The influence of shape and size of silyl units on the properties of bent-core liquid crystals - From dimers via oligomers and dendrimers to polymers

Article abstract of DOI:10.1039/b706923e

A series of oligomeric and dendritic liquid crystalline materials combining mesogenic bent-core units and silicon-containing linking units (siloxanes and carbosilanes) was synthesized and investigated. The influence of the size and shape of the silyl units upon the mesophase type and the switching behaviour was investigated. Linear oligosiloxane based dimers and dendritic carbosilanes show ferroelectric and non-classical antiferroelectric switching monolayer smectic phases with dark conglomerate textures coexisting with achiral domains, whereas cyclic oligosiloxane and silsesquioxane derivatives have antiferroelectric switching and birefringent smectic phases with double layer structure. It is concluded that ferroelectricity and layer distortion require a certain size of the silyl groups and a monolayer structure where each layer is decoupled by sublayers formed by the silylated units. The Royal Society of Chemistry 2007.

Full text of DOI:10.1039/b706923e

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/materials | Journal of Materials Chemistry
The influence of shape and size of silyl units on the properties of bent-core
liquid crystals—from dimers via oligomers and dendrimers to polymers{
Christina Keith,a Gert Dantlgraber,a R. Amaranatha Reddy,{a Ute Baumeister,b Marko Prehm,ab
Harald Hahn,c Heinrich Langc and Carsten Tschierske*a
Received 8th May 2007, Accepted 25th June 2007
First published as an Advance Article on the web 11th July 2007
DOI: 10.1039/b706923e
A series of oligomeric and dendritic liquid crystalline materials combining mesogenic bent-core
units and silicon-containing linking units (siloxanes and carbosilanes) was synthesized and
investigated. The influence of the size and shape of the silyl units upon the mesophase type and the
switching behaviour was investigated. Linear oligosiloxane based dimers and dendritic
carbosilanes show ferroelectric and non-classical antiferroelectric switching monolayer smectic
phases with dark conglomerate textures coexisting with achiral domains, whereas cyclic
oligosiloxane and silsesquioxane derivatives have antiferroelectric switching and birefringent
smectic phases with double layer structure. It is concluded that ferroelectricity and layer distortion
require a certain size of the silyl groups and a monolayer structure where each layer is decoupled
by sublayers formed by the silylated units.
recently. It turned out that siloxane9 as well as carbosilane
1. Introduction
groups2b,6,10 have
a great influence on the mesophase
behaviour of bent-core molecules.11 It was shown that
increasing the number of dimethylsilyl groups per bent-core
mesogenic moiety gives rise to a transition from antiferro-
electric switching to ferroelectric switching.2b
Currently there is great scientific interest in bent-core liquid
crystalline materials. The special properties of these liquid
crystalline (LC) phases, namely antiferroelectric and ferro-
electric switching, lead to the potential for numerous applica-
tions.1 In these smectic LC materials the polar order occurs
due to the restricted rotation of the bent-core molecules
leading to a polar direction parallel to the layers. If the
molecules are additionally tilted with respect to the layer
planes (SmCPA/FE phases), the layers possess an inherent
chirality. In these LC phases layer normal, tilt direction and
polar direction describe either a right- or left-handed system
(see Fig. S1 and S2 in the ESI{).1 The occurrence of chiral
superstructures in these mesophases formed by achiral
molecules is of significant general scientific interest. Research
in this field has concentrated mainly on low molecular
weight bent-core mesogens, but also a few dimesogens2,3 and
polymers4,5 have been reported so far. First examples of
dendrimers based on carbosilane cores6,7 and diaminobutane-
poly(propylideneimine)dendrimer (DAB) cores8 with bent-
core mesogenic units at the periphery were also reported
Herein a series of new carbosilane and oligosiloxane derived
oligomesogens and dendritic molecules with bent-core units
are compared. Two types of spacers, an aliphatic -(CH2)11-
spacer (compounds Xb) and another one incorporating an
additional ether oxygen (CH2)3O(CH2)6 (compounds Xa) were
used to couple the bent-core moieties to the Si-containing
groups. It was found that flexible structures, such as
dendrimers and linear oligosiloxanes, favour ferroelectric
switching and dark textures, whereas more rigid silicon con-
taining groups, like cyclic siloxanes12 and silsesquioxanes,13
preferably form strongly birefringent smectic phases with
antiferroelectric switching behaviour. Molecules based on
linear or branched silyl groups organize in monolayer
structures where each layer of the bent-core units is separated
by silyl sublayers. In contrast, cyclic units give rise to double
layer structures where only every second interlayer interface is
formed by silyl sublayers, and this difference leads to the
distinct properties. Moreover, all dendrimers 1/na and also the
siloxane based dimesogen 5b form mesophases appearing dark
between crossed polarizers with coexisting chiral and achiral
domains, whereas the carbosilane based dimesogens 6a and 7a
as well as the side chain polymer 8b show dark mesophases
exhibiting exclusively chiral domains.
aInstitute of Chemistry, Organic Chemistry, Martin-Luther-University
Halle-Wittenberg, Kurt-Mothes-Str. 2, D-06120, Halle, Germany.
E-mail: carsten.tschierske@chemie.uni-halle.de; Fax: +49 345 5527 346;
Tel: +49 345 5525 664
bInstitute of Chemistry, Physical Chemistry, Martin-Luther-University,
Halle-Wittenberg, Mu¨hlpforte 1, D-06108, Halle, Germany
cInstitute of Inorganic Chemistry, Technical University Chemnitz,
Faculty of Sciences, Strasse der Nationen 6, D-09111, Chemnitz,
Germany
2. Results and discussion
{ Electronic supplementary information (ESI) available: Figures
explaining the phase structures and origin of chirality in the
mesophases of bent-core molecules, additional XRD patterns and
textures, table with X-ray data, synthetic procedures and analytical
data. See DOI: 10.1039/b706923e
2.1. Synthesis
The compounds discussed in this paper were synthesized
by hydrosilylation of the olefin-terminated bent-core
mesogens Ena and Enb,6b,9a respectively, with appropriate
{ Present address: Dept. of Chemistry & Biochemistry, University of
Colorado at Boulder, Boulder, CO 80309-0215, USA.
3796 | J. Mater. Chem., 2007, 17, 3796–3805
This journal is ß The Royal Society of Chemistry 2007

View Article Online
H-functionalized dendritic carbosilanes (to give compounds
1/na and 2b) or H-functionalized linear or cyclic oligo(silox-
anes) (to give compounds 3b–5b) using Karstedt’s catalyst (see
Scheme 1).14 Due to the distinct reactivity of carbosilanes and
siloxanes in hydrosilylation reactions6b the olefin Enb was used
for the synthesis of siloxane derivatives (compounds 2b–5b),
whereas the allyl ether Ena was used for the synthesis of the
carbosilane derived dendrimers 1/na. The Si–H terminated
dendritic carbosilanes used for the hydrosilylation reaction
were synthesized in
a repetitive manner following a
procedure described by Muzafarov et al.15 The H-termi-
nated siloxanes, i.e. 1,3,5,7-tetramethylcyclotetrasiloxane
(Lancaster), 1,3,5,7,9,11,13,15-octakis(dimethylsiloxy)pentacy-
clo-[9.5.1.13,9.15,15.17,13]octasilane (Aldrich) and the linear
and at both ends H-terminated oligosiloxane with on average
eight siloxane units (Aldrich) were commercially available. The
synthesized compounds are depicted in Scheme 2 and the
mesogenic properties of these compounds are summarized in
Table 1. The dendrimers 1/15a, 1/4a6a and 1/4b,6b the
carbosilane derived dimesogens 6a and 7a2b as well as the
polymer 8b4a were reported previously and these compounds
are incorporated for comparison. All other compounds are
new. The compounds were separated from by-products and
purified by repeated column chromatography. Analytical
characterization was carried out by TLC, NMR (1H, 13C,
29Si) and elemental analysis (see ESI{). In the case of the cyclic
compounds 3b and 4b a complete functionalization of the SiH
groups could not be achieved. According to the analysis of the
1H-NMR spectra (Si–H at 5.4 ppm) there is 0.5 Si–H per
molecule 3b, which means that it actually represents a 1 : 1
mixture of the tetramesogen (shown in Scheme 2) and the
corresponding trimesogen with one unfunctionalized Si–H
unit. The silsesquioxane derivative 4b is also a 1 : 1 ratio of the
completely substituted octamesogen and the corresponding
heptamesogen with one Si–H unit. This is considered in the
calculation of the Si/bc ratios in Table 1. For all other
compounds and polymers neither additional SiH signals nor
Scheme 1 Structure and LC phases of the olefins Ena and Enb used
for the synthesis of compounds 1–86b,9a and the influence of the
structure of the spacer unit upon the properties of the trisiloxane
substituted mesogenic molecules Sia, Sib and Sic.9a,c Abbreviations:
Cr = crystalline solid, Iso = isotropic liquid; SmCPA = antiferroelectric
switching and tilted smectic phase with birefringent texture;
Scheme 2 Structures of the discussed oligomesogens; the dendrimers
1/4a,6b 1/4b6a and 1/15a,6b the carbosilane based dimesogens 6a and
7a9a and the polymer 8b4a were reported in earlier communications and
are included for comparison; in the dendrimer 1/15a, there are 16
branches, 15 of them are functionalized with bent-core units and one
by a OMe group; the cyclic oligomesogens 3b and 4b represent 1 : 1
mixtures of the shown molecules with corresponding molecules having
one bent-core unit less.
[*]
SmCPA = SmCPA phase with dark texture showing domains of
opposite handedness (dark conglomerate texture); values in parenth-
eses indicate monotropic (metastable) mesophases.
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 3796–3805 | 3797

View Article Online
Table 1 Mesophases of compounds 1–8a
Compd
T/uC [DH/kJ mol21
]
Si/bc
1.3
d/nm
Ps/nCcm22
[*N]
[*N]
[*N]
[*N]
1/3a
Cr
66
[40.8]
73
,20
65
SmCPA9
106
Iso
3.95
800
[56.5]
111
47
1/4a6b
1/4b6a
1/6a
Cr
Cr
Cr
SmCPA9
g
SmCPA9
Iso
SmCRPFE
Iso
1.25
1.25
1.7
4.0
4.1
700
1400
750
139
Iso
115
[16.9]
79
79
[87.5]
113
129
1/15a6b
2b
Cr
Cr
SmCPA9
Iso
Iso
1.9
1.3
4.45
4.4
600
980
[*]
SmCPA9/FE
[22.4]
93
[40.2]
141
[40.0]
53
[61.7]
143
[89.5]
149
[152.0]
130
3b
4b
5b
Cr
Cr
Cr
SmCPA
Iso
Iso
Iso
1.1
2.1
4
7.8
700
300
650
SmCPA
8.0
[*N]
SmCPFE
4.55
[7.6]
98
57
67
[42.8]
95)
98
[*]
6a2b
7a2b
8b4a
a
Cr
Cr
Cr
(SmCPFE
[*]
SmCPFE
(g
Iso
Iso
SmCPFE
3.5
4
16
4.8
4.8
7.3
660
880
500
[*]
29)
119
Iso
Abbreviations: Si/bc = ratio of the number of Si-atoms per bent-core mesogenic unit; the incomplete functionalization of the Si–H groups in
the cases of 1/15a (one Si–OMe unit) and the cyclic oligomesogens 3b and 4b (see text, section 2.1) is considered in this calculation; d = layer
spacing; Ps = spontaneous polarization; SmCRPFE = ferroelectric switching smectic phase with random tilt direction and polar direction;
[*N]
SmCPA9
tristable switching) with dark texture showing chiral ([*]) and achiral ([N]) domains; SmCPFE ferroelectric switching polar and tilted smectic
phase with dark conglomerate texture; g = glass transition temperature; for the other abbreviations, see Scheme 1.
= non-classic antiferroelectric switching tilted smectic phase (A9 = only one polarization current peak, but optically observed
[*]
signals due to protons of the double bonds of residual
olefinic precursors (which would be clearly visible in the range
in-plane order and an organization of the siloxane units in
distinct layers. Only one diffuse scattering around d = 0.5 nm is
observed for the carbosilane derivatives (dendrimers 1/na,
dimesogens 6a,7a and trimesogen 2b). As the mean distance
between the carbosilane chains is approximately the same as
that between alkyl chains the formation of a distinct carbo-
silane layer cannot be indicated by X-ray evidence in the case
1
5–6 ppm) could be detected in the H-NMR spectra.
2.2. Mesomorphic properties
The mesomorphic properties of the compounds discussed
here were investigated by polarized light microscopy on a hot
stage (Mettler FP 82 HAT), differential scanning calorimetry
(DSC-7, Perkin-Elmer), X-ray diffraction, as well as by
electrooptical experiments in transparent ITO-coated glass
cells (EHC, Japan).
Generally, the viscosity of the liquid crystalline phases
increases with the number of bent-core mesogenic units, and
hence molecules with a smaller number of mesogenic units give
more specific textures and can more easily be investigated than
those with a larger number.
2.2.1. X-Ray scattering. All compounds show relatively
broad regions of liquid crystalline phases representing layer
structures with layer distances between d = 4.0 nm and d =
4.8 nm, with the exceptions of polymer 8b and cyclosiloxanes
3b and 4b which have much larger layer distances of d = 7.3,
7.8 and 8.0 nm, respectively (see Table 1 and ESI{ Table S1).
For the oligomesogens and dendrimers 1/na and 2b as well as
for the dimesogens 6a and 7a the layer distance is only slightly
dependent on the structure of the silyl unit and much smaller
than the overall size of the oligomeric molecules, indicating
that the layers are formed by the organization of the individual
bent aromatic cores. These aromatic cores form layers which
are separated by aliphatic layers containing the terminal chains
as well as the spacer units, and layers of the silyl groups as
shown in Fig. 1a. The diffraction patterns of the siloxane
derivatives (with the exception of the cyclosiloxanes 3b and 4b)
show two maxima of the outer diffuse scattering around 0.44–
0.5 nm and 0.7–0.8 nm indicating fluid smectic phases without
Fig. 1 a) Organization of silyl-substituted bent-core molecules in
triply segregated layer structures where silyl units, bent-core units and
the aliphatic parts (spacers plus terminal alkyl chains) are organized
in separate layers (d = layer distance); a synpolar (ferroelectric)
organization is shown, as proposed for the SmCsPF layer stacks; b)
sponge-like distorted layer structure where the aromatic bent cores are
organized on the shown surface.
3798 | J. Mater. Chem., 2007, 17, 3796–3805
This journal is ß The Royal Society of Chemistry 2007

View Article Online
Fig. 3 2D XRD pattern of a partially surface-aligned sample of 3b at
135 uC on cooling: a) small-angle region, b) full scattering range, the
scattering of the isotropic liquid is subtracted to enhance the effect of
the anisotropic distribution of the wide-angle scattering in the SmCPA
phase; from the positions of the maxima of the diffuse wide angle
scattering an average tilt angle of the molecules with respect to the
layer normal of 36u can be calculated (see also ESI Fig. S3). The higher
intensity of the maxima at 130 and 314u compared with the other two
indicates a synclinic tilt of the molecules in different layers.
Fig. 2 Organization of the cyclic siloxane derivatives 3b and 4b in
double layer structures where silyl units, spacers, bent cores and
terminal alkyl chains are organized in separate layers.
of carbosilanes. In the case of polymer 8b the large layer
distance (d = 7.3 nm) is due to the high dilution of the bent-
core mesogenic units by the polymer backbone (there are 15
OSiMe2 units for each bent-core unit), which leads to a
significant expansion of the oligo(siloxane) sublayers.4a
angle of 36u. The outer diffuse scattering of compound 3b is
located at d = 0.44 nm. This value is very typical for the mean
distance between alkyl groups. There is only a very weak
additional diffuse scattering centred on the equator with a
maximum at ca. d = 0.85 nm which can be assigned to the
mean distance between the cyclic oligosiloxane units.
However, for the cyclic siloxanes 3b and 4b, the much larger
layer distances cannot be explained in this way. The doubling
of the layer distance indicates that in these mesophases the
complete oligomesogens act as mesogenic units determining
the layer distance. From the viewpoint of the bent-core units
these mesophases represent double layer structures. These are
the first examples where an amphiphilic molecular structure
provided by silyl groups leads to a double layer organization of
bent-core units, in all previous cases double layer formation
was suppressed by competing steric effects of the silyl groups.§
It is thought that this double layer structure results from the
organization of aliphatic spacer units and terminal alkyl chains
in distinct separate layers (see Fig. 2). In this way the layers of
the bent-core units are separated by two distinct types of non-
aromatic layers, one composed exclusively of aliphatic end-
chains and the other one containing sublayers formed by the
silyl groups and aliphatic sublayers formed by the aliphatic
spacers. In the layers of the terminal alkyl chains partial
overlapping of the chains (intercalation) of adjacent layers is
possible, which leads to stronger coupling of the layers.
The 2D X-ray diffraction pattern of a partially aligned
sample of compound 3b is shown in Fig. 3. The intensity
distribution of the wide-angle scattering indicates a synclinic
tilt of the molecules in the surface-aligned sample with a tilt
In the X-ray diffraction pattern of the SmCPA phase of
the silsesquioxane derivative 4b (see Fig. 4, curve a and ESI{
Fig. S4) there is also only one outer diffuse scattering located
at d = 0.49 nm, but in this case there is an intense additional
diffuse scattering at about 1.4 nm. The diffuse scatterings are
retained in the isotropic liquid state (see Fig. 4, curve b). This
diffuse small angle scattering is attributed to the mean distance
between the silsesquioxane units.13 Hence, in all cases there is a
segregation of the silyl units, indicated by the presence of an
additional diffuse scattering with a position depending on the
size of these units.
The type of smectic phase seems to be influenced by the
flexibility of the silyl units. For molecules with flexible linear
§ In this article monolayer and double layer refers to the organization
of the individual bent aromatic cores and not to the organization of the
complete oligomesogens. Because there is only the possibility of a
partial intercalation of the terminal alkyl chains, but not of the bent
cores, assignment as ‘‘double layer’’ smectic phases is preferred with
respect to the term ‘‘bilayer’’ phases. Double layer structures have also
been observed for bent-core mesogens with a polar terminal group.8
There is also a report about double layer formation in the smectic
phases of symmetric bent-core molecules; in this case a SmCG structure
was assumed to be responsible for a doubling of the layer distance.22
Fig. 4 XRD patterns of 4b: a) SmCPA phase at T = 144 uC with
maxima of the diffuse scatterings at 1.43 nm and 0.49 nm, b) isotropic
liquid at T = 150 uC, c) partly crystallized sample at 138 uC (see also
ESI Fig. S4).
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 3796–3805 | 3799

View Article Online
or branched silyl chains the terminal alkyl chains can easily be
accommodated in the layers formed by the spacer unit which
leads to a monolayer structure with small d values. The cyclic
silyl units provide more constraints which inhibit the mixing of
the spacers and the terminal chains, and hence the layer
distance is nearly doubled.
2.2.2. Textures. Though all mesophases are simple layer
structures without in-plane order and with strongly tilted
organization of the molecules in the layers, the textures and
also the switching behaviour in electric fields are quite different
for the distinct compounds. The textures observed for the
oligomesogens and dendrimers 1/3a,1/4a, 1/6a and 2b strongly
depend on the cooling rate used for investigation. Fast cooling
of the samples (cooling rates .10 K min21) leads to
birefringent non-specific grainy textures (ESI{ Fig. S5b)
whereas slow cooling (cooling rates ,1 K min21) leads to
textures which appear nearly completely dark between crossed
polarizers (see Fig. 5b, this photomicrograph is heavily
overexposed to make the domain structure visible, there are
also few islands with birefringent texture). If polarizer and
analyzer are decrossed, then dark and bright domains become
visible (see Fig. 5a,c). The dark or bright appearance of the
domains reverses if the decrossing direction of polarizer and
analyzer is reversed, indicating a conglomerate of chiral
domains with opposite chirality (dark conglomerate phase).
There are also areas where the brightness does not change.
These domains are discussed as optically inactive, e.g. racemic
structures. For many compounds these optically inactive
regions appear as dark lines at the boundaries between the
domains of opposite chirality sense (crossed polarizers, see
Fig. 5b), due to the mixing of both domains at these interfaces.
However, for the dendrimers 1/na there are also larger
optically inactive areas. These mesophases which are optically
isotropic and show textures composed of chiral as well as
achiral domains are assigned with the superscript ‘‘[*N]’’. The
same combination of chiral and racemic textures was observed
for the dimesogen 5b based on a relatively long oligosiloxane
connecting unit (see Fig. 5d–f) and for the largest dendrimer
1/15a, but in these two cases the dark textures were always
formed, independent of the cooling rates (no birefringent
textures were obtained).
[*N]
Fig. 5 Textures: a–c) of the SmCPA9 phase of 1/4a after cooling
with a rate of 1 K min21 at 105 uC (mixture of dark texture and small
[*N]
low birefringent islands); d–f) of the SmCPFE
phase of the
dimesogen 5b at 120 uC; b,e) between crossed polarizers, a,d) between
slightly decrossed polarizers (ca. 5u) in one direction and c,f) in the
other direction, indicating chiral domains of opposite handedness (+)
and (2) (sign was arbitrarily chosen) and racemic domains (¡). All
photomicrographs are heavily overexposed to make the different
domains clearly visible; moreover the exposure times are different;
between crossed polarizers only the racemic regions are completely
dark, whereas the optically active areas appear brighter due to the
rotation of the plane of polarization; the reason for the bluish
appearance of the chiral domains under decrossed polarizers is not
completely clear, it might be due to light scattering at the disordered
micro domains.
The carbosilane based dimesogens 6a and 7a and the
polymer 8b show dark textures with chiral domains exclu-
sively, i.e. no racemic domains or birefringent textures could
be observed under any condition (assigned with [*]).
The silsesquisoxane derivative 4b exhibits a non-specific low
birefringent texture (ESI{ Fig. S5a) and the cyclic tetrasiloxane
derivative 3b shows a highly birefringent texture (Fig. 6a) as
typically observed for the olefins Ena and Enb without silyl
units. This texture is typical for antiferroelectric switching
(B2 type) SmCPA phases. Remarkably, there is a correlation
between observed texture and switching behaviour.
Classical and non-classical antiferroelectric switching. For the
cyclic siloxane compound 3b with birefringent texture clear
antiferroelectric switching, characterized by the occurrence of
two well separated polarization current response peaks per half
period of the applied triangular wave voltage, was observed
(Fig. 6b).
2.2.3. Electrooptical investigations. All dimesogens, oligo-
mesogens and dendrimers have ferroelectric or antiferroelectric
switching smectic phases with high polarization values, for
most compounds in the range between 650 and 1000 nC cm22
(see Table 1).
This antiferroelectric switching can also be seen in optical
investigations. Slow cooling of the isotropic liquid under a dc
field (¡38 V, 6 mm polyimide-coated ITO cells) yielded two
types of circular domains, A and B (see Fig. 7a). One is highly
3800 | J. Mater. Chem., 2007, 17, 3796–3805
This journal is ß The Royal Society of Chemistry 2007

View Article Online
Fig. 6 SmCPA phase of the cyclic tetrasiloxane 3b: a) strongly
birefringent texture (at T = 135 uC), b) current response to a triangular
wave field (U = 200 Vpp, f = 10 Hz, R = 5 KV, T = 130 uC, 6 mm
polyimide-coated ITO cell), the two peaks clearly indicate tristable (i.e.
antiferroelectric) switching behaviour.
Fig. 7 Tristable switching as observed for the SmCPA phase of the
cyclic siloxane 3b between crossed polarizers at 130 uC (dc field
experiments, cells as used in Fig. 6b, polarizers are horizontal and
vertical) and models showing the organisation of the bent-core units
(views along the bending direction, the apex of the bent cores point
either into the projection plane, indicated by crosses, or out of the
projection plane, indicated by dots); strongly birefringent domains A
and models at the left: homogeneously chiral domain; low birefringent
domains B and models at the right: racemic domain; in the models
red and blue colours indicate the chirality sense of the layers which
results from the fact that layer normal, tilt direction and polar
direction represent either a right-handed or left-handed system (see
Fig. S2 in the ESI).
birefringent with extinction crosses inclined with respect to the
directions of the polarizers (domain A), indicating a synclinic
tilted organization of the molecules (tilt direction is identical in
all layers, SmCsPF structure). The other type (domain B) is of
low birefringence and the dark extinction crosses are parallel
to polarizer and analyzer, indicating an organization with an
equal number of layers with opposite tilt direction. On
terminating the applied field the extinction crosses of the
synclinic domains (A) become low birefringent and relax to
positions which coincide with the direction of the crossed
polarizers to form a uniform texture (Fig. 7b). This confirms
the tristable, i.e. antiferroelectric, switching process, where the
field induced synclinic and synpolar SmCsPF domains (A)
relax to an overall anticlinic and antipolar ground state
structure by rotation on a cone.
axis is parallel to the polarizers. This racemic field induced
structure is assigned as [SmCsPF]aPS and it is composed of
mesoscopic chiral domains with (+)-SmCsPF and (2)-SmCsPF
structure. At zero voltage the [SmCsPF]aPS structure relaxes to
a structure where SmCsPF layer stacks adopt a synclinic and
antipolar correlation. As relaxation of (+)-SmCsPF and (2)-
SmCsPF layer stacks should take place with equal probability
(see arrows in Fig. 7a,b at the right) microdomains with
opposite tilt direction are randomly mixed (leading to a
structure combining [SmCsPF]sPA and [SmCsPF]aPS organiza-
tion, see Fig. 7b, right), so that the average optical axis is
perpendicular to the layer planes and hence the extinction
crosses maintain their position parallel to polarizer and
No change can be seen in the low birefringent domains (B)
with extinction crosses coinciding with the directions of the
polarizers (see Fig. 7b). These domains are usually assigned to
an anticlinic SmCaPF structure, where the tilt direction
changes from layer to layer (see ESI{ Fig. S1). It is also
possible that in these field-induced domains the tilt direction
changes between meso-scale synclinic SmCsPF layer stacks
with anticlinic (SmCaPF) interfaces as already discussed for
other silylated bent-core molecules9e,g and as shown in Fig. 7a
at the right. At each of these interfaces the tilt direction of the
synclinic domains changes, so that the average main optical
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 3796–3805 | 3801

View Article Online
analyzer." The reorganization in the strongly birefringent
domains A can be understood in a similar way by a
reorganization of the uniform SmCsPF structure with forma-
tion of a [SmCsPF]aPA structure at 0 V (see Fig. 7b, model at
the left). The antiferroelectric switching of compound 3b can
satisfactorily be explained with the simple model based on a
SmCsPF or SmCaPF organization. However, the observation
that the same type of optical appearance of the switching
process is found for other silylated compounds having only
one polarization peak is in favour for the model based on layer
stacks. Moreover, this model can satisfactorily explain the
observed transition from antiferroelectric to ferroelectric
switching behaviour (see below). In addition, the organization
in SmCsPF layer stacks is in line with other observations made
in different series of silylated bent-core molecules.9e–h Also the
density distribution of the diffuse scattering in the X-ray
diffraction pattern of 3b (see Fig. 3b) confirms a synclinic tilted
organization in this mesophase. It seems that under the
experimental conditions of the X-ray experiments a locally
unequal distribution of the SmCsPF stacks with opposite tilt
direction is achieved.
[*]
Fig. 8 Current response of the SmCPFE phase of the dimesogen 5b
to an applied triangular wave field (U = 200 Vpp, f = 2 Hz, R = 5 KV,
T = 110 uC, 5 mm ITO cell).
7a, but with shorter spacer length) and it is assigned as non-
classical antiferroelectric switching (SmCPA9 phases).2b For the
cyclosiloxane 3b and the silsesquioxane 4b there are much
smaller layer stacks composed of only a few layers (4b) or
double layers (3b, see model in Fig. 2), and therefore for these
compounds a classical antiferroelectric switching, characterized
by two separate repolarisation current peaks, is observed.
Also the mesophase exhibited by the trimesogen 2b
combining a siloxane linking unit with branchedI and silyl-
substituted alkyl spacers is characterized by a dark mesophase
with domains of opposite handedness. During electrooptical
experiments also for this compound only one peak per half
period of the current response curve could be observed. For
this compound no circular domains could be grown and
therefore no clear decision can be made if this is a non-classical
antiferroelectric switching as found for compounds 1/na or a
ferroelectric switching smectic phase as observed for com-
pounds 5–8 (see next section). Therefore this phase is
The same kind of optical appearance as found for
compound 3b was observed for the switching of the
silsesquioxane derivative 4b and the dendrimers 1/na.
Similarly to the cyclic oligosiloxane 3b, two repolarisation
current peaks were detected for the silsesquioxane 4b, whereas
for all dendrimers 1/na only a single peak was found in the
current response curves. The reason for this single peak is not
completely clear, but it can be explained with a mesophase
structure composed of SmCsPF layer stacks (see above). In
such a superstructure only the interfaces between these stacks
are changed in the switching process. If these stacks are
sufficiently large only a few of these interfaces have to be
changed and the voltage required for switching these layer
stacks becomes small. In addition, the polar interactions with
the substrate surfaces can stabilize the polar (ferroelectric)
organization. Though the polar structure cannot be retained at
0 V the threshold for switching becomes very small and the two
peaks expected for antiferroelectric switching merge to only
one. This type of switching was also observed for some
carbosilane based dimesogens (related to compounds 6a and
[*]
tentatively assigned as a SmCPA9/FE phase.
Ferroelectric switching. A single polarization current peak
(see Fig. 8) was observed for the dimesogens 5b,6a and 7a and
the polymer 8b with high Si/bc ratios. This peak does not split
under a modified triangular wave voltage. For these com-
pounds also the optical investigation of the switching process
indicates ferroelectric switching. As shown in Fig. 9a–c for the
mesophase of the dimesogen 5b with a long oligosiloxane
spacer and in Fig. 9d–f for the siloxane based side-chain
polymer 8a,4a all circular domains have extinction crosses
inclined with the directions of polarizer and analyzer,
indicating a uniform SmCsPF structure. It seems that in these
cases the driving force for a uniform SmCsPF organization is
sufficiently strong that SmCaPF interfaces are completely
suppressed. Even after switching off the applied field the
extinction crosses retain their position, i.e. there is no relaxa-
tion to an antiferroelectric ground state. For the dimesogen 5b
" If an anticlinic SmCaPF structure would be assumed for these field
induced domains, where the tilt direction changes from layer to layer
(see ESI{ Fig. S1), then an AF switching process by rotation on a cone
should lead to a synclinic ground state structure at zero voltage
(SmCsPA, see ESI{Fig. S1). In this structure the extinction brushes
should be inclined with polarizer and analyzer, which is not the case
(see Fig. 7b). However, also in this SmCaPF structure, relaxation can
take place with an equal probability for layers with opposite tilt
direction (for the (+)- and (2)-SmCP layers). This leads to SmCsPA
microdomains with opposite tilt direction which are randomly mixed,
so that the average optical axis is perpendicular to the layer planes and
the extinction crosses can maintain their position parallel to polarizer
and analyzer also in this case. A switching around the long axis is in
principle also possible, but during this switching process the synclinic
structure of the domains A in Fig. 7a should be maintained, i.e. no
rotation of the extinction crosses should be observable. Hence,
reorganization by collective rotation around the long axis is unlikely.
Moreover, all biphenyl derived phenyl benzoate based bent-core
materials 1–8 show a rotation on a cone. Rotation around the long axis
is only observed if a second bulky silyl group is introduced9e or if the
alkyl chain is replaced by a more bulky perfluoroalkyl chain.9d
I The secondary C atoms in this molecule provide a chiral structure
but all stereogenic centres are racemic.
3802 | J. Mater. Chem., 2007, 17, 3796–3805
This journal is ß The Royal Society of Chemistry 2007

View Article Online
related compounds with alkyl spacers. The ether oxygen atoms
in the spacer units can increase the flexibility of these spacers
due to the reduced rotational barriers around the CH2–O
bonds.16 This is assumed to give rise to the reduction of the
transition temperatures.
The comparison in Scheme 1 also indicates that for
molecules incorporating only one bent-core unit the change
of the structure of the spacer units has no effects on the
mesophase type. There also seems to be no influence of the
parity of the spacer units on the mesophase structure. For
example, compounds Sib and Sic shown in Scheme 1 both
have a dark conglomerate phase and show ferroelectric
switching with comparable values of the spontaneous
polarization (Ps = 800–1000 nC cm22). However, an influence
of the type of spacer unit could be found for oligomesogens.
For example, compound 1/4a with a spacer unit incorporating
an ether oxygen has a non-classic antiferroelectric switching
[*N]
dark conglomerate phase (SmCPA9 phase). In contrast 1/4b
with an alkyl spacer6a forms
a ferroelectric switching
mesophase with a texture as typical for SmA phases (fan-like
texture, homeotropic alignment and oily streaks), though polar
order and a tilted organization of the molecules were indicated
by electrooptical investigations and X-ray scattering, respec-
tively. This phase is assigned as SmCRPFE, a ferroelectric
switching smectic phase with overall random tilt direction and
random polar direction.**6a
3. Conclusions
A series of different bent-core based oligomesogens and
dendrimers with oligosiloxane or carbosilane linking units is
compared with respect to the liquid crystalline properties. All
compounds form tilted smectic phases with antiferroelectric
ground state structures. However, the textures and the
switching behaviour are strongly influenced by the size and
shape of the silicon containing units. Antiferroelectric switch-
ing SmCPA phases with birefringent textures are formed by
molecules where the ratio of silicon atoms to bent-core units
(Si/bc ratio, see Table 1) is 1 : 1 or less. With increasing ratio
these birefringent mesophases are replaced by mesophases with
dark texture. Simultaneously antiferroelectric switching is
replaced by ferroelectric switching. For a Si/bc ratio of .1 : 1
to 2 : 1 (see Table 1) the switching is still antiferroelectric
though only one peak is observed in the polarization current
response curves (non-classical antiferroelectric switching,
SmCPA9 phases). For larger ratios ferroelectric switching is
observed. Beside the number of silyl groups, the shape of these
units also has a significant influence. It seems that linear or
branched oligosiloxane or carbosilane units are especially
efficient in inducing optically isotropic mesophases and ferro-
electric switching, whereas compounds 3b and 4b with rigid
cyclic units have a certain tendency to retain birefringent
textures and antiferroelectric switching. This is especially
Fig. 9 Bistable switching as observed between crossed polarizers for
[*]
the SmCPFE phases: a)–c) of the dimesogen 5b at T = 110 uC in a
5 mm ITO cell; d)–f) polymer 8b at T = 95 uC in 6 mm polyimide coated
ITO cell and models showing the reorganisation of the bent-core units:
from a) to b) and from d) to e) there is no rotation of the dark
extinction crosses, whereas there is a rotation of the extinction crosses
from b) to c) and e) to f); the reduced birefringence of the 0 V state of
the polymer 8b4a (see e) indicates some relaxation in the centre of the
sample, whereas there is no relaxation for the dimesogen 5b (see b);
positions of polarizer and analyzer are horizontal and vertical.
there is no reduction of the birefringence (see Fig. 9b) which
indicates a stable ferroelectric ground state structure through-
out the whole cell. For the polymer 8b,4a as for the dimesogens
6a and 7a only a slight reduction of the birefringence is seen for
the 0 V state (see Fig. 9e). It is thought that this reduction of
the birefringence, which takes place without changing the
position of the extinction crosses, is due to the occurrence of
some layer distortion. This means that in the middle of the
sample, where the influence of the surfaces is only weak, a
strongly distorted, and hence optically isotropic, layer struc-
ture similar to that found in the dark conglomerate ground
state textures, is formed.9c
** It is assumed that in this mesophase microscopically sized domains
with distinct tilt direction and polar directions are organized in layers.
Hence this mesophase can be regarded as a less distorted variant of the
dark conglomerate phases, where the polar direction and tilt direction
are randomized, but the layers are still flat and non-distorted. In the
original paper the mesophase was abbreviated as SmCPR.6a
2.2.4. Influence of the spacer unit. As shown in Scheme 1 the
mesophase stability of the bent-core molecules with ether
oxygen in the spacer is generally a bit lower (10–20 K) than for
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 3796–3805 | 3803

View Article Online
evident for the silsesquioxane derivative 4b which shows
classical antiferroelectric switching though the ratio Si/bc is
2.1, a value for which ferroelectric or non-classic (single-peak)
antiferroelectric switching would be expected. It seems that the
flexibility of linear and branched silicon containing groups
provides a major contribution to the special mesophase
modifying properties of the silyl units. The reduced flexibility
of the cyclic silyl units seems to inhibit the intercalation of the
terminal alkyl chains with the layers of the aliphatic spacer
units and this leads to a phase structure where the bent-core
units adopt a double layer organization. In this double layer
structure only every second non-aromatic layer incorporates a
silyl sublayer (see Fig. 2). The presence of a second type of
non-aromatic layer comprising only the terminal alkyl chains
increases the coupling between the layers of the bent aromatic
cores. This is achieved by the possibility of interlayer
fluctuations (fluctuations of the molecules by intercalation of
the terminal alkyl chains) which favour the antiferroelectric
structure. These fluctuations can be suppressed by the oligo-
siloxane layers and these siloxane layers reduce the driving
force for antiferroelectric organization. Hence, the decrease of
the number of siloxane layers changes the switching behaviour
from ferroelectric to antiferroelectric.
polarization in the layers and with increasing number of
layers within the SmCsPF layer stacks. The structure is
homogeneously chiral (SmCsPF) and areas with uniform
handedness can be distinguished as areas with uniform optical
rotation. Macroscopic areas with opposite chirality occur with
equal probability leading to the typical dark conglomerate
texture. However, due to the decoupling of the layers by the
silyl groups the SmCsPF structure is not completely uniform. If
there are exclusively SmCaPA defects a homogeneous chiral
structure is retained. However, if there are additional SmCsPA
defects, the chirality changes at these defects. If there are a
relatively large number of such (racemic) defects, then the
optically isotropic mesophase is achiral.18 In most cases,
however, these achiral domains coexist with homogeneously
chiral domains (see for example Fig. 5). In addition, in some
regions the rotation power of chiral domains seems to be
smaller than in the uniformly chiral domains. This is possible if
there are only few racemic interfaces and the number of layers
with opposite chirality is unequal (optically active, but not
homogeneously chiral domains). The coexistence of chiral and
achiral domains is a typical feature of all dendrimers 1/na.
However it is also observed for dimesogen 5a with a long
oligosiloxane spacer, whereas polymer 8b and the carbosilane
based dimesogens (6a,7a) show exclusively chiral domains.
Hence, the molecular structural reason for the occurrence of
achiral domains in some cases is not completely clear.
The observations made herein completely fit with the
recently developed structural models for the smectic phases
of end-silylated bent-core molecules9c–f and carbosilane
based dimesogens.2b Accordingly, the fundamental structure
is SmCsPF and the stability of this structure is increased with
rising degree of silylation. However, due to the layer
decoupling effect of the silyl groups the SmCsPF structure is
not uniform under all conditions. Often, there are SmCsPF
layer stacks which are separated by defects with another layer
correlation. Under an applied electric field there are SmCaPF
interfaces which, above a certain concentration, give rise to a
change of the position of the extinction brushes in circular
domains from inclined with respect to the positions of
polarizer and analyzer as typically found for uniform
SmCsPF phases to a position parallel to polarizer and analyzer
(see Fig. 7a, domain B). This is found if the size of the uniform
SmCsPF layer stacks becomes smaller than the wavelength of
light. With increased tendency to form a uniform SmCsPF
organization these coinciding extinction brushes are replaced
by inclined extinction brushes (see Fig. 9). Increasing tendency
to form a uniform SmCsPF structure leads to larger SmCsPF
layer stacks which can more efficiently couple to the polar
substrate surfaces which stabilizes the ferroelectric states and
leads to ferroelectric switching. Hence, it is a surface stabilized
ferroelectric switching.
In summary, in a series of papers the influence of silyl units
(oligosiloxanes and carbosilanes) at the periphery of bent-core
mesogens on the occurrence of ferroelectricity and super-
molecular chirality was studied in a systematic way, starting
from single molecules with one9a,c,10a or two silyl end
groups9b,e via dimesogens,2 trimesogens, oligomesogens to
dendrimers6 and side-chain polymers.4a It turned out that
there is a continuous transition from classical antiferroelectric
switching via non-classical antiferroelectric switching to
surface stabilized ferroelectric switching with increasing
decoupling of the layers by the silyl units and that this is
related to a change of the texture from birefringent textures to
chiral and achiral dark textures. There is no strong ‘‘dendrimer
effect’’ or ‘‘polymer effect’’ on the properties, though cyclic
units which reduce the flexibility can compensate the ‘‘silyl
effect’’. All experimental findings are in accordance with a
model where the molecules in these mesophases are organized
in SmCsPF layer stacks with defect layers. Increasing the
degree of silylation is thought to stabilize the SmCsPF
organization and this gives rise to ferroelectric switching by
surface coupling, whereas in the ground state the mesoscopic
polar order in these layer stacks gives rise to layer distortion,
which leads to the dark appearance of the mesophases.
Chirality in these domains is thought to arise from the layer
chirality in the SmCsPF structure.19 However, it is not yet clear
if there is a coupling of this layer chirality with molecular
conformational chirality.1b,20 In the smectic phases of the
biphenyl based bent-core molecules with one silyl-substituted
end the switching process usually takes place by rotation on a
cone, independent of the size of the silyl group. If silyl groups
are attached to both ends,9b,e then switching by rotation
around the long axis is the usual switching process, which
reverses the chirality of the mesophases. In this case there is a
In the absence of an external field the overall structure is
assumed to be antiferroelectric. However, the presence of the
polar SmCsPF layer stacks and the steric layer frustration by
the relatively large silyl groups are responsible for the layer
distortion leading to strongly deformed layers as shown in
Fig. 1b.9c,17 These smectic phases appear dark between crossed
polarizers if the size of the uniformly oriented areas becomes
smaller than the wavelength of light. Hence, without an
external field the polar directions of the ferroelectric clusters
are spread in all directions and the overall structure is apolar.
The tendency for layer deformation increases with rising
3804 | J. Mater. Chem., 2007, 17, 3796–3805
This journal is ß The Royal Society of Chemistry 2007

View Article Online
C. Keith, R. Amaranatha Reddy, A. Hauser, U. Baumeister and
C. Tschierske, J. Am. Chem. Soc., 2006, 16, 3051; (d)
R. Amaranatha Reddy, G. Dantlgraber, U. Baumeister and
C. Tschierske, Angew. Chem., Int. Ed., 2006, 45, 1928; (e)
C. Keith, R. Amaranatha Reddy, U. Baumeister, H. Kresse,
J. Lorenzo Chao, H. Hahn, H. Lang and C. Tschierske, Chem.–
Eur. J., 2007, 13, 2556; (f) C. Keith, G. Dantlgraber,
R. Amaranatha Reddy, U. Baumeister and C. Tschierske, Chem.
Mater., 2007, 19, 694; (g) R. Amaranatha Reddy, U. Baumeister,
C. Keith and C. Tschierske, J. Mater. Chem., 2007, 17, 62; (h)
R. Amaranatha Reddy, U. Baumeister, C. Keith, H. Hahn,
H. Lang and C. Tschierske, Soft Matter, 2007, 2, 558.
strong additional influence of the size of the silyl groups and
the temperature. Large silyl groups or increased temperatures
can lead to the complete loss of polar order, i.e. to a transition
to smectic phases without polar order in the layers and to
modulated smectic phases,9e similar to those seen for rod-like
LC materials.
Hence, the work on silylated bent-core mesogens has
contributed to the understanding of the relation between
molecular structure and properties in this class of compounds.
In order to get comparable results only one type of biphenyl
based bent-core units21 was used in most investigations.
Modification of the bent-core units will lead to a broader
variety of materials with new phase structures and application
properties.1b,9g,h
10 Carbosilanes: (a) C. Keith, R. Amaranatha Reddy, H. Hahn,
H. Lang and C. Tschierske, Chem. Commun., 2004, 1898; (b)
Y. Zhang, M. Wand, M. J. O’Callaghan, C. M. Walker and
W. Thurmes, Ferroelectrics, 2006, 344, 11.
11 (a) J. E. Mark, Acc. Chem. Res., 2004, 37, 946; (b) D Guillon,
M. A. Osipov, S. Mery, M. Siffert, J.-F. Nicoud, C. Bourgogne and
P. Sebastiao, J. Mater. Chem., 2001, 11, 2700.
12 Cyclic oligosiloxane with rod-like mesogenic units: (a)
D. M. Medeiros, M. A. Hale, R. J. P. Hung, J. K. Leitko and
C. G. Wilson, J. Mater. Chem., 1999, 9, 1453; (b) T. E. Mann and
D. Lacey, Liq. Cryst., 2000, 27, 299; (c) T. E. Mann and D. Lacey,
Liq. Cryst., 2003, 30, 1159.
13 Silsesquioxanes with rod-like mesogenic units: (a) G. Mehl and
J. W. Goodby, Angew. Chem., Int. Ed. Engl., 1996, 35, 2641; (b)
G. Mehl and I. M. Saez, Appl. Organomet. Chem., 1999, 13, 261; (c)
R. Elsa¨ßer, G. H. Mehl, J. W. Goodby and D. J. Photinos, Chem.
Commun., 2000, 851; (d) I. M. Saez, J. W. Goodby and
R. M. Richardson, Chem.–Eur. J., 2001, 7, 2758.
14 (a) P. B. Hitchcock, M. F. Lappert and J. W. Warhurst, Angew.
Chem., Int. Ed. Engl., 1991, 30, 438; (b) I. Ojima, in The Chemistry
of Organo Silicon Compounds, ed S. Patai and Z. Rappoport,
John Wiley, New York, 1989; (c) F. Faglioni, M. Blanco, W. A.
Goddart, III and D. Sauders, J. Phys. Chem. B, 2002, 106, 1714.
15 A. M. Muzafarov, O. B. Gorbatsevich, E. A. Rebrov, G. M.
Ignat’eva, T. B. Chenskaya, V. D. Myakushev, A. F. Bulkin and
V. S. Papkov, Vysokomol. Soedin., Ser. A, Ser. B, 1993, 35, 1867.
16 (a) E. Kleinpeter, H. Ko¨hler, C. Tschierske and H. Zaschke,
J. Prakt. Chem., 1988, 330, 484; (b) E. Kleinpeter, H. Ko¨hler,
A. Lunow, C. Tschierske and H. Zaschke, Tetrahedron, 1988, 44,
1609.
17 M. Spannuth, L. Hough, D. Coleman, C. Jones, M. Nakata,
Y. Takanishi, H. Takezoe, J. Watanabe, C. Tschierske,
E. Ko¨rblova, D. Walba and N. A. Clark, poster presented at the
9th International Liquid Crystal Conference, Ljubljana, July 4–9,
2004, Book of Abstracts, SYN-P026.
References
1 (a) D. R. Link, G. Natale, R. Shao, J. E. Maclennan, N. A. Clark,
E. Ko¨rblova and D. M. Walba, Science, 1997, 278, 1924; (b)
R. Amaranatha Reddy and C. Tschierske, J. Mater. Chem., 2006,
16, 907; (c) H. Takezoe and Y. Takanishi, Jpn. J. Appl. Phys.,
Part 2A, 2006, 45, 597.
2 Dimesogens based on Si-containing spacers: (a) G. Dantlgraber,
S. Diele and C. Tschierske, Chem. Commun., 2002, 2768; (b)
C. Keith, R. Amaranatha Reddy, U. Baumeister and C. Tschierske,
J. Mater. Chem., 2006, 16, 3444.
3 (a) B. Kosata, G.-M. Tamba, U. Baumeister, K. Pelz, S. Diele,
G. Pelzl, G. Galli, S. Samaritani, E. V. Agina, N. I. Boiko,
V. P. Shibaev and W. Weissflog, Chem. Mater., 2006, 18, 691; (b)
S. Umadevi, B. K. Sadashiva, H. N. Shreenivasa Murthy and
V. A. Raghunathan, Soft Matter, 2006, 2, 210.
4 Side chain polymers: (a) C. Keith, R. Amaranatha Reddy and
C. Tschierske, Chem. Commun., 2005, 871; (b) R. Achten,
A. Koudijs, M. Giesbers, R. Amaranatha Reddy, T. Verhulst,
C. Tschierske, A. T. M. Marcelis and E. J. R. Sudho¨lter, Liq.
Cryst., 2006, 33, 681.
5 Main chain polymer: (a) C.-D. Keum, A. Kanazawa and T. Ikeda,
Adv. Mater., 2001, 13, 321; (b) S. Demel, C. Slugovc, F. Stelzer,
K. Fodor-Csorba and G. Galli, Macromol. Rapid Commun., 2003,
24, 636; (c) A. C. Sentman and D. L. Gin, Angew. Chem., Int. Ed.,
2003, 42, 1815; (d) E.-J. Choi, J.-C. Ahn, L.-C. Chien, C.-K. Lee,
W.-C. Zin, D.-C. Kim and S.-T. Shin, Macromolecules, 2004, 37,
71; (e) J. Barbera, N. Gimeno, L. Monreal, R. Pinol, M. B. Ros
and J. L. Serrano, J. Am. Chem. Soc., 2004, 126, 7190.
6 Carbosilane dendrimers with bent-core units: (a) G. Dantlgraber,
U. Baumeister, S. Diele, H. Kresse, B. Lu¨hmann, H. Lang and
C. Tschierske, J. Am. Chem. Soc., 2002, 124, 14852; (b) H. Hahn,
C. Keith, H. Lang, R. Amaranatha Reddy and C. Tschierske, Adv.
Mater., 2007, 18, 2629.
18 G. Liao, S. Stojadinovic, G. Pelzl, W. Weissflog, S. Sprunt and
A. Ja´kli, Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat.
Interdiscip. Top., 2005, 72, 021710.
19 (a) L. E. Hough and N. A. Clark, Phys. Rev. Lett., 2005, 95,
107802; (b) L. E. Hough, C. Zhu, M. Nakata, N. Chattham,
G. Dantlgraber, C. Tschierske and N. A. Clark, Phys. Rev. Lett.,
2007, 98, 037802.
20 (a) D. J. Earl, M. A. Osipov, H. Takezoe, Y. Takanishi and
M. R. Wilson, Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat.
Interdiscip. Top., 2005, 71, 021706; (b) D. M. Walba, L. Eshdat,
E. Ko¨rblova and R. K. Shoemaker, Cryst. Growth Des., 2005, 5,
2091.
21 (a) D. Shen, S. Diele, I. Wirth and C. Tschierske, Chem. Commun.,
1998, 2573; (b) D. Shen, A. Pegenau, S. Diele, I. Wirth and
C. Tschierske, J. Am. Chem. Soc., 2000, 122, 1593.
7 Carbosilane dendrimers with calamitic units: S. A. Ponomarenko,
N. I. Boiko, V. P. Shibaev, R. M. Richardson, I. J. Whitehouse,
E. A. Rebrov and A. M. Muzafarov, Macromolecules, 2000, 33,
5549.
8 D. Kardas, M. Prehm, U. Baumeister, D. Pociecha,
R. Amaranatha Reddy, G. H. Mehl and C. Tschierske, J. Mater.
Chem., 2005, 15, 1722.
9 Siloxanes: (a) G. Dantlgraber, A. Eremin, S. Diele, A. Hauser,
H. Kresse, G. Pelzl and C. Tschierske, Angew. Chem., Int. Ed.,
2002, 41, 2408; (b) C. Keith, R. Amaranatha Reddy, U. Baumeister
and C. Tschierske, J. Am. Chem. Soc., 2004, 126, 14312; (c)
22 J. P. Bedel, J. C. Rouillon, J. P. Marcerou, H. T. Nguyen and
M. F. Achard, Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat.
Interdiscip. Top., 2004, 69, 061702.
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 3796–3805 | 3805