Merlet, D. J. Magn. Reson. 2006, 183, 6067. (21) Bikash, B.; Suryaprakash, N. J. Phys. Chem. A 2007, 111, 52115217. (22) Norwood, T. J. Prog. Nucl. Magn. Reson. Spectrosc. 1992, 24, 295375. (23) Bodenhausen, G. Prog. Nucl. Magn. Reson. Spectrosc. 1981, 14, 137-173. (24) Sinton, S.; Pines, A. Chem. Phys. Lett. 1980, 76, 263-267. (25) Warren, W. S.; Pines, A. J. Am. Chem. Soc. 1981, 103, 16131617. (26) Braunschweiler, L.; Bodenhausen, G.; Ernst, R. R. Mol. Phys. 1983, 48, 535-560. (27) Bax, A.; Freeman, R.; Frenkiel, T. A.; Levitt, M. H. J. Magn. Reson. 1981, 43, 478-483. (28) Sørenson, O. W.; Eich, G. W.; Levitt, M. H.; Bodenhausen, G.; Ernst, R. R. Prog. Nucl. Magn. Reson. Spectrosc. 1983, 16, 163-192. (29) Weitekamp, D. P. AdV. Magn. Reson. 1983, 11, 111-274. (30) Weitekamp, D. P.; Garbow, J. R.; Pines, A. J. Magn. Reson. 1982, 46, 529-534. (31) Sørenson, O. W.; Levitt, M. H.; Ernst, R. R. J. Magn. Reson. 1983, 55, 104-113. (32) Rance, M.; Sørensen, O. W.; Leupin, W.; Kogler, H.; Wu¨thrich, K.; Ernst, R. R. J. Magn. Reson. 1985, 61, 67-80. (33) Murdoch, J. B.; Warren, W. S.; Weitekamp, D. P.; Pines, A. J. Magn. Reson. 1984, 60, 205-235. (34) Munowitz, M.; Pines, A. AdV. Chem. Phys. 1987, 66, 1-152. (35) Barbara, T. M.; Tycko, R.; Weitekamp, D. P. J. Magn. Reson. 1985, 62, 54-64. (36) Pople, J. A.; Schneider, W. G.; Bernstein, H. J. High Resolution Nuclear Magnetic Resonance; McGraw-Hill: New York, 1959. (37) Emsley, J. W.; Lindon, J. C. NMR Spectroscopy using liquid crystal solVents; Pergamon Press: Oxford, U.K., 1975.