Analysis of intramolecular dynamic processes in enantiomeric diaryl atropisomers and related derivatives by2H NMR spectroscopy in polypeptide liquid crystals

Article abstract of DOI:10.1002/chem.200601284

We demonstrate the analytical potential of ²H-{¹H} NMR spectroscopy in weakly ordering, chiral lyotropic liquid crystals made of poly(γ-benzyl-L-glutamate) (PBLG) dissolved in chloroform or dichloromethane for investigating the intra

Full text of DOI:10.1002/chem.200601284

DOI: 10.1002/chem.200601284
Analysis of Intramolecular Dynamic Processes in Enantiomeric Diaryl
Atropisomers andRelatedDerivatives by 2H NMR Spectroscopy in
Polypeptide Liquid Crystals
[b]
Olivier Lafon,[a] Philippe Lesot,*[a] Chun-An Fan,[b] andHenri B. Kagan
Abstract: We demonstrate the analyti-
phenomena are observed, simulation of
the experimental 2H-{1H} lineshapes
using a formalism tailored for two deu-
terons undergoing mutual exchange
allows the rate constants and the acti-
vation parameters for the internal rota-
tion processes to be calculated. Experi-
mental values of DH° have been com-
pared with data evaluated by molecular
modelling calculations and the activa-
tion parameters are discussed for the
various compounds. It is shown that
these polypeptide mesophases have no
significant impact on the interconver-
sion dynamics of these compounds. In
contrast with the nematic thermotropic
phases, Hallerꢀs equation cannot be
used to predict the evolution of the
quadrupolar splittings (DnQ values),
and hence the order parameters, versus
T in the PBLG mesophases. For these
particular lyotropic systems, it is shown
that an exponential function of the
2
cal potential of H-{1H} NMR spectros-
copy in weakly ordering, chiral lyotrop-
ic liquid crystals made of poly(g-ben-
A
chloroform or dichloromethane for in-
vestigating the intramolecular dynamic
processes of four deuterated diaryls
C
(derivatives of 1-(4’-methylphenyl)-
naphthalene). When the rotation of the
form
DnQ[Hz]=C”exp
A
2
2
ꢀ
aryl groups about the sp sp bond is
sufficiently slow relative to the NMR
timescale, the method allows the spec-
tral discrimination of enantiomeric
atropisomers or enantiotopic directions
in the prochiral derivatives. The effect
of the position of substituents on the
phenyl group on the conformational
dynamics of these compounds has been
examined as well as the nature of the
organic co-solvent. When coalescence
provides excellent agreement between
the experimental and expected DnQ
values. Analysis of the results reported
in this work suggests that orientation
and chiral discrimination phenomena
in these lyotropic solvents could be
treated separately because they would
involve different interaction mecha-
nisms.
Keywords: atropisomerism · chiral-
ity · conformational analysis · ex-
change processes · kinetic and acti-
vation parameters · liquid crystals ·
NMR spectroscopy
Introduction
tation about a single covalent bond is sufficiently hindered
such that different stereoisomers can be isolated.[1–3] Typical
examples are tetra-ortho-substituted diaryls. In these com-
pounds, the bulkiness of the substituents determines the
The use of chiral atropisomers in asymmetric synthesis has
successfully been investigated during the last two de-
cades.[1,2] Atropisomerism is a particular class of stereoiso-
height of the barrier to the rotation about the sp sp single
bond that corresponds to the activation energy for the inter-
conversion between the enantiomeric or diastereoisomeric
atropisomers. The ability to separate their signals using
chromatographic techniques (HPLC, GC) or NMR spectros-
copy is directly related to the exchange rate constant, that
itself depends on the magnitude of the barrier to rotation,
DH°, and the sample temperature, T.
2
2
ꢀ
merism involving molecular structures in which internal ro-
[a] Dr. O. Lafon, Dr. P. Lesot
Laboratoire de Chimie Structurale Organique
ICMMO UMR CNRS 8182
Bât. 410, UniversitØ de Paris-Sud, 91405 Orsay (France)
Fax : (+33)169-15-81-05
E-mail: philesot@icmo.u-psud.fr
NMR spectroscopy is exquisitely well-suited to the analy-
sis of dynamic phenomena (tautomerism, hindered rotation,
etc.). The method relies on the fact that dynamic processes
modify the resonance lineshapes when the exchange or in-
terconversion rate constant, k, is comparable to the magni-
[b] Dr. C.-A. Fan, Prof. H. B. Kagan
Laboratoire de Catalyse MolØculaire, ICMMO UMR CNRS 8182
Bât. 420, UniversitØ de Paris-Sud, 91405 Orsay (France)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
3772
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3772 – 3786

FULL PAPER
tude of the NMR interactions (in frequency units).[4,5] In iso-
tropic NMR methods, only the isotropic part of the chemical
shift or scalar coupling provides information on the dynamic
processes,[4,5] even though original approaches involving dif-
fusion NMR experiments (DOSY spectroscopy) have re-
cently been explored.[6] This restriction can limit the effi-
ciency of the method if the dynamic process does not suffi-
ciently affect NMR interactions and thereby modify the
NMR spectra. To overcome this problem, the use of NMR
spectroscopy in magnetically aligned media was proposed in
the early seventies.[7–9] In particular, proton-decoupled deu-
terium NMR (2H-{1H} NMR) spectroscopy of isotopically
enriched compounds embedded
chiral LCs, enantiomers or enantiotopic directions in prochi-
ral molecules are on average oriented differently and hence
their residual anisotropic magnetic interactions are likely to
be different.[15,16,18]
Very recently, we reported the first example of the analy-
sis of rotational isomerism phenomena using NMR spectros-
copy in CLCs.[19] In this work, we extensively explore the an-
alytical potential of 2H NMR spectroscopy in polypeptide
mesophases for the investigation of the intramolecular dy-
namic process of four deuterated diaryls derived from 1-(4’-
methylphenyl)naphthalene and denoted 1–4. The structures
and the numbering are given in Figure 1. From a stereo-
in liquid crystals (LCs) was suc-
cessfully applied to the analysis
of various types of dynamic
processes.[7,8]
Although
the
molecules
under investigation have to be
isotopically labelled, the analy-
sis of molecular interconversion
processes using 2H-{1H} NMR
spectroscopy of deuterated mol-
ecules in liquid crystals possess-
es several practical advantag-
es.[7,8] Firstly, the spectra are
free of background signals orig-
inating from the LC. Secondly,
the dynamic effects on the
NMR lineshapes and the coa-
lescence phenomena can be
readily interpreted because the
spectra are only dominated by
residual quadrupolar interac-
tions. Thirdly, the spectral sepa-
Figure 1. a) Enantiomeric conformers associated with the 1-bromo-3-deuterio-5-methyl-2-(1’-naphthyl)benzene
(1) and the 1-bromo-3-deuterio-2-methyl-5-(1’-naphthyl)benzene (2). The absolute configuration of each enan-
tiomer is defined by application of the Cahn–Ingold–Prelog priority rules.[3] The notation (a–j and a’–g’) ap-
plies to all compounds. b) Definition of the inter-ring torsion angle (b-a-a’-b’), f, used to calculate the poten-
tial energy profiles. c) Elements of symmetry for 1-(2’,6’-dideuterio-4’-methylphenyl)naphthalene (3) when the
2
rations between exchanging H
anisotropic signals can be much
larger than those observed in
ꢀ
dihedral angle, f, is equal to 908. In this case, the C D directions are enantiotopic. d) Enantiomeric rotamers
associated with 1-(2’-deuterio-4’-methylphenyl)naphthalene (4).
1
isotropic H or 13C NMR spec-
tra, thus allowing a much wider
range of dynamic processes to be studied.[7,8] Fourthly, the
chemical point of view, the monobromide aromatic deriva-
tives, 1 and 2, are chiral compounds for non-planar confor-
mations. On the other hand, the non-planar, di- and mono-
deuterated derivatives, 3 and 4, can be defined as prochiral
and chiral by virtue of the isotopic substitution H/D, respec-
tively.
2
simulation of dynamic H-{1H} NMR spectra requires only
the solution of Bloch–McConnell-type equations rather than
the full density matrix[9] and so calculations of the kinetic
rate constants and activation parameters are facilitated.
Until now such investigations were restricted to molecules
dissolved in achiral thermotropic nematics such as EBBA,
phase V or ZLI 2452.[9–12] In particular, the method was used
to study the ring inversion of perdeuterated (halo)cyclohex-
anes,[9,11] p-dioxane,[12] and cis-decalin.[10] However, this ap-
proach should basically be irrelevant in the analysis of inter-
conversions between enantiomers or enantiotopic directions
because their spectral discrimination is impossible in an
achiral LC. To overcome this problem, it is necessary to
In an introductory part, we will briefly present some as-
2
pects of dynamic H NMR spectroscopy in CLCs. Then the
discussion will be divided into three subsections. Firstly, we
will analyse the conformational behaviour of compounds 1–
4 using molecular modelling calculations (MMCs). Secondly,
2
the H NMR results in PBLG mesophases will be presented.
In particular, we will examine the influence of the position
of bromine and deuterium atoms on the barrier to rotation.
Thirdly, we will investigate the effect of organic co-solvents
on the conformational dynamics of these biaryls. For all
compounds, the activation parameters (DH°, DS° and
2
record H NMR spectra using chiral liquid crystals (CLCs),
such as those made of poly(g-benzyl-l-glutamate) (PBLG)
dissolved in chloroform or dichloromethane.[13–17] In these
Chem. Eur. J. 2007, 13, 3772 – 3786
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3773

P. Lesot et al.
DG°(T)) will be determined by 2H NMR spectroscopy in
CLCs or by other methods such as HPLC and subsequently
compared with those predicted by MMCs.
Theoretical Aspects
2
Enantiodiscrimination using H NMR in chiral liquidcrys-
tals: In deuterium (spin I=1) NMR spectroscopy using
liquid crystals as solvent, the partially averaged magnetic in-
teractions are dominated by the quadrupolar interac-
2
tion.[20,21] The H-{1H} NMR spectra of deuterated solutes or
of solutes at the natural abundance level consist of the sum
of quadrupolar doublets in accord with the number of non-
equivalent deuterons.[16] In CLCs, the existence of enantiose-
lective interactions leads enantiomers or enantiotopic direc-
ꢀ
tions (for instance C D bonds) in prochiral molecules to be
Figure 2. a) Schematic structures of 1D 2H NMR spectra of two deuter-
ons (A and B) undergoing mutual exchange in the three possible ex-
change regimes. We assume that the signs of DnAQ and DnQB are positive, j
DnAQ j>jDnBQ j, and jDnQA jꢀjDnQB j is smaller than jDnQA j and jDnBQ j.
b) Identical to a) but for DnQA >0 and DnBQ <0. The arrows indicate the
mutually exchanged transitions. The reduction of both DnQ values and
linewidths at high temperature mimics the decrease in orientational
order when T is increased.
or pro-R
oriented differently on average, and hence SR
¼
SS pro-S, where S is an order parameter [see Eq. (2)].[13,16]
Clearly, NMR analysis of these systems provides informa-
tion that is not accessible by NMR analysis in achiral LCs.
or
Enantiodifferentiation observed in H-{1H} NMR spectra is
2
based on differences between the quadrupolar splittings,
jDnAQꢀDnBQ j, expressed in Hz, where the subscripts A and B
correspond to the stereochemical descriptors R and S for
enantiomers and pro-R and pro-S for enantiotopic direc-
tions. In both cases, the quadrupolar splittings are given by
Equation (1),[14,15,21] where KD is the 2H quadrupolar cou-
ed symmetrically relative to the offset resonance frequency,
wo, irrespective of the sign of the quadrupolar splittings, as
shown in Figure 2. Readers interested in the theoretical for-
malism used in this work can refer to the Supporting Infor-
mation.
A or B
pling, S
is the internuclear order parameter associated
CꢀD
A or B
ꢀ
with the C D bond, given by Equation (2), and qCꢀD is the
[14,15]
ꢀ
angle between the C D axis and the magnetic field axis.
In practice, when SCAꢀD differs from SCBꢀD, then two sharp
2
quadrupolar doublets centred generally on the same chemi-
Computer simulation of the H NMR spectral lineshapes
2
cal shift (small H chemical shift anisotropy) are observed
and their fits with experimental resonances allow the kinetic
constant, k, at the various temperatures explored to be de-
termined. However, as the kinetic rate constant, the quadru-
polar splittings and the relaxation time of deuterons are
temperature-dependent, an accurate determination of k
from simulations requires a good assessment of the parame-
(Figure 2).[14,15] The relatively large magnitude of the KD pa-
rameter (160–185 kHz) contributes to the success of the
method.
3
2
DnAQ or B
¼
KDSACꢀoDr B
ð1Þ
A
B
*
ters Dw=2p(DnQ/2ꢀDnQ/2) and 1/T2 at each temperature
(see the Supporting Information).[7–9,24] Nevertheless, suffi-
3cos2 ðq
Þꢀ1
A or B
*
ciently accurate values of 1/T2 can be inferred from the line-
A or B
CꢀD
CꢀD
2
ð2Þ
S
¼ h
i
widths of analogous samples (at the same T) in which ex-
change is not observed (achiral mesophase). If the differ-
2
Dynamic NMR spectroscopy of spin I=1 nuclei andactiva-
tion parameters: The theory of dynamic H NMR spectros-
ence, Dw, is directly measured from H NMR spectra in the
2
slow exchange regime, the Dw values in the fast exchange
regime will be obtained by extrapolation of the evolution of
DnQ values in the slow exchange regime, as we will describe
in the discussion below.[12,24]
In a second step, the set of k parameters extracted from
simulated H NMR spectra is analysed by plotting an Eyr-
ingꢀs plot, namely the natural logarithm of kNh/RT against
1/T, before and after the coalescence temperature (see Fig-
ure 6b).[5a] Indeed, the dependence of k(T) on temperature
is theoretically described by Eyringꢀs Equation (3),[25,26]
where DG° is defined by Equation (4), R=8.31 JKꢀ1 molꢀ1,
N=6.02”1023 molꢀ1 and h=6.62”10ꢀ34 Js.
copy in oriented solvents has been described by various au-
thors.[7,22] For our purpose, we restricted the framework to
two non-coupled, non-equivalent deuterons in mutual inter-
conversion (denoted A and B) and undergoing only Zeeman
and quadrupolar interactions. In our case the two sites will
correspond to either enantiomers (A, BꢁR or S) or enan-
tiotopic directions (A, Bꢁpro-R or pro-S) and an exchange
process between two “equally populated sites” only is con-
sidered.[23] Owing to the absence of a difference in the
chemical shift between the exchanging sites (wA0 =woB =wo),
the positions of the two transitions (w0,1 and wꢀ1,0) are locat-
2
3774
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3772 – 3786

Intramolecular Dynamic Processes in Diaryl Atropisomers
FULL PAPER
ꢀ
ꢁ
RT
Nh
DG°
The evolution of Epot for 1 and 2 between 0 and 3608 is
displayed in Figure 3. For a simple comparison of data, the
maximum for each curve has been set to 0. Data relative to
these plots are listed in Table 1. For 1 and 2, the curves are
ð3Þ
ð4Þ
kðTÞ ¼
exp
ꢀ
RT
DG°ðTÞ ¼ DH°ꢀTDS°
Interestingly, the rate constant, k(Tc) (in sꢀ1), and the free
energy of activation at the coalescence temperature Tc,
DG°(Tc) (in kJmolꢀ1), can be rapidly deduced from the
measurement of the half-difference of values of jDnQ j
2
(jDDnQ/2j=jDnAQ/2ꢀDnQB/2j) in the H NMR spectra below
Tc.[27] Indeed, at this particular temperature and assuming
*
identical T2 for both exchanging deuterons, we can write
Equations (5) and (6).[2,3,5a,25]
2p  jDnAQ=2ꢀDnBQ=2j
ð5Þ
pffiffiffi
kðTcÞ ¼
2
2
pffiffiffi
ꢀ
ꢁ
RTc
Nh
2
DG°ðTcÞ ¼ RTc  ln
Â
ð6Þ
p  jDnAQ=2ꢀDnBQ=2j
Figure 3. Potential energy profiles of 1 and 2 versus the dihedral angle f
(0–3608). The maximum for each curve was set to 0.
Results andDiscussion
Conformational analysis of diaryl derivatives 1–4 by MMCs:
We have determined the evolution of the potential energy,
Epot, of 1–4 as a function of the inter-ring dihedral angle,
f(b-a-a’-b’), by MMCs (see Figure 1b). The calculations
were carried out with the Hyperchem 5.0 software package
using the semi-empirical AM1 force field.[28,29] More sophis-
ticated approaches such as ab initio (STO-3G) or DFT
methods[30] were not used because the relative energies ob-
tained with the AM1 and STO-3G methods for some specif-
ic values of f differed by less than 3%. Hence it was perti-
nent to plot Epot =f(f) by using the less time-consuming
method. Owing to the Born–Oppenheimerꢀs approximation,
the classical quantum-chemical software packages are not
able to take into consideration isotope effects and hence
they cannot distinguish between hydrogen and deuterium
atoms. To mimic the decrease in the average distance of a
Table 1. Data extracted from the potential energy profile analysis of
compounds 1–4.
[a]
[a]
Compound fmin
Epot(08)[b]
Epot(fmin
)
Epot(1808)[b] DH°[c]
[8]
[kJmolꢀ1
]
[kJmolꢀ1
]
[kJmolꢀ1
0
]
[kJmolꢀ1
]
1
2
83.1
ꢀ9.4
ꢀ87.5
+78.1
+46.3
57.5/ <ꢀ0.05
ꢀ46.30
122.5
3
4
57.5/ <ꢀ0.05
122.5
ꢀ45.9
ꢀ45.9
0
0
+45.9
+45.9
57.5/ <ꢀ0.05
122.5
[a] The minimal energy conformer was calculated by using a convergence
criterion equal to 0.012 kJmolꢀ1. [b] Each point of the curve was calculat-
ed with a convergence criterion equal to 0.04 kJmolꢀ1. [c] Here, we
define DH° as being equal to Epot(08)ꢀEpot(fmin).
ꢀ
ꢀ
C D bond, we have artificially shortened the C H distance
to 107.3pm at the site of isotopic substitution. [31]
symmetrical about f=1808 and the regions of the curve to
the two sides of the maximum correspond to the enantio-
meric conformers which may or may not be isolable, de-
pending on the energy difference, DH°, between the mini-
mum and the maximum potential energy. In the case of 1,
steric hindrance to rotation produces the highest energy
rotamer for f=1808, when the bromide interacts with the
aromatic hydrogen atom denoted as i (see Figure 1). Quanti-
tatively, the difference in Epot between the maximum-energy
planar rotamers obtained at f=0 and 1808 is equal to
ꢀ9.4 kJmolꢀ1. Hence, the enantiomeric interconversion pro-
cess is energetically favoured when the bromine atom inter-
acts with the aromatic hydrogen atom b (f=08). Conse-
quently, DH°(1) is defined as the difference between the
The results of the MMCs are discussed in two parts. First,
we compare the potential energy profiles of diaryl bromide
derivatives 1 and 2, which are both chiral with non-planar
geometries (see Figure 1a) but differ from each other by
virtue of the position of the bromine atom in the methyl-
phenyl group. The different positions of the bromine atom
in 1 and 2 should affect the height of the barrier to internal
ꢀ
rotation about the a a’ bond. Hence, the characteristics of
the H NMR spectra in chiral LCs should be significantly
2
modified in terms of chiral discrimination at a given temper-
ature, for instance, at room temperature. Secondly, we dis-
cuss the case of di- and monodeuterated derivatives 3 and 4
to analyse the effect of isotopic (H/D) substitution on con-
formational exchange.
energy of the conformer at f=08 and the minimum of Epot
.
Chem. Eur. J. 2007, 13, 3772 – 3786
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3775

P. Lesot et al.
In the case of 2, the energies of the rotamers at f=0 (or
3608) and 1808 can be considered to be equal
(<0.05 kJmolꢀ1) and hence no preferential planar confor-
mer exists. This behaviour indicates that the steric hindrance
due to the bromine atom in the meta position is negligible
compared with the steric hindrance of a bromine atom in
the ortho position and does not affect the barrier to rotation
compared with a non-substituted derivative.
the b’ and f’ positions are identical, the evolution of Epot for
3 versus f is symmetrical about f=908 and very similar to
the evolution of Epot for 2 (see Figure SI-2 in the Supporting
Information). In fact the regions to each side of this angle
correspond to enantiomeric pairs. In this particular case, it
should be emphasised that the conformers at f and 360ꢀf
are also mirror images and so the evolution of Epot between
0 and 3608 is also symmetrical about f=1808. Once again,
the regions to the two sides of the absolute maximum corre-
spond to enantiomeric pairs. Steric hindrance produces the
highest energy rotational conformers for f=0 (360) and
1808. Owing to the absence of stabilisation by p orbital over-
lap when the phenyl and naphthyl rings are orthogonal, the
lowest energy conformers are not found at 908 (see
Table 1).[3] The barrier to rotation, DH°, is evaluated as
+45.9 kJmolꢀ1, while the barrier to interconversion between
the enantiomeric conformer pairs, f1/f2 or f3/f4, is less than
+2 kJmolꢀ1. The slightly smaller value of DH° for 3 relative
to 2 (0.4 kJmolꢀ1) indicates that the bromine atom in the
meta position has only a very small effect on the height of
the barrier to rotation.
The inter-ring torsion angle corresponding to the lowest
energy conformation (fmin) results from the balance be-
tween the steric repulsion due to the ortho substituents and
the p-orbital overlap that produces maximum resonance sta-
bilisation when the aromatic rings are co-planar.[3] The Epot
curve for 1 is quite flat in the angular regions around 90 and
2708, with two minima located at f1(1)=83.18 and f2(1)=
276.98. This pattern of behaviour is generally observed when
strong steric repulsion exists between bulky ortho substitu-
ents.[3] For 2, the evolution of Epot shows four local minima
(f1(2), f2(2), f3(2) and f4(2)) at f=57.5 and 122.58 and the
complementary angles 302.5 and 237.58. The presence of
two local minima located symmetrically about 908 is due to
the effect of resonance stabilisation by p orbital overlap at
these four angles. This effect is measured at around
2 kJmolꢀ1. Finally, the large difference in activation energy,
DH°, between 1 and 2 (ꢂ32 kJmolꢀ1) originates from the
difference in steric hindrance due to the bulkiness of the
bromine atom (vdW radius=185 pm) relative to the hydro-
gen atom (vdW radius=120 pm).
Compared with 1 and 2, the dideuterated compound 3 is
symmetrical with two small-volume substituents on the
phenyl ring (two deuterons). From a chemical point of view,
this molecule can be defined as prochiral because if we are
able to dissymmetrise the phenyl ring through an appropri-
ate chemical reaction, we obtain a chiral compound for all
the dihedral angles f differing from 0 and 1808 (planar
structure).[3] From a stereochemical point of view, this mole-
cule is characterised by a plane of symmetry (s) when f=
908 (see Figure 1c).
As a result of dissymmetrisation of the phenyl ring, the
diaryl 4 does not possess a plane of symmetry for the con-
formation f=908 and it can be defined as a chiral derivative
by virtue of the isotopic substitution (H/D). This molecule
presents enantiomeric forms when f is different to 0 and
1808. In the past, some rigid and flexible chiral molecules
due to isotopic substitution have been investigated by 2H
NMR spectroscopy in polypeptide LCs, but derivatives ex-
hibiting a conformational exchange detectable by NMR
spectroscopy have not yet been explored.[32,33] In addition, it
is worthwhile comparing the results obtained for 3 and 4 in
2
order to analyse possible steric isotope effects on H NMR
spectra.[34] The MMCs carried out on 4 show that the poten-
tial energy profiles for 3 and 4 are identical. No deviation of
the f angles or of the DH° values for 3 and 4 occurs when
ꢀ
ꢀ
one of two C D bonds is replaced by a C H bond in the
structure.
Depending on the constant k, which is related to the mag-
nitude of DH° and T, the stereochemical relationship be-
NMR study of 1-bromo-3-deuterio-5-methyl-2-(1’-naphthyl)-
benzene (1): The 2H-{1H} NMR spectrum of compound 1
dissolved in PBLG/CHCl3 at room temperature shows two
quadrupolar doublets with a difference in quadrupolar split-
tings, jDnAQꢀDnBQ j, equal to 257 Hz (Figure 4a). As the abso-
lute configuration of each doublet is unknown, the stereode-
scriptors R and S were replaced by the arbitrary notations A
and B. The presence of two doublets suggests a priori a dif-
ꢀ
tween the two C D directions in 3 is not the same and so
gives different 2H-{1H} anisotropic spectra. Theoretically,
three stereochemical situations exist. Firstly, if the rotation
ꢀ
around the aryl aryl bond is fast, no spectral discrimination
ꢀ
between the C D directions is possible and so a single dou-
blet is expected. Secondly, if the rotation is sufficiently hin-
ꢀ
dered, the C D directions should be distinguished if their
ꢀ
orientational order parameters are sufficiently different in
ference in the orientation of the C D direction in each en-
the PBLG chiral phase. As the average conformation of 3 is
antiomer. To confirm this, we recorded the 2H-{1H} NMR
spectrum of 1 dissolved in an achiral polypeptide oriented
phase made of an equimassic mixture of PBLG and PBDG
(enantiomer of PBLG), both dissolved in CHCl3 (see
Table 4 below). In this mixture, denoted “PBG”, the spectral
discrimination of the enantiomers or enantiotopic directions
was eliminated.[35] Experimentally, the two doublets collapse
into a single doublet. This second step therefore shows un-
ambiguously that the enantiomers of 1 are spectrally dis-
ꢀ
of Cs symmetry, the C D directions can be considered as
enantiotopic.[3,16] In principle, two quadrupolar doublets are
2
pro-R
pro-S
expected in the H-{1H} spectrum if S
¼S
. Thirdly, if
CꢀDi
CꢀDi
the rotation is impossible and if the dihedral angle f is not
908 the molecule should adopt C1 symmetry and then two
ꢀ
enantiomeric forms exist in which the C D directions are
now diastereotopic. In this particular situation, four quadru-
polar doublets should be observed. As the substituents at
3776
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3772 – 3786

Intramolecular Dynamic Processes in Diaryl Atropisomers
FULL PAPER
does not modify the equilibri-
um constant in favour of one of
the two enantiomers.
In a second series of NMR
experiments, we recorded the
2H-{1H} NMR spectra of
1
versus T (224–350 K) to check
if the coalescence phenomenon
could be observed (see Fig-
ure 4b). The evolution of the
difference in quadrupolar split-
tings, jDnAQꢀDnBQ j, the average
linewidth (measured on the
four components) and the aver-
age
quadrupolar
splitting,
DnaQverage =jDnQA +DnBQ j/2, is pro-
vided in the Supporting Infor-
mation (see Figure SI-3). Anal-
ysis of the spectra shows that
enantiomers are differentiated
over the entire range of T. This
result also includes the case in
which one singlet and one dou-
blet are simultaneously ob-
served, as in the spectrum re-
corded at T=345 K (see the
inset in Figure 4b). This situa-
Figure 4. a) 61.4 MHz 2H-{1H} NMR spectra of 1 recorded at T=302 K in the PBG/CHCl3 achiral phase (top)
and in the PBLG/CHCl3 chiral phase (bottom). b) Graphical evolution of positive and negative components of
doublets A (outer) and B (inner) of 1 in the PBLG/CHCl3 phase versus T. The dashed line corresponds to an
experimental value of DnaQverage, namely the evolution that should be observed in the achiral mesophase assum-
ing the same sign for values of DnQ. The insert shows the spectrum recorded at 345 K.
ꢀ
tion arises when the C D inter-
nuclear axis of one of the two
enantiomers is aligned on aver-
age along the magic angle (q=
A
G
U
54.78) direction [Eq. (2)]. As a consequence, even at very
high temperatures, the coalescence phenomenon was not ob-
served.
lecular conformational effects or mesophase-related phe-
nomena). The magnitude of the doublet splitting in the
PBG mesophase (jDnPQBG j=250 Hz) implies that the sign of
the quadrupolar splittings in PBLG are identical, either pos-
itive or negative. Opposite signs would have led to a value
of around 130 Hz. Ideally, the anticipated value should be
equal to 225 Hz, but differences in the sample preparation
can yield these small differences (see below). Finally, note
that enantiomers of 1 have also been discriminated by
proton-decoupled 13C NMR spectroscopy.[14,36] Remarkably,
this example illustrates that NMR spectroscopy in CLCs can
be used to analyse chiral molecules without any stereogenic
centre and devoid of conventional reactive functionality for
which classical NMR tools are rather limited.[15]
As expected for any solute in a LC, the quadrupolar split-
tings increase monotonically as T decreases. This is a conse-
quence of the reduction in the mobility of molecules in the
mesophase as T is lowered. For a simple quantitative evalua-
tion of the evolution of jDnaQverage j, jDnQA j, jDnQB j or
jDnAQꢀDnBQ j versus T, we can calculate the relative spectral
¯
variation, DX/X, expressed in %, of these quantities be-
tween Tmin and Tmax by using Equation (7), where X repre-
sents jDnQaverage j, jDnQA j, jDnQB j or jDnQAꢀDnBQ j. From this
simple evaluation, we find that the relative spectral varia-
tions of jDnAQꢀDnQB j and jDnaQverage j between Tmin and Tmax
are equal to 38 and 160%, respectively. This large difference
in the relative variation of these quantities suggests that
chiral discrimination and orientation mechanisms are two
separate phenomena. This assumption is supported by other
results reported in this paper.
From a conformational analysis point of view, we show
here that the interconversion between enantiomers of 1 is
2
sufficiently slow on the H NMR timescale to allow separa-
Tmin
max
DX
XT ꢀX
tion of their deuterium signals at room temperature. This
result is consistent with the DH° value calculated by MMCs.
Within experimental error, integration of the peak areas of
A and B leads to identical values, thus indicating that the
mixture is still racemic in the chiral LC phase. This implies
that K=kR!S/kS!R =1 and thus shows that the mesophase
ð7Þ
½%Š ¼ 100 Â
ðXT þ XT Þ=2
ꢀ
max
min
X
The evolution of jDnQaverage j (and associated order parame-
ters) as well as jDnQA j and jDnBQ j versus T (in K) over a
range of around 130 K can be fitted using an exponential
Chem. Eur. J. 2007, 13, 3772 – 3786
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3777

P. Lesot et al.
ꢀ
function of the form given by Equation (8), where R=
8.314 JKꢀ1 molꢀ1 and the constants C and E are the parame-
ters to be evaluated. Values of C and E are expressed in Hz
and Jmolꢀ1, respectively, and are listed in Table 2.
nitude of DnQ for a given C D bond, and consequently the
associated SCꢀ order parameter, versus T can be related to
D
the “strength” of the interaction potential (van der Waals)
between this molecular direction and its environment. The
negative sign for E indicates an attractive potential between
ꢀ
the PBLG helices and the C D bond. In addition, the E
ꢀ
ꢁ
values (in kJmolꢀ1) obtained are comparable to energies as-
sociated with van der Waals interactions and are very close
to the thermal energies (RT),
E
RT½KŠ
DnQ½HzŠ ¼ C  exp ꢀ
ð8Þ
which vary from 1.771 (213K)
Table 2. C and E parameters for fitting the evolution of jDnaQverage j, jDnAQ j and jDnQB j using Equation (8).
to 2.910 kJmolꢀ1 (350 K) in this
study.[38] To confirm these re-
sults, further investigations on a
wide range of solutes are under-
way.
Entry
Compd
Polypeptide/
co-solvent
C prefactor
E factor
[kJmolꢀ1
jDnaQverage
Average
E factor
[kJmolꢀ1
c parameter
[Hz]
]
j
jDnaQverage
j
jDnaQverage
j
]
jDnAQ j/jDnBQ j
[a]
jDnAQ j/jDnBQ j
jDnAQ j/jDnBQ j
1
2
3
4
1
1
2
3
PBLG/CHCl3
PBLG/CH2Cl2
PBLG/CHCl3
PBLG/CHCl3
4.43
15.1/0.49
9.43
21.2/2.1
128.3
204.4/77.6
81.7
143.6/59.6
82.3
ꢀ9.866
ꢀ7.881/ꢀ13.635
ꢀ8.253
ꢀ10.758
ꢀ8.709
ꢀ3.585
ꢀ4.081
0.992
0.981/0.995
0.999
0.999/0.995
0.999
0.997/0.999
0.998
0.997/0.996
0.998
In contrast to jDnaQverage j, the
variation in jDnAQꢀDnQB j versus
T is relatively small and its evo-
lution can be fitted using a
linear function. In this example,
we found that jDnAQꢀDnQB j
[Hz]=ꢀ0.934T[K]+542.2, with
c=0.989. The existence of dis-
ꢀ7.347/ꢀ10.071
ꢀ3.583
ꢀ3.063/ꢀ4.108
ꢀ4.337
ꢀ3.493/ꢀ4.668
ꢀ4.442
5
6
3
3
PBG/CHCl3
PBLG/CH2Cl2
–
121.9
177.0/197.0
69.4
ꢀ2.272
ꢀ1.299
0.996
0.993/0.954
0.998
ꢀ2.275/ꢀ0.323
ꢀ4.585
tinct evolutions for jDnQaverage
j
7
4
PBLG/CHCl3
ꢀ3.986
and jDnAQꢀDnBQ j is somewhat
consistent with the large differ-
ences in the relative variations
of these quantities. Again, these
facts are new arguments in
135.8/66.4
ꢀ3.543/ꢀ4.429
0.992/0.999
8
4
PBLG/CHCl3
84.3[b]
ꢀ4.259[b]
–
0.999[b]
[a] Average E factor for jDnQA j and jDnQB j. [b] Corrected values for jDnaQverage j [see Eq. (9)].
The factor E is the same for fitting the evolution of jDnAQ j,
jDnBQ j or the positive, A(+) or B(+), and negative, A(ꢀ) or
B(ꢀ), resonances. In contrast, the prefactor C associated
with the positive and negative resonances is obtained by di-
viding the value of C obtained for DnAQ or DnQB by a factor of
+2 and ꢀ2, respectively. Except in one case (Table 2,
entry 6), we observe that the outer doublet, DnAQ, is charac-
terised by a higher C and a lower E than the inner doublet.
This situation is expected if the variations in DnAQ and DnQB
are rather similar between Tmin and Tmax. Equation (8) dif-
fers from Hallerꢀs equation, S=So(1ꢀT/T*)a, which is classi-
cally used to fit the evolution of the order parameters versus
T for nematic liquid crystals.[37] In fact, this equation cannot
be used to fit anisotropic parameters in the case of organic
solutions of PBLG because the nematic–isotropic transition
temperature (N!I) cannot be determined (due to the evap-
oration of co-solvent). As we will see below, this kind of
evolution has been ascertained using two liquid-crystalline
phases irrespective of the compound (1–4) (see Table 2). For
all of them, the factor E relative to the evolution of
jDnaQverage j is approximately the average of factors E deter-
mined for jDnAQ j and jDnQB j. Thus, the discrepancy between
the factors E obtained for 1 is less than 10%. This behav-
iour suggests that the enantiomeric discrimination in the
CLC may be described as a deviation from the average
order existing in the achiral oriented phase.
favour of the separation of the mechanisms for orientation
and chiral discrimination, which can be seen as distinct phe-
nomena and so treated separately.
The absence of coalescence signals at 350 K implies that
Tc is at a higher temperature and hence the activation pa-
2
rameters cannot be calculated using 1D H NMR spectra in
the PBLG/CHCl3 solvent. Another co-solvent with a higher
boiling point, such as DMF, could be used in order to record
2H NMR spectrum above T=350 K, but this change has not
been explored because the magnitude of jDnAQ/2ꢀDnQB/2j in
this new oriented solvent is a priori unknown. To estimate
the exchange rates, we have explored the use of 2D ex-
change spectroscopy (EXSY 2D experiments).[22,39] Howev-
er, irrespective of the mixing time used, no exchange cross-
peaks between the two quadrupolar doublets were observed
experimentally on the 2D map, very likely because the acti-
vation barrier is too high to be determined by this approach.
In fact, because Equations (4) and (6) are equal at Tc, we
may predict Tc of 1 by making some reasonable assump-
tions. By using the hypothesis in which the quantities
jDnAQ/2ꢀDnBQ/2j at T>350 K, DH° and DS° are known, we
can graphically resolve this equality to determine Tc and
DG°(Tc).[40] The magnitude of jDnQA/2ꢀDnBQ/2j above 350 K
can be extrapolated from the exponential fit of the entities
DnAQ and DnBQ (see Table 2, entry 1). If the DH° value can be
set to the value determined by the MMCs (+78.1 kJmolꢀ1),
we ignore the DS° value for 1. However, we can assume
that DS° is close to the data obtained for analogous com-
Remarkably, Equation (8) is similar to the Boltzmannꢀs
relation. In fact, we can infer that the evolution of the mag-
3778
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3772 – 3786

Intramolecular Dynamic Processes in Diaryl Atropisomers
FULL PAPER
Table 3. Comparison of the data evaluated by molecular modelling calculations and experimental parameters for compounds 2, 3 and 4.
Data evaluated from
molecular modelling
Data determined from NMR experiment in CLC
Compd.
f
[8]
DH°
Polypeptide/
co-solvent
Tc
[K]
jDDnQ/2j
k(Tc)[a]
DG°(Tc)[b]
[kJmolꢀ1
DH°[c]
DS°[c]
DG°(Tc)[c]
[kJmolꢀ1
[kJmolꢀ1
]
[Hz]
[sꢀ1
]
]
[kJmolꢀ1
]
[Jmolꢀ1 Kꢀ1
]
]
2
3
3
4
57.5/122.5 46.3PBLG/CHCl
250ꢃ2 175ꢃ5
245ꢃ2 102ꢃ5
245ꢃ2 106ꢃ5
245ꢃ2 98ꢃ5
3
ꢃ1289 48.5ꢃ0.5
44.7ꢃ0.5
44.2ꢃ0.5
43.6ꢃ0.5
44.3ꢃ0.5
ꢀ18ꢃ2
ꢀ19ꢃ2
ꢀ16ꢃ2
ꢀ20ꢃ2
49.0ꢃ0.5
48.9ꢃ0.5
47.6ꢃ0.5
49.2ꢃ0.5
3
57.4/122.6 45.9
57.4/122.6 45.9
57.4/122.6 45.9
PBLG/CHCl3
PBLG/CH2Cl2
PBLG/CHCl3
226ꢃ11 48.6ꢃ0.5
23 ꢃ11 5 48.5ꢃ0.5
218ꢃ11 48.7ꢃ0.5
[a] Values calculated from Equation (5). [b] Values calculated from Equation (6). [c] Values derived from the Eyringꢀs plot.
pounds (such as 2). By using this argument, the DS° value
was set to ꢀ19 Jmolꢀ1 Kꢀ1 (see Table 3). This approach leads
to
a
value for Tc of 415.5 K and DG°(Tc) of
+86.0 kJmolꢀ1.[41]
As an alternative to the NMR approach, we explored the
analytical potential of chiral HPLC. Thus it was possible to
separate the enantiomers of 1 by using a chiral stationary
phase (Chiralcel ODH column) at 298 K with hexane as
eluent (flow 0.5 mLminꢀ1). The chromatogram (not shown)
displays two sharp peaks (with retention time=35.3 and
46.6 min) separated by a rather flat “plateau”. These partic-
ular chromatograms are generally recorded for molecules
undergoing an interconversion between enantiomers during
their elution on a resolving enantiopure stationary phase.
The effect is referred to as “peak coalescence of the second
type”.[42] By using the technique developed by Schurig and
co-workers, it is possible to simulate the chromatogram of 1
in order to evaluate DG°(298 K)=+91.4ꢃ0.6 kJmolꢀ1.[43]
By assuming that DS° =ꢀ19 Jmolꢀ1 Kꢀ1, we can estimate
DH° to be +85.7 kJmolꢀ1 from Equation (4). The differ-
ence between the data determined by this approach and
that deduced by MMCs is about 10%. Although the DS°
term is generally small for intramolecular exchanges in
small molecules, the discrepancy in DH° can mainly be at-
tributed to the approximation made for DS°. Here we prob-
ably underestimate DS° because 1 exhibits a much greater
steric hindrance than other compounds. However, the rela-
tively good agreement between the two results is satisfacto-
ry. Another approach would involve evaluating DS° by com-
bining the value of DH° determined by MMCs and DG°
value derived by HPLC at 298 K. By using Equation (4) we
found DS° =ꢀ45 Jmolꢀ1 Kꢀ1. This value is two times higher
than those determined by NMR analysis for 2 or 3. By using
this new value, we can evaluate Tc for compound 1 to be
473K and DG°(Tc) to be +99.4 kJmolꢀ1. Reasonably, the
correct value of DS° is intermediate between the two values
Figure 5. a) Experimental 61.4 MHz 2H-{1H} NMR spectra of 2 dissolved
in the PBLG/CHCl3 phase recorded at different temperatures as indicat-
ed on the right side. b) Simulated spectra computed with the specific rate
constant, k, indicated on the right side.
the NMR timescale to prevent spectral enantiodiscrimina-
tion in PBLG at room temperature. This result is in agree-
ment with the smaller barrier to internal rotation in 2 rela-
tive to 1, as evaluated by MMCs (see Table 1). Spectrally,
this single quadrupolar doublet corresponds to the average
of the quadrupolar splittings arising from the R and S iso-
2
mers. In fact, this H NMR spectrum is identical to that ex-
pected in the achiral oriented solvent, PBG. This conclusion
will be illustrated for diaryl 3.
To observe the coalescence phenomenon for 2, we have
to reduce the interconversion rate by lowering the tempera-
ture. Figure 5a shows the experimental evolution of the ani-
2
sotropic H-{1H} NMR spectra of 2 versus T. For a better
visualisation, the evolution of DnQ values over the tempera-
ture range 224–350 K is shown in Figure 6a (see also Figure
SI-3b in the Supporting Information). While significant line-
broadening is observed below 270 K, the coalescence phe-
nomenon occurs at Tc =250ꢃ2 K. Here, the phenomenon
can be clearly identified because a single doublet with very
broad components (ꢂ130 Hz linewidth) is recorded. This
spectral lineshape indicates that the signs of the DnQ values
are the same for both enantiomers (as for 1) at low temper-
atures (see Figure 2a). Thus it is possible to determine the
ꢀ1
given above. By using the average value (ꢀ3 2 Jmol Kꢀ1),
Tc =442.5 K and DG°(Tc)=+92.3kJmol ꢀ1 are obtained.
NMR study of the 1-bromo-3-deuterio-2-methyl-5-(1’-naph-
thyl)benzene (2): In contrast to compound 1, the 2H-{1H}
NMR spectrum of 2 recorded at room temperature (302 K)
shows a single quadrupolar doublet with a splitting of
538 Hz (see Figure 5a). Consequently the interconversion
between the enantiomeric forms of 2 is sufficiently fast on
Chem. Eur. J. 2007, 13, 3772 – 3786
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3779

P. Lesot et al.
allows DG°(Tc) as well as other activation parameters to be
calculated more accurately from NMR results. This suppos-
es, however, that we are able to determine the rate con-
2
stants for each temperature by fitting the experimental H
NMR signals to the theoretical lineshapes [see Equa-
tions (SI-4) and (SI-5) in the Supporting Information].
Below Tc, the quantity jDnAQ/2ꢀDnBQ/2j is directly measured
from the experimental spectra (see Figure 5a). Above Tc,
this quantity has to be extrapolated from the evolution of
the A and B doublets observed in the slow exchange regime,
as if the coalescence phenomenon did not exist.[12,24] In this
hypothetical case, two doublets are expected owing to chiral
discrimination. Regarding the previous results obtained for
1, we have assumed that the four transitions, wAꢀ1,0, wꢀB1,0, w0A,1
and wB0,1, follow an exponential evolution over the entire
temperature range. Thus, it is possible to predict the fre-
quency evolution of these transitions versus T using Equa-
tion (8) and data extracted from the slow exchange regime.
The parameters C and E associated with DnAQ and DnQB
values are listed in Table 2 (entry 3). As an illustration, the
hypothetical evolution of the four components above Tc is
shown in Figure 6a (continuous lines). Thus, these extrapola-
tions permit the term jDnQA/2ꢀDnBQ/2j to be calculated in the
fast exchange regime.[24] By assuming that sufficiently accu-
*
rate values of 1/T2 can be extracted from the linewidths of
analogous samples for which no exchange process is ob-
served (PBG phase), this strategy allows the variables of
Equations (SI-4) and (SI-5) in the Supporting Information,
which are actually kept constant during the fitting proce-
dure, to be predicted.
Figure 6. a) Graphical evolution of the positive and negative components
of the doublets A (outer) and B (inner) of 2 dissolved in the PBLG/
CHCl3 phase versus T. The continuous lines represent the predicted evo-
lution of the components for both doublets versus T if no coalescence
phenomenon exists (see text). The dashed lines simulate the average evo-
lution of DnQ (derived from the predicted evolution) that should be ideal-
ly observed in the achiral oriented phase, assuming identical signs for the
values of DnQ. b) Eyringꢀs plot obtained for 2. The full line corresponds
to the linear fit of data (c=0.999).
As expected, the points on the Eyringꢀs plot (Figure 6b)
form a straight line, the slope and y intercept of which
permit DH° and DS°, respectively, to be determined.[5] Pa-
rameters derived from the Eyringꢀs plot are DH° =(+44.7ꢃ
0.5) kJmolꢀ1, DS° =(ꢀ18.0ꢃ0.5) Jmolꢀ1 Kꢀ1 and finally
DG° =(+49.0ꢃ0.5) kJmolꢀ1 at Tc. The discrepancy between
DG°(Tc) deduced from the Eyringꢀs plot and calculated
using Equation (6) mainly arises from the uncertainty in Tc.
The experimental DH° value differs by less than 2 kJmolꢀ1
(<4%) from the theoretical value determined by MMCs
(see Table 3), thus indicating the reliability of the two meth-
ods. The small magnitude of DS° is reasonable because the
transition state of small-to-medium-sized molecules under-
going an intramolecular exchange is not more significantly
ordered (from the entropic point of view) than the ground
state.[5a] Under these conditions, the entropy change, DS°,
between the ground state (the lowest energy rotamer) and
the transition state (the highest energy rotamer) is rather
small or even close to zero.[5a] Also we found that the rota-
tion rate is around 1.3”104 sꢀ1 at 302 K, thus implying that
spectral enantiodiscrimination could be observed at this
temperature only if jDnQA/2ꢀDnBQ/2j was much larger than
5850 Hz [Eq. (5)]! Finally, the excellent agreement between
the predicted and experimental DH° values indicates that
the intermolecular forces between the solute and the PBLG
phase do not significantly influence the conformational in-
terconversion process of 2 (compared with an isolated mole-
relative sign of the DnQ values without recording data in the
PBG mesophase. Below Tc, two resolved quadrupolar dou-
blets, the linewidths of which decrease with T (until 224 K),
are observed, thus demonstrating that 1) the average orien-
tation of each enantiomer is different in the mesophase and
2) their lifetime is now sufficiently long on the NMR time-
scale to observe their respective NMR signal. This notable
result highlights the ability of the polypeptide helices to dis-
tinguish between enantiomers of axially chiral derivatives
even if the barrier to rotation is not enough to discriminate
them at room temperature.
Although the half-difference of the DnQ values is temper-
ature-dependent, its evolution is very weak below Tc. Conse-
quently, by adopting an average value for jDnAQ/2ꢀDnQB/2j
equal to (175ꢃ5) Hz, we can evaluate k(Tc) to be (389ꢃ
12) sꢀ1 and DG°(Tc) to be (+48.5ꢃ0.5) kJmolꢀ1 for 2 using
Equations (5) and (6) and Tc =(250ꢃ2) K. The analysis of
Eyringꢀs plot over a sufficient temperature range around Tc
3780
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3772 – 3786

Intramolecular Dynamic Processes in Diaryl Atropisomers
FULL PAPER
cule). This new study confirms other recent results involving
complex calculations.[44]
From the point of view of orientational behaviour, the
comparison between the evolution of the spectral data
versus T for 1 and 2 is noteworthy (see Figure 4b and Fig-
ure 6a). Indeed, as the main directions of electrical field gra-
ꢀ
dient (EFG) tensor for the C D bonds in 1 and 2 are
almost parallel, they provide the same information in terms
of order parameters relative to the magnetic field axis and
hence it is possible to interpret the differences observed.
Over the temperature range explored (224–350 K), the
global evolution of jDnaQverage j and jDnQA/2ꢀDnQB/2j (partly
obtained by extrapolation) for 2 is similar to the evolution
observed for 1, but differences in their respective variations
exist. If the relative variation of jDnAQꢀDnQB j for 2 and 1 is
similar (29 instead of 38%), the relative variation of
jDnaQverage j is very different (67 instead of 160%). From a
theoretical point of view, these results suggest that the chiral
discrimination mechanisms for 2 and 1 are similar over a
large temperature range and lead to spectral separation of
similar magnitude. Thus the values of jDnAQꢀDnQB j for 2 and
1 are equal to 314 and 353 Hz, respectively, at 213 K, and
204 and 266 Hz at 350 K. In other words, the position of the
bromine atom has only a small effect on the spectral separa-
tion, thus suggesting that the discrimination mechanisms are
more sensitive to a global shape recognition rather than to
local changes in the molecular structure. This conclusion
confirms recent work on the discrimination of chiral hydro-
carbons that indicated that shape recognition plays a crucial
role in the chiral discrimination mechanisms of apolar enan-
tiomers.[15] In contrast, the molecular orientation mecha-
nisms seem to be more sensitive to the global geometry of
the solute since the position of the bromide on the phenyl
ring significantly affects the magnitude of jDnaQverage j. This
result could be related to the fact that these mechanisms are
more sensitive to electrostatic intermolecular interactions,
for which the position of the bromine atom plays an impor-
tant role. This analysis of the results supports again the idea
that the chiral discrimination and orientation phenomena do
not follow the same laws and could be treated using differ-
ent interaction mechanisms.
Figure 7. Experimental 61.4 MHz 2H-{1H} NMR spectra of 3 dissolved in
a) the PBLG/CHCl3 phase and b) the PBG/CHCl3 phase at different tem-
peratures as indicated. Note the difference in linewidth at 245 K.
Tc indicates two distinct averages for the orientation of the
pro-R
C D directions (SCꢀD ¼SpCrꢀo-DS). To confirm this result, we re-
ꢀ
corded the H-{1H} NMR spectra of 3 in the achiral phase
2
PBG/CHCl3 over the same temperature range (see Fig-
ure 7b and SI-5b in the Supporting Information). A single
pro-R
pro-S
doublet (S
=S
) was observed irrespective of the tem-
CꢀD
CꢀD
perature. This last result shows that the two doublets ob-
served at low T in PBLG do not originate from a particular
reorientation phenomenon of the solute, but arise from the
ꢀ
orientational non-equivalence of the C D directions in the
chiral mesophase. Thus we assess experimentally the enan-
tiotopic character of these directions when the rotation is
slow relative to the NMR timescale and so demonstrate that
the molecular structure resulting from averaging of all ro-
tamers has Cs symmetry.
By using an averaged value of jDDnQ/2j=102ꢃ5 Hz, we
calculated k(Tc)=(226ꢃ11) sꢀ1 and DG°(Tc)=(+48.6ꢃ
0.5) kJmolꢀ1 from Equations (5) and (6). By applying the
procedure described above for 2, the k values extracted
2
from each simulated H-{1H} NMR spectrum were analysed
using the Eyringꢀs plot. The evolution of the average of the
DnQ values (dashed lines) extrapolated from data measured
in the slow regime correctly fits the evolution of jDnQaverage
j
(measured below and above Tc) from 213to 305 K, but di-
verges slightly at higher temperatures (see Figure SI-4 in the
Supporting Information). The discrepancy measured at
350 K is 38 Hz, that is, an error of around 10%. This effect
is mainly related to the small number of points used for fit-
ting data during the slow exchange regime (five points). As
the temperature range used to evaluate k varies from 224 to
302 K, this divergence has only a small impact on the accu-
racy of k. Parameters derived from the Eyringꢀs plot are
given in Table 3. Once again the experimental DH° value
differs by less than 2 kJmolꢀ1 (<4%) from the value ob-
tained by MMCs. The rotation rate calculated at 302 K (k=
1.2”104 sꢀ1) is very similar to the value found for 2, thus
confirming that the bromine atom in the meta position does
NMR study of 1-(2’,6’-dideuterio-4’-methylphenyl)naphtha-
lene (3): With regard to the MMC results, the spectral be-
haviour of 3 should be a priori very similar to that of 2 in
spite of their stereochemical differences. Figure 7a shows
2
some anisotropic H-{1H} NMR spectra of 3 recorded in the
PBLG phase between 213and 350 K (see also Figure SI-4
and SI-5a in the Supporting Information). Similarly to 2, we
observe a sharp quadrupolar doublet (linewidth<5 Hz) at
302 K and above, while the coalescence phenomenon is
reached at Tc =(245ꢃ2) K. This Tc value is slightly lower
than that for 2 because the difference in DnQ values for the
2
2
ꢀ
ꢀ
C D directions is now equal to 204 instead of 348 Hz. As
not hinder rotation about the sp sp bond.
before, the coalescence lineshape corresponds to the case in
which the quadrupolar doublets have identical signs at low
temperatures. The presence of two resolved doublets below
As in preceding examples, the evolution of jDnaQverage j for
3 dissolved in the PBLG and PBG phases can be fitted with
good agreement by using Equation (8) (c>0.997). Compari-
Chem. Eur. J. 2007, 13, 3772 – 3786
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3781

P. Lesot et al.
son of the results obtained in the two phases indicates that
parameters C and E differ by less than 1 and 3%, respec-
tively (see Table 2, entries 4 and 5). These minor discrepan-
cies arise mainly from small differences in the sample prepa-
ration (slight changes in the composition) as well as in dif-
ferences in the average degree of polymerisation (DP) of
the polypeptides used (see Table 4). As in the case of 1, the
evolution of the difference of DnQ values is linear below Tc.
Again, we confirm the assumption that mechanisms for ori-
entation and discrimination are not the same. In addition, it
appears that the discrimination phenomenon could be seen
as a simple perturbation of the orientation phenomenon. If
this was not the case, the E fitting factors associated with
the evolution of jDnAQ j, jDnQB j and jDnQaverage j (in the PBG
or PBLG phase) versus T should be very different. Except
for the cases in which the angular terms qACꢀD and qCBꢀD are
ciently to induce differences in the activation parameters
larger than experimental error. The evolution of jDnQaverage
j
versus T for 4 (dashed line) and 3 (continuous line) is plot-
ted in Figure 8. As can be seen, small differences exist, in
i
i
rather close to the magic angle, the above statements could
explain why the signs of the quadrupolar splittings for
ꢀ
mono
A
tions are generally identical.
Figure 8. Comparison of jDnaQverage(3)j and jDnaQverage(4)j with and without
corrections. Note the excellent fit between values of 4 (after correction)
and 3.
Experimentally, spectral discrimination of the enantiomer-
ic rotamers (f1/f2 and f3/f4) was not observed in the spectra
at low temperatures, even by recording the H NMR spec-
2
trum at 200 K (not shown). Below this temperature, the
sample starts to form a gel phase, thus preventing the use of
PBLG as a weakly ordering liquid crystal with a low viscosi-
ty. This result is in accord with the MMCs since the height
of the barrier between the most populated enantiomeric
conformers, f1/f2 or f3/f4, is only around 2 kJmolꢀ1. Thus, if
we assume that jDDnQ/2j could vary between 50 and
110 Hz, we can predict that the second coalescence phenom-
enon should be observed between 11.2 and 11.6 K! These
temperatures are technically impossible to reach with cur-
rent NMR probes and is evidently not applicable to a
PBLG sample. Finally, note that the evolution of the spectra
could also be explained by assuming a potential energy pro-
file with only just two minima set at 90 and 2708, corre-
sponding to a prochiral conformer of Cs symmetry.
particular, at high temperatures. This effect primarily results
from minor differences in the sample preparation or the
macroscopic homogeneity of the phase. To correct these ef-
fects and obtain a more reliable comparison, a correction to
jDnaQverage(4)j can be applied by weighting each value with a
parameter that is directly related to the liquid-crystalline
phase. With this aim, we used the 13C–1H dipolar coupling,
1DCH, of the organic co-solvent (here, chloroform), which
can be measured from the signals from 13C satellites in the
1H NMR spectra.[15] The correction applied to jDnQaverage(4)j
is given by Equation (9).
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
1
DCHð3Þ
DCHð4Þ
jDncQorrected averageð4Þj ¼ jDnQaverageð4Þj Â
ð9Þ
ꢂ 1
NMR study of 1-(2’-deuterio-4’-methylphenyl)naphthalene
(4): The evolution of the 2H-{1H} NMR spectra (quadrupolar
splitting, linewidths) for compound 4 versus T is very similar
to that observed for 3 (see Figure SI-6 in the Supporting In-
formation). The coalescence phenomenon is observed at
(245ꢃ2) K and the average value of jDDnQ/2j measured
below Tc (on three points) is equal to (98ꢃ5) Hz (instead of
102 Hz). This small deviation can be attributed to small dif-
ferences in the sample composition, but also to the line-
widths measured at Tc (around 25 Hz). By using these
values, we obtain k(Tc)=(218ꢃ10) sꢀ1 and DG°(Tc)=
(+48.4ꢃ0.5) kJmolꢀ1. The activation parameters derived
from the Eyringꢀs plot analysis are listed in Table 3. Within
experimental error, we can conclude that the activation pa-
rameters for 3 and 4 are identical. This means that the con-
formational dynamics are the same, or at least, the replace-
ment of deuterium by the hydrogen atom does not modify
the conformational behaviour of 4 compared with 3 suffi-
Applying this correction harmonises the two sets of data.
Quantitatively, differences of 3and 2% in the C and E pa-
rameters, respectively, are found when the evolution of
jDnaQverage(3)j and jDnaQverage(4)j, after corrections, are fitted
using Equation (8) (see Table 2, entries 4 and 8). From an
orientational point of view, we have thus demonstrated that
the origin of the discrimination of enantiomers of chiral
molecules by virtue of isotopic substitution (H/D) is a con-
ꢀ
sequence of the discrimination of enantiotopic C D direc-
tions in related, prochiral molecules.
Influence of the co-solvent on the conformational dynamics:
As an extension of this study, we investigated the influence
of the nature of the organic co-solvent on the orientational
behaviour and the intramolecular dynamic processes of
compounds 1 and 3. For this purpose, we replaced chloro-
2
form by dichloromethane and recorded the H NMR spectra
3782
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3772 – 3786

Intramolecular Dynamic Processes in Diaryl Atropisomers
FULL PAPER
over the maximal temperature range 213–344 K. It is diffi-
cult to heat the PBLG sample in dichloromethane above
345 K owing to the boiling point of this co-solvent. Figure 9a
quadrupolar doublets below Tc. By using an average value
of jDDnQ/2j=(106ꢃ5) Hz, we obtain k(Tc)=(235ꢃ11) sꢀ1
and DG°(Tc)=(+48.5ꢃ0.5) kJmolꢀ1. As before, the evolu-
tion of the average of the DnQ values extrapolated from
data recorded in the slow exchange regime (dashed line) fits
the evolution of jDnaQverage j (measured below and above Tc)
with an error <5% between 213and 305 K, but starts to di-
verge for temperatures above 305 K. At 340 K, the discrep-
ancy is less than 40 Hz, that is, an error of 14%. The activa-
tion parameters derived from the Eyringꢀs analysis are in-
cluded in Table 3. Although the DG°(Tc) value is slightly
smaller than other DG°(Tc) values, we estimate that the de-
viation between the activation parameters obtained in the
PBLG/CHCl3 and PBLG/CH2Cl2 phases is not sufficient to
confirm that the nature of the co-solvent affects the dynam-
ics of the compounds investigated here.
As before, the evolution of jDnaQverage j for both compounds
can be fitted using Equation (8). The C and E factors are
listed in Table 2 (entries 2 and 6). The E factor, which re-
flects the decrease in the exponential function, is globally
reduced by around 17% for 1 and 48% for 3 compared with
values in the PBLG/CHCl3 mesophase. These findings are
consistent with the relative variations in jDnaQverage j, which
are equal to 131% for 1 and 44% for 3 between Tmin
=
224 K and Tmax =344 K instead of 156 and 83% measured in
the PBLG/CHCl3 phase. Note also that there is only a small
variation in DnBQ (Figure 9b), but it can still be fitted with an
exponential function with a very small E factor, while the
evolution of DnAQ shows a significant variation between Tmin
and Tmax. The differential spectral evolution between the
inner and outer doublets results from a different evolution
ꢀ
of orientational behaviour between the enantiotopic C D
directions. This evolution, not observed for the other sam-
ples, is rather difficult to interpret because we probe only
two directions of the five independent internuclear direc-
tions required to determine correctly Saupeꢀs matrix.[20,21]
Figure 9. a) Evolution of the positive and negative components of dou-
blets A (outer) and B (inner) of 1 in the PBLG/CH2Cl2 phase versus T.
b) As for a), but for compound 3. See the caption to Figure 6a for a defi-
nition of the curves and graphical points.
ꢀ
Put simply, we can assume that the C D direction associat-
ed with DnBQ tends to reorient towards the magnetic field
axis as T increases, thus balancing the global decrease in the
molecular order parameters due to faster molecular mo-
tions.
and 9b show the spectral evolution of the positive and nega-
tive components of the quadrupolar doublets of 1 and 3 as a
function of temperature. In both cases, the global evolution
of the DnQ values is identical to that obtained when chloro-
form was employed as co-solvent. Thus for 3, the coales-
cence effect is visible at (245ꢃ2) K, while no effect is ob-
served for 1 in the temperature range explored. By using
the same procedure as described above for evaluating Tc for
1 dissolved in the PBLG/CHCl3 phase and by setting DS° to
(ꢀ32ꢃ2) Jmolꢀ1 Kꢀ1, we evaluate Tc and DG°(Tc) to be
442.5 K and +92.3kJmol ꢀ1, respectively. These data are
similar to those obtained in PBLG/CHCl3, thus reflecting
the fact that the values of jDDnQ/2j are small and similar in
both media at high temperatures. Similarly to preceding ex-
amples, the coalescence phenomenon for 3 is characterised
by a single doublet with broad components and then by two
Conclusion
2
Herein, we have demonstrated the robustness of H NMR
spectroscopy in weakly ordering, chiral LCs for the investi-
gation of the intramolecular dynamic processes of new
diaryl atropisomers and related prochiral derivatives. This
approach is a powerful method for calculating the kinetic
and activation parameters of enantiomers or enantiotopic
directions undergoing intramolecular conformational ex-
change because their spectral discrimination is much more
efficient in an oriented polypeptide solvent compared with
isotropic NMR methods dedicated to chiral analysis.
Beyond the advantages offered by this analytical method,
the interpretation of these new experimental results has en-
Chem. Eur. J. 2007, 13, 3772 – 3786
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3783

P. Lesot et al.
hanced our understanding of the orientation and chiral dis-
crimination mechanisms in polypeptide LCs. First, we have
demonstrated the excellent agreement between the theoreti-
cal evaluation of DH° and experimental values extracted by
NMR analysis using two different organic co-solvents. This
result explicitly indicates that the intermolecular forces be-
tween the solute and the polypeptide oriented phase do not
significantly influence the conformational interconversion
process of molecules undergoing an intramolecular dynamic
process (compared with an isolated molecule). The study of
a possible deviation of activation parameters by changing
the nature of polypeptide used (poly(g-ethyl-l-glutamate) or
poly(e-carbobenzyloxy-l-lysine)) is currently underway. Sec-
ondly, in the temperature range explored, it appears that the
evolution of DnQ values (in Hz) versus T (in K) can be
fitted using an exponential function of the form DnQ =C”
exp(ꢀE/RT), whereas the evolution of the difference in DnQ
values between enantiomers or two enantiotopic directions
is linear. This situation suggests (at least for the compounds
investigated here) that the mechanisms of orientation and
discrimination could be treated separately because the con-
tributions of shape recognition and purely electrostatic inter-
actions differ in both mechanisms. Hence they can be seen
as uncorrelated phenomena. With regard to these results, it
seems that chiral discrimination might be seen as a simple
perturbation of the orientation mechanism. This assessment
is important for a better modelling of the PBLG system. It
is also important to establish if the exponential function
used for predicting the evolution of DnQ values versus T can
be applied to a wide range of solutes. The question is now
open.
volving no practical methodologies used in chemical labora-
tories, but to be related only to the stability scale of the vari-
ous possible isomers of a given compound.
Experimental Section
Synthesis of deuterated compounds: Scheme 1a and 1b show the synthet-
ic strategies leading to compounds 1–4. The detailed description of the
syntheses is reported in the Supporting Information.
Finally, this work points out a recurrent semantic problem
between the communities of organic chemists and spectro-
scopists about the definition of the intrinsic stereoisomeric
nature (chiral, prochiral)[3] of a chemical species and how a
molecule should be considered. This problem (already
pointed out by several authors in various contexts[16,45–47]
)
arises with the compounds investigated here, for which the
stereoisomeric nature of the molecules can be discussed as a
function of the temperature of the experiment. It raises the
basic question as to whether the classical definition of ster-
eoisomerism or pro-stereoisomerism should be imposed by
the reality of synthetic chemistry, based either on the possi-
bility of isolating a molecule[3,45,46] or disposing of a chemical
reaction able to synthesise enantiomers from a prochiral
molecule,[16,47] or the reality of the spectroscopic methods,
which depend on the experimental conditions or the obser-
vational technique itself.[46] With regard to these two points
of view, the definition introduced by Eliel[46] in which “a
chemical species” is defined “as a molecule at or very near
a minimum in a potential energy hypersurface” seems a
priori to be more satisfactory because this concept is inde-
pendent of the method of observation and experimental
conditions as far as possible. Under this condition, two
chemical species are identical when the interconverting
energy barrier is less than RT and (stereo)isomeric when the
barrier is higher. This definition offers the advantage of in-
Scheme 1. a) Synthesis of compounds 1, 3 and 4. b) Synthesis of com-
pound 2.
Practical aspects of NMR spectroscopy in orientedsolvents : The PBLG
and PBDG polymers are commercially available from Sigma Corp. The
preparation of fire-sealed oriented samples can be found in referen-
ces [14] and [15]. The samples were made by mixing the solute (ca.
15 mg), polypeptide (ca. 100 mg) and co-solvent (ca. 440 mg). The exact
composition of each sample is listed in Table 4. The 2H-{1H} NMR spec-
tra were recorded on a 400 MHz Bruker Avance spectrometer equipped
with a 5 mm QXO probe. The temperature was controlled by a BVT
3200 variable-temperature controller over the range of 200 (ꢀ738C) to
350 K (+778C). Note that the sample homogeneity is destroyed by boil-
3784
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3772 – 3786

Intramolecular Dynamic Processes in Diaryl Atropisomers
FULL PAPER
Table 4. Composition of the liquid-crystalline NMR samples investigated.
[16] a) C. Aroulanda, D. Merlet, J.
Courtieu, P. Lesot, J. Am. Chem.
Soc. 2001, 123, 12059—12066;
b) P. Lesot, D. Merlet, M. Sarfati,
J. Courtieu, H. Zimmermann, Z.
Luz, J. Am. Chem. Soc. 2002, 124,
10071–10082.
[17] C. Aroulanda, M. Sarfati, J. Cour-
tieu, P. Lesot, Enantiomer 2001,
6, 281–287.
[18] Note that the case of chiral, sub-
stituted cyclohexanes is very pe-
culiar because the exchange rate
between enantiomers can be de-
termined without discriminating
the enantiomeric forms. Indeed
for these compounds, the inter-
NMR
sample
Solute
Polymer
DP[a]
Co-solvent Solute Polymer Co-solvent
Amount of
polymer [wt%]
[mg][b]
[mg][b]
[mg][b]
1
2
3
4
5
6
7
8
1
1
1
2
3
3
3
4
PBLG
PBLG/PBDG 782/914
782
CHCl3
CHCl3
CH2Cl2
CHCl3
CHCl3
CHCl3
CH2Cl2
CHCl3
15
12
10
12
11
10
10
11
101
51+53440
107
101
101
50+51
99
99
440
18.2
18.7
PBLG
PBLG
PBLG
782
782
782
442
440
441
439
441
442
19.1
18.3
18.3
18.4
18.0
18.0
PBLG/PBDG 782/914
PBLG
PBLG
782
782
[a] DP: average degree of polymerisation of the polypeptide used (PBLG and PBDG). [b] To an accuracy of
ꢃ1 mg.
ing the co-solvent above 350 K for CHCl3 and above 344 K for CH2Cl2.
The temperature was calibrated using standard procedures. Mixtures
were left for 10–15 min to equilibrate at the sample temperature before
recording spectra. All 1D 2H NMR spectra were recorded with 1024
scans of 2048 data points. The tuning and matching of the deuterium coil
were optimised at each temperature. The classical WALTZ-16 sequence
was used to decouple protons (<0.4 W of power).
conversion between enantiomers also leads to the interconversion of
diastereotopic directions (axial and equatorial bonds) that are none-
quivalent in achiral LCs.
[19] P. Lesot, O. Lafon, C.-A. Fan, H. B. Kagan, Chem. Commun. 2006,
389–391.
[20] C. Zannoni in Nuclear Magnetic Resonance of Liquid Crystals (Ed.:
J. W. Emsley), NATO, Dordrecht, 1985, pp. 1–31.
[21] J. W. Emsley, J. C. Lindon, NMR Spectroscopy Using Liquid Crystal
Solvents, Pergamon Press, Oxford, 1975.
[22] a) B. Halle, Prog. Nucl. Magn. Reson. Spectrosc. 1996, 32, 137–159;
b) Z. Luz in NMR of Ordered Liquids (Eds.: E. E. Burnell, C. A. de
Lange), Kluwer, Dordrecht, 2002, pp. 419–448.
Acknowledgements
[23] M. H. Levitt, Spin Dynamics, Wiley, 2005, Chichester, pp. 492–494.
[24] E. Gelerinter, Z. Luz, R. Poupko, H. Zimmermann, J. Chem. Phys.
1990, 92–93, 8845–8850.
[25] a) J. A. Pople, W. G. Schneider, H. J. Bernstein, High Resolution
NMR, McGraw-Hill, New York, 1959, pp. 218–230; b) G. Binsch,
Top. Stereochem. 1968, 3, 97–192.
The authors are grateful to Prof. Dr. V. Schurig and Dr. O. Trapp for cal-
culating the free energy of activation from the HPLC study. Also they
cordially thank Dr. R. Paugam for stimulating discussions. O.L. and C.-
A.F. acknowledge MNESER for a Ph.D. grant and the CNRS of Gif-sur-
Yvette for their financial support, respectively.
[26] H. Eyring, Chem. Rev. 1935, 35, 65–77.
[27] This approach is called the “coalescence temperature method”.
[28] M. J. S. Dewar, K. M. Dieter, J. Am. Chem. Soc. 1986, 108, 8075–
8086.
[29] HyperchemTM, Professional 6.1, Hypercube, Inc, 1115 NN 4th
street, Gainsville, Florida, 32601 (USA).
[1] P. Llyod-Williams, E. Giralt, Chem. Soc. Rev. 2001, 30, 145–157, and
references therein.
[2] M. Oki, Top. Stereochem. 1983, 14, 1–82, and references therein.
[3] A. L. Eliel, A. H. Wilen, Stereochemistry of Organic Compounds,
Wiley, New York, 1994.
[30] a) R. F. Stewart, J. Chem. Phys. 1970, 52, 431–438; b) L. J. Sham, W.
Kohn, Phys. Rev. 1965, 140, 1697–1705.
[4] H. S. Gutowsky, D. W. McCall, C. P. Slichter, J. Chem. Phys. 1953,
21, 279–292.
[31] a) K. L. Servis, R. L. Domenick, J. Am. Chem. Soc. 1986, 108, 2211–
2214; b) S. Berger, H. Kꢂnzer, Tetrahedron 1983, 39, 1327–1329;
c) G. S. Pawley, E. A. Yeats, Acta Crystallogr. Sect. B 1969, 25,
2009–2013.
[32] A. Meddour, I. Canet, A. Loewenstein, J.-M. PØchinØ, J. Courtieu, J.
Am. Chem. Soc. 1994, 116, 9652–9656.
[5] a) A. D. Bain, Prog. Nucl. Magn. Reson. Spectrosc. 2003, 39, 63 –
103, and references therein; b) S. Grilli, L. Lunazzi, A. Mazzanti, M.
Pinamonti, Tetrahedron 2004, 60, 4451–4458; c) D. Casarini, C. Co-
luccini, L. Lunazzi, A. Mazzanti, J. Org. Chem. 2005, 70, 5098–5102.
[6] a) C. S. Johnson, J. Magn. Reson., Ser. A 1993, 102, 214–218; b) E. J.
Cabrita, S. Berger, P. Bräuer, J. Kärger, J. Magn. Reson. 2002, 157,
124–131; c) P. Thureau, B. Ancian, S. Viel, A. ThØvand, Chem.
Commun. 2006, 200–202.
[33] P. Lesot, O. Lafon, J. Courtieu, P. BerdaguØ, Chem. Eur. J. 2004, 10,
3741–3746.
[34] L. S. Bartell, J. Am. Chem. Soc. 1961, 83, 3567–3571.
[35] C. Canlet, D. Merlet, P. Lesot, A. Meddour, A. Loewenstein, J.
Courtieu, Tetrahedron: Asymmetry 2000, 11, 1911–1918.
[36] a) P. Lesot, D. Merlet, A. Meddour, A. Loewenstein, J. Courtieu, J.
Chem. Soc., Faraday Trans. 1995, 91, 1371–1375; b) A. Meddour, P.
BerdaguØ, A. Hedli, J. Courtieu, P. Lesot, J. Am. Chem. Soc. 1997,
119, 4502–4508.
[37] a) I. Haller, Prog. Solid State Chem. 1975, 10, 103–118; b) H.
Kneppe, V. Reiffenrath, Chem. Phys. Lett. 1982, 87, 59–62; c) M. L.
Magnuson, B. M. Fung, J.-P. Bayle, Liquid crystals 1995, 19, 823–
832.
[7] R. Poupko, Z. Luz in Encyclopedia of Nuclear Magnetic Resonance
(Eds.: G. M. Grant, R. K. Harris), Wiley, Chichester, 1996, pp. 1783–
1797, and references therein.
[8] Z. Luz in Nuclear Magnetic Resonance of Liquid Crystals (Ed.: J. W.
Emsley), NATO, Dordrecht, 1985, pp. 315–340.
[9] R. Poupko, Z. Luz, J. Chem. Phys. 1981, 75, 1675–1681.
[10] C. Boeffel, Z. Luz, R. Poupko, H. Zimmermann, J. Am. Chem. Soc.
1990, 112, 7158–7163.
[11] a) S. Ternieden, D. Muller, K. Muller, Liquid Crystals 1999, 26, 759–
769; b) S. Ternieden, D. Zauser, K. Muller, Liquid Crystals 2000, 27,
1171–1182.
[38] J. Israelachvili, Intermolecular and Surface Forces, Academic Press,
San Diego, 1991.
[12] M. E. Moseley, R. Poupko, Z. Luz, J. Magn. Reson. 1982, 48, 354–
360.
[39] a) J. Jeener, B. H. Meier, P. Bachmann, R. R. Ernst, J. Chem. Phys.
1979, 71, 4546–4553; b) C. Boeffel, Z. Luz, R. Poupko, A. J. Vega,
Isr. J. Chem. 1989, 29, 283–296.
[40] In practice, we plot the two right sides of Equations (4) and (6)
versus Tc and then determine the intersection point of the two plots.
[13] I. Canet, J. Courtieu, A. Loewenstein, A. Meddour, J.-M. PØchinØ, J.
Am. Chem. Soc. 1995, 117, 6520–6526.
[14] M. Sarfati, P. Lesot, D. Merlet, J. Courtieu, Chem. Commun. 2000,
2069–2081.
[15] P. Lesot, M. Sarfati, J. Courtieu, Chem. Eur. J. 2003, 9, 1724–1745.
Chem. Eur. J. 2007, 13, 3772 – 3786
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3785

P. Lesot et al.
[41] This implies that the coalescence phenomenon cannot be reached at
350 K and hence the data derived from the MMCs are consistent
with the NMR results. At 350 K, the value of DG° determined from
Equation (4) (DG° =+84.7 kJmolꢀ1) is too high compared with the
value calculated with Equation (6) (DG° = +70.9 kJmolꢀ1).
[42] a) V. Schurig, W. Bꢂrkle, J. Am. Chem. Soc. 1982, 104, 7573–7580;
b) F. Gasparrini, D. Misiti, M. Pierini, C. Villani, Tetrahedron:
Asymmetry 1997, 8, 2069–2073; c) M. Jung, V. Schurig, J. Am.
Chem. Soc. 1992, 114, 529–534.
[44] a) J. W. Emsley, P. Lesot, D. Merlet, Phys. Chem. Chem. Phys. 2004,
6, 522–530; b) J. W. Emsley, P. Lesot, J. Courtieu, D. Merlet, Phys.
Chem. Chem. Phys. 2004, 6, 5331–5337.
[45] K. Mislow, P. Bickart, Isr. J. Chem. 1977, 17, 1–6.
[46] E. L. Eliel, Isr. J. Chem. 1977, 17, 7–11.
[47] a) S. Fujita, J. Am. Chem. Soc. 1990, 112, 3390–3397; b) S. Fujita,
Tetrahedron 2000, 56, 735–740.
[43] W. Bꢂrkle, H. Karfunkel, V. Schurig, J. Chromatogr. A 1984, 288, 1–
14.
Received: September 6, 2006
Published online: January 16, 2007
3786
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3772 – 3786