Blue-light-induced processes in a series of azobenzene poly(ester imide)s

Article abstract of DOI:10.1016/j.jphotochem.2017.07.047

We report on a series of photoresponsive azobenzene poly(ester imide)s thermally converted from their precursor poly(ester amic acid)s. Thermal properties as well as ability for efficient photoinduced chromophore orientation and polymer mass transport under linearly polarized blue irradiation were investigated, focusing on the effect of azo chromophore location within polymer chain. The amount of chromophore alignment studied in the experiment on photoinduced birefringence differed by a factor of l.5 between the materials. Despite the fact that the generated birefringence was of the same order of magnitude, very large differences in efficiency of surface relief grating formation were seen. Specific features of photoresponsive behaviour of the polymers were shown to correlate with the presence or the absence of intrachain C[sbnd]H?π interaction between azobenzene hydrogen and aromatic imide structure, which was revealed by density functional theory (DFT) calculations. Advantageous chromophore location next to the imide groups for effective light-driven processes was shown.

Full text of DOI:10.1016/j.jphotochem.2017.07.047

Accepted Manuscript
Title: Blue-light-induced processes in a series of azobenzene
poly(ester imide)s
Authors: Anna Kozanecka-Szmigiel, Jolanta Konieczkowska,
Dariusz Szmigiel, Jerzy Antonowicz, Jan Małecki, Ewa
Schab-Balcerzak
PII:
S1010-6030(17)30911-5
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.jphotochem.2017.07.047
JPC 10773
To appear in:
Journal of Photochemistry and Photobiology A: Chemistry
Received date:
Revised date:
Accepted date:
28-6-2017
30-7-2017
31-7-2017
Please cite this article as: Anna Kozanecka-Szmigiel, Jolanta Konieczkowska, Dariusz
Szmigiel, Jerzy Antonowicz, Jan Małecki, Ewa Schab-Balcerzak, Blue-light-induced
processes in a series of azobenzene poly(ester imide)s, Journal of Photochemistry and
Photobiology A: Chemistryhttp://dx.doi.org/10.1016/j.jphotochem.2017.07.047
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Blue-light-induced processes in a series of azobenzene poly(ester imide)s
Anna Kozanecka-Szmigiel1*, Jolanta Konieczkowska2, Dariusz Szmigiel3, Jerzy Antonowicz1,
Jan Małecki4, Ewa Schab-Balcerzak2
1Faculty of Physics, Warsaw University of Technology, 75 Koszykowa Str., 00-662 Warszawa,
Poland
2Centre of Polymer and Carbon Materials, Polish Academy of Sciences, 34 M. Curie-
Sklodowska Str., 41-819 Zabrze, Poland
3Institute of Electron Technology, al. Lotnikow 32/46, 02-668 Warszawa, Poland
4Institute of Chemistry, University of Silesia, Szkolna 9, 40-006 Katowice, Poland
1

Highlights
Three poly(ester imide)s with different azobenzene location in backbone.
Similar magnitude of light-induced birefringence in all the materials.
Very different efficiency in light-induced polymer mass transport.
Intrachain C-H… interaction influences chromophore ability for photoisomerization.
2

Abstract
We report on a series of photoresponsive azobenzene poly(ester imide)s thermally converted
from their precursor poly(ester amic acid)s. Thermal properties as well as ability for efficient
photoinduced chromophore orientation and polymer mass transport under linearly polarized
blue irradiation were investigated, focusing on the effect of azo chromophore location within
polymer chain. The amount of chromophore alignment studied in the experiment on
photoinduced birefringence differed by a factor of l.5 between the materials. Despite the fact
that the generated birefringence was of the same order of magnitude, very large differences in
efficiency of surface relief grating formation were seen. Specific features of photoresponsive
behaviour of the polymers were shown to correlate with the presence or the absence of
intrachain C-H… interaction between azobenzene hydrogen and aromatic imide structure,
which was revealed by density functional theory (DFT) calculations. Advantageous
chromophore location next to the imide groups for effective light-driven processes was
shown.
Keywords: azo polymers; azo polyimides; trans-cis photoisomerization; photoinduced optical
anisotropy; surface relief grating
* corresponding author. Tel.: +48 22 234 5230; fax +48 226 282 171.
E-mail address: annak@if.pw.edu.pl
1. Introduction
Azobenzene polymers are very important light-responsive materials exhibiting
photoinduced orientation [1,2], photoinduced mass transport [3-6] or photomechanical effects
[7,8] upon irradiation with polarized light. In spite of extensive investigations carried out
during the past two decades, azo polymers still fascinate and are the subject of either
fundamental considerations or practical research focused on their use in optical data storage
and processing, nanofabrication, nanoscale machines, photoalignment of liquid crystals or
optical power limiting [9-14].
The photoresponsive behaviour of azo polymers arises from clean and reversible trans-cis
photoisomerization reactions of the azobenzene derivatives followed by molecular motions
appearing at different length scales [15]. Under linearly polarized irradiation of a proper
wavelength, multiple trans-cis-trans convertions are accompanied by the gradual
reorientation of the azo dyes into positions perpendicular to the polarization direction. The
process gives rise to photoinduced optical anisotropy (POA) i.e., dichroism and birefringence
3

evidenced as differences in material absorbance and refractive index depending on a
polarization direction of the passing light. Moreover, efficient trans-cis-trans cycling upon
exposure to interference pattern of light may result in macroscopic motion of the polymeric
material over micron-distances. The motion produces stable sinusoidal deformations, i.e.,
surface relief gratings (SRG), on the initially flat polymer layer.
The efficiency of photo-orientation and mass transport processes is strongly related to
the azo polymer structure and may achieve very different values: extraordinary large,
moderate or even non-detectable, what may be a surprising observation [16,17]. The
relationships between the structural features and the optical response are known to be very
complicated due to complexity of polymeric materials. Moreover, performing the comparative
studies between various polymeric systems is non-trivial due to difficulty in separating a
certain structural detail as a single factor affecting the photoinduced process.
Extensive research on amorphous azo polymers differing in chemical structure, glass
transition temperature or azobenzene content allowed for establishing some general rules
governing the efficiency or dynamics of the light-induced phenomena [15,18]. Nevertheless,
the current knowledge appears to be insufficient in predicting or explaining the
photoresponsive behaviour of many polymeric azo systems. Among them are azo polyimides,
which possesses a complex chemical architecture of a repeating unit.
Polyimides are a group of polymers exhibiting excellent thermal stability, high
mechanical strength and high chemical resistance [19,20]. They offer attractive hosts for
functionalized or supramolecular azo systems [21,22]. In contrast to flexible main-chain
polymers, azobenzene polyimides show a lower efficiency of photoinduced processes [15,
23]. Nevertheless, a temporal stability of the generated photoalignment and a thermal stability
of the inscribed gratings in azo polyimides are often significantly larger than in other families
of azo polymers. The long-term stability of the POA makes the azo polyimides very attractive
candidates in the area of liquid crystal alignment [20,24], while the stability of SRG at high
temperatures allows for avoiding an additional grating fixation treatment [25-27]. Azo
polyimides however, usually exhibit poor solubility, which leads to some difficulties in their
processing. This probably comprises one of the reasons for significantly less literature reports
on azo polyimides than on other classes of macromolecules. This also results in worse
understanding of the role of particular structural feature in enabling or restricting the light-
driven molecular motions.
Continuing our efforts to obtain processable azo polyimides exhibiting an enhanced
temporal and thermal stability of the photoinduced properties, we present here a series of
4

amorphous side-chain poly(ester imide)s possessing non-substituted azobenzene
chromophores attached to the backbone via one of their phenyl rings. We aim in getting an
insight into relationships between the azobenzene location within the backbone and the
efficiency of light-induced molecular orientation and polymer chain movement. We involve
DFT calculations in order to find possible reasons for the observed differences in
photoresponsive behaviour of the materials.
2. Materials and Methods
2.1. Materials
Resorcinol, m-phenylenediamine, trimelitic anhydride chloride (TMACl), N-methyl-2-
pyrolidone (NMP), tetrahydrofuran (THF) were purchased from Sigma-Aldrich Chemical Co.
Methanol, pyridine, aniline were purchased from Avantor (POCH). Detailed synthesis and
characterization of 2,4-diaminoazobenzene (denoted as diamine DA(H)) and
diesterdianhydride without azobenzene group (denoted as anhydride DB) have been reported
previously [28,29].
2.2. Synthesis of 2,4-dihydroxyazobenzene
The 2,4-dihydroxyazobenzene was synthesized according to the procedure published
previously [30]. Aniline (30 mmol) was dissolved in the mixture of deionized water (40 ml)
and concentrated HCl (18 ml). The solution was stirred in an ice bath (0-5 oC) and a solution
of sodium nitrate (30 mmol) in 5 ml of deionized water was added by dropping. Next, a
solution of resorcinol (30 mmol) in 50 ml of methanol has been added dropwise to the
solution of aniline. The mixture was stirred for 30 minutes. Finally, the solution was
neutralized by two portions of sodium acetate (42 mmol per portion) added in 30-min
intervals. The mixture was further neutralized to pH=6-7 by adding 10 % solution of sodium
hydroxide. The product was filtrated and washed with deionized water.
Yield 90 %. 1H NMR (DMSO-d6, δ, ppm): 6.35 (s, ArH, 1H), 6.50
(dd, ArH, 1H), 7.44-7.47 (t, ArH, 1H), 7.51-7.54 (t, ArH, 2H), 7.68 (d, ArH, 1H), 7.84
(d, ArH, 2H), 10.59 (s, OH, 1H), 12.43 (s, OH, 1H). FTIR (KBr, cm-1): 3358 (OH), 1591
(-N=N-). Anal. Calcd. for C12H10N2O2 (214.22 g/mol) calcd./found (%): C, 67.28/66.73; N,
13.08/13.15; H, 4.70/4.48. Mp= 164 oC. UV-vis λ= 388 nm (ε= 1,5∙105 l∙mol-1∙cm-1).
2.3. Synthesis of azo-dianhydride
5

The azo-dianhydride (DA(H)) was synthesized according to the procedure published
previously [30]. Trimellitic anhydride acid chloride (24 mmol) was dissolved in 25 ml of THF
in a round-bottomed flask equipped with a magnetic stirrer and dropping funnel. The solution
was stirred at 80 oC under argon atmosphere. The prepared mixture of 2,4-
dihydroxyazobenzene (16 mmol) and pyridine (2.2 ml) in 25 ml of THF was added dropwise
to the flask. The solution was stirred and heated under reflux for 1 h. After cooling down to
room temperature, pyridine hydrochloride was precipitated. The solution was filtered and
THF was evaporated. During evaporation, orange azo-dianhydride was precipitated. The
product was washed several times with acetic anhydride and dried in a vacuum oven at 50 oC
for 24 h.
1
Yield 30%. H NMR (DMSO-d6, δ, ppm): 7.48 (d, ArH, 1H), 7.53 (d, ArH, 2H), 7.67
(d, ArH, 2H), 7.85 (d, ArH, 1H), 8.00 (d, ArH, 1H), 8.30 (t, ArH, 2H), 8.67 (d, ArH, 4H),
8.71 (d, ArH, 1H). FTIR (KBr, cm-1): 1855, 1779 (C=O in anydride); 1740 (-O-CO-); 1588 (-
N=N-). Anal. Calcd. for C30H14N2O10 (562.44g/mol) calcd./found (%): C, 64.44/63.11; N,
o
4.98/4.60; H, 2.51/2.00. Mp= 238 C. UV-vis λ= 262, 340, 480 nm (ε= 0,2∙105, 0,8∙105,
0,7∙105 l∙mol-1∙cm-1).
2.4. Synthesis of poly(ester amic acid)s (PAA-X)
The poly(amic acid)s (PAAs) were synthesized by low-temperature polycondensation
reaction of a dianhydride (DA or DA(H)) with m-phenylenediamine or 2,4-
diaminoazobenzene. The preparation steps were as follows: 1 mmol of diamine was dissolved
in NMP (20 % solution). Next, 1 mmol of dianhydride was gradually added in three portions
o
over a period of 2h. The reaction was carried out at 60 C for 24 h. The product was
o
precipitated in a solution of water/methanol (1:1 v/v). PAAs were dried at 50 C for 24 h in
vacuum.
1
PAA-1 was obtained from dianhydride DA and diamine DA(H). H NMR (DMSO-d6,
δ, ppm): 7.50-7.53 (m, ArH, 3H); 7.66 (d, ArH, 2H); 7.86 (d, ArH, 1H); 7.91 (d, ArH, 1H);
8.03 (d, ArH, 1H); 8.37 (d, ArH, 2H); 8.43-8.47 (qd, ArH, 2H); 13.80 (s broad, ArH, 2H).
FTIR (KBr, cm–1): 3650-2680 (broad COOH); 1711 (C=O in ester group), 1665 (C=O in
COOH); 1627 (C=O in amide group). Anal. Calcd. for [C36H22N4O10]n (670.60 g/mol)
calcd./found (%): C, 64.48/65.08; N, 8.35/8.21; H, 3.31/3.37.
1
PAA-2 was obtained from dianhydride DA(H) and m-phenylenediamine. H NMR
(DMSO-d6, δ, ppm): 7.32 (s, ArH, 1H); 7.45 (s, ArH, 1H); 7.52 (s, ArH, 2H); 7.58 (s, ArH,
1H); 7.70 (s, ArH, 2H); 7.81 (s, ArH, 1H); 7.97 (s, ArH, 1H); 8.08 (s, ArH, 1H); 8.19 (s,
6

ArH,1H); 8.29 (s, ArH, 2H); 8.42 (s, ArH, 4H); 8.63 (d, ArH, 1H); 10.60 (d, NH, 2H); 13.15
(s broad, COOH, 2H). FTIR (KBr, cm–1): 3660-2480 (broad COOH); 1742 (C=O in ester
group), 1726 (C=O in COOH); 1663 (C=O in amide group). Anal. Calcd. for [C36H22N4O10]n
(670.60 g/mol) calcd./found (%): C, 64.48/64.24; N, 8.35/7.75; H, H, 3.31/3.00.
1
PAA-3 was obtained from dianhydride DA(H) and diamine DA(H). H NMR (DMSO-
d6, δ, ppm): 7.36 (s, ArH, 1H); 7.41 (s, ArH, 2H); 7.51-7.53 (d, ArH, 4H); 7.60 (s, ArH, 2H);
7.68 (s, ArH, 2H); 7.72 (s, ArH, 1H); 7.85 (s, ArH, 1H); 7.90 (s, ArH, 1H); 7.98 (s, ArH, 1H);
8.10-8.14 (m, ArH, 1H); 8.38 (s, ArH, 1H); 8.46-8.48 (m, ArH, 2H); 8.67-8.67 (m, ArH, 2H);
8.71-8.75 (s, ArH, 1H); 11.04 (s, NH, 1H); 11.89 (s, NH, 1H); 13.63 (s broad, COOH, 2H).
FTIR (KBr, cm–1): 3680-2695 (broad COOH); 1740 (C=O in ester group), 1728 (C=O in
COOH); 1680 (C=O in amide group). Anal. Calcd. for [C42H26N6O10]n (772.69 g/mol)
calcd./found (%): C, 65.12/66.31; N, 10.85/10.46; H, 3.38/3.09.
2.5. Preparation of poly(ester imide)s (PI-X)
Poly(ester imide)s were prepared by casting the poly(amic acid) solution (0.1 g of
o
polymer in 1.5 ml of NMP) on glass slides and subsequent drying at 50 C to remove the
residual solvent. Next, the PAAs were converted to the PIs by three-stage heating of the
prepared slides in a vacuum oven in the temperature range of 180 – 240 oC depending on the
azo polymer chemical structure. The PAA films were heated initially at the onset temperature
of the imidization process for 45 minutes, then at the temperature of the maximum imidization
o
rate (that is at 218, 180 and 189 C for PAA-1, PAA-2 and PAA-3, respectively) for 35
minutes and finally, at the endset temperature of the imidization process for 25 minutes.
1
PI-1: H NMR (DMSO-d6, δ, ppm): 7.51 (s, ArH, 2H); 7.58-7.63 (m, ArH, 3H); 7.69
(s, ArH, 2H); 7.74 (s, ArH, 1H); 7.85 (d, ArH, 1H); 7.98 (d, ArH, 1H); 8.20 (d, ArH, 2H);
8.53-8.65 (m, ArH, 5H). FTIR (KBr, cm–1): 1782, 1722 (C=O in imide group); 1597 (N=N);
1363 (C-N stretching); 721 (C-N deformation). Anal. Calcd. for [C36H18N4O8]n (634.55
g/mol) calcd./found (%): C, 68.14/67.31; N, 8.83/8.46; H, 2.86/2.09.
1
PI-2: H NMR (DMSO-d6, δ, ppm): 7.49-7.53 (m, ArH, 3H), 7.66 (d, ArH, 2H), 7.85
(dd, ArH, 1H), 7.90 (d, ArH, 1H), 8.02 (d, ArH, 2H), 8.11-8.16 (q, ArH, 2H), 8.35 (d, ArH,
1H), 8.35-8.45 (qd, ArH, 5H). FTIR (KBr, cm–1): 1784, 1721 (C=O in imide group); 1596
(N=N); 1365 (C-N stretching); 723 (C-N deformation). Anal. Calcd. for [C36H18N4O8]n
(634.55 g/mol) calcd./found (%): C, 68.14/67.78; N, 8.83/9.85; H, 2.86/3.55.
PI-3: 1H NMR (DMSO-d6, δ, ppm): 7.52 (s, ArH, 6H); 7.69 (s, ArH, 5H); 7.86 (d, ArH,
1H); 8.00 (d, ArH, 2H); 8.05 (t, ArH, 1H); 8.26 (s, ArH, 2H); 8.56-8.69 (m, ArH, 5H). FTIR
7

(KBr, cm–1): 1782, 1722 (C=O in imide group); 1595 (N=N); 1358 (C-N stretching); 720 (C-
N deformation). Anal. Calcd. for [C42H22N6O8]n (738.66 g/mol) calcd./found (%): C,
68.29/69.18; N, 11.38/11.31; H, 3.00/3.63.
2.6.Characterization
1H NMR spectra have been recorded on an Avance II 600 MHz Ultra Shield Plus
(Bruker) spectrometer in DMSO-d6 using TMS as internal standard. The infrared (IR) spectra
of polymers were acquired with a Nicolet 6700 FTIR apparatus (Thermo Scientific) using
KBr pallets. UV-Vis spectra were measured with a V-570 UV-Vis-NIR spectrophotometer
(Jasco Inc.). The X-ray diffraction patterns of solid samples were recorded using CuKα
radiation on a wide-angle HZG-4 diffractometer (Carl Zeiss Jena) working in the typical
Bragg geometry. The reduced viscosity of azo poly(esterimide)s was measured in NMP at a
concentration of 0.2 g (100 mL)−1 at 25 oC using an Ubbelohde viscometer.
Thermogravimetric analyses (TGA) was performed with a Mettler Toledo TGA/DSC STARe
system over the temperature range from 25–800 C at a heating rate of 10 min-1 and a
constant nitrogen flow of 60 mL min-1. Differential scanning calorimetry (DSC) was
performed with a TA-DSC 2010 apparatus (TA Instruments) in a temperature range from 25
C to 300 C and the heating rate of 20 C min−1 under a nitrogen flow of 50 mL min-1. The
glass transition temperatures of azo poly(ester imide)s were determined as the midpoint of the
increase in the specific heat associated with the transition observed during the heating cycle
carried out after imidization of the precursor samples in the first heating run. The modulated
differential scanning calorimetry (MDSC) experiments were performed with DSC Q2000 (TA
Instruments) calorimeter equipped with liquid nitrogen cooling system and operating in a
heat-flux mode. Sample cell was constantly fluxed with high purity helium gas at a constant
flow rate of 25 mL/min. The MDSC experiments were carried out using 2 °C/min-heating
rate, 1 °C sinusoidal temperature modulation amplitude and 120 s modulation period. The
MDSC data were analyzed using TA Universal Analysis software.
2.7. Measurements of light-responsive behaviour
Photoinduced birefringence measurements were performed for polymer films cast on
glass substrates using the experimental setup reported in our previous work [22]. The
birefringence was induced by the excitation beam at 445 nm and was simultaneously probed
by a smaller in diameter 690-nm beam, both from single mode diode lasers. The polarization
directions of the beams were set at 45o with respect to each other. The intensity of the
8

excitation beam was 75 mW/cm2. The transmission T of the probing beam through the
polymer sample placed between two crossed polarizers was detected by a silicon
photodetector and measured using a lock-in amplifier (modulation frequency of the probing
beam was 2590 Hz). The photoinduced birefringence was calculated from the formula
λ
[31]: Δn =
arcsin T , where d is the thickness of the polymer film,  is the probing beam
πd
wavelength.
Holographic grating recording was studied in a setup presented in Fig. 1 with 442-nm
writing beam from a He-Cd laser. Two interfering beams were obtained using a 50:50
o
beamsplitter and two reflecting mirrors directing the beams to cross at the angle of ca. 30 .
Such an irradiation geometry resulted in a grating spacing =0.85 m (according to
=/(2sin(/2)), where  is the writing wavelength,  is the intersection angle). The
polarization directions of the beams were orthogonal and set at +45o or – 45o with respect to
the vertical direction. The polarization plane and intensity of both the writing beams were
controlled independently by a combination of a half-wave plate and a polarizer introduced in
the path of each beam. The intensity of a single beam was 35 mW/cm2 measured behind the
polarizer. The process of the holographic grating build-up was monitored by a horizontally
polarized 690-nm beam (from a diode laser) normally incident on the sample. The intensity of
the +1 order diffracted beam from the grating in the transmission mode was recorded during
ca. 80 min. After irradiation, topography of the polymer surface was examined by a Dektak
XT stylus profiler with 0.2 m radius diamond tipped stylus and a stylus force of 1 mg. Scan
duration was chosen so that a traversing resolution was better than 0.03 m.
Fig. 1. Experimental set-up for holographic grating recording
9

In the photoinduced birefringence measurements the PI-1, PI-2, PI-3 films of the
thickness of 0.84, 0.66, 0.45 m, respectively were used. The experiment on holographic
grating recording was performed for thicker layers of 1.5, 1.7 and 1.2 m thickness in the case
of PI-1, PI-2 and PI-3, respectively. The film thickness was determined using the Dektak XT
stylus profiler.
3. Results and discussion
Chemical structures of the prepared poly(amic acid)s and poly(ester imide)s are
illustrated in Fig. 2. The azo polymers contained the azobenzene groups with one of the
aromatic rings being an integral part of the polymer main chain. The chromophores were
situated between the amic acid or imide groups in PAA-1 and PI-1 or between the ester
groups in PAA-2 and PI-2. PAA-3 and PI-3 contained two chromophores in a repeating unit
placed between the ester and amic acid or imide groups, respectively.
Fig. 2. Chemical structures and codes of the studied poly(amic acid)s and poly(ester imide)s
3.1. Thermal imidization of PAAs
10

The imidization behaviour of the precursor samples was studied by conventional and
modulated DSC techniques (see Supplementary Material).
3.2. Polymer characterization
1
The results of H NMR and FTIR spectroscopy confirmed the designed chemical
structures of the azo poly(amic acid)s and their conversion to the corresponding azo
polyimides after thermal cyclization (see Supplementary Material).
To get an insight into possible geometry of the polymer chains, the geometry
calculations of polymer fragments consisting of three units were carried out. The quantum
theoretical calculations were performed using density functional theory (DFT), with an
exchange correlation hybrid functional B3LYP and the basis 6-31G++ for all atoms.
Calculations were carried out with use of Gaussian09 program [32]. The geometries of these
compounds were optimized in vacuum and no imaginary frequencies were found for any of
these species, computed at Ci symmetry, by vibrational analysis. The optimized geometries of
the polymers together with some intramolecular hydrogen bonds are shown in Fig. 3.
PI-1
11

PI-2
PI-3
Fig. 3. Calculated geometries of the azobenzene poly(ester imide)s trimers
In all the cases, hydrogen bonds between the protons from phenyl in polymer backbone and
oxygen of the carbonyl group in the ester linkages or in the imide rings occur. Additionally, in
the PI-2 and PI-3 systems the C-H… interactions between the protons of phenyl rings in
azobenzene and aromatic imide structures are observed. These interactions are responsible for
a larger distortion of the PI-2 and the PI-3 chains in comparison to that of PI-1. The C-H…
interactions affect the structural stiffness and may influence the chromophore ability for
photoisomerization.
Figure S4 presents exemplary X-ray diffraction patterns for the studied azo polymers. A
broad diffraction peak of the diffusion type centered on ca. 23o confirmed perfectly
amorphous nature of all the materials.
The solubility of the prepared poly(amic acid)s was examined qualitatively by
dissolution of 5 mg of polymer in 1 ml of organic solvent (at room temperature and the
boiling temperatures of the solvents) and was compared with the solubility of the resulting
azo poly(ester imide)s. The results are collected in Table S1. All the polymers were soluble in
polar solvents such as NMP and DMSO at room temperature and showed a partial solubility
in THF, CHCl3 and cyclohexanone. Good solubility in organic solvents may be connected
with an incorporation of the ester groups in the main chain and azobenzene groups as side
chains [30, 21]. In general, poly(amic acid)s were characterized by a better solubility than
polyimides due to the open imide rings. The PAA-1 and PAA-3 poly(amic acid)s exhibited a
better solubility in cyclohexanone than PIs; PAA-1 was soluble in THF unlike its
12

corresponding PI. The worst solubility among the studied poly(amic acid)s was found for
PAA-2. The reduced viscosity (ηred) of the resulting poly(esterimide)s were similar (Table S1)
and suggested low molecular weight of the investigated materials and thus their oligomeric
nature.
3.3. Thermal properties of poly(ester imide)s
Thermal properties of poly(ester imide)s, that is glass transition temperature (Tg) and
thermal stability were evaluated by differential scanning calorimetry and thermogravimetric
analysis. Thermal stability was characterized by the temperatures of the 5 and 10 % (T5%,
T10%) weight loss, which define the beginning of thermal decomposition, and by temperatures
of the maximum rate of polymer degradation (Tmax). Thermogravimetric and derivative
thermogravimetric curves of azo poly(ester imide)s are shown in Fig. S5. The values of glass
transition temperatures of the azo poly(ester imide)s were in the range 174 –206 oC (Table 1).
Table 1. Thermal properties of the azo poly(ester imide)s
Char yield [%]
Polymer
Tg *[oC]
T5% [oC]
T10% [oC]
Tmax [oC]
at 800 oC
174
206
195
345
246
310
372
345
336
409; 623
416; 639
367; 617
43
36
48
PI-1
PI-2
PI-3
* determined from the heating cycle at 20 C min−1 carried out after imidization of the precursor samples during
the first heating run
Larger Tgs observed for PI-2 and PI-3 than for PI-1 appear to correlate with the presence of
C-H... interactions (Fig. 3), which affect the structural stiffness. The lowest Tg found for the
polyimide with the azo chromophores placed between the imide rings is in accordance with
the observation found recently for a series of poly(amide imide)s [33]. The studied poly(ester
imide)s were characterized by high values of T5% and T10%, well separated from the Tgs (Fig.
S5, Table 1). This is an important property when practical applications are considered, as it
allows for thermal erasure of photo-induced anisotropy or surface relief gratings at
temperatures near the Tg without material degradation. Similarly to other functionalized azo
polymers, all the studied poly(ester imide)s exhibited two degradation steps of the maximum
rate at Tmax (Table 1). The first step in the range 367 – 416 oC may be connected with thermal
degradation of the N=N bond in the azobenzene group and the second one, in the range 439 –
639 oC is attributed to the degradation of the main chains [34, 35]. The investigated poly(ester
imide)s exhibited residual weight at 800 oC in the range of 36 – 48 %.
13

3.4. Photoinduced birefringence
The UV-vis absorption spectra of the azo poly(ester imide) films on glass substrates are
presented in Fig. 4.
Fig. 4. UV-vis absorption spectra of the poly(ester imide) films onto glass substrates
The spectra showed similar characteristics. The intense band at ca. 380 nm was attributed to
-* electronic transition of the trans azobenzene moiety, while the less intense band
extending from ca. 410-500 nm was assigned to n-* electronic transition. The utilised
writing wavelength (i.e., 445 nm) excited mainly the latter type of transitions of trans
chromophores, while the probing beam wavelength was located outside the absorption band.
The determined film absorption coefficients at 445 nm (445) were 1.2·104 cm-1, 1.0·104 cm-1,
2.0·104 cm-1 for PI-1, PI-2 and PI-3, respectively. In accordance with the expectations,
similar values of the absorption coefficients were found for poly(ester imide)s containing one
azo chromophore per a repeating unit, while the largest 445 exhibited the polymer with two
azo chromophores in the repeating unit.
The photoinduced birefringence growth curves recorded within ca. 20 min-irradiation
with the 445-nm excitation beam of 75 mW/cm2 intensity are compared in Fig. 5 a. The
photoinduced birefringence relaxation curves recorded after turning off the excitation beam
are presented in Fig. 5 b. (Figure 5 b presents an additionally measured birefringence decrease
in the PAA-1 film).
14

Fig. 5. (a) Birefringence growth curves upon 445-nm beam of 75 mW/cm2 intensity for the studied azo
poly(ester imides), (b) the normalized birefringence decreases after turning off the excitation beam (together
with the birefringence decrease for PAA-1).
The largest photoinduced birefringence of 0.036 was generated in PI-3 having two
chromophores per a structural unit, while the lowest of 0.02 was achieved in PI-2. The
photoinduced birefringence of 0.02 is similar in magnitude to the nmax reported recently for a
family of azo poly(ester imide)s containing a flexible aliphatic group within the backbone
[36]; the n value of 0.036 may be considered high for azo polymers belonging to polyimides
[37]. Interestingly, the final light-induced birefringence achieved in PI-1 was only 17 % lower
than the maximum birefringence observed for PI-3. On the other hand, the nmax in PI-1
evidently exceeded the maximum n generated in PI-2 possessing a similar absorption
coefficient at 445 nm. For PI-1 saturation of the photo-orientation process was observed,
contrary to PI-2 and PI-3, for which the birefringence signals were still increasing with time
even at the end of the experiment.
Considering the results of DFT calculations, it is reasonable to suspect that differences
in the magnitude of the photoinduced birefringence between PI-2 and PI-1 arise from the
presence of the C-H… interactions between the protons of phenyl rings in azobenzene and
aromatic imide structures in PI-2. The interactions may restrict the chromophore mobility
leading to a less efficient and longer in time process of molecular reorientation towards
directions perpendicular to the electric field vector of the excitation beam. This is what was
observed in our measurement. The lack of birefringence saturation in PI-3 and its nmax
slightly exceeding the nmax of PI-1 (despite a larger chromophore content) can also be
explained by the presence of the mentioned intrachain hydrogen bonds found in the PI-3
system.
15

As shown in Fig. 5 b a relatively slow decrease in the generated birefringence after
ceasing the excitation beam was observed for all the azo poly(ester imide)s. The relaxation of
the induced molecular order has been attributed to the thermal cis-trans isomerization as well
as thermal reorientation of azo chromophores with a possible motion of polymer chain
segments [38]. Low birefringence relaxation is a characteristic feature of side-chain
polyimides possessing rigidly attached azo-dyes via one of their phenyl rings [33,36,37].
Nevertheless, some differences in the magnitude of birefringence decrease were observed
depending on details of the chemical architecture of the studied materials. As it can be noticed
from Fig. 5 b, the birefringence decrease by ca. 7 % within 33 min was observed for PI-2 and
PI-3. A visibly more stable birefringence (4-% decrease during 33 min) was found for PI-1. It
is worth pointing that the birefringence decrease in the PAA-1 precursor was similar in
magnitude to the decrease in PI-2 and PI-3. (The maximum birefringence in PAA-1 was
0.024 - the result not shown in Fig. 5 a). Comparison between the birefringence relaxation
curves for the PAA-1 precursor and the PI-1 polyimide additionally indicates that the
presence of the imide ring next to the azobenzene chromophore had an advantageous effect on
the photoinduced birefringence stability in the dark.
3.5. Surface relief grating inscription
The experiment on diffraction grating inscription by two interfering beams was
performed under similar excitation conditions as these used in the photoinduced birefringence
measurements (i.e., almost identical excitation wavelengths and similar excitation beam
intensities). Figure 6 presents the growth in the diffraction efficiency () of 690-nm beam
recorded as a function of time. Examination of the surface topography after irradiation
confirmed the grating inscription in PI-1 and PI-3; surprisingly, for PI-2 no surface
modulation was observed. An exemplary roughness of PI-3 film before irradiation and the
profiles of the inscribed SRGs are presented in Fig. 7.
16

Fig. 6. Growths of the first-order diffraction efficiencies (defined as the ratio between the intensities of the first-
order diffracted beam and the incident beam)
Fig. 7. Surface topology of the (a) non-irradiated PI-3 film, (b) PI-3 film after irradiation with the interfering
beams, (c) PI-1 film after irradiation with the interfering beams. (Graphs from the acquisition data system of
Dektak XT profiler)
In general, the diffraction efficiency of scattered light results from presence of two
phase gratings, i.e., volume birefringence - and thickness gratings, of contributions dependent
on the experimental arrangement and azo polymer architecture [39,40]. In the case of the PI-2
polymer, the maximum-recorded diffraction efficiency was only 0.05 %. As no surface relief
was observed for this material, the recorded weak diffraction signal was entirely due to bulk
birefringence grating. Visibly larger diffraction efficiencies were measured for two other azo
17

poly(ester imide)s. The final value of  was twice higher for PI-1 than for PI-3. At the end of
the experiment the diffracted signals were still increasing indicating that the surface relief
grating formation was not saturated. As shown in Fig. 6, the diffraction efficiency measured
for the PI-1 and PI-3 samples increased monotonously with time but with a different
dynamics. The observed effect might arise from mutual phase shifts between the growing
refractive index and relief gratings, which might additionally change with time [40,41].
The period of the surface grating spacing (i.e., 1.0 m and 0.85 m for PI-1 and PI-3,
respectively) was consistent with the period of the interference pattern. We would like to
mention that SRGs of such a period gave the first order diffraction beam only, appearing at a
large angle of about 50o. The grating modulation profile in PI-1 had a very regularly spaced
sinusoidal shape with the amplitude of 25 nm (Fig. 7 c). The grating profile found for PI-3
was not so regular being affected by initial surface roughness of few nanometers (Fig. 7 a). As
maximum values of photoinduced birefringence were quite similar in these materials, a
stronger contribution to final diffraction efficiency in the case of PI-1 was due to the SRG
than the bulk refractive index grating.
It is known that in high-Tg polymers, formation of the surface relief proceeds with a
lower efficiency [18,37]. Although the presence of SRG on the film surface of PI-1 and PI-3
was relatively easy verified with the naked eye, the measured values of modulation depths
appeared to be low, when compared with the modulation amplitudes reported for other azo
polymers [15,18]. The modulation amplitude for PI-1 is however, comparable to that reported
for a series of poly(ester imide)s possessing the azo chromophores with the electron-
withdrawing substituents attached to the backbone via a long spacer [16]. The modulation
amplitude of the SRG inscribed in PI-1 is also similar to the amplitude of the surface grating
inscribed in high-Tg azocarbazole-based polyimide [37]. It is worth noting that the experiment
on holographic gratings recording was studied here with the writing beams of rather low
intensities as compared with other experiments carried out for azo polyimides [16,37]. One
may suspect that optimization of the recording conditions would result in deeper SRGs.
However, our intention was to apply the same irradiation conditions in both experiments on
photoresponsive behaviour. From a combined analysis of the results on the photoinduced
birefringence generation - and SRG formation processes some interesting points arise. In spite
of relatively efficient photoisomerization cycles (probed in the photoinduced birefringence
experiment) in all the materials, large differences in the efficiency of the polymer movement
were found. A very low height of SRG in PI-3 (having two chromophores per a structural
unit) and the lack of surface modulation in PI-2 could be explained by the orientation theory
18

of optical deformations in glassy azo polymers [42]. It was recently shown by Saphiannikova
et al. [43] that the proposed orientation approach agrees well with the results of experimental
and computer simulation studies. According to the theory, reorientation of chromophores is
accompanied by the appearance of the mechanical stress, which may be comparable to the
yield stress [42,43], and thus may lead to irreversible surface deformations. Using this
approach the lack of SRG in PI-2 indicates that the light-induced molecular order does not
generate the stress needed for deforming the PI-2 surface. The needed stress is generated in
PI-3, which possesses twice larger chromophore concentration, and in PI-1, which exhibits a
similar value of nmax as PI-3. Different modulation amplitudes of the inscribed SRGs in
these two materials appear to be related to differences in their yield stresses (of a lower value
in the case of PI-1). Lacking the additional measurements of the yield stress we believe that
the Tg value for PI-1 being the lowest for the whole polymer series and the absence of C-
H... interaction indicated by DFT calculations are in accordance with the prediction of a
lower yield stress of PI-1 than of PI-3.
4. Conclusions
A family of three functionalized azobenzene poly(ester imide)s was prepared from
their precursor poly(ester amic acid)s. Differences in their chemical architecture referred to
the chromophore position within the backbone: either between the ester or the imide groups as
well as to the chromophore concentration (one or two per a structural unit). The most
important results arising from the performed studies are as follows:
o
i. High glass transition temperatures in the range of (175-205) C well separated from the
decomposition temperatures were observed for all the materials. The lowest Tg value was
found for the polymer possessing the azo chromophores located between the imide groups
(i.e., PI-1).
ii. DFT calculations revealed the presence of C-H… interactions between protons of phenyl
rings in azobenzene and aromatic imide structures in the system containing the chromophores
exclusively between the ester groups (i.e., PI-2) and in the system with two chromophores per
a structural unit (i.e., PI-3).
iii. Saturation and a relatively large photoinduced birefringence in PI-1, noticeably exceeding
the birefringence induced in PI-2 were attributed to the absence of the intrachain C-
H…interaction in PI-1.
19

iv. The most efficient polymer mass transport in PI-1 was correlated with its relatively large
photoinduced birefringence and the lowest glass transition temperature among the series. The
lack of mass transport in PI-2 was correlated with its lowest value of photoinduced
birefringence and the highest Tg among the series. Both observations appear to be in
accordance with the orientation theory of optical deformations in glassy azo polymers by
indicating the necessity of generating sufficiently large degree of molecular order for
initiating macroscopic chain movement.
Efficient photoinduced chromophore ordering and its slow relaxation after irradiation
observed for PI-1 make this azo polymer an attractive candidate for certain applications in
photonics such as e.g. photoalignment of liquid crystals.
Acknowledgements
This study was partially supported by the National Science Centre, Poland under Grant No.
UMO-2016/21/N/ST5/03037. Calculations have been carried out in Wroclaw Centre for
Networking and Supercomputing (http://www.wcss.wroc.pl), grant no. 18.
References
[1] T. Todorov, L. Nikolova, N. Tomova, Appl. Opt. 23 (1984) 4309.
[2] M. Dumont and Z. Sekkat, Proc. SPIE 1774 (1992) 188.
[3] P. Rochon, E. Batalla, A. Natansohn, Appl. Phys. Lett. 66 (1995) 136.
[4] D. Y. Kim, S. K. Tripathy, L. Li, J. Kumar, Appl. Phys. Lett. 66 (1995) 1166.
[5] H. Ishitobi, S. Shoji, T. Hiramatsu, H.B. Sun, Z. Sekkat, Opt. Express 16 (2008) 14106.
[6] H. Leblond, R. Barille, S. Ahmadi-Kandjani, J.-M. Nunzi, E. Ortyl, S. Kucharski, J. Phys.
B 42 (2009) 205401.
[7] D. Bublitz, M. Helgert, B. Fleck, L. Wenke, S. Hvilstedt, P.S. Ramanujam, Appl. Phys. B:
Lasers Opt. 70 (2000) 863.
[8] Y. Yu, M. Nakano, T. Ikeda, Nature 425 (2003) 145.
[9] S. Hvilstedt, C. Sanchez, R. Alcala, J. Mater. Chem. 19 (2009) 6641.
[10] T. Alasaarela, D. Zheng, L. Huang, A. Priimagi, B. Bai, A. Tervonen, S. Honkanen, M.
Kuittinen, J. Turunen, Opt. Lett. 36 (2011) 2411.
[11] A. Kravchenko, A. Shevchenko, V. Ovchinnikov, A. Priimagi, M. Kaivola, Adv. Mater.
23 (2011) 4174.
[12] V. Balzani, A. Credi, M. Venturi, Chem. Soc. Rev. 38 (2009) 1542.
[13] O.Yaroshchuk, Y. Reznikov, J. Mater. Chem. 22 (2012) 286.
20

[14] B. Rajashekar, S. Limbu, K. Aditya, G. N. Rao, S.S.S. Sai, Photochem. Photobiol. Sci.
12 (2013) 1770.
[15] A. Natansohn, P. Rochon, Chem. Rev. 102 (2002) 4139.
[16] E. Schab-Balcerzak, M. Siwy, M. Kawalec, A. Sobolewska, A. Chamera, A. Miniewicz,
J. Phys. Chem. A. 113 (2009) 8765.
[17] A. Kozanecka-Szmigiel, J. Konieczkowska, K. Switkowski, J. Antonowicz, B.
Trzebicka, D. Szmigiel, E. Schab-Balcerzak, J. Photochem. Photobiol. A 318 (2016) 114.
[18] „Smart Light-Responsive Materials: Azobenzene-Containing Polymers and Liquid
Crystals” Y. Zhao, T. Ikeda, Eds.; John Wiley & Sons, New York 2009.
[19] G.-S. Liou, H.-J. Yen, Polym. Sci.: A Comprehensive Ref. 5 (2012) 497.
[20] D.-J. Liaw, K.-L. Wang, Y.-C. Huang, K.-R. Lee, J.-Y. Lai, C.-S. Ha, Progress in
Polymer Science 37 (2012) 907.
[21] I. Sava, A. Burescu, I. Stoica, V. Musteata, M. Cristea, I. Mihaila, V. Pohoatab, I.
Topalab RSC Adv. 5 (2015) 10125.
[22] J. Konieczkowska, H. Janeczek, J. Małecki, B. Trzebicka, D. Szmigiel, A. Kozanecka-
Szmigiel, E. Schab-Balcerzak, Polymer 113 (2017) 53.
[23] D. Sek, E. Grabiec, A. Miniewicz, A. Sobolewska, Proc. SPIE 5724 (2005) 311.
[24] R. Weglowski, W. Piecek, A. Kozanecka-Szmigiel, J. Konieczkowska, E. Schab-
Balcerzak, Opt. Mater. 49 (2015) 224.
[25] N. Zetssu, T. Seki, Macromolecules 37 (2004) 8692.
[26] W. Li, S. Nagano, T. Seki, N J. Chem. 33 (2009) 1343.
[27] H. Takase, A. Natansohn, P. Rochon, Polymer 44 (2003) 7345.
[28] E. Schab-Balcerzak, M. Siwy, B. Jarzabek, A. Kozanecka-Szmigiel, K. Switkowski, B.
Pura, J. Appl. Polym. Sci. 120 (2011) 631.
[29] D. Sek, E. Schab-Balcerzak, E. Grabiec, Polymer 39 (1998) 7001.
[30] E. Schab-Balcerzak, H. Janeczek, P. Kucharski, J. Appl. Polym. 118 (2010) 2624.
[31] M. Born, E. Wolf, Principles of Optics 1999 7th edn (Cambridge: Cambridge University
Press).
[32] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G.
Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H.P.
Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T.
Vreven, Jr. J.A. Montgomery, J.E. Peralta, F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers,
K.N. Kudin, V.N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J.C.
21

Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B.
Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J.
Austin, R. Cammi, C. Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski,
G.A. Voth, P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, O. Farkas, J.B.
Foresman, J.V. Ortiz, J. Cioslowski, D.J. Fox (2009) Gaussian, Inc., Gaussian 09, Revision
A.1, Wallingford CT
[33] J. Konieczkowska, H. Janeczek, A. Kozanecka-Szmigiel, E. Schab-Balcerzak. Mater.
Chem. Phys. 180 (2016) 203.
[34] I. Sava, L. Sacarescu, I. Stoica, I. Apostol, V. Damian, N. Hurduc, Polym. Int. 58 (2009)
163.
[35] I. Sava, A. Burescu, G. Lisa, Polym. Bull. 71 (2014) 1359.
[36] A. Kozanecka-Szmigiel, J. Konieczkowska, D. Szmigiel, K. Switkowski, M. Siwy, P.
Kuszewski, E. Schab-Balcerzak, Dyes and Pigm., 114 (2015) 151.
[37] J. P. Chen, F. L. Labarthet, A. Natansohn, P. Rochon, Macromolecules 32 (1999) 32
8572.
[38] M.S. Ho, A. Natansohn, P. Rochon Macromolecules 28 (1995) 6124.
[39] N. K. Viswanathan, S. Balasubramanian, L. Li, S.K. Tripathy. J Kumar, Jpn. J. Appl.
Phys. 38 (1999) 5928.
[40] A. Sobolewska, A. Miniewicz, J. Phys. Chem. B, 111 (2007) 1536.
[41] N. Reinke, A. Draude, T. Fuhrmann, H. Franke, R.A. Lessard, Appl. Phys. B 78 (2004)
205.
[42] V. Toshchevikov, M. Saphiannikova, G.Heinrich, J. Phys. Chem. B 113 (2009) 113
5032.
[43] M. Saphiannikova, V. Toshchevikov, Journal of the SID 23/4 (2015) 146.
22