Layer frustration, polar order and chirality in liquid crystalline phases of silyl-terminated achiral bent-core molecules

Article abstract of DOI:10.1002/chem.200600876

A novel class of bent-core molecules with oligo(siloxane) or carbosilane units at both ends was synthesized and the self-organization of these molecules was investigated by polarizing microscopy. DSC, X-ray scattering, dielectric and electrooptical method

Full text of DOI:10.1002/chem.200600876

DOI: 10.1002/chem.200600876
Layer Frustration, Polar Order and Chirality in Liquid Crystalline Phases of
Silyl-Terminated Achiral Bent-Core Molecules
Christina Keith,[a] Ramaiahgari Amaranatha Reddy,[a, d] Marko Prehm,[a]
Ute Baumeister,[b] Horst Kresse,[b] Jessie Lorenzo Chao,[b] Harald Hahn,[c]
Heinrich Lang,[c] and Carsten Tschierske*[a]
Abstract: A novel class of bent-core
molecules with oligo(siloxane) or car-
phases is surface stabilized ferroelectric
and, depending on the conditions, two
distinct slow relaxation processes to
nonpolar structures were observed. It
is proposed that the smectic phases are
built up by chiral and polar SmCsPF
layer stacks which are separated by an-
ticlinic interfaces. If the size of these
layer stacks is sufficiently large a cou-
pling to the substrate surfaces takes
place and ferroelectric switching is ob-
served. It is also suggested that the
sponge-like layer distortion, occurring
in the low birefringent mesophases, is
due to an escape from the local polar
order within these SmCsPF layer stacks.
For compounds with larger silylated
units a steric frustration arises, which
leads to layer modulation (columnar
ribbon phases) and this is associated
with a transition from ferroelectric to
antiferroelectric switching. All com-
pounds show a switching of the mole-
cules around the long axis which re-
verses the layer chirality.
AHCTREUNG
bosilane units at both ends was synthe-
sized and the self-organization of these
molecules was investigated by polariz-
ing microscopy, DSC, X-ray scattering,
dielectric and electrooptical methods.
Depending on the size of the silicon-
containing segments, smectic and col-
umnar liquid crystalline phases are
formed. Most smectic phases are low
birefringent and composed of macro-
scopic domains of opposite handedness
(dark conglomerate phases). The
switching process in these smectic
Keywords: ferroelectricity · liquid
crystals · mesophases · self-assem-
bly · supramolecular chemistry
Introduction
The design of programmed molecules which can organize in
a predictable way with formation of complex superstructures
is an attractive target of contemporary research.[1] Crystal
engineering, for example, has significantly advanced during
the last decade by using rigid molecules having a well de-
fined shape and functionality.[2] Such molecules play an im-
portant role in coordination polymers and metal-organic
frameworks,[3] as well as for the construction of polygonal
grids[4] and cage structures.[5] Moreover, rigid segments are
important building blocks for liquid crystalline (LC) materi-
als.[6] In all these cases the shape of the rigid segments and
their appropriate functionalization determine the morpholo-
gy of the resulting self-assembled superstructure. Chevron-
shaped molecules, for example, can be organized in quite
different modes, such as supramolecular rhombuses, hexa-
gons, helical superstructures,[7–9] zigzag patterns[10] (Figure 1)
or cages.[11] Another mode of organization is the formation
of densely packed layers or ribbons with a uniform direction
of the chevrons (see bottom of Figure 1). These systems ex-
hibit LC properties if flexible alkyl chains are attached to
[a] Dr. C. Keith, Dr. R. Amaranatha Reddy, Dipl.-Chem M. Prehm,
Prof. Dr. C. Tschierske
Institute of Organic Chemistry
Martin-Luther-University Halle-Wittenberg
Kurt-Mothes-Strasse 2, 06120 Halle (Germany)
Fax : (+49)345-552-7346
E-mail: carsten.tschierske@chemie.uni-halle.de
[b] Dr. U. Baumeister, Prof. Dr. H. Kresse, Dipl.-Chem J. L. Chao
Institute of Physical Chemistry
Martin-Luther-University Halle-Wittenberg, Mühlpforte
06108 Halle (Germany)
[c] Dr. H. Hahn, Prof. Dr. H. Lang
Faculty of Sciences, Department of Chemistry
Institute of Inorganic Chemistry
Strasse der Nationen 62, 09111 Chemnitz (Germany)
[d] Dr. R. Amaranatha Reddy
present address:
Department of Chemistry and Biochemistry
University of Colorado at Boulder, Boulder, CO 80309-0215 (USA)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author: Additional
figures and table with X-ray data.
2556
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2556 – 2577

FULL PAPER
electric (SmCaPA and SmCsPA). Usually the antiferroelectric
states (SmCaPA and SmCsPA) are more stable and can be
switched into the corresponding ferroelectric states (SmCsPF
and SmCaPF, respectively) under the influence of an external
electric field. Upon switching off the field the polar ferro-
electric (FE) states usually relax back to the nonpolar anti-
ferroelectric (AF) states. This switching process includes
three stable states and is therefore assigned as an AF
switching process. In few special materials, however, the
switching can take place by a direct switching between the
two polar FE states without relaxation to an AF state,
which is assigned as FE switching.[15] Hence, these bent-core
LC provide a new source of polar ordered soft matter with
FE and AF properties.
In addition, the tilted organization of these achiral mole-
cules in polar layers gives rise to a C2 symmetry of the
layers, which lacks a mirror symmetry and therefore these
layers are inherently chiral, providing a new source of supra-
molecular chirality (more specific: superstructural chirali-
ty).[14] Layer normal, tilt direction and polarization vector
describe either a right handed or a left handed system (see
Figure 2c). Changing either the direction of the polarization
Figure 1. Examples of supramolecular structures formed by chevron-like
molecules. These structures can be found in coordination polymers, dis-
crete supramolecular aggregates, etc. The individual molecules can be or-
ganized, for example, by metal coordination, hydrogen bonding or van
der Waals forces.
the ends of the rigid units. Liquid crystals (LC) have impor-
tant application properties as these materials can easily
change their configuration under the influence of different
external stimuli. This unique combination of long range
order and local mobility pro-
vides the foundation of numer-
ous technological applications
of such materials, as for exam-
ple in display devices.[6]
Within the field of LC mate-
rials chevron-shaped molecules,
usually designated as bent-core
LC or banana-shaped LC, are
of special current interest.[12,13]
For such molecules
a polar
order results from the packing
of the bent-core units in layers
where the bent structure re-
stricts their rotation around the
long axis, leading to a directed
packing with a polar direction
(P) parallel to the layer planes,
giving rise to polar smectic
phases (SmP). Remarkably, in
most cases, the molecules are
additionally tilted in these polar
layers and this increases the
structural diversity in these self-
organized systems. In adjacent
Figure 2. Organization of bent-core molecules in polar smectic phases: a) FE and AF switching of bent-core
molecules (side views). b) The four classical arrangements in tilted polar smectic phases. The subscripts “s”
and “a” indicate the correlation of the tilt direction in adjacent layers: s=synclinic means an identical tilt di-
rection; a=anticlinic indicates an opposite tilt direction. Subscripts “A” and “F” indicate the correlation of the
polar direction in adjacent layers. PA is indicative for an antipolar structure (polar direction alternates); this
layers tilt direction and polar structure is macroscopically nonpolar and usually assigned as “antiferroelectric”. PF stands for a synpolar
structure (polar direction is identical); this structure is macroscopically polar and usually assigned as “ferro-
electric” (PF). The grey-scale indicates the chirality of the layers. The SmCaPA and SmCsPF structures are ho-
mogeneously chiral (identical chirality sense in the layers), the SmCsPA and SmCaPF structures are macroscopi-
cally racemic (alternating chirality sense in the layers). Only two layers are shown to represent the structure,
direction can be either identical
or opposite, leading to in total
four different phase structures,
as shown in Figure 2.[14] Two of
but in these classical models the indicated alternation of polar order or tilt direction takes place between each
of the individual layers. c) Origin of the superstructural chirality within the smectic phases of bent-core mole-
cules. Layer normal, tilt direction and the polar axis define either a right-handed coordinate system (+) shown
in dark grey, whereas in the mirror image these vectors define a left-handed system (ꢀ), shown in pale grey.
This code is maintained throughout the manuscript. d) Switching on a cone reverses polar direction and tilt di-
rection and hence, retains the layer chirality. e) Switching around the long axes of the molecules reverses only
them are macroscopically polar,
assigned
as
ferroelectric
(SmCsPF and SmCaPF), the
others are macroscopically non-
polar and assigned as antiferro- the polar direction and in this way reverses layer chirality.
Chem. Eur. J. 2007, 13, 2556 – 2577
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2557

C. Tschierske et al.
or the tilt direction changes the
chirality sense of the layers,
whereas changing both retains
the chirality sense. The SmCsPA
and SmCaPF phases have oppo-
site chirality sense in adjacent
layers, leading to macroscopi-
cally racemic superstructures,
whereas in the SmCaPA and
SmCsPF phases the layers have
identical chirality sense and
hence, these structures are ho-
mogeneously chiral and
a
mirror image exists for each of
these structures (see Figure 2b).
In the (+)- and (ꢀ)-SmCsPF
phases, for example, either the
tilt direction or the polar direc-
tion is opposite (see Figure S1,
Supporting Information). The
polar direction can be reversed
by means of electric fields and
for this collective process there
are two distinct switching mech-
anisms. If the switching takes
place by collective rotation on a
cone, which is usually the case,
tilt direction and polar direction
are changed simultaneously and
Figure 3. Comparison of the structures of distinct types of bent-core molecules: classical bent-core molecules
I,[24] bent-core molecules having only one silyl unit (compounds II),[19a–f] bent-core molecules having silyl and
perfluoroalkyl end groups (compounds III);[19g] the molecules Si1–CSi3, having silyl groups at both ends are re-
ported herein.[16]
layer chirality is preserved (see Figure 2d).[14] However, if
the molecules reorganize by a collective rotation around the
long axis, only the polar direction is reversed and this gives
rise to an inversion of the layer chirality (see Figure 2e).
This provides the possibility of field induced transitions
from a racemic to a uniformly chiral structure and for ho-
mogeneously chiral structures the chirality sense can be re-
versed by means of electric fields.[16,17] This unique combina-
tion of properties makes this new class of LC materials es-
pecially interesting for numerous applications, for example
as switchable ferroelectric, NLO active and chiroptical ma-
terials.[18]
We have recently enhanced the complexity of bent-core
molecules by introduction of oligosiloxane and carbosilane
units.[19] It turned out that such silyl groups have a signifi-
cant impact on the properties of the mesophases of these
molecules. These units change the switching behavior from
AF to FE and also the optical appearance of the mesophase
is completely changed. The highly birefringent textures usu-
ally observed for smectic phases of bent-core molecules with
simple alkyl chains, like I (Figure 3) are replaced by low bi-
refringent or optically isotropic phases which appear dark
between crossed polarizers. Moreover, these mesophases are
composed of a 1:1 mixture of macroscopically chiral do-
mains with opposite handedness.[19] In crystalline systems
this type of racemic structure, where enantiomers are organ-
ized in separate enantiomorphic crystals is termed conglom-
erate.[20] Hence, this type of low-birefringent mesophases
with a local chiral structure has been assigned as “dark con-
glomerate phase” and is indicated by the superscript “[*]”,
which indicates that the mesophase has a chiral superstruc-
ture though the molecules themselves are achiral.[21] These
mesophases have been found for few distinct classes of
bent-core materials,[22,23] but in the class of silylated bent-
core molecules they are dominating.[19] Neither the origin of
ferroelectricity in some of these mesophases, nor the origin
of the spontaneous achiral symmetry breaking with forma-
tion of macroscopic homogeneously chiral domains, nor the
driving forces for layer distortion, leading to optical isotropy
of these smectic phases are fully understood at present.
Herein we propose a model of the “dark conglomerate
phases” found for the silylated bent-core molecules, based
on a mesophase structure composed of chiral and polar meso-
A
arises due to surface coupling which occurs if these polar
SmCsPF layer stacks exceed a certain critical size, and that
layer distortion is due to the escape from the mesoscale po-
larity within these polar layer stacks. The investigations
were done with a new class of bent-core mesogens, namely
compounds Si1–Si3, Si3i and CSi3 with structures shown in
Figure 3. In contrast to the earlier reported compounds with
only one silyl group[19] these molecules have silicon contain-
ing groups at both ends, which has a significant impact upon
their self-organization, due to the segregation, the layer de-
coupling and steric effects of the silyl groups.[19] Some of the
materials show a temperature dependent change of the
2558
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2556 – 2577

Liquid Crystals
FULL PAPER
mode of molecular self organization from a sponge-like net
of layers in the “dark conglomerate phases” to ribbon-like
aggregates in the modulated smectic (columnar phases).
This is associated with a change of the switching process
from FE to AF. Moreover, in most mesophases formed by
these molecules, switching takes place by rotation around
the long axis which switches the chirality of the LC system.
The generalization of these results improves the knowledge
about the self organization principles in polar soft matter
systems, providing fundamental clues for a target specific
design of new bent-core mesogenic materials.
phase (Colr) composed of ribbons with an antipolar order of
the ribbons in a 2D lattice (a=3.9, b=4.9 nm), which is sta-
bilized by the efficient escape from a macroscopic polar
order provided by this organization.[27] As most B1 phases,
also the columnar phase of En does not show any polar
switching. In this Colr phase the terminal chains are in con-
tact with the bent aromatic cores at the interfaces between
the ribbons (see Figure 4c). Silylation of the terminal double
bonds gives rise to an increase of the incompatibility of the
terminal chains with these cores, which destabilizes the
ribbon-like organization. Therefore, this columnar phase is
[
replaced by a layer structure (SmCPFE *]) without these un-
favorable inter-ribbon interfaces, if ethyldimethylsilyl units
(compound Si1) or pentamethyldisiloxane-1-yl units (com-
pound Si2) are attached to both ends. For compounds with
larger silicon-containing groups (heptamethyltrisiloxane
units in compounds Si3 and Si3i[16,28] and carbosilane units
in CSi3), however, a steric frustration arises within the
layers and the layer structure becomes frustrated again. This
steric frustration leads to oblique columnar phases (Colob,
ColobPA), which replace the smectic phases.
Results and Discussion
Synthesis
The synthesis of compound En with two terminal double
bonds was achieved by esterification of 3,4’-biphenyldiol[24]
with 4-[4-(10-undecenyloxy)benzoyloxy]benzoic acid (see
Schem 1).[25,26] In the final step En was hydrosilylated with
appropriate H-silanes according to procedures described
earlier, yielding the siloxane derivatives Si2–Si3, Si3i[16] and
the carbosilanes Si1 and CSi3.[19] Isolation and purification
was achieved by repeated chromatography and crystalliza-
tion.
Smectic phases of compounds Si1 and Si2
Investigation of the ground-state structures: For compounds
Si1 and Si2 X-ray investigations indicate a smectic phase
without in-plane order over the whole mesomorphic temper-
ature range. The layer distance, determined by X-ray scat-
tering, is significantly smaller than the calculated molecular
length and this is in line with a tilted organization of the
molecules in these smectic phases. Only one maximum of
the diffuse scattering at d=0.46 nm is observed for Si1 with
ethyldimethylsilyl end groups (see Figure S2a, Supporting
Information), whereas the wide angle region of Si2 is char-
acterized by a diffuse scattering with two maxima (see
Figure S2b, Supporting Information), one at d=0.45 nm cor-
responds to the mean distance between the alkyl and the
aryl segments, and a second one corresponding to d=
0.63 nm, assigned to the mean distance between the nano-
A
Scheme 1. Synthesis of compounds Si1–Si3, Si3i and CSi3: i) DCC,
organization of Si2 in the smectic phase is shown in Fig-
ure 5b. Due to the segregation of the terminal silyl groups
there are two different mean distances between the hydro-
carbon units (cross-sectional area = 0.21 nm2) and between
the siloxane units (cross-sectional area = 0.38 nm2) which
leads to a steric frustration within these layers. Tilting of the
aromatic units and folding of the alkyl chains can partly
compensate this difference, but nevertheless it can be as-
sumed that the packing of the aromatic cores and alkyl
chains is disturbed in comparison to conventional bent-core
mesogens and in comparison to the monosilylated com-
pounds II where antiparallel packing can compensate this
difference (Figure 5a).[19f,29]
DMAP, CH2Cl2, 2 08C, 24 h;[26] ii) EtMe2SiH (Si1), Me3Si
ACHTREUNG
1,2: Si2, Si3), or (Me3SiO)2MeSiH (Si3i), or Me3Si
ACHTREUNG
(CSi3), Karstedtꢁs catalyst, toluene, 208C, 24 h.[19] DCC= dicyclohexyl-
carbodiimide, DMAP=dimethylaminopyridine.
Investigation of the LCbehaviour—Comparison of the
olefin terminated compound En with the silylated molecules
Sin
The synthesized compounds were investigated by polarizing
microscopy, differential scanning calorimetry (DSC), X-ray
diffraction and by detailed electrooptical investigations. In
Table 1 the influence of the size of the Si-containing units
upon the liquid crystalline properties is shown.
Upon cooling compounds Si1 and Si2 between crossed po-
larizers, the mesophases appear as fractal nuclei which coa-
lesce to grainy unspecific and low-birefringent textures (see
Figure 6b). Upon rotating the analyzer by small angle (5–
The nonsilylated and terminally unsaturated compound
En shows the typical textural features of the so-called B1
phases (see Figure 4a,b). This is a rectangular columnar
Chem. Eur. J. 2007, 13, 2556 – 2577
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2559

C. Tschierske et al.
Table 1. Mesophases, phase transition temperatures, transition enthalpies and lattice parameters of the bent-core molecules under investigation.[a]
R3Si
T/8C
d, a, b/nm, g/8
a=3.9, b=4.9
l/nm[c]
DH/kJmolꢀ1[b]
En
Si1
Si2
Si3
--
Cr 99
43.6
Cr 111
46.2
Cr 80
14.8
Cr 89
10.5
Colr 146 Iso
18.8
5.6
6.2
6.6
7.2
A
EtMe2Si-
Me3SiOMe2Si-
SmCPFE [*] 159 Iso
27.1
4.4 (1208C)
4.5 (95 8C)
SmCPFE [*] 162Iso
26.0
SmCPFE 126
<0.1
Me3Si
A
SmCPA 143[d]
ColobPA 147[e]
–
Colob 155 Iso
20.7
d=4.8 (1208C)
a=5.8, b=5.1,
g=114.5 (1488C)
d=4.7 (1108C)
a=2.5, b=4.65,
g= 107 (1308C)[g]
d=5.1 (1208C)
a=2.8, b=5.1,
g=108 (1448C)
1.0
Si3i
(Me3SiO)2MeSi-
Cr 78
12.7
USmCPA 127
1.3
ColobPA 157 Iso[f]
6.6
7.8
24.0
CSi3
Me3Si
[(CH2)3 Me2Si]2-
Cr 60
7.2
SmCPFE 132[h]
–
SmCPA 139
ColobPA 150 Iso
17.0
1.7
[a] Peak temperatures, determined by DSC, first heating scan, 10 Kminꢀ1; abbreviations: Colr =nonswitchable rectangular columnar phase (B1-type
phase); SmCPFE =FE switching and birefringent polar smectic phase, SmCPFE[*] =FE switching dark conglomerate phase, *] indicates the formation of
[
optically active domains, though the molecules themselves are achiral, the virgin ground-state structure in both SmCPFE phases is assumed to be
[SmCsPF]aPA (see Figure 8c); SmCPA =AF switching smectic phase; USmCPA =AF switching undulated smectic phase; Colob =nonswitching oblique col-
umnar phase; ColobPA =AF switching oblique columnar phase. [b] Values in the lower lines in italics. [c] Determined using CPK models and assuming a
V-shape with a bending angle of 1208 and a most stretched conformation of the rod-like wings (alkyl chains in all-trans conformation). [d] Phase transi-
tion on cooling: 1368C. [e] Determined by electrooptical investigations in a polyimide coated ITO cell (6 mm). [f] See ref. [16], the slightly higher transi-
tion temperatures in comparison to the reference values are due to the higher purity of the 1,1,1,2,3,3,3-heptamethyltrisiloxane used for the repeated syn-
thesis of this compound. [g] See ref. [28]. [h] Determined by electrooptical investigations in a noncoated ITO cell (6 mm) on heating, on cooling the
phase transition is seen at 1278C; a peak at T=1278C (DH=1.4 kJmolꢀ1) is also seen in the DSC cooling curve, but not in the heating scan.
chiral domains with opposite chirality sense. Hence, these
mesophases show the typical features of “dark conglomerate
phases”.
Up to now the precise structure of these “dark conglomer-
ate phases” is not clear. The low birefringence could be ex-
plained with a SmCaPA phase structure with a 458 tilted or-
ganization of the molecules and a bending angle of the aro-
matic cores of about 1098 (orthoconic anticlinic director
structure)[31,32] as recently proposed for a field-induced
“dark-conglomerate phase”.[23c] Also for compounds Si1 and
Figure 4. B1 mesophase of En. a),b) Optical photomicrographs (crossed
Si2 the optically measured tilt angles (see Section on Field-
induced textures) are in the range of 40–458. However, in
contrast to the field induced “dark conglomerate phase” re-
ported in ref. [23c] these mesophases represent virgin struc-
tures, obtained without previous application of an electric
field and hence their structure could be different. A low bi-
refringent mesophase can alternatively result from a layer
distortion which leads to a mesoscopically disordered struc-
ture, either composed of randomly ordered fragments of
these layers,[33] or due to a random (sponge-like) deforma-
polarizers) obtained at 1448C. c) Possible model of the organization of
compounds En in the B1 phase; the terminal chains (not shown) fill the
space between the ribbons formed by the bend aromatic cores.
108) dark and bright domains become visible (see Fig-
ure 6a). If the analyzer is rotated into the opposite direction
the brightness of the domains is reversed (see Figure 6c),
but the light transmission does not change if the sample is
rotated.[30] This indicates a conglomerate of macroscopically
2560
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2556 – 2577

Liquid Crystals
FULL PAPER
Figure 5. Models of the organization of the siloxane substituted bent-core molecules a) antiparallel packing of the monosilylated compounds II. b) Pack-
ing of the molecules Sin and CSi3 with silyl units at both ends as proposed for the arrangement of the molecules in the smectic phases (SmCPFE
,
SmCPFE[*] and SmCPA) and in the ribbons of the Colob phases. c) Model of the sponge phase composed of deformed stacks of layers (only one layer is
shown).
al polar smectic phases of other bent-core molecules is also
in line with the proposed sponge-like 3D organization.[35]
As explained in the introduction, chirality in the polar
smectic phases of achiral bent-core molecules arises from
the tilted organization of the molecules in polar layers (see
Figure 2). In the SmCaPA and SmCsPF phases the layers have
Figure 6. SmCPFE[*] phase of compound Si2 as seen at 1508C under a po-
larizing microscope a) between slightly uncrossed polarizers (arrows at
identical chirality sense and hence, these structures are ho-
mogeneously chiral. Hough and Clark[36] proposed that the
optical activity of the “dark conglomerate phases” arises
from this layer chirality.[37] It requires the phase structure
being homogeneously chiral and the birefringence of the
mesophase itself is sufficiently small. This means that the
“dark conglomerate phases” can either have an anticlinic
and antiferroelectric SmCaPA structure or a synclinic and fer-
roelectric SmCsPF structure whereas the SmCsPA and
SmCaPF phases are racemic and therefore optically inactive.
Hence, only the kind of tilt correlation or the type of polar
correlation should be known in order to assign the meso-
phase type, but this turned out to be very difficult. It is not
possible to determine the layer correlation of the smectic
phases under discussion by X-ray scattering experiments, be-
cause the distorted structure of these mesophases does not
allow the growth of uniformly aligned samples for recording
2D diffraction patterns. Optical investigations also give no
direct insight into the organization in the ground state, be-
cause well developed birefringent textures can only be ob-
tained if the mesophases were grown under a DC electric
field, which stabilizes the synpolar (FE) organization of the
molecules. This means that the original mesophase structure
is changed during investigation.
However, in the case of compounds Si1 and Si2, in con-
trast to the related compounds II with only one silicon con-
taining end group, shearing of the sample removes the chiral
domains and gives rise to a birefringent texture with oily
streaks and a schlieren texture between the streaks. The
schlieren texture indicates an optically biaxial smectic phase,
and the relatively low birefringence of the homeotropically
aligned regions in the texture of Si2 (see Figure 6e, regions
between the stripes) suggests an anticlinic tilt correlation in
this mesophase. This would be in favor of a SmCaPA struc-
ture, if the classical models shown in Figure 2b are consid-
the right indicate the positions of the polarizers in a)–e). b) Same region
between crossed polarizers. c) Same region between slightly uncrossed
polarizers, direction of the polarizers is reversed with respect to a).
d) Coexistence of the conglomerate texture and oily streaks at the begin-
ning of shearing; e) oily streaks and low birefringent schlieren-texture
after completion of shearing.
tion of the layers.[19f,34] In this case optical isotropy is inde-
pendent of the tilt angle and it is possible for all polar smec-
tic phases shown in Figure 2b if the size of the domains with
uniform director orientation is smaller than the wave length
of light. There are indeed strong arguments that the “dark
conglomerate phases” of the silylated bent-core molecules
represent smectic phases with a strongly folded and nonre-
gular organization of the smectic layers, similar to sponge
phases known from lyotropic systems (see Figure 5c).[19f,34]
At first, the layer reflections are not resolution limited and
indicate a relatively short correlation length of the smectic
layers. Secondly, in the absence of an electric field aligned
samples could not be obtained for any of these smectic mes-
ophases. Thirdly, the distorted structure is also obvious from
the microscopic observation of the growing process, which is
characterized by a fractal growing with a grainy and nonspe-
cific appearance of the mesophase which is not typical for
smectic phases with flat layers. Finally, in the case of a com-
pound of type II, with a trisiloxane unit at only one end
saddle-splay structures, confirming the sponge-like structure,
were visualized by TEM of freeze-fracture samples.[34] The
typical properties of compounds Si1 and Si2, seen in X-ray
diffraction and optical experiments are very similar to those
observed for related monosilylated bent-core molecules II
and therefore a sponge-like distorted structure is assumed
also for the “dark conglomerate phases” of these com-
pounds. Moreover, the significantly higher viscosity of the
mesophases of Si1 and Si2 in comparison to the convention-
Chem. Eur. J. 2007, 13, 2556 – 2577
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2561

C. Tschierske et al.
ered. However, the switching behavior is ferroelectric,
which is in conflict with a classical SmCaPA phase structure
as discussed on the basis of the results of electrooptical in-
vestigations in the next Section.
synclinic organization would lead to a more birefringent
schlieren texture in the shear-induced homeotropic regions
(see Figure 6e). Investigations of carbosilane-derived bent-
core mesogenic dimers[19h] have shown that with increasing
number of dimethylsilyl groups within the spacer units,
there is a continuous transition from typical AF to surface-
stabilized FE switching and it was proposed that this is due
to a transition from an organization of alternating single
layers (classical SmCaPA structure as shown in Figure 2b) to
an alternating organization of stacks of synpolar layers.[19h]
Hence, a possible ground-state structure of compounds Si1
and Si2 could also be composed of SmCsPF clusters with a
mesoscopic size. Figure 8c shows such a structure composed
of SmCsPF layer stacks with an anticlinic and antipolar cor-
relation between them ([SmCsPF]aPA structure).[40] This
Electrooptical investigations of the smectic phases of com-
pounds Si1 and Si2: Switching experiments in electric fields
can give additional information concerning the organization
of the molecules.[14] The switching behavior under an applied
triangular wave field and the optical observation of the tex-
tural changes in DC-field experiments are important tools.
Switching process under a triangular wave field: For the
smectic phases of compounds Si1 and Si2 only one sharp
single current response peak in the half period of a triangu-
lar-wave voltage was found over the whole mesomorphic
temperature range (Figure 7a). The calculated polarization
mesophase would have a low birefringence if the size of the
A
SmCsPF layer stacks is sufficiently small. Because only the
layer correlation of the interfaces between the layer stacks
is changed during the switching process, and not the interfa-
ces between each layer as in the classical models, a compa-
ratively low voltage is required for switching. Hence, the FE
states can be more easily stabilized by polar interactions
with the substrate surface, leading to a bistable switching be-
tween the two surface stabilized FE states which give rise to
only a single peak in the switching current curves. This
means that the virgin ground-state structure could be overall
AF and the actually observed switching is surface stabilized
FE. In order to get additional information optical investiga-
tions of the switching process were carried out.
Field-induced textures: The textures of compounds Si1 and
Si2 are significantly changed under the influence of external
electric fields. Compound Si1 will be discussed in more
detail. For this compound, after application of an AC or DC
field, the “dark conglomerate texture” becomes strongly bi-
refringent and the chiral domains disappear. However, the
textures obtained in this way are nonspecific. Well devel-
oped textures with circular domains were obtained on cool-
ing Si1 from the isotropic liquid state under an applied elec-
tric field above a certain threshold voltage. This means that
under the electric field the sponge-like disordered layer
structure of the ground state is unfolded and transformed
into nondistorted layers.[19,22] Because the layer polarization
is parallel to the layer planes (along the bend-direction) the
electric field, which is applied between the surfaces, induces
a FE state with bookshelf geometry where the layers are
perpendicular to the substrates. In the field induced circular
domains the layers form concentric cylinders with the layer
normal parallel to the substrates.
The actually observed type of circular domains depends
on the characteristics of the electric field as well as on the
applied voltage. Under a triangular wave AC field, inde-
pendent of the voltage, exclusively low birefringent (grey)
domains with extinction crosses parallel to polarizer and an-
alyzer are formed (see Figure 10a). This indicates an anti-
clinic organization (either between single layers or between
SmCsPF layer-stacks). The situation is different if a DC field
Figure 7. Switching current response of compound Si1 in a noncoated
6 mm ITO cell (100 Vpp, 10 Hz, 5 kW) on applying a) an alternating
simple and b) modified triangular-wave voltage at T=1508C. This type
of curves can be observed over the whole temperature range of the mes-
ophase between 65 and 1508C.
values are around 850 nCcmꢀ2. A modification of the trian-
gular-wave field by a delay at zero voltage,[38] using noncoat-
ed ITO cells,[39] leaves the switching behavior unchanged,
only one sharp peak is observed at all temperatures (Fig-
ure 7b). This means that immediately after zero voltage
crossing a direct switching between the FE states with oppo-
site polarity takes place and this is a strong indication of a
bistable, that is, FE or FE-like switching process.
This FE switching suggests a synclinic SmCsPF structure
(see Figure 8d) as a possible ground-state organization for
the smectic phases of compounds Si1 and Si2. However, this
2562
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2556 – 2577

Liquid Crystals
FULL PAPER
crosses are inclined with the di-
rections of the polarizers by
about 40–458. Hence, the mole-
cules in these domains seem to
have 40–458 tilted synclinic or-
ganization. However, there are
many domains with lower bire-
fringence which have extinction
crosses with smaller angles (e.g.
domain B)[41] and also few very
low birefringent domains with
extinction crosses parallel to
the directions of the polarizers
(C) can be found. The occur-
rence of these quite different
domains can be interpreted on
the basis of a structure com-
posed of SmCsPF layer stacks
which are separated by anticlin-
ic and synpolar interfaces
(SmCaPF interfaces), as shown
in Figure 8b. Here, the tilt di-
rection changes between adja-
cent layer stacks. Due to the
coupling of tilt direction and
layer chirality, this structure can
also be regarded as an alternat-
ing array of enantiomeric (+)-
and (ꢀ)-SmCsPF layer stacks. If
there is an equal number and
size of the oppositely tilted
SmCsPF stacks, and if the size of
these stacks is smaller than the
wave length of light, then the
position of the extinction
crosses coincides with the posi-
tions of polarizer and analyzer
and the birefringence is low
Figure 8. Proposed modes of organization of compound Si1 depending on the conditions. The assignment of
the phases is shown at the bottom, where the structure within the layer-stacks (SmCsPF) is given in square
brackets and the following subscripts define the correlation of the interfaces between the layer-stacks (a=anti-
clinic, s=synclinic, A=antipolar, S=synpolar). a) Synclinic and antipolar [SmCsPF]sPA structure as obtained in
the relaxation process after switching off an applied AC voltage at T < 1108C. b) Apparently anticlinic and
synpolar [SmCsPF]aPS structure as obtained by cooling Si1 under an applied AC-voltage, both structures are
racemic. c) Two enantiomeric [SmCsPF]aPA structures as obtained either by cooling from the isotropic liquid
state without applied field or in the relaxation process after switching off an applied AC voltage at T >
1108C, it seems that the formation of this structure is associated with a layer deformation into a dark texture
for which a sponge-like deformation of the layers is proposed; the chiral domains are microscopic in size (the
phase appears optically inactive) if the phase is formed by relaxation from structure b) and macroscopic chiral
upon cooling from the isotropic liquid (dark conglomerate texture). d) Uniform synclinic and synpolar SmCsPF
structure as obtained upon cooling under a DC electric field with sufficient strength. The position of the ex-
tinction crosses with respect to the positions of polarizer and analyzer (arrows), as seen between crossed polar-
izers is shown at the top for each of the structures. The fundamental modes of organization shown in this
Figure are also valid for the other compounds Sin and for CSi3.
is applied. Here the position of the extinction crosses addi-
tionally depends on the applied voltage. At low voltage
(0.5 Vmmꢀ1) a low birefringent texture with coincident ex-
tinction crosses (as observed under an AC field) is formed
exclusively. As the voltage is increased (0.9 Vmmꢀ1), few
highly birefringent synclinic domains appear, where the ex-
tinction crosses are inclined with the positions of polarizer
and analyzer. These synclinic domains become the majority
at higher voltage and finally (at 3.5 Vmmꢀ1) replace nearly
all other domains (Figure 9a). Interestingly, the tilt angle of
the extinction crosses is not uniform and once formed
during the growing process under a certain voltage it cannot
be modified by increasing the voltage.
The inclined position of the majority of dark extinction
brushes obtained under a DC field (Figure 9a) indicates a
field induced synpolar smectic phase with a predominately
uniform tilt direction of the molecules in adjacent layers
(SmCsPF phase). In the highly birefringent domains (e.g.
domain A with blue birefringence color) the extinction
(Figure 9a, domain C). This racemic structure is assigned as
[SmCsPF]aPS.[40] If there is a nonequal distribution of oppo-
sitely tilted (+)- and (ꢀ)-SmCsPF stacks, then, depending on
their ratio, the extinction crosses can adopt different angles
with respect to the polarizers (domain B). This structure,
which is nonhomogeneously chiral is assigned as (X)-
[SmCsPF]aPS, where (X) indicates the configurational inho-
mogeneity of the system.[42] If the uniform SmCsPF stacks
have a macroscopic size, then the position of the extinction
crosses becomes ca 458 inclined with the positions of the po-
larizers and synclinic domains (either (+)- or (ꢀ)-SmCsPF)
can be observed (domain A).
This model is identical with a model proposed by Folcia
et al.[43] The authors suggest that the classical model of the
SmCaPF structure, where polar layers strictly alternate (see
Figure 2b) would be in conflict with the Curie principle and
therefore a regular domain structure composed of alternat-
ing SmCsPF layer stacks was proposed for the field induced
SmCaPF structure of bulk samples.
Chem. Eur. J. 2007, 13, 2556 – 2577
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2563

C. Tschierske et al.
Figure 9. Field-induced textures of compound Si1 as seen between
crossed polarizers (polarizer and analyzer are positioned horizontally and
vertically, respectively); a),b) ring-like circular domain structure after
cooling the mesophase (1 Kminꢀ1) under an applied DC field (12V) at
1588C (noncoated ITO cells, 6 mm); dark areas are residues of the iso-
tropic liquid phase. c) Low birefringent texture as obtained after switch-
ing off the external electric field at 1508C. d) Texture as obtained after
application of the electric field with opposite sign; the angle of the ex-
tinction crosses with respect to the position of the polarizers is changed
as seen by comparison with the dotted line which indicates the initial po-
sition of the extinction crosses in b).
Under a stereochemical point of view the [SmCsPF]aPS or-
ganization can be regarded as pseudoracemate, which is a
racemic form between the conglomerates (SmCsPF and
SmCaPA phases = association of uniformly chiral layers)
and the racemic structures (SmCsPA and SmCaPF phases =
association of layers with unlike chirality).[20] This means
that the energetic difference for the association of layers
with the same or different chirality sense is relatively small.
It seems that for the compounds under discussion the ener-
getic difference between synclinic and anticlinic organiza-
tion could indeed be rather small, due to the bulky Si-con-
taining groups at both ends which can decouple the individ-
ual layers by the nearly isotropic oligosiloxane sublayers.
Therefore, it is proposed that the SmCsPF structure is the
preferred fundamental organization in the mesophases of
the silylated bent-core molecules,[19f] but also some anticlinic
and/or antipolar interlayer interfaces coexist in a thermody-
namic equilibrium. This gives rise to a variety of different
textures. The growing of the mesophase under an electric
DC field provides a preferred uniform tilt direction due to
surface alignment, leading to a texture with predominately
inclined extinction crosses. Depending on the strength of
these alignment forces and the uniformity of the resulting
domains, the observed extinction crosses can adopt different
Figure 10. Textures of compound Si1 as seen between crossed polarizers
(polarizer and analyzer are positioned horizontally and vertically, respec-
tively); a) low birefringent texture with apparently anticlinic domains as
obtained under the applied AC field (6 mm noncoated ITO cell, 100 Vpp
,
f=10 Hz) at all temperatures. b) Dark texture ca. 10 s after switching off
the field at T=1408C. c) Highly birefringent texture characterized by
synclinic domains ca. 10 s after switching off the field at T=808C (the
brightness of a–c is scaled differently); application of an AC field to the
textures shown in b) and c) immediately restores the texture a); the
insets in a)–c) show the textures at the same temperatures, but in poly-
imide coated ITO cells (6 mm, 120 Vpp, f=100 Hz).
positions (see texture in Figure 9a). However, under a dy-
namic AC field a relatively large number of anticlinic layers
is formed and this gives rise to a reduction of the size of the
SmCsPF layer stacks and exclusively a low birefringent tex-
ture with extinction crosses parallel to the polarizers is ob-
served (see Figure 10a).
Optical observation of the switching process: In a texture
composed of predominately synclinic domains as obtained
for Si1 under a DC field (Figure 9b), switching off the ap-
2564
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2556 – 2577

Liquid Crystals
FULL PAPER
plied DC field at 1508C distorts the circular domains. The
overall birefringence is strongly reduced and some regions
become completely isotropic. However, the position of the
extinction crosses in the residual birefringent rings is not
changed (see Figure 9c). Application of an electric DC field
with opposite direction restores the birefringent circular do-
mains, but the birefringence and the angle of the extinction
crosses with respect to the polarizers are reduced in compar-
ison to the starting situation (Figure 9d). This indicates that
the field induced and surface stabilized synclinic structure is
partly lost if the alignment field is removed. After several
switching cycles (usually less than 10) all extinction crosses
adopt positions completely parallel to polarizer and ana-
lyzer. A possible explanation could be that after removal of
the external field a partial relaxation of the nearly uniform
SmCsPF structure (see Figure 8d) to an antipolar and anti-
clinic [SmCsPF]aPA structure (see Figure 8c) takes place by
rotation on a cone. This locally polar structure (SmCsPF-
layer stacks) adopts a sponge-like distorted structure which
is optically isotropic and reduces the overall birefringence at
0 V. In addition, this relaxation seems to be irreversible and
once the [SmCsPF]aPA structure is formed it can only be
switched into the corresponding synpolar and anticlinic
domain structure ([SmCsPF]aPS, see Figure 8b) and this re-
quires a rotation around the long axis. Hence, depending on
the mesophase structure there are two distinct switching/re-
laxation processes. It also seems that the relaxation process
does not take place in the whole sample. Whereas the uni-
form SmCsPF structure is retained at parts of the surfaces,
there is a relaxation in the center of the sample. For this
reason the position of the extinction crosses does not
change after switching off the field, though the birefringence
is reduced. After switching on the field again, the residues
of the surface stabilized SmCsPF structure (inclined extinc-
tion crosses) and the field induced [SmCsPF]aPS structure
(coinciding extinction crosses) coexist and this gives rise to a
reduced average angle of the extinction crosses with respect
to the polarizers. Apparently, the number of uniformly tilted
SmCsPF layers at the surface is reduced in each switching
cycle and this leads to a reduction of the angle of the extinc-
tion crosses with respect to the polarizers until they coincide
with polarizer and analyzer. At lower temperatures (T <
1108C) the switching process becomes more complicated
and highly birefringent domains occur beside the low bire-
fringent domains.
optical isotropic regions. The growth of the dark areas sets
in immediately after switching off the field and, depending
on the temperature and the frequency of the applied
field,[44] takes several seconds to minutes until the final com-
pletely dark state is reached. This type of optical response
can be assigned to a relaxation of the surface stabilized FE
and racemic [SmCsPF]aPS structure (Figure 8b) into a disor-
dered structure, which is macroscopically nonpolar and ap-
pears optically isotropic. It is assumed that the relaxation
leads to an overall antiferroelectric [SmCsPF]aPA structure
(see Figure 8c) by a collective rotation of the molecules
around their long axes. As shown in Figure 8b,c there is an
equal probability of relaxation of the (ꢀ)-SmCsPF layer
stacks with formation of a (+)-[SmCsPF]aPA structure (follow
the dashed line i in Figure 8) and of (+)-SmCsPF layer
stacks with formation of a (ꢀ)-[SmCsPF]aPA structure (follow
the dashed line ii in Figure 8). Hence, the reorganization of
the SmCsPF layer stacks into an overall AF structure leads
to a macroscopically racemic mixture of microscopically
chiral (+)- and (ꢀ)-(SmCsPF)aPA domains. This means that
after relaxation, the overall structure should be similar to
the virgin ground-state structure, that is, the flat layers are
folded into a sponge-like structure, which is optically iso-
tropic. However, no chiral domains could be detected, be-
cause the size of the uniformly chiral domains obtained
after switching off the field is too small to be distinguished
optically. It seems that macroscopic chiral domains (“dark
conglomerate textures”) are only formed by growing this
mesophase from the isotropic liquid.
As the temperature is reduced, the dark regions formed
after switching off the field become smaller, the relaxation
process becomes slower and no change of the texture can be
seen if the field is switched off at temperatures close to
1108C. It seems that at reduced temperature the more dense
packing of the bent cores, coupled with an increased viscosi-
ty makes the rotation around the long axis more difficult.
At T < 1108C another relaxation process is observed. At
this temperature a slow transition to a strongly birefringent
texture with extinction brushes inclined with polarizer and
analyzer takes place in the zero-field state, indicating a tran-
sition to
a purely synclinic phase structure (see Fig-
ure 10c).[44] A relaxation of the [SmCsPF]aPS phase (Fig-
ure 8b) into a [SmCsPF]sPA structure[40] shown in Figure 8a,
with synclinic but antipolar boundaries is a possible mecha-
nism of this relaxation. The [SmCsPF]sPA structure is a mac-
roscopically nonpolar (AF-like) and racemic structure. Re-
application of the electric field immediately restores the low
birefringent texture with coincident extinction crosses (Fig-
ure 10a) as typical for the field induced [SmCsPF]aPS phase
(Figure 8b). This change of the orientation of the extinction
crosses indicates that at this temperature the relaxation pro-
cess should take place by rotation of the molecules on a
cone.
In order to clarify the situation, temperature dependent
electrooptical experiments were carried out with textures
grown under an AC field in a noncoated 6 mm ITO cell.
Under these conditions only extinction crosses coinciding
with the positions of the polarizers were obtained (see Fig-
ure 10a, proposed structure is [SmCsPF]aPS). Depending on
the temperature two slow relaxations to other textures can
be observed. At temperatures above 1108C the dark extinc-
tion crosses do not change their position if the applied volt-
age is switched off, but the brightness of large areas of the
texture significantly decreases (see Figure 10b) and in most
regions the birefringent texture is completely replaced by
Hence, the switching is surface stabilized FE at all tem-
peratures. However, the FE states are metastable and after
switching off the external field, two distinct relaxation pro-
cesses to macroscopically nonpolar structures take place.
Chem. Eur. J. 2007, 13, 2556 – 2577
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2565

C. Tschierske et al.
Depending on the temperature the relaxation process takes
place either by collective rotation around the long axis or
on a cone, which gives rise to the distinct appearance of the
macroscopically nonpolar states. Because these relaxation
processes are slow always only a single peak, resulting from
the fast surface-stabilized FE switching process is found in
the switching experiments under a triangular wave field.
Furthermore, it seems that the formation of the sponge-like
dark texture is only possible if the SmCsPF layer stacks
adopt an anticlinic and antipolar correlation, that is, if the
overall phase structure is homogeneously chiral. If an anti-
polar but synclinic structure, that is, a racemic structure is
formed (at T < 1108C) the birefringence is retained. For
this reason, it seems that the formation of the dark conglom-
erate structure is bound either to a chiral superstructure or
to the presence of some anticlinic interfaces.
Influence of surfaces: Switching in polyimide-coated mea-
surement cells: It should be pointed out that surface interac-
tions are of great importance for the actually observed
switching behavior, and therefore any change of these surfa-
ces can lead to different results. In polyimide-coated ITO
cells, for example, no relaxation to a low birefringent texture
can be observed at any temperature (see inset in Fig-
ure 10b), and the relaxation to the synclinic structure is also
much slower. In addition, there is also an influence of these
surface layers on the repolarization current curves. Only one
switching current peak is observed under all conditions in
ITO cells without additional coating layer. However, if poly-
imide-coated cells are used, the repolarization current
curves are different. In these cells also a single peak is ob-
served under a simple triangular wave field. However,
under a modified triangular-wave field (when a delay is in-
troduced at zero voltage) the single peak splits into two
nearly identical peaks at high temperature, close to the
clearing temperature. Upon reducing the temperature the
two peaks become unequal in size and only at sufficiently
low temperature a single peak is found (see Figure 11a–c).
This indicates that at high temperature the switching occurs
around zero-voltage crossing[45] whereas at low temperature
switching occurs after sign reversal of the applied electric
field (as observed in the noncoated cells at all tempera-
tures).
Figure 11. Switching current response curves of compound Si1 in a 6 mm
polyimide coated ITO cell (ꢁ 105 V, 1 Hz, 5 kW) on applying an alternat-
ing simple (regime of the downward current peak) and modified triangu-
lar-wave voltage (regime of the upward current peak/s) at a) T=1408C,
b) T=858C, c) T=658C.
The influence of the polyimide coating layer can be un-
derstood if it is considered that the internal electric field is
influenced by the electric double layer formed at the surfa-
ces. The field produced by this double layer is opposite to
the applied external electric field, and hence, the internal
field actually interacting with the LC is different from the
applied (external) field (see Figure 12).[46]
Figure 12. Influence of the electric double layer on the effective internal
field (E) a) with applied external field and b) after switching off the ex-
ternal field.
Vꢀ2d0=e0 s þ P0
ness (d’) and the dielectric constant (e’) of the insulating
alignment layer, on the related properties of the LC (d, e)
and on the surface density (d) of ionic charges (P0).[46] Thus,
the effect of the double layer is especially strong in ITO
cells with an additional insulating polyimide alignment
layer, if ionic impurities are present, and if the threshold
ð1Þ
E ¼
d þ 2d0ðe=e0Þ
e
According to Equation (1) the strength of the resulting in-
ternal electric field (E) depends not only on cell thickness
(d) and applied external voltage (V), but also on the thick-
2566
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2556 – 2577

Liquid Crystals
FULL PAPER
voltage of the switching process itself is low, as observed for
many of the Si-containing bent-core molecules. Because the
induced internal field is opposite in sign to the external
field, the internal zero voltage is obtained before terminat-
ing the external field and the sign of the internal field is al-
ready reversed at zero external field (see Figure 12).[46] If
the threshold voltage (Vth) required for the switching pro-
cess is smaller than the internally induced voltage (Vint), the
around the molecular long axis at all temperatures. This
should be due to the additional space available between the
bent aromatic cores, as a consequence of the more bulky si-
lylated end groups in this compound.
Smectic and columnar phases of compounds Si3, Si3i and
CSi3
switching process starts before crossing zero voltage (Vint
>
Ground-state structures of Si3: Compound Si3 with a trisi-
loxane unit shows a completely different and more complex
behavior than Si1 and Si2 with smaller Si-containing units.
On cooling from the isotropic liquid state the mesophase
grows with a strongly birefringent texture containing sharp
lines and leaf-like patterns and these features are indicative
of a columnar phase. On further cooling, a phase transition
can be observed at about 1368C, which is seen by a signifi-
cant decrease of the birefringence and viscosity (see
Figure 13a,b). Upon heating this transition takes place at
1438C (see Figure 13c). Hence, the high temperature phase
can be significantly supercooled. On further cooling, at
1268C there is an additional transition which is reversible
and indicated only by a slight increase of the birefringence.
At this temperature the switching-process changes from AF
to FE (see next Section), but there is no clear peak in the
DSC curves (Figure 13c) associated with this transition.
Over the whole mesomorphic region the two-dimensional
X-ray diffraction pattern of an aligned sample shows two
Vth). In this case the peak splits into two in a modified trian-
gular wave field. The splitting is clearly visible at high tem-
perature and gradually disappears on decreasing tempera-
ture as the threshold voltage increases at low temperature
and Vth becomes larger than Vint. This explains why in the
switching experiments with Si1 in polyimide-coated ITO
cells and at high temperatures no relaxation to a dark tex-
ture can be found upon terminating the field (see insets in
Figure 10a,b). At zero external voltage the induced field
switches some of the FE layers into the FE state with oppo-
site polarity (peak before 0 V crossing). As switching takes
place around the long axis, it does not change the texture,
but it decreases the overall macroscopic polarity and this
seems to reduce the driving force for the slow relaxation
processes to the dark texture ([SmCsPF]aPA structure). At re-
duced temperature the uniform polar structure is largely re-
tained (only one peak after 0 V crossing) and a slow relaxa-
tion to the synclinic AF structure can be observed (see inset
in Figure 10c). More generally, it is important to consider
the effects of alignment layers in electrooptical investiga-
tions to exclude misinterpretations.
Electrooptical investigations of Si2: The switching behavior
of the smectic phase of Si2 is similar to that observed for
Si1. For this compound, too, only one sharp current re-
sponse peak in the half period of a triangular-wave voltage
was found which only splits under a modified triangular
wave voltage if polyimide coated ITO cells were used. Opti-
cal experiments show that the conglomerate texture is re-
placed by birefringent textures upon application of an elec-
tric field, but the textures are less well developed than in
the case of Si1. The most obvious difference is seen in
switching experiments, where, independent on the condi-
tions, coinciding domains ([SmCsPF]aPS) always coexist with
inclined (uniform SmCsPF) domains, indicating a stronger
preference for a uniform synclinic organization. In addition,
no relaxation can be observed in the anticlinic domains, nei-
ther to the dark texture, nor to highly birefringent synclinic
domains. Only in the field induced synclinic domains some
reduction of the birefringence is found after switching off
the external field. It seems that the surface stabilization of
the FE structure is more efficient and there is only some re-
laxation to the dark domain structure in the center of the
sample.[19f] This means that the increase of the volume frac-
tion of the silyl units stabilizes the uniform SmCsPF struc-
ture. The extinction crosses do not rotate, either by switch-
ing off the field or by applying the opposite field. This indi-
cates a switching process which takes place by rotation
Figure 13. Optical photomicrographs (crossed polarizers) and DSC
curves obtained for the mesophases of compound Si3: a) Colob phase at
1498C, characterized by a strongly colored and highly birefringent tex-
ture. b) SmCPA phase as obtained on cooling the Colob phase to 1308C,
which represents
a low birefringent grey texture. c) DSC traces
(10 Kminꢀ1, first heating- and cooling runs).
Chem. Eur. J. 2007, 13, 2556 – 2577
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2567

C. Tschierske et al.
diffuse maxima in the wide-angle region (see Fig-
ure 14a,b,d,e). The diffuse character of these reflections
perature (T > 1478C) this phase does not show any re-
sponse on an applied electric field (Colob), whereas at lower
temperature an AF switching process was found (ColobPA).
The inset of polar order is not associated with any change in
the X-ray diffraction pattern and it also cannot be detected
in the DSC traces.
ACHTREUNG
confirms the LC character of the mesophases. The diffuse
maximum at about 0.47 nm, corresponding to the mean dis-
tance within the aromatic/aliphatic parts, is split of into two
crescent-like halos in the 2D diffraction patterns of an
aligned sample. These diffuse halos are inclined with meridi-
an and equator which indicates a tilted organization of the
molecules with a tilt angle of about 458. At high tempera-
ture (T > 1438C), the intensity of both maxima is different
At 1438C, at the transition from the columnar phase to
the low temperature phase,[47] the cross reflections corre-
sponding to the 2D lattice completely disappear and only
layer reflections (d=4.8 nm at 1208C) are retained as shown
in Figure 14d,f, which indicate the smectic nature of this
mesophase. From the position of the wide angle reflections
a tilted organization (tilt angle of ca. 458) can be concluded.
The diffraction pattern does not significantly change by fur-
ther reduction of the temperature until crystallization
occurs. In the smectic phases the layer reflections are unusu-
ally broad (see Figure 14f). This would be in line with an un-
dulated smectic phase and in the case of the isomeric com-
pound Si3i the undulated structure was additionally con-
firmed by the observation of satellites occurring beside
these layer reflections.[16] For compound Si3, however, based
on X-ray data, there is no unambiguous proof of a layer un-
dulation because at the transition from the columnar to the
smectic phase the uniform alignment obtained in the colum-
nar phase is partially lost so that any extension of the layer
reflections parallel to the equator would be overlapping
with a broadening of these reflections due to multidomain
formation. For this reason there is also no clear difference
seen in the shape of the layer reflections at the temperature
where the switching changes from AF to surface stabilized
FE (see next Section). These smectic mesophases are as-
signed as SmCPA and SmCPFE, respectively.
(see Figure 14b, and Figure S2d, Supporting Information,
A
thick black line), which most probably arises from a synclin-
ic tilt of the molecules, whereas below 1438C the intensity
becomes nearly equal (Figure 14e, and Figure S2d, Support-
ing Information, thick red line) which is in line with a transi-
tion to a mesophase with an anticlinic tilt organization, as
also identified in optical experiments (see below). In the
whole mesomorphic range the inner diffuse scattering corre-
sponding to a mean distance of d=0.70 nm shows that the
siloxane units are segregated and form distinct sublayers.
Within these layers the siloxane units are strongly disor-
dered as indicated by the nearly ring-like shape of this scat-
tering (see Figure 14a,d). There is only a slightly reduced in-
tensity of this scattering around and at the meridian (see in-
tensity distribution along c shown in Figure S2d, Supporting
Information).
The small angle scattering also changes with temperature.
At T > 1438C the X-ray diffraction pattern (Figure 14c)
shows reflections positioned on the meridian as well as out
of equator and meridian, which indicate a columnar meso-
phase. These reflections can be indexed to an oblique lattice
with the lattice parameters, a=5.8, b=5.1 nm and g=114.58
(see Table S1, Figure S3g, Sup-
porting Information). Accord-
ingly, this mesophase is a modu-
lated smectic phase with obli-
que 2D lattice, where the smec-
tic layers are fragmented into
ribbons (ribbon phase). The lat-
tice parameter a corresponds to
the width of the ribbons (ca. 6–
7 molecules per unit cell, that
is, there are about 6–7 mole-
cules in the cross-section of
each ribbon, see Table S2, Sup-
porting Information) and
b
compares to the thickness of
the ribbons. The value b=
5.1 nm is less than the calculat-
ed molecular length (l=
7.2nm), which is due to the sig-
nificant tilt of the molecules
with respect to the direction b.
Because of this 2D lattice the
phase is assigned as Colob. In
electrooptical experiments (see
Figure 14. Two-dimensional X-ray diffraction patterns of an aligned sample of compound Si3 at 1488C a)–c)
and at 1208C d)–f), on cooling: a), d) original patterns, b),e) the scattering of the isotropic liquid subtracted
from the original patterns to enhance the effect of the nonisotropic distribution of the outer diffuse scattering,
the next section) at higher tem- c), f) reflections in the small-angle region, see also Table S1 and Figure S3, Supporting Information.
2568
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2556 – 2577

Liquid Crystals
FULL PAPER
The smectic phases (SmCPA and SmCPFE) possess schlie-
ren textures with relatively low birefringence, the same as
obtained for Si2 after shearing. Remarkably, there is no for-
mation of a chiral or nonchiral dark texture upon cooling
from the ColobPA phase. It seems that an optically isotropic
(sponge-like) phase cannot be formed upon cooling from a
columnar (modulated smectic) phase. Nevertheless, there
seems to be an inherent tendency for layer deformation and
layer distortion in these smectic phases as in X-ray experi-
ments, at the transition from the columnar to the smectic
phases, the well aligned samples of the columnar phases
always irreversibly loose their alignment.
Electrooptical investigations of Si3: On cooling the isotropic
liquid under
a triangular-wave electric field of about
ꢁ105 V (6 mm) at a frequency of 10 Hz the columnar phase
appears with a mosaic-like texture, but no polarization cur-
rent response can be observed, as shown in Figure 15a. The
textural pattern obtained under these conditions is indistin-
guishable from the natural texture of this mesophase and in-
dicates that the mesophase structure is not influenced by the
field. In this texture, domains in which the ribbons are
aligned parallel or perpendicular to the analyzers are bright,
whereas those which are inclined with the positions of the
polarizers are dark and this confirms a synclinic tilted orga-
nization in this mesophase. The absence of a polarization
current response confirms the missing long range polar
order. That means that the rotation of the molecules around
the long axis is nearly free. This is feasible as there is suffi-
cient space provided for the bent cores by the bulky hepta-
methyltrisiloxane units at both terminal positions.
Upon cooling, below 1478C, that is, in the temperature
region of the ColobPA phase two polarization current peaks
appear, accompanied by a textural change as shown in Fig-
ure 15b. The two peaks in each half period of the applied
triangular wave voltage indicate an AF switching process
(Ps =300 nCcmꢀ2 at 1408C) between 147 and 1368C (transi-
tion temperature to the smectic phase on cooling). The dark
areas within the texture remain at the same position, which
means that the tilt remains synclinic. The synclinic tilt is fur-
ther confirmed by DC field experiments where the extinc-
tion crosses in circular domains are inclined with the posi-
tions of polarizer and analyzer (see inset in Figure 15b). In
addition, no change of the position of the extinction crosses
can be seen either by removing the field or by reversing the
field (see Figure S4a–c, Supporting Information). This indi-
cates that the AF switching process takes place by an elastic
rotation of domains around the long axes. It seems that at
reduced temperature, within the temperature range of the
columnar phase, due to the reduced space required by the
end-groups a polarization along the bent direction results,
leading to an inset of polar order. In this ColobPA phase the
switching process takes place by collective rotation around
the long axis and the polarization values observed in this
phase are relatively small.
Figure 15. Switching current response traces (left) obtained under a trian-
gular-wave field (ꢁ105 V, 10 Hz, 6 mm polyimide coated ITO cell) and
the corresponding textures (right) of the mesophases of compound Si3 a)
Colob phase at 1488C. b) ColobPA phase at 1448C. c) SmCPA phase at
1308C. d) SmCPFE phase at 1008C. e) Plot of the spontaneous polariza-
tion of the mesophases of compound Si3 depending on the temperature.
On further cooling, at the transition to the smectic phase
at 1368C, the birefringence drops and the dark and bright
areas in the texture are exchanged (Figure 15c). Also, in the
circular domains the extinction crosses rotate to positions
Chem. Eur. J. 2007, 13, 2556 – 2577
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2569

C. Tschierske et al.
coincident with the polarizers (compare insets in Fig-
ure 15b,c). This indicates a transition from the synclinic or-
ACHTREUNG
ganization to an organization with an alternating tilt-direc-
tion either between the individual layers (SmCaPF, see Fig-
ure 2b) or between stacks of synclinic layers ([SmCsPF]aPS
structure, see Figure 8b). The AF switching behavior is re-
tained at this phase transition but the position of the repola-
rization current peaks is changed. The two widely separated
peaks, seen in the ColobPA phase, disappear and two closer
peaks appear in the smectic phase (in the two phase range
all four repolarization peaks can be observed, see Figure S5,
Supporting Information). This indicates that the energy dif-
ference between the two FE states and the AF ground-state
structure is reduced. Also in the temperature range of the
smectic phase the position of the extinction crosses does not
change during the switching process. Hence, though the
phase structure is changed, the AF switching still takes
place by rotation around the long axis. In addition, a strong
increase in the polarization value is observed at this transi-
tion. The polarization value saturates at Ps =650 nCcmꢀ2
with further decreasing temperature (see Figure 15e).
At 1268C one of the polarization peaks (that one occur-
ring at higher voltage) disappears, indicating a transition to
a FE switching process (Figure 15d). This transition cannot
be observed by X-ray diffraction or DSC (see Figure 13c),
but in the field induced textures, a slight increase of the bi-
refringence can be seen (compare Figure 15c and d). The
field induced circular domains which coincide with the di-
rections of the polarizers do not change their position at this
temperature, and during switching no change of the position
of the extinction crosses can be found (see Figure S4d–f,
Supporting Information). This means that at 1268C the sta-
bility of the FE states is further increased and causes a FE
switching which also takes place as a collective rotation
around the long axis. It seems that at this temperature the
surface stabilization of the more polar FE structure becomes
dominating, leading to a surface stabilized FE switching,
similar to that observed for compounds Si1 and Si2.
Figure 16. a) Relaxation frequencies and b) limits of the dielectric per-
mittivities as obtained for compound Si3 from the fit to two Cole–Cole
[48]
equations as described in ref.
The phase transitions are indicated as
vertical lines. Phase regions: 1=Colob, 2=ColobPA, 3=SmCPA.
phase with decreasing temperature until the onset of polar
order (transition to ColobPA). This is in line with the pro-
posed model for the Colob to ColobPA transition (see preced-
ing section). In this temperature range the effects in the die-
lectric permittivity e1 are low and hence it is experimentally
difficult to separate the small low frequency relaxation from
the conductivity. The picture changes dramatically at the
ColobPA to SmCPA transition. The stepwise increase of fR2 at
this transition is associated with the transition from a ribbon
structure, where the inter-ribbon boundaries restrict the ro-
tation, to infinite layers where the rotation around the mo-
lecular long axis is less hindered. From the decrease of fR2
with temperatures in the smectic phases an activation
energy of 56 kJmolꢀ1 was calculated for the rotation around
the long axis. Furthermore, a pronounced low frequency ab-
sorption appears. Remarkably, there is no visible change of
fR1 and fR2 at 1268C where the transition from AF to FE
switching is observed in electrooptical experiments. The de-
pendence of the dielectric constants e on the temperature is
shown in Figure 16b. The most remarkable feature seen in
Figure 16b is the strong increase of the dielectric permittivi-
ties e1 and e0 at the transition from the columnar to the
smectic phases, whereas no significant change can be ob-
served at the transitions Colob to ColobPA and SmCPA to
SmCPFE. This indicates that the positive dipole coupling is
strongly enhanced at the transition from the columnar to
Dielectric investigations: Dielectric investigations were car-
ried out with a much lower external electrical voltage.
Therefore, the orientation of the molecules and the struc-
ture of the phases are not disturbed. This method is sensitive
to changes in the short range order within the mesophases.
Data of compound Si3 (Figure 16) indicate two main relaxa-
tion processes, one associated with the fast rotation of the
molecules around the long axis (fR2) and the second one at
lower frequency is a collective process (fR1). The complex fit
of the raw data to two Cole–Cole mechanisms, conductivity
and electric double layer results in the limits of the dielec-
tric permittivities e0ꢀe2 and the related relaxation frequen-
cies fR at different temperatures.[48] As seen in Figure 16a,
there is a strong temperature dependence of the relaxation
frequency fR2. The rotation of the molecules is very fast in
the isotropic phase (it cannot be seen in our time window),
decreases stepwise at the Iso-Colob transition and becomes
significantly slower in the temperature range of the Colob
2570
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2556 – 2577

Liquid Crystals
FULL PAPER
the smectic phase, whereas no significant change takes place
at the transition from AF to FE switching in the smectic
phase region. This observation confirms the proposed organ-
ization of the molecules in these smectic phases. The high
dielectric permittivities of the low frequency limit e0 indicate
a positive coupling of the molecular dipole moments in
large clusters. Also the high values of the dielectric permit-
tivities e1 in the SmCPA phase cannot be satisfactorily ex-
plained with a SmCaPA structure where the polar direction
changes from layer to layer. It is more in line with the pro-
posed organization of the molecules in alternating SmCsPF
layer stacks ([SmCsPF]aPA structure, see Figure 8c). In the
temperature range of the SmCPA phase there is a relatively
strong increase of e0 with decreasing temperature which
should be due to a decrease of rotational disorder of the
molecules with temperature. This is also related to the sig-
nificant increase of the spontaneous polarization from about
300 nCcmꢀ2 at the ColobPA to SmCPA transition to about
600 nCcmꢀ2 at the SmCPA to SmCPFE transition (see Fig-
ure 15e) and the further increase within the SmCPFE phase.
It seems that at the SmCPA to SmCPFE transition at about
1268C a sufficiently dense packing of the molecules is ach-
ieved which enhances the polarization and allows a coupling
of these polar layer stacks to the polar cell surfaces. This sta-
bilizes a uniform SmCsPF structure also at zero voltage,
which is seen as a transition from AF to surface stabilized
FE switching. Accordingly, there is no change of the funda-
mental phase structure at the AF to FE transition; there is
only a reduction of the degree of rotational disorder due to
the changing packing density of the bent-core units. The
slight increase of the birefringence at the transition to the
SmCPFE phases (see Figure 15c,d) might also be due to the
increase of polar order.
layer distortion by these branched end-chains seems also to
be responsible for the maintenance of AF switching at lower
temperature (see explanation given in Section on Effects on
the switching process).
The mesophases of CSi3: The carbosilane derivative CSi3
also behaves similar to Si3 (see Table 1). There is a high
temperature ColobPA phase which is replaced at lower tem-
perature by two smectic phases with apparently anticlinic or-
ganization. There is a change of the position of the repolari-
zation current peaks at the transition from the columnar to
the smectic phase and in the temperature region of this
smectic phase there is a transition from AF to FE switching
at 1328C (see Figure S6a–c, Supporting Information).[49] The
switching takes place around the long axis in the whole mes-
omorphic temperature range. In contrast to Si3, but similar
to Si3i AF switching is found over the whole temperature
range of the Colob phase (ColobPA). However, the textures
are less specific (see Figure S7, Supporting Information) and
for X-ray scattering no well aligned samples of the meso-
phases have been obtained. Only one very weak nonlayer
reflection has been found in the X-ray diffraction pattern of
the columnar phase (see Figure S8, Supporting Informa-
tion). The assignment of these reflections to a distinct obli-
que lattice (a=2.8, b=5.1 nm, see Table 1 and Table S1,
Supporting Information) is more or less tentative and done
in analogy to the corresponding phases of Si3 and Si3i. Fur-
thermore, in the field aligned samples no clear textural
changes take place at the ColobPA to SmCPA transition and
in the temperature range of the ColobPA phase the synclinic
domains formed by cooling under a DC field (or by cooling
without applied field) are replaced by apparently anticlinic
domains under a triangular wave voltage, as typically ob-
served for the smectic phases of Si1 and Si2. It seems that in
the case of compound CSi3 the alignment field can easily
remove the layer modulation of the ColobPA phase which
would require a synclinic tilt.
The mesophases of Si3i: Compound Si3i which has
branched heptamethyltrisiloxane units instead of the linear
ones in Si3 was reported earlier.[16] This compound has also
a ColobPA high temperature phase (a=2.5, b=4.65 nm; g=
1078)[28] and a smectic phase (USmCPA) at lower tempera-
ture. As reported for Si3 there is a transition from a synclin-
ic organization in the ColobPA phase to an apparently anti-
clinic organization in the smectic phase, a change of the po-
sition of the repolarization current peaks at the phase transi-
tion, and the switching takes place by rotation around the
long axis in both mesophases. Hence, the branched com-
pound Si3i is similar to the isomeric nonbranched com-
pound Si3, but there are also differences. Namely, the size
of the ribbons in the ColobPA phase of Si3i is much smaller
than in the ColobPA phase of Si3.[28] Only 2–3 molecules are
organized in the cross-section of these ribbons. In contrast
to Si3 the switching is AF over the whole temperature range
of the columnar and the smectic phases, that is, there is no
transition to a FE switching phase at low temperature and
no loss of the polar order at high temperature. The smaller
size of the ribbons in the ColobPA phase seems to compen-
sate the larger effective size of the branched heptamethyltri-
siloxane units with respect to the linear ones. The stronger
Conclusions
The development of the mesophases in the series of com-
pounds Si1–Si3, CSi3 and other silicon containing bent-core
molecules should mainly result from two competitive forces
provided by the silyl units, namely layer decoupling and
steric frustration.
The effect of layer decoupling on the structure of the smec-
tic phases
Ferroelectric versus antiferroelectric switching: An impor-
tant effect of the Si-containing units is their tendency to seg-
regate into distinct isotropic sublayers which decouple the
layers of the bent-core units. This effect increases with elon-
gation of the silylated units and it has a large impact upon
switching behavior and interlayer correlation. Usually, the
AF organization in the liquid crystalline phases of bent-core
Chem. Eur. J. 2007, 13, 2556 – 2577
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2571

C. Tschierske et al.
molecules allows an easy fluctuation of the molecules from
layer to layer, because the wings of the bent cores are or-
ganized in a synclinic fashion at the inter-layer interfaces
(see Figure 17a). These fluctuations, which stabilize the AF
state which is actually [SmCsPF]aPA for the compounds
under discussion.
From the experimental results it seems that there is some
preference for a synclinic tilted and synpolar (FE) organiza-
tion in the mesophases of these silylated compounds. Hence,
the fundamental phase structure is SmCsPF, but as the layers
are strongly decoupled by the Si-containing sublayers, there
seems to be a relatively low energy difference between syn-
clinic and anticlinic organization and therefore some anti-
clinic interfaces can coexist in a thermodynamic equilibrium
with the synclinic correlated layers. Depending on the pre-
cise molecular structure and the experimental conditions,
the size of the SmCsPF layer stacks could be different, which
leads to distinct types of switching current response curves.
With increasing size of the SmCsPF layer stacks the switch-
ing process changes from classical AF (two well separated
peaks, switching of relatively thin SmCsPF layer stacks) via
single-peak-AF (single peak at 0 V crossing, switching of
larger SmCsPF layer stacks[19h]) to surface stabilized FE
(single peak after 0 V crossing, switching between surface
stabilized macroscopically polar SmCsPF structures).
Surface alignment forces can also stabilize a uniform syn-
clinic organization during growth of the phase from the iso-
tropic liquid state under a static electric field (DC field).
The ring-like appearance of the synclinic domains obtained
under these conditions can be explained by the presence of
synpolar and anticlinic SmCaPS interfaces in these field in-
duced samples ([SmCsPF]aPS structure). If the concentration
of anticlinic defect-layers is sufficiently high that the size of
the synclinic stacks of the layers becomes smaller than the
wave length of light, then only an average optical axis paral-
lel to the layer normal is seen. This gives rise to a low bire-
fringent texture with extinction crosses coinciding with the
directions of the polarizers as seen for the mesophases
grown under a dynamic AC field. If an unequal number of
differently tilted layers are combined, the extinction crosses
adopt an angle which is intermediate between these two
cases.
Figure 17. Effect of segregated siloxane sublayers upon the out-of-plane
interlayer fluctuations (dashed arrows) in the SmCPA and SmCPF phases
of a) conventional bent-core molecules of type I (see Figure 3) and b)
bent-core mesogens with micro segregated siloxane or carbosilane sub-
layers (compounds Sin, CSi3, II and III); these sublayers suppress the
fluctuations and therefore reduce the entropic penalty of the SmCPF or-
ganization with respect to the SmCPA structure.
organization entropically, are more difficult in the FE struc-
tures, where the wings of the bent cores are organized in an
anticlinic fashion at these interfaces and hence disturb these
fluctuations. This leads to an entropic penalty for the FE
structure which is especially strong for conventional banana
molecules with linear alkyl chains. Branching of the alkyl
chains[15a–c,h,l] or any other structural change at the ends of
these chains, which reduces the conformational order at the
interlayer interfaces reduces these fluctuations, reduces the
entropic penalty for the FE structure and makes the FE
structure more favorable (see Figure 17b). The nearly iso-
tropic siloxane sublayers are especially effective in this re-
spect and after alignment in an electric field, surface stabi-
lized FE switching is observed for most of the smectic
phases of the silylated compounds.[19] On the other hand, in-
teractions between the chain ends can stabilize the FE struc-
ture of alkyl substituted molecules energetically,[50] but this
effect is assumed to be less important for the siloxane deriv-
atives, due to the disordered (nearly isotropic) character of
the oligosiloxane sublayers which are located at the interlay-
er interfaces. Therefore, the virgin ground-state structure of
the silylated bent-core molecules should be overall AF, but
FE states can easily be achieved by surface stabilization.[51]
In order to distinguish these surface stabilized FE switching
mesophases from those with virgin FE ground-state struc-
tures (polarization modulated SmCsPF phases[52] or SmCaPF
phases[15b,c,l] with polarization splay between the cell surfa-
“Dark conglomerate” versus birefringent mesophases:
Though the [SmCsPF]aPA ground-state structure is macro-
scopically nonpolar, there is a local polarity within the
SmCsPF layer stacks which can be minimized by a splay of
these layers, producing a sponge-like structure. The decou-
pling of the layers by the siloxane sublayers and the steric
frustration arising from the space required by the silyl units
can additionally contribute to the destabilization of flat
layers and supports layer deformation.
Remarkably, the [SmCsPF]sPA structure (Figure 8a) does
not relax into the dark texture, whereas the [SmCsPF]aPA
structure often does, and this suggests that anticlinic interfa-
ces are important for the formation of these dark phases. It
is known that these anticlinic layers disfavor a regular 2D
modulation of the SmCsPF layers into polarization modulat-
ed smectic phases (B7-type mesophases). Hence, it is sug-
gested that the formation of sponge-like disordered phases
is an alternative way for SmCsPF structures to escape from a
ces[53,15c]), these mesophases are assigned as SmCPFE
/
[
SmCPFE *], as suggested recently.[13c] In this notation the sub-
script “FE” indicates the experimentally observed switching
behavior and not the mesophase structure in the ground
2572
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2556 – 2577

Liquid Crystals
FULL PAPER
(mesoscopic) polar order if anticlinic defect layers cannot be
suppressed. The [SmCsPF]aPA structure is homogeneously
chiral and the optical activity of this mesophase can be de-
tected as regions with uniform optical rotation. The chirality
can easily be observed, because the sponge-like structure
itself is optically isotropic.[36] These sponge-like structures
represent random 3D-networks of layers and therefore a
portion of the molecules is always in direct contact with the
substrate surfaces. For this reason, surface alignment effects
can fix the chirality sense once established in the growing
process over macroscopic regions, leading to a stable struc-
ture composed of domains with opposite chirality sense. De-
pending on the number of layers within the SmCsPF stacks
and the degree of polar order within the layers, the layer
distorting force, that is, the desire to escape from mesoscale
polar order, can have different strength. Hence, the “dark
conglomerate texture” can be very robust as observed for
most monosilylated compounds of type II. For compounds
Si1 and Si2, however, this texture is less stable and can be
disrupted by mechanical shear forces, which lead to birefrin-
gent textures.
As also the switching behavior is dependent on size and
polarity of the SmCsPF layer stacks (see previous Section)
“dark conglomerate phases” can show either FE or non-
classical types of AF switching (single peak at 0 V or two
relatively close peaks).[13b] Only if the size and polarity of
the layer stacks is below a certain critical value, then the
driving force for layer distortion is too small. In this case ex-
clusively birefringent textures and classical AF switching be-
havior (two well separated peaks) can be observed (B2
phases).
It is also remarkable that the birefringent textures ob-
tained after mechanical shearing of Si1 and Si2 do not spon-
taneously reorganize to the sponge-like structure, whereas
this structure is immediately formed from the field induced
[SmCsPF]aPS structure after removal of the alignment field
(see the section on optical observation of the switching pro-
cess). A possible explanation could be that in the field in-
duced structure there are relatively large uniformly aligned
polar regions (ferroelectric clusters) which provide a strong
driving force for layer distortion. In the sheared samples the
FE clusters could be much smaller and hence, there might
be a weaker driving force for layer distortion.
If there are exclusively SmCaPA type defect layers the
SmCsPF layer-stacks have a uniform chirality sense and the
macroscopic chiral domains can be observed optically
(“dark conglomerate phases”). If there are additional
SmCaPS interfaces, or if (+)- and (ꢀ)-[SmCsPF]sPA structures
are mixed, then the chiral domains cannot be distinguished
optically and low birefringent textures without chiral do-
mains can be observed.[19i,33] Such textures were observed
for compound Si1 after switching off the AC field at T >
1108C, for carbosilane based dendrimers with bent-core
units[19i] and occasionally also in the smectic phases occuring
below nematic phases of bent-core molecules.[33]
these optically isotropic mesophases of bent-core molecules.
Moreover, it seems that beside the four classical structures
shown in Figure 2b there is a broad spectrum of additional
polar smectic mesophase structures composed of layer
stacks of different size (see Figure 8a–c).[54]
Size effects of the silyl groups
Effects on the phase structure: Beside the layer decoupling,
a second important effect of the Si-containing units is due to
their size which becomes especially important for large Si-
substituents at elevated temperature. In the compounds Si1–
Si3 and CSi3 with bulky silyl units at both ends there is no
possible way for the escape from this steric stress (Fig-
ure 5b). This leads to a frustration of the smectic layers with
formation of columnar ribbon phases (Colob) at higher tem-
perature (compounds Si3, Si3i and CSi3). The size of the
ribbons is reduced with increasing steric frustration (com-
pare the linear compound Si3 and the branched compound
Si3i). Because the molecules are tilted in the layers, the rib-
bons organize into an oblique lattice. In these oblique
ribbon phases any anticlinic organization is highly unfavora-
ble due to packing constrains at the inter-ribbon interfaces.
Therefore, all anticlinic defects are squeezed out and the or-
ganization of the molecules in these columnar phases is ex-
clusively synclinic.
As the temperature is reduced, this steric effect becomes
smaller, the inter-ribbon interfaces disappear and infinite
smectic layers are formed. In these smectic phases (SmCPA,
USmCPA) anticlinic interfaces become possible which imme-
diately gives rise to a [SmCsPF]aPA structure of these smectic
phases, indicated by a texture with extinction crosses paral-
lel to the polarizers.
Effects on the switching process: The steric effects provided
by the Si-containing end-groups should also be responsible
for the change of the switching process, which can take
place either by a collective rotation on a cone or around the
long axis (Figure 2d,e). As a close packing is not possible
for molecules with silyl groups at both ends more space be-
comes available between the bent cores and therefore not
only in the columnar phases where the switching around the
cone is inhibited by the inter-ribbon interfaces, but also in
the smectic phases of compounds Si2, Si3, Si3i and CSi3 the
switching always takes place by rotation around the long
axis (with exception of the switching of compound Si1 with
the smallest silyl group at T < 1108C). This type of switch-
ing is of special interest as it changes the chirality sense of
the layers and hence allows the switching of the superstruc-
tural chirality.
For compound Si3 the steric effects are even large enough
to give rise to a transition from an AF switching (ColobPA)
to a nonswitching columnar (Colob) phase upon rising the
temperature. This means that there is a temperature depen-
dent transition from a polar ordered to a conventional non-
polar phase. Though this was previously reported for smectic
phases,[55] this phenomenon is rare in columnar phases.
Hence, the proposed models can satisfactorily explain the
experimental findings associated with the occurrence of
Chem. Eur. J. 2007, 13, 2556 – 2577
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2573

C. Tschierske et al.
There is only one report about such a transition in a hexago-
nal columnar phase of a polycatenar bent-core molecule,
but the proposed phase structure of these Colh phases is dif-
ferent from that one of the Colob phases reported herein.[56]
Especially in the smectic phase occurring directly below
the ColobPA phase there is still a significant steric frustration
which separates the bent-cores of the individual molecules
in the smectic layers. In these smectic phases the layer polar-
ization is comparatively small. Dielectric measurements
(Figure 16) indicate that the rotation around the long axis is
quite easy and hence the polarization within the SmCsPF
overlapping of about one half of the rigid bent cores at the
interfaces between adjacent ribbons. In this way the inter-
ribbon interfaces of the AF arrangements Figure 18a and d
become different. Only the reversal of the polar directions
within the modulated layers (along direction a, as in Fig-
ure 18d) allows a parallel packing of the rod-like wings of
the aromatic cores at the interfaces between the ribbons.
This enables a continuous packing of the rod-like wings in
adjacent ribbons and therefore this organization seems to be
favorable. In contrast, if the polar direction in adjacent rib-
bons would be synpolar (FE, see Figure 18b,c), then the
rod-like wings are not parallel at these interfaces and there-
fore unfavorable packing results at these interfaces. Hence,
it can be concluded that in the ColobPA phases the polar di-
rection changes within the modulated layers from ribbon to
ribbon, as shown in Figure 18d. This also means that a FE
organization, which requires a uniform polar order along di-
rection a (Figure 18b,c) is less favorable than the AF struc-
ture (Figure 18d). Therefore, the switching process changes
from FE in smectic phases (continuous layers) to AF in the
Colob type ribbon phases. For the same reason the switching
of the AF ground state into the FE state requires significant
energy and the threshold voltage for this switching process
becomes much higher than in the smectic phases.
It can be concluded, that the fundamental mode of switch-
ing of the silylated bent-core molecules with a sufficiently
large number of dimethylsilyl groups is surface stabilized FE
(SmCPFE phases). AF switching is the result of the steric
frustration arising from very large silyl groups, either leading
to the formation of unfavorable inter-ribbon interfaces
(ColobPA phases), or to the separation of the bent cores in
the layers which reduces the polar order and hence decreas-
es the surface coupling (SmCPA and USmCPA phases).
In summary, a significant step forward was achieved in
the understanding of the self organization of bent-core mol-
ecules in general. The proposed models, based on a meso-
phase structure composed of chiral and polar mesoscale
SmCsPF layer stacks with anticlinic defects, provide a quite
uniform picture and allow the explanation of the different
observations made within the distinct series of silylated
bent-core compounds reported herein and earlier.[16,19] It
contributes to a fundamental understanding of the dark con-
glomerate phases and the transition from AF to FE switch-
ing. It also provides new insights into the general relations
between structure and properties of chevron shaped mole-
cules which allow a more efficient design of soft matter ma-
terials with ferroelectric/antiferroelectric properties and
switchable supramolecular (superstructural) chirality for use
in future applications.
Figure 18. Modes of organization in the ColobPA phases, the number of
molecules in the ribbons and the tilt direction of the molecules with re-
spect to b were arbitrarily chosen (the molecules could also be organized
more parallel to b). a) AF organization with alternation of the polar di-
rection between the modulated layers. b) FE organization. c) FE organi-
zation with opposite polarity and chirality. d) AF organization with alter-
nation of the polar direction from ribbon to ribbon within the modulated
layers.
layer stacks of the [SmCsPF]aPA structure is not large enough
to allow a sufficiently strong coupling of these layer stacks
to the surfaces and the switching process is AF. Only upon
further reduction of the temperature the polarization in-
creases and at a certain temperature the polar surface inter-
actions become strong enough to induce FE switching.
Effects of the phase structure on the switching behavior
An additional important effect of the organization of the
molecules in a ribbon structure is the change of the switch-
ing process from surface stabilized FE in the smectic phases
(SmCPFE) to AF in the columnar phases (ColobPA). In order
to understand this it must be considered that there are two
principally different organizations which can lead to antifer-
roelectricity in the ribbon phases. Either the polar direction
changes along the direction b between the modulated layers
as shown in Figure 18a, or the antipolar order is realized
within the modulated layers from ribbon to ribbon (i.e.,
along the direction a), as shown in Figure 18d. As seen in
Figure 18 the ribbon structure can be achieved if there is an
Experimental Section
Methods: Determination of phase transition temperatures was done by
polarizing microscopy (Optiphot-2polarizing microscope, Nikon in con-
junction with a FP 82HT hot stage, Mettler) and confirmed by DSC
(DSC-7, Perkin-Elmer, scanning rate 10 Kminꢀ1). Powder X-ray investi-
2574
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2556 – 2577

Liquid Crystals
FULL PAPER
gations were carried out with a Guinier film camera and a Guinier goni-
ometer (both by Huber) with samples kept in glass capillaries (1=
1 mm) in a temperature-controlled heating stage using quartz-mono-
chromatized CuKa radiation (30 to 60 min exposure time, calibration of
(d, 3J=8.7 Hz, 2H, Ar-H), 7.21 (m, 1H, Ar-H), 6.97 (d, 4H, 3J=8.9 Hz,
Ar-H), 4.04 (t, J=6.6 Hz, 4H, OCH2), 1.81 (m, 4H, OCH2CH2), 1.47 (m,
3
4H, OCH2CH2CH2), 1.27 (m, 28H, CH2), 0.50 (m, 2H, SiCH2), 0.07 [s,
18H, Si-(CH3)3], 0.04 [s, 12H, Si-(CH3)2], 0.02[s, 1H2 , Si-(CH 3)2];
13C NMR (125 MHz, CDCl3): d = 164.49 (4C), 164.47, 164.33, 163.84,
155.45 (2C), 151.35, 150.66, 142.09, 138.04, 132.42, 131.84 (8C), 129.87,
128.33 (2C), 126.82 (2C), 124.71, 122.12 (4C), 122.08 (4C), 120.97,
120.63, 120.45, 114.43, 68.40 (2C), 33.44 (2C), 29.62 (2C), 29.58 (2C),
29.56 (2C), 29.39 (2C), 29.37 (2C), 29.09 (2C), 25.99 (2C), 23.23 (2C),
18.29 (2C), 1.81 (6C), 1.27 (4C), 0.20 (4C); 29Si NMR (99.3 MHz,
the film patterns with the powder pattern of Pb(NO3)2). 2D patterns for
R
aligned samples on a glass plate on a temperature-controlled heating
stage (alignment at the sample/glass or at the sample/air interface) were
recorded with a 2D detector (HI-STAR, Siemens). Electrooptical meas-
urements were carried out in commercial ITO cells (E. H. C. Corp.; spac-
ing: 5 mm or 6 mm). The switching polarization was measured by means
of the triangular wave voltage method.[57] Dielectric investigations were
carried out in the range from 1 Hz to 10 MHz using a Solartron-Schlum-
berger Impedance Analyzer Si 1260 and a Chelsea Interface. A brass cell
coated with gold (d=0.2mm) was used as capacitor.
CDCl3):
d
=
7.47, 7.03, ꢀ21.06; elemental analysis calcd (%) for
C76H110O14Si6: C 64.46, H 7.83; found: C 64.29, H 7.6.
CSi3: 1H NMR (400 MHz, CDCl3): d = 8.29 (d, 3J=8.9 Hz, 2H, Ar-H),
8.28 (d, 3J=8.7 Hz, 2H, Ar-H), 8.14 (d, 3J=8.7 Hz, 4H, Ar-H), 7.65 (d,
3J=8.9 Hz, 2H, Ar-H), 7.51 (d, 3J=5.2Hz, 2H, Ar-H), 7.44 (m, 1H, Ar-
H), 7.37 (d, 3J=8.7 Hz, 2H, Ar-H), 7.36 (d, 3J=8.7 Hz, 2H, Ar-H), 7.28
(d, 3J=8.7 Hz, 2H, Ar-H), 7.22 (m, 1H, Ar-H), 6.97 (d, 4H, 3J=8.9 Hz,
Synthesis and analytical data
En: 3,4’-Biphenyldiol[24] (0.5 g, 2.7 mmol), 4-[4-(undec-10-enyloxy)ben-
zoyloxy]benzoic acid[25] (2.32 g, 5.7 mmol), DCC (1.17 g, 5.7 mmol) and
DMAP (0.05 g) were dissolved in dry CH2Cl2 (80 mL) and stirred at
room temperature for 24 h. After evaporation of the solvent the crude
product was purified by column chromatography (CHCl3, Rf =
0.22).Yield: 1.89 g (72%), colourless solid. 1H NMR (400 MHz, CDCl3):
d = 8.29 (m, 4H, Ar-H), 8.14 (d, 3J=8.5 Hz, 4H, Ar-H), 7.65 (d, 3J=
8.7 Hz, 2H, Ar-H), 7.51 (m, 2H, Ar-H), 7.47 (m, 1H, Ar-H), 7.37 (m,
4H, Ar-H), 7.31 (d, 3J=8.7 Hz, 2H, Ar-H), 7.21 (m, 1H, Ar-H), 6.97 (d,
3J=8.9 Hz, 4H, Ar-H), 5.81 (m, 2H, CH=CH2), 4.95 (m, 4H, CH=CH2),
4.04 (t, 3J=6.6 Hz, 4H, OCH2), 2.02 (m, 4H, CH2-CH=CH2), 1.81 (m,
4H, OCH2CH2), 1.33 (m, 20H, CH2).
3
Ar-H), 4.04 (t, J=6.5 Hz, 4H, OCH2), 1.81 (m, 4H, OCH2CH2), 1.47 (m,
4H, OCH2CH2CH2), 1.32(m, 38H, CH ), 0.53 (m, 16H, SiCH2), 0.46 (m,
2
4H, SiCH2), ꢀ0.04 [s, 18H, Si
C
(CH3)2];
13C NMR (125 MHz, CDCl3): d
163.83, 155.43 (2C), 151.34, 150.65, 142.06, 138.01, 132.41, 131.82 (4C),
131.57 (4C), 129.85, 128.31 (2C), 126.85, 126.81, 124.68, 124.61, 122.11
(2C), 122.07 (2C), 120.96, 120.62, 120.44, 114.42, 68.39 (2C), 33.70 (2C),
29.63(2C), 29.59 (2C), 29.55 (2C), 29.37 (2C), 29.29 (2C), 29.09 (2C),
G
25.98 (2C), 23.92 (2C), 21.36 (2C), 20.10 (2C), 20.03 (2C), 20.00 (2C),
18.41 (4C), 15.38 (2C), ꢀ1.54 (6C), ꢀ3.18 (4C), ꢀ3.26 (4C); 29Si NMR
(99.3 MHz, CDCl3): d = 1.59, 0.99, 0.57; elemental analysis calcd (%) for
C88H134O10Si6: C 69.5, H 8.88; found: C 70.05, H 8.86.
Si1: Under an argon atmosphere En (300 mg, 0.31 mmol) was dissolved
in anhydrous toluene (5 mL), ethyldimethylsilane (120 mg, 1.3 mmol) and
one drop of Karstedtꢁs catalyst (platinum–divinyltetramethyl–siloxane
complex in xylene, Gelest Inc.) was added to this mixture. After reaction
was completed (detection by TLC, ca. 36 h) the solvent was evaporated
and the crude product was purified by centrifugal preparative thin layer
chromatography (silica gel, CHCl3) followed by recrystallisation from
ethyl acetate to give Si1 (60 mg, 0.14 g, 32%). 1H NMR (400 MHz,
Acknowledgements
The work was supported by the DFG (GRK 894/1) and the Fonds der
Chemischen Industrie; R.A.R is grateful to the Alexander von Humboldt
Foundation for the research fellowship.
3
CDCl3): d = 8.29 (d, J=8.7 Hz, 2H, Ar-H), 8.15 (d, 3J=8.7 Hz, 4H, Ar-
H), 7.63 (d, 3J=8.5 Hz, 2H, Ar-H), 7.49 (d, 3J=4.9 Hz, 2H, Ar-H), 7.44
(m, 1H, Ar-H), 7.36 (d, 3J=8.7 Hz, 2H, Ar-H), 7.27 (d, 3J=8.7 Hz, 2H,
Ar-H), 7.21 (m, 1H, Ar-H), 6.97 (d, 4H, 3J=8.9 Hz, Ar-H), 4.04 (t, 3J=
6.5 Hz, 4H, OCH2), 1.82(m, 4H, OCH 2CH2), 1.45 (m, 4H,
OCH2CH2CH2), 1.25 (m, 30H, CH2), 0.44 (m, 4H, SiCH2), 0.05 [s, 36H,
[1] a) J.-M. Lehn, Proc. Natl. Acad. Sci. USA 2002, 99, 4763–4768;
b) J.-M. Lehn, Science 2002, 295, 2400–2403.
[2] V. A. Blatov, L. Carlucci, G. Ciani, D. M. Proserpio, CrystEngComm
2004, 6, 377–395.
Si-(CH3)3], ꢀ0.02(s, 6H, Si-CH 3); 13C NMR (125 MHz, CDCl3): d
=
164.32 (4C), 164.18 (2C), 163.73 (2C), 155.36 (2C), 151.27, 150.59,
142.00, 137.96, 132.34 (4C), 131.75 (4C), 129.78, 128.25 (2C), 126.78
(2C), 124.62, 122.05 (4C), 122.01 (4C), 120.95, 120.55, 120.39, 114.40,
68.44 (2C), 33.29 (2C), 29.69 (2C), 29.65 (4C), 29.44 (4C), 29.19 (2C),
26.08 (2C), 23.16 (2C), 17.74 (2C), 1.97 (12C), ꢀ0.14 (2C); 29Si NMR
[3] B. Moulton, M. J. Zaworotko, Chem. Rev. 2001, 101, 1629–1658.
[4] a) D. M. Bassani, J.-M. Lehn, K. Fromm, D. Fenske, Angew. Chem.
1998, 110, 2534–2537; Angew. Chem. Int. Ed. 1998, 37, 2364–2567;
b) P. N. W. Baxter, J.-M. Lehn, G. Baum, D. Fenske, Chem. Eur. J.
2000, 6, 4510–4517; c) E. Breuning, G. S. Hanan, F. J. Romero-Sal-
guero, A. M. Garcia, P. N. W. Baxter, J.-M. Lehn, E. Wegelius, K.
Rissanen, H. Nierengarten, A. van Dorsselaer, Chem. Eur. J. 2002,
8, 3458–3466.
[5] S. R. Seidel, P. J. Stang, Acc. Chem. Res. 2002, 35, 972–983.
[6] Handbook of Liquid Crystals, Vol. 1–3 (Eds.: D. Demus, J. Goodby,
G. W. Gray, H.-W. Spiess, V. Vill), Wiley-VCH, Weinheim, 1998.
[7] a) T. Yamamoto, A. M. Arif, P. Stang, J. Am. Chem. Soc. 2003, 125,
12309–12317; b) M. Fujita, Chem. Soc. Rev. 1998, 27, 417–425.
[8] a) M. T. Stone, J. M. Heemstra, J. S. Moore, Acc. Chem. Res. 2006,
39, 11–20; b) H.-J. Kim, W.-C. Zin, M. Lee, J. Am. Chem. Soc. 2004,
126, 7009–7014.
[9] B. Hasenknopf, J.-M. Lehn, Helv. Chim. Acta 1996, 79, 1643–1650.
[10] H. Abourahma, B. Moulton, V. Kravtsov, M. J. Zaworotko, J. Am.
Chem. Soc. 2002, 124, 9990–9991.
[11] M. Tominaga, K. Suzuki, T. Murase, M. Fujita, J. Am. Chem. Soc.
2005, 127, 11950–11951.
[12] T. Niori, T. Sekine, J. Watanabe, T. Furukawa, H. Takezoe, J. Mater.
Chem. 1996, 6, 1231–1233.
[13] Reviews: a) G. Pelzl, S. Diele, W. Weissflog, Adv. Mater. 1999, 11,
707–724; b) C. Tschierske, G. Dantlgraber, Pramana 2003, 61, 455–
481; c) R. Amaranatha Reddy, C. Tschierske, J. Mater. Chem. 2006,
(99.3 MHz, CDCl3):
d = 3.6; elemental analysis calcd (%) for
C70H90O10Si2: C 73.26, H 7.90; found: C 73.56, H 7.78.
Si2: 1H NMR (400 MHz, CDCl3): d = 8.29 (d, 3J=8.9 Hz, 2H, Ar-H),
8.28 (d, 3J=8.9 Hz, 2H, Ar-H), 8.14 (d, 3J=8.5 Hz, 4H, Ar-H), 7.65 (d,
3J=8.7 Hz, 2H, Ar-H), 7.51 (d, 3J=4.8 Hz, 2H, Ar-H), 7.44 (m, 1H, Ar-
H), 7.37 (d, 3J=8.7 Hz, 2H, Ar-H), 7.36 (d, 3J=8.9 Hz, 2H, Ar-H), 7.29
(d, 3J=8.7 Hz, 2H, Ar-H), 7.22 (m, 1H, Ar-H), 6.97 (d, 4H, 3J=9.1 Hz,
Ar-H), 4.04 (m, 4H, OCH2), 1.82(m, 4H, OCH 2CH2), 1.45 (m, 4H,
OCH2CH2CH2), 1.25 (m, 28H, CH2), 0.49 (m, 4H, SiCH2), 0.05 [s, 18H,
Si-(CH3)3], 0.02[s, 12H, Si-(CH 3)2]; 13C NMR (125 MHz, CDCl3): d =
164.47 (4C), 164.45, 164.31, 163.83, 155.44 (2C), 151.34, 150.65, 142.07,
138.02, 132.41, 131.83 (8C), 129.86, 128.32 (2C), 126.85, 126.82, 124.69,
122.11 (4C), 122.07 (4C), 120.96, 120.62, 120.44, 114.42, 68.39 (2C), 33.40
(2C), 29.61 (2C), 29.56 (2C), 29.55 (2C), 29.37 (2C), 29.09 (2C), 25.98
E
(2C), 23.26 (2C), 23.17 (2C), 18.37, 18.13, 1.97 (6C), 0.35 (4C); 29Si
NMR (99.3 MHz, CDCl3): d = 6.16, 5.34; elemental analysis calcd (%)
for C72H98O12Si4: C 68.21, H 7.79; found: C 68.67, H 7.67.
Si3: 1H NMR (400 MHz, CDCl3): d = 8.29 (d, 3J=8.7 Hz, 2H, Ar-H),
8.28 (d, 3J=8.7 Hz, 2H, Ar-H), 8.14 (d, 3J=8.7 Hz, 4H, Ar-H), 7.65 (d,
3J=8.5 Hz, 2H, Ar-H), 7.49 (d, 3J=5.2Hz, 2H, Ar-H), 7.44 (m, 1H, Ar-
H), 7.37 (d, 3J=8.7 Hz, 2H, Ar-H), 7.36 (d, 3J=8.7 Hz, 2H, Ar-H), 7.29
Chem. Eur. J. 2007, 13, 2556 – 2577
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2575

C. Tschierske et al.
16, 907–961; d) M. B. Ros, J. L. Serrano, M. R. de la Fuente, C. L.
Folcia, J. Mater. Chem. 2005, 15, 5093–5098; e) H. Takezoe, Y. Taka-
nishi, Jpn. J. Appl. Phys. 2006, 45, 597–625.
E. Kçrblova, R. K. Shoemaker, Cryst. Growth Des. 2005, 5, 2091–
2099, and references given in [30].
[22] a) J. Thisayukta, Y. Nakayama, S. Kawauchi, H. Takezoe, J. Wata-
nabe, J. Am. Chem. Soc. 2000, 122, 7441–7448; b) J. Thisayukta, H.
Kamee, S. Kawauchi, J. Watanabe, Mol. Cryst. Liq. Cryst. 2000, 346,
63–75; c) H. N. Shreenivasa Murthy, B. K. Sadashiva, Liq. Cryst.
2002, 29, 1223–1234; d) J. Ortega, C. L. Folcia, J. Etxebarria, N.
Gimeno, M. B. Ros, Phys. Rev. E 2003, 68, 011707; e) W. Weissflog,
M. W. Schrçder, S. Diele, G. Pelzl, Adv. Mater. 2003, 15, 630–633;
f) M. W. Schrçder, S. Diele, G. Pelzl, U. Dunemann, H. Kresse, W.
Weissflog, J. Mater. Chem. 2003, 13, 1877–1882; g) M. W. Schrçder,
G. Pelzl, U. Dunemann, W. Weissflog, Liq. Cryst. 2004, 31, 633–637;
h) W. Weissflog, S. Sokolowski, H. Dehne, B. Das, S. Grande, M. W.
Schrçder, A. Eremin, S. Diele, G. Pelzl, H. Kresse, Liq. Cryst. 2004,
31, 923–933; i) V. Kozmik, A. Kovarova, M. Kuchar, J. Svoboda, V.
Novotna, M. Glogarova, J. Kroupa, Liq. Cryst. 2006, 33, 41–56.
[23] Field induced SmCP[*] phases: a) G. Heppke, D. D. Parghi, H.
Sawade, Liq. Cryst. 2000, 27, 313–320; b) J. Etxebarria, C. L. Folcia,
J. Ortega, M. B. Ros, Phys. Rev. E 2003, 67, 042702; c) A. Jakli, Y. -
M. Huang, K. Fodor-Csorba, A. Vajda, G. Galli, S. Diele, G. Pelzl,
Adv. Mater. 2003, 15, 1606–1610.
[24] D. Shen, A. Pegenau, S. Diele, I. Wirth, C. Tschierske, J. Am. Chem.
Soc. 2000, 122, 1593–1601.
[25] R. Achten, A. Koudijs, M. Giesbers, A. T. M. Marcelis, E. J. R.
Sudhçlter, Liq. Cryst. 2005, 32, 277–285.
[26] C. Tschierske, H. Zaschke, J. Prakt. Chem. 1989, 331, 365–366.
[27] However, the molecular length (l=5.6 nm, determined as described
in footnote [c] in Table 1) is significantly larger than the parameter
b and this is quite unusual in this case. A possible explanation could
be that the molecules are not strictly orthogonal arranged within the
ribbons, but adopt a tilted organization where the tilt direction
changes randomly along the direction c. It is also possible that there
is some intercalation of the terminally unsaturated chains which
leads to a reduction of the efficient molecular length.
[14] a) D. R. Link, G. Natale, R. Shao, J. E. Maclennan, N. A. Clark, E.
Kçrblova, D. M. Walba, Science 1997, 278, 1924–1927; b) D. M.
Walba, Top. Stereochem. 2003, 42, 475–518.
[15] Examples of FE switching bent-core materials: a) D. M. Walba, E.
Kçrblova, R. Shao, J. E. Maclennan, D. R. Link, M. A. Glaser, N. A.
Clark, Science 2000, 288, 2181–2184; b) E. Gorecka, D. Pociecha, F.
Araoka, D. R. Link, M. Nakata, J. Thisayukta, Y. Takanishi, K. Ishi-
kawa, J. Watanabe, H. Takezoe, Phys. Rev. E 2000, 62, R4524–
R4527; c) M. Nakata, D. R. Link, F. Araoka, J. Thisayukta, Y. Taka-
nishi, K. Ishikawa, J. Watanabe, H. Takezoe, Liq. Cryst. 2001, 28,
1301–1308; d) J. P. Bedel, J. C. Rouillon, J. P. Marcerou, M. La-
guerre, H. T. Nguyen, M. F. Achard, J. Mater. Chem. 2002, 12, 2214–
2220; e) R. Amaranatha Reddy, B. K. Sadashiva, J. Mater. Chem.
2002, 12, 2627–2642; f) R. Amaranatha Reddy, B. K. Sadashiva,
Liq. Cryst. 2003, 30, 1031–1050; g) S. Rauch, P. Bault, H. Sawade,
G. Heppke, G. G. Nair, A. Jakli, Phys. Rev. E 2002, 66, 021706;
h) K. Kumazawa, M. Nakata, F. Araoka, Y. Takanishi, K. Ishikawa,
J. Watanabe, H. Takezoe, J. Mater. Chem. 2004, 14, 157–164; i) R.
Amaranatha Reddy, V. A. Raghunathan, B. K. Sadashiva, Chem.
Mater. 2005, 17, 274–283; j) H. Nadasi, W. Weissflog, A. Eremin, G.
Pelzl, S. Diele, S. Das, S. Grande, J. Mater. Chem. 2002, 12, 1316–
1324; k) J. P. Bedel, J. C. Rouillon, J. P. Marcerou, H. T. Nguyen,
M. F. Achard, Phys. Rev. E 2004, 69, 061702; l) T. S. K. Lee, S. Heo,
J. G. Lee, K.-T. Kang, K. Kumazawa, K. Nishida, Y. Shimbo, Y. Ta-
kanishi, J. Watanabe, T. Doi, T. Takahashi, H. Takezoe, J. Am.
Chem. Soc. 2005, 127, 11085–11091.
[16] Preliminary communication (compound Si3i):[28] C. Keith, R. Amar-
anatha Reddy, U. Baumeister, C. Tschierske, J. Am. Chem. Soc.
2004, 126, 14312–14313.
[17] a) M. W. Schrçder, S. Diele, G. Pelzl, W. Weissflog, ChemPhysChem
2004, 5, 99–103; b) R. Amaranatha Reddy, M. W. Schrçder, M.
Bodyagin, H. Kresse, S. Diele, G. Pelzl, W. Weissflog, Angew. Chem.
2005, 117, 784–788; Angew. Chem. Int. Ed. 2005, 44, 774–778;
c) J. P. Bedel, J. C. Rouillon, J. P. Marcerou, H. T. Nguyen, M. F.
Achard, Phys. Rev. E 2004, 69, 061702; d) E. Gorecka, N. Vaupotic,
D. Pociecha, M. Cepic, J. Mieczkowski, ChemPhysChem 2005, 6,
1087–1093.
[28] Re-indexation with improved X-ray data gave an oblique lattice
with smaller lattice parameters than reported in a previous commu-
nication (ref. [16])
[29] The fact that only one diffuse scattering is found for the carbosilane
Si1 does not necessarily mean that in this case the Si-containing
units are not segregated. The ethyldimethylsilyl groups (and also the
carbosilane units in CSi3) provide the same mean distance as the
alkyl chains. For this reason in the case of the investigated carbosi-
lanes, the presence of a segregated structure cannot be deduced
from the profile of the wide angle scattering.[19h]
[18] a) A. Jakli, Liq. Cryst. Today 2002, 11, 1–5; b) A. Jakli, L.-C. Chien,
D. Krüerke, H. Sawade, G. Heppke, Liq. Cryst. 2002, 29, 377–381;
c) F. Kentischer, R. Macdonald, P. Warnick, G. Heppke, Liq. Cryst.
1998, 25, 341–347.
[19] a) G. Dantlgraber, A. Eremin, S. Diele, A. Hauser, H. Kresse, G.
Pelzl, C. Tschierske, Angew. Chem. 2002, 114, 2515–2518; Angew.
Chem. Int. Ed. 2002, 41, 2408–2412; b) G. Dantlgraber, S. Diele, C.
Tschierske, Chem. Commun. 2002, 2768–2769; c) G. Dantlgraber,
U. Baumeister, S. Diele, H. Kresse, B. Lühmann, H. Lang, C.
Tschierske, J. Am. Chem. Soc. 2002, 124, 14852–14853; d) C. Keith,
R. Amaranatha Reddy, H. Hahn, H. Lang, C. Tschierske, Chem.
Commun. 2004, 1898–1899; e) C. Keith, R. Amaranatha Reddy, C.
Tschierske, Chem. Commun. 2005, 871–873; f) C. Keith, R. Amara-
natha Reddy, A. Hauser, U. Baumeister, C. Tschierske, J. Am.
Chem. Soc. 2006, 128, 3051–3066; g) R. Amaranatha Reddy, G.
Dantlgraber, U. Baumeister, C. Tschierske, Angew. Chem. 2006, 118,
1962–1967; Angew. Chem. Int. Ed. 2006, 45, 1928–1933; h) C.
Keith, R. Amaranatha Reddy, U. Baumeister, H. Hahn, H. Lang, C.
Tschierske, J. Mater. Chem. 2006, 16, 3444–3447; i) H. Hahn, H.
Lang, C. Keith, R. Amaranatha Reddy, C. Tschierske, Adv. Mater.
2006, 18, 2629–2633.
[30] In thicker samples of compounds Si1 and Si2 a typical bluish color is
seen. This type of “bluish conglomerate phases” was first reported
for some bent-core molecules incorporating naphthalene units[22a]
and it was also seen for the soft crystalline B4[*] phases: T. Sekine,
T. Niori, J. Watanabe, T. Furukawa, S. W. Choi, H. Takezoe, J.
Mater. Chem. 1997, 7, 1307–1309; H. Niwano, M. Nakata, J. Thi-
sayukta, D. R. Link, H. Takezoe, J. Watanabe, J. Phys. Chem. B
2004, 108, 14889–14896; Thisayukta, H. Takezoe, J. Watanabe, Jpn.
J. Appl. Phys.2001, 40, 3277–3287; F. Araoka, N. Y. Ha, Y. Kinoshi-
ta, B. Park, J. W. Wu, H. Takezoe, Phys. Rev. Lett. 2005, 94, 137801.
It might arise from a light scattering at disordered smectic domains.
[31] K. D’Have, A. Dahlgren, P. Rudquist, J. P. F. Lagerwall, G. Ander-
sson, M. Matuszczyk, S. T. Lagerwall, R. Dabrowski, W. Drzewinski,
Ferroelectrics 2000, 244, 415–428.
[32] For bent-core molecules a deviation from exactly 458 tilt angle (q) is
possible if the bending angle (a) is changed according to:
a=2tanꢀ1
(1/cosq).
A
[20] E. L. Eliel, S. H. Wilen, L. N. Mander, Stereochemistry of Organic
Compounds, Wiley, New York, 1994, p. 159.
[33] G. Liao, S. Stojadinovic, G. Pelzl, W. Weissflog, S. Sprunt, A. Jꢂkli,
Phys. Rev. E 2005, 72, 021710.
[21] There is a second different type of mesophases showing this type of
“dark conglomerate textures”. In contrast to the mesophases dis-
cussed herein, which are simple layer structures without in-plane
order, this second type has additional order and is assumed to be a
soft crystalline phase, assigned as B4[*], see D. M. Walba, L. Eshdat,
[34] M. Spannuth, L. Hough, D. Coleman, C. Jones, M. Nakata, Y. Taka-
nishi, H. Takezoe, J. Watanabe, C. Tschierske, E. Kçrblova, D.
Walba, N. A. Clark, Poster Presented at the 9th International Liquid
Crystal Conference, Ljubljana, July 4–9, 2004, Book of Abstracts,
SYN-P026.
2576
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2556 – 2577

Liquid Crystals
FULL PAPER
[35] Remarkably, also the birefringent textures obtained after shearing
have a significantly lower viscosity than the original dark conglomer-
ate phases of compounds Si1 and Si2.
[36] L. E. Hough, N. A. Clark, Phys. Rev. Lett. 2005, 95, 107802.
[37] In a previous paper (see ref. [19f]) we have supported this proposal
by the observation of a tilt-angle dependent inversion of the optical
rotation, which is proposed by this model.
[47] No supercooling of the phase transition Colob–SmCPA was observed
in X-ray experiments. Probably, the stronger surface interactions in
the thin measurement cells (5–10 mm) used for the optical and elec-
trooptical investigation provide a strong influence of these surfaces,
which seems to stabilize the columnar organization. In X-ray diffrac-
tion experiments capillaries (Guinier method) or open droplets (2D
X-ray) were used where the influence of the surfaces is diminished.
[48] H. Kresse, H. Schlacken, U. Dunemann, M. W. Schrçder, G. Pelzl,
Liq. Cryst. 2002, 29, 1509–1512.
[38] M. F. Achard, J. Ph. Bedel, J. P. Marcerou, H. T. Nguyen, J. C. Rouil-
lon, Eur. Phys. J. E 2003, 10, 129–134.
[39] In polyimide-coated cells the switching behavior can be more com-
plex in some cases, due to the effect of an induced electric double
layer, which will be discussed in section on the influence of surfaces.
[40] In this notation the structure of the layer stacks is given in brackets
and the following subscripts define the correlation between the
layer stacks (a=anticlinic, s=synclinic, A=antipolar, S=synpolar).
[41] This kind of field induced texture was observed earlier in an AF
switching smectic phase (SmCPA): M. Zennyoji, Y. Takanishi, K. Ish-
ikawa, J. Thisayukta, J. Watanabe, H. Takezoe, J. Mater. Chem. 1999,
9, 2775–2778.
[42] The symbol X, according to the rules used by Beilstein, replaces ste-
reochemical descriptors, in cases of configurational uncertainty and
inhomogenity, see Stereochemical descriptors in Beilstein Handbook
of Organic Chemistry, Vol 18, Part 1, 4th ed., 5th Suppl., Springer,
Berlin, 1986, p. LXXV.
[43] a) C. L. Folcia, J. Ortega, J. Etxebarria, Liq. Cryst. 2003, 30, 1189–
1191; b) P. Pyc, J. Mieczkowski, D. Pociecha, E. Gorecka, B. Donnio,
D. Guillon, J. Mater. Chem. 2004, 14, 2374–2379.
[44] The relaxation time decreases with increasing frequency of the ap-
plied AC field. The relaxation is very slow (i.e., it takes some mi-
nutes for completion) if the sample is aligned in a low frequency tri-
angular wave field (10 Hz) and it becomes faster (several seconds) if
the frequency is higher (100–500 Hz).
[45] Such a zero voltage centered single peak switching was initially as-
signed to a superparaelectric switching process where macroscopic
FE domains orient under the influence of the field, but are random
without applied field: see reference [38]. The slow relaxation of the
polar states to the dark texture can be regarded as such a type of re-
organization. The dependence of the position of the peaks on the
type of ITO-cell would mean that the fast switching process changes
from FE in noncoated cells to superparaelectric in polyimide-coated
cells. This is however in contradiction with the optical investigations,
which indicate that in the polyimide coated cells, where a centered
peak is observed, the birefringent texture is retained, whereas the
(slow) relaxation to the random organization of the FE domains
takes place in the noncoated ITO cells where a noncentered peak is
observed. This makes an explanation of the peak position based on
the effect of the internal (effective) field more reasonable.
[46] S. Rauch, C. Selbmann, P. Bault, H. Sawade, G. Heppke, O. Mo-
rales-Saavedra, M. Y. M. Huang, A. Jakli, Phys. Rev. E 2004, 69,
021707.
[49] In polyimide coated ITO cells with the same thickness the transition
was found at 1268C, indicating for this compound a significant influ-
ence of the surfaces on this phase transition.
[50] a) B. Zenks, M. Cipec, SPIE Proceedings 1998, 3318, 68; b) T. Mat-
sumoto, A. Fukuda, M. Johno, Y. Motoyama, T. Yui, S-S. Seomun,
M. Yamashita, J. Mater. Chem. 1999, 9, 2051–2080; c) M. A. Osipov,
A. Fukuda, Phys. Rev. E 2000, 62, 3724–3735.
[51] Based on the investigations reported herein it is suggested that also
the monosilylated compounds II adopt a [SmCsPF]aPA ground-state
structure in the dark conglomerate phases, instead of the initially
proposed uniform SmCsPF structure.[19f]
[52] D. A. Coleman, J. Fernsler, N. Chattham, M. Nakata, Y. Takanishi,
E. Korblova, D. R. Link, R.-F. Shao, W. G. Jang, J. E. Maclennan, O.
Mondainn-Monval, C. Boyer, W. Weissflog, G. Pelzl, L.-C. Chien, J.
Zasadzinski, J. Watanabe, D. M. Walba, H. Takezoe, N. A. Clark,
Science 2003, 301, 1204–1211.
[53] F. Araoka, H. Hoshi, H. Takezoe, Phys Rev. E 2004, 69, 051704.
[54] A related sequence of structures resulting from SmCaPA layer stacks
might also be possible for other materials.
[55] a) I. Kovalenko, W. Weissflog, S. Grande, S. Diele, G. Pelzl, I. Wirth.
Liq. Cryst. 2000, 27, 683–687; M. W. Schrçder, S. Diele, N. Pancen-
ko, W. Weissflog, G. Pelzl, J. Mater. Chem. 2002, 12, 1331–1334;
b) J. Mieczkowski, K. Gomola, J. Koseska, D. Pociecha, J. Szydlow-
ska, E. Gorecka, J. Mater. Chem. 2003, 13, 2132–2137; c) A.
Eremin, H. Nadasi, G. Pelzl, S. Diele, H. Kresse, W. Weissflog, S.
Grande, Phys. Chem. Chem. Phys. 2004, 6, 1290–1298; d) H. N.
Shreenivasa Murthy, B. K. Sadashiva, Liq. Cryst. 2004, 31, 567–578;
e) W. Weissflog, U. Dunemann, M. W. Schrçder, S. Diele, G. Pelzl,
H. Kresse, S. Grande, J. Mater. Chem. 2005, 15, 939–946; f) R.
Amaranatha Reddy, B. K. Sadashiva, U. Baumeister, J. Mater.
Chem. 2005, 15, 3303–3316; g) J. P. F. Lagerwall, F. Giesselmann,
M. D. Wand, D. M. Walba, J. Mater. Chem. 2004, 14, 3606–3615.
[56] E. Gorecka, D. Pociecha, J. Mieczkowski, J. Matraszek, D. Guillon,
B. Donnio, J. Am. Chem. Soc. 2004, 126, 15946–15947.
[57] K. Miyasato, S. Abe, H. Takezoe, A. Fukuda, E. Kuze, Jpn. J. Appl.
Phys. 1983, 22, L661–L663.
Received: June 21, 2006
Published online: December 22, 2006
Chem. Eur. J. 2007, 13, 2556 – 2577
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2577