Cyclic azobenzene-containing side-chain liquid crystalline polymers: Synthesis and topological effect on mesophase transition, order, and photoinduced birefringence

Article abstract of DOI:10.1021/ma100246c

A cyclic side-chain liquid crystalline polymer (SCLCP) bearing azobenzene mesogens, namely, poly{6-[4-(4-methoxyphenylazo)phenoxy]hexyl methacrylate} (PAzoMA), was successfully synthesized by using "click" cyclization of the linear polymer precursor with alkyne and azide end groups. Samples of cyclic-PAzoMA of various molecular weights were prepared, characterized, and studied in comparison with their linear counterparts. The results show that the topological constraint arising from the tortuosity of the ring structure and the absence of chain ends in cyclic-SCLCPs affects profoundly the liquid crystalline (LC) phase transitions (temperature, enthalpy, and entropy) of mesogenic side groups and that this topological effect is more prominent for smaller SCLCP rings. Moreover, the photoinduced anisotropy in films as a result of the trans-cis photoisomerization of azobenzene mesogens was investigated, and cyclicPAzoMA was found to behave differently from linear-PAzoMA. On the one hand, the cyclic polymer exhibits a nonmonotonic rise and erasure of birefringence upon linearly and circularly polarized excitation (488 nm), respectively, in contrast with the linear polymer displaying monotonic changes. On the other hand, unlike the linear polymer, the photoinduced orientation of azobenzene mesogens in cyclic-PAzoMA cannot be enhanced upon annealing in the nematic phase. All these manifestations of the topological constraint suggest that cyclization offers a new way to change the coupling between mesogenic side groups and chain backbone of SCLCPs, and their interplay under the additional topological effect may generate new behaviors that are of interest to be explored.

Full text of DOI:10.1021/ma100246c

3
664 Macromolecules 2010, 43, 3664–3671
DOI: 10.1021/ma100246c
Cyclic Azobenzene-Containing Side-Chain Liquid Crystalline Polymers:
Synthesis and Topological Effect on Mesophase Transition, Order, and
Photoinduced Birefringence
†
†
†
‡
Dehui Han, Xia Tong, Yi Zhao, Tigran Galstian, and Yue Zhao*
,†
†
D eꢀ partement de chimie, Universit eꢀ de Sherbrooke, Sherbrooke, Qu eꢀ bec, Canada J1K 2R1, and
‡
Centre d’Optique, Photonique et Lasers, D eꢀ partement de Physique, Universit eꢀ Laval, 2375 Rue de la Terrasse,
Qu eꢀ bec, Canada G1V 0A6
Received February 3, 2010; Revised Manuscript Received March 12, 2010
ABSTRACT: A cyclic side-chain liquid crystalline polymer (SCLCP) bearing azobenzene mesogens,
namely, poly{6-[4-(4-methoxyphenylazo)phenoxy]hexyl methacrylate} (PAzoMA), was successfully synthe-
sized by using “click” cyclization of the linear polymer precursor with alkyne and azide end groups. Samples
of cyclic-PAzoMA of various molecular weights were prepared, characterized, and studied in comparison
with their linear counterparts. The results show that the topological constraint arising from the tortuosity of
the ring structure and the absence of chain ends in cyclic-SCLCPs affects profoundly the liquid crystalline
(
LC) phase transitions (temperature, enthalpy, and entropy) of mesogenic side groups and that this
topological effect is more prominent for smaller SCLCP rings. Moreover, the photoinduced anisotropy in
films as a result of the trans-cis photoisomerization of azobenzene mesogens was investigated, and cyclic-
PAzoMA was found to behave differently from linear-PAzoMA. On the one hand, the cyclic polymer exhibits
a nonmonotonic rise and erasure of birefringence upon linearly and circularly polarized excitation (488 nm),
respectively, in contrast with the linear polymer displaying monotonic changes. On the other hand, unlike the
linear polymer, the photoinduced orientation of azobenzene mesogens in cyclic-PAzoMA cannot be
enhanced upon annealing in the nematic phase. All these manifestations of the topological constraint suggest
that cyclization offers a new way to change the coupling between mesogenic side groups and chain backbone
of SCLCPs, and their interplay under the additional topological effect may generate new behaviors that are of
interest to be explored.
1
5
Introduction
on the ring) and polymers bearing these cyclic oligomers as side
1
6
groupsoron the main chain. Overall, however, cyclic functional
or stimuli-responsive polymers have not been much explored
until now, and their full potential in the development of new
functional materials remains to be unveiled.
This paper reports the synthesis and characterization of, to our
knowledge, the first cyclic side-chain liquid crystalline polymer
The importance of cyclic polymers in biological and macro-
1-3
molecular sciences has long been recognized.
Compared with
linear polymers, there are relatively few reports on cyclic poly-
mers, and the number of cyclic polymers studied is limited as well,
partly due to the more demanding synthetic procedures. In recent
years, there has been a growing interest in cyclic polymers thanks
to the development of a number of efficient synthetic methods
(
cyclic-SCLCP) with pendant azobenzene mesogenic groups. The
4-11
successful synthesis of cyclic-SCLCP was achieved by using the
aforementioned “click” cyclization method. We anticipated that
the topological constraint of cyclic polymers could affect the LC
phase transitions of mesogenic side groups. By having azo-
benzene mesogens, the effect of the cyclic structure on the
photoinduced anisotropy upon trans-cis photoisomerization
could also be investigated. As reported in this paper, the results
of our comparative study with cyclic-SCLCP and their linear
precursors (linear-SCLCP) revealed profound, and sometimes
surprising, topological effects on the mesophase transitions and
the photoinduced anisotropy.
that make their preparation more accessible.
Of them, the
“
click” cyclization of R-alkynyl-ω-azido heterodifunctional
linear polymers under high dilution is particularly attractive,
since it can be applied to many polymers synthesized by using
6-9
atom transfer radical polymerization (ATRP).
Cyclic poly-
mers display distinct topological features; the tortuosity of the
chain ring structure and the absence of chain ends could make
them to behave differently from their linear counterparts both
1-12
in solution and in the melt.
In solution, the most investigated
properties of cyclic polymers are the hydrodynamic volume,
radius of gyration, and viscosity. Noticeably, a recent study of
Xu et al. found different thermal phase transition behaviors of
cyclic poly(N-isopropylacrylamide) (PNIPAM) in aqueous solu-
Experimental Section
7
tion as compared to the linear polymer. In the bulk, in addition
1
. Polymer Synthesis. The synthetic route to the cyclic azo-
to the viscosity and thermostability, the glass transition tempera-
benzene-containing SCLCP is depicted in Scheme 1. More
details are given below.
1
3
ture (T ) (e.g., for cyclic polystyrene) and the melting tempera-
g
1
4
ture (T ) (mainly for cyclic poly(ethyleneoxide)) havealsobeen
Materials. Prior to use, tetrahydrofuran (THF, 99%) was
refluxed with sodium and a small amount of benzophenone and
distilled, triethylamine (TEA) (Aldrich, g99) was refluxed with
p-toluenesulfonyl chloride (Fluka, g99%) and distilled, and N,
N-dimethylformamide (DMF, 99.8%) was dried with anhydrous
m
studied. In the case of liquid crystalline polymers (LCPs), Percec
et al. studied cyclic LC oligomers (with up to 4-5 mesogenic units
*Corresponding author. E-mail: yue.zhao@usherbrooke.ca.
pubs.acs.org/Macromolecules Published on Web 03/29/2010
r 2010 American Chemical Society

Article
Scheme 1. Synthetic Route to Cyclic Azobenzene-Containing
Macromolecules, Vol. 43, No. 8, 2010 3665
twice, the linear SCLCP precursor, linear PAzoMA-N , was
collected (yellow solid, 0.43 g, yield 86%). H NMR (600 MHz,
3
1
Side-Chain Liquid Crystalline Polymers
δ ppm, CDCl ): 7.83 (b, 4H, o-Ar H to -NdN-), 6.94 (b, 4H,
3
m-Ar H to -NdN-), 4.60 (b, 2H, COOCH
2
CtCH), 3.95 (b,
), 1.96 (b, 6H, OOC(CH ),
CH CH CH O and main chain
4
H,OCH
.2-2.0 (b, 10H, OCH
2
), 3.84 (b, 3H, OCH
CH CH
3
3 2
)
1
2
2
2
2
2
2
CH ), 0.8-1.1 (b, 3H, main chain CH ).
2
3
Synthesis of Cyclic Side-Chain Liquid Crystalline Polymer
Cyclic-PAzoMA). 160 mL of DMF was added into a 250 mL
(
three-necked round-bottom flask, followed by three freeze-
pump-thaw cycles for degassing. CuBr (205.6 mg, 1.4 mmol)
and PMDETA (247.8 mg, 1.4 mmol) were then charged into the
2 3
flask under protection of N flow. Linear-PAzoMA-N (0.10 g,
-
3
7
tion was thoroughly deoxygenated by bubbling with N for 1 h.
.14ꢀ10 mmol) was dissolved in DMF (40 mL), and the solu-
2
Using a pressure-equalizing addition funnel, the polymer solu-
tion was added into the CuBr/PMDETA reaction mixture at
1
00 ꢀC very slowly, over 24 h. After the addition of polymer
solution was completed, the reaction was allowed to proceed for
another period of 24 h. Afterward, DCM was used to extract the
polymer after cooling the reaction mixture to room temperature.
The organic phase was washed several times with saturated
NaHSO solution, dried over MgSO , and concentrated with a
4
4
rotary evaporator. The resulting cyclic polymer, cyclic-PAzo-
MA, was then purified by precipitation in methanol (yellow
1
solid, 0.072 g, 72% yield). H NMR (600 MHz, δ ppm, CDCl
3
):
7
-
3
.83 (b, 4H, o-Ar H to -NdN-), 6.94 (b, 4H, m-Ar H to
NdN-), 5.20 (b, 2H, COOCH ), 3.95 (b, 4H,OCH ),3.84 (b,
H, OCH ), 1.96 (b, 6H, OOC(CH ),1.2-2.0 (b, 10H, OCH
2
2
3
3
)
2
2
-
CH
2
CH
2
CH
2
CH
2
CH
2
O and main chain CH ), 0.8-1.1 (b, 3H,
2
main chain CH3).
1
magnesium sulfate and distilled under reduced pressure. Dichloro-
methane (DCM) was distilled from CaH . Copper(I) bromide
2. Characterization. H NMR spectra were recorded on a
Bruker 600 MHz spectrometer using deuterated chloroform as
solvent and tetramethylsilane as internal standard. The spectra
were used to determine the number-average molecular weights
n
(M ) of the linear polymer precursors. A Waters size exclusion
chromatograph (SEC) instrument, equipped with a Waters 410
2
0
(
CuBr, 99.999%), R-bromoisobutyryl bromide (98%), N,N,N ,
0
00
N ,N -pentamethyldiethylenetriamine (PMDETA, 99%), and
propargyl alcohol (99%) were purchased from Aldrich and
used directly. The monomer with an azobenzene group, 6-[4-(4-
methoxyphenylazo)phenoxy]hexyl methacrylate (AzoMA), was
differential refractometer detector and a Waters 996 photodiode
array detector, was also utilized to measure the number- and
weight-average molecular weights (M and M ) as well as the
1
7
synthesized by using a literature method.
Synthesis of ATRP Initiator: Propargyl 2-Bromoisobutyrate.
Into a 250 mL round-bottom flask with a magnetic stirrer,
propargyl alcohol (7.0 g, 130 mmol), TEA (17.5 mL, 130 mmol),
and CH Cl (100 mL) were added. After the mixture was cooled
n
w
polydispersity index (PDI). The SEC measurements were con-
duced at 35 ꢀC using one column (Waters Styragel HR4E,
7.8 mmꢀ300 mm, 5 μm beads), polystyrene (PS) standards for
2
2
-1
to 0 ꢀC, R-bromoisobutyryl bromide (28.8 g, 130 mmol) was
added dropwise over a period of 30 min. The reaction mixture
was then brought back to room temperature and stirred for 24 h.
The salt formed was removed by filtration, and the filtrate was
washed with NaCl solution several times. The organic phase was
calibration, and THF as the eluent (flow rate: 1.0 mL min ). A
TA Q200 differential scanning calorimeter (DSC) was used to
investigate the phase transition behaviors, using indium as the
calibration standard and a heating or cooling rate of 10 ꢀC
-
1
min . The glass transition temperature (T ) was measured as
g
dried over MgSO
pressure, the colorless propargyl 2-bromoisobulyrate was obta-
4
. After the removal of solvent under reduced
the midpoint of the change in heat capacity, while mesophase
transition temperatures were taken as the maximum of the
respective endothermic peak. Polarizing optical micrographs
(POM) were obtained using a Leitz DMR-P microscope equip-
ped with an Instec hot stage. UV-vis spectra were recorded with
a Varian 50 Bio spectrophotometer, while Fourier transform
infrared (FTIR) spectra were recorded on a Nicolet AVATAR
370 DTGS FTIR spectrometer.
1
ined by distillation under reduced pressure. H NMR (300 MHz,
CDCl
CH CtCH); 1.96 (6H, C(CH )
3
), δ (TMS, ppm): 4.80 (2H, COOCH
2
); 2.50 (1H, COO-
2
3 2
).
Synthesis of Linear Polymer Precursor (Linear-PAzoMA-N3).
Propargyl 2-bromoisobutyrate (0.01 g, 0.05 mmol), CuBr (7.06 mg,
0
0
1
.05 mmol), AzoMA (1.5 g, 3.8 mmol), PMDETA (8.50 mg,
.05 mmol), and THF (4.0 mL) were added successively into a
0 mL flask. The reaction mixture was degassed by three pump-
3. Optical Measurements. Two types of optical measurements
were performed in this study. On the one hand, the dynamic
process of photoinduced birefringence in thick films (thickness
∼10 μm) of cyclic- and linear-PAzoMA was monitored at room
temperature by using a polarimetric optical setup described
2
thaw cycles, backfilled with N , and placed in an oil bath
thermostated at 70 ꢀC for 1.5 h. It was then diluted with THF
and passed through a column of neutral alumina to remove the
metal salt. After precipitation by adding the polymer solution
of THF into methanol, the yellow polymer PAzoMA-Br was
collected by filtration and then dried under vacuum overnight
1
8
elsewhere. In essence, a polymer film, cast on a glass slide
and preirradiated with a UV lamp to reach the photostationary
cis-rich state, was placed between two crossed polarizers and
þ
(
0.63 g, yield 42%). Afterward, azide group-ended polymer was
obtained as follows. PAzoMA-Br (0.50 g, 0.04 mmol), NaN
0.023 g, 0.36 mmol), and DMF (5 mL) were added into a 10 mL
exposed to a linearly polarized Ar ion laser beam (λ=488 nm)
3
with a small incident angle to the film normal. The polarization
of the excitation beam was set at 45ꢀ with respect to the crossed
polarizers, the excitation beam had a spot of about 2 mm in dia-
meter, and its power was adjusted to be 42 mW. The photo-
(
round-bottom flask with a magnetic stirrer, and the reaction
mixture was stirred for 24 h at room temperature. After puri-
fication by precipitation of the polymer solution into methanol
þ
induced birefringence in the area hit by the Ar laser beam resulted

3
666 Macromolecules, Vol. 43, No. 8, 2010
Table 1. Characteristics of Linear- and Cyclic Azobenzene-Containing Side-Chain Liquid Crystalline Polymers
Han et al.
a
b
c
M
(g mol
n
M
(g mol
n
ΔSsn
ΔHni
1-
ΔSni
-
1
)
-1
c
ΔHsn (J g-1
-1 -1
-1 -1
sample
)
PDI
T
g
(ꢀC)
T
sn (ꢀC)
)
(J g
K
)
T
ni (ꢀC)
(J g
)
(J g
K
)
-
-
-
-
-
-
3
3
2
3
3
3
-3
-3
linear-PAzoMA12
cyclic- PAzoMA12
linear-PAzoMA18
cyclic- PAzoMA18
linear-PAzoMA35
cyclic- PAzoMA35
a
4750
7130
6010
1.15
1.15
1.18
1.16
1.10
1.09
68.8
56.1
76.4
57.6
78.9
60.3
82.6
74.3
89.2
74.5
91.3
85.9
3.27
2.19
5.47
2.66
2.92
2.27
9.19 ꢀ 10
6.30 ꢀ 10
1.50 ꢀ 10
7.65 ꢀ 10
8.01 ꢀ 10
6.32 ꢀ 10
120.9
106.4
129.0
105.9
131.3
120.2
3.51
1.87
5.02
1.49
2.50
1.68
8.91 ꢀ 10
5630
9170
8250
14000
12160
4.93 ꢀ 10
-2
1.25 ꢀ 10
3.93 ꢀ 10
6.18 ꢀ 10
-3
-3
-3
13860
4.27 ꢀ 10
M
n
: number-average molecular weight; PDI: polydispersity index; T
g
: glass transition temperature; Tsn, ΔHsn, and ΔSsn: smectic-to-nematic phase
transition temperature, enthalpy, and entropy, respectively; Tni, ΔHni, and ΔSni: nematic-to-isotropic phase transition temperature, enthalpy, and
entropy, respectively. From H NMR spectra. From SEC measurements
b
1
c
Figure 1. Size-exclusion chromatograph curves: (A) linear- and cyclic azobenzene-containing side-chain liquid crystalline polymers of different
molecular weights; (B) linear precursor and the sample obtained by reaction at a high precursor concentration (about 100 times the concentration used
for cyclization) showing interchain coupling.
in the polarization variation of a probe light. A weak He-Ne
laser (633 nm, 4 mW) was used to provide the probe beam at
normal incidence, having little effect on the isomerization of
azobenzene mesogens. With the crossed polarizers, the polar-
ization change of the probe light led to a transmission intensity
variation measured by means of a photodetector, which allowed
the photoinduced birefringence to be calculated. In a typical
experiment, when the photoinduced birefringence reached stabi-
lization state (under the linearly polarized excitation), the exci-
tation beam was turned off for a period of time for the observa-
tion of its stability in time. This was followed by the application
Results and Discussion
. Synthesis of Cyclic Side-Chain Liquid Crystalline Poly-
mers. Three samples of different molecular weights of linear-
PAzoMA, with a terminal alkyne and an azide group, were
synthesized using ATRP followed by a chain-end functiona-
1
lization reaction (Scheme 1). M of these linear polymer
n
1
precursors were determined from their H NMR spectra (see
below) and also from SEC using PS standards. The results
are given in Table 1; the number in the acronym of each
sample is the number of monomer units calculated from the
þ
of a circularly polarized excitation beam of the Ar ion laser
(
NMR-based M . Under the used conditions, in particular
n
obtained by means of adding a quarter-wave plate on the
the high dilution of linear polymer precursor ensured by its
very slow addition into the reaction solution, the intrachain
click reaction of the alkyne and azide end groups of the
linear-PAzoMA precursors resulted in their cyclization,
giving rise to cyclic-PAzoMA. The successful cylization,
rather than condensation, was revealed by the results obta-
ined using a number of characterization methods. Figure 1a
shows the SEC curves of all three cyclic polymers and their
linear counterparts. In all cases, the cyclic polymer displayed
a longer retention time (larger retention volume) than the
linear counterpart, meaning that the cyclic polymer has a
smaller hydrodynamic volume than the linear one. The
difference in retention time is most prominent between the
optical path of the same excitation beam), and the dynamic
process of erasure of the orientation axes of azobenzene meso-
gens in the film plane was also recorded.
On the other hand, thin polymer films (thickness ∼40 nm)
were prepared on a quartz plate, for which the photoinduced
orientation of azobenzene mesogens at room temperature and
the effect of thermal annealing into the LC phases could be
characterized by using an order parameter calculated from the
linear dichroism of the absorption band of trans-azobenzene
at around 360 nm. In this experiment, all films were also pre-
irradiated with a UV light (λ=360 nm, 10 mW cm-2, 1 min) at
room temperature before being exposed to linearly polarized
visible light (λ= 440 nm, 1.4 mW cm-2, 30 min). The used UV
and visible light were obtained by using a spot curing system
samples with the largest size (linear- and cyclic-PAzoMA35)
(
Novacure 2100) combined with interference filters (10 nm
while it is less important for the samples with a lower M ,
n
which is consistent with reports on conventional polymers.
Without ruling out the possibility of having traces of linear
species in the cyclic samples, the display of longer retention
times, together with the spectral evidence of an effective click
2
bandwidth, Oriel). The films were then annealed in an oven
with a preset temperature for 10 min before being taken off
from the oven for polarized UV-vis measurements at room
temperature.

Article
Macromolecules, Vol. 43, No. 8, 2010 3667
1
Figure 2. (A) H NMR and (B) infrared spectra of linear- and cyclic-
PAzoMA35
Figure 3. DSC heating (A) and cooling (B) curves of linear- and cyclic
azobenzene-containing side-chain liquid crystalline polymers of diffe-
rent molecular weights (second scan).
.
reaction between alkyne and azide groups (see below), clearly
indicates the absence of any meaningful amount of linear
polymers. The apparent changes in molecular weight from
the SEC measurements are shown in Table 1. We also con-
ducted a control test to observe the difference between
cyclization (intrachain coupling) under the high dilution
condition and condensation (interchain coupling) at a high
concentration of the polymer precursor. Figure 1b compares
the SEC curves of linear-PAzoMA18 and the sample obta-
ined after 15 h of reaction at a precursor concentration about
for the cyclic polymer. All the results indicate that the “click”
cyclization of R-alkynyl-ω-azido heterodifunctional linear
polymers under high dilution is an efficient way to prepare
cyclic azobenzene-containing SCLCPs.
2. Thermal Mesophase Transitions. As compared to linear
SCLCP, the topological constraint subjected by cyclic SCLCP
would understandably influence the order and organization
of the side-group mesogens. This was indeed observed from
changes in mesophase transition temperature, enthalpy, and
entropy. Figure 3 show the DSC heating and cooling curves
(second scan) of the three linear- and cyclic-PAzoMA; the
different thermal behaviors are evident. The used linear SCLCP
1
00 times that used for preparing the cyclic polymers in
Figure 1a (all other conditions were kept the same). In this
case, the sample displayed shorter retention times than the
linear precursor, indicating the formation of species with an
increased molecular weight that could only be formed by
19
is known to display a nematic and smectic-A phase; on
heating to and cooling from the isotropic phase, the phase
transition endotherms (peaked at T and T ) and exotherms
interchain coupling. The apparent M increased from 9170
n
sn
ni
to 14 750 and the PDI from 1.18 to 1.28.
H NMR and infrared spectroscopic measurements also
(T and T ), respectively, are visible. For the cyclic SCLCP,
in ns
1
the two mesophase transitions remained but occurred at
significantly lower temperatures and with broadened transi-
tion endothermic or exothermic peaks. The smectic-to-nematic
and nematic-to-isotropic phase transition temperatures on
confirmed the occurrence of the cycloaddition reaction
between the alkyne and azide groups. Figure 2 shows the
H NMR and IR spectra of linear- and cyclic-PAzoMA35 as
1
example. From NMR, on the one hand, the methylene
protons next to the alkyne group had a signal at ∼4.6 ppm.
heating (Tsn and T ) as well as the corresponding enthalpies
ni
(ΔHsn and ΔH ) and entropies (ΔS and, ΔS ), calculated
ni
sn
ni
(
This peak was used to determine M of the linear polymer
n
according to ΔSsn(or ni) = ΔHsn(or ni)/Tsn(or ni), are also sum-
marized in Table 1. And to better illustrate the effect of
molecular weight on the differences between linear- and
cyclic-PAzoMA, their Tg and two mesophase transition
temperatures measured on the heating scan are plotted as a
through comparison of integrals.) This characteristic signal
disappeared for the cyclic polymer which displayed a new
signal (weak but noticeable) at ∼5.2 ppm, attributed to
3
,6
methylene protons next to 1,2,3-trazole. From IR, on the
other hand, it is seen that the characteristic absorption band
function of M (NMR) in Figure 4. It can be seen that all cyclic
n
polymers have lower T and LC phase transition temperatures
g
-
1
of azide groups at ∼2110 cm of the linear polymer was absent

3
668 Macromolecules, Vol. 43, No. 8, 2010
Han et al.
Figure 4. Glass and mesophase transition temperatures (on heating) vs
molecular weight for linear- and cyclic azobenzene-containing side-
chain liquid crystalline polymers.
than their linear counterparts. With respect to cyclic-PAzo-
MA , the decreases appear to be more important for the two
3
5
cyclic polymers of smaller sizes, revealing a greater topo-
logical constraint effect on smaller cycles of SCLCP. How-
ever, cyclic-PAzoMA18 and cyclic-PAzoMA12 displayed
similar mesophase transition temperatures. From Table 1,
it can also be noticed that the topological constraint effect in
cyclic-PAzoMA led to a reduction of the mesophase transi-
tion enthalpy and entropy. Qualitatively, the smaller phase
transition enthalpies suggest a decreased molecular order of
mesogens in the LC phases of cyclic polymers, weakening the
intermolecular interactions. As for the smaller phase transi-
tion entropies, this may be accounted by a more important
reduction of polymer chain mobility in the isotropic phase
Figure 5. Polarizing optical micrographs of linear- and cyclic-PAzo-
MA35 recorded, uponcoolingfrom isotropic phase, at twotemperatures
i6 n7 tμh me .nematic and smectic phase, respectively. Image area: 67 μm ꢀ
with respect to the nematic phase (for ΔS ) and in the
On the one hand, the kinetics of photoinduced birefringence
in thick polymer films (thickness ∼10 μm) at room tempera-
ture was monitored upon excitation at 488 nm using an Ar
ni
nematic phase with respect to the smectic phase (for ΔSsn).
We carried out powder X-ray diffraction measurements on
linear- and cyclic-PAzoMA, but the results did not show any
noticeable difference on the ordering of smectic layers,
noting that even for linear-PAzoMA, smectic layers were
difficult to develop to be observed by X-ray diffraction.
Figure 5 shows an example of polarizing optical micro-
graphs with linear- and cyclic-PAzoMA . They were taken
1
8
ion laser. On the other hand, the development of photo-
induced orientation of azo mesogens in thin films (thickness
∼40 nm) upon annealing in the LC phases was measured by
recording polarized UV-vis spectra. In both experiments,
prior to the excitation using a linearly polarized visible light,
all films were preirradiated with an unpolarized UV lamp to
obtain the photostationary cis-rich state for azo mesogens,
which is a known treatment for azo SCLCP whose trans-
3
5
in the course of cooling from the isotropic phase, at two
temperatures in the nematic and smectic phase, respectively.
Knowing that the smectic layers were difficult to develop in
these polymers, the micrographs in the smectic phase were
recorded after 24 h of annealing at the chosen temperature
2
3,27,28
azobenzene absorbs at around 360 nm.
Before report-
ing and discussing the results, we mention that the UV-vis
absorption spectra of linear- and cyclic-PAzoMA are quite
similar, and the trans-cis photoisomerization occurs both in
solution and in the solid state. An example of the UV-vis
spectra of thin films of linear- and cyclic-PAzoMA35 is
shown in Figure 6. The two films were cooled from isotropic
phase to room temperature before recording the spectra.
Prior to UV irradiation, the same spectral shapes indicate a
similar aggregation sate of azo mesogens in the linear and
cyclic polymer, displaying a broad absorption band peaked
at about 360 nm (π-π* transition of nonaggregated trans
azo) with blue-shifted (H-aggregation) and red-shifted
(J-aggregation) shoulders. In both cases, the photostation-
ary cis-reach state could readily be obtained after 1 min UV
(
84 ꢀC for linear-PAzoMA and 77 ꢀC for cyclic-PAzoMA ).
35
35
For both samples, the transition from nematic to smectic
phase resulted in a change in the LC texture; however, des-
pite the thermal treatment, no clear focal conic texture could
be observed for the smectic phase, and the samples showed
mainly a schlieren texture of nematic phase. The difference
between the linear and cyclic polymer is clear, particularly in
the nematic phase, where the cyclic sample displayed more
threads under crossed polarizers, indicating more and smaller
nematic domains with changing LC directors. This may also be a
manifestation of the topological constraint effect that hampers
the LC order and causes more defects in the cyclic SCLCP.
-
2
3
. Photoinduced Anisotropy. The use of SCLCP bearing
irradiation (360 nm, 10 mW cm ) inducing the trans-cis
isomerization. From the decrease of absorption of the trans
isomer, being accompanied by the increase of absorption of
cis isomer peaked at ∼450 nm (n-π* transition), the popula-
tion of cis-azo mesogens is about 77% for the linear polymer
and 72% for the cyclic polymer. From this photostationanry
azo mesogens also made it possible to investigate the effect of
topological constraint arising from chain cyclization on the
photoinduced anisotropy as a result of the trans-cis isomeri-
zation of the chromophore, which is a major interest for azo
2
0-28
polymers.
Two types of experiments were performed.

Article
Macromolecules, Vol. 43, No. 8, 2010 3669
Figure 6. UV-vis spectra of thin films of linear- and cyclic-PAzoMA35
recorded at room temperature before and after an UV irradiation
-
2
(
360 nm, 10 mW cm , 1 min), leading to the photostationary state
that is rich in cis-azobenzene.
state, excitation with a linearly polarized visible light (usually
>440 nm) can bring azo mesogens in the cis form back to the
trans form, with concomitant orientation of azo mesogens
perpendicularly to the polarization of the visible light.
The kinetics of photoinduced birefringence at room tem-
perature for linear and cyclic polymers of two different sizes,
PAzoMA35 and PAzoMA , are compared in Figure 7. A
18
typical experiment consisted in first applying a linearly pola-
rized 488 nm light to induce the birefringence, then turning
off the excitation to examine the stability of the birefringence,
and finally applying a circularly polarized light to “erase” the
2
9
Figure 7. Dynamic processes of photoinduced birefringence of (A) linear-
and cyclic-PAzoMA35 and (B) linear- and cyclic-PAzoMA18. The turn-
on and turn-off times of the linearly polarized excitation light as well as
the turn time of the circularly polarized light are indicated by arrows.
“
in-plane” anisotropy. For all films, the small birefrin-
gence existing prior to the linearly polarized excitation was
normalized to zero to better show the part of the photo-
induced birefringence. From the results, it becomes immedi-
ately clear that the cyclic polymers behave differently from
their linear precursors. For the two polymer sizes investi-
gated, upon the laser excitation, the photoinduced birefrin-
gence in the linear polymers grows monotonically and
reaches a greater value than in the cyclic polymers. This
indicates that the greater topological constraint in cyclic
polymers makes the orientation of azo mesogens more
difficult to develop. More interestingly, the dynamic pro-
cesses are also very different. For the two linear polymers,
they have a similar two-process anisotropy increase and
erasure, with PAzoMA18 exhibiting faster times and higher
birefringence than PAzoMA , likely due to an effect of poly-
down before the final stabilization of the signal. More
strikingly, when exposed to circularly polarized light, there
is a fast transient increase of the apparent birefringence
(increase in transmission of the probe light under crossed
polarizers), followed by a partial decrease of the birefrin-
gence. Such nonmonotonic behaviors are also visible for the
smaller cyclic-PAzoMA18. The nonmonotonic increase in
birefringence suggests that there may be at least two mecha-
nisms that are acting in opposite directions during linearly
polarized excitation, while the fast increase in transmission
upon circularly polarized excitation suggests the occurrence
30
of transient out-of-plane reorientation of azo mesogens.
3
5
mer size. For both birefringence induction upon linearly
polarized excitation and its subsequent erasure by circularly
polarized light, the signal changes are quite monotonic over
time, and the fast and slow processes can be well fitted with
two exponentials, giving rise to their rate ratio and signal
ratio (relative contribution to the anisotropy). For linear-
PAzoMA , these two ratios are 5.7 and 1.6 for birefringence
Although the exact mechanisms accounting for the peculiar
behaviors of cyclic polymers are unclear at this time, the fact
that these are observed only after cyclization point to an
origin of topological constraints.
Another experimental design also confirmed the topolo-
gical constraint effect on the photoinduced anisotropy in
cyclic azo SCLCPs. Thin films were prepared on quartz
plates to exhibit an absorbance at 360 nm suitable for mea-
suring the orientation of trans azo mesogens. They were
preirradiated at room temperature with unpolarized UV
light to reach the photostationary cis-rich state (this should
also create an anisotropy axis perpendicular to the film plane
3
5
induction and 5.7 and 1.3 for its erasure, while with linear-
PAzoMA , they are 6.0 and 1.3 for birefringence induction
1
8
and 8.5 and 1.0 for erasure (curve fittings not shown in the
figure for the sake of clarity).
By contrast, the cyclic polymers display more complicated
dynamic processes, and the signal changes appear nonmono-
tonic. In the case of cyclic-PAzoMA , upon linearly polari-
2
9
since the light was normally incident) and then exposed to
linearly polarized visible light to induce an “in-plane” orien-
tation of trans azo mesogens in the direction perpendicular
to the polarization of the excitation visible light. This
photoinduced molecular orientation in thin films could be
3
5
zed excitation, there is a fast increase in birefringence, and
then it is decreased before increasing again more slowly; after
about 140 s of excitation, the increase rate is further slowed

3
670 Macromolecules, Vol. 43, No. 8, 2010
Han et al.
Figure 9. Plots of (A) order parameter vs temperature and (B) absor-
Figure 8. Plots of (A) order parameter vs temperature and (B) absor-
bance at 360 nm vs temperature for linear- and cyclic-PAzoMA35
bance at 360 nm vs temperature for linear- and cyclic-PAzoMA18
.
.
behavior can be explained from the change in absorbance
of the 360 nm band. The drop of order parameter coincides
with a drop of the absorbance, which indicates that the
reorganization of azo mesogens in the smectic phase resulted
in a homeotropic alignment. It appears that in the thin films,
for which the interaction of polymers with the substrate
surface would be important, smectic layers with azo meso-
gens preferentially aligned perpendicular to the film surface
are energetically more stable. In this temperature range, the
difference between cyclic and linear polymers, if there was
any, could not be assessed within experimental error (about
20% on the value of order parameter) due to the low orien-
tation degree. When the films were further annealed at higher
temperatures, into the nematic phase, the mesophase transi-
tion turn azo mesogens to lie in the film plane, as can be seen
from the recovered absorbance in all samples. In the nematic
phase, the orientation of azo mesogens in the two linear
polymers was much enhanced as revealed by the risen order
parameter (error <10%). In this case, the initial photoin-
duced orientation of azo mesogens served as a template to
amplify the macroscopic orientation, leading to a mono-
domain in the nematic phase. By sharp contrast, with the two
cyclic polymers, this photoinduced orientation could not
develop in the nematic phase; the small order parameter
indicates a polydomain morphology in the nematic phase,
revealing again the topological constraint effect that hinders
the ordering and organization of mesogenic side groups. In
all cases, when heated to the isotropic phase, the orientation
of azo mesogens was erased.
measured by recording polarized UV-vis spectra with the
polarization of the spectrophotometer’s beam set to be paral-
lel (A ) and perpendicular (A ) to the polarization of the
^
excitation visible light; from the dichroic ratio of the 360 nm
band (R=A /A ) the order parameter S, characterizing the
^
average orientation of trans azo mesogens, could be calcu-
lated according to S=(R-1)/(R þ 2). A well-known pro-
perty of azo SCLCPs is that when such a film is heated
to above T of the polymer and annealed in a LC phase,
g
the chain mobility allows mesogenic side groups to better
organize, and during this process, the initial photoinduced
orientation of azo mesogens could serve as seed and lead
20-28
to a much enhanced orientation of the mesogens.
We
carried out the experiments with the two cyclic polymers
and their linear precursors. The results are summarized in
Figure 8 for linear- and cyclic-PAzoMA35 and in Figure 9 for
linear- and cyclic-PAzoM . Figures 8a and 9a are plots of order
18
parameter as a function of annealing temperature (10 min at
each temperature before cooling to room temperature for the
measurements), while Figures 8b and 9b are plots of absorbance
of the 360 nm band (calculated from (A þ 2A )/3) vs tempera-
^
ture. Similar observations can be made for the two sizes of
polymers, and the same conclusions can be reached.
With the used thin films, all samples had a similar photo-
induced orientation of azo mesogens initially. As the films
were annealed at T > T , over a temperature range in the
g
smectic A phase (about 60-80 ꢀC), they all showed a dec-
rease of the order parameter. This apparently surprising

Article
Macromolecules, Vol. 43, No. 8, 2010 3671
Conclusions
(2) Kricheldorf, H. R. J. Polym. Sci., Part A: Polym. Chem. 2010, 48,
2
51–284.
A cyclic SCLCP bearing azobenzene mesogens was synthe-
sized, characterized, and studied for the first time. We showed
that the “click” cyclization of R-alkynyl-ω-azido heterodifunc-
tional linear SCLCPs under high dilution is an effective way to
synthesize cyclic SCLCPs. The ring structure-imposed topological
constraint effect on LC phase transitions (temperature, enthalpy,
and entropy), LC order, and the photoinduced orientation of
azobenzene mesogens was evidenced by comparing the behaviors
of cyclic-PAzoMA with the linear polymer precursor. Cyclic
polymers exhibited a lower T , T , and T as well as smaller
phase transition enthalpy and entropy than the linear polymer
counterparts, the differences being more important for smaller
ring sizes. As for the photoinduced birefringence, it was smaller in
the cyclic polymers than in the linear polymers, and the dynamic
processes of birefringence induction and erasure upon linearly
and circularly polarized excitation, respectively, were very differ-
ent as well. While linear-PAzoMA showed monotonic changes
over time, cyclic-PAzoMA displayed nonmonotonic changes.
Moreover, for cyclic-PAzoMA, the photoinduced orientation
of azobenzene mesogens inscribed at room temperature could
not be enhanced by subsequent annealing in the nematic phase,
contrary to the linear polymers. All these results indicate that the
topological constraint in cyclic SCLCPs affects profoundly the
arrangement and organization of mesogenic side groups as well
as their orientation processes in response to an external stimulus
such as light. This finding points out that chain cyclization of
linear SCLCPs is a way to change the coupling of mesogenic
side groups with chain backbone; the interplay of the two
constituents under the additional topological constraint may be
explored to generate new properties or functions. More studies
are underway in our laboratory, and the results will be published
at due time.
(
3) Laurent, B. A.; Grayson, S. M. Chem. Soc. Rev. 2009, 38, 2202–2213.
(4) Bielawski, C. W.; Benitez, D.; Grubbs, R. H. Science 2002, 297,
2041–2044.
(5) Schappacher, M.; Deffieux, A. Science 2008, 319, 1512–1515.
(
6) Laurent, B. A.; Grayson, S. M. J. Am. Chem. Soc. 2006, 128, 4238–
239.
4
(
(
7) Xu, J.; Ye, J.; Liu, S. Macromolecules 2007, 40, 9103–9110.
8) Eugene, D. M.; Grayson, S. M. Macromolecules 2008, 41, 5082–5084.
(9) Hoskins, J. N.; Grayson, S. M. Macromolecules 2009, 42, 6406–
6413.
(10) Hu, J.; Zheng, R.; Wang, J.; Hong, L.; Liu, G. Macromolecules
2
11) Dong, Y.-Q.; Tong, Y.-Y.; Dong, B.-T.; Du, F.-S.; Li, Z.-C.
g
sn
ni
009, 42, 4638–4645.
(
Macromolecules 2009, 42, 2940–2948.
(12) Beaucage, G.; Kulkarni, A. S. Macromolecules 2010, 43, 532–537.
(13) Santangelo, P. G.; Roland, C. M.; Chang, T.; Cho, D.; Roovers, J.
Macromolecules 2001, 34, 9002–9005.
(14) Viras, K.; Yan, Z.-G.; Price, C.; Booth, C.; Ryan, A. J. Macro-
molecules 1995, 28, 104–109.
15) Percec, V.; Kawasumi, M. Macromolecules 1992, 25, 3851–3861.
16) Percec, V.; Asandei, A. D.; Chu, P. Macromolecules 1996, 29, 3736–
(
(
3
750.
(
17) Ringsdorf, H.; Schmidt, H. W. Makromol. Chem. 1984, 185, 1327–
1334.
(18) (a) Zhao, Y.; Bai, B.; Asatryan, K.; Galstian, T. Adv. Funct. Mater.
2
003, 13, 781–788. (b) Cui, L.; Zhao, Y.; Yavrian, A.; Galstian, T.
Macromolecules 2003, 36, 8246–8252.
19) Walther, M.; Faulhammer, H.; Finkelmann, H. Macromol. Chem.
(
Phys. 1998, 199, 223–237.
(
20) Ichimura, K. Chem. Rev. 2000, 100, 1847–1873.
(21) Natansohn, A.; Rochon, P. Chem. Rev. 2002, 102, 4139–4175.
(22) (a) Ikeda, T. J. Mater. Chem. 2003, 13, 2037–2057. (b) Yu, H.; Naka,
Y.; Shishido, A.; Ikeda, T. Macromolecules 2008, 41, 7959–7966.
(c) Yu, H.; Shishido, A.; Li, J.; Kamata, K.; Iyoda, T.; Ikeda, T. J. Mater.
Chem. 2007, 17, 3485–3488.
(
(
(
(
23) (a) Seki, T. Bull. Chem. Soc. 2007, 80, 2084–2109. (b) Uekusa, T.;
Nagano, S.; Seki, T. Macromolecules 2009, 42, 312–318.
24) Wang, D.; Ren, H.; Wang, X.; Wang, X. Macromolecules 2008, 41,
382–9388.
25) Yager, K. G.; Barrett, C. J. Curr. Opin. Solid State Mater. Sci.
001, 5, 487–494.
26) Zhao, Y., Ikeada, T., Eds.; Smart Light-Responsive Materials:
Azobenzene-Containing Polymers and Liquid Crystals; John Wiley:
Hoboken, NJ, 2009.
27) Wu, Y.; Zhang, Q.; Kanazawa, A.; Shiono, T.; Ikeda, T.; Nagase,
Y. Macromolecules 1999, 32, 3951–3956.
9
Acknowledgment. We acknowledge the financial support
from the Natural Sciences and Engineering Research Council
of Canada (NSERC) and le Fonds qu ꢀe b ꢀe cois de la recherche
sur la nature et les technologies of Qu ꢀe bec (FQRNT). Y.Z.
is a member of the FQRNT-funded Center for Self-Assembled
Chemical Structures.
2
(
(
28) Tong, X.; Cui, L.; Zhao, Y. Macromolecules 2004, 37, 3101–3112.
29) Yavrian, A.; Galstian, T.; Piche, M. Opt. Eng. 2002, 41, 852–855.
References and Notes
(
(30) Saad, B.; Galstyan, T.; Denariez-Roberge, M. M.; Dumont, M.
(1) Semlyen, J. A. Cyclic Polymers, 2nd ed.; Klumer Academic: Dordrecht,
The Netherlands, 2000.
Opt. Commun. 1998, 151, 235–240.