Photoinduced Alignment of Ferroelectric Liquid Crystals Using Azobenzene Polymer Networks of Polyethers and Polyepoxides

Article abstract of DOI:10.1021/ma034885l

Several divinyl ether and diepoxide monomers bearing an azobenzene moiety were synthesized and used to investigate the photoalignment of a ferroelectric liquid crystal (FLC) with no use of surface orientation layers. Azobenzene polymer networks, obtained by cationic polymerization of the monomers dissolved in the FLC host and exposed to linearly polarized irradiation, were found to be able to induce and stabilize a bulk alignment of FLC. Photoinduced reorientation of the FLC can also be achieved by changing the polarization of irradiation light, but in contrast with the use of chiral azobenzene poly(meth)acrylates,¹ the results suggest a mechanism based on the formation of an anisotropic azobenzene polyether or polyepoxide network for the commanding effect on the FLC alignment.

Full text of DOI:10.1021/ma034885l

Macromolecules 2003, 36, 9033-9041
9033
Photoinduced Alignment of Ferroelectric Liquid Crystals Using
Azobenzene Polymer Networks of Polyethers and Polyepoxides
Son ia Se´vign y,† Lu c Bou ch a r d ,‡ Sh a h r ok h Mota llebi,‡ a n d Yu e Zh a o*,†
De´partement de chimie, Universite´ de Sherbrooke, Sherbrooke, Que´bec, Canada J 1K 2R1, and
St-J ean Photochimie, 725 Trotter, St-J ean-sur-Richelieu, Que´bec, Canada J 3B 8J 8
Received J une 27, 2003; Revised Manuscript Received September 19, 2003
ABSTRACT: Several divinyl ether and diepoxide monomers bearing an azobenzene moiety were
synthesized and used to investigate the photoalignment of a ferroelectric liquid crystal (FLC) with no
use of surface orientation layers. Azobenzene polymer networks, obtained by cationic polymerization of
the monomers dissolved in the FLC host and exposed to linearly polarized irradiation, were found to be
able to induce and stabilize a bulk alignment of FLC. Photoinduced reorientation of the FLC can also be
achieved by changing the polarization of irradiation light, but in contrast with the use of chiral azobenzene
poly(meth)acrylates,1 the results suggest a mechanism based on the formation of an anisotropic azobenzene
polyether or polyepoxide network for the commanding effect on the FLC alignment.
In tr od u ction
photoalignment of FLC could also be achieved by using
azobenzene polyether or polyepoxide networks, but the
mechanism of action may be different from the chiral
polyacrylates or polymethacrylates.1
In polymer-stabilized liquid crystals (PSLC), a poly-
mer network of low concentration exerts the effect of
stabilizing the LC alignment or textures induced by
surface orientation layers or an electric field.2,3 In the
case of ferroelectric liquid crystals (FLC), the use of a
polymer network is also aimed at improving the shock
resistance of the FLC cell whose bulk alignment is
induced by rubbed surfaces.4 Alignment has always
been an important issue for LC in general and for FLC
technologies in particular. Development of novel meth-
odology, different from the established surface-stabilized
FLC,5 is of interest from both fundamental and applied
points of view.6-9
Exp er im en ta l Section
1. Ma ter ia ls. Mercury(II) acetate (98+%), aminobenzoni-
trile (98%), isoamyl nitrite (97%), resorcinol (98%), and bro-
mohexene (95%) were used as received from Aldrich. 3-Chlo-
roperoxybenzoic acid (MCPBA, Alfa Aesar, 70-77%), diphenyl-
iodonium hexafluorophosphate (TCI America, 97%), potassium
carbonate (BDH, 99%), potassium iodide (BDH, 99%), acetic
acid (Fisher, 99%), and hydrochloric acid (EMD, 36-38%) were
used as received. Prior to use, 2,2′-azobis(isobutyronitrile)
(AIBN, Polysciences) was recrystallized from ethanol, 6-chlo-
rohexanol (Aldrich, 96%) mixed with dichloromethane was
neutralized by washing with 10% NaHCO3, and ethyl vinyl
ether (Aldrich, 99%) was distilled to remove the peroxides (bp
35-36 °C). 4-(4-Hydroxyphenylazo)benzoic acid was prepared
using the method reported elsewhere.1 All solvents were
commercially available and used as received.
The ferroelectric liquid crystal (FLC) mixture used was
Felix-015 (Clariant) that has the following phase transitions:
Cr -12 °C SC* 72 °C SA 83 °C N* 85 °C Iso. Its spontaneous
polarization and rotational viscosity measured at 25 °C are
+33 nC/cm2 and 80 mPa, respectively (date provided by
Clariant). As compared with the FLC used in the study with
methacrylate- and acrylate-based chiral azobenzene polymers,1
this FLC has a very narrow temperature range for the N*
phase and a strong spontaneous polarization. In the case of
surface-stabilized FLC, the absence of an N* phase is known
to make the alignment more difficult to achieve. The choice of
a different commercially available FLC aimed at investigating
the general validity of this optical alignment approach for FLC.
2. Syn th esis of Azoben zen e Mon om er s. Two divinyl
ethers and two diepoxides were synthesized. Their chemical
structures are shown in Figure 1. Basically, ether1 is similar
to epoxide1 in structure, differing only in the reactive group
and the number of methylene units in the spacer. The same
can be noted between ether2 and epoxide2, which have the
two reactive groups linked to the same aromatic ring. Their
synthetic schemes and details are given below.
In a previous study,1 we reported on and discussed
the use of some chiral azobenzene-containing polyacry-
late and polymethacrylate networks to optically align
FLC with no need for surface orientation layers. The
photoisomerization and the related photoalignment of
azobenzene moieties under linearly polarized irradia-
tion10 are the basis of this optical and rubbing-free
approach.11 This method is different from the photo-
alignment of LC by coating the substrate surfaces with
photochromic molecules such as azobenzene polymers,
i.e., the so-called commanding surfaces developed by
Ichimura and co-workers.12 The azobenzene polymer
network, which is dispersed in the FLC host, provides
a volume effect on the latter, even though the azoben-
zene moieties on the surface of the network may play a
central role in some systems.1
All the azobenzene polymer networks studied so far
were either polyacrylates or polymethacrylates prepared
from free radical polymerization.1,11 In pursuit of our
efforts aimed at understanding the photoalignment
mechanisms and fully assessing the effectiveness of this
optical approach, we have synthesized a series of new
azobenzene divinyl ether and diepoxide monomers and
investigated the use of their polymer networks, pre-
pared through cationic polymerization, to optically align
FLC. The results reported in this paper show that the
2.1. P r ep a r a tion of Eth er 1. (6-Ch lor oh exyloxy)eth en e
(2): Ethyl vinyl ether (249.24 g, 3.45 mol) was reacted with
neutralized 6-chorohexanol (79.43 g, 0.58 mol) and mercury
acetate (3.46 g, 0.01 mol). The reaction was performed at reflux
for 6 h. Then 10 g (0.07 mol) of anhydrous potassium carbonate
was added to the solution. The product was first separated
† Universite´ de Sherbrooke.
‡ St-J ean Photochimie.
* Corresponding author: e-mail yue.zhao@usherbrooke.ca.
10.1021/ma034885l CCC: $25.00 © 2003 American Chemical Society
Published on Web 11/07/2003

9034 Se´vigny et al.
Macromolecules, Vol. 36, No. 24, 2003
F igu r e 1. Chemical structures of azobenzene divinyl ether and diepoxide monomers.
Sch em e 1. Syn th etic Rou te to 4-(4-Vin yloxym eth oxyp h en yla zo)ben zoic Acid 6-Vin yloxyh exyl Ester (Eth er 1)
from remaining ethyl vinyl ether through distillation under
reduced pressure (bp 85-87 °C, 8 mmHg) and then puri-
fied by silica gel column chromatography using AcOEt/hexane
(1:9) as eluent. A yellow oily liquid was obtained; yield 30%
(22.3 g, 0.14 mol); mp -21 °C. MS (m/e): 162 (M+). λmax
(THF): 256 nm. 1H NMR (δ, acetone): 1.30-1.41 (4H, m,
Cl-CH2-CH2-CH2-CH2), 1.53-1.62 (2H, m, O-CH2-CH2),
1.65-1.74 (2H, m, Cl-CH2-CH2), 3.59-3.65 (2H, m, Cl-CH2),
3.92-3.94 (2H, m, O-CH2), 4.12-4.17 (4H, m, O-CHdCH2),
6.44-6.51 (1H, m, O-CH).
reaction lasted 14 h at room temperature. To collect the
product, the reaction mixture was added to a mixture of acetic
acid (3.75 g) and water (500 g), the solution was filtered, and
the product, a red powder, was rinsed with water and dried.
Yield: 81% (16.4 g, 68.6 mmol); mp: decomposed at 210 °C
1
prior to melting. MS (m/e): 239 (M+). λmax(THF): 400 nm. H
NMR (δ, acetone): 6.26 (1H, s, aromatic H ortho to both
C-OH), 6.49-6.52 (1H, m, aromatic H ortho to C-OH and
para to the other), 7.62-7.64 (1H, m, aromatic H meta to
C-OH), 7.77-7.80 (2H, m, aromatic H ortho to C-CN), 7.85-
7.88 (2H, m, aromatic H meta to C-CN), 9.60 (2H, OH).
4-[2,4-Bis(6-vin yloxyh e xyloxy)p h e n yla zo]b e n zon i-
tr ile (Eth er 2): The reaction procedure was similar to ether1,
except that 4-(2,4-dihydroxyphenylazo)benzonitrile (3), instead
of 4-(4-hydroxyphenylazo)benzoic acid (1), was used to react
with 2. For purification, the monomer, a red powder, was
washed several times with 2-propanol. Yield: 50%; mp: 52°C.
4-(4-Vin yloxym eth oxyp h en yla zo)ben zoic Acid 6-Vin -
yloxyh exyl Ester (Eth er 1): 4-(4-Hydroxyphenylazo)benzoic
acid (370 mg, 1.54 mmol) (1) and (6-chlorohexyloxy)ethene (750
mg, 4.63 mmol) (2) were added to a mixture of 5 mL of DMSO
and 25 mL of DMF containing K2CO3 (640 mg, 4.63 mmol)
and KI (770 mg, 4.63 mmol). The reaction was performed at
80 °C for 14 h. The solution was cooled, and chloroform was
added prior to extraction of the product with water. The
organic phase was collected and dried over anhydrous mag-
nesium sulfate. After removing the solvent, an orange powder
was obtained with a yield of 34% (150 mg, 0.3 mmol); mp 77
1
MS (m/e): 438 (M+). H NMR (δ, DMSO): 1.34-1.49 (8H, m,
C-O-CH2-CH2-CH2-CH2), 1.55-1.63 (4H, m, CH2dCH-
O-CH2-CH2), 1.72-1.82 (4H, m, C-O-CH2-CH2), 3.60-3.67
(4H, m, C-O-CH2), 3.82-3.95 (4H, m, CH2dCH-O-CH2),
4.06-4.21 (4H, m, CH2dCH-O), 6.41-6.48 (2H, m, CH2dCH-
O), 6.75-6.76 (1H, d, J ) 2.3 Hz, aromatic H ortho to both
C-O-CH2), 6.96-6.99 (1H, d, J ) 8.4 Hz, aromatic H ortho
to one C-O and para to the other), 7.64-7.67 (1H, d, J ) 9.1
Hz, aromatic H meta to C-O), 7.86-7.89 (2H, d, J ) 8.5 Hz,
aromatic H ortho to CN), 7.99-8.02 (2H, d, J ) 8.5 Hz,
aromatic H meta to CN).
1
°C. MS (m/e): 494 (M+). H NMR (δ, DMSO): 1.44-1.56 (8H,
m, C(O)-O-CH2-CH2-CH2-CH2 and C-O-CH2-CH2-CH2-
CH2), 1.65-1.86 (8H, m, C(O)-O-CH2-CH2-CH2-CH2-CH2
and C-O-CH2-CH2-CH2-CH2-CH2), 3.58-3.70 (4H, m,
CH-O-CH2), 3.87-3.92 (2H, m, C-O-CH2), 4.07-4.14 (4H,
m, O-CHdCH2), 4.26-4.35 (2H, t, J ) 6.6 Hz, C(O)-O-CH2),
6.42-6.49 (2H, m, O-CHdCH2), 7.06-7.08 (2H, m, aromatic
H ortho to C-O), 7.90-7.95 (4H, m, aromatic H ortho to
C-NdN), 8.14-8.17 (2H,m, aromatic H ortho to C(O)-O).
2.2. P r ep a r a tion of Eth er . 2. 4-(2,4-Dih yd r oxyp h en yl-
a zo)ben zon itr ile (3): Aminobenzonitrile (10.0 g, 84.8 mmol)
was dissolved in ethanol (25.0 g) cooled in an ice-acetone bath,
to which a cold solution of HCl (10.02 g, 101.7 mmol) was
added drop-by-drop to keep the temperature under 5 °C,
followed by addition of isoamyl nitrite (10.42 g, 88.98 mmol).
Afterward, a solution of resorcinol (9.33 g, 84.8 mmol) dissolved
in ethanol (40.0 g) was mixed with the first solution, and the
2.3. P r ep a r a tion of Ep oxid e1 a n d Ep oxid e. 2. 2-(4-
Br om obu tyl)oxir a n e (4): Bromohexene (50 mg, 0.3 mmol)
was reacted with 3-chloroperoxybenzoic acid (70-77% pure,
165 mg, 0.71 mmol) in dichloromethane (5 mL) at reflux for 3
h. Then dichloromethane (30 mL) was added prior to washing
with aqueous solution of K2CO3 (1 N) and then with water.
The organic phase was collected and dried over anhydrous
magnesium sulfate. The product obtained was a yellow oily
liquid. Yield: 95% (52 mg, 0.28 mmol). Mp: -22 °C. MS (m/e):
1
179 (M+). λmax(THF): 240 and 290 nm. H NMR (δ, acetone):

Macromolecules, Vol. 36, No. 24, 2003
Alignment of Ferroelectric Liquid Crystals 9035
Sch em e 2. Syn th etic Rou te to 4-[2,4-Bis(6-vin yloxyh exyloxy)p h en yla zo]ben zon itr ile (Eth er 2)
Sch em e 3. Syn th etic Rou te to 4-[4-(4-Oxir a n ylbu toxy)p h en yla zo]ben zoic Acid 4-Oxir a n ylbu tyl Ester (Ep oxid e1)
a n d 4-[2,4-Bis(4-oxir a n ylbu toxy)p h en yla zo]ben zon itr ile (Ep oxid e2)
1.77-1.85 (2H, m, Br-CH2-CH2-CH2), 1.90-2.03 (2H, m,
Br-CH2-CH2-CH2-CH2), 2.22-2.28 (2H, m, Br-CH2-CH2),
2.96-3.01 (2H, m, O-CH2 from epoxide), 3.12-3.22 (1H, m,
O-CH from epoxide), 3.83-3.91 (2H, m, Br-CH2).
tion temperature through AIBN radicals acting as electron
donors to the iodonium salt.13 The typical procedure is
described in the following example.
In a small flask, ether1 (2.5 mg, 0.005 mmol), FLC (47.5
mg), AIBN (0.5 mg), and DiPhIF6P (0.5 mg) were dissolved in
acetone (about 0.1 mL) at room temperature to obtain a
homogeneous solution. The solvent was allowed to evaporate
at room temperature first under the atmosphere pressure for
3-5 h and then in a vacuum oven for 1 h. A drop of the freshly
dried mixture was then placed between two quartz plates,
warmed to 50 °C, and finally compressed manually to produce
a film of about 60 mm2 large and with a thickness around 5
µm. Afterward, the sample was moved into a microscope hot
stage and, without delay, exposed to linearly polarized light
(normal incidence) while being heated, at a rate of about 20
°C/min, to 110 °C for 10 min. Finally, under irradiation the
mixture was cooled, at a rate of about 5 °C/min, to 83 °C for
10 min before being cooled to room temperature. After the
preparation procedure was completed, the irradiation light was
turned off. All samples prepared under these conditions were
first examined on a polarizing optical microscope to observe
the photoalignment of FLC and then used in different experi-
ments as shown in the paper.
The concentration of azobenzene monomer in the FLC was
between 1 and 10 wt %. Linearly polarized irradiation light
was obtained by using an UV-vis curing system (Novacure)
combined with an UV or visible interference filter (10 nm
bandwidth, Oriel) and an UV linear dichroic polarizer (Oriel).
Both polarized UV irradiation at 360 nm (∼10 mW/cm2) and
polarized visible exposure at 400 nm (∼5 mW/cm2) were
produced. In all experiments, samples with ether1 and ep-
oxide1 were irradiated at 360 nm while those with ether2 and
epoxide2 were irradiated at 400 nm.
4-[4-(4-Oxir a n ylbu toxy)p h en yla zo]ben zoic Acid 4-Ox-
ir a n ylbu tyl Ester (Ep oxid e1): The reaction procedure was
similar to ether1, except that 2-(4-bromobutyl)oxirane (4),
instead of (6-chlorohexyloxy)ethene (2), was reacted with 1.
The monomer, an orange powder, was recrystallized in metha-
nol and dried under vacuum. Yield: 30%; mp: 62 °C. MS
(m/e): 491 (M+). 1H NMR (δ, DMSO): 1.46-1.59 (8H, m, C(O)-
O-CH2-CH2-CH2-CH2 and C-O-CH2-CH2-CH2-CH2)
1.73-1.81 (4H, m, C(O)-O-CH2-CH2 and C-O-CH2-CH2)
2.43-2.50 (4H, m, O-CH2 from epoxide), 2.65-2.68 (2H, m,
O-CH from epoxide), 4.08-4.12 (2H, t, J ) 6.4 Hz, C-O-
CH2), 4.23-4.33 (2H, t, J ) 6.4 Hz, C(O)-O-CH2), 7.13-7.16
(2H, d, J ) 7.0 Hz, aromatic H ortho to C-O), 7.88-7.96 (4H,
m, aromatic H ortho to NdN), 8.11-8.14 (2H, d, J ) 8.5 Hz,
aromatic H meta to NdN).
4-[2,4-Bis(4-oxir a n ylb u t oxy)p h en yla zo]b en zon it r ile
(Ep oxid e2): A similar reaction procedure was employed as
for epoxide1, except that 4-(2,4-dihydroxyphenylazo)benzoni-
trile (3), instead of 1, was reacted with 2-(4-bromobutyl)oxirane
(4). The monomer was purified by washing with hexane. A red
powder was obtained with a yield of 28%; mp 66 °C. MS (m/e):
435 (M+). 1H NMR (δ, DMSO): 1.44-1.63 (8H, m, C-O-CH2-
CH2-CH2-CH2), 1.75-1.86 (4H, m, C-O-CH2-CH2), 2.41-
2.50 (4H, m, C-O-CH2), 2.64-2.70 (2H, m, O-CH from
epoxide), 4.08-4.23 (4H, m, C-O-CH2), 6.61-6.65 (1H, m,
aromatic H ortho to both C-O), 6.78 (1H, s, aromatic H ortho
to one C-O and para to the other), 7.65-7.68 (1H, d, J ) 2.7
Hz, aromatic H meta to CN), 7.84-7.89 (2H, d, J ) 8.5 Hz,
aromatic H ortho to CN), 7.99-8.02 (2H, d, J ) 8.3 Hz,
aromatic H meta to CN).
4. Ch a r a cter iza tion s. Mon om er s: The four azobenzene
monomers and the compounds used for their synthesis were
3. P r ep a r a tion of Op tica lly Align ed F LC. Photoaligned
FLC was prepared by polymerizing azobenzene monomers
dissolved in the FLC host while exposing the mixture to
linearly polarized UV or visible light. Unlike the free radical
polymerization used for acrylate and methacrylate monomers,1
cationic polymerization was required for divinyl ether and
diepoxide monomers.13,14 The thermal initiator used in this
study was composed of 50/50 (w/w) of diphenyliodonium
hexafluorophosphate (DiPhIF6P) and AIBN, and its concentra-
tion was 2 wt % with respect to the total amount of the
mixture. AIBN was used to diminish the cationic polymeriza-
1
characterized using several techniques. Their H NMR spectra,
in either acetone or DMSO, were recorded on a Bruker-AC300
(300 MHz) or a Bruker DMX-600 (600 MHz) spectrometer.
Mass spectra were recorded on a gas chromatography-mass
spectrometer (GC-MS), an Agilent GC6890 system coupled to
a detector Agilent MSD5973 with a HP-5 column using helium
as carrier gas. The thermal behavior was investigated by
means of a differential scanning calorimeter (Perkin-Elmer
DSC-7), using indium as the calibration standard and a
heating or cooling rate of 10 °C/min. All compounds were found

9036 Se´vigny et al.
Macromolecules, Vol. 36, No. 24, 2003
F igu r e 2. UV-vis spectra of the monomers in tetrahydrofu-
ran solution.
to display only a crystalline phase before melting. Their
reported melting temperature was determined as the maxi-
mum of the endothermic peak during the second heating scan.
UV-vis spectra in THF solution were recorded on a Hewlett-
Packard 8452A diode array spectrophotometer.
Mixtu r es of Mon om er /F LC a n d P olym er ized Sa m p les:
The cationic polymerization of azobenzene monomers dissolved
in the FLC was examined by means of DSC. Typically, a
freshly dried mixture (about 15 mg) was inserted into a DSC
pan, and the first heating scan was recorded up to 200 °C to
complete the polymerization. Then the sample was cooled to
room temperature, and a second heating scan was conducted
to observe the changes. A polarizing optical microscope (Leitz
DMR-P, magnification 100×) equipped with an Instec hot
stage was used to examine the photoalignment of FLC after
the irradiation light was turned off at room temperature (see
the conditions described above for the preparation of optically
aligned samples). Optical micrographs were taken using a
Leica DC-300 digital camera. The photoisomerization of
azobenzene monomers (before polymerization) and polymers
(after polymerization) in the FLC was also investigated using
the UV-vis spectrophotometer. UV-vis spectra of thin films
(about 5 µm thick) cast between two quartz plates were
recorded immediately following the UV or visible irradiation.
For polarized UV-vis spectra, a polarizer (Oriel) was placed
in front of the optically aligned sample. Scanning electron
microscopy (SEM) was used to observe the azobenzene polymer
networks. For these experiments, samples between quartz
plates were dipped in hexane to slowly remove the FLC; the
two plates were then carefully separated, and the polymer
network remaining on both substrates was dried before being
examined on a scanning electron microscope (J EOL J SC-
840A). Finally, photoaligned samples for the electrooptic effects
measurements were prepared between glass plates coated with
conducting indium-tin oxide (ITO) using the procedure de-
scribed above. The measurements were conducted using a
photodetector (Displaytech) mounted on the optical microscope
and collected to a digital oscilloscope (Tektronix, TDS 210),
and a high-voltage waveform generator (WFG500, FLC Elec-
tronics) was used to apply the ac field.
F igu r e 3. UV-vis spectra of 5% azobenzene monomers in the
ferroelectric liquid crystal host before and after irradiation:
(a) 4-(4-vinyloxymethoxyphenylazo)benzoic acid 6-vinyloxy-
hexyl ester (ether1) and (b) 4-[2,4-bis(4-oxiranylbutoxy)phen-
ylazo]benzonitrile (epoxide2).
moiety. Reversible trans-cis photoisomerization was
observed in solution for all monomers. In the FLC, the
situation is a little bit complicated for ether1 and
epoxide1 due to the overlapping of the absorption band
of the FLC. Figure 3a shows an example with a mixture
containing 5% ether1. The absorption maximum of
azobenzene moieties is noticeable as a shoulder over the
strong absorption of the FLC host. Nevertheless, the
trans-to-cis photoisomerization takes place in the mix-
ture, since the absorption around 450 nm, which is
mostly contributed by the n-π* transition of cis isomer,
increases on UV irradiation. By contrast, as is noted in
Figure 3b, the photoisomerization of azobenzene in
epoxide2, likewise for ether2, can easily be seen since
their absorption peak is well separated from that of
FLC. Table 1 collects the absorption maxima of all
monomers in THF and in the FLC before and after
polymerization as well as their melting temperature. It
is noted that the absorption of all monomers shift to
higher wavelengths in the FLC and that a further red
shift seems to occur after polymerization for ether1 and
epoxide1. Without ruling out possible effects from light
scattering and interaction of the chromophore with the
FLC, the increase in absorption wavelength is probably
caused by aggregation of azobenzene moieties in the
FLC, considering the high azobenzene monomer con-
centrations used.15
The occurrence of cationic polymerization of all mono-
mers in the FLC, with the used DiPhIF6P/AIBN
(50/50) initiator, was confirmed by DSC measurements.
Figure 4 shows the heating curves (first scan, 10
°C/min) of the pure FLC and the mixtures with 10% of
azobenzene monomers. Several observations can be
made. First, polymerization of all monomers is revealed
by the exotherms around 115 °C, and for the two ethers,
a second exotherm appears at higher temperature. The
diminution of polymerization temperature by AIBN is
clear because with pure DiPhIF6P, the polymerization
exotherm was observed at about 200 °C. Second, prior
Resu lts a n d Discu ssion
1. Mon om er s, P olym er s, a n d P h otoisom er iza -
tion . Before studying the photoalignment of FLC, the
photochromic properties of the azobenzene monomers
and their cationic polymerization in the FLC host were
investigated. Shown in Figure 2 are the UV-vis spectra
of the monomers in THF solution. The absorptions of
trans-azobenzene (π-π* transition) of ether2 and ep-
oxide2 appear at higher wavelengths than ether1 and
epoxide1 because of stronger electron donor, electron
acceptor, and the 2,4,4′ substitution on the azobenzene

Macromolecules, Vol. 36, No. 24, 2003
Alignment of Ferroelectric Liquid Crystals 9037
Ta ble 1. Ch a r a cter istics of th e Azoben zen e Mon om er s
a
b
monomer
λTHF (nm)
λFLC (nm)
λFLC (nm)
Tm (°C)
4-(4-vinyloxymethoxyphenylazo)benzoic acid 6-vinyloxyhexyl ester (ether1)
4-[2,4-bis(6-vinyloxyhexyloxy)phenylazo]benzonitrile (ether2)
4-[4-(4-oxiranylbutoxy)phenylazo]benzoic acid 4-oxiranylbutyl ester (epoxide1)
4-[2,4-bis(4-oxiranylbutoxy)phenylazo]benzonitrile (epoxide2)
363
391
362
390
∼378
∼386
77
52
62
66
399
399
∼374
398
∼385
398
a
b
Before polymerization. After polymerization.
contrast, the same polymerization process under ir-
radiation leads to a bulk alignment of smectic domains
in the direction perpendicular to the polarization of
irradiation light (Figure 5c). Again, the action of azoben-
zene moieties is clearly at the origin of the photoalign-
ment of FLC. Among the four monomers, and with the
FLC used, ether1 and epoxide1 were found to be most
effective in inducing the photoalignment. The typical
results of photoalignment of FLC achieved using various
concentrations of ether1 and epoxide1 are presented in
Figures 6 and 7, respectively. In both cases, good bulk
alignment was obtained with 1-5% azobenzene net-
work, but the number of defects increases with the
polymer concentration. The quality of photoalignment
with 1% epoxide1 is actually close to that obtained by
surface orientation layers of a 5 µm cell. We also used
0.5% ether1 and epoxide1, but this concentration seemed
to be too low to sustain a bulk alignment of FLC; only
partial alignment was observed. No bulk alignment
could be induced with 10% exther1 or epoxide1 because
of phase separation, which is consistent with the result
obtained with chiral azobenzene polyacrylates and
polymethacrylates.1 On the other hand, ether2 and
epoxide2 were less efficient for photoalignment of FLC
than ether1 and epoxide1, with the alignment observed
only at low monomer concentrations (1-3%). This
difficulty may be related to their molecular structures
(Figure 1), as both have the two reactive groups linked
to the same phenyl unit which may sterically hinder
the movement of the azobenzene moiety.
F igu r e 4. DSC curves (first heating scan, 10 °C/min) for the
pure ferroelectric liquid crystal and its mixtures with 10%
azobenzene monomers.
to polymerization of the monomer, the shift to lower
temperatures for the phase transitions of the FLC, as
compared with the pure FLC, reflects the interaction
between the two components. It is apparent that the
disruption effect of the two ethers was greater than the
two epoxides. Third, after polymerization (the second-
scan curves are not shown), interactions between the
azobenzene polymers and the FLC can still be noticed
from changes in the transition peaks. On the basis of
the above results, in the preparation of optically aligned
FLC, the mixture exposed to linearly polarized irradia-
tion was heated to 110 °C, in the isotropic phase, for
10-15 min to complete the polymerization process. We
have no explanation for the second exothermic peak
observed for the two ethers in Figure 4, but what
matters for the present study is that under the used
experimental conditions subsequent DSC scans showed
no reaction exotherm, implying the depletion of the
monomer during the polymerization.
2. P h otoa lign m en t of F LC. Similar to acrylate- and
methacrylate-based polymers,1,11 azobenzene-containing
polyether and polyepoxide networks are able to induce
photoalignment of FLC without the use of surface
orientation layers. Polarizing optical micrographs are
presented in Figure 5 for the mixture with 3% ether2.
Prior to heating, the film prepared by compressing a
freshly prepared homogeneous mixture between two
quartz plates appears birefringent, displaying no phase
separation or developed smectic texture (Figure 5a). In
the absence of irradiation, polymerization in the isotro-
pic phase followed by cooling to room temperature
results in no bulk alignment of smectic domains, but
the smectic texture is now visible (Figure 5b). By
Similar to chiral diacrylate or dimethacrylate mono-
mers,1 using azobenzene divinyl ethers and diepoxides,
no enhanced phase separation of the polymer network
from FLC was observed in the aligned SC* phase.
Moreover, exposing an optically aligned sample to
irradiation light with its polarization changed by 90°
could also rotate the bulk alignment of FLC by 90°. A
marked difference, however, is that the orientational
memory effect was observed with azobenzene polyethers
and polyepoxides, contrary to chiral polyacrylates or
polymethacrylates.1 That is, after isotropization of an
aligned sample in the isotropic phase, the bulk align-
ment of FLC can be recovered on cooling to the LC
phases in the absence of irradiation. Such orientational
memory effect is the signature of the formation of an
anisotropic polymer network.2,3,16 Unfortunately, poly-
ether and polyepoxide networks were very difficult to
observe by a scanning electron microscope (SEM) after
extraction of the FLC in hexane; most of the polymer
seemed to be washed out during the process. This
suggests that these polymer networks may be extremely
fine and well dispersed in the FLC host, which is indeed
supported by the observation that, unlike samples with
chiral polyacrylates or polymethacrylates,1 no azoben-
zene polyether or polyepoxide networks were discernible
on bright field optical microscope. The networks found
on the substrates after removal of FLC do display a very
different morphology, as is seen from the SEM pictures
in Figure 8. Instead of polymer threads or beads,1,16 the

9038 Se´vigny et al.
Macromolecules, Vol. 36, No. 24, 2003
F igu r e 5. Polarizing optical micrographs for a film of ferroelectric liquid crystal with 3% 4-[2,4-bis(6-vinyloxyhexyloxy)phenylazo]-
benzonitrile (ether2): (a) before polymerization and irradiation, (b) after polymerization without irradiation, and (c) after
polymerization under linearly polarized irradiation. Picture area: 780 µm × 625 µm.
F igu r e 8. Scanning electron microscope (SEM) images of
azobenzene polyether and polyepoxide networks remained on
the substrate after extraction of the ferroelectric liquid crystal
in hexane. The scale bar is 1 µm.
a similar chemical structure to ether1 and epoxide1 was
found to be unable to induce the photoalignment of FLC.
Some factors may be responsible for the different
behavior. First, the FLC hosts used in the two studies
are different for the reason explained earlier. The differ-
F igu r e 6. Photoalignment of ferroelectric liquid crystal
achieved using various concentrations of 4-(4-vinyloxymethox-
yphenylazo)benzoic acid 6-vinyloxyhexyl ester (ether1). Picture
area: 780 µm × 625 µm.
ent FLC with a different monomer may well develop
different intermolecular interactions that affect the
interaction and compatibility between the azobenzene
polymer and FLC. Second, the diacrylate and dimethacry-
late monomers were polymerized through free radical
polymerization,1 while cationic polymerization was used
for the ether and epoxide monomers. The rate of chain
growth through ionic reactive species is generally faster
than that via free radicals.17 This difference, together
with the effect of different polymer and FLC interac-
tions, would result in different polymer network mor-
phologies, which is indeed the case as revealed by SEM.
Third, ether1 and epoxide1 have longer flexible spacers
than the dimethacrylate monomer whose short spacer
also contains the chiral center. This structural difference
may provide the azobenzene moiety in ethre1 and
epoxide1 with greater freedom of movement, which is
important for photoinduced orientation.
Now we discuss the commanding effect observed with
networks of polyethers and polyepoxides. Even though
the orientational memory effect suggests the formation
of an anisotropic azobenzene polymer network in the
FLC, the photoinduced reorientation of FLC upon
subsequent irradiation must be the consequence of a
reorientation of azobenzene moieties on the polymer
network. More experiments on photoinduced reorienta-
tion of FLC were performed in order to get clues about
the commanding effect provided by azobenzene moieties
or an anisotropic network. A couple of examples are
discussed below. In the experiment shown in Figure 9,
part of an optically aligned sample with 5% ether1 was
subjected to a second irradiation whose polarization was
rotated by 90° (the other part was covered by an
F igu r e 7. Photoalignment of ferroelectric liquid crystal
achieved using various concentrations of 4-[4-(4-oxiranylbu-
toxy)phenylazo]benzoic acid 4-oxiranylbutyl ester (epoxide1).
Picture area: 780 µm × 625 µm.
networks from 3% ether1 andepoxide2 are apparently
formed by connection of thin leaves (∼200 nm thick).
Another note regarding the previous study using
chiral azobenzene polyacrylates and polymethacrylates1
needs to be made. One dimethacrylate monomer having

Macromolecules, Vol. 36, No. 24, 2003
Alignment of Ferroelectric Liquid Crystals 9039
F igu r e 10. Polarizing optical micrographs showing changes
after isotropization of a sample with 1% 4-[4-(4-oxiranylbu-
toxy)phenylazo]benzoic acid 4-oxiranylbutyl ester (epoxide1)
containing two areas with perpendicularly aligned domains.
The sketch depicts the reorientation of smectic domains and
the possible relaxation of the azobenzene polymer network (see
text for details). Picture area: 940 µm × 780 µm.
F igu r e 9. Polarizing optical micrographs showing the pho-
toinduced reorientation of ferroelectric liquid crystal for a
sample with 5% 4-(4-vinyloxymethoxyphenylazo)benzoic acid
6-vinyloxyhexyl ester (ether1). The sketch depicts the reori-
entation of smectic domains and the possible change in the
azobenzene polymer network (see text for details). Picture
area: 940 µm × 780 µm.
procedure as in Figure 9. Again, the photoinduced
reorientation of azobenzene moieties is responsible for
the result. Nevertheless, the apparently sharper bound-
ary between the two areas with perpendicularly aligned
domains is in contrast with the result in Figure 9 and
hints at the effect of an anisotropic network. This is
because at this low polymer network concentration, a
change in network anisotropy in the area subjected to
the second irradiation would be expected to affect less
the covered area due to the low network density. For
the experiment, the film was then heated to the isotropic
phase (10 min at 100 °C) and cooled back to room
temperature. While the perpendicular bulk alignment
was recovered in the two areas, with the boundary
remained visible, the smectic domains in both areas
changed their orientation direction to about 45° with
respect to the boundary. This phenomenon may be the
manifestation of evolution of the whole anisotropic
polymer network crossing the boundary, which deter-
mined the recovered alignment of FLC. A plausible
option for the relaxation of the anisotropic network
during the isotropization process is schematically il-
lustrated in the figure, being a compromise for the
different anisotropies in the two areas initially imposed
by the two irradiations. On the other hand, without
invoking a dominant effect by an anisotropic network,
the result may come from relaxation of aligned smectic
domains in the presence of a polymer network. In this
case, the sketch simply presents the changes in domain
alignment direction.
Polarized UV-vis spectroscopic measurements con-
firmed that the photoinduced reorientation of FLC is
accompanied by reorientation of azobenzene moieties on
the polymer network. Plotted in Figure 11 is an example
of the angular dependence of the absorbance of azoben-
zene, measured at 398 nm for a sample with 1%
epoxide2 subjected to two irradiations with the polariza-
tion changed by 90°, the angle being that between the
polarization of the beam of spectrophotometer and the
polarization of irradiation light used for the first ir-
radiation. It is seen that the orientation of azobenzene
moieties is in the same direction as FLC domains, i.e.,
perpendicular to the polarization of irradiation light,
and that the reorientation of azobenzene after the
aluminum sheet), while the sample being heated to and
cooled from the isotropic phase. A 90° reorientation of
smectic domains was achieved in the irradiated area,
with the reoriented domains appearing dark under
crossed polarizers. Interestingly, smectic domains in the
nonirradiated area were severely altered by the reori-
entation in the irradiated area in such way that they
changed the alignment direction gradually across the
boundary. The photoinduced 90° reorientation of azoben-
zene moieties in the irradiated area is responsible for
the reorientation of FLC in the irradiated area and may
lead to the observed changes in the nonirradiated area
through two possible ways. First, the photoinduced
reorientation of azobenzene moieties changes the ani-
sotropy of the polymer network; that is, the network is
“stretched” in the same direction as the azobenzene
groups and reoriented by 90° during the second irradia-
tion. As a result of this, the left side of the network in
the covered area would be dragged to follow, creating a
curvature that dictates the gradual change in alignment
of the smectic domains through the boundary. The
second possible mechanism is that the 90° reorientation
of azobenzene moieties in the irradiated area induces a
reorientation of the surrounding FLC molecules, which
results in the 90° reorientation of smectic domains in
this area. Because of the strong cooperative movement
that characterizes liquid crystals, the propagation of this
reorientation effect leads to the gradual changes in
domain alignment direction through the boundary. In
the latter case, the sketch in Figure 9 simply illustrates
the alignment and interaction of smectic domains aris-
ing from the reorientation of azobenzene moieties in the
irradiated area. As mentioned earlier, SEM could not
confirm the formation of an anisotropic network for the
reason explained, but the orientational memory effect
suggests its existence. What actually happened may
well be a combination of the two mechanisms that both
find the origin in the photoinduced reorientation of
azobenzene moieties on the polymer network.
Another experiment (Figure 10) is also worth being
presented. A 90° reorientation was first achieved for
part of a sample containing 1% epoxide1, using the same

9040 Se´vigny et al.
Macromolecules, Vol. 36, No. 24, 2003
F igu r e 12. Electrooptical effects for photoaligned ferroelectric
liquid crystal with different network concentrations of 4-(4-
vinyloxymethoxyphenylazo)benzoic acid 6-vinyloxyhexyl ester
(ether1). Applied ac field: triangular wave, 100 Hz, 20 V (peak-
to-peak).
F igu r e 11. Angular dependence of the absorbance of azoben-
zene moieties on the network at 398 nm for a sample with 1%
4-[2,4-bis(4-oxiranylbutoxy)phenylazo]benzonitrile (epoxide2)
after two irradiations with the polarization changed by 90°.
demonstrate unambiguously that using the photoalign-
ment of an azobenzene polymer network to induce bulk
alignment of FLC represents a generic approach, even
though different mechanisms of action may be involved.
In the case of azobenzene polyethers and polyepoxides,
and achiral polyacrylates as well,11 the results suggest
the formation of an anisotropic polymer network that
induces the alignment of FLC and provides the char-
acteristic orientational memory effect. By contrast, an
isotropic polymer network of colloidal particles may be
formed with a chiral dimethacrylate.1 In that case, the
photoalignment of azobenzene moieties on the surface
of the network commands the alignment of FLC, and
the orientational memory effect is absent. Using some
azobenzene monomers, good-quality bulk alignment of
FLC could be achieved with as low as 1% of monomer,
which points out the potential of discovering new
azobenzene systems that may satisfy the requirements
for practical uses of photoaligned FLC.
second irradiation is obvious. A slight decrease in
absorbance of azobenzene is noted after the second
irradiation. This may be explained by the fact that data
in Figure 11 were obtained from two separate sets of
measurements. After the second irradiation, during
which the sample was placed inside the hot stage for
heating and cooling, the sampling area for the recording
of polarized spectra may not be exactly at the same
position as in the first series of measurements after the
first irradiation. No noticeable development of azoben-
zene moieties aligned perpendicular to the film plane,
which may also decrease the absorption of azobenzene,
was observed in separate test experiments by exposing
samples to different irradiations.
Similar to the use of chiral azobenzene polyacrylates
and polymethacrylates,1 photoinduced alignment of FLC
with polyether and polyepoxide networks was difficult
to achieve on indium-tin oxide (ITO)-coated glass
plates. However, when the ITO surface was treated with
ozone, which increases the surface tension, reasonably
good bulk alignment could be obtained. Figure 12 shows
an example of the electrooptical effects of the photo-
aligned FLC with 1-5% ether1. Even though the exact
data may vary from one series of samples to another
due to a varying quality of bulk alignment, the results
are representative and some trends can be noticed.
Generally, the optical contrast increases with decreasing
the network concentration because of reduced light
scattering. The hysteresis loop also becomes larger for
a higher concentration of network. As compared with
surface-aligned pure FLC with a comparable quality of
alignment, the presence of a polymer network slightly
diminishes the optical contrast and enlarges the hys-
teresis loop.
Ack n ow led gm en t. We thank the Natural Sciences
and Engineering Research Council of Canada and le
Fonds que´be´cois de la recherche sur la nature et les
technologies (Que´bec) for financial support.
Refer en ces a n d Notes
(1) Leclair, S.; Mathew, L.; Gigue`re, M.; Motallebi, S.; Zhao, Y.
Macromolecules 2003, 36, 9024.
(2) Yang, D.-K.; Chien, L. C.; Fung, Y. K. In Liquid Crystals in
Complex Geometries Formed by Polymer and Networks;
Crawford, G. P., Zumer, S., Eds.; Taylor & Francis: London,
1996; pp 103-143.
(3) Hikmet, R. A. M. Adv. Mater. 1995, 7, 300.
(4) Guymon, C. A.; Dougan, L. A.; Martens, P. J .; Clark, N. A.;
Walba, D. M.; Bowman, C. N. Chem. Mater. 1998, 10, 2378.
(5) Clark, N. A.; Larerwall, S. T. Appl. Phys. Lett. 1980, 36. 899.
(6) Molsen, H.; Kitzerow, H.-S. J . Appl. Phys. 1994, 15, 710.
(7) Mao, G.; Wang, J .; Ober, C. K.; Brehmer, M.; O’Rourke, M.
Con clu sion s
J .; Thomas, E. L. Chem. Mater. 1998, 10, 1538.
(8) Semmler, K.; Finkelmann, H. Macromol. Chem. Phys. 1995,
196, 3197.
(9) Brehmer, M.; Zentel, R. Macromol. Chem. Phys. 1994, 195,
In this paper, azobenzene-containing divinyl ether
and diepoxide monomers were synthesized and used for
optically aligned FLC in the absence of surface orienta-
tion layers. Together with the previous studies,1,11 we
have performed a systematic investigation on various
types of azobenzene polymer networks. The results
1891.
(10) See for example: (a) Natansohn, A.; Rochon, P. Chem. Rev.
2002, 102, 4139. (b) Kim, D. Y.; Li, L.; J iang, X. L.;
Shivshankar, V.; Kumar, J .; Tripathy, S. K. Macromolecules

Macromolecules, Vol. 36, No. 24, 2003
Alignment of Ferroelectric Liquid Crystals 9041
1995, 28, 8835. (c) Eich, M.; Wendorff, J . H. Makromol. Chem.
Rapid Commun. 1987, 8, 59. (d) Wu, Y.; Zhang, Q.; Kanaza-
wa, A.; Shiono, T.; Ikeda, T.; Nagase, Y. Macromolecules
1999, 32, 3951. (e) Fischer, T.; Lasker, L.; Stumpe, J .;
Kostromin, S. Photochem. Photobiol. A: Chem. 1994, 80, 453.
(11) Zhao, Y.; Paiement, N. Adv. Mater. 2001, 13, 1891.
(12) Ichimura, K. Chem. Rev. 2000, 100, 1847.
(14) Broer, D. J .; Lub, J .; Mol, G. N. Macromolecules 1993, 26,
1244.
(15) Menzel, H.; Weichart, B.; Schmidt, A.; Paul, S.; Knoll, W.;
Stumpe, J .; Fischer, T. Langmuir 1994, 10, 1926.
(16) Zhao, Y.; Chenard, Y. Macromolecules 2000, 33, 5891.
(17) Odian, G. Principles of Polymerization, 3rd ed.; J ohn Wiley
& Sons: New York, 1991.
(13) J onsson, H.; Andersson, H.; Sundell, P.-E.; Gedde, U. W.;
Hull, A. Polym. Bull. (Berlin) 1991, 25, 641.
MA034885L