Solute-solvent interactions and chiral induction in liquid crystals

Article abstract of DOI:10.1021/ja051589a

The induction of a cholesteric phase by doping an achiral nematic liquid crystal with an enantiopure solute is a phenomenon that, as in all general supramolecular phenomena of chiral amplification, depends in a subtle way on intermolecular interactions. T

Full text of DOI:10.1021/ja051589a

Published on Web 07/29/2005
Solute-Solvent Interactions and Chiral Induction in Liquid
Crystals
Giorgio Celebre,*,† Giuseppina De Luca,† Michela Maiorino,† Francesca Iemma,‡
Alberta Ferrarini,*,§ Silvia Pieraccini,| and Gian Piero Spada*,|
Contribution from the Dipartimento di Chimica, UniVersita` della Calabria, Via P. Bucci, 87036
Rende (CS), Italy, Dipartimento di Scienze Farmaceutiche, UniVersita` della Calabria, 87036
Rende (CS), Italy, Dipartimento di Scienze Chimiche, UniVersita` di PadoVa, Via Marzolo 1,
35131 PadoVa, Italy, and Dipartimento di Chimica Organica “A. Mangini”, Alma Mater
Studiorum - UniVersita` di Bologna, Via San Giacomo 11, 40126 Bologna, Italy
Received March 12, 2005; E-mail: giorgio.celebre@unical.it; alberta.ferrarini@unipd.it; gianpiero.spada@unibo.it
Abstract: The induction of a cholesteric phase by doping an achiral nematic liquid crystal with an enantiopure
solute is a phenomenon that, as in all general supramolecular phenomena of chiral amplification, depends
in a subtle way on intermolecular interactions. The micrometric helical deformation of the phase director in
the cholesteric phase is generated by the interplay of anisotropy and chirality of probe-medium interactions.
In the case of a flexible chiral dopant, the solvent can influence the twisting power in two ways, difficult to
disentangle: it is responsible for the solute orientational order, an essential ingredient for the emergence
of phase chirality; but also it can affect the dopant conformational distribution and then the chirality of the
structures present in the solution. In this work we have investigated methyl phenyl sulfoxide, a flexible,
chiral molecule that, when dissolved in different nematics, can produce cholesteric phases of opposite
handedness. This peculiar, intriguing sensitivity to the environment makes MPS a suitable probe for a
thorough investigation of the effects of solute-solvent interactions on chiral induction in liquid crystals.
NMR experiments in various nematic solvents have been performed in addition to twisting power
1
measurements. From the analysis of partially averaged H-1H and 13C-1H dipolar couplings, the effects
of solvent on solute conformation and orientational order are disentangled, and this information is combined
with the modeling of the chirality of intermolecular interactions, within a molecular field theory. The integration
of different techniques allows an unprecedented insight into the role of solvent in mediating the chirality
transfer from molecule to phase.
1. Introduction
about the stereochemistry of dopants, the main goal being the
understanding of the relation between the cholesteric handedness
and a stereochemical descriptor of the molecular chirality.3 In
a macroscopic picture, cholesteric induction can be explained
in terms of the competition between the chiral strength, which
favors a twist deformation of the mesophase director, and the
elastic restoring torque, which opposes the deformation.4 The
latter contribution is amenable to measurement and can be
considered a property of the liquid crystal solvent at the given
temperature, at least for dilute solutions. On the contrary, the
chiral strength depends on the solute-solvent pair, and a
molecular statistical analysis allows its interpretation in terms
Chiral nonracemic dopants, when dissolved in nematic
solvents, have the ability of inducing a helical deformation of
the mesophase director on the length scale of micrometers.1 The
resulting chiral nematic (cholesteric) phase is characterized by
the magnitude and sign of the helical pitch p. For dilute solutions
the wavenumber q ) 2π/p is a linear function of the solute
molar fraction x and the enantiomeric excess r:
q ) 2π/p ) 2πâ rx
(1)
The proportionality factor â is called the helical twisting power
(HTP) and is a specific function of the solute-solvent pair. It
is taken as positive or negative, according to the handedness
(right or left, respectively) of the cholesteric helix.2 The study
of induced cholesteric phases has led to relevant information
(3) The chiral doping of nematic phases and its application to stereochemical
problems has been reviewed in: Gottarelli, G.; Spada, G. P. In Materials-
Chirality: Volume 24 of Topics in Stereochemistry; Green, M. M., Nolte,
R. J. M., Meijer, E. W. Eds.; Wiley: Hoboken, NJ, 2003; pp 425-455.
Also the chiral transfer in ferroelectric liquid crystals has been recently
proposed for the configuration assignment: Hartley, C. S.; Wang R. Y.;
Lemieux, R. P. Chem. Mater. 2004, 16, 5297-5303; Lemieux, R. P. Acc.
Chem. Res. 2001, 34, 845-853. For recent papers on the application of
chiral doping of nematics, see: van Delden, R. A.; Feringa, B. L. Angew.
Chem., Int. Ed. 2001, 40, 3198. Kuball, H. G.; Tu¨rk, O. Pol. J. Chem.
1999, 73, 209. Pieraccini, S.; Gottarelli, G.; Labruto, R.; Masiero, S.;
Pandoli, O.; Spada, G. P. Chem.sEur. J. 2004, 10, 5632-5639. van Delden,
R. A.; Mecca, T.; Rosini, C.; Feringa, B. L. Chem.sEur. J. 2004, 10, 61.
(4) Vertogen, G.; de Jeu, W. H. Thermotropic Liquid Crystals. Fundamentals;
Springer: Berlin, 1988.
† Dipartimento di Chimica, Universita` della Calabria.
‡ Dipartimento di Scienze Farmaceutiche, Universita` della Calabria.
§ Universita` di Padova.
| Universita` di Bologna.
(1) Friedel, G. Ann. Phys. Paris 1922, 18, 273.
(2) (a) Solladie´, G.; Zimmermann, R. Angew. Chem., Int. Ed. Engl. 1984, 23,
348. (b) Baessler, H.; Labes, M. M. J. Chem. Phys. 1970, 52, 631. (c)
Spada, G. P.; Proni, G. Enantiomer 1998, 3, 301.
9
11736
J. AM. CHEM. SOC. 2005, 127, 11736-11744
10.1021/ja051589a CCC: $30.25 © 2005 American Chemical Society

Solute−Solvent Interactions and Chiral Induction
A R T I C L E S
of chirality and anisotropy of solute-solvent interactions.5,6 In
this sense HTP can be viewed as a sensitive probe of
intermolecular forces, reflecting features, like their anisotropy,
which are not accessible by experiments in isotropic liquids,
where anisotropies are completely washed out. Unfortunately
the experimental observable conveys only average information
on the interplay between chirality and anisotropy of the
interactions experienced by a given solute. Therefore the
understanding of the molecular mechanism requires a theoretical
analysis, with a detailed modeling of solute and solvent, as
needed when dealing with chiral properties. Actually, it has been
seen that in most cases the twisting ability scales with solvent
in a similar way for different chiral dopants. This has suggested
that the specific features of solvent should play a minor role in
the phenomenon of chiral induction in liquid crystals. Namely,
a simple interpretation of the experimental behavior can be
gained in terms of a phenomenological model, wherein chirality
and anisotropy of interactions are parametrized in terms of those
of the molecular surface of the dopant, while solvent only enters
through its macroscopic properties, i.e., orienting strength and
twist elastic constant.5
Recently, a strong solvent dependence of the helical twisting
power, comprising changes in handedness, has been observed
for a set of flexible alkyl aryl sulfoxides.7 It was seen that in
cyanobiphenyl, cyanophenylcyclohexyl, and cyanobicyclohexyl
solvents these dopants induce chiral nematic phases of handed-
ness opposite to that observed in phenylbenzoate and benzylidene-
aniline solvents. This fact clearly reflects changes in solute-
solvent interactions. A reason for the high solvent sensitivity
of the HTP of alkyl aryl sulfoxides probably resides in their
structure, characterized by the simultaneous presence of an
aromatic ring and a nonnegligible electric dipole along the SO
bond, which can rotate with respect to each other. In the case
of rigid alkyl aryl sulfoxides a different behavior was observed,
with homochiral cholesteric phases induced by a given en-
antiomer: this suggested that changes of cholesteric handedness
with solvent might originate from variations in the torsional
potential about the OS-CC bond.
Figure 1. The molecule of 13C-labeled methyl phenyl sulfoxide (MPS)
with the reference system (x,y,z) and the torsional angle φ with its positive
sense of rotation. The value φ ) 0 is assumed for the S-O bond in the
plane of the aromatic ring.
directly related to the magnitude and orientation of internuclear
vectors connecting the interacting nuclei (see Appendix, section
A2, in the Supporting Information). For this reason reduced
dipolar couplings are also used for the structure determination
of proteins.8 In the case of small molecules as MPS, if a
sufficient number of magnetically active nuclei is available, both
the torsional potential and the order parameter (or Saupe) matrix
S can be obtained as a function of the torsional angle by fitting
the experimental data.9 This detailed knowledge allows us to
disentangle the effect of solvent on torsional potential and
orientational order. The information derived from NMR has been
used within a molecular theory connecting HTP with the
orientational order and chirality of MPS; the latter is calculated
as a function of the torsional angle according to the surface
chirality method.5
The body of the paper has been organized in order to favor
the illustration of results, avoiding detailed descriptions of
important, but perhaps boring, technical aspects; these have been
moved to a long Appendix available in the Supporting Informa-
tion, where interested readers will find all details (for the sake
of brevity, henceforth the frequent calls to the Appendix will
be shortened in this way: App. - s. An - in the SI; where
App. is for Appendix, s. An indicates the n-th section of
Appendix, and SI is, of course, for Supporting Information).
In this paper we report on a more systematic investigation,
which has been carried out with the purpose of getting a deeper
insight into the role of solvent in chiral induction. We have
focused on methyl phenyl sulfoxide (MPS): this compound,
which was not considered in the past studies,7 is the simplest
member of the series of flexible alkyl aryl sulfoxides. The HTP
of the S enantiomer has been measured in various liquid crystal
solvents, chosen for their different structure and dielectric
properties. In addition, an extensive liquid crystal NMR
(LXNMR) analysis has been performed on MPS with 13C-
labeled methyl group, a strategic position for conformational
investigations; henceforth we will refer to this molecule as MPS-
13C. NMR in partially ordered phases is a powerful tool for
2. Results and Discussion
2.1. NMR Experiments and Analysis. NMR experiments
(see Experimental Section) have been carried out at room
temperature on the racemic mixture of MPS-13C (Figure 1)
dissolved in the achiral nematic solvents ZLI-1132, EBBA, and
CCN55 (Chart 1). These nematics, beside being well-known
and widely used as solvents for NMR studies,9 are also
representative of the classes of liquid crystals used for HTP
measurements (see section 2.2). The results obtained for the
racemic compound can be extended to the enantiopure dopant
in the induced cholesteric phase. In fact, the NMR spectra in
the achiral nematic phases, as well as all nonchiral properties,
are identical for the two enantiomers; furthermore, the choles-
teric pitch has a length of the order of micrometers, and the
cholesteric phase induced by a given enantiomer is locally very
similar to the nematic phase formed by the racemic mixture.
1
the conformational analysis of solutes: the spectra of /2 spin
1
nuclei (typically, H and 13C) are strongly affected by dipolar
(or direct) couplings Dij, representing a rich source of structural,
conformational, and orientational information since they are
(5) Ferrarini, A.; Moro, G. J.; Nordio, P. L. Phys. ReV. E 1996, 53, 681.
Ferrarini, A.; Janssen, F.; Moro, G. J.; Nordio, P. L. Liq. Cryst. 1999, 26,
201.
(6) Emelyanenko, A. V.; Osipov, M. A.; Dunmur, D. A. Phys. ReV. E 2000,
(8) Trempe, J.-F.; Gehring, K. In NMR of Ordered Liquids; Burnell, E. E., de
Lange, C. A., Eds.; Kluwer: Dordrecht, 2003.
(9) Celebre, G.; Longeri, M. In NMR of Ordered Liquids; Burnell, E. E., de
Lange, C. A., Eds.; Kluwer: Dordrecht, 2003.
62, 2340.
(7) Pieraccini, S.; Donnoli, M. I.; Ferrarini, A.; Gottarelli, G.; Licini, G.; Rosini,
C.; Superchi, S.; Spada, G. P. J. Org. Chem. 2003, 68, 519.
9
J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005 11737

A R T I C L E S
Chart 1
Celebre et al.
Experimental and calculated spectra of MPS-13C are shown
in Figure 2. Their analysis was carried out in stages, with an
iterative procedure (see App. - s. A1 - in the SI): the chemical
represent the corresponding Cartesian components of the dipolar
coupling tensor between the i and j nuclei. This tensor depends
on the molecular conformation if the two nuclei are located in
different moieties of the solute, i.e., one in the methyl and the
other in the aromatic group. Elements of the dipolar coupling
tensors are calculated on the basis of the molecular geometry
(see App. - s. A3 - in the SI), whereas elements of Saupe
matrices, Súú(φ), and torsional angle distributions in the nematic
phases, p(φ), are derived from a fit of the partially averaged
dipolar couplings by assuming a model for their dependence
upon the φ angle. The methodology, based on the application
of the so-called additive potential (AP) model, has been
presented elsewhere,9 and it is summarized in App. - s. A2 -
in the SI. In short, within this approach the torsional potential
is expressed as a superposition of an isotropic part, which would
be present in a virtual isotropic phase of the nematic solvent at
the experiment temperature and a term accounting for how well
a given conformation can be accommodated in the nematic
environment. The latter is approximated as the sum of inde-
pendent contributions from each “rigid” moiety in the molecule.
The AP method has been criticized11,12 because it suffers, in
principle, from a drawback (see App. - s. A2 - in the SI, for
a detailed discussion). Anyway, this drawback could become
really serious in particular when very flexible solutes (with many
torsional degrees of freedom, as, for example, long alkyl chains)
are treated: for such systems other models (as, e.g., the Chord
Model12) can be used. On the other hand, as well testified by a
wide range of literature in the field, the AP model has been
shifts obtained in this way and the partially averaged DH-H
,
DC-H dipolar couplings are reported in Table 1. The Dij’s in
CCN55 are all positive, and most of them are smaller in
magnitude than those observed in the other solvents: this is
due to the fact that the director aligns perpendicular to the
magnetic field in CCN55 and parallel to it in the other solvents
(the diamagnetic anisotropy ∆ø is negative for CCN55, while
it is positive for EBBA and ZLI-1132).10
The measured Dij dipolar couplings are averages over all
possible molecular conformations, defined by the torsional angle
φ, and all orientations of the molecule relative to the external
applied magnetic field B0. They can be expressed in the
following form:9
Dij ≡ D˜ ijZZ
)
1
2
3
(3 cos2 R - 1) ‚
S (φ) Dij,ςς(φ) p(φ) dφ (2)
∫
∑
ςς
[
] [
]
2
ς
where R is the angle between B0, defining the Z direction in
the laboratory frame, and n, the director of the mesophase. The
function p(φ) is the torsional angle probability distribution, and
Súú(φ)’s are the principal elements of the Saupe matrix S
describing the alignment of MPS to the nematic director.4
The element Súú(φ) is a function of the solute conformation,
since in general both the degree of alignment and the alignment
axes vary when conformation is changed. The Dij,ςς(φ)'s
(11) Photinos, D. J. In NMR of Ordered Liquids; Burnell, E. E., de Lange, C.
A., Eds.; Kluwer: Dordrecht, 2003.
(12) Samulski, E. T. In NMR of Ordered Liquids; Burnell, E. E., de Lange, C.
A., Eds.; Kluwer: Dordrecht, 2003.
(10) de Gennes, P. G.; Prost, J. The Physics of Liquid Crystals; Oxford
Science: Oxford, 1993.
9
11738 J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005

Solute−Solvent Interactions and Chiral Induction
A R T I C L E S
Table 1. Chemical Shifts and Dipolar Couplings of MPS-13C in
ZLI-1132, EBBA, and CCN55
MPS-13C
in ZLI1132
MPS-13C
in EBBA
MPS-13C
in CCN55
{Dij}/{Hz}
D9,10
D9,11
D9,12
D9,13
D9,14
D10,11
D10,12
D10,14
D11,14
D14,15
D8,9
-1543.36 ( 0.06
-231.55 ( 0.19
-61.53 ( 0.06
-24.18 ( 0.08
-77.17 ( 0.12
-474.03 ( 0.17
-24.18 ( 0.08
-95.45 ( 0.12
-82.24 ( 0.05
-1477.79 ( 0.03
-42.72 ( 0.31
-27.55 ( 0.31
-23.86 ( 0.13
-1146.44 ( 0.07
-896.45 ( 0.14
-115.23 ( 0.41
0.13 ( 0.16
630.57 ( 0.18
96.83 ( 0.61
26.19 ( 0.24
14.91 ( 0.24
66.30 ( 0.25
210.62 ( 0.67
14.91 ( 0.24
48.61 ( 0.26
40.26 ( 0.14
548.02 ( 0.08
24.16 ( 0.69
14.92 ( 0.63
11.40 ( 0.37
427.52 ( 0.17
a
60.37 ( 0.16
-99.15 ( 0.21
6.72 ( 0.29
60.37 ( 0.16
-61.42 ( 0.21
-51.63 ( 0.09
-506.06 ( 0.06
-31.84 ( 0.58
-15.72 ( 0.44
-14.83 ( 0.19
-378.06 ( 0.10
a
D8,10
D8,11
D8,14
MPS-13C
in ZLI1132
MPS-13C
in EBBA
MPS-13C
in CCN55
{∆νij}
/
{Hz
}
9,11
10,11
14,11
-77.37 ( 0.18
-83.55 ( 0.19
-71.16 ( 0.40
-100.34 ( 0.29
117.81 ( 0.44
45.98 ( 0.41
-1371.51 ( 0.05 -2190.79 ( 0.09 -1482.74 ( 0.23
a D9,13 ) D10,12 assumed.
Figure 3. Probability distribution p(φ) for the OS-CC dihedral angle of
MPS in the nematic phase of the solvents EBBA, ZLI-1132, and CCN55.
virtual isotropic phase of the nematic solvent at the experiment
temperature, denoted as piso(φ), can be derived.9 This corre-
sponds to the torsional angle distribution devoid of effects of
orientational order, as would be in a conventional liquid sharing
the same physical properties of the nematic solvent (average
dielectric permittivity, density, etc.) with the exception of the
ordering power. Although piso(φ) is in principle different from
p(φ) (see eq A2.12, App. - s. A2 - in the SI), for small
molecules the differences have been generally found to be
small13 or negligible.14 Also for MPS we have found piso(φ) ≈
p(φ): this indicates that orientational order has only minor
effects on the conformational distribution.
The conformational distributions satisfy the relation pS(φ) )
pR(-φ) where the apexes denote the S and R enantiomers; their
plots in the three solvents are shown in Figure 3 (there and in
the following only the range 0°-180° will be shown, because
of the symmetry of the torsional potential with respect to the
aromatic plane).
1
Figure 2. Experimental and calculated H-LXNMR spectra of MPS-13C
in the nematic solvents EBBA, ZLI-1132, and CCN55.
largely used in the past, and it proved its effectiveness and
reliability (also compared with results from other techniques)
especially when small flexible molecules were studied (see, e.g.,
ref 9 and reference therein, where, inter alia, the cases of anisole
and ethylbenzene are illustrated). Since, in the present work,
we deal with MPS, a very small flexible molecule with, in
practice, just one torsional degree of freedom, AP can represent
an effective method to treat the case.
The orientational and conformational analysis has been carried
out as described in App. - s. A4 - in the SI. It is worth
emphasizing that, from the use of the AP model in analyzing
NMR dipolar couplings, also the probability distribution in a
The torsional angle distribution is found to be practically the
same in both the cyano derivatives ZLI-1132 (a solvent with
(13) Celebre, G.; De Luca, G.; Emsley, J. W.; Foord, E. K.; Longeri, M.;
Lucchesini, F.; Pileio, G. J. Chem. Phys. 2003, 118, 6417.
(14) Celebre, G.; De Luca, G.; Longeri, M.; Pileio, G.; Emsley, J. W. J. Chem.
Phys. 2004, 120, 7075.
9
J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005 11739

A R T I C L E S
Celebre et al.
longitudinal dipole, CN ) and CCN55 (a solvent with transverse
||
dipole, CN ), with maxima corresponding to conformers with
the S-O bond close to the ring plane. Potential energy minima
for conformations with the S-O bond approximately in the ring
plane are also predicted by quantum mechanical calculations
in vacuo.7 So, the CN liquid crystals, irrespective of CN-dipole
direction, do not seem to strongly perturb φmin, the position of
the minimum in the torsional potential. On the contrary,
theoretical calculations15 give an energy barrier of 34 kJ/mol,
by far higher than the value obtained in our CN solvents, which
is about 4 kJ/mol (Table A3, App. - s. A4 - in the SI). The
torsional distribution obtained for EBBA is different from that
for CN solvents, being characterized by a pronounced maximum
for conformations with the aromatic ring roughly bisecting the
O-S-CH3 angle. The height of the barrier in EBBA (15 kJ/
mol) is larger than that in CN solvents but still significantly
lower than the value predicted in vacuo. Even considering that
the fitted torsional potential is less accurate in regions corre-
sponding to low probability conformations, the results obtained
in all the investigated solvents indicate that the MPS minimum
energy conformations in solution are less confined than those
predicted by calculations for an isolated molecule. In ref 15
the authors themselves maintain that their theoretical energy
barrier V2 is probably overestimated with respect to experimental
results in solution (ref 15 and references therein); nevertheless,
the significant differences found between our values of V2 and
φmin from different liquid crystals (Table A3, App. - s. A4 -
in the SI) and theory have prompted us to a series of checks to
validate the reliability of the results obtained. To do this, a large
number of tests has been performed (see App. - s. A4 - in
the SI, where the results of the tests are given in detail): the
tests confirm the full inadequacy of theoretical values of V2
and φmin to fit experimental dipolar couplings and the substantial
differences among the three nematic solvents.
The results obtained for MPS suggest that in this case
electrostatic interactions with solvent do not strongly affect the
torsional potential: ZLI-1132 and CCN55, having very different
electrostatic characteristics, share indeed the same conforma-
tional distribution p(φ). About MPS in EBBA, a possible
explanation for the non-negligible differences with the other
solvents (mainly concerning the shape of the probability
distribution function and the location of the conformational
minima) lies, in our opinion, in short-range solute-solvent
interactions, essentially consisting in steric repulsions, which
strongly depend on the geometry of the interacting molecules.
The second piece of information derived from the LXNMR
analysis concerns the orientational order of the solute in nematic
solvents. The Saupe matrix S accounts for the degree of
alignment (magnitude of the principal values) and the preferred
orientations of the solute (direction of the principal axes) in
the liquid crystal environment. Both depend on the solute
conformation. We shall denote the principal alignment axes of
a given conformer as a,b,c, with the labels chosen so that Scc g
Figure 4. Principal values of the ordering matrix of MPS as a function of
the torsional angle φ, as obtained from the NMR measurements in the
solvents ZLI-1132 and CCN55 and EBBA.
magnitude) element of the Saupe matrix is positive or disklike
if it is negative. For a pure rod it would simply be Scc ) -2Saa
) -2Sbb, whereas for a pure disk it would be Sbb ) -2Saa
)
-2Scc. Figure 4 shows the principal elements of the Saupe
matrix in the three solvents as a function of the torsional angle
φ. Figure 5 displays the principal axes of the Saupe matrixes
obtained for three selected values of the torsional angle φ, that
is φ ) 10° and φ ) 50°, which correspond to high probability
conformations respectively in CN solvents and in EBBA, in
addition to φ ) 100°. We can see that the degree of alignment
and alignment axes are quite different in the three solvents. It
appears from the plots in Figure 4 that MPS has a significantly
lower degree of order in EBBA, which cannot be ascribed to a
higher value of the reduced temperature Tr ) T/TNI, since the
spectra were recorded at room temperature and the nematic-
isotropic transition temperatures TNI are quite similar (339, 342,
and 345 K for CCN55, EBBA, and ZLI-1132, respectively).
This implies that the Tr values are very close (in particular, Tr
(CCN55) ≈ 0.88; Tr (EBBA) ≈ 0.88; Tr (ZLI1132) ≈ 0.87),
and this condition justifies a direct comparison of conformational
and orientational behaviors in the different solvents.
Saa g Sbb. This means that a molecule tends to orient in such a
way that the director lies on the ac plane, with preference for
alignment of the c axis, while the b axis tends to stay
perpendicular to the director. The orientational behavior of a
given conformer can be classified as rodlike if the largest (in
(15) Benassi, R.; Mucci, A.; Schenetti, L.; Taddei, F. THEOCHEM 1989, 184,
261.
9
11740 J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005

Solute−Solvent Interactions and Chiral Induction
A R T I C L E S
Figure 5. Principal alignment axes of MPS at selected values of the torsional angle, φ ) 10°, φ ) 50°, and φ ) 100°, as obtained from the NMR measurements
in the nematic solvents ZLI-1132, CCN55, and EBBA. The planes ac (panel A) and bc (panel B) are shown; the b axis tends to lie perpendicular to the local
director and along the helix axis in the chiral nematic phase. The framed structures correspond to high probability conformations in the three solvents.
The plots in Figure 4 indicate a clear rodlike behavior of the
solute in EBBA; the projections in Figure 5 show that in this
solvent the axis of preferential alignment to the director is not
far from the Ph-S bond, irrespective of the conformation. The
direction of the other principal alignment axes in the molecule
changes with conformation, in such a way that the b axis, which
has the strongest tendency to lie normal to the director, remains
roughly parallel to the SO bond. On the contrary, when MPS is
dissolved in the CN solvents, a basically disklike orientational
behavior appears from Figure 4, along with a more pronounced
dependence of alignment on the molecular conformation (rod/
disk changes in solute’s ordering, on passing from the solvent
EBBA to the solvent ZLI-1132, were also found in the past for
small rigid molecules such as 1,4-difluorobenzene and p-
benzoquinone16). In ZLI-1132 a strong difference in the align-
ment of the three axes (large biaxiality) is observed, and Figure
5 shows a definite preference of the director to lie on the plane
of the aromatic ring. An even clearer disklike behavior emerges
from the plot of order parameters in CCN55, with only a small
preference for alignment to the director of the c over the a axis.
Moreover, Figure 5 shows that the orientation of the (ac) plane,
which preferentially lies parallel to the nematic director, strongly
depends on the molecular conformation; in contrast to the other
two solvents, the orientation of the c axis can be far from parallel
to the Ph-S bond. The difference between ZLI-1132 and
CCN55 in orienting MPS probably has to be ascribed to
(16) Celebre, G.; De Luca, G. Chem. Phys. Lett. 2003, 368, 359.
9
J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005 11741

A R T I C L E S
Celebre et al.
Table 2. Helical Twisting Power (S Enantiomer) and Chirality
Parameter Qh (Obtained as Described in the Text), for MPS in
Different Solvents
the relation among HTP, orientational order, and conformational
distribution is not straightforward. To establish this link,
calculations based on the surface chirality method 5 have been
performed. According to this model the helical twisting power
can be expressed as
1
â
/
µ
m-
Qh
/Å3
cyanophenylcyclohexyl
-2.0 (ZLI-1132)
-2.3 (ZLI-1083)
+0.4 (MBBA/EBBA)
+1.3 (CCN55)
<0a (ZLI-1132)
benzylideneaniline
CN -cyanobicyclohexyl
+0.7 (EBBA)
+3.5 (CCN55)
â ) (RTê/2πK22Vm)Q
(3)
CN -cyanobicyclohexyl
-2.7 (ZLI-2395)
-2.2 (E7)
||
CN -cyanobiphenyl
where R is the gas constant, T is temperature, while K22 and Vm
are the twist elastic constant and molar volume of the solution,
which for high dilution practically coincide with those of the
solvent. The parameter ê represents the orienting strength of
the liquid crystal environment, and Q is the chirality parameter,
which depends on the coupling of chirality and orientational
behavior of the dopant. It is defined as
||
phenylbenzoate
+0.9 (ZLI-1052)
a A negative value is obtained, of the order of magnitude of the numerical
error (∼0.1 Å3).
electrostatic interactions.17 The two solvents have indeed large
and opposite dielectric anisotropies (ꢀ| - ꢀ ) -8.2 for CCN55
and +10.3 for ZLI-1132, at 20 °C), due to the presence of a
strong electric dipole roughly perpendicular to the alignment
axis in one case, and parallel to it in the other.4 The multisided
role played by electrostatic interactions appears from the
LXNMR analysis: on one hand they do not seem to influence
the conformational distributions of MPS in ZLI-1132 and
CCN55, but on the other hand they are decisive in governing
the orientational order in the two solvents. More difficult to
figure out is the origin of the different orientational behaviors
of MPS in EBBA compared with those in the CN solvents,
because of the more pronounced differences between the two
kinds of solvents. Besides electrostatics (EBBA has a small
negative dielectric anisotropy, deriving form the presence of a
relatively small dipole moment, roughly perpendicular to the
long molecular axis), also structural features should be taken
into account: aromatic-aromatic interactions and steric repul-
sions involving the broad, flat, rigid core of EBBA are expected
to play a significant role.
With reference to the phenomenon of chiral induction, which
is the main object of our investigation, it is worth recalling that
the axis of the cholesteric helix is perpendicular to the local
director, and the solute tends to align to what we have denoted
as the b axis. Considering now the three framed structures in
Figure 5, which correspond to high probability geometries in
the three solvents, we can see that the average molecular
orientation with respect to the helix axis is quite different in
the three cases. This result alone is not sufficient to predict the
twisting ability of MPS in the three solvents, but at least it
suggests that different helicities might be transferred to the
medium in the three cases.
Q ) -x(2/3) (Qaa Saa + Qbb Sbb + Qcc Scc)
(4)
where Súú are principal elements of the Saupe ordering matrix
S, and Qúú are the corresponding components of the helicity
tensor Q. This describes the helicity of the molecular surface
and accounts for the chirality of intermolecular interactions. The
component Qúú describes the helicity of the molecular surface
of the dopant as viewed along the ú axis. As a matter of fact,
contributions of different sign and magnitude to HTP can derive
from different regions in the molecule.18 The Q tensor depends
on the molecular conformation; therefore also the chirality
parameter is a function of the torsional angle, Q ) Q(φ). The
measured helical twisting power is proportional to Qh, which is
defined as the average over the torsional angle, weighted by
the probability distribution p(φ):
Qh ) Q(φ) p(φ) dφ
(5)
∫
Given the geometry of the dopant, the molecular surface is
defined and the chirality tensor Q can be calculated; this is a
function of the molecular shape, then of the torsional angle.
The chirality parameter Q(φ) has been calculated for each
conformation according to eq 4, by using for each solvent the
principal order parameters Súú(φ) obtained by LXNMR, together
with the corresponding components of the chirality tensor, Qúú
(φ), evaluated on the basis of the molecular surface. It is worth
pointing out that the tensor components Qcc, Qbb, and Qaa can
be significantly different for the three solvents since, as we have
seen, the orientation of the a,b,c axes change with solvent. It
follows that, due to the different orientation, the solute in a given
conformation is expected to have different twisting abilities in
the three solvents.
2.2. HTP Measurements and Predictions of the Surface
Chirality Method. Helical twisting power of the (S)-MPS
enantiomer has been measured in various solvents. The HTP
values are reported in Table 2. In all cases a small HTP is found,
with a sign which depends on the solvent. As for other flexible
alkyl aryl sulfoxides,7 opposite handedness is found in CN
solvents (ZLI-2395, ZLI-1083, ZLI-1132, and E7) and phenyl-
benzoates or benzylideneanilines. Moreover, opposite handed-
The Q(φ) profiles calculated according to eq 4 for ZLI-1132,
CCN55, and EBBA are shown in Figure 6. In the same plots
also the torsional angle distributions from LXNMR are shown,
to highlight the range of φ values which give the strongest
contribution to the average chirality parameter Qh, according to
eq 5. The importance of a change of alignment axes and order
parameters appears from comparison of the Q(φ) profiles. Those
obtained in CCN55 and ZLI-1132 look quite different; in
particular we can see that the stable conformations of MPS,
which correspond to the same geometry in these two solvents,
are characterized by negative values of the chirality parameter
||
ness is also measured in CN and CN solvents.
||
As we have seen above, differences in orientational order
and conformational distribution of MPS in the three classes of
solvents have emerged from the NMR analysis. However, due
to the complexity of the mechanism of cholesteric induction,
(17) di Matteo, A.; Todd, S. M.; Gottarelli, G.; Solladie´, G.; Williams, V. E.;
Lemieux, R. P.; Ferrarini, A.; Spada, G. P. J. Am. Chem. Soc. 2001, 123,
7842.
(18) Todd, S. M.; Ferrarini, A.; Moro, G. J. Phys. Chem. Chem. Phys. 2001, 3,
5535.
9
11742 J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005

Solute−Solvent Interactions and Chiral Induction
A R T I C L E S
The Qh values in the three solvents, calculated with eq 5 by
averaging over the conformational distribution, are reported in
Table 2. According to eq 3 the Qh value is proportional to the
helical twisting power, with a proportionality factor which
depends on solvent. By using reasonable values (T ∼300 K,
K22 ∼3 × 10-12 N, Vm ∼3 × 10-4 m3, ê ∼3 × 10-2 Å-2) we
obtain â [µm-1]/Q [Å3] ≈ order of the unity. We can see that
the predicted values have the correct order of magnitude and,
more important, they show a clear trend with solvent, going
from negative to positive on moving from CN to CN through
||
EBBA, in agreement with experiment. The change in sign more
clearly appears from comparison of the Q values for the stable
conformations in Figure 6. A closer comparison between
calculated and measured quantities is made difficult by a number
of reasons. First of all, the surface chirality method is based on
a simple parametrization of the chirality of intermolecular
interactions. In addition, the unavoidable uncertainty in torsional
angle distribution p(φ) and order parameters Súú entering eqs 4
and 5, especially in correspondence of the less stable conforma-
tions, can affect the calculated Qh values. Finally, precise
estimates of the parameters entering eq 3 are not available,
especially for the commercial mixtures. Therefore accurate
predictions of helical twisting power, a property resulting from
the fine balance of contributions, cannot be expected. Neverthe-
less, unambiguous results have been obtained, which allow us
to explain the experimental findings.
3. Conclusion
Cholesteric induction originates from the interplay of the
anisotropy and chirality of the interactions experienced by the
chiral dopant in the liquid crystal phase. If the dopant is a
flexible molecule, the solvent influences the observed HTP in
different interlinked ways, which cannot be easily disentangled.
On one side solvent affects the torsional potential and the
conformational distribution, so the chirality of the structures
present in the solution. On the other side, it is responsible for
orientational order, which is essential for the emergence of the
phase chirality. In this work we have considered the chiral
dopant methyl phenyl sulfoxide in different solvents, which can
be grouped into three classes: weakly polar solvents with broad
aromatic cores (phenylbenzoates and benzylideneanilines) and
more polar and flexible solvents, with the electric dipole either
longitudinal or transversal (cyanobicyclohexyls and cyanophen-
ylcyclohexyls). We have seen that MPS has low twisting ability,
and the cholesteric phase induced by the (S) enantiomer is left-
handed in cyanobicyclohexyls and cyanophenylcyclohexyls
solvents with longitudinal dipole, and right-handed in the other
cases.
Figure 6. Dependence on the torsional angle φ of the chirality parameter
Q calculated as explained in the text for the nematic solvents ZLI-1132,
CCN55, and EBBA (solid lines). Dotted lines represent the probability
distribution for the OS-CC dihedral angle, p(φ).
in ZLI-1132, whereas positive values are predicted in CCN55.
This result is in agreement with the sign of the helical twisting
observed in the two solvents, and the origin of the behavior
can be explained by the different alignment of MPS. If we
compare the framed configurations in Figure 5, we can see that
the molecule, viewed along the axis of the cholesteric helix,
looks different in ZLI-1132 and in CCN55. The Q(φ) profile
obtained for MPS in EBBA is different from both those obtained
in the other solvents. The angular dependence of Q is roughly
the opposite of that in ZLI-1132 (with maxima in EBBA roughly
corresponding to minima in ZLI-1132, and vice versa); the
profile looks more similar to the angular dependence in CCN55,
but with smaller values altogether. In the case of EBBA also a
different conformational distribution has to be taken into
account. However even in this case a major role is played by
the orientational behavior: we can see in Figure 5 that the
dopant orientation in EBBA is different, in different respects,
from both those in ZLI-1132 an in CCN55. A positive HTP is
predicted for the stable conformations of MPS in EBBA, again
in agreeement with experiment.
With the help of a molecular theory it has been possible to
use the results of the LXNMR analysis, from which conforma-
tional distribution and orientational order of the dopant can be
obtained, to make predictions of HTP in the three classes of
solvents. In this way a detailed description of the phenomenon
is obtained; a complex picture emerges from the analysis,
wherein the multiform role played by solvent appears. In the
following, the results of our investigation will be summarized,
without neglecting a number of unexplained issues which
stimulate future work.
(i) The conformational distribution of MPS is very similar
in cyanobicyclohexyl and cyanophenylcyclohexyl solvents,
9
J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005 11743

A R T I C L E S
Celebre et al.
gel (70-230 mesh) and with hexane as eluent. The obtained thioanisole-
R-13C weighed 1.11 g (8.9 mmol) (74%). MS (m/z): 125 (100% M+),
109 (38%), 92 (30%), 78 (32%).
irrespective of the orientation of the dipole in the solvent
molecule, whereas significant changes are observed in solvents
with a broader aromatic core and a lower value of the average
dielectric permittivity, like phenylbenzoates and benzylidene-
anilines. The results obtained suggest that packing effects,
determined by short-range steric interactions which strongly
depend on the structure of solute and solvent, might be important
in determining the stability of MPS conformations. Further work
is needed to gain a deeper insight into this aspect, which goes
beyond the main aim of the present work.
(ii) The orientational order of MPS is different in the three
classes of solvents and appears to depend on both electrostatic
and structural properties of solvent. Not only the degree of order
but also the alignment axes of the solute change with solvent.
This kind of behavior has already been observed for other small
solutes in nematic solvents,16 and the molecular mechanism
behind it is not fully clear yet.
(iii) As the main result of this work, we have shown that the
change of cholesteric handedness with solvent is substantially
driven by a change in orientation of the dopant. Namely, this is
the only reason for the cholesteric helix inversion on going from
ZLI-1132 to CCN55. In the case of EBBA also a variation in
the conformational distribution has to be considered, but again
the different alignment is the main reason for the change of
handedness with respect to ZLI-1132. The orientational differ-
ences in the three solvents, illustrated in Figure 5, might appear
not very significant; indeed, the system under investigation gives
an example of a peculiar feature of chiral properties, i.e., the
dramatic effects of even small variations at the molecular level.
By a proper combination of experiments and theory, we have
been able to identify the subtle differences in solute-solvent
interactions which underlie the puzzling behavior of the chiral
solute MPS in nematic solvents. A given enantiomer can
potentially induce left- and right-handed chiral nematic phases;
the actual handedness is selected by solvent, mainly through
its orienting peculiarities, even though effects on the confor-
mational distribution can also play some role. This is a general
rule, which holds for all chiral dopants; however changes in
alignment and conformational stability with solvent are much
weaker in the case of bulky molecules, with clear orientational
preferences, which can only be slightly modified by the
structural features of the solvent.
Methyl Phenyl Sulfoxide-R-13C: A solution of metaperiodate (1.44
g, 6.7 mmol) in 10 mL of bidistilled cooled water to 0 °C was added
to thioanisole-R-13C (0.80 g, 6.4 mmol) and allowed to react for 3 h.
Then the reaction mixture was submitted to the following treatments:
addition of bidistilled water (10 mL), extraction with chloroform,
anhydrification with Na2SO4, evaporation of the solvent, and column
chromatography with Merck 60 silica gel (70-230 mesh) and with
hexanes-ethyl acetate (40:60) as eluent. The solid methyl phenyl
sulfoxide-R-13C weighed 0.86 g (6.1 mmol) (95%). MS (m/z): 141
(1% M+ + 1), 140 (91%), 125 (100%), 109 (11%), 97 (47%), 91 (11%),
77 (50%).
4.2. Helical Twisting Powers. Cholesteric pitches and handedness
were obtained at room temperature using the lens version of the
Granjean-Cano method.20
4.3. NMR Experiments. LXNMR experiments were performed by
using the racemic mixture of MPS-13C. Samples were prepared by
dissolving approximately 10 wt % of MPS-13C in the nematic solvents
ZLI1132, EBBA, and CCN55 (all purchased from Merck Ltd.). The
1
13C- and H-LXNMR spectra were recorded at 298 K (a) on a Bruker
AC 300, working at 7.04 T (MPS-13C in ZLI-1132 and CCN5) and
(b) on a Bruker AVANCE 500 working at 11.7 T (MPS-13C in EBBA)
on samples contained in 5 mm o.d. sample tubes. The free induction
decays were stored in 32 kWords of computer memory giving a Hz/pt
precision ratio on measuring the peak position of 0.49, 0.21, and 0.25,
respectively, in ZLI1132, EBBA, and CCN55.
4.5. Surface Chirality Calculations. According to eq 4, evaluation
of the chirality parameter Q(φ) for a given molecular geometry requires
the chirality tensor Q and the ordering matrix S. The elements of the
chirality tensor, which depend on the chirality of the molecular surface,
are calculated in the following way.5 Given the nuclear positions for a
specified value of the torsional angle φ, the molecular surface is
generated. This is defined as the surface drawn by the center of a bead
rolling on the assembly of interlocking van der Waals spheres centered
on the nuclei and is approximated by a set of triangles, obtained with
the algorithm developed by Sanner et al.21 The following radii have
been assumed for calculating the surface of MPS: rCH(aromatic) ) 2 Å,
rC(methyl) ) 1.85 Å, rH(methyl) ) 1.2 Å, rS ) 2 Å, rO ) 1.5 Å. The
geometric parameters are specified in Table A2 of App. - s. A3 - in
the SI; the same values have been used for the analysis of NMR data
and for the calculation of the tensor Q(φ). For the elements of the Saupe
matrix appearing in eq 4, the values derived from the NMR analysis
were used for each solvent.
Acknowledgment. The paper is dedicated to Professor
Giovanni Gottarelli, author of seminal works on cholesteric
induction, on the occasion of his 65th birthday. This work has
been supported by MIUR (PRIN2003 “Cristalli Liquidi e
Macromolecole per Nanostrutture Organizzate”). A.F. acknowl-
edges additional support from FIRB2001 (RBNE01P4JF). F.I.
acknowledges support from Unical University Funds. The
authors gratefully acknowledge Prof. M. Longeri (LXNMR,
S.C.An. Lab., Dipartimento di Chimica, Universita` della Cala-
bria) for stimulating discussions and Professor Carlo Rosini
(Universita` della Basilicata) for supplying enantiopure MPS.
4. Experimental Section
4.1. Synthesis. (S)-Methyl phenyl sulfoxide has been prepared
according to the procedure described in ref 19. Racemic methyl phenyl
sulfoxide-R-13C (MPS-13C) has been obtained via oxidation of thio-
anisole-R-13C according to the following procedure.
Thioanisole-R-13C: Thiophenol (1.32 g, 12.0 mmol) was allowed
to react with sodium methoxide (16.0 mmol) in dry methanol (40 mL)
at -10 °C. The reaction mixture was stirred and cooled for an additional
10 min, and 2.00 g (14.0 mmol) of iodomethane-13C were slowly added.
After stirring for 1 h at room temperature and removal of the solvent
at reduced pressure, the residue was hydrolyzed with water (100 mL)
and the aqueous phase was extracted with ether (4 × 70 mL). All the
organic phases were combined, dried (Na2SO4), and freed of the solvent
at reduced pressure to afford the thioanisole-R-13C as a liquid which
was further purified by column chromatography with Merck 60 silica
Supporting Information Available: Appendix containing
details about the following: (A1) Analysis of NMR Spectra;
(A2) Theoretical Background; (A3) Geometries of Rigid
Subunits; (A4) Conformational and Orientational Analysis. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(19) Donnoli, M. I.; Superchi, S.; Rosini, C. J. Org. Chem. 1998, 63, 9392.
(20) Heppke, G.; Oestreicher, F. Z. Naturforsch. 1977, 32a, 899. Gottarelli, G.;
Samor`ı, B.; Stremmenos, C.; Torre, G. Tetrahedron 1981, 37, 395.
(21) Sanner, M. F.; Spenher, J.-C.; Olson A. Biopolymers 1996, 38, 305.
JA051589A
9
11744 J. AM. CHEM. SOC. VOL. 127, NO. 33, 2005