Liquid crystalline behavior of tetraaryl derivatives of benzo[c]cinnoline, tetraazapyrene, phenanthrene, and pyrene: The effect of heteroatom and substitution pattern on phase stability

Article abstract of DOI:10.1039/b615545f

A series of closely related tetrasubstituted derivatives of benzo[c]cinnoline (1), tetraazapyrene (2), phenanthrene (3), and pyrene (4) were investigated for their mesogenic properties using thermal, optical, spectroscopic, and powder XRD analyses. Only three 3,4-dioctyloxyphenyl derivatives exhibited mesogenic properties. Substitution of N for CH (3 → 1 and 4 → 2 pairs) and also increase of the core element size (1 → 2 and 3 → 4 pairs) significantly increases the mesophase stability. The findings and observed trends were rationalized by analysis of conformational properties which included calculation of the planarization energy, and modeling of aliphatic chain density and fill fractions. MO calculations showed that the tetraaza derivative 2c is significantly electron deficient and suitable for electron conductive materials. The Royal Society of Chemistry.

Full text of DOI:10.1039/b615545f

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/materials | Journal of Materials Chemistry
Liquid crystalline behavior of tetraaryl derivatives of benzo[c]cinnoline,
tetraazapyrene, phenanthrene, and pyrene: the effect of heteroatom and
substitution pattern on phase stability{{
b
Monika J. Sienkowska,§a John M. Farrar,"a Fan Zhang,I Sharat Kusuma,a Paul A. Heineyb and
Piotr Kaszynski*a
Received 26th October 2006, Accepted 11th December 2006
First published as an Advance Article on the web 15th January 2007
DOI: 10.1039/b615545f
A series of closely related tetrasubstituted derivatives of benzo[c]cinnoline (1), tetraazapyrene (2),
phenanthrene (3), and pyrene (4) were investigated for their mesogenic properties using thermal,
optical, spectroscopic, and powder XRD analyses. Only three 3,4-dioctyloxyphenyl derivatives
exhibited mesogenic properties. Substitution of N for CH (3 A 1 and 4 A 2 pairs) and also
increase of the core element size (1 A 2 and 3 A 4 pairs) significantly increases the mesophase
stability. The findings and observed trends were rationalized by analysis of conformational
properties which included calculation of the planarization energy, and modeling of aliphatic chain
density and fill fractions. MO calculations showed that the tetraaza derivative 2c is significantly
electron deficient and suitable for electron conductive materials.
In this context, we designed a class of compounds of
Introduction
structure I in which the parent heterocycles, 1a and 2a, are
Over the past decade, discotic liquid crystals2,3 have been
substituted by four phenyl groups bearing flexible alkoxy
recognized4 as potential materials for energy harvesting,5 light
substituents to promote mesogenic behavior. For the purpose
emitting diodes,6,7 and organic electronics.8 All these applica-
of the present investigations, the octyl group was selected as an
tions involve charge transport in the bulk phase, and for this
alkyl chain. Derivatives 1 and 2 provide a rare opportunity for
reason the phase structure, electronic properties of the
comparison of discogenic properties of heterocyclic systems
aromatic core, and the molecular packing are particularly
with their carbocyclic analogs. Therefore, derivatives of
relevant to the material’s performance. Discotics with
phenanthrene (3a) and pyrene (4a) were also included in the
electron-rich cores are relatively numerous and many show
studies.
facile hole transport properties.9 On the other hand, there are
relatively few p-deficient discotic mesogens suitable for
efficient electron transport. The majority of them are based
on nitrogen-containing polycyclic heteroaromatics such
as tricycloquinazoline,10,11 hexaazatriphenylene,12–14 and
hexaazatrinaphthylene.15–17 It can be expected that other
heterocycles such as benzo[c]cinnoline (1a) and 4,5,9,10-
tetraazapyrene (2a) are also suitable structural elements for
n-type organic semiconductors. Moreover, these two hetero-
cycles provide convenient models for our design of discotic
radicals in which the –NLN– group is a substitute for the spin-
bearing thioaminyl group.18,19
aOrganic Materials Research Group, Department of Chemistry,
Vanderbilt University, Nashville, TN, 37235, USA
bDepartment of Physics and Astronomy, University of Pennsylvania,
Philadelphia, PA, 19104, USA
{ This work has been described in part previously; see ref. 1.
{ Electronic supplementary information (ESI) available: XRD, DSC,
synthetic and computational details including archives for the MP2/6-
31G(d) results, colour photomicrographs, and literature redox data.
See DOI: 10.1039/b615545f
§ Current address: University of Pennsylvania, Philadelphia, PA
19104, USA.
" Current address: Abraham Baldwin Agricultural College, Tifton,
GA 31793, USA.
Here we report detailed studies of the influence of the
number and distribution of the octyloxy chains on the solid
and mesogenic properties in series 1–4. We also analyze the
I Current address: Argonne National Laboratory, Argonne, IL 60439,
USA.
effect of a replacement of the bridging –NLN– group in
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 1399–1411 | 1399

View Article Online
of biphenyls 8. The overall yields for tetraazapyrenes 2b–2d
were higher than those in the first method and ranged from
20% to 29%. Thus, the overall synthesis of 2 was significantly
improved in the second method by the ease of the reaction
monitoring, purification of products and their intermediates,
and generally higher yields.
All products 1–4 were purified by gradient column
chromatography using a hexanes–CH2Cl2 mixture as eluent.
However, the chromatographic purification of tetraazapyrene
derivatives 2 was not trivial, due to a number of side products
formed during the reduction, and the apparently low stability
of tetraazapyrene derivatives on silica gel. Chromatographic
purification was followed by recrystallization, typically from
hexane, ethanol–toluene, or 1-propanol except for 1d, which
appeared to be a liquid at room temperature.
Scheme 1
heterocycles 1 and 2 with the –CHLCH– group in hydro-
carbons 3 and 4 on the stability of liquid crystalline phases.
Unfortunately, 1,3,6,8-tetrakis(3,5-dioctyloxyphenyl)phe-
nanthrene (3d) could not be obtained in the pure form.
According to the NMR analysis, 3d was obtained as an oily
mixture of several species and could not be separated either by
chromatography or by recrystallization. However, based on
trends in thermal behavior of other derivatives in the series
(vide infra), it was concluded that 3d would most likely not
exhibit liquid crystalline behavior, and it was not pursued
further.
Results
Synthesis
Tetraarylarenes. Tetraarylarenes 1–4 were prepared from the
corresponding tetrabromoarenes 1e–4e and appropriate boro-
nic esters 5 using the general Suzuki coupling procedure20
(Scheme 1). Yields of the coupling products were improved by
using DME as a solvent instead of toluene. In each reaction,
biphenyls 6 were obtained as byproducts in the homocoupling
reaction of boronic esters 5.
The purity of all teraarylated arenes 1–4 was confirmed
by normal phase analytical HPLC, with a hexane–CH2Cl2
mixture as an eluent.
Tetrahaloarenes. 1,3,6,8-Tetrabromobenzo[c]cinnoline22 (1e,
Scheme 3), was synthesized according to a modified literature
method. Catalytic reduction of 2,29-dinitrobiphenyl23 (9)
followed by bromination of the resulting 2,29-diaminobiphenyl
(10) in a mixture of acetic and hydrochloric acids gave the
tetrabromo derivative 11 in 64% overall yield. The diamine 11
was oxidized with peracetic acid to form 1e in 88% yield.22
The synthesis of 1,3,6,8-tetrabromo-4,5,9,10-tetraazapyrene
(2e, Scheme 3) was attempted in a similar way to the
preparation of 1e starting from 2,29,6,69-tetranitrobiphenyl21
(12), which was prepared in the Ullmann coupling of 2,6-
dinitrochlorobenzene. Catalytic reduction of 12 gave tetra-
amine21 13 as a dark red solid in 88%–90% yield. Without
further purification, 2,29,6,69-tetraaminobiphenyl (13) was
brominated in a mixture of acetic and hydrochloric acids to
yield 61%–66% of 2,29,6,69-tetraamino-3,39,5,59-tetrabromobi-
phenyl (14). The product was purified chromatographically
with CH2Cl2 as an eluent. The subsequent closure of the azo
bridges by peracetic acid oxidation was not trivial. A brownish
suspension of tetraamine 14 in acetic acid turned black after
the addition of peracetic acid and the elevation of the
temperature to 60 uC. After 2–3 days of stirring at 60 uC the
precipitate was filtered off and the resulting black micro-
crystalline material, presumably product 2e, was washed with
hot chloroform several times. The NMR data of the black
solids, which were partially soluble in CDCl3 and obtained
from different trial reactions, were inconsistent and hard to
unambiguously interpret. Elemental analysis data, obtained
for both the sample filtered off directly from the reaction
mixture and for the sample washed three times with chloro-
form, showed a general consistency with the composition of
Most reactions were completed overnight. Coupling reac-
tions using the crude 1,3,6,8-tetrabromo-4,5,9,10-tetraazapyr-
ene (2e) yielded tetraarylarenes 2b–2d in yields below 3%.
Therefore, an alternative method was developed for the
preparation of these derivatives (Scheme 2), based on the
reported preparation of the parent tetraazapyrene 2a.21 Thus,
the Suzuki coupling reaction of tetrachloride 7 and appro-
priate boronic ester 5 gave tetraarylbiphenyl 8, which was
easily purified by column chromatography. The two azo
bridges were closed by the reduction21 of 8 with freshly
prepared Raney nickel under basic conditions to give tetra-
azapyrenes 2b–2d. The literature procedure was modified by
using 1-propanol instead of ethanol, to increase the solubility
Scheme 2
1400 | J. Mater. Chem., 2007, 17, 1399–1411
This journal is ß The Royal Society of Chemistry 2007

View Article Online
Scheme 5
Boronic esters. Boronic esters 5 were prepared according to a
general method,27 in which the appropriate octyloxybromo-
benzenes 18, prepared by O-alkylation of bromophenols 17,
were treated with n-butyllithium followed by triisopropyl
borate (Scheme 5). Esterification of the resulting boronic
acids28 with ethylene glycol afforded boronic esters 5, which
were easy to purify by distillation and/or recrystallization.
5-Bromopyrocatechol (17c) was prepared from 5-bromo-
salicylaldehyde using the literature procedure.29
3,5-Dioctyloxybromobenzene (18d) was prepared in two
ways. In the first method, 3,5-dimethoxyaniline was diazotized
and converted to bromo-3,5-dimethoxybenzene (19).30 The
Scheme 3
the tetrabromide 2e. Both analyses showed that the content of
carbon was higher by about 1.3%–1.8% relative to that
expected for the pure compound. Also, the filtrate obtained
after the separation of the black solid (presumably 2e), poured
into water gave a brown solid, which was difficult to purify
and analyze. The use of this brown material for Suzuki
coupling did not give the expected tetraarylated product, as
confirmed by TLC. The black microcrystalline material gave
less than 3% of the expected 1,3,6,8-tetraaryl-4,5,9,10-tetra-
azapyrene in Suzuki reaction with boronic ester 5.
31
methyl groups were removed with BBr3 (Scheme 6) and the
resulting 5-bromoresorcinol (17d) was alkylated with 1-bro-
mooctane to form 18d in an overall yield of 22%. Given the
relatively high cost for the aniline, we explored another route
to 18d starting from 3,5-dihydroxybenzoic acid (Scheme 7).
According to a literature procedure, 3,5-dihydroxybenzoic
acid was converted to its methyl ester,31 which was O-alkylated
with 1-bromooctane and subsequently hydrolyzed.32 The
resulting 3,5-dioctyloxybenzoic acid (20) was transformed to
18d using a Barton reaction procedure for an analogous
compound.33 The intermediate acid chloride was obtained
The preparation of 1,3,6,8-tetrabromophenanthrene (3e)
is described elsewhere,24 and the synthesis of 4e followed
literature procedure.25
The preparation of tetrachloride 7 was attempted by
stoichiometric Ullmann coupling reaction24 of commercially
available 1,3,5-trichloro-2,6-dinitrobenzene (15) with half an
equivalent of freshly prepared copper in 5-tert-butyl-m-xylene
(Scheme 4). However, low regioselectivity and undesired
side reactions significantly lowered the yield of pure 7 and
complicated its purification. Therefore, 1-bromo-3,5-dichloro-
2,6-dinitrobenzene26 (16) in dry xylenes was used instead for
Ullmann coupling giving 7 in 33% yield (or 72% based on
recovered starting 16).
Scheme 6
Scheme 4
Scheme 7
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 1399–1411 | 1401

View Article Online
Table 1 Transition temperatures and enthalpies for derivatives 1–4a
b
c
d
Cr
Cr
Cr
Cr1
133
59
I
I
I
Cr1
Cr1
Cr
19 Cr2
5
92 Colh 168
12
I
[Colh 32]b
I
20
1
2
3
4
161
39
46 Cr2
3
95 Colh 199
69 11
I
Cr
45
92
I
c
152
73
101
44
I
67 Cr2 95 Cr3 116 Id Cr
43 50
73 Colh 87
55 24
I
Cr
81
25
[Colh 25]b
I
—
a
Transition temperatures (uC) and enthalpies (in italics, kJ mol21) were determined by DSC (10 uC min21; 5 uC min21 for 2) in the heating
b
c
d
mode: Cr = crystalline; Colh = columnar hexagonal; I = isotropic. Virtual isotropic transition. Not investigated. Melting at 89 uC followed
by crystallization at 95 uC. See text for details.
from acid 20 as a dark oil, which partially decomposed upon
attempted vacuum distillation. Therefore, the acid chloride
was used as a crude product, giving the bromide 18d in a
moderate overall yield of 41% based on 20.
regions of a uniaxial phase (Fig. 1a). Upon cooling, 1c formed
a rigid phase, which optically was not different from the
preceding mesophase. This observation was confirmed by
DSC, and on the basis of XRD analysis (vide infra) it was
tentatively assigned as a columnar crystal (Crcol). The intense
dark red color of tetraazapyrene derivative 2c did not allow for
a detailed optical analysis of its texture.
Liquid crystalline properties
Tetraaryl derivatives 1–4 (except for 3d) and biphenyls 6
were investigated by polarized optical microscopy (POM)
and differential scanning calorimetry (DSC) for mesogenic
behavior. Five of them were subsequently studied by variable
temperature powder X-ray diffraction (XRD). Results of DSC
analysis for 1–4 are collected in Table 1 and XRD data are
shown in Table 2.
Other compounds in the series 1–4, except for 1d, showed
only melting to an isotropic phase upon heating, and crystal-
lization upon cooling, often exhibiting significant supercooling
of the isotropic phase. Benzo[c]cinnoline derivative 1d exists
as an isotropic liquid. Analysis of biphenyls 6b–6d showed
no liquid crystalline behavior even in samples supercooled by
up to 10 uC.
Polarized optical microscopy (POM). Optical investigations
showed that only three compounds with 3,4-dioctyloxyphenyl
substituents, 1c, 2c, and 4c, exhibit liquid crystalline behavior.
The mesophase of 2c and 4c supercooled forming a glass at
room temperature. Neither phenanthrene derivative 3b nor 3c
exist as liquid crystalline material. Mesophases obtained on
cooling of the three mesogenic compounds exhibited fan- or
flower-shaped textures, which are characteristic for columnar
phases34 (Fig. 1). The texture of benzo[c]cinnoline derivative 1c
also displayed domains typical for homeotropically oriented
Differential scanning calorimetry (DSC). DSC analysis
confirmed optical observations that only 1c, 2c, and 4c exhibit
liquid crystalline properties (Table 1). Benzo[c]cinnoline 1c
showed three transitions. On cooling, the 168 uC transition
showed practically no hysteresis, while the two others
exhibited a modest supercooling by 6 uC for the 92 uC
transition, and 10 uC for the 19 uC transition (Fig. 2). This is
consistent with the appearance of progressively more viscous
and organized phases. The enthalpy of the clearing transition
1402 | J. Mater. Chem., 2007, 17, 1399–1411
This journal is ß The Royal Society of Chemistry 2007

View Article Online
The first DSC heating curve of pyrene 4c showed a broad
transition at about 60 uC followed by two well-defined broad
transitions at 73 uC and at 87 uC. The cooling trace displayed
only one transition at 77 uC. This shows a bigger hysteresis of
the isotropic–discotic transition as compared with the results
for 1c and 2c. No crystallization peak was observed at a
scanning rate of 10 uC min21, and the hexagonal phase most
likely formed a glass phase at room temperature. This is
consistent with microscopic observations (vide supra). On the
second heating curve obtained immediately after the cooling
experiment, only a transition at 87 uC with an enthalpy of
18 kJ mol21 was detected.
DSC thermograms of the remaining compounds generally
showed only transitions corresponding to melting and crystal-
lization processes, often with significant supercooling. The
only exceptions are benzo[c]cinnoline 1d and pyrene 4b. The
former exists as an isotropic liquid and DSC analysis did not
detect any transition upon cooling to 2120 uC. The thermal
behavior of pyrene derivative 4b is exceptional in the series,
since the compound exhibits rich crystalline polymorphism
with a peculiar melting–crystallization transition at 89 uC
and 95 uC (Fig. 3). Solid and liquid coexist in this range of
temperatures and are clearly observed under the microscope. A
similar phenomenon was observed previously for a calamitic
liquid crystal.36 Upon cooling, the isotropic liquid of 4b was
supercooled by about 45 uC and the material partially crystal-
lized over a range of 30 uC. Subsequent heating of the sample
revealed a continued crystallization in the range of 35 uC–
60 uC. The Cr–Cr transition at 67 uC observed for the virgin
sample was absent but the melting-crystallization and the
subsequent melting transitions were fully reproduced. Similar
behavior of 4b was recently reported in the literature.37 The
published DSC data indicate that the sample was largely
glassified during the crystallization process (crystallization
peak at 51 uC). Also the peaks with negative enthalpies at
89.5 uC and 112.5 uC in the reported37 DSC trace are most
likely due to evaporation of solvents (ethanol and toluene)
trapped in the solid during crystallization. This underscores
the necessity of verifying compound purity with combustion
analysis.
Fig. 1 Optical textures of (a) 1c at 155 uC and (b) 4c at 75 uC. Both
textures were obtained upon cooling. Magnification 6300.
at 168 uC is 20 kJ mol21, which is consistent with that for other
discotic materials.35
The first heating DSC curve of tetraazapyrene 2c also
showed three transitions, two of which were not observed
in the cooling cycle. Consequently, the mesophase was
supercooled to ambient temperature, presumably forming a
glass. The subsequent heating showed a broad annealing
peak and partial crystallization at about 85 uC followed by
melting at 99 uC.
Fig. 2 DSC traces of benzo[c]cinnoline 1c. The heating rate was
10 uC min21
Fig. 3 DSC traces of pyrene derivative 4b. The heating rate was
10 uC min21
.
.
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 1399–1411 | 1403

View Article Online
Table 2 X-Ray diffraction data
˚
d/A
˚
Cell const./A
Sample
Temp/uC
phase
Colh
hkl
1c
130
100
110
200
halo
001
100
210
halo
001
100
110
200
210
300
220
400
500
halo
001
25.3
14.6b
12.7b
4.55b
3.5b
25.2
9.5
a = 29.2a
2c
141
80
Colh
Colh
a = 29.1a
a = 28.9a
4.4
,3.7
25.6
14.1
12.5
9.5
8.2
6.9
6.2
4.9
4cc
Fig. 4 Clearing temperature TI taken as the peak of the transition vs.
mol fraction of 1d in 1c and 4d in 4c. The extrapolated virtual clearing
temperatures: 32 ¡ 11 uC for 1d (r2 = 0.97) and 26 ¡ 5 uC for 4d
(r2 = 0.95).
4.55
3.9
Binary mixtures. Virtual isotropic transitions of the non-
mesogenic 3,5-dioctyloxy derivatives d were estimated by
analysis of binary mixtures with their mesogenic 3,4-dioctyloxy
analogs c using POM and DSC. Both types of analyses
demonstrated broad clearing transitions for solutions con-
taining about 10 mol% and 20 mol% of d in c. A plot of the
transition peak (DSC) vs. concentration for each pair of
analogs is shown in Fig. 4.
a
b
Indexed to a hexagonal lattice: a = d100 6 2/!3. The wide-angle
data for 1c were collected using the Scintag system (see ESI), while
the other data were collected with the HiStar area detector. Ref. 38.
c
annealing at 80 uC. This indicates that, in contrast to the
mono-octyloxy derivatives, compounds 1c–3c form poorly
ordered crystalline phases, presumably due to lower symmetry
of the molecules and a stronger tendency to disorder.
Both benzo[c]cinnoline 1d and pyrene 4d decrease clearing
temperatures of their 3,4-dioctyloxy derivatives with increasing
concentration. For the tetraazapyrene, 2, the data appear to
not follow the sequence observed for 1 and 4. Linear
extrapolation of the data by setting the intercept for each
function at the value for the pure host gave virtual transition
temperatures for the non-mesogenic derivatives: 32 uC for 1d
and 26 uC for 4d. The same analysis for 2d using only the zero
and high concentration data points gave the virtual clearing
temperature of 277 uC.
In the isotropic phase the benzo[c]cinnoline and phenan-
threne derivatives exhibited a typical diffuse feature at
˚
about 22 A, which corresponds to short-range intermolecular
correlations. In addition, phenanthrene derivatives 3b (161 uC)
and 3c (137 uC) showed a broad feature with a maximum at
about 4.5 A, which for benzo[c]cinnolines 1b (140 uC) and 1c
(179 uC) presumably appears outside the measurement window
(2h . 23u).
Analysis of benzo[c]cinnoline 1c and tetraazapyrene 2c in
temperature regions identified to be liquid crystalline showed
diffraction patterns that can be ascribed to a two-dimensional
hexagonal columnar lattice.39 Both compounds display a sharp
˚
Powder X-ray diffraction (XRD). Five compounds in series
1–3 were investigated by powder X-ray diffraction to deter-
mine their phase structure. Data were collected at sample–
detector distances of 54 and 124 cm at several temperatures
chosen typically in the middle of the phase range determined
by DSC. XRD results for the mesogenic derivatives 1c and 2c
are collected in Table 2 and compared to those reported
elsewhere for 4c.38
˚
high intensity peak at d of about 25 A indexed as the (100)
peak (Fig. 5) which corresponds to a cell parameter a of about
˚
29 A (Table 2). The benzo[c]cinnoline 1c also displays two
much weaker higher order reflections in the expected spacing
In a typical experiment, freshly prepared powder samples
were analyzed at 25 uC. The non-mesogenic compounds 1b, 3b,
and 3c were then analyzed at 80 uC, and subsequently in the
isotropic phase. The mesogenic benzo[c]cinnoline derivative 1c
was analyzed in the isotropic phase, cooled to a mesophase,
and finally measured again at room temperature. The tetra-
azapyrene 2c was analyzed only in the heating mode.
Analysis of freshly prepared samples at 25 uC showed a
single sharp reflection in the small angle region and a variety of
features at higher angles in all five compounds. Indexing of the
first few diffraction peaks allow for the tentative determination
˚
of two out of the three unit cell dimensions for 1b (a = 31.8 A
˚
˚
˚
and b = 29.2 A) and 3b (a = 21.5 A and b = 26.5 A). The
diffractograms for 1c, 2c, and 3c had only poorly resolved
features in the wide angle region, which became sharper upon
Fig. 5 X-Ray diffraction pattern for 2c at 141 uC.
1404 | J. Mater. Chem., 2007, 17, 1399–1411
This journal is ß The Royal Society of Chemistry 2007

View Article Online
Table 3 Rotational transition state energies for phenyl derivatives 21a
ratio of d100/!3 and d100/2, which are indexed as the (110) and
(200) peaks.40 These reflections are too weak to observe in
the diffractogram of 2c, which shows only a weak (210) peak
DE{
DFT
MP2
DH{
DFT
MP2
DG{
DFT
MP2
h
DFT
MP2
˚
at d = 9.5 A (Fig. 5).
Compound
The wide-angle data for 1c and 2c show a diffuse peak at
˚
21a
Csb
4.0
6.6
3.5
6.1
4.9
7.4
43
46.5
4.5 A which arises from correlations between molten alkyl
chains. The diffractogram40 of 1c also shows a sharper reflec-
˚
tion at 3.5 A which corresponds to the distance between
cores. Unfortunately, the maximum of this peak for 2c lies
˚
outside of the measurement window (,3.7 A). The presence
of this reflection as well as its sharpness indicates the degree
of intra-columnar correlations of the hexagonal columnar
phase.39
21b
C1
12.6
15.2
12.1
14.7
13.4
16.0
55
57 (exp. 60–65)c
b
Upon cooling to ambient temperature, the wide-angle
diffraction pattern shows that the hexagonal mesophase of
tetraazapyrene 2c crystallized. In contrast, the diffraction
pattern of benzo[c]cinnoline 1c changed very little and no
additional reflections appeared in the WAXS diffractogram.
The only noticeable change is a split of the 100 peak. This
suggests the formation of a glass phase with a rectangular or
oblique structure, tentatively assigned as Colr.
21c
D2h
2.1
3.9
1.6
3.4
2.9
4.8
38
45 (exp. 45)d
b
Molecular modeling
a
DFT calculation at the B3LYP/6-31G(d) level of theory and MP2
at the MP2(fc)/6-31G(d) level with DFT thermodynamic corrections.
Energies in kcal mol21 and angles in u. The GS geometry dihedral
To gain a better understanding of the observed phase behavior
of and XRD results for compounds in series 1–4, we per-
formed detailed conformational analysis of relevant molecular
fragments and also derivatives 2 as representative for the
series.
b
angle h is defined by marked atoms in each structure. Symmetry of
c
the TS. Ref. 41. Ref. 42.
d
observed for biphenyl (21c). These values are moderately
larger for cinnoline 21a, but significantly greater for 1-phe-
nylnaphthalene (21b). A comparison of the computational
results with scant literature data shows that the MP2 method
gives more accurate results than the DFT method. Thus,
biphenyl was found to have a 45u dihedral angle between the
benzene ring planes in the gas phase,42 while in the solid state
the crystal packing forces largely overcome the hydrogen–
hydrogen interactions and biphenyl is nearly planar.43 In
1-phenylnaphthalene (21b) this dihedral angle is larger and
between 59u–66u, according to other quantum mechanical
calculations44 and solution phase experiments.41,45
Inspection of the general molecular model I (vide supra)
shows that the overall geometry of compounds 1–4 depends
mainly on three factors: i) orientation of the phenyl substituent
relative to the central core, ii) orientation of the alkoxy
substituents relative to each other and to the phenyl ring, and
iii) conformational properties of the alkyl chains. In addition,
the energy required to place the phenyl ring co-planar with the
central core (planarization energy) indicates the compressi-
bility of the stacks and minimization of the core–core mean
distance in the columnar phase.
The phenyl–central core conformational properties and the
height of the rotational barrier were assessed at the DFT and
MP2 levels of theory using compounds 21a–21c as models
representing the key fragment of the molecule (Fig. 6).
Results collected in Table 3 show that the smallest deviation
from co-planarity and smallest planarization energy are
Detailed analysis of the computational results shows that in
…
…
the equilibrium geometry the non-bonding N H and H
H
distances in the cinnoline and naphthalene derivatives, 21a and
˚
21b, are calculated at about 2.6 A at the MP2 level of theory,
while the DFT method predicts these distances shorter by
…
˚
0.05 A. The H H distance calculated for biphenyl by the two
˚
methods is shorter (2.45 A and 2.37 A) and comparable with
˚
˚
the sum of van der Waals radii (2.4 A). This difference in the
…
H
H separations is consistent with the smaller dihedral angle
h for biphenyl (21c, 45u) than for 1-phenylnaphthalene (21b,
57u). In the rotational transition state, cinnoline 21a and
…
biphenyl 21c adopt planar structures in which the N H and
…
˚ ˚
H distances decrease to 2.08 A and 1.95 A, respectively,
H
according to the MP2 results. Steric interactions of the
hydrogen atoms in the naphthalene derivative 21b prevent
the formation of the planar transition state. Instead, the
naphthalene ring is warped (h = 8.5u at DFT and h = 11.7u at
MP2) and the H H distance is very short (1.62 A at DFT and
˚
1.68 A at MP2).
Fig. 6 Phenyl derivatives 21 (outlined) as fragments of 1c (21a: X =
N, Y = H; 21c: X = H, Y = N), 2c (21a: X = Y = N), 3c (21b: X = CH,
Y = H; 21c: X = H, Y = CH), and 4c (21b: X = Y = CH).
…
˚
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 1399–1411 | 1405

View Article Online
which compares well with the calculated 46.5u for phenylcinno-
line (21a, Table 3).
Results in Fig. 8 show that for the all-trans conformers, the
width of the molecule is smallest for the 3,4-derivative 3,4-t,t
˚
˚
(13.5 A) and the largest for 3,5-syn-t,t (17.5 A) and it scales
with sinh where h is the angle between the phenyl ring and the
heterocycle. Introduction of the gauche conformation at the
C1–C2 bond, at a cost of about 0.9 kcal mol21 per bond,
significantly lowers the molecular extent in the z direction. The
molecular thickness for 3,4-g,t and 3,5-g,g derivatives with two
Fig. 7 Arrangement of two alkoxy substituents. Only the first two
carbon atoms are shown for simplicity.
˚
gauche conformations is about 9 A. It can be estimated that for
˚
the molecular width of 3.7 A, consistent with the XRD data,
Conformational properties of the alkoxy chains were
assessed based on literature reports, and their impact on the
molecular shape of 1–4 was evaluated using molecular
mechanics. A literature search revealed several solid-state
structures containing either the 3,4-dialkoxyphenyl (II) or 3,5-
dialkoxyphenyl (III) substituents (Fig. 7). In the former the
two vicinal alkoxy groups are coplanar with the benzene
ring and the alkyl chains extending in opposite directions
(Fig. 7).46–48 To improve crystal packing, one46,48 or both47
alkyl chains adopt a gauche conformation so that the two
chains can propagate parallel to each other. The two solid state
structures of 3,5-dialkoxyphenyl (III) reported in the litera-
ture49,50 indicate that the two substituents are coplanar and
propagate in the same direction (a syn orientation). The other
two symmetric orientations of the alkoxy chains (out and in)
have not been reported.
the angle h should be about 18u, instead of 50u calculated for
an isolated molecule.
Projections in Fig. 8 hint at difficulties in accommodation of
3,5-dioctyloxyphenyl substituents in the planarized conforma-
tions (small h). Analysis suggests that a significant number of
gauche conformations are required for achieving geometry
favorable for the formation of a mesophase. In contrast, in
the 3,4-dioctyloxy derivatives one gauche conformation in the
3-octyloxy chain is sufficient for significant ‘‘flattening’’ of
the molecule and the 4-octyloxy chains, that are projected
outwards, do not hinder the planarization of the phenyl–
heterocycle torsional angle h. A projection of a geometry-
optimized molecule of 2c with one possible orientation of the
3,4-octyloxyphenyl rings is shown in Fig. 9. It is conceivable
that the second molecule will adopt an orientation 90u rotated
relative to the first molecule to fill in the gap. Such a possibility
was suggested for another derivative of pyrene.37
Orientation of the alkoxy substituents shown in Fig. 7
suggested the starting points for molecular geometry optimiza-
tions of tetraazapyrenes 2c and 2d as representatives for the
series. For simplicity only one dialkoxyphenyl substituent
was included, and the structures were optimized using the
universal force field (UFF) algorithm. Initially, both alkyl
chains were set in all-trans conformation coplanar with the
phenyl ring and results of geometry optimization for the 3,4-
dioctyloxy- and syn and in 3,5-dioctyloxyphenyl derivatives
are shown in Fig. 8. Subsequently a gauche conformation
was introduced between the C1 and C2 atoms in the octyloxy
chain at the 3 position of the 3,4-derivative, and in both chains
of the 3,5-derivative. In all calculations, the torsional angle
between the phenyl ring and the heterocycle was about 50u,
Electronic structure
To estimate charge accepting abilities of the discotics, energies
for the two frontier molecular orbitals (FMOs) of the central
core elements 1a–4a and a pair of triphenylene–hexaazatri-
phenylene were calculated at the MP2/6-31G(d) level of theory.
Results in Fig. 10 show that among the three hydrocarbons,
pyrene (4a) has the highest lying HOMO (26.87 eV) and hence
it is the best electron donor. Conversely, its tetraaza analog,
2a, has the lowest LUMO energy (0.30 eV) among the three
heterocycles, and it appears to be the best electron acceptor in
the series. The substitution of N atoms into pyrene shifts the
Fig. 8 Selected conformers of model for 2c and 2d viewed along the plane containing the central heterocycle. For clarity only one
dioctyloxyphenyl substituent is used in the modeling. Geometry was optimized with universal force field for alkyl chains in trans and gauche
conformation imposed for the C1–C2 bond.
1406 | J. Mater. Chem., 2007, 17, 1399–1411
This journal is ß The Royal Society of Chemistry 2007

View Article Online
Fig. 11 UV-Vis spectra for 4-octyloxyphenyl derivatives of benzo[c]
cinnoline 1b, tetraazapyrene 2b, phenanthrene 3b, and pyrene 4b
recorded in cyclohexane. Each spectrum is plotted in 0–105 molar
absorptivity scale and the spectra are stacked for clarity. The vertical
line marks the UV-Vis border.
Fig. 9 A view of the most extended confromation of 2c along the z
axis perpendicular to the heterocycle plane. Conformation of each
substituent corresponds to 3,4-g,t in Fig. 8. The geometry of 2c-g,t was
constrained in the Cs symmetry and minimized using the PM3 method.
˚
The inscribed circle has a diameter of 36 A.
groups results in a red-shift of their absorption bands. Among
the four compounds, only the phenanthrene derivative 3b is a
colorless solid. The longest wavelength absorption bands in 1b
and 4b tail in to the visible region (the gray vertical line in
Fig. 11) giving them a yellowish color. Compound 2b has two
intense absorption bands in the visible region which give it the
dark red color.
FMOs by about 1.5 eV, which is about 50% more than
observed in the phenanthrene–benzo[c]cinnoline pair. Trends
shown in Fig. 10 are consistent with generally observed
trends in experimental redox potentials, electron affinities, and
ionization potentials (for details see ESI{).
The moderate intensity (log
e =
2.4)51 Lb band in
Electronic absorption spectroscopy
benzo[c]cinnoline 1a appears to be shifted in 1b by about
30 nm to 373 nm (log e = 4.46). Substitution in the parent
tetraazapyrene 2a red-shifts its two longest wavelength
absorption bands, at 373 nm (log e = 3.79, EtOH) and
391 nm (log e = 3.87, EtOH),52 by about 70 nm in 2b. In the
parent phenanthrene 3a the longest wavelength Lb band
(log e = 2.4)51 is replaced with an intense band with a
maximum at 335 nm (log e = 4.53) in 3b. This band is likely to
be related to the La band of 3a red-shifted in 3b by about
40 nm. The lowest energy absorption band for pyrene 4b has a
maximum at 390 nm (log e = 4.64), which is shifted by 55 nm
UV-Vis absorption spectra for the 4-octyloxyphenyl deriva-
tives 1b–4b in cyclohexane are shown in Fig. 11. Substitution
of the parent compounds 1a–4a with four 4-octyloxyphenyl
1
relative to the allowed La transition or 35 nm relative to the
1
forbidden Lb band in pyrene (4a).53
Discussion
Analysis of results in Table 1 demonstrates the significance of
conformational, steric, and packing density (space filling)
effects on mesophase stability. A comparison of the effective-
ness of the substituents b–d in phase stabilization shows that
only 3,4-dioctyloxyphenyl (c) induces mesogenic behavior
presumably by striking a balance between the density of the
alkyl chains and their preferred orientation. The octyl group of
the mono-octyloxyphenyl substituent (b) propagates in the
plane close to that of the central core but the alkyl chain
density (fill fraction) is apparently too low to support a
Fig. 10 The HOMO and LUMO energy levels for hydrocarbons and
their aza analogs obtained at the MP2/6-31G(d) level. Numerical
values are listed in the ESI.
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 1399–1411 | 1407

View Article Online
38
˚
mesogenic state. Addition of a second chain in the 3-position
in c provides the necessary density for appearance of a
mesophase apparently at the modest expense of one gauche
conformation (Fig. 8). A similar induction of a columnar
mesophase was observed for other pairs of compounds.38,54
Moving the octyloxy group from position 4 to 5 in substituent
3.7 A. These large differences in the core–core distance can
be related to the planarization energy for each phenyl group.
Thus, the largest intermolecular separation is observed for
pyrene, which is consistent with the steepest rotational
potential for the Ph–pyrene bond with the maximum at
about 15 kcal mol21. In contrast, for tetraazapyrene and
benzo[c]cinnoline this potential energy curve has a maximum
of only about 6 kcal mol21 and the dihedral angle between the
phenyl and the core can be compressed more easily.
d
increases steric interference of the alkyl chains and
hinders molecular planarization. This leads to a significant
broadening of the molecule and to a lower fill fraction.
Consequently, the mesophases are dramatically destabilized by
136 uC for 1d and 64 uC for 4d, as estimated from binary
mixtures. In these derivatives, the contact between the rigid
cores is so difficult that compounds have either low melting
points (e.g. 2d) or are liquids at ambient temperature (1d, see
also ref. 38). Also, the experienced difficulties with crystal-
lization and purification of 3d can be largely ascribed to the
significantly hindered core–core contact. The unusually low
enthalpy of melting for 4d is presumably also due to the poor
packing and loose core–core contact.
Modeling results show that dimensions for molecules with
˚
the most extended chains (cf. Fig. 9) are about 7 A larger than
the intercolumnar distance established by XRD. This suggests
that the last few methylene units are ‘‘melted’’ and folded up,
which is consistent with the observed diffraction patterns and
abnormal intensity of the (110) and (200) peaks for 2c and 4c.
The low intensity, or even absence, of (110) and (200)
reflections in the X-ray diffraction pattern of 2c (Fig. 5) was
unexpected, although similar situations have been previously
observed in molecules with unusual charge distributions.55–58
Indeed, for columnar derivatives of [5]helicene56 and triphe-
nylene57 the intensity of the (110) reflection was almost zero,
the (200) was more intense, and the strongest higher-order
reflection was the (210). A very similar situation is observed in
the diffraction pattern of 2c in which only the (210) higher-
order reflection is clearly visible and the (110) and (200) are at
the noise level. In the diffraction pattern of 4c, a close analog
of 2c, the (110) and (210) peaks are very weak.38
The significance of the high fill fraction for mesogenic
compounds is demonstrated by a pair of discogens 1c and 2c.
At first glance it would appear that the moderate dipole
moment of the benzo[c]cinnoline ring and more compressible
biphenyl-type torsional angles in 1c would result in higher
stability of the mesophase. Experiment shows, however, that
incorporation of another azo group into the central hetero-
cycle increases the clearing temperature by over 30 uC in 2c, in
spite of the loss of the net molecular dipole moment and
increase of the planarization barrier for two substituents in 2c.
It is plausible that this stabilization of the mesophase is related
to the higher fill fraction of the central core in 2c than in 1c.
This appears to be valid also for the pair 3c and 4c for which
the difference in the mesophase stability is at least 60 uC.
The third type of structural effect on the mesophase
thermodynamic stability is the heteroatom substitution in
the central core. In a pair of such discogens, 2c and 4c, the
observed effect is pronounced. The substitution of the four CH
groups in 4c with nitrogen atoms in 2c increases the clearing
temperature by over 110 uC. This is consistent with the results
for the triphenylene–hexaazatriphenylene pair in which the
clearing temperature difference is 59 uC.54 This significant
stabilization may be due to stronger quadrupolar interactions
between the cores in the former, and may also result from
A plausible explanation of this apparent anomaly in the
diffraction pattern of 2c is provided by the predicted X-ray
scattering intensity calculated from the square of the Fourier
transform of the molecular charge density.59 Calculations
performed for the most extended conformer of 2c-g,t (Fig. 9)
show intensities of the higher order reflections typical for the
columnar phase: the (110) and (200) signals in the high
amplitude region and the (210) is in the minimum of the struc-
ture factor (Fig. 12). This is observed for 1c but is inconsistent
with the results for 2c. Folding up of the alkyl chain ends
…
fewer H H steric interactions. The latter gives rise to greater
compressibility of the Ph–core dihedral angle, and conse-
quently closer contact of the rigid cores within the column. A
comparison of the core–core correlation peak in the wide-angle
region of the XRD data provides information on the intra-
column packing, and ability of the molecules to adopt the
planar conformation. The core–core separation appears to be
˚
the largest for pyrene derivative 4c (3.9 A), while for
benzo[c]cinnoline 1c the mean distance is the smallest in the
39
series (3.5 A) and typical for discogens. (For tetraazapyrene
˚
2c the 001 peak presumably is below 3.7 A.) The core–core
˚
Fig. 12 Relative calculated scattered intensity for 2c-g,t (a) with most
extended alkyl chains (Fig. 9) and conformer 2c-melted (b) with ends
of alkyl chains ‘‘melted’’ and folded back (Fig. 13). The intensity was
calculated from the square of the Fourier transform of the cylindrically
averaged molecular charge density, including the 1/(sinh sin2h) Lorentz
factor. The vertical bars show the positions of expected reflections for
the Colh phase of 2c.
separation typically shows only a weak temperature depen-
dence so the comparison of the data for pyrene 4c, taken close
to the transition point, and 1c and 2c further away from the
clearing point should still hold. Even the carbazole analog of
4c at 120 uC near the clearing point has a core–core spacing of
1408 | J. Mater. Chem., 2007, 17, 1399–1411
This journal is ß The Royal Society of Chemistry 2007

View Article Online
˚
in 2c-g,t and decreasing the molecular diameter to 29 A,
condensed central core. It is also expected that the permanent
dipole moment in benzo[c]cinnoline derivatives 1 and high
quadrupolar moment in tetraazapyrene 2 further stabilize the
liquid crystalline phases, but these effects appear to be
secondary to geometrical and steric factors. Consequently,
tetraazapyrene derivative 2c has the most stable and
phenanthrene derivative 3c has the least stable mesophases in
the series.
consistent with the unit cell parameter a, results in a reduction
of the mean charge density radius and a concomitant increase
in the angles of characteristic features in the structure factor.
This shifts a minimum in the structure factor to the region near
the (110) and (200) peaks, and increases the amplitude of the
(210) peak. Thus, increasing the density of gauche bonds in the
alkyl chain has the general effect of reducing the mean radius
of the charge density, reducing the intensity of the (110) and
(200) peaks and increasing the amplitude of the (210) peak.
Given the small number of measured diffraction intensities
and the large number of spatial degrees of freedom for the
molecule, we did not attempt to refine molecular parameters to
fit the data. Nonetheless, the observed trend in the calculated
structure factor and molecular shape is consistent with the
experimentally observed intensity of higher order reflections
and the unit cell parameter a.
Calculations demonstrated that high electron deficiency is
achieved by using the azo group as part of the heterocycle
skeleton. Thus, tetraazapyrene discogen 2c, containing two
such groups, has a low-lying LUMO and appears to be
suitable for electron transport. However, the azo moiety shifts
the electronic absorption to the visible range which may be
undesirable for some applications.
Analysis of XRD patterns demonstrated that the intensity of
the higher order reflections may differ significantly from
typical and expected values, and can be related to the
molecular structure and packing in the columnar phase.
Conclusions
A series of compounds 1–4 provided a rare opportunity to
demonstrate the significance of the fill fraction in stabilization
of discotic phases. Analysis suggested that the fill fraction is
affected by the conformational properties of the Ph–core
junction and the distribution and density of the alkyl chains.
The presence of N atoms lowers planarization barriers, the 3,4
substitution pattern allows for the phenyl rings to planarize
without significant alkyl chain–alkyl chain interactions and
necessity of a high density of gauche conformations. In
addition, the higher fill fraction is obtained with a more
Computational details
Quantum-mechanical calculations were carried out with the
B3LYP/6-31G(d) and MP2(fc)/6-31G(d) methods using the
Linda-Gaussian 98 package60 on a Beowulf cluster of 14
processors. Geometry optimizations were undertaken using
appropriate symmetry constraints and default convergence
limits. Vibrational frequencies were calculated with the DFT
method and were used to characterize the nature of the
stationary points and to obtain thermodynamic parameters.
Zero-point energy (ZPE) corrections were scaled by 0.9806.61
The DFT-optimized ground and transition state structures
were used as starting points for MP2 geometry optimizations.
Geometry optimizations were carried out using the Universal
Force Field algorithm supplied with the Cerius2 suite of
programs.
Experimental
The purity of samples was determined by HPLC analysis using
a normal phase column and a gradient hexane–CH2Cl2
mixture. Samples of liquid crystalline materials were prepared
using a pair of virgin microscopic cover glass plates and
observed in polarized light using a PZO Biolar PI microscope
consisting of polarizer, analyzer, and the HS1 Instec hot stage.
Samples were heated and cooled at rates of 5 uC min21 or
10 uC min21. Thermal data were obtained on a TA Instrument
DSC-2920 calibrated with an indium standard. Preliminary
XRD measurements on 1c employed a Scintag X1 Advanced
Diffraction System. Further measurement, carried out at the
University of Pennsylvania, Philadelphia, employed a Bruker-
Nonius FR591 fine-focus generator with a Cu target (l =
˚
1.542 A) mirror-monochromator optics, and a Bruker HiStar
wire detector. 2D diffraction patterns were subsequently
reduced to plots of scattered intensity versus 2h, where h is
related to d-spacing by d = l/2sin(h). In comparison with the
Scintag measurements, data collected with the HiStar detector
had higher angular resolution and lower spectral contamina-
tion but did not extend to as high an angular range.
Fig. 13 Conformer 2c-melted with the ‘‘melted’’ alkyl chain ends of
2c-g,t viewed along the z axis perpendicular to the heterocycle plane
(top) and along the heterocycle plane (bottom). The radius of the
˚
inscribed circle is 29 A.
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 1399–1411 | 1409

View Article Online
M. Wagner, K. Mu¨llen, P. Samori and J. P. Rabe, J. Mater. Chem.,
Acknowledgements
2000, 10, 1519–1525; S. J. Lock, J. W. Goodby, M. Hird and
K. J. Toyne, J. Mater. Chem., 1995, 5, 2175–2182; M. C. Artal,
K. J. Toyne, J. W. Goodby, J. Barbera´ and D. J. Photinos,
J. Mater. Chem., 2001, 11, 2801–2807.
This project was supported by the NSF grants (CHE-9528029
and CHE-0096827). We thank Mr. Xuan Sun and Mr. Andy
Wenger for their technical assistance. Crystallographic work at
the University of Pennsylvania was supported by the MRSEC
program of the National Science Foundation (Grant #
DMR05-20020).
29 H. D. Dakin, Am. Chem. J., 1909, 42, 477–487; T. V. Hansen and
L. Skattebøl, Tetrahedron Lett., 2005, 46, 3357–3358.
30 M. R. Detty and B. J. Murray, J. Am. Chem. Soc., 1983, 105,
883–890.
31 J. H. Birkinshaw and A. Bracken, J. Chem. Soc., 1942, 368–370.
32 G. Pickaert and R. Ziessel, Synthesis, 2004, 2716–2726.
33 G. C. Dol, P. C. J. Kamer and P. W. N. M. van Leeuwen, Eur. J.
Org. Chem., 1998, 359–364.
References
34 I. Dierking, Textures of Liquid Crystals, Wiley-VCH, Weinheim,
2003.
35 V. Vill, LiqCryst 4.6, LCI Gmbh, Hamburg, 2004 and references
therein.
36 P. Kaszynski, S. Pakhomov, K. F. Tesh and V. G. Young, Jr.,
Inorg. Chem., 2001, 40, 6622–6631.
1 In part described in: M. J. Sienkowska, Dissertation, Vanderbilt
University, 2004; J. M. Farrar, Dissertation, Vanderbilt University,
2001.
2 S. Chandrasekhar, in Handbook of Liquid Crystals, ed. D. Demus,
J. Goodby, G. W. Gray, H.-W. Spiess and V. Vill, Wiley-VCH,
New York, 1998, vol. 2B, pp. 749–780.
37 V. de Halleux, J.-P. Calbert, P. Brocorens, J. Cornil, J.-P. Declercq,
J.-L. Bre´das and Y. Geerts, Adv. Funct. Mater., 2004, 14,
649–659.
38 M. J. Sienkowska, H. Monobe, P. Kaszynski and Y. Shimizu,
J. Mater. Chem., 2007, DOI: 10.1039/b612253a.
39 S. K. Prasad, D. S. S. Rao, S. Chandrasekhar and S. Kumar, Mol.
Cryst. Liq. Cryst., 2003, 396, 121–139.
40 The wide-angle data for 1c were collected using the Scintag system
(see ESI{).
41 C. L. Cheng, D. S. N. Murthy and G. L. D. Ritchie, J. Chem. Soc.,
3 S. Chandrasekhar and S. K. Prasad, Contemp. Phys., 1999, 40,
237–245.
4 For a recent overview see: M. O’Neill and S. M. Kelly, Adv.
Mater., 2003, 15, 1135–1146. See also ref. 9.
5 G. Lu¨ssem and J. H. Wendorff, Polym. Adv. Technol., 1998, 9,
443–460.
6 I. H. Stapff, V. Stu¨mpflen, J. H. Wendorff, D. B. Spohn and
D. Mo¨bius, Liq. Cryst., 1997, 23, 613–617.
7 I. Seguy, P. Jolinat, P. Destruel, J. Farenc, R. Mamy, H. Bock, J. Ip
and T. P. Nguyen, J. Appl. Phys., 2001, 89, 5442–5448.
8 L. Schmidt-Mende, A. Fechtenko¨tter, K. Mu¨llen, E. Moons,
R. H. Friend and J. D. MacKenzie, Science, 2001, 293, 1119–1122.
9 N. Boden and B. Movaghar, in Handbook of Liquid Crystals, ed.
D. Demus, J. Goodby, G. W. Gray, H.-W. Spiess and V. Vill,
Wiley-VCH, New York, 1998, vol. 2B, pp. 781–798.
10 S. Kumar, E. J. Wachte1 and E. Keinan, J. Org. Chem., 1993, 58,
3821–3827.
Faraday Trans. 2, 1972, 68, 1679–1690.
42 A. Almenningen, O. Bastiansen, L. Fernholt, B. N. Cyvin,
S. J. Cyvin and S. Samdal, J. Mol. Struct., 1985, 128, 59–76.
43 G.-P. Charbonneau and Y. Delugeard, Acta Crystallogr., Sect. B,
1977, 33, 1586–1588.
44 J. Cioslowski, P. Piskorz, G. Liu and D. Moncrieff, J. Phys. Chem.,
1996, 100, 19333–19335.
45 E. D. Bergmann, M. Rabinovitz, M. J. Aroney, R. J. W. Le Fevre´,
L. Radom and G. L. D. Ritchie, J. Chem. Soc. B, 1968, 1551–1554.
46 H. Allouchi, J.-P. Bideau and M. Cotrait, Acta Crystallogr., Sect.
C, 1992, 48, 1037–1041.
47 H. Allouchi, M. Cotrait, D. Guillon, B. Heinrich and H. T. Nguyen,
Chem. Mater., 1995, 7, 2252–2258.
11 N. Boden, R. C. Borner, R. J. Bushby and J. Clements, J. Am.
Chem. Soc., 1994, 116, 10807–10808.
12 T. Ishi-i, K. Murakami, Y. Imai and S. Mataka, Org. Lett., 2005,
7, 3175–3178; T.-H. Chang, B.-R. Wu, M. Y. Chiang, S.-C. Liao,
C. W. Ong, H.-F. Hsu and S.-Y. Lin, Org. Lett., 2005, 7,
4075–4078.
48 S. H. Eichhorn, A. J. Paraskos, K. Kishikawa and T. M. Swager,
J. Am. Chem. Soc., 2002, 124, 12742–12751.
49 G. Bocelli and M.-F. Grenier-Loustalot, Acta Crystallogr., Sect. C,
1987, 43, 1221–1223.
50 P. B. Rheiner and D. Seebach, Chem.–Eur. J., 1999, 5, 3221–3236.
51 D. N. de Vries Reilingh, R. P. H. Rettschnick and G. J. Hoytink,
J. Chem. Phys., 1971, 54, 2722–2727.
52 P. F. Holt and R. Oakland, J. Chem. Soc., 1964, 6090–6094.
53 A. C. Benniston, A. Harriman, D. J. Lawrie and S. A. Rostron,
Eur. J. Org. Chem., 2004, 2272–2276.
13 K. Pieterse, P. A. van Hal, R. Kleppinger, J. A. J. M. Vekemans,
R. A. J. Janssen and E. W. Meijer, Chem. Mater., 2001, 13,
2675–2679.
14 R. I. Gearba, M. Lehmann, J. Levin, D. A. Ivanov, M. H. J. Koch,
J. Barbera´, M. G. Debije, J. Piris and Y. H. Geerts, Adv. Mater.,
2003, 15, 1614–1618.
15 G. Kestenont, V. de Halleux, M. Lehmann, D. A. Ivanov,
M. Watson and Y. H. Geerts, Chem. Commun., 2001, 2074–2075.
16 C. W. Ong, S.-C. Liao, T. H. Chang and H.-F. Hsu, J. Org. Chem.,
2004, 69, 3181–3185.
17 M. Lehmann, G. Kestemont, R. G. Aspe, C. Buess-Herman,
M. H. J. Koch, M. G. Debije, J. Piris, M. P. de Haas, J. M.
Warman, M. D. Watson, V. Lemaur, J. Cornil, Y. H. Geerts,
R. Gearba and D. A. Ivanov, Chem.–Eur. J., 2005, 11, 3349–3362.
18 P. Kaszynski, Molecules, 2004, 9, 716–724.
19 P. Kaszynski, in Magnetic Properties of Organic Materials, ed.
P. M. Lahti, Marcel Dekker, Inc., New York, 1999, pp. 305–324.
20 N. Miyaura, T. Yanagi and A. Suzuki, Synth. Commun., 1981, 11,
513–519.
54 N. Boden, R. J. Bushby, G. Headdock, O. R. Lozman and
A. Wood, Liq. Cryst., 2001, 28, 139–144.
55 S. H. J. Idziak, N. C. Maliszewskyj, G. B. M. Vaughan,
P. A. Heiney, C. Mertesdorf, H. Ringsdorf, J. P. McCauley and
A. B. Smith, III, J. Chem. Soc., Chem. Commun., 1992, 98–99;
J. Malthete, A.-M. Levelut and J.-M. Lehn, J. Chem. Soc., Chem.
Commun., 1992, 1434–1436.
56 C. Nuckolls and T. J. Katz, J. Am. Chem. Soc., 1998, 120,
9541–9544.
21 H. Stetter and M. Schwarz, Chem. Ber., 1957, 90, 1349–1351.
22 J. F. Corbett and P. F. Holt, J. Chem. Soc., 1961, 3695–3699.
23 A. E. Blood and C. R. Noller, J. Org. Chem., 1957, 22, 711–712.
24 J. M. Farrar, M. Sienkowska and P. Kaszynski, Synth. Commun.,
2000, 30, 4039–4045.
57 M. T. Allen, S. Diele, K. D. M. Harris, T. Hegmann, B. M. Kariuki,
D. Lose, J. A. Preece and C. Tschierske, J. Mater. Chem., 2001, 11,
302–311.
58 S. Kumar, J. J. Naidu and D. S. S. Rao, J. Mater. Chem., 2002, 12,
1335–1341.
25 K. Ogino, S. Iwashima, H. Inokuchi and Y. Harada, Bull. Chem.
Soc. Jpn., 1965, 38, 473–477.
59 J. Als-Nielsen and D. McMorrow, Elements of Modern X-ray
Physics, Wiley, New York, 2001.
26 J. T. Manka, F. Guo, J. Huang, H. Yin, J. M. Farrar,
M. Sienkowska, V. Benin and P. Kaszynski, J. Org. Chem.,
2003, 68, 9574–9588.
27 G. W. Gray, M. Hird, D. Lacey and K. J. Toyne, J. Chem. Soc.,
Perkin Trans. 2, 1989, 2041–2053.
60 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, Jr.,
R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam,
A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi,
V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli,
C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala,
Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck,
28 4-Octyloxyphenylboronic acid and 3,4-dioctyloxyphenylboronic
acid are known: T. Yatabe, M. A. Harbison, J. D. Brand,
1410 | J. Mater. Chem., 2007, 17, 1399–1411
This journal is ß The Royal Society of Chemistry 2007

View Article Online
K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz,
A. G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz,
I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith,
M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez,
M. Challacombe, P. M. W. Gill, B. G. Johnson, W. Chen,
M. W. Wong, J. L. Andres, M. Head-Gordon, E. S. Replogle and
J. A. Pople, GAUSSIAN 98 (Revision A.9), Gaussian, Inc.,
Pittsburgh, PA, 1998.
61 A. P. Scott and L. Radom, J. Phys. Chem., 1996, 100,
16502–16513.
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 1399–1411 | 1411