Enantiomeric and enantiotopic analysis of cone-shaped compounds with C3 and C3v symmetry using NMR spectroscopy in chiral anisotropic solvents

Article abstract of DOI:10.1021/ja020269d

We describe the enantiomeric and enantiotopic analysis of the NMR spectra of compounds derived from the functionalized cone-shaped core, cyclotriveratrylenes (CTV), dissolved in weakly oriented lyotropic chiral liquid crystals (CLCs) based on organic solutions of poly-γ-benzyl-L-glutamate. The CTV core lacks prostereogenic as well as stereogenic tetrahedral centers. However, depending on the pattern of substitution, chiral and achiral compounds with different symmetries can be obtained. Thus, symmetrically nonasubstituted CTVs (C₃ symmetry) are optically active and exhibit enantiomeric isomers, while symmetrically hexasubstituted (C_3v symmetry) derivatives are prochiral and possess enantiotopic elements. In the first part we use ²H and ¹³C NMR to study two nonasubstituted (-OH or -OCH₃) CTVs, where the ring methylenes are fully deuterated, and show for the first time that the observation of enantiomeric discrimination of chiral molecules with a 3-fold symmetry axis is possible in a CLC. It is argued that this discrimination reflects different orientational ordering of the M and P isomers, rather than specific chiral short-range solvent-solute interactions that may affect differently the magnetic parameters of the enantiomers or even their geometry. In the second part we present similar measurements on hexasubstituted CTV with flexible side groups (-OC(O)CH₃ and the, partially deuterated bidentate, -OCH₂CH₂O-), having on the average C_3v symmetry. No spectral discrimination of enantiotopic sites was detected for the -OC(O)CH₃ derivative. This is consistent with a recent theoretical work (J. Chem. Phys. 1999, 111, 6890) that indicates that in C_3v molecules no chiral discrimination between enantiotopic elements, based on ordering, is possible. In contrast, a clear splitting was observed in the ²H spectra of the enantiotopic deuterons of the side groups in the tri(dioxyethylene)-CTV. It is argued that this discrimination reflects different ordering characteristics of the various, rapidly (on the NMR time scale) interconverting conformers of this compound. Assuming two twisted structures for each of the dioxyethylene side groups, four different conformers are expected, comprising two sets of enantiomeric pairs with, respectively, C₃ and C₁ symmetries. Differential ordering and/or fractional population imbalance of these enantiomeric pairs leads to the observed spectral discrimination of sites in the side chains that on average form enantiotopic pairs.

Full text of DOI:10.1021/ja020269d

Published on Web 07/31/2002
Enantiomeric and Enantiotopic Analysis of Cone-Shaped
3
Compounds with C and C3v Symmetry Using NMR
Spectroscopy in Chiral Anisotropic Solvents
Philippe Lesot,*,† Denis Merlet, Muriel Sarfati, Jacques Courtieu,
†
†
†
‡
,§
Herbert Zimmermann, and Zeev Luz*
Contribution from the Laboratoire de Chimie Structurale Organique, CNRS ESA 8074, ICMO,
B aˆ t. 410, UniVersit e´ de Paris-Sud, 91405 Orsay, France, the Max-Planck-Institut f u¨ r
Medizinische Forschung, Abteilung Biophysik, Jahnstrasse 29, 69120 Heidelberg, Germany, and
the Weizmann Institute of Science, Department of Chemical Physics, 76100 RehoVot, Israel
Received February 20, 2002. Revised Manuscript Received May 13, 2002
Abstract: We describe the enantiomeric and enantiotopic analysis of the NMR spectra of compounds derived
from the functionalized cone-shaped core, cyclotriveratrylenes (CTV), dissolved in weakly oriented lyotropic
chiral liquid crystals (CLCs) based on organic solutions of poly-γ-benzyl-L-glutamate. The CTV core lacks
prostereogenic as well as stereogenic tetrahedral centers. However, depending on the pattern of substitution,
chiral and achiral compounds with different symmetries can be obtained. Thus, symmetrically nonasubstituted
3
CTVs (C symmetry) are optically active and exhibit enantiomeric isomers, while symmetrically hexasub-
stituted (C3v symmetry) derivatives are prochiral and possess enantiotopic elements. In the first part we
use H and 13C NMR to study two nonasubstituted (-OH or -OCH
2
3
) CTVs, where the ring methylenes are
fully deuterated, and show for the first time that the observation of enantiomeric discrimination of chiral
molecules with a 3-fold symmetry axis is possible in a CLC. It is argued that this discrimination reflects
different orientational ordering of the M and P isomers, rather than specific chiral short-range solvent-
solute interactions that may affect differently the magnetic parameters of the enantiomers or even their
geometry. In the second part we present similar measurements on hexasubstituted CTV with flexible side
groups (-OC(O)CH
3
and the, partially deuterated bidentate, -OCH
2
CH
2
O-), having on the average C3v
derivative.
symmetry. No spectral discrimination of enantiotopic sites was detected for the -OC(O)CH
3
This is consistent with a recent theoretical work (J. Chem. Phys. 1999, 111, 6890) that indicates that in C3v
molecules no chiral discrimination between enantiotopic elements, based on ordering, is possible. In contrast,
2
a clear splitting was observed in the H spectra of the enantiotopic deuterons of the side groups in the
tri(dioxyethylene)-CTV. It is argued that this discrimination reflects different ordering characteristics of the
various, rapidly (on the NMR time scale) interconverting conformers of this compound. Assuming two twisted
structures for each of the dioxyethylene side groups, four different conformers are expected, comprising
two sets of enantiomeric pairs with, respectively, C and C symmetries. Differential ordering and/or fractional
3 1
population imbalance of these enantiomeric pairs leads to the observed spectral discrimination of sites in
the side chains that on average form enantiotopic pairs.
Introduction
Functionalized cyclotriveratrylenes (CTV), also referred to
as tribenzocyclononene (TBCN), can exist in two geometrical
isomers, crown and saddle, as shown in Figure 1.1-7 The saddle
form is highly flexible and, in solution, it undergoes fast
pseudorotation between its various conformations, leading to
an average nonchiral species. The crown form, on the other
5
hand, is rigid and, depending on the mode of substitution, can
exhibit C1, Cs, C3, or C3V symmetry. From a stereochemical point
of view, the latter two symmetries are of interest with regard to
their chiral (or prochiral) properties. Thus, although the crown
core lacks prostereogenic or stereogenic tetrahedral centers,
symmetrically nonasubstituted CTVs (C3 symmetry) are chiral,
exhibiting enantiomeric isomers (see compounds 1 and 2 in
Figure 2). According to their stereochemical helicity, these are
*
Corresponding authors. P.L.: E-mail, philesot@icmo.u-psud.fr; Fax,
3
9
3 (0)1-69-15-47-70. Z.L.: E-mail, Ciluz@wisemail.weizmann.ac.il; Fax,
72-8-934-4123.
†
University of Paris-Sud.
The Max Planck Institute.
The Weizmann Institute.
‡
§
(
1) Poupko, R.; Luz, Z.; Spielberg, N.; Zimmermann, H. J. Am. Chem. Soc.
1
989, 111, 6094.
(
(
2) Zimmermann, H.; Poupko, R.; Luz, Z.; Billard, J. Z. Naturforsch 1985,
4
0a, 149.
(6) Collet, A.; Gabard, J.; Jacques, J.; Cesario, M.; Guilhem, J.; Pascard, C. J.
Chem. Soc., Perkin Trans. 1981, 1, 1630.
(7) Collet, A In ComprehensiVe Supramolecular Chemistry; Atwood, J. L.,
Davies, J. E. D., V o¨ gtle, F., Eds.; Pergamon: New York, 1996; Vol. 6, p
281.
3) Zimmermann, H.; Bader, V.; Poupko, R.; Wachtel, E. J.; Luz, Z. Manuscript
in preparation, 2002.
4) Collet, A.; Gabard, J. J. Org. Chem. 1980, 45, 5400.
5) Collet, A. Tetrahedron 1987, 43, 5725.
(
(
10.1021/ja020269d CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 10071-10082
9
10071

A R T I C L E S
Lesot et al.
chiral, short-range solute-solvent interactions that may affect
differently the magnetic parameters (for example, differential
electronic shielding) or even the geometry of the chiral centers
(
vibrations). In practice, however, except for very special cases,8
the spectral resolution is often too low to exhibit separated
signals for the different enantiomers. An alternative approach
9
-11
is to employ a chiral liquid crystal as a solvent.
In such
solutions, in addition to (the negligible) specific solute-solvent
interactions, orientational ordering of the solute also contributes
to the chiral discrimination. As it turns out, in many situations
it is the dominant discrimination mechanism providing the
doubling of peaks, where the conventional approach of using
isotropic solvents failed. In the present work, we apply this
method by using as solvent a lyotropic liquid crystalline solution
of poly-γ-benzyl-L-glutamate (PBLG) in organic solvents, such
as chloroform (CHCl3) or dimethylformamide (DMF).1 Under
proper experimental conditions of concentration and tempera-
ture, such solutions form a cholesteric phase with a relatively
long pitch. Under the effect of the strong magnetic field (Bo)
of the NMR spectrometer, the cholesteric pitch in these solutions
unwinds, yielding well-aligned chiral-nematic phases with the
Figure 1. Structure of substituted cyclotriveratrylene (CTV) derivatives
in the saddle (a) and in the crown conformation (b).
0,11
1
2
director parallel to Bo. For comparative purposes, similar
solutions containing racemic (compensated) mixtures of PBLG
and PBDG (poly-γ-benzyl-D-glutamate), which are nematic but
not chiral, have also been used. In the following we denote such
solutions as PBG. In these solutions the solute molecules
exchange rapidly between the PBLG and PBDG vicinities,
resulting in identical average magnetic interactions for the two
1
1,13
enantiomers.
Solutes dissolved in PBLG or PBG solutions become aligned
and, although this alignment is usually rather small,1
0,11
they
often exhibit characteristic NMR features of partially ordered
systems. These features are the extra splittings or shifts due to
anisotropic interactions, such as dipolar couplings, quadrupolar
couplings (for spin I > 1/2), and chemical shift anisotropies
1
4
(
CSA). For our purpose, carbon-13 and deuterium NMR are
1
0
13
1
particularly useful in this respect. The C- H dipolar
interaction over a single bond is on the order of 30 kHz, the
deuterium quadrupole coupling in a C-D bond is about 180
Figure 2. Structure of four CTV derivatives investigated in this work: (a)
nonahydroxy- and nonamethoxy-CTV (1 and 2; the two enantiomeric
structure are shown), (b) hexaacetyloxy-CTV (3) with average C3V symmetry,
and (c) tri(dioxyethylene)-CTV (4) with average C3V symmetry.
13
kHz, while the C CSA for aromatic carbons is about 100 ppm
(
translated to ∼10 kHz in commonly used spectrometers). Thus,
even if the orientational order parameter of the solutes is only
-
2
-4
about 10 to 10 , extra splitting and shifts (compared with
NMR in isotropic solvent) may be observed in their high-
resolution NMR spectra. The special chiral discrimination power
of CLC, such as the PBLG solutions, results from selective
effects on the ordering parameter of chiral solutes. Due to the
large effect of orientational ordering on the NMR spectra
labeled using the M or P descriptors.5,6 On the other hand,
symmetrically hexasubstituted CTVs (C3V symmetry; see com-
pound 3 in Figure 2) are prochiral and possess enantiotopic
elements. According to IUPAC nomenclature, enantiotopic
elements are “constitutionally identical atoms or groups in
molecules, which are related by symmetry elements of the
second kind only (mirror plane, inversion center, or rotation-
reflection axis)”. Such elements can in principle be discriminated
by spectroscopic methods, such as NMR, in a chiral environ-
ment. These two classes of CTV derivatives of the crown form,
therefore, merit special attention, and in the present work we
investigate the NMR spectra of some selected nona- and
hexasubstituted derivatives of this core in chiral liquid crystalline
solvents. Such studies are possible because the crown and saddle
forms of CTV are long-lived and do not interconvert in solution,
15
(
compared to that induced by specific solute-polymer interac-
(
8) (a) Parker, D. Chem. ReV. 1991, 91, 1441. (b) Rothchild R. Enantiomer
2000, 5, 457.
(9) (a) Sackmann, E. E.; Meiboom, S.; Snyder, L. C. J. Am. Chem. Soc. 1968,
90, 2183. (b) Tracey A. S.; Diehl, P. FEBS Lett. 1975, 59. (c) Courtieu, J.;
Bayle, J. P.; Lafontaine, E. J. Am. Chem. Soc. 1989, 111, 8294.
10) Sarfati, M.; Lesot, P.; Merlet, D.; Courtieu, J. J. Chem. Commun. 2000,
(
2069, and references therein.
(
11) Aroulanda, C.; Sarfati, M.; Courtieu, J.; Lesot, P. Enantiomer 2001, 6, 281.
12) Czarniecka, K.; Samulski, E. T. Mol. Cryst. Liq. Cryst. 1981, 63, 205.
(
(13) Canlet, C.; Merlet, D.; Lesot, P.; Meddour, A.; Loewenstein, A.; Courtieu,
J. Tetrahedron: Asymmetry 2000, 11, 1911.
(
14) Emsley, J. W.; Lindon J. C. In NMR Spectroscopy Using Liquid Crystal
SolVents; Pergamon Press: Oxford, 1975.
3
,5
even on the time scale of many months at room temperature.
(
15) (a) Lesot, P.; Gounelle, Y.; Merlet, D.; Loewenstein, A.; Courtieu, J. J.
Phys. Chem. 1995, 99, 14871. (b) Lesot, P.; Merlet, D.; Rantala, T. P.;
Jokisaari, J.; Emsley, J. W.; Courtieu, J. Phys. Chem. A 1997, 101, 5719.
In principle, any isotropic chiral solvent could lead to NMR
spectral discrimination of enantiomeric solutes through specific
10072 J. AM. CHEM. SOC.
9
VOL. 124, NO. 34, 2002

NMR Analysis of Cone-Shaped Compounds
A R T I C L E S
tions), even small differences in the ordering of corresponding
chiral elements may result in sufficient spectral discrimination
to be observed in particular using deuterium NMR. This is
manifested in doubling (or partial doubling) of the NMR spectra
of the chiral solutes in comparison with those recorded in a
corresponding racemic PBG solution.
The discrimination effect can be divided into two categories.
The first one concerns optically active compounds. In this case
the two enantiomers will in general have slightly different
orientational order parameters resulting, in favorable situations,
in doubling of the NMR spectrum due to the two optical isomers.
In the first part of the Results and Discussion, we examine the
case of nonasubstituted CTV derivatives (C3 symmetry) with
side chains R ) OH and OCH3 (compounds 1 and 2 in Figure
axes frame of the ordering tensor.21 Thus, enantiotopic sites in
such molecules will, in general, exhibit resolved NMR signals
in CLC solutions, purely by ordering, even in the absence of
specific chiral, short-range solute-solvent interactions. Similar
situations apply to molecules with Cs, S4, or D2d symmetry.2
In contrast, in molecules with C3V symmetry (or other point
groups with C3 or higher order symmetry axes) the ordering
tensor is axially symmetric with the unique principal direction
(z) parallel to the molecular C3 (or Cn, n > 3) axis. Thus, even
though the effective symmetry of such a molecule, when
dissolved in a CLC, reduces to C3 (or Cn), the orientational
ordering tensor remains axially symmetric with z still parallel
to the unique axis. Consequently, on the basis of ordering
considerations alone, enantiotopic sites in such molecules will
not become nonequivalent in CLC solutions. Regarding dis-
crimination due to specific chiral, short-range solute-solvent
interactions as unlikely, we therefore expect no discrimination
in the NMR spectra of such sites in chiral liquid crystalline
solutions. In the second part of the Results and Discussion we
test this conclusion by studying two examples of hexasubstituted
CTV derivatives, i.e. hexaacetyloxy-CTV (compound 3) and
tri(dioxyethylene)-CTV (compounds 4, 10% deuterated in the
ethylene groups). In both compounds the side chains are highly
flexible, rapidly (NMR wise) interconverting between different
conformations, leading to an average C3V symmetry. Indeed,
no discrimination was found between the enantiotopic sites of
compound 3 in the PBLG solutions. However, somewhat
unexpectedly, well-resolved signals were observed for the
enantiotopic ethylene deuterons in the dioxyethylene side groups
of compound 4. We argue that this discrimination reflects
selective ordering of the various (chiral) conformers in the
oriented PBLG system.
0-23
2). Both compounds were deuterated in the ring methylenes and
as spectroscopy tools we use proton-decoupled deuterium and
2
1
13
1
10,16,17
carbon-13 NMR ( H-{ H} and C-{ H}).
The NMR
spectra of the M and P isomers in PBLG/DMF solutions are
indeed well-resolved, while no such discrimination is observed
in chiral isotropic solvents. We argue that this chiral discrimina-
tion is due to selective ordering rather than specific, short-range
chiral solute-solvent interactions such those that can differently
affect the magnetic parameters of two enantiomers in chiral
isotropic solvents.
The second category of chiral discrimination relates to
18,19
prochiral molecules.
In this case the stereochemical dis-
crimination concerns enantiotopic elements and is brought about
by reducing the symmetry of the orientational distribution
function (eliminating symmetry elements of the second kind)
2
0
compared to their symmetries in achiral solvents. This
reduction in symmetry may, depending on the original molecular
symmetry, partially lift the restrictions on the orientation of the
principal coordinate system of the ordering tensors in the
molecular frame. This, in turn, will cause enantiotopic elements
to become nonequivalent in the CLC solution, resulting in
doubling of their NMR spectra. In a recent paper a compre-
hensive theoretical analysis of this effect was given, including
a full classification of all molecular point groups into those
Experimental Section
Synthesis. The synthesis of isotopically normal nonahydroxy-,
nonamethoxy- and hexaacetyloxycyclotriveratrylenes (1, 2, and 3) was
as described earlier.2 Compounds 1 and 2 deuterated in the crown-
ring methylenes were prepared as reported in ref 1. Compound 4 was
prepared both in the normal form as well as statistically deuterated
,3
20
whose symmetries are, or are not, affected by a uniaxial CLC.
(
∼10%) in the dioxyethylene side groups. The latter was prepared
Thus, for example, the principal directions of the ordering tensor
in molecules with C2V symmetry are fixed to lie along the C2
according to the five-step reaction shown in Scheme 1. The normal
compound was obtained by trimerization of normal 1,4-benzodioxane-
6-methanol, which was prepared by reduction of the commercially
available 1,4-benzodioxane-6-carboxaldehyde (Aldrich). In the follow-
ing we describe the synthetic steps leading to the deuterated compound
(z) axis and perpendicular to the two mirror planes (x, y). The
effective symmetry of such molecules in a CLC reduces to C2
and only the z-direction of the ordering tensor remains fixed
4
.
(parallel to the C2 axis), while the x and y axes are undetermined
1
,2-Ethanediol (∼10% Deuterated). A solution of 65 g of diethyl-
oxalate (Aldrich) in 100 mL of dry ether was slowly added to a slurry
of LiAlH (18 g) and LiAlD (2 g) in 1300 mL of ether. The mixture
was refluxed for 3 h, followed by hydrolysis with water containing
0% deuterium and saturated with Na SO . The slurry was filtered,
and the residue was boiled with 300 mL of THF and filtered. This
process was repeated twice. The filter cake was then treated with H
SO (2 N) for complete hydrolysis, and then extracted again with THF.
by any symmetry consideration. Pairs of enantiotopic sites in
such molecules become nonequivalent in CLC solutions,
because they are not symmetry related anymore in the principal
4
4
1
2
4
(
16) (a) Meddour, A.; Canet, I.; Loewenstein, A.; P e´ chin e´ , J. M.; Courtieu, J.
J. Am. Chem. Soc. 1995, 116, 9652. (b) Canet, I.; Courtieu, J.; Loewenstein,
A.; Meddour, A.; P e´ chin e´ , J.-M. J. Am. Chem. Soc. 1995, 117, 6520.
17) (a) Lesot, P.; Merlet, D.; Meddour, A.; Loewenstein, A.; Courtieu J. Faraday
Trans. 1995, 91, 1371. (b) Meddour, A.; Berdagu e´ , P.; Hedli, A.; Courtieu,
J.; Lesot, P. J. Am. Chem. Soc. 1997, 119, 4502.
2
-
(
4
The combined filtrates were evaporated under vacuum and the oily
residue purified by vacuum distillation to yield 24.0 g of 1,2-ethanediol
(∼10% deuterated).
(
18) Eliel, A. L.; Wilen, A. H. In Stereochemistry of Organic Compounds; John
Wiley and Sons: New York, 1994.
(19) (a) Fujita, S. J. Am. Chem. Soc. 1990, 112, 3390. (b) Fujita, S. Tetrahedron
1
991, 47, 31. (c) Fujita, S. Tetrahedron 2000, 56, 735. (d) Hanson, K. R.;
1
,2-Dibromoethane (∼10% Deuterated). The 24.0 g of the above
Eliel, A.; Mislow, K.; Siegel, J. J. Am. Chem. Soc. 1984, 106, 3319. (e)
Mislow, K, Siegel, J. J. Am. Chem. Soc. 1984, 106, 3319. (f) Hirschmann,
H.; Hanson, K. R. Tetrahedron 1974, 30, 3649. (g) Prelog, V.; Helmchen,
G. HelV. Chim. Acta 1972, 55, 2581. (h) Hanson, K. R. J. Am. Chem. Soc.
glycol and 5.3 g of red phosphorus were heated and stirred at 140 °C,
(21) Merlet, D.; Emsley, J. W.; Jokisaari, J.; Kaski, J. P.C.C.P. 2001, 3, 4918.
(22) Merlet, D.; Loewenstein, A.; Smadja, W.; Courtieu, J.; Lesot, P. J. Am.
Chem. Soc. 1998, 120, 963.
(23) Aroulanda, C.; Merlet, D.; Courtieu, J.; Lesot, P. J. Am. Chem. Soc. 2001,
123, 12059.
1
966, 88, 2731. (i) Mislow, K. Raban, M. In Topics in Stereochemistry
967, 1, 1. (j) Hanson, K. R. J. Am. Chem. Soc. 1966, 88, 2731.
1
(
20) Merlet, D.; Emsley, J. W.; Lesot, P.; Courtieu, J. J. Chem. Phys. 1999,
1
11, 6890.
J. AM. CHEM. SOC.
9
VOL. 124, NO. 34, 2002 10073

A R T I C L E S
Lesot et al.
Table 1. Compositions of Liquid-Crystalline NMR Samples Investigated
a
solute/mgb
polymer/mgb
cosolvent/mgb
solute
sample
polymer
DP (PBLG/PBDG)
cosolvent
% of polymer by wt
c
1
1
2
3
4
5
6
6
PBLG
PBLG
PBG
562
562
562/914
562/914
562
DMF
DMF
DMF
DMF
CHCl3
CHCl3
CHCl3
3.6
4.2
3.4
4.2
25.0
129.2
129.9
65.4/65.4
66.2/66.2
100.1
301.0
29.8
29.8
29.7
29.4
16.7
18.3
18.3
c
2
c
1
306.0
313.6
475.1
450.1
450.3
c
2
PBG
3
4
4
PBLG
PBLG
PBG
d
562
562/914
5.4
5.4
102.0
51.2/51.2
d
a
DP: degree of polymerization. b The accuracy of the weights is (0.5 mg. c The solutes 1 and 2 are dideuterated on the three methylene bridges. d To
eliminate residual solid particles, 12.0 mg of solute 4 have been dissolved in 1 g of CHCl3 and then filtered; 450 mg of this solution was added to the
polymer.
Scheme 1. Synthesis of Tri(dioxyethylene-2,2,3,3-d2)-CTV
3
CDCl /TMS: aromatic singlet at δ ) 6.81 ppm; crown-ring methylenes;
two doublets (AX spin system) at δ ) 4.52 and 3.46 ppm, JAX ) 13.8
Hz; hydrogens in the dioxyethylene side chain, a symmetrical multiplet
at δ ) 4.13 ppm.)
As indicated above, isotopically normal tri(dioxyethylene)-CTV was
prepared by trimerization of normal 1,4-benzodioxane-6-methanol,
which was obtained by reduction of the commercially available 1,4-
benzodioxane-6-carboxaldehyde. The procedure was as follows: 10 g
of the latter dissolved in dry ether (200 mL) was slowly added under
stirring to a slurry of 3 g of LiAlH
mixture was refluxed overnight and then hydrolyzed with H
4
in ether at room temperature. The
O,
2
acidified, and separated by repeated extractions. The ether extracts were
dried and the solvent finally removed, resulting in 8 g of a colorless
oil (TLC: one spot, no traces of the aldehyde, silica/CH Cl ). This
2 2
product was then trimerized using the same procedure described for
the synthesis of the tri(dioxyethylene)-CTV (∼10% deuterated).
Sample Preparation. The various liquid-crystalline NMR samples
investigated in this work were prepared using a standard procedure
described elsewhere.10 Note that all homopolymers used are com-
mercially available from Sigma and all 5 mm o.d. NMR tubes were
sealed to avoid solvent evaporation and centrifuged back and forth until
an optically homogeneous birefringent phase was obtained. The exact
composition of each oriented NMR sample (sample 1-7) is shown in
Table 1. Except for compound 3, the amount of CTV derivatives used
in the oriented sample studied does not exceed 1% in weight due to
the rather low solubility of this kind of solute in organic solutions of
PBLG. For the PBLG/DMF system, it was necessary to heat the mixture
slightly in order to dissolve the polypeptide during sample preparation
more easily.10
while dropwise adding 20.0 g of bromine. The reaction mixture was
kept at this temperature for another hour and then cooled to room
temperature. The reaction mixture was diluted with water and ether,
the solid residues were filtered, the ether layer was separated, and the
solvent was evaporated. After distillation, 23 g of the 1,2-dibromoethane
(
∼10% deuterated) was obtained.
,4-Benzodioxane-6-carboxylic Acid Ethyl Ester (∼10% Deuter-
ated). To 11.5 g of 3,4-dihydroxybenzoic acid ethyl ester (Aldrich) in
0 mL of ethanol (95%) was added 9.2 g of solid KOH while being
1
NMR Spectroscopy. Deuterium and carbon-13 NMR spectra in
oriented solvents were collected at 9.4 T on a high-resolution Bruker
DRX 400 spectrometer equipped with a standard variable-temperature
unit (BVT 3000) and using either a selective deuterium probe operating
at 61.4 MHz for deuterium or inverse multinuclear probe (BBI)
operating at 100.13 MHz for carbon-13. The proton-decoupled deute-
rium and carbon-13 spectra were recorded with 8 µs pulse duration
6
stirred. The mixture was slowly heated while 23 g of the 1,2-dibromo-
ethane (∼10% deuterated) was dropwise added and the mixture refluxed
for 6 h. The hot mixture was filtered, the residue was boiled with ethanol
and filtered again, and the combined filtrates were evaporated. An oil
was obtained that was subject to column chromatography (silica, CH
2
-
1
3
(
{
90° pulse) and 6 µs pulse duration (ca. 70°), respectively. For C-
CH /n-hexane 8/2) to yield 4.7 g of pure (one spot on TLC) 1,4-ben-
2
1
2
1
H} and H-{ H} experiments, the protons were decoupled using the
zodioxan-6-carboxylic acid ethyl ester (∼10% deuterated in the
broadband composite pulse sequence WALTZ-16. For the carbon-13
spectra, proton irradiation was applied during the relaxation delay period
so as to benefit from the nuclear Overhauser effect. Other experimental
NMR parameters or details are given in the legend of the figures.
dioxyethylene side chains).
1
,4-Benzodioxane-6-methanol (∼10% Deuterated). A solution of
.5 g of the above ester in 50 mL of dry ether was slowly added to a
slurry of 1 g of LiAlH in 200 mL of ether. After refluxing for 2 h the
compound was hydrolyzed using 10% H SO . After the usual workup,
.2 g of 1,4-benzodioxane-6-methanol (∼10% deuterated) was isolated
TLC: one spot, silica/CH Cl ).
Tri(dioxyethylene)-CTV (∼10% Deuterated). A 3 g portion of the
above compound was heated and stirred with 60 mL of 10% H SO
4
4
2
4
Results and Discussion
3
(
2
2
Chiral Discrimination of Enantiomers in Nonasubstituted
CTV with C3 Symmetry. In this subsection we discuss the
NMR spectra of compounds 1 and 2 (perdeuterated in the ring
methylene; see Figure 2) dissolved in PBLG solutions. Both
2
4
for 6 h. The resulting solid was filtered, washed with water, and dried
under vacuum. The solid was trituriated with 80 mL of benzene
1
derivatives are chiral (C symmetry) and thus, as synthesized,
overnight and isolated. The mass spectrum and H NMR were consistent
3
with the desired product and showed approximately 10% statistical
deuteration in the dioxyethylene side chains. ( H NMR, 500 MHz,
they consist of a racemic mixture of two enantiomers, M and
P. Before recording their NMR spectra in the CLC solvents,
1
10074 J. AM. CHEM. SOC.
9
VOL. 124, NO. 34, 2002

NMR Analysis of Cone-Shaped Compounds
A R T I C L E S
larger than usually observed for proton-decoupled deuterons in
lyotropic PBG solutions. We attribute the extra broadening to
in
out
unresolved geminal D -D dipolar interactions. The full
i
Q
quadrupolar splittings, ∆ν , of these doublets are directly
related to the orientational order parameter, Szz, of the molecular
C3-symmetry axis by
2
i
i
Q
3
2
3 cos â - 1
∆
ν ) S Q
(1)
zz C-D
(
)
2
i
i
where â is the angle between the C-D bond direction and the
molecular C3 axis, and QC-D is the deuterium quadrupole
2
coupling constant, QC-D ) e qQ/h. For aliphatic deuterons this
constant is ∼170 kHz. In eq 1 the quadrupolar coupling tensor
is assumed to be axially symmetric with the unique axis parallel
in
out
to the C-D bond direction. The angles â and â for the inner
and outer deuterons were estimated earlier from X-ray and NMR
measurements, in various substituted CTV derivatives to be
1
approximately 38° and 66°, respectively. Hence, from eq 1,
in
2
1
we can identify the larger and smaller splittings with D and
Figure 3. 61.4 MHz H-{ H} spectrum of the hexadeuterated nonahy-
droxy-CTV (left) and hexadeuterated nonamethoxy-CTV (right), at 310 K
in the achiral PBG/DMF phase (a and b) and in the chiral PBLG/DMF
phase (c and d). The spectra c, d, a, and b (samples 1, 3, 2, and 4 in Table
out
D , respectively as shown in Figure 3. Moreover, it follows
that ∆ν and ∆ν have opposite signs, but on the basis of the
deuterium NMR spectra shown in this figure alone, we cannot
in
out
Q
Q
1
) were recorded by adding 21 000, 7200, 30 000, and 5200, scans,
respectively. For all spectra, zero-filling to 32K data points was used to
increase the digital resolution, and a Gaussian filtering was also applied to
improve the spectral appearance. Peaks marked with solid circles are
tentatively assigned to the saddle conformer. The natural abundance
deuterium signals of DMF are labeled by asterisks.
decide which is positive and which is negative. In the next
section we show (on the basis of the proton-coupled C NMR
13
spectrum) that the sign of Szz for compound 3 dissolved in
PBLG/CHCl3 is negative, indicating that it prefers to align with
its C3 axis perpendicular to the director of the liquid-crystalline
phase. It is safe to assume that the same preference applies also
to other CTV derivatives dissolved in such solvents. Thus,
assuming a negative sign for Szz of compounds 1 and 2 in the
we tried to see whether any chiral discrimination is observed
in their NMR spectra when dissolved in isotropic chiral solvents.
To that end, we used solutions of compound 1 in the enantio-
merically pure solvent 3-methyl-2-butanol with 10% acetone,
as well as in a chloroform solution of an optically pure chiral-
in
Q
out
Q
PBG/DMF solutions, it follows that ∆ν is negative and ∆ν
is positive, as indicated in Table 2. Using the above quoted
8
lanthanide shift reagent. No extra splitting was observed in the
i
Q
value for QC-D, we can use the values of ∆ν to calculate the
proton and carbon-13 spectra in both these solvents.
molecular order parameters, Szz of the solutes in these racemic
For the measurements in the CLC solutions we used proton-
decoupled deuterium and carbon-13 NMR in two solvents,
PBLG/CHCl3 (18%) and PBLG/DMF (29.8%). No discrimina-
tion was observed in the chloroform solutions, but both
compounds exhibited clear chiral discrimination in the proton-
decoupled deuterium and carbon-13 spectra in the DMF
solutions (samples 1 and 2 in Table 1). For comparison, we
also performed measurements in PBG/DMF solutions for which
no chiral differentiation is expected. These samples contain a
racemic mixture of PBLG and PBDG with the same total
concentration of the polymer as in the chiral PBLG solutions
in
out
solutions. The average values (over D and D ) obtained at
10 K are -0.030 and -0.019 for compounds 1 and 2,
3
respectively. These values are quite large for PBG solutions
and reflect the strong tendency of the CTV core to align in such
solvents. A close examination of the spectra in the racemic
solvent shows that the centers of the quadrupolar doublets, due
in
out
to D and D , do not exactly coincide. Rather, the center of
in
the D doublet is slightly shifted (∼0.58 ppm) to low field
out
compared to that of D , indicating that the inner hydrogen of
the ring methylenes is less shielded than the outer ones. Within
the experimental errors due to the line width and considering a
small solvent effect induced by the phase, the shift observed in
CLC is essentially identical to that measured in high-resolution
proton NMR in isotropic solvents (e.g., CHCl3).
(samples 3 and 4 of Table 1). Examples of the deuterium spectra,
recorded at 310 K, in both the racemic and chiral solvents are
shown in Figure 3.
We first discuss the top traces in the figure, which correspond
to the racemic solution. Each spectrum consists of two intense
quadrupolar doublets ascribed to the diastereotopic inner and
Referring next to the spectra in the PBLG/DMF solutions
(bottom traces in Figure 3), we note that each of the doublets
in
out
splits into two, corresponding to the two enantiomers, which
become inequivalent in a chiral environment. The splittings are
small but quite apparent and reflect the chiral discrimination
power of the CLC solvent with respect to the CTV enantiomers.
The degree of discrimination can be quantified using the
“differential ordering effect” (DOE) parameter. For an axially
outer deuterons, which we label D and D in Figure 2. This
result is fully consistent with that recorded in nonchiral,
1
thermotropic nematic solvents. Some weak peaks at the center
region of the spectra are assigned to natural abundance deuterons
of the DMF solvent (namely three quadrupolar doublets centered
on three distinct chemical shifts) and, for compound 1, also to
a small amount of the saddle isomer, as indicated in the caption
of the figure. The width of the lines (∼30 Hz) is somewhat
i
Q
M
i
P
symmetric molecule for which the signs of (∆ν ) and (∆ν )
Q
are identical and assuming that the chiral discrimination is
J. AM. CHEM. SOC.
9
VOL. 124, NO. 34, 2002 10075

A R T I C L E S
Lesot et al.
Table 2. Quadrupolar Splittings in Hz Associated with Din and Dout in the Compounds 1-d6 and 2-d6 in the PBLG/DMF and PBG/DMF
Phase
inner deuterons
outer deuterons
M or P
M or P
solute
sample
solvent
∆νQ /Hz
DOE/%
∆νQ /Hz
DOE/%
1
1
PBLG/DMF
-2928 ( 19
3018 ( 19
-3076 ( 35
-2973.1)a
-1678 ( 14
2005 ( 16
-1916 ( 15
-1841)a
3.0 ( 1.5
+2030 ( 15
+2090 ( 15
+2131 ( 30
(+2059.8)a
+1188 ( 11
+1425 ( 12
+1359 ( 11
(+1306)a
2.9 ( 1.4
-
3
2
4
PBG/DMF
PBLG/DMF
PBG/DMF
0.0
0.0
(
2
17.7 ( 1.6
0.0
18.1 ( 1.8
0.0
-
(
a
The values in parentheses correspond to the average of quadrupolar splittings measured in the PBLG/DMF phase.
entirely due to ordering, the expression for the DOE is
isomer increases by the same amount as its decrease for the
other. This is expected on the basis of similar observation in
other chiral systems, but it is not required by any symmetry
considerations.
i
Q
M
i
Q
P
(
(
(
(
∆ν ) - (∆ν )
DOE ) 2 ×
× 100
i
|
|
|
i
M
i
Q
P
∆ν ) + (∆ν )
Q
Encouraged by the positive results (enantioselectivity) of the
deuterium spectra, we looked for chiral discrimination in the
M
P
S ) - (S )
zz
zz
1
3
1
10,17
)
2 ×
× 100
(2)
C-{ H} NMR of compounds 1 and 2.
For this nucleus
M
P
|
S ) + (S )
zz
zz
the useful order-dependent interaction is the CSA, which is
particularly large for doubly and triply bonded, as well as
aromatic, carbons. In practice, the spectral enantiodiscrimination
is manifested by peak doubling in CLC solution as in the
deuterium spectra. If we assume that the discrimination is
entirely due to the orientational dependent anisotropic part of
the chemical shift tensors, a line splitting between the M and P
enantiomers of
The results are 2.9% and 18.1% for compounds 1 and 2,
respectively (see Table 2). The figures can safely be compared,
since they are derived from spectra taken under the same
experimental conditions (solvent, concentration, temperature).
It is noteworthy that the DOE parameter for 1 is considerably
smaller than that for 2, even though the quadrupole splittings
in the latter are smaller. This indicates that the factors determin-
ing the chiral discrimination and the ordering are not necessarily
the same. The smaller DOE in CTV-OH compared with CTV-
OCH3 seems surprising at first glance. In fact, similar relations
were found earlier for chiral solutes such as ethers or esters,
compared to their hydroxy analogues in PBLG/DMF, while the
opposite relations were found in PBLG/CHCl3 or PBLG/CH2-
M
P
i
i
M
i
P
|νi - ν | ) ν |(δ ) - (δ ) |
(3)
L
aniso
aniso
i
is expected, where δ
is the anisotropic chemical shift for a
aniso
given carbon i (see below) and νL is the carbon-13 Larmor
frequency.
For the measurements we used samples 1 and 2 (PBLG/
DMF), for which the enantiomeric discrimination in the
deuterium spectra was apparent. No spectral differentiation was
found for compound 1, but very pronounced doubling of some
of the aromatic carbon signals was observed for compound 2.
Examples of spectra for this compound are shown in Figure 4.
The upper and lower traces were recorded at 310 K in a
PBLG/DMF solution (sample 2) and at room temperature in
chloroform, respectively. Only the spectral range 110-140 ppm,
including four (out of the six) aromatic peak, is shown. The
strong background signal in the 125-135 ppm range of the CLC
2
4
Cl2 solutions. The effect could be explained in terms of
molecular affinity. In the present case we may argue that the
stronger affinity between the (achiral) DMF solvent and the polar
CTV-OH solute, compared with that of CTV-OCH3, reduces
the affinity of the former with the chiral polypeptide, thus
decreasing its DOE in comparison with that of the nonhydroxylic
CTV derivative.
It is interesting to note that for both compounds, within the
in
experimental accuracy (line width), the DOE values for D and
out
D
are essentially identical, differing by less than 3%. This
fact lends support to the conclusion that the chiral discrimination
is predominantly due to differential ordering.15 If, instead, it
solution is due to the natural abundance aromatic carbon-13 in
the flexible side chains of the PBLG molecule.1
7,25
Some
where due to selective, chiral short-range solute-solvent
interaction on, for example, the quadrupolar coupling, the inner
and outer deuterons would be expected to exhibit completely
different degrees of enantioselectivity. Thus, we conclude that
the discrimination is brought about by small but different
changes in Szz of the two isomers in the chiral solvent. This
impurity lines are also observed. The signals of the CTV solute
are quite weak, due to its low solubility (about one weight
percent), but they are apparent and can readily be identified by
comparison with those in the chloroform solution. The labeling
of these peaks is according to the numbering system of Figure
2. Their assignment is tentative and is based on additive rules
in
conclusion also allows us to identify the outer doublets of D
2
6
out
for benzene substituents. The assignment of the C-H carbon
C-5) was confirmed by the appearance of a doublet in the
and D with one enantiomer and likewise the two inner
(
doublets with the other enantiomer. It is noteworthy that the
average splitting of the corresponding quadrupolar doublets for
a given deuteron is almost identical to its splitting in the racemic
solution. This indicates that in the chiral solution Szz for one
undecoupled C-13 spectrum of the chloroform solution. Well-
resolved, albeit small, splittings of 6, 8, and 13 Hz are measured
(
25) Mirau, P. A.; Bovey, F. A. J. Am. Chem. Soc. 1986, 108, 5130.
26) Kalinowski, H. A.; Berger, S.; Braun. S. In Carbon-13 NMR Spectroscopy;
Wiley: New York, 1988.
(
(24) Unpublished results.
10076 J. AM. CHEM. SOC.
9
VOL. 124, NO. 34, 2002

NMR Analysis of Cone-Shaped Compounds
A R T I C L E S
i
bb
i
cc
i
i
δ
+ δ ) 0, and R and â are the azimuthal and polar angles
i
of the molecular symmetry axis (C3) in the PAS . For aromatic
carbons, the c direction (most shielded) is usually normal to
the molecular plane. Moreover, its value is often quite different
i
aa
i
bb
from the in-plane components, while δ and δ are often
2
9
quite close in magnitude. For a rough estimate of the order-
dependent chemical shifts of the aromatic carbons, we can
therefore assume that the tensor is approximately axially
i
aa
i
i
cc
symmetric (δ ) δ ) -δ /2), so that
bb
2
i
i
2
3
i
3 cos â - 1
δ
) S ∆δ
(5)
aniso
zz
(
)
2
i
i
cc
i
aa
i
where ∆δ ) [δ - (δ + δ )/2]. For the CTV core, the
bb
1
i
angle â is about 43°, while the ∆δ values depend somewhat
on the substituents, ranging from about -150 ppm for the C-H
atoms, through ca. -130 ppm for the ternary (C-C) atoms to
ca. -80 ppm for oxygen-substituted (C-O) carbons.29 Taking
the Szz value from the deuterium results yields, respectively,
i
0
.9, 0.8, and 0.6 ppm for δ
of the three types of carbons,
aniso
respectively.
This corresponds to overall shifts of 90-60 Hz (for a
spectrometer frequency of 100 MHz) compared with their
respective isotropic values. Finally, assuming that the chiral
discrimination results entirely from ordering, we can use the
DOE value determined from the deuterium spectra (18%) to
calculate chiral discriminations of about 16 Hz for the C-H
carbons, 14 Hz for the ternary carbons, and 11 Hz for the C-O
carbons. Considering the crudeness of the approximation made,
the comparison with the experimental results (Figure 4) is not
unsatisfactory. The order of magnitude agreement, in fact,
strongly supports the conclusion that the chemical shift dis-
crimination in the PBLG solution is driven by differential
ordering. Also, the lack of chiral discrimination in the carbon-
13 spectrum of sample 1 (compound 1) is not surprising. A
similar order of magnitude calculation predicts a maximum
splitting (for the C-H carbon) of ∼3 Hz, which is well within
the experimental line width (6-8 Hz). On the other hand, the
discussion above also shows that the application of carbon-13
NMR for quantitative analysis of the discrimination, for
example, determining the degree of selective ordering, is more
subtle than for deuterium, because it requires an exact knowl-
edge of the chemical shift tensors, which are in general not
simply related to the molecular structure.
Figure 4. 100.6 MHz 13C-{ H} signals of the aromatic C-1, C-2, C-3,
1
and C-5 carbons of the hexadeuterated nonamethoxy-CTV (sample 2) at
3
10 K in the PBLG/DMF phase (a) and at room temperature in an isotropic
solution of CHCl3 (b). Signals of carbons C-4 and C-6 located at around
52 ppm are not shown, because they do not exhibit enantiomeric
discrimination. Spectra a and b were recorded by adding 27 000 and 10 000
scans, respectively. For both a recycle delay of 1.8 s was used. No filtering
was employed. The peak assignment corresponds to the numbering system
shown in Figure 2.
1
for carbons 2, 3, and 5, but none are measured for the other
aromatic carbons. The chemical shift differences evidently
reflect the discrimination of the CTV-OCH3 enantiomers in
the chiral liquid-crystalline solvent. An exact analysis of the
chemical shifts is not as straightforward as for the quadrupolar
splittings in the deuterium spectra. This is because the carbon-
1
3 results are quite sensitive to the CSA parameters (principal
values and principal directions). While the principal values of
the chemical shift tensors can often readily be obtained from
magic angle spinning (MAS) measurements,27 the accurate
determination of the principal axis system (PAS) in the
molecular frame is more involved and as a rule requires single-
crystal measurements as a function of the magnetic field
orientation. We shall therefore reverse the procedure and attempt
to calculate the expected enantiomeric discrimination in the
carbon-13 spectra from the earlier analysis of the deuterium
results and crude estimates of the CSA for the aromatic carbons.
Spectral Discrimination of Enantiotopic Elements in
Hexasubstituted CTV with C3W Symmetry. In this section we
explore possible spectral discrimination of enantiotopic elements
of hexasubstituted CTV with C3V symmetry in PBLG liquid-
crystalline solutions. These molecules possess equivalent groups
of atoms related only by the mirror planes in the molecule
(enantiotopic elements), for example, the pairs of aromatic
carbons 1/6, 2/5, and 3/4, or groups linked to these atoms (see
Figure 2). These pairs become nonequivalent in chiral circum-
stances and in principle should be distinguished in NMR
i
For a solute molecule with a C3 (or higher) symmetry axis, δ aniso
2
8
is given by
2
i
i
2
i
cc
1
2
i
i
bb
3 cos â - 1
δ
aniso ) Szz[3
δ - (δ + δ )
+
(
aa
)
(
)
2
1
2
i
i
2
i
i
(
δ - δ )(sin â )cos 2R (4)
aa
bb
]
1
8
spectroscopy using chiral solvents. However, as indicated in
the Introduction, specific chiral, short-range solute-solvent
interactions are usually too weak to exhibit a measurable
enantiotopic discrimination, while discrimination by orienta-
i
aa
i
bb
i
cc
where δ , δ , and δ are the principal components of the
carbon-13 anisotropic chemical shift tensors in their respective
principal axis system (PAS ) with δ > δ > δ and δ
i
i
aa
i
bb
i
cc
i
aa
+
(
27) Herzfeld, J.; Berger, A. E. J. Chem. Phys. 1980, 73, 6021.
28) Snyder, L. C. J. Chem. Phys. 1965, 43, 4041.
(29) Duncan T. M. In Principal Components of the Chemical Shifts Tensors. A
Compilation, 2nd ed.; Farragut Press: Madisson, 1977.
(
J. AM. CHEM. SOC.
9
VOL. 124, NO. 34, 2002 10077

A R T I C L E S
Lesot et al.
related to the aromatic C-H carbons (carbons 2/5) of compound
3
. Trace a, exhibiting a doublet of triplets, is a undecoupled
carbon-13 spectrum recorded in a CDCl3 solution at room
temperature. The larger doublet splitting is due to scalar coupling
1
with the adjacent proton, JC-H ) +160 Hz, where the positive
1
17,26
sign is the commonly accepted sign for JC-H.
The smaller
splitting is due to scalar interaction with the nearby ring
3
methylene, JC-H ) 6.0 Hz. Spectrum b is a natural abundance
deuterium spectrum of compound 3 dissolved in PBLG/CHCl3
3
0
(
sample 5) at 300 K. The outer doublet (open circles)
corresponds to the aromatic C-D deuteron and exhibits a
C-D
Q
splitting of |∆ν | ) 1930 ( 30 Hz. The other doublets in
the spectrum are due to natural abundant deuterons in the side
chain methyls (solid circles) and the chloroform solvent
(asterisks). From the structure of the CTV (crown) core (Figure
2
) it follows that the aromatic C-D bonds are perpendicular to
C-D
the molecular C3 axis (â
) 90°). We can thus use eq 1 to
calculate the magnitude of the order parameter, yielding
Szz| ) 0.015. The more interesting spectrum is that shown in
part c. It is the proton-coupled carbon-13 spectrum of carbons
/5 in the same PBLG/CHCl3 solution (sample 5) and recorded
|
2
at the same temperature (300 K) as the deuterium spectrum of
b. The structure of this spectrum can be interpreted in terms of
1
3
1
1
Figure 5. Series of NMR spectra associated with the aromatic C-H carbon
the following three-spin system, C- HA‚‚‚ HA′ (see drawing
in Figure 5), where the C-HA pair belongs to one benzene ring
and HA′ to the nearby ring. A quantitative analysis of this
spectrum in terms of the dipolar and scalar interactions between
the three nuclei allows us to determine both the magnitude and
sign of the molecular order parameter, Szz.
13
1
(
2/5) of compound 3, recorded at 300 K. (a) 100.6 MHz C-{ H} spectrum
(
2900 scans) in CDCl3. The large doublet originates from the scalar
1
couplings JC-H, while the triplet structure is due to scalar couplings with
3
2
1
the nearby methylene protons JC-H. (b) 61.4 MHz H-{ H} signals at
natural abundance (160 000 scans) in PBLG/CHCl3. The chloroform signals
are labeled with asterisks. The quadrupolar doublet associated with the
deuterons of the methyl and aromatic groups are labeled by solid and open
circles, respectively. No signals are observed for the deuterons of the ring
methylenes. (c) 100.6 MHz C-{ H} (30 000 scans) in the PBLG/CHCl3
phase. An exponential filtering (LB ) 4 Hz) was applied to reduce the
spectral noise. (d) Simulated proton-coupled carbon-13 spectrum obtained
with the spectal data reported in the text. The spectrum is displayed using
a line width of 2 Hz. (e) Same as d, calculated with a line width of 25 Hz.
The spin Hamiltonian of this three-spin system for a molecule
that dissolves in a nematic solvent is
1
3
1
1
3
1
C
H
A
A′
1
A
Hˆ ) -ν S - ν (I + I ) + ( J
- 2DC-H)S I -
L
z
L
z
z
C-H
z z
A A′
z
1
4
A A′
+ -
A A′
- +
2
DH-H I I - (I I + I I ) (6)
tional order is not possible for molecules with C3V symmetry.20
The latter law applies to rigid molecules with fixed symmetry.
The question posed here, however, is whether it also applies to
[
z
]
where the first two terms correspond to the Zeeman energies
of the carbon-13 and hydrogen nuclei, respectively, and the next
two terms are the C-H and H -H interactions. We have
“
flexible” molecules, namely molecules undergoing rapid in-
terconversion between several, not necessarily symmetric,
conformations, even though their ensemble average has a C3V
symmetry. The question is relevant to the substituted CTV
derivative, since although the core is rigid, the substituents are
usually flexible. We have studied two hexasubstituted CTV
compounds with (average) C3V symmetry, hexaacetyloxy-CTV
A
A
A′
neglected long-range scalar and dipolar couplings (over more
than a single bond). Recalling the 3-fold molecular symmetry
of CTV and that the dipolar tensor is axially symmetric, the
dipolar couplings in eq 7 are given by an expression similar to
eq 1,
(
compound 3) and tri(dioxyethylene)-CTV (compound 4),
2
M-N
partially deuterated to ∼10% in the dioxyethylene groups. No
chiral discrimination was detected in the proton-decoupled
deuterium and carbon-13 (both in natural abundance) of a
solution of compound 3 in a PBLG/CHCl3 solution. However,
a clear spectral enantiodistinction was observed for the enan-
tiotopic deuterons in the ethylene groups of the dioxanne chain
in compound 4. In section II(a) we discuss this effect in terms
of differential ordering of the various conformers of this
compound. Before doing so, however, we briefly describe the
spectra obtained for compound 3 in the lyotropic PBLG/CHCl3
solution. As mentioned earlier, this allowed us to determine the
magnitude and sign of the order parameter of this solute in the
PBLG mesophase.
KM-N
3 cos â
- 1
DM-N ) Szz
(7)
(
)
3
2
r
M-N
where rM-N is the M to N distance (in Å), âM-N is the angle
between the rM-N direction and the molecular C3 axis, and KM-N
is a constant proportional to the magnetic dipoles of the M and
3
3
N nuclei, KC-H ) 30 188 Hz.Å and KH-H ) 120 060 Hz.Å .
M-N
Since â
is the same for both pairs of interacting nuclei
) 90°), it follows that both interactions have the same
sign. The Hamiltonian in eq 6 can be solved analytically. For
the carbon-13 spectrum it yields a symmetric five-line spec-
M-N
(â
(
30) (a) Lesot, P.; Merlet, D.; Loewenstein, A.; Courtieu, J. Tetrahedron :
Asymmetry 1998, 9, 1871. (b) Merlet, D.; Loewenstein, A.; Smadja, W.;
Courtieu, J.; Lesot, P. J. Am. Chem. Soc. 1998, 120, 963.
NMR Spectra and Order Parameter of Hexaacetyloxy-
CTV in PBLG/CHCl3. In Figure 5 is shown a set of spectra
10078 J. AM. CHEM. SOC.
9
VOL. 124, NO. 34, 2002

NMR Analysis of Cone-Shaped Compounds
A R T I C L E S
13
C
trum centered around ν , featuring (a) an inner doublet at
L
1
(
1/2( JC-H - 2DC-H), each with the relative intensity 1, (b)
an outer doublet located at (1/2[( JC-H - 2DC-H) + 4D
each with the relative intensity ( JC-H
( JC-H - 2DC-H) + 4D
1
2
2
1/2
] ,
H-H
2DC-H) /
] (smaller than 1), and (c) a
1
2
-
1
2
2
[
H-H
2
1
central singlet with relative intensity 8D
/[( JC-H -
H-H
2
2
2
DC-H) + 4D
]. Despite the poor resolution, the main
H-H
features of the expected spectrum can readily be recognized in
trace c of Figure 5. To facilitate the analysis, we first estimate
the magnitude of the DC-H from the measured quadrupolar
splitting of the corresponding C-D deuteron (trace b of Figure
5
). From eqs 1 and 7 it follows that for the aromatic C-H/
C-D
C-D fragment the ratio ∆νQ /DC-H is fixed by the quadru-
3
31
polar and dipolar constants at (3/2)QC-D/(KC-H/r ) ≈ 12.
C-H
C-D
Thus, from the measured |∆νQ | (1930 Hz), |DC-H| is
estimated to be about 170 Hz. Referring next to trace c, we can
readily recognize the two (unresolved) doublets (a and b) and
the center peak (c). From the splitting of the inner doublet we
Figure 6. Schematic representation of twisted dioxyethylenebenzene
fragments of compound 4. The right twist (a) and a left twist (b) are shown.
1
calculate | JC-H - 2DC-H| ≈ +150 Hz. Since we know that
1
JC-H ) +160, DC-H must be ca. +150 Hz. The alternative
solution of |DC-H| ∼ 0 Hz is ruled out by a comparison with
RRR and LLL comprising an enantiomeric pair (which will be
denoted arbitrarily M and P) and likewise RLL and LRR. In a
nonchiral liquid-crystalline solution there is no discrimination
between enantiomeric pairs. The M and P conformers have the
same fractional population, the same order parameter, and, from
the NMR point of view, the same magnetic parameters for
corresponding nuclei. Moreover, the fast interconversion be-
tween the various conformations results in an average C3V
symmetry. This statement applies, in fact, to each class of
conformers (C3 and C1) separately. Hence, sites in the molecule
that are related by the average C3V symmetry are chemically
equivalent (as in isotropic medium), but if they are related by
reflection, they are also enantiotopic. In a chiral liquid-crystalline
solvent the chemical equivalence of such sites is not preserved.
This is because, in principle, neither the fractional population
nor the ordering of the M and P enantiomeric conformers need
to be the same, nor even the magnetic parameters of their
corresponding nuclei. These effects may lead to chiral spectro-
scopic discrimination of enantiotopic sites, as we have indeed
observed in the case of compound 4 embedded in the PBLG
solvent.
C-D
Q
|
∆ν |. From this value of DC-H, using rC-H ) 1.09 Å, we
obtain Szz ≈ -0.013. Thus, both dipolar couplings, DC-H and
DH-H as well are positive, while Szz is negative. This means
that the molecular C3-symmetry axis of compound 3 prefers to
be aligned perpendicular, rather than parallel, to the director in
the PBLG/CHCl3 phase. We assume that the same applies to
the other CTV derivatives in the lyotropic PBLG solutions.
Finally a best-fit analysis was carried out on the experimental
spectrum, yielding DC-H ) +159 ( 2 Hz and DH-H ) +49 (
1
Hz. The calculated best-fit spectrum, using an effective line
width of 25 Hz, is shown in trace e of Figure 5. When the same
parameters are used with a line width of only 2 Hz, the
calculated spectrum shown in d is obtained. We believe that at
least some of the broadening is due to dipolar interaction with
remote protons in the molecule. Using the final result DC-H
and eq 7, we obtain Szz ) -0.0146, which is essentially identical
C-D
to the absolute value calculated above from ∆νQ (|Szz| )
0
.015).
Deuterium NMR Spectra and Order Parameter of Tri-
(dioxyethylene)-CTV in PBLG/CHCl3. Tri(dioxyethylene)-
In Figure 7 are shown deuterium NMR spectra of solutions
of tri(dioxyethylene)-CTV enriched to 10% deuterium in the
ethylene side groups (compound 4), recorded at 300 K in achiral
PBG/CHCl3 (top) and chiral PBLG/CHCl3 (bottom) solutions
CTV (compound 4) consists of the rigid CTV core to which
three dioxyethylene bridges are symmetrically linked. The
dioxyethylene bridges can acquire two states, corresponding to
a right twist and a left twist, as displayed in Figure 6, which
we label hereafter R and L, respectively. From the high-
resolution NMR spectra of this compound we know that these
states rapidly interconvert, so that hydrogens A and A′ are on
average equivalent and likewise hydrogens B and B′. When all
possible twisted states are considered, four types of conformers
for the tri(dioxyethylene)-CTV molecules are obtained. Two of
them, labeled RRR and LLL, have C3 symmetry, where the
triplets of letters indicate the twist of the dioxyethylene groups
sequentially around the CTV core. The two other conformers
are RLL and LRR with C1 symmetry. The latter two are
structurally 3-fold degenerate, RLL, LRL and LLR, and
similarly, LRR, RLR, and RRL. Each conformer is chiral with
(samples 6 and 7, respectively). We first discuss the spectrum
observed in the racemic solution, which exhibits three quadru-
polar doublets. The more intense one, centered at 6.5 ppm
(labeled by asterisks), is due to the natural abundant deuterons
in the chloroform solvent. The two weaker doublets (open and
solid circles), centered at 3.5 ppm are ascribed to the A(A′)
and B(B′) deuterons in the enriched dioxyethylene groups (see
Figure 6). This spectrum is consistent with the average C3V
symmetry of the compound, according to which sites A and A′
are equivalent, and likewise B and B′.
To quantitatively analyze the spectrum in the racemic solution
it is sufficient to consider the average over a single pair of R
and L conformations and assume an overall C3V symmetry for
the tri(dioxyethylene)-CTV molecule. Thus, the average quad-
(31) Sarfati, M.; Courtieu, J.; Lesot, P. Chem. Commun. 2000, 1113.
J. AM. CHEM. SOC.
9
VOL. 124, NO. 34, 2002 10079

A R T I C L E S
Lesot et al.
Table 3. Geometrical Parameters Related to the C-D Bonds in
the Oxyethylene Side Groups, in the R and L Conformations of
Compound 4a
C−D bond
A
A′
B
B′
R
a
θ (i)
74.7
-34.9
8.59
86.5
110.3
-0.3197
8.59
-86.5
74.7
34.9
44.2
0.2709
171.4
83.1
104.9
-34.8
(-)137.3
0.3112
104.9
34.8
171.4
-83.1
136.8
0.2984
R
æ (i)
L
θ (i)
L
æ (i)
R
b
â (i)
R
F (i)
a
All angles are in degrees. θ and æ are the polar and azimuthal angles
of the indicated C-D bonds in the frame of the dioxyethylenebenzene
fragment (Figure 6). In this frame, c is perpendicular to the benzene plane,
b is in the direction linking the two aromatic C-H bonds, and a completes
the frame to a right-handed system (Figure 2). â is the angle between the
indicated C-D bond and the molecular C3 axis and F is the second-order
Legendre function (eq 9). In the calculation of the â values, an angle of
43° between C3 and Z was assumed. Only values for the R conformation
are given. Those for the L conformation are related to the latter by
L
R
L
R
â (A) ) â (A′) and â (A′) ) â (A) and similar relations apply for the
B and B′ deuterons. b The sign (-) indicates that the angle refers to an
axial C-D bond, pointing downward and inside the CTV core.
functional theory) program.34 The angular orientations of the
ethylene C-D bonds, θ and æ in these fragments are sum-
marized in Table 3. Note that the calculations do not yield a
perfect symmetric twist, but the deviations are minute. Perfect
n
n
symmetry would require that θ (A,A′) ) 180° - θ (B′,B) and
2
1
Figure 7. 61.4 MHz H-{ H} spectra of compound 4, with statistically
n
n
n
æ (A,A′) ) -æ (B′,B). Using these results, the angles â (i)
between the C-D bond directions and the molecular C3 axis
were computed assuming that the latter is inclined by 43° to
the benzene ring normal (c in Figure 2). The results for the
various â (i) values and corresponding F (i) functions are also
given in Table 3. Only the â (i) values for the n ) R
conformation are included. Those for the L conformation are
related to the former, by â (A) ) â (A′), â (A′) ) â (A) and
a similar relation exists for the B and B′ deuterons.
(
∼10%) deuterated dioxyethylene side chains, recorded at 300 K. (a) In
the chiral solvent PBLG/CHCl3. (b) In the achiral solvent PBG/CHCl3. The
spectra a and b were recorded by adding 20 000 and 10 000 scans,
respectively. For both spectra, zero-filling to 32K data points was used to
increase the digital resolution and a Gaussian filtering was applied to
improve the spectral appearance. The quadrupolar doublets associated with
1
n
n
n
(
A and A′) and (B and B′) deuterons are labeled with open and solid circles,
respectively. Peaks with asterisks correspond to the natural abundance
deuterium signals of chloroform.
L
R
L
R
rupole splittings of the A (or B) deuterons can be written as32,33
For the calculations it is sufficient to consider the enantio-
meric pair, RRR (say M) and LLL (P), both of which have C3
symmetry. Identical conclusions would be reached if the C1
isomers were included in the analysis, since the average over
RLL, LRL, and LLR also corresponds to C3 symmetry and
likewise the average over LRR, RLR, and RRL. Considering
first the results for the PBG solution, where the M and P
3
n(i)
〈
Q
i〉
n
n
n
∆
ν ) f S
Q
F (i)
(8)
∑
zz( ) C-D
n
2
where
2
n
3
cos â (i) - 1
M
P
n
conformation have the same ordering (S ) S ) Szz), the
F (i) )
(9)
zz
zz
2
M
P
same fractional populations (f ) f ) 1/2) and using the entries
from Table 3, we obtain from eq 9
In these equations the index, i, labels the deuterons A and A′
(
or B and B′), the summation index, n, indicates whether the
〈
A〉
〈A′〉
M
M
∆ν
) ∆νQ ) 127.5S [F (A) + F (A′)]
n
Q zz
twist of the dioxyethylene bridge is R or L, f is the fractional
population of the two twists, and â (i) the angle between the
n
)
-6.22S (in kHz)
(10a)
(10b)
zz
C3 axis and the C-D bond direction of the ith deuteron in the
nth conformer. In eq 8 we included the possibility that QC-D
may, in principle, depend on the deuterium site (A or B) and
conformation (R or L), hence the superscript n(i). On the basis
of the discussion above we consider this effect too small to be
of importance here, and in the following we will drop the
superscript and use the quadrupolar coupling constant at 170
kHz for all deuterons. To proceed we need to estimate the â (i)
values. This we have done in two steps. First, we calculated
the geometry of the twisted dioxyethylene-benzene fragments
and
〈
Q
B〉
〈B′〉
M
M
∆ν ) ∆νQ ) 127.5S [F (B) + F (B′)]
zz
)
+77.72S (in kHz)
zz
This result immediately allows us to identify the doublets
with the larger and smaller splittings (solid and open circles in
Figure 7) with the deuterons A(A′) and B(B′), respectively. To
proceed we use eq 10b for ∆ν . The corresponding equation
for ∆ν is quite unreliable, since the numerical coefficient in
n
〈
B〉
Q
(R and L of Figure 6) using the Gaussian 98 DFT (density
〈
Q
A〉
(
32) Emsley, J. W.; Luckhurst, G. R.; Stockley, C. P. Proc. R. Soc. London A
982, 381, 117.
33) Emsley, J. W.; Luckhurst, G. R. Mol. Phys. 1980, 41, 19.
it results as a small differences between similarly large numbers
1
(
(34) Gaussian 98, Revision A.7; Gaussian Inc., Pittsburgh, PA, 1998.
10080 J. AM. CHEM. SOC.
9
VOL. 124, NO. 34, 2002

NMR Analysis of Cone-Shaped Compounds
A R T I C L E S
and is therefore subject to a very large uncertainty. Thus, taking
enantiomers of compounds 1 and 2 (see DOE values in Table
2). We cannot rule out, however, the possibility that the
discrimination is due to imbalance in the population of the M
and P enantiomers or to a combined effect of the latter and
selective ordering. The mechanism discussed above can only
affect the flexible part of the molecule. In the approximation
used to derive eq 12, the geometrical factor vanishes for atoms
in the rigid part of the molecule, and therefore, no discrimination
based on this mechanism is expected in the NMR spectra of
nuclei from this part. In principle, other mechanisms could lead
to discrimination of enantiotopic elements. For example, the
twisted side chains can induce magnetic nonequivalence of the
substituted aromatic carbons (3,6) that could lead to NMR
spectroscopic discriminations.
〈
B〉
|
∆ν | ) 793 Hz yields |Szz| ≈ 0.010, where, based on the
Q
earlier results, the sign is most likely negative.
We next discuss the spectrum of tri(dioxyethylene)-CTV in
the chiral solvent (bottom trace in Figure 7), where both the
A,A′ and B,B′ enantiotopic deuterons exhibit chiral discrimina-
〈
Q
A〉
〈A′〉
〈
Q
B
〉
〈B′〉
tion with ∆ν * ∆νQ and ∆ν
* ∆νQ . Although this
discrimination is not forbidden by symmetry, its origin is
puzzling. To discuss the mechanism of this discrimination and
estimate its extent, we use again eq 9. Here too it is sufficient
to consider the enantiomeric conformer pair RRR (M) and LLL
(P). Thus, the difference in the quadrupolar splittings between
deuterons A and A′ can be written as
〈
Q
A-A′〉
〈A〉
〈A′〉
∆
∆ν
) ∆νQ - ∆νQ
Finally, the question may be asked, why was there no
discrimination observed in the carbon-13 and deuterium spectra
of hexaacetyloxy-CTV (compound 3) in the chiral solvent
PBLG/CHCl3 (sample 5). After all, here too the side chains are
flexible and can lead to many, including chiral, conformations.
We believe that the effect certainly exists, but it is simply too
small to be experimentally observed. As we saw in the
discussion above, the spectral discrimination reflects small
differences in the ordering and/or the population distribution
of the conformers. Apparently these effects tend to cancel when
a large number of conformers is involved in the averaging.
M
M
M
P
P
P
)
(3/2){[f S Q F + f S Q F (A)] -
zz C-D zz C-D
Μ
M
M
P
P
P
[
φ S Q F (A′) + f S Q F (A′)]} (11)
zz C-D
zz C-D
〈
Q
B-B′〉
〈B〉
〈B′〉
and similarly for ∆∆ν
) ∆νQ - ∆νQ . From eq 11 the
discrimination can result from a selective effect on the fractional
population (fM * f ) and/or the ordering (S * S ). In princi-
P
M
zz
P
zz
ple, the discrimination can also result from chiral, short-range
solute-solvent interactional effects on the magnetic and geo-
n(i)
metrical parameters (QC-D and â ), but as before, we reject
M
P
this possibility as unlikely. By symmetry we set â (A) ) â (A′),
Summary and Conclusions
M
P
â (A′) ) â (A) (see Figure 6 and Table 3). Equation 11 thus
In the present work we extend earlier studies of chiral
discrimination, by NMR spectroscopy, in the lyotropic chiral
nematic mesophases of PBLG to solutes derived from the crown
form of CTV. This core does not possess stereogenic tetrahedral
centers, but depending on the substitution pattern, its derivatives
can be chiral, prochiral, or achiral. In the present work we
concentrated on two types of substituted CTV derivatives having
C3 and C3V symmetries. In the first part of the paper we discuss
the (proton decoupled) deuterium and carbon-13 NMR spectra
of nonasubstituted CTV derivatives with C3 symmetry. Depend-
ing on the nature of the cosolvent, in particular its polarity, very
significant chiral discrimination could be obtained, demonstrat-
ing the power of the method for studies of such chiral
compounds and its potential for analytical application to these
systems.
becomes
〈
Q
A-A′〉
∆
∆ν
)
M
M
P
P
M
M
2
55(f S - f S )[F (A) - F (A′))] (in kHz) (12)
zz zz
〈
Q
B-B′〉
n
with a similar equation for ∆ν
values from Table 3 yields
. Substituting the F (i)
〈
A-A′〉
Q
M
M
zz
P
P
zz
∆
∆ν
) (150(f S - f S ) (in kHz)
(13a)
and
〈
B-B′〉
Q
M
M
P P
∆
∆ν
) (3.26(f S - f S ) (in kHz) (13b)
zz zz
where the sign is undetermined because we did not commit
In the second part of the paper we discuss substituted CTV
with flexible side chains having on average C3V symmetry. In
solution these compounds consist of mixtures of rapidly (NMR
wise) interconverting conformers, with in general lower sym-
metries than C3V, including chiral enantiomeric pairs. Nonstatis-
tical distribution or selective ordering of these enantiomeric
conformers may result in spectral discrimination of some sites,
which on the average are enantiotopic. It is important to point
out that, based on population or ordering selectivity alone, only
sites in the flexible part of the molecule can be spectrally
discriminated. No discrimination of enantiotopic elements in
the rigid part is possible by this mechanism alone. The question
of whether the discrimination observed in the flexible part is
predominantly due to selective ordering and/or deviation from
statistical population distribution of enantiomeric isomers
remains open. On the basis of comparison with the results for
the C3 compound in the first part of the paper, it appears that
the former mechanism dominates the discrimination. If indeed
ourselves to the absolute configuration of the M and P
〈
A-A′〉
enantiomers. The calculations suggest that ∆∆ν
and
Q
〈
B-B′〉
∆∆ν
have opposite signs, but this conclusion depends
Q
M
P
very delicately on the difference [F (B) - F (B′)], which is
minute compared to the magnitude of each element separately
(
see Table 3). For this reason we continue the analysis with the
〈
A-A′〉
less vulnerable eq 13a for ∆ν
. Assuming that the domi-
Q
nant effect on the discrimination is due to selective ordering,
M
rather than an imbalance in the population distribution (S
*
-
zz
P
M
P
〈A-A′〉
M
S , f ) f ) 1/2), eq 13a becomes |∆∆ν
| ) 75|(S
zz
Q
zz
P
S )| (in kHz). The two quadrupolar doublets of the A and A′
zz
deuterons in Figure 7 have splittings of 40 and 10 Hz. If both
〈
Q
A-A′〉
splittings have the same sign, |∆∆ν
| ) 30 Hz, while if
| ) 50 Hz. Substituting these
values (with |Szz| ) 0.011, calculated from the average of B
〈
Q
A-A′〉
their sign is opposite, |∆∆ν
M
zz
P
zz
and B′ splittings) yields |(S - S )/Szz| ) 0.037 and 0.060,
respectively. Both values are on the order found for the
J. AM. CHEM. SOC.
9
VOL. 124, NO. 34, 2002 10081

A R T I C L E S
Lesot et al.
this is so, the example discussed in the second part of the paper
is the first direct observation of conformation dependent ordering
in a CLC. Until now the arguments of an orientation-
conformation dependence have been invoked to explain some
unexpected deviations between theoretical and experimental
the possibility to differentiate between them. The first class
contains all nonplanar rigid derivatives of C , C , S , and D
2d
s
2V
4
symmetry. Typical examples of enantiotopic distinction for this
molecular class can be found in refs 20, 21, and 23. The second
one contains all flexible derivatives exhibiting one of these
symmetries on average on the NMR time such as benzyl alcohol
or ethyl alcohol reported in refs 22 and 23. The third and the
last one contain all flexible molecules (such as compound 4)
of which the average symmetry differs from Cs, C2V, S4, and
D2d but exhibit a sufficiently large differential ordering for their
enantiomeric conformers in the course of their stereochemical
dynamic.
results (after fastidious calculations involving various elaborated
models),3
3,37
but to the best of our knowledge, a direct exper-
imental evidence has not been reported so far. Conclusively,
NMR in PBLG solutions should provide an interesting tool for
investigating the orientation-conformation correlation in flex-
ible molecules.
Contrary to simple stereochemical considerations, we have
shown that all elements (nuclei, groups, internuclear directions)
defined as enantiotopic, according to the symmetry or substitu-
The next step of this work will consist of quantifying the
molecular orientational ordering dependence on molecular
conformation in the case of two flexible enantiomers embedded
in a chiral oriented solvent. This research is currently underway.
18,19
tion criteria customary employed as a test,
cannot be
experimentally differentiated using NMR in a CLC system as
far as selective ordering is only concerned. From an NMR in
CLC point of view, it now becomes possible to divide solutes
exhibiting enantiotopic elements in three classes according to
Acknowledgment. This work was partly supported by the
German-Israeli foundation, G.I.F. Project NO-I-558.219.05/97.
We thank B. Epel and P. Carl for carrying out the DFT
calculations.
(
35) (a) Lesot, P.; Merlet, D.; Loewenstein, A.; Courtieu, J. Tetrahedron : Asym.
1
998, 9, 1871. (b) Merlet, D.; Loewenstein, A.; Smadja, W.; Courtieu, J.;
Lesot, P. J. Am. Chem. Soc. 1998, 120, 963.
(36) Sarfati, M.; Courtieu, J.; Lesot, P. Chem. Commun. 2000, 1113.
37) Photinos, D. J.; Samulski, E. T.; Toriumi, H. J. Phys. Chem. 1990, 94,
(
4
688.
JA020269D
10082 J. AM. CHEM. SOC.
9
VOL. 124, NO. 34, 2002