Are two azo groups better than one? Investigating the photoresponse of polymer-bisazobenzene complexes

Article abstract of DOI:10.1021/cm5023129

Azobenzene chromophores are an ideal choice for material applications where functionality needs to be activated in a precise remote-controlled fashion. The azobenzene stimuli-response falls into two categories, either based on efficient trans-to-cis photoisomerization and a high cis yield enabling on-off type functions, or relying on a fast trans-cis-trans cycling creating motion in the material system. Herein, we show that using bisazochromophores instead of the more common monoazobenzene derivatives makes a difference in the performance of light-responsive azopolymers, more specifically in photo-orientation and all-optical surface patterning. Our findings point out that polymer-bisazobenzene complexes are an attractive alternative as high-performance photoreponsive materials and that although their properties are highly sensitive to the extent of conjugation in the system, they can be designed into relatively transparent films with high performance for all-optical patterning.

Full text of DOI:10.1021/cm5023129

Article
pubs.acs.org/cm
Are Two Azo Groups Better than One? Investigating the
Photoresponse of Polymer-Bisazobenzene Complexes
Jaana Vapaavuori,*,†,‡,∥ Alexis Goulet-Hanssens,*,§,⊥,∥ Ismo T.S. Heikkinen,† Christopher J. Barrett,§
and Arri Priimagi†,#
†Department of Applied Physics, Aalto University, P. O. Box 13500, FI-00076 Aalto, Finland
‡Department of Chemistry, University of Montreal, C.P. 6128, Succursale Centre-Ville, Montreal, Quebec H3C 3J7, Canada
§Department of Chemistry, McGill University, 801 Sherbrooke Street West, Montreal, Quebec H3A 0B8, Canada
ABSTRACT: Azobenzene chromophores are an ideal choice
for material applications where functionality needs to be
activated in a precise remote-controlled fashion. The
azobenzene stimuli-response falls into two categories, either
based on efficient trans-to-cis photoisomerization and a high cis
yield enabling on−off type functions, or relying on a fast
trans−cis−trans cycling creating motion in the material system.
Herein, we show that using bisazochromophores instead of the
more common monoazobenzene derivatives makes a differ-
ence in the performance of light-responsive azopolymers, more
specifically in photo-orientation and all-optical surface
patterning. Our findings point out that polymer-bisazobenzene complexes are an attractive alternative as high-performance
photoreponsive materials and that although their properties are highly sensitive to the extent of conjugation in the system, they
can be designed into relatively transparent films with high performance for all-optical patterning.
deformations in films,12,13 long-range ordering in liquid
INTRODUCTION
■
crystalline environments,14−18 and changes to chirality.19,20
Stimuli-response is a critical feature in many smart materials,1
This demonstrates that one can still engineer systems exhibiting
and light is the best suited stimulus to deliver a localized, dosed
large material changes with classically suboptimal isomerization
input with high spatiotemporal control. Azobenzene has been
conditions. Controlling material properties using photoinduced
extremely prevalent in the literature as a reliable, reusable
motions driven by trans−cis−trans cycling is of interest for
photoswitch that can affect large molecular and macromolecular
fabricating dynamic photomechanical materials, such as light-
changes upon irradiation with light.2,3 Azobenzene′s reversible
powered oscillators,21 and for the inscription of light-induced
trans−cis photoisomerization yields a cis state that is metastable
surface-relief gratings (SRGs) for photonics and nanostructur-
and will thermally reconvert to the trans state. Functionalizing
ing applications.22 Improving both types of switching is highly
polymers with these chromophores provides materials with
interesting surface- as well as bulk-switching properties.4,5
desired, and a comprehensive understanding on the con-
nections between the azo chromophore, their environment, and
Chromophore incorporation can be achieved through various
means, either by covalently linking the chromophores to the
polymer chain, doping them in the system or making use of
supramolecular interactions to localize the azobenzene
the efficiency of photoinduced phenomena is critical for
effective use of azobenzene systems in materials applications.
Polyconjugated azobenzenes have been used as dyes that
molecules.6 To optimize the light-triggered work of azo
have a potential to achieve larger light-induced molecular
changes than an individual chromophore. These polyazos can
materials, both the chromophores and their attachment to a
achieve a high cis content if they are electronically decoupled
system need to be designed to optimize their desired function.
Achieving successful light-response in materials can rely
either by adding in a twist angle between neighboring dyes7,23
to prevent conjugation in the system, or changing the position
either on a system that can attain a high photostationary state
of polyazobenzene from a para−para orientation to a para−
(light pushing a chromophore to an ”on”/”off” state) or rapid
meta.24 In addition to loading more isomerizing units into a
cycling between isomers to drive material response (light
activating a material system to an ”on” state).4,7−9 Remarkable
material without aggregation, bisazo architectures can provide
larger volume change upon isomerization and the possibility to
changes in materials properties can be observed, even if the
separately address both azo groups25 compared to monoazos.
extent of isomerization is heavily hindered. For instance,
crystals of pure azobenzenes, bearing an ultrahigh optical
density can be bent with light.10,11 Likewise, brush architectures
and liquid-crystalline assemblies having large chromophore
content can be switched to impart large macromolecular
Received: June 26, 2014
Revised: August 14, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/cm5023129 | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
Although polyazo chromophores are extremely interesting as
drivers for photoinduced motions, there are few systematic
studies on the structure−property relations in bisazo systems.
Meng and co-workers demonstrated the importance of using
bisazochromophores in photo-orientation,26 and Wang and co-
workers made an extensive structure−property study of all-
optical surface patterning,27 both using covalent azopolymers
bearing bisazochromophore pendant units in the side chains. In
2010, using supramolecular hydrogen-bonded polyazo com-
plexes, Wu and co-workers reported on high and stable
birefringence values upon hydrogen bonding of an azo dye to
the side-chain of a polymer already containing an azo unit.28
The same year, bisazobenzene complexes hydrogen-bonded
directly to a poly(4-vinylpyridine) (p4VP) polymer backbone
being capable for efficient SRG formation were introduced,29
followed by optimization of the hydrogen-bonded bisazoben-
zene chromophore, affording SRGs that can be inscribed over
large range of wavelengths.30
EXPERIMENTAL SECTION
■
Synthesis of o-MHA and m-DY7. 1H NMR spectra were
acquired at 300 K, on a Varian-Mercury 300 or 400 MHz spectrometer
with 13C NMR spectra being acquired on a Varian-Mercury 300 MHz
NMR. Chemical shifts are reported in ppm on the δ scale using the
solvent signal for reference. High resolution mass spectrometry (HR-
MS) was acquired on a Thermo Scientific Exactive Plus Orbitrap.
Samples were ionized using either atmospheric pressure chemical
ionization (APCI) or electrospray ionization (ESI). All observed ions
in positive and negative ionization modes are reported. All chemicals
were obtained from Sigma-Aldrich Corporation and were used without
further purification.
Synthesis of o-MHA, 2-Methyl-4-(phenyldiazenyl)phenol.31 In a
100 mL round-bottom flask equipped with a stir bar, 15 mL of water
and 15 mL of acetone was set to cool in an ice bath. Once the solvent
temperature was below 5 °C, 456 μL of aniline (5 mmol, 1 equiv) was
added followed by 0.414 g of NaNO2 (6 mmol, 1.2 equiv) and 2.8 mL
of 5 M HCl added dropwise (14 mmol, 2.8 equiv). The solution was
left to stir for 30 min to diazotize the aniline, while in a second 250 mL
round-bottom flask equipped with a stir bar, 40 mL of water and 20
mL of acetone was set to cool in an ice bath. To this second flask, 515
μL of ortho-cresol (5 mmol, 1 equiv), 1.06 g of Na2CO3 (10 mmol, 2
equiv), and 0.8 g of NaOH (20 mmol, 4 equiv) were added and set to
stir in an ice bath to cool to below 5 °C. After the 30 min preactivation
of the aniline, the contents of the first flask were added slowly to the
second flask and left to stir for a further hour. After this time, the
reaction mixture was extracted with 150 mL of ethyl acetate. The red
organic phase was washed twice with 200 mL of water and then once
with 100 mL of brine. The organic phase was dried with MgSO4,
filtered, and evaporated. The crude product was purified by silica
column chromatography using a gradient of 0−30% ethyl acetate in
hexanes. Dark orange crystals. Yield: 0.148 g, 14%.
Exactly how bisazo chromophores affect material changes
remains unclear, as most systems explored to date are para-
oriented molecules (including Disperse Yellow 7 (DY7), Figure
1), which behave as one large conjugated photosystem.24 Given
1H NMR (300 MHz, chloroform-d) δ, ppm: 7.88 (d, J = 7.03 Hz, 1
H), 7.71−7.79 (m, 2 H), 7.41−7.53 (m, 3 H), 6.87 (d, J = 8.50 Hz, 1
H), 5.38 (br. s., 1 H), 2.34 (s, 3 H).
13C NMR (300 MHz, chloroform-d) δ, ppm: 156.8, 152.7, 146.9,
130.3, 129.1, 125.4, 124.7, 123.3, 122.5, 115.3, 15.9.
HR-MS (APCI, 4 kV) m/z: [M + H]+ calcd for C13H13N2O,
213.1022; found, 213.1020.
Synthesis of 1-(3-Nitrophenyl)-2-phenyldiazene.32 In a 30 mL vial
equipped with a stir bar, 0.365 g of nitrosoaniline (3.4 mmol, 1 equiv)
was added followed by 0.602 g of 3-nitroaniline (4.4 mmol, 1.3 equiv).
The blend was dissolved to a clear green solution with 15 mL of acetic
acid and set to stir for 7 days, as it slowly turned orange. The product
was purified by adding the acid solution to a separatory funnel with 80
mL of ethyl acetate and 300 mL of water. The red organic phase was
washed twice with 200 mL of water and then once with 100 mL of
brine. The organic phase was dried with MgSO4, filtered, and
evaporated. The crude product was purified by silica column
chromatography using a gradient of 0−30% ethyl acetate in hexanes.
Yellow crystals. Yield: 0.638 g, 83%.
Figure 1. Structures of the studied compounds.
the success of DY7 in SRG formation,29 we wanted to
investigate the benefits of breaking this conjugation and gaining
a system with two decoupled, potentially separately addressable
isomerizing units. To achieve this, we designed and synthesized
a new bisazo photosystem with the second azobenzene group
oriented meta to the first (m-DY7, Figure 1). The placement of
the second azobenzene in a meta position relative to the first in
m-DY7 splits the bisazobenzene structure into two electroni-
cally independent azo units. To help elucidate the behavior of
DY7 and m-DY7 and the consequences of this electronic
independence, ortho-methyl hydroxy azobenzene (o-MHA,
Figure 1) was used as a monoazo reference because the
structure of the bonding unit to p4VP is similar to o-MHA in
both bisazo dyes. We chose a noncovalent functionalization
strategy, employing a supramolecular hydrogen-bonding
approach to form complexes with the phenolic OH proton
and the electron lone-pair of pyridine in p4VP. This allows for
comparative studies on these chromophores which can all be
bound to the same polymer.29 Using this supramolecular
approach provides batch-to-batch reproducibility of molecular
weight and polydispersity among all the samples studied.
1H NMR (300 MHz, chloroform-d) δ, ppm: 8.69 (t, J = 2.05 Hz, 1
H), 8.28 (ddd, J = 8.12, 2.27, 1.02 Hz, 1 H), 8.19−8.25 (m, 1 H),
7.88−8.00 (m, 2 H), 7.66 (t, J = 8.05 Hz, 1 H), 7.49−7.58 (m, 3 H).
13C NMR (300 MHz, chloroform-d) δ, ppm: 152.9, 152.1, 149.0,
132.1, 129.9, 129.2, 129.2, 124.9, 123.3, 117.0.
HR-MS (APCI, 4 kV) m/z: [M]+ calcd for C12H9N3O2, 250.0587;
found, 250.0597.
Synthesis of 3-(phenyldiazenyl)aniline.33 In a 250 mL round-
bottom flask equipped with a stir bar, 0.5 g of 1-(3-nitrophenyl)-2-
phenyldiazene (2.2 mmol, 1 equiv) was added followed by 3.97 g of
Na2S·9H2O (16.5 mmol, 7.5 equiv). The blend was dissolved in 54 mL
of dioxane, 14 mL of ethanol and 3.5 mL of water and set to reflux at
90 °C for 4 h. The product was purified by adding the solution to a
separatory funnel with 100 mL of ethyl acetate and 300 mL of water.
The red organic phase was washed twice with 200 mL of water and
then once with 100 mL of brine. The organic phase was dried with
MgSO4, filtered, and evaporated. The crude product was purified by
B
dx.doi.org/10.1021/cm5023129 | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
Figure 2. (a) UV−vis spectra of DY7, m-DY7, and o-MHA in THF. (b) Fitting of the m-DY7 spectrum with two gaussians and comparison to the
spectra of o-MHA and 3-aminoazobenzene; spectra of (c) p(DY7), (d) p(m-DY7), and (e) p(o-MHA) in thin films in the dark and under
illumination of 365 and 488 nm. The strong peaks evident particularly in (c) result from scattering of the pump light into the spectrometer.
silica column chromatography using a gradient of 0−50% ethyl acetate
in hexanes. Red solid. Yield: 0.289 g, 67%.
13C NMR (300 MHz, chloroform-d) δ, ppm: 153.8, 152.7, 147.3,
130.9, 129.9, 129.1, 122.8, 117.9, 115.0, 107.4.
1H NMR (300 MHz, chloroform-d) δ, ppm: 7.87−8.02 (m, 2 H),
7.47−7.58 (m, 3 H), 7.41 (dt, J = 7.90, 1.17 Hz, 1 H), 7.33 (t, J = 7.75
Hz, 1 H), 7.20−7.28 (m, 1 H), 6.81 (ddd, J = 7.75, 2.34, 1.02 Hz, 1
H), 3.79 (br. s., 2 H).
HR-MS (ESI, 4 kV) m/z: [M + H]+ calcd for C12H12N3, 198.1026;
found, 198.1023.
Synthesis of m-DY7, 2-Methyl-4-((3-(phenyldiazenyl)phenyl)-
diazenyl)phenol.31 In a 100 mL round-bottom flask equipped with
C
dx.doi.org/10.1021/cm5023129 | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
a stir bar, 40 mL of water and 20 mL of acetone was set to cool in an
ice bath. Once the solvent temperature was below 5 °C, 0.289 g of 3-
(phenyldiazenyl)aniline (1.45 mmol, 1 equiv) was added followed by
0.12 g of NaNO2 (1.75 mmol, 1.2 equiv) and 2 mL of 2 M HCl was
added dropwise (4 mmol, 4 equiv). The solution was left to stir for 30
min to diazotize the aniline, while in a second 250 mL round-bottom
flask equipped with a stir bar, 40 mL of water and 20 mL of acetone
was set to cool in an ice bath. To this second flask, 150 μL of ortho-
cresol (1.45 mmol, 1 equiv), 0.120 g of Na2CO3 (3 mmol, 2.1 equiv),
and 0.318 g of NaOH (6 mmol, 4 equiv) were added and set to stir in
an ice bath to cool to below 5 °C. After the 30 min preactivation of the
3-(phenyldiazenyl)aniline, the contents of the first flask were slowly
added to the second flask and left to stir for a further hour. After this
time, the reaction mixture was extracted with 150 mL of ethyl acetate.
The red organic phase was washed twice with 200 mL of water and
then once with 100 mL of brine. The organic phase was dried with
MgSO4, filtered, and evaporated. The crude product was purified by
silica column chromatography using a gradient of 0−40% ethyl acetate
in hexanes. Dark orange solid. Yield: 0.184 g, 40%.
the transmission direction of the polarizer/analyzer, and the
birefringence values (Δn) were calculated from
π|Δn|d
λ
I = I0sin2
(1)
in which I stands for measured transmitted intensity, I0 for transmitted
intensity of the parallel polarizers setup prior to starting the writing
process, d for the thickness of the film, and λ for the wavelength of the
measuring beam.
Surface Patterning Studies. The surface-relief grating inscription
was performed using circularly polarized light with a wavelength of 488
nm from a Coherent Innova 70 Ar+ laser with an intensity of 300 mW/
cm2. The grating patterns were inscribed on the thin film samples
using a Lloyd mirror interferometer setup, in which half of the incident
beam hits directly the sample and the other half is reflected to the
surface from a mirror perpendicular to the sample. The period of the
resulting interference pattern can be tuned by changing the angle
between the inscription beam and the sample normal. In this study, an
angle of 13° was used in order to produce a grating with a 1 μm
period.
The formation of the SRGs was monitored in real time by
measuring the first-order diffraction efficiency of a horizontally
polarized probe beam from a 635 nm diode laser with a photodiode
as a function of time. The diffraction efficiency was calculated by
dividing the power of the diffracted beam by the power of the same
beam transmitted through the sample before the SRG inscription. The
surface profiles of the inscribed gratings were characterized using a
Veeco Dimension 5000 atomic-force microscope (AFM), with images
of the SRGs being acquired two weeks after inscription.
1H NMR (300 MHz, chloroform-d) δ, ppm: 8.40 (d, J = 2.05 Hz, 1
H), 7.93−8.06 (m, 3 H), 7.81 (s, 1 H), 7.77 (dd, J = 8.49, 2.34 Hz, 1
H), 7.65 (d, J = 7.90 Hz, 1 H), 7.44−7.60 (m, 3 H), 6.89 (d, J = 8.49
Hz, 1 H), 5.55 (br. s., 1 H), 2.34 (s, 3 H).
13C NMR (300 MHz, chloroform-d) δ, ppm: 157.0, 153.5, 153.4,
152.6, 146.9, 131.3, 129.6, 129.1, 129.1, 125.5, 125.1, 124.7, 124.6,
123.5, 123.0, 120.7, 116.0, 115.3, 15.9.
HR-MS (ESI, 4 kV) m/z: [M − H]− calcd for C19H15N4O,
315.1252; found, 315.1251.
Kinetic Studies in Solution. A known mass of all small molecules
were optically characterized by dissolving in a volumetric flask in dry
spectral grade THF. The absorbance spectra were recorded on Cary
Bio 300 UV/vis spectrophotometer. Solution-based kinetic data were
recorded on the same instrument after irradiating the sample with a
405 nm diode laser with an intensity of 11 mW/cm2 for 6 min, after
which time all samples had achieved their photostationary states. The
time-dependent rise in absorbance was tracked at the λmax of the trans
state, giving rise to a first order exponential growth of the trans state.
Preparation of the Thin Film Samples. Poly(4-vinylpyridine)
(p4VP, Polymer Source, Inc., Mn = 3200 g/mol, Mw = 3800 g/mol),
Disperse Yellow 7 (DY7, Sigma-Aldrich, 95%), as well as the
synthesized m-DY7 and o-MHA were separately dissolved in
tetrahydrofuran (THF, Sigma-Aldrich, >99.9%) and stirred for 2 h.
Polymer−azobenzene complexes p(DY7), p(m-DY7), and p(o-MHA),
with a nominal mixing ratio of 0.5 chromophores per polymer repeat
unit, were prepared by mixing the requisite amount of polymer and
dye stock solutions and stirring these complexed solutions for 2 h
before spin-coating. SRG inscription was performed on spin-coated
thin films (1000 rpm for 60 s) on clean microscope slides. UV−vis
studies were carried out on thin films spin-cast on quartz glass from 2
wt % complex solutions, whereas SRG studies were conducted with
samples from 4 wt % solutions. Film thicknesses were measured with a
DEKTAK 6 M surface profiler.
RESULTS AND DISCUSSION
■
UV−Visible Spectroscopy and Photoisomerization.
Figure 2a presents the UV−visible spectra of DY7, m-DY7,
and o-MHA in THF both before (all-trans state) and after
(photostationary state) illumination with a 405 nm laser. The
initial spectra of m-DY7 and o-MHA show absorption maxima
at 356 and 354 nm, respectively, being blue-shifted ∼35 nm as
compared to the absorption maximum of DY7, measured at
390 nm. This is due to the longer conjugation length in DY7,
which increases the delocalization of the π-electrons and thus
lowers the energy of the π−π* transition.34 Additionally, due to
the hypsochromic shift of the π−π* transition in m-DY7 and o-
MHA as compared to DY7, the low-intensity n-π* absorption
band around 440 nm is visible in these traces. Figure 2b shows
that the absorption spectrum of m-DY7 can be fitted with two
Gaussians, centered at 316 and 361 nm, indicating that the
meta-substitution geometry electronically decouples the two
azobenzenes units. By comparing the Gaussian fits to the
spectra of o-MHA (λmax 354 nm) and 3-aminoazobenzene (λmax
318 nm) as model compounds for the two azobenzenes
constituting m-DY7, we note that the Gaussian at 361 nm can
be assigned to the OH-capped azo and 316 nm to the H-
capped azo of the m-DY7. The absorbance at 440 nm results
from the combination of the n−π* transitions of both
azobenzene units.
UV−Vis Studies on Thin Films. The UV−vis spectra of thin films
were measured in the dark and under irradiation with a fiber-optic
spectrometer (Ocean Optics USB2000+). The illumination wave-
lengths were 365 nm (Thorlabs M365L2) and 488 nm (circularly
polarized light from Coherent Innova 70 Argon-ion laser). To study cis
isomer half-lives in the solid state, films were first pumped to the cis
state (365 nm) and the thermal cis−trans relaxation was monitored by
measuring the transmittance of a weak probe beam (390
5 nm)
In order to gain insight into the achievable photoconversion
yield of the chromophores in solution, we studied their spectral
changes upon irradiation (Figure 2a). The minimum cis content
can be obtained by taking the ratio of the λmax absorption of the
photostationary state over the all-trans state. This does not
preclude a residual cis absorption; however, this remains the
simplest method to compare solution and film data, and as
such, the photoconversion yield in all cases refers to a
minimum cis yield. The photostationary state of DY7 is
dissimilar compared to those of m-DY7 and o-MHA, in which a
obtained from a band-pass filtered, fiber-coupled Xenon light source
(Abet Technologies LS150). To improve the signal-to-noise ratio, the
probe was chopped at 200 Hz frequency and data was collected
through a lock-in amplifier (Stanford Research Systems Model SR830
DSP).
Birefringence Studies. Photo-orientation was performed by
exciting the chromophores with linearly polarized 365 and 488 nm
light and measured by following the transmission of 820 nm diode
laser through a polarizer−sample−analyzer configuration. The
direction of the pump beam was set to 45 degs with respect to
D
dx.doi.org/10.1021/cm5023129 | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
significant decrease in the π−π* transition (minimum photo-
conversion yields being 78% and 96%, respectively) and a
concomitant rise in absorbance of the n−π* transition at
around 445 nm is observed. The spectral characteristics of the
latter two are similar to the photochemical properties of the 4-
hydroxy-substituted azobenzenes, as reported previously by
Brode and colleagues.35 Conversely, the photostationary state
of DY7 simply decreases the intensity of π−π* absorption band
at 390 nm (by 16%) with no evident changes in the shape of
the spectrum.
In terms of potential photonic applications, it is pertinent to
study the photochemical properties of the azobenzene units in
thin polymeric films. Hydrogen bonding of alcohol-function-
alized azobenzenes to the amine lone-pair electrons of poly(4-
vinylpyridine) is an easy and versatile way of controlling
azobenzene doping in the solid state without harmful
chromophore−chromophore aggregation destroying the pho-
tonic properties.6,29,36−39 On the basis of earlier studies, a
mixing ratio of 50 mol % was selected as this ratio combines the
high mass percentage of photoactive units in the material while
still providing enough free volume, prohibiting hindered
isomerization.29 Figures 2c−e show the spectra of the thin
films of the three studied compounds as complexed to p4VP
and the spectral changes under illumination with 365 or 488
nm light. Although the spectral maxima have red-shifted a few
nanometers and broadened (Table 1), similar initial spectral
photoconversion yield of 13%. In general, we relate this low
photoconversion, observed both in thin films and in solution, to
the reduced quantum yield seen in para-conjugated bisazo-
benzenes.24
In measuring the cis-state decay in THF solution (Table 1),
we observed that the half-life of DY7 is on a time scale of
minutes, whereas the cis half-lives of o-MHA and m-DY7 are
measured in hours, indicating that the cis isomer for the latter
dyes is much more stable than that of DY7. These thermal
decay kinetics in solution are of first order for all dyes, including
the bisazomolecules DY7 and m-DY7, suggesting that there are
no thermal isomerizations of both azo ends occurring at the
same time. Given the qualitatively similar spectral changes upon
irradiation in m-DY7 and o-MHA in THF as well as between
the thin films of p(m-DY7) and p(o-MHA), the photo-
conversion in m-DY7 is likely to occur predominantly in the
OH-capped azo portion of the molecule. Although the
isomerizing azo unit is the same for m-DY7 and o-MHA, the
shorter cis lifetime of the former can be related to more
sterically hindered cis state of the m-DY7 as compared to o-
MHA. When the supramolecular complex of these hydrogen-
bonded dyes are spin-cast and annealed to a solid state, we
observe a shortening of the cis lifetimes of all studied
compounds. Individually, the effect of isomerization in a
solid-state polymer matrix is much more drastic in o-MHA and
m-DY7, bringing the cis lifetimes to the same order of
magnitude with DY7 due to more constrained polymer
environment surrounding them.
Material Testing. Photo-Orientation. Illuminating an azo-
containing material with linearly polarized light leads to the
buildup of photoinduced anisotropy, rendering the initially
anisotropic materials birefringent.40 Figure 3a presents the
photoinduced birefringence as a function of time, the relaxation
of the induced orientation after turning off the illumination, as
well as the rerandomization of the molecular order under
circularly polarized pump for all the studied complexes. Table 2
summarizes the obtained birefringence values and their long-
term stability obtained by dividing the relaxed value by the
saturated birefringence. Similar to spectroscopic studies, photo-
orientation behavior of p(m-DY7) and p(o-MHA) is highly
dependent on the irradiation wavelength as demonstrated by
Figure 3b for p(m-DY7), whereas for p(DY7), birefringence
can be inscribed in a similar qualitative fashion at both
wavelengths.
Table 1. Half-lives T1/2, Maximum Absorption Wavelengths
λmax, and Minimum Photoconversion Yields γ for the
Chromophores in THF and Polymer−Chromophore
a
Complexes in Thin Films
chromophore
DY7
m-DY7
o-MHA
T1/2 in THF (min)
λmax in THF (nm)
5.83
390
162.9
356
266.6
354
T
1/2, thin film (min)
2.60
391
8.80
361
6.50
361
λmax, thin films (nm)
γ405(%) in THF
15.5
13.3
12.6
78.0
38.0
17.9
95.9
63.3
17.1
γ365 (%), thin film
γ488 (%), thin film
a
Minimum photoconversion yield is calculated from the maximum
trans content at the photostationary state at a given irradiation
wavelength.
The birefringence results can be rationalized as follows: as
opposed to p(DY7), two different mechanisms producing
photoinduced anisotropy can be identified for p(m-DY7) and
p(o-MHA) depending on the irradiating wavelength. This is
particularly evident in Figure 3b, which shows the difference in
build-up for photo-orientation for p(m-DY7) when illuminated
at 365 and at 488 nm. Illuminating that sample at 365 nm
produces a high cis-isomer yield and, at the same time, does not
drive efficient photoisomerization back to the all-trans state. A
linearly polarized writing beam, thus, results in angularly
dependent cis-concentration in the film (angular hole
burning),41,42 observed as relatively low saturated birefringence
values and the fast saturation. In contrast, irradiation with 488
nm drives the photoisomerization in both directions for all the
complexes, leading gradually to orientation of the molecular
axis perpendicular to pump polarization resulting in higher
birefringence values than the angular hole burning mecha-
nism.41,42 The wavelength-dependence of the dominant
mechanism for p(m-DY7) and p(o-MHA) is evident also
features as observed in THF are conserved in the solid state.
We note that the minima at 365 nm and 488 nm in Figures 2c−
e result from scattering of the pump light into the detector from
the sample surface.
The solid-state absorption spectra under illumination with
365 and 488 nm further highlight the differences of p(m-DY7)
and p(o-MHA) as compared to p(DY7). As the former systems
show separated π−π* and n−π* bands, we can address these
transitions separately with 365 or 488 nm light. Figures 2d and
e illustrate the photoconversion yields obtained for p(m-DY7)
and p(o-MHA). When illuminating at 365 nm, although lower
than when illuminating by 405 nm in THF solution, the
photoconversion remains significant, whereas for 488 nm
illumination the photoconversion yield remains below 20%.
This demonstrates that for those two systems, we can easily
tweak the photoinduced properties by changing the wavelength
of the photoisomerizing light. As illustrated in Figure 2c, the
spectral changes in p(DY7) are independent of the irradiation
wavelength, both 365 and 488 nm result in a modest
E
dx.doi.org/10.1021/cm5023129 | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
Figure 3. Normalized birefringence for thin films of (a) p(DY7), p(m-DY7), and p(o-MHA) inscribed at 488 nm light and for (b) p(m-DY7) at
inscribed both at 365 and 488 nm light.
Table 2. Saturated Birefringence Values Δn and Their Long-
Term Stability S in Thin Films of Polymer-Azobenzene
Complexes of Thickness d As Inscribed at Different
Wavelengths
light, accentuating the fact that it is the trans−cis−trans cycling,
as opposed to trans−cis isomerization, that drives the photo-
orientation in amorphous azopolymers. When comparing the
saturated birefringence values, they progress in the order of
p(DY7) > p(m-DY7) > p(o-MHA). The fact that p(DY7),
consisting of the rod-like DY7 molecules, results in the highest
birefringence, which is over 2-fold the value of p(m-DY7) and
over 4-fold the value of p(o-MHA), is to be expected based on
its large aspect ratio.26 Also, the larger size of m-DY7
chromophores as compared to o-MHA results in larger
birefringence values. In this sense, using bisazobenzene
chromophores provides additional advantage to photo-orient-
able materials.
sample
d (nm)
Δn488
S488 (%)
Δn365
S365 (%)
p(DY7)
195
140
170
0.035
0.014
0.008
82
73
59
0.017
0.007
0.004
74
47
21
p(m-DY7)
p(o-MHA)
based on the remnant birefringence values, being significantly
less stable upon 365 nm irradiation, whereas the stability of
p(DY7) is rather pump wavelength independent.
The drop in the measured birefringence when irradiation is
ceased (60s in Figure 3a) is due to (i) thermal back
isomerization and (ii) thermal randomization of the anisotropic
molecular alignment in the initially amorphous polymer
systems. Because cis half-life for all the studied molecules in
the solid state are similar, the long-term stability values
measured on the order of several minutes remain comparable.
Hence, it is clear that the angular diffusion of the
Upon irradiation with 488 nm, the build-up dynamics for
birefringence is rather similar for all the compounds, rendering
the comparison of photo-orientation efficiency more mean-
ingful. We note that at this wavelength, the absorption
coefficient of the materials is low, and even for p(DY7), the
majority of the pump light propagates through the sample. It is
interesting that for all the samples, this poorly absorbed pump
light gives rise to higher birefringence than the 365 nm pump
Figure 4. First-order diffraction efficiency upon SRG inscription at 488 nm light (a) as a function of time and (b) as a function of absorbed energy
for thin films of polymer−azobenzene complexes.
F
dx.doi.org/10.1021/cm5023129 | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
bisazomolecules in p(DY7) and in p(m-DY7) upon thermal
cis−trans relaxation is much lower than for the p(o-MHA). This
is also supported by the fact that the buildup of birefringence
for p(o-MHA) is remarkably faster than for the bisazobenzene
complexes.
but comprising two azobenzene units, is shown to provide the
most efficient SRG formation in this normalized comparison,
indicating that a given light flux creates the largest material
motions.
The higher SRG formation efficiency of p(m-DY7) as
compared to p(o-MHA) can be rationalized by at least two
factors. First, although it is in principle the same hydroxy-
capped azo that is isomerizing in both m-DY7 and o-MHA,
there is a possibility for the m-aminoazo head to act as an
energy donor upon excitation, effectively increasing the
isomerization yield of the hydroxyazobenzene of the
bisazobenzene.9,43 Second, the free-volume sweep upon
photoisomerization of m-DY7 is larger than that of o-MHA
creating more motion into the material. This follows previous
reports that bulkier azobenzenes result in more efficient SRG
formation in amorphous azo-polymers44 and in molecular
glasses.45
Meng and co-workers, as well as Cojocariu and co-workers
linked the higher stability as well as slower buildup of
birefringence in bisazochromophores to the idea that
concurrent photoisomerization of both azo bonds is required
to produce any motion of the chromophore with respect to its
initial orientation.26,34 Our results, as well as the results of
Bleg
́
er and co-workers,7,23 challenge this hypothesis because
driving the bisazomolecules, especially the conjugated ones
such as DY7 to cis−cis state, is very demanding. We address the
higher stability (as well as the slower build-up kinetics) of the
bisazochromophores to the steric hindrances that the longer-
axis bisazo molecules face when reorienting upon thermal
isomerization.
Most commonly, in SRG-forming amorphous azopolymers,
push−pull-type azobenzenes with high absorption at blue-green
wavelengths and high overlap of the π−π* and n-π* bands are
used.22 p(m-DY7) and p(o-MHA), having well-separated π−π*
and n−π* bands, are rather unconventional for SRG formation,
yet they confirm that it is not the extent of initial absorption,
but efficient trans−cis−trans cycling that is key to efficient SRG
formation in the context of amorphous materials.46 Because the
initial absorption at writing wavelength, 488 nm, is already
lower than at the π−π* absorption, the thickness of the films is
not an issue and the films can, for example, be easily
manipulated from the backside if needed. Although p(DY7)
is more efficient in terms of absolute SRG formation efficiency,
in terms of how efficiently the absorbed photons are transferred
into mass transport of the polymer, p(m-DY7) is clearly
superior to the conjugated p(DY7). We link this reduced
performance to the lower photoconversion observed in both
the solution state and thin films in these dyes.
All-Optical Surface Patterning. When irradiating an
azopolymer thin film with an interference pattern of two
circularly polarized laser beams, the film deforms and forms a
surface-relief pattern with the periodicity dictated by the light
interference pattern. Figure 4a illustrates the growth of the first-
order diffraction efficiency of the gratings upon SRG inscription
at 488 nm. The original film thickness as well as final
modulation depth of the gratings after an illumination of 1800 s
at 300 mW/cm2 are listed in Table 3. The SRG formation
Table 3. Obtained SRG Modulation Depths Λ after the
Irradiation of 1800 s at 488 nm for Thin Films of Polymer−
Azobenzene Complexes of Thickness d
sample
d (nm)
Λ (nm)
p(DY7)
340
550
620
315
250
170
p(m-DY7)
p(o-MHA)
efficiency at 488 nm develops in the order of p(DY7) > p(m-
DY7) > p(o-MHA), the modulation depths of the gratings
being 315, 250, and 170 nm, respectively. We note, however,
that because p(DY7) has a much higher extinction coefficient at
488 nm than p(m-DY7) and p(o-MHA), the SRG formation in
p(DY7) is driven by a remarkably larger number of photon
absorptions than in the other studied materials.
A better way to consider SRG formation efficiency between
materials with different optical densities is shown in Figure 4b,
in which diffraction efficiency is plotted as a function of
absorbed energy by the surface layer of the material. The
energy, dE, absorbed during the time t, by a differentially thin
layer of material, dz, having an area of A and when illuminated
with an intensity Ibeam, can be expressed by the equation
CONCLUSIONS
■
We have demonstrated the power of rational molecular design
for the buildup of fundamental understanding on light-
responsive polymer materials. By varying between meta- and
para-substitution of the second azobenzene in bisazobenzene
molecules we show drastic differences in isomerization and
thermal reconversion. In a materials context, namely, in the
hydrogen-bonded polymer-azobenzene complexes, meta-sub-
stitution allows the two linked molecules to act independently
allowing them to be addressed at different wavelengths, whereas
the para-linkage causes the molecules to act as one large azo.
Highly conjugated para-substituted bisazobenzenes offer
long aspect ratios and, thus, are seen as preferable for photo-
orientation as compared to analogous monoazo derivatives.
Changing the position of the additional azo unit to the meta-
position, results in more efficient SRG formation than the
corresponding monoazo. The photoinduced motions in
azopolymers are dictated by the properties of the photoactive
units and their interactions with the local environment. The
supramolecular polymer-meta-bisazobenzene complexes are an
attractive alternative for optimizing macromolecular motions in
a manner that is far more optically transparent than current
systems allowing for further progress in both the design of
molecular chromophores as well as materials containing these
switches.
dE
dz
= αtAIbeam
(2)
in which α denotes the absorption coefficient of the material at
the wavelength in question. Figure 4b shows the diffraction
efficiency of all three complexes as a function of absorbed light,
which gives the absorption-normalized SRG formation
efficiency, and shows that in this comparison, p(DY7) is in
fact the least efficient from the studied complexes. This is
different from the non-normalized time-domain measurement
(Figure 4a), in which p(DY7) was shown to be the most
efficient. p(m-DY7), sharing the same azo unit than p(o-MHA),
G
dx.doi.org/10.1021/cm5023129 | Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
(23) Bleg
́
er, D.; Liebig, T.; Thiermann, R.; Maskos, M.; Rabe, J. P.;
AUTHOR INFORMATION
■
Hecht, S. Angew. Chem., Int. Ed. 2011, 50, 12559−12563.
(24) Cisnetti, F.; Ballardini, R.; Credi, A.; Gandolfi, M. T.; Masiero,
S.; Negri, F.; Pieraccini, S.; Spada, G. P. Chem.Eur. J. 2004, 10,
2011−2021.
Corresponding Authors
*E-mail: jaana.vapaavuori@umontreal.ca.
*E-mail: alexis.goulet-hanssens@mail.mcgill.ca.
(25) Robertus, J.; Reker, S. F.; Pijper, T. C.; Deuzeman, A.; Browne,
W. R.; Feringa, B. L. Phys. Chem. Chem. Phys. 2012, 14, 4374−4382.
(26) Meng, X.; Natansohn, A.; Rochon, P. Polymer 1997, 38, 2677−
2682.
(27) Wang, X.; Yin, J.; Wang, X. Polymer 2011, 52, 3344−3356.
(28) Wu, S.; Duan, S.; Lei, Z.; Su, W.; Zhang, Z.; Wang, K.; Zhang,
Q. J. Mater. Chem. 2010, 20, 5202−5209.
(29) Vapaavuori, J.; Priimagi, A.; Kaivola, M. J. Mater. Chem. 2010,
20, 5260−5264.
Present Addresses
⊥Department of Chemistry, Humboldt-Universitat zu Berlin,
̈
Brook-Taylor-Straße 2, 12489 Berlin, Germany.
#Department of Chemistry and Bioengineering, Tampere
University of Technology, P.O. Box 541, FI-33101 Tampere,
Finland.
Author Contributions
∥J.V. and A.G.H. contributed equally to this work.
(30) Koskela, J. E.; Vapaavuori, J.; Hautala, J.; Priimagi, A.; Faul, C. F.
J.; Kaivola, M.; Ras, R. H. A. J. Phys. Chem. C 2012, 116, 2363−2370.
(31) Leriche, G.; Budin, G.; Brino, L.; Wagner, A. Eur. J. Org. Chem.
2010, 2010, 4360−4364.
Notes
The authors declare no competing financial interest.
(32) Asanuma, H.; Liang, X.; Komiyama, M. Tetrahedron Lett. 2000,
41, 1055−1058.
ACKNOWLEDGMENTS
■
(33) Samanta, S.; Beharry, A. A.; Sadovski, O.; McCormick, T. M.;
Babalhavaeji, A.; Tropepe, V.; Woolley, G. A. J. Am. Chem. Soc. 2013,
135, 9777−9784.
J.V. is grateful for postdoctoral grant funded by Emil Aaltonen
foundation. A.G.H. acknowledges FQRNT for a B2 doctoral
scholarship as well as a B3 postdoctoral fellowship. A.P.
acknowledges the financial support of Emil Aaltonen
foundation. This work made use of the Aalto University
Nanomicroscopy Center (Aalto-NMC) premises.
(34) Cojocariu, C.; Rochon, P. Macromolecules 2005, 38, 9526−9538.
(35) Brode, W.; Gould, J.; Wyman, G. J. Am. Chem. Soc. 1952, 74,
4641−4646.
(36) Priimagi, A.; Cattaneo, S.; Ras, R. H. A.; Valkama, S.; Ikkala, O.
Chem. Mater. 2005, 17, 5798−5802.
(37) Priimagi, A.; Kaivola, M.; Rodriguez, F. J.; Kauranen, M. Appl.
Phys. Lett. 2007, 90, 121103.
REFERENCES
■
(1) Theato, P.; Sumerlin, B. S.; O’Reilly, R. K.; Epps, T. H., III Chem.
Soc. Rev. 2013, 42, 7055−7056.
(38) Priimagi, A.; Vapaavuori, J.; Rodriguez, F. J.; Faul, C. F. J.;
Heino, M. T.; Ikkala, O.; Kauranen, M.; Kaivola, M. Chem. Mater.
2008, 20, 6358−6363.
(2) Dugave, C.; Demange, L. Chem. Rev. 2003, 103, 2475−2532.
(3) Russew, M.-M.; Hecht, S. Adv. Mater. 2010, 22, 3348−3360.
(4) Mahimwalla, Z.; Yager, K. G.; Mamiya, J.-I.; Shishido, A.;
Priimagi, A.; Barrett, C. J. Polym. Bull. 2012, 69, 967−1006.
(5) Yu, H.; Ikeda, T. Adv. Mater. 2011, 23, 2149−2180.
(6) Priimagi, A.; Kaivola, M.; Virkki, M.; Rodríguez, F. J.; Kauranen,
M. J. Nonlinear Opt. Phys. Mater. 2010, 19, 57−73.
(39) Vapaavuori, J.; Valtavirta, V.; Alasaarela, T.; Mamiya, J.-I.;
Priimagi, A.; Shishido, A.; Kaivola, M. J. Mater. Chem. 2011, 21,
15437−15441.
(40) Natansohn, A.; Rochon, P. Chem. Rev. 2002, 102, 4139−4175.
(41) Sekkat, Z.; Wood, J.; Knoll, W. J. Phys. Chem. 1995, 17226−
17234.
(7) Bleg
Hecht, S. J. Phys. Chem. B 2011, 115, 9930−9940.
(8) Bleger, D.; Yu, Z.; Hecht, S. Chem. Commun. 2011, 47, 12260−
12266.
(9) Yu, Z.; Hecht, S. Angew. Chem., Int. Ed. 2013, 52, 13740−13744.
(10) Bushuyev, O. S.; Tomberg, A.; Frisc c, T.; Barrett, C. J. J. Am.
Chem. Soc. 2013, 135, 12556−12559.
́
er, D.; Dokic, J.; Peters, M. V.; Grubert, L.; Saalfrank, P.;
(42) Dumont, M.; El Osman, A. Chem. Phys. 1999, 245, 437−462.
(43) Bandara, H. M. D.; Burdette, S. C. Chem. Soc. Rev. 2012, 41,
1809−1825.
(44) Barrett, C. J.; Natansohn, A. L.; Rochon, P. L. J. Phys. Chem.
1996, 3654, 8836−8842.
(45) Wang, X.; Yin, J.; Wang, X. Langmuir 2011, 27, 12666−12676.
(46) Goldenberg, L. M.; Gritsai, Y.; Stumpe, J. J. Opt. 2011, 13,
075601.
́
̌ ̌ ́
i
(11) Bushuyev, O. S.; Singleton, T. A.; Barrett, C. J. Adv. Mater.
2013, 25, 1796−1800.
(12) Hosono, N.; Kajitani, T.; Fukushima, T.; Ito, K.; Sasaki, S.;
Takata, M.; Aida, T. Science 2010, 330, 808−811.
(13) Ikeda, T.; Ube, T. Mater. Today 2011, 14, 480−487.
(14) Lee, K. M.; Koerner, H.; Vaia, R. A.; Bunning, T. J.; White, T. J.
Soft Matter 2011, 7, 4318−4324.
(15) Lee, K. M.; Smith, M. L.; Koerner, H.; Tabiryan, N.; Vaia, R. A.;
Bunning, T. J.; White, T. J. Adv. Funct. Mater. 2011, 21, 2913−2918.
(16) Li, W.; Dohi, T.; Hara, M.; Nagano, S.; Haba, O.; Yonetake, K.;
Seki, T. Macromolecules 2012, 45, 6618−6627.
(17) Haque, H. A.; Kakehi, S.; Hara, M.; Nagano, S.; Seki, T.
Langmuir 2013, 29, 7571−7575.
(18) Iamsaard, S.; Aßhoff, S. J.; Matt, B.; Kudernac, T.; Cornelissen,
J. J. L. M.; Fletcher, S. P.; Katsonis, N. Nat. Chem. 2014, 6, 229−235.
(19) Takaishi, K.; Muranaka, A.; Kawamoto, M.; Uchiyama, M. J. Org.
Chem. 2011, 76, 7623−7628.
(20) Takaishi, K.; Muranaka, A.; Kawamoto, M.; Uchiyama, M. Org.
Lett. 2012, 14, 276−279.
(21) White, T. J.; Tabiryan, N. V.; Serak, S. V.; Hrozhyk, U. A.;
Tondiglia, V. P.; Koerner, H.; Vaia, R. A.; Bunning, T. J. Soft Matter
2008, 4, 1796−1798.
(22) Priimagi, A.; Shevchenko, A. J. Polym. Sci., Part B: Polym. Phys.
2014, 52, 163−182.
H
dx.doi.org/10.1021/cm5023129 | Chem. Mater. XXXX, XXX, XXX−XXX