Millisecond photorefractivity with novel dicyanomethylenedihydrofuran- containing polymers

Article abstract of DOI:10.1039/c2jm31320k

New multifunctional copolymers containing carbazole units and high loads of dicyanomethylenedihydrofuran (DCDHF) were synthesized and used to prepare blends for photorefractive (PR) purposes. The materials response, which is strongly dependent on the glass-transition temperature (T_g), was thoroughly analyzed by holography, conductivity and ellipsometry measurements in order to both determine the limiting factors and optimize the performance. Materials that have a T_g around room temperature show strongly hindered chromophore orientation, which is avoided by lowering the T_g down to 6 °C, without compromising the PR effect or the material stability. A further DCDHF-containing homopolymer without carbazole was synthesized and characterized, showing an inferior PR response, which is attributed to a beneficial role for charge generation and transport of the attached carbazole in the copolymers. The new blends strongly improve the structural properties of previous DCDHF-based materials, allowing application of fields well above 100 V μm^-1 and preventing beam fanning. Outstanding PR performance was achieved, with fast buildup and erasure times of a few tens of milliseconds (even at low recording intensities), large refractive index modulation (over 10^-2) and two-beam coupling gain (above 350 cm^-1). Such performance is among the best reported for PR materials based on multifunctional and nonlinear polymers and comparable to standard PR composites.

Full text of DOI:10.1039/c2jm31320k

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
Journal of
Materials Chemistry
Cite this: J. Mater. Chem., 2012, 22, 12220
www.rsc.org/materials
PAPER
Millisecond photorefractivity with novel dicyanomethylenedihydrofuran-
containing polymers
a
Francisco Gallego-Gomez,† Julio C. Alvarez-Santos, Jose L. Rodrıguez-Redondo, Enrique Font-Sanchis,
b
b
b
ꢀ
ꢀ
ꢀ
Jose M. Villalvilla, Angela Sastre-Santos, Marıa A. Dıaz-Garcıa and Fernando Fernandez-Lazaro
ꢀ
a
b
a
b
ꢀ
*
*
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
Received 2nd March 2012, Accepted 23rd April 2012
DOI: 10.1039/c2jm31320k
New multifunctional copolymers containing carbazole units and high loads of
dicyanomethylenedihydrofuran (DCDHF) were synthesized and used to prepare blends for
photorefractive (PR) purposes. The materials response, which is strongly dependent on the glass-
transition temperature (Tg), was thoroughly analyzed by holography, conductivity and ellipsometry
measurements in order to both determine the limiting factors and optimize the performance. Materials
that have a Tg around room temperature show strongly hindered chromophore orientation, which is
avoided by lowering the Tg down to 6 ꢀC, without compromising the PR effect or the material stability.
A further DCDHF-containing homopolymer without carbazole was synthesized and characterized,
showing an inferior PR response, which is attributed to a beneficial role for charge generation and
transport of the attached carbazole in the copolymers. The new blends strongly improve the structural
properties of previous DCDHF-based materials, allowing application of fields well above 100 V mmꢁ1
and preventing beam fanning. Outstanding PR performance was achieved, with fast buildup and
erasure times of a few tens of milliseconds (even at low recording intensities), large refractive index
modulation (over 10ꢁ2) and two-beam coupling gain (above 350 cmꢁ1). Such performance is among the
best reported for PR materials based on multifunctional and nonlinear polymers and comparable to
standard PR composites.
load of NLO moieties (chromophores) for large EO modulation,
and, second, low glass-transition temperatures (Tg) in order to
profit from birefringence (BR) modulation, which also contri-
butes to photorefractivity (orientational enhancement mecha-
nism,3 OEM). All of these requisites make the design of an
optimized material a challenging task.
Introduction
The photorefractive (PR) effect is a nonlinear optical (NLO)
process that induces a change in the material refractive index by
nonuniform illumination, arising from the coexistence of charge
photogeneration, carrier transport and trapping and electrooptic
(EO) response.1 PR materials could be useful for applications
like high-density holographic data storage, image processing,
phase conjugated mirrors and lasers and real-time imaging.
Among them, organic PR materials (mainly polymers) offer low
cost and easy processability and benefit from their typically low
dielectric constants (allowing for stronger internal fields) and
versatile composition.2 The complex nature of the PR effect
makes it necessary to combine all required functionalities in the
same material. Particularly suitable are materials with, first, high
So far, composites, in which a host polymer is doped by the
distinct functional species, allowing the separate improvement of
each molecule type and easy Tg adjustment, are the most
successful approach.4,5 However, being a physical mixture,
composites can undergo crystallization and phase separation,
which restricts the use of a high percentage of NLO chromo-
phores and limits the mechanical stability against electrical
breakdown. Alternatively, monolithic materials, like fully func-
tionalized (FF) polymers6 and low molecular weight glasses,7,8
are formed by one single component combining all of the
required functionalities, so that instability problems are greatly
reduced. Nevertheless, besides the need for important synthetic
effort, the price to be paid is a much lower tunability of their
properties, especially charge generation (with adequate photo-
sensitizing groups) and Tg. As an intermediate option, multi-
functional polymers,9,10 containing both charge transporting and
NLO subunits, also prevent phase separation (even at high
a
ꢀ
Instituto Universitario de Materiales de Alicante and Dpto. Fısica
Aplicada, Universidad de Alicante, 03080 Alicante, Spain. E-mail: maria.
diaz@ua.es; Fax: +34 96 590 9726; Tel: +34 96 590 3543
b
ꢀ
ꢀ
ꢀ
ꢀ
Division de Quımica Organica, Instituto de Bioingenierıa, Universidad
Miguel Hernaꢀndez, 03202 Elche, Spain. E-mail: fdofdez@umh.es; Fax:
+34 96 665 8351; Tel: +34 96 665 8405
† Current address: Instituto de Ciencia de Materiales de Madrid, CSIC;
28049 Madrid, Spain.
12220 | J. Mater. Chem., 2012, 22, 12220–12228
This journal is ª The Royal Society of Chemistry 2012

View Article Online
chromophore contents) through
a much less demanding
Experimental
synthesis. Photogeneration and Tg can be adjusted by separately
adding a photosensitizer (in low amounts) and a plasticizer
(normally a nonpolar molecule) without compromising the
material stability. However, in spite of notable research work,
these strategies are generally in clear disadvantage in comparison
to the standard composite approach.
Synthetic procedures
NMR spectra were taken on a 300 MHz Bruker AC-300, using
deuterated chloroform as the solvent and tetramethylsilane as
the internal reference. UV-vis absorption measurements were
taken on a Thermo Spectronic Helios g spectrophotometer with
chloroform as the solvent. Ultraviolet data were used to deter-
mine copolymer composition. Infrared measurements were taken
with a Fourier Transform Thermo Nicolet IR 200 spectrometer
in ATR mode with a germanium window. Glass transition
temperatures (Tg) were measured by differential scanning calo-
rimetry (DSC) with a PerkinElmer Pyris 1 analyzer during the
second heating run (20 K minꢁ1) under nitrogen. Gel permeation
chromatography (GPC) was carried out using Agilent Zorbax
PSM silanized columns. N,N-Dimethylformamide (DMF) with
0.1% of LiBr was used as the solvent and the measurements were
Here, we undertake the synthesis of two multifunctional
methacrylate copolymers, 1 and 2 (Fig. 1), that have an attached
carbazole, as a widely used hole-conducting molecule,11 and
hydroxyl derivatives of 2-dicyanomethylene-3-cyano-4-styryl-
dihydrofuran, taking advantage of the high hyperpolarizability
of dicyanomethylenedihydrofuran (DCDHF).12 DCDHF-based
glasses are among the best PR monolithic materials reported to
date, generally showing high optical gain coefficients G (up to
13,14
250 cmꢁ1 at ꢂ30 V mmꢁ1
)
and diffraction efficiencies (near
total diffraction at ꢂ25 V mmꢁ1).14,15 However, they exhibit
a slow PR response (in the best cases with buildup times of ꢂ1 s
at 30 V mmꢁ1 13,16,17 and erase times of several seconds13,15), low
dielectric breakdown (often below 50 V mmꢁ1) and frequently
strong beam fanning7,13,16 or thermal instability.7,13 Only one PR
polymer bearing DCDHF has been reported so far;18 a FF
ꢀ
ꢁ1
done at 70 C with a flow rate of 1.0 mL min using a diode
array detector. The columns were calibrated with narrow
distribution standards of polystyrene. The methacrylate of 2-
hydroxyethylcarbazole was synthesized as reported in the
literature.19
polymer exhibiting good steady-state performance (G
¼
180 cmꢁ1 and total diffraction at 50 V mmꢁ1), but a slow
response (>1 s) and low dielectric breakdown. Since the typical
slow response is attributed to limited chromophore orienta-
tion,8,13,18 we focus on the achievement of relatively soft poly-
mers and subsequent preparation of PR blends with optimized
Tg. Fast erasure, as a relevant technological parameter, was also
aimed for. High polymer solubility and processability, which
should minimize beam fanning and dielectric breakdown, were
also targeted. We further synthesized homopolymer 3 (Fig. 1),
which is identical to 2 but without carbazole, to evaluate the
convenience of anchoring together charge transport and NLO
moieties. A detailed analysis of the PR performance, supported
by photoconductivity and ellipsometric experiments, was per-
formed in order to understand the mechanisms at play and the
improving factors of our materials
(E)-2-{40-[200-(4000-Cyano-5000-dicyanomethylene-2000,2000-
dimethyl-2000,5000-dihydrofuran-3000-yl)vinyl]-20,50-dimethylphe-
noxy}ethyl methacrylate (5). Compound 420 (1.04 g, 2.8 mmol)
was dissolved in 25 mL of dry THF under an argon atmosphere.
Then, methacryloyl chloride (0.41 g, 4.4 mmol) and dry trie-
thylamine (0.37 g, 3.6 mmol) were added and the reaction
mixture was held at room temperature for 96 hours. The reaction
mixture was filtered and THF was removed under reduced
pressure. The crude solid was dissolved in ethyl acetate and the
solution was washed twice with 2 M sodium hydroxide solution,
then with 0.1 M HCl and finally with distilled water. The organic
layer was dried over anhydrous sodium sulfate and filtered, and
the solvent was removed under reduced pressure. The product
was purified by chromatography through a silica gel column
(dichloromethane). Yield: 51%. Melting point: 115–116 ꢀC.
Fig. 1 Chemical structures of polymers 1–3 synthesized in this work.
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 12220–12228 | 12221

View Article Online
1H-NMR d: 7.56 (d, 1H, J ¼ 16.1 Hz), 7.16 (s, 1H), 6.96 (s, 1H),
6.85 (d, 1H, J ¼ 16.1 Hz), 6.14 (br, 1H), 5.59 (br, 1H), 4.53 (t, 2H,
J ¼ 5.2 Hz), 4.32 (t, 2H, J ¼ 5.2 Hz), 2.58 (s, 3H), 2.17 (s, 3H),
1.95 (br, 3H), 1.40 (s, 6H). IR: 2930, 2231, 1709, 1584, 1528,
solvent was removed under reduced pressure. The product was
purified by chromatography through a silica gel column
ꢀ
(dichloromethane). Yield: 1.43 g (89%). Boiling point > 350 C
(dec.). 1H-NMR d: 8.10 (d, 2H, J ¼ 7.5 Hz), 7.44 (m, 4H), 7.23 (t,
2H, J ¼ 7.8 Hz), 6.08 (br, 1H), 5.54 (br, 1H), 4.37 (t, 2H, J ¼ 7.2
Hz), 3.54 (t, 2H, J ¼ 6.7 Hz), 1.93 (br, 3H), 1.85–1.35 (m, 8H).
IR: 2936, 2858, 1717, 1637, 1597, 1485, 1454, 1324, 1296, 1164,
1507, 1455, 1305, 1264, 1165, 1098, 963, 855 cmꢁ1
.
(E)-6-{40-[200-(4000-Cyano-5000-dicyanomethylene-2000,2000-
dimethyl-2000,5000-dihydrofuran-3000-yl)vinyl]-20-hexyloxyphenoxy}
hexyl methacrylate (7). Compound 621 (0.9 g, 1.8 mmol) was
dissolved in 15 mL of dry THF under an argon atmosphere.
Then, methacryloyl chloride (0.56 g, 6.1 mmol) and dry trie-
thylamine (0.57 g, 5.5 mmol) were added and the reaction
mixture was held at room temperature for 96 hours. The reaction
mixture was filtered and THF was removed under reduced
pressure. The crude solid was dissolved in ethyl acetate and the
solution was washed twice with 2 M sodium hydroxide solution,
then with 0.1 M HCl and finally with distilled water. The organic
layer was dried over anhydrous sodium sulfate and filtered, and
the solvent was removed under reduced pressure. The product
was purified by chromatography through a silica gel column
1120, 1031, 941, 815, 751, 725, 651 cmꢁ1
.
General procedure for polymerization. Azobisisobutyronitrile
(AIBN, 2 wt%) was added to a 1 M solution of methacrylic
monomers (in the case of copolymers, the ratio carbazole/chro-
mophore was 1.0/1.2) in dry DMF. The polymerization medium
was bubbled with an argon stream for 10 minutes before heating
ꢀ
at 60 C for 24 hours. The reaction was stopped by cooling to
room temperature. Cold methanol was added to precipitate the
polymers from the obtained gels. Polymers 2 and 3 were dis-
solved in THF and reprecipitated with MeOH.ꢀFinally, they were
filtered and dried overnight in an oven at 90 C under reduced
pressure. Polymer 1, being insoluble, had to be mechanically
broken and was washed for 24 hours with cold methanꢀol.
Finally, it was filtered and dried overnight in an oven at 90 C
under reduced pressure.
ꢀ
1
(dichloromethane). Yield: 59%. Melting point: 132–133 C. H-
NMR d: 7.57 (d, 1H, J ¼ 16.2 Hz), 7.22 (dd, 1H, J1 ¼ 8.5 Hz, J2 ¼
1.7 Hz), 7.10 (d, 1H, J ¼ 1.7 Hz), 6.91 (d, 1H, J ¼ 8.5 Hz), 6.85 (d,
1H, J ¼ 16.2 Hz), 6.09 (br, 1H), 5.55 (br, 1H), 4.16 (t, 2H, J ¼ 6.6
Hz), 4.08 (t, 2H, J ¼ 6.6 Hz), 4.05 (t, 2H, J ¼ 6.5 Hz), 1.91–1.25
(br, 25H), 0.91 (t, 3H, J ¼ 6.9 Hz). IR: 2928, 2866, 2226, 1736,
1562, 1540, 1461, 1380, 1317, 1261, 1149, 1142, 1117, 1015, 756,
Polymer 1. Yield: 72%. IR: 2227, 1729, 1674, 1570, 1528, 1504,
1460, 1311, 1261, 1095, 753, 726 cmꢁ1. Tg: 152 C.
ꢀ
634 cmꢁ1
.
Polymer 2. Yield: 64%. 1H-NMR d: 7.99, 7.36, 7.14, 7.00, 6.72,
4.11, 3.92, 3.83, 1.68, 1.43, 1.29, 1.01, 0.86 ppm. IR: 2940, 2863,
2231, 1729, 1574, 1538, 1465, 1274, 1150, 756, 731 cmꢁ1. UV-vis l
(log 3): 243 (1.20), 265 (1.08), 296 (0.94), 333 (0.53), 347 (0.60),
475 nm (0.95). Tg: 75 ꢀC. Mn: 28 000 g molꢁ1. Mw: 66 000 g
molꢁ1. Mw/Mn: 2.3.
6-(90H-Carbazol-90-yl)hexyl methacrylate (8). This compound
was prepared using a slightly modified procedure from that
described elsewhere.9,22 Carbazole (2 g, 12 mmol), sodium
hydride (0.56 g, 14 mmol) and a catalytic amount of potassium
iodide were dissolved in dry DMF (42 mL) under an argon
atmosphere. After twenty minutes stirring at room temperature,
6-chlorohexanol (1.64 g, 12 mmol) dissolved in dry DMF (8 mL)
Polymer 3. Yield: 43%. 1H-NMR d: 7.57, 7.22, 7.10, 6.91, 6.85,
4.16, 4.08, 4.05, 1.91, 1.73, 1.62, 1.55, 1.33, 1.25, 0.91 ppm. IR:
2927, 2860, 2227, 1726, 1567, 1532, 1465, 1382, 1327, 1267, 1153,
1143, 1111, 1020, 615 cmꢁ1. UV-vis l (log 3): 240 (1.04), 277
ꢀ
was added dropwise. The reaction mixture was heated at 90 C
for 21 hours. After cooling to room temperature, water was
added to the reaction mixture to remove excess sodium hydride.
The mixture was extracted twice with ethyl acetate and the
organic layer was washed with diluted hydrochloric acid, then
with water, and dried over anhydrous sodium sulfate before
removal of solvent under reduced pressure. The crude product
was purified by chromatography through a silica gel column (n-
hexane-ethyl acetate, 2/1 v/v). The yield of 6-(90H-carbazol-90-yl)
hexan-1-ol was 1.97 g (61%). Melting point: 110–111 ꢀC. 1H
NMR d: 8.10 (d, 2H, J ¼ 7.8 Hz), 7.44 (m, 4H), 7.23 (m, 2H), 4.31
(t, 2H, J ¼ 8.1 Hz), 3.60 (t, 2H, J ¼ 6.2 Hz), 1.47 (m, 8H). IR:
3346, 2937, 2862, 1710, 1453, 1326, 1277, 1237, 1074, 1054, 1033,
1002, 752, 725, 650 cmꢁ1. This product (1.3 g, 4.8 mmol) was
dissolved in dry THF (50 mL) under an argon atmosphere. Then,
methacryloyl chloride (1.53 g, 14.6 mmol) and dry triethylamine
(1.47 g, 14.6 mmol) were added and the reaction mixture was
heated at 50 ꢀC for 72 hours. After cooling, the reaction mixture
was filtered and THF was removed under reduced pressure. The
crude solid was dissolved in ethyl acetate and the solution was
washed twice with 2 M sodium hydroxide solution, then with
0.1 M HCl and finally with distilled water. The organic layer was
dried over anhydrous sodium sulfate and filtrated, and the
(0.83), 350 (0.64), 474 nm (1.08). Tg: 80 ꢀC. Mn: 62 000 g molꢁ1
Mw: 84 000 g molꢁ1. Mw/Mn: 1.3.
.
Materials preparation and samples fabrication
Different blends were prepared by separately dissolving each
component in dry dichloromethane and mixing them in adequate
proportions. The solvent was evaporated by slow dropping on
two glass plates heated at 50 ꢀC. Thereafter, composites were
homogenized by mechanical pounding between the two plates at
90–120 ꢀC. Material Tg’s were measured by DSC in a Mettler
Toledo TGA/DSC 1 apparatus during the second heating cycle
(20 K minꢁ1) under static air. Sandwiꢀch-like samples of thickness
d ¼ 37 mm were fabricated at ꢂ100 C by pressing the material
between two indium tin oxide (ITO) electrodes and prompt
cooling down to room temperature. The thickness was guaran-
teed using fiber-glass spacers. The moderate thickness ensured
optical quality and high transmittance of the samples. Absor-
bances, A, were measured on fresh samples with a Helios Gamma
UV-vis spectrophotometer to obtain the absorption coefficient
a according to a ¼ ln 10 A/d.
12222 | J. Mater. Chem., 2012, 22, 12220–12228
This journal is ª The Royal Society of Chemistry 2012

View Article Online
Photoconductivity and ellipsometry measurements
where Q is the PR phase-shift, i.e., the phase difference between
the index grating and the light interference pattern, and l is the
light wavelength. In low-Tg PR materials,3 Dn is proportional to
the external field E and the developing space-charge field ESC, the
latter being in turn proportional to the conductivity
contrast M.1
DC conductivities were measured by applying a field of 22 V
mmꢁ1 to the sample in the dark (sdark) and under illumination
(slight) with a He–Ne laser beam (633 nm, full-width half-
maximum of 10 mm2 and intensity of 50 mW cmꢁ2). The current,
i, was measured after 200 s with a Keithley 600B electrometer to
calculate the conductivities from sdark, light ¼ id/(VS), where S ¼
4.2 mm2 is the electrode area and d ¼ 37 mm is the sample
thickness. Photoconductivity was defined as sph ¼ slight ꢁ sdark
Dn was obtained from h according to eqn (2). For evaluation
of the PR dynamics, the resulting time-dependent Dn curves were
fitted by a bi-exponential function, with fast and slow time
constants sfast and sslow
.
and conductivity contrast as M ¼ (slight ꢁ sdark)/slight
.
Field-induced birefringence was measured by null-trans-
mission ellipsometry.23 A weak 785 nm beam impinged on the
sample (at internal incidence angle of 32ꢀ) placed between two
crossed polarizers (set at ꢃ45ꢀ with respect to the plane of inci-
dence). The transmitted beam intensity by applying an external
field to the sample was monitored. Time-dependent birefringence
curves were fitted by a monoexponential function with time
Results and discussion
DCDHF-containing polymers
Polymers were prepared by radical polymerization of the corre-
sponding methacrylates, using AIBN as the initiator. In a first
attempt, we selected polymer 1 as the target expecting that the
short alkyl chains would ensure enough solubility while mini-
mizing the weight of inert material. In this way, chromophore 420
was acylated in 51% yield by treatment with methacryloyl chlo-
ride in THF. The resulting ester 5 was copolymerized with 2-
(carbazol-90-yl)ethyl methacrylate19 to afford 1 in 72% yield
(Scheme 1). Polymer 1, which presented ethyl connectors,
happened to be insoluble, thus precluding complete characteri-
zation and any further material preparation. We decided to
introduce longer alkyl chains to improve the solubility of the
resulting polymers while lowering the Tg, although this would
reduce the percentage of active moieties in the material. This
strategy proved to be correct and allowed us the preparation of
the readily soluble polymer 2 (Scheme 2). Thus, reaction of
chromophore 621 with methacryloyl chloride yielded ester 7,
which was copolymerized in the same conditions as earlier,9
leading to 2 in 64%. In order to study the effect of attaching the
carbazole units to the polymer backbone, we also synthesized the
soluble polymer 3 in 43% starting from 7 (Scheme 2).
constant sBR
.
Holographic characterization
The composites were characterized by standard degenerate four-
wave-mixing (DFWM) and two-beam coupling (TBC) experi-
ments at room temperature (ꢂ22 ꢀC). Thereby, gratings were
written by two p-polarized 633 nm write beams (WB’s) with equal
internal intensities (160 mW cmꢁ2 each) in a typical tilted
recording geometry with internal WB angles q1 ¼ 26.7ꢀ and q2 ¼
33.3ꢀ (grating tilt angle J ¼ 30ꢀ and grating spacing L ¼ 3.2 mm).
Gratings were read out by a weak (1.2 mW cmꢁ2) p-polarized
633 nm read beam (RB), counterpropagating to WB1. Time-
resolved measurements were standardly carried out by pre-illu-
minating the sample with the fringe pattern (i.e., both WB’s on) in
the absence of an external field for 15 min and switching the field
on/off to write/erase the grating (E-switch method). Alternatively
(see below subsection on Photorefractive Characterization), the
sample was pre-poled by applying the field in the presence of only
one write beam (WB2) for 15 min and WB1 was switched on/off to
write/erase the grating (WB1-switch method). Grating values
reached after 200 s were taken as steady-state data.
1
The H NMR spectra of both 2 and 3 showed the disappear-
ance of the olefinic protons (at ca. 6.1 and 5.5 ppm) present in the
starting methacrylates (Fig. 2). UV-vis spectra of polymers 2 and
3 resembled those of the constitutive monomers (Fig. 3). Polymer
3, which has a higher quantity of chromophore (DCDHF) per
concentration unit, showed higher absorption in the 400–600 nm
range than copolymer 2. The comparison with the spectra of the
starting methacrylates allowed for the calculation of the
composition of copolymer 2, yielding a carbazole/chromophore
The PR performance was characterized by the internal
diffraction efficiency h (ratio of the diffracted RB intensity to the
sum of the diffracted and the transmitted RB intensities) and the
gain coefficient G calculated according to1
ꢀ
ꢁ
ꢂ
ꢁ
ꢂꢃ
1
I1
I2
G ¼
cosq1 ln
ꢁ cosq2 ln
(1)
d
I1ð0Þ
I2ð0Þ
where I1(2)(0) and I1(2) are the WB1(2) intensities after the sample
before and during grating recording, respectively. h and G are
related to the refractive index modulation Dn of the PR grating
by1,24
!
pdDn
h ¼ sin2
(2)
1=2
lðcosq1 cosq2Þ
and
2p
G ¼
Dn sin Q
(3)
l
Scheme 1 The synthesis of insoluble polymer 1.
J. Mater. Chem., 2012, 22, 12220–12228 | 12223
This journal is ª The Royal Society of Chemistry 2012

View Article Online
Fig. 3 Absorption spectra of (a) monomers 7 and 8, and (b) copolymer 2
and homopolymer 3. Inset: absorption coefficients of composites M2-54
and M3-42 measured at 37 mm samples.
Scheme 2 The synthesis of polymers 2 and 3.
Fig. 2 1H NMR spectra of copolymer 2 and homopolymer 3.
12224 | J. Mater. Chem., 2012, 22, 12220–12228
This journal is ª The Royal Society of Chemistry 2012

View Article Online
ratio of 1.0/1.9. The insolubility of polymer 1 precluded the
estimation of the composition by UV-vis. In this case, we used
the intensity of selected IR bands, obtaining a carbazole/chro-
mophore ratio of 1.0/0.9. The solubility of polymers 2 and 3
allowed for the determination of their molecular weights. Thus, 2
exhibited values of 66 000 and 28 000 for Mw and Mn, with
a polydispersity of 2.3. On the other hand, 3 showed a Mw value
of 84 000 and a Mn value of 62 000, being the polydispersity 1.3.
From the Mn values, we estimated that 3 was statistically
composed of chains of 116 units, while the chains of 2 contained
around 21 carbazole and 40 chromophore units.
Composite materials
PR materials M2 with varying Tg’s close to room temperature
were prepared by mixing polymer 2 with different amounts of the
plasticizer N-ethylcarbazole (ECZ), which also contributed to
photoconduction. All materials were sensitized with 1% of 2,4,7-
trinitro-9-fluorenone (TNF), which provides efficient carrier
photogeneration via charge transfer with the carbazole units. The
resulting composites were named after the polymer percentage
(Table 1). A further composite was analogously prepared from
polymer 3 (M3-42, Table 1) for direct comparison with the best
performing M2 material (M2-54), both having similar Tg’s (6–
Fig. 4 Temporal evolution of (a) diffraction efficiency and (b) gain
coefficient of gratings in M2 composites with different Tg’s (Table 1).
Gratings were recorded using the E-switch method (with an applied field
of 60 V mmꢁ1).
constants are displayed in Fig. 5(a) and (b) against the materials’
Tg, which evidently was a crucial parameter for the PR perfor-
mance. Indeed, gratings in harder composites showed lower Dn
and slower rates (larger buildup times, both sfast and sslow) while,
by lowering the material’s Tg, both grating strength and speed
improved due to enhanced chromophore motion. However, by
further decreasing the Tg towards 0 ꢀC the speed barely
increased, while the grating modulation Dn worsened. Such
behavior indicates that chromophore orientation was hindered in
8
ꢀC) and, simultaneously, equally large percentages of both
carbazole (ꢂ45 wt%) and NLO (ꢂ30 wt%) units. Note that some
percentage of carbazole units in the M3 composite was provided
by the addition of poly(N-vinylcarbazole) (PVK) in order to,
first, match the Tg of the M2 blend and, second, achieve a similar
percentage of polymeric matrix (ꢂ50 wt%). It is remarkable that,
despite the high loads of the strongly polar DCDHF (up to ꢂ40
wt%), samples exhibited neither phase separation nor aggrega-
tion over more than two years. This represents a clearly superior
stability compared to most host–guest materials, typically
restricted to a few weeks or months.2
ꢀ
the harder materials (Tg > 20 C, although being slightly above
RT), leading to restricted OEM (i.e. poor BR contribution to the
grating) and slow response. In the softer materials, chromophore
mobility clearly improved, allowing larger modulation Dn with
faster speed. The saturation of the speed improvement suggests
that chromophore orientation was no longer a limiting factor for
Tg ¼ 6 ꢀC or less. This was confirmed by the dynamics of chro-
mophore alignment from transient ellipsometry, described by the
time constant sBR (Fig. 5b); sBR was comparable to sslow in the
harder composites (Tg ¼ 23 and 31 ꢀC) and further decreased in
the softer composites, being even lower than sfast. Thus, by
softening the material, the PR response speed was increasingly
determined by the space-charge field buildup, rather than by the
chromophore orientational speed.
Photorefractive characterization
We investigated the PR performance of M2 materials by DFWM
and TBC experiments. Both grating strength and dynamics
strongly varied between composites, some of them exhibiting
high and fast diffraction and optical gain at moderate applied
fields (Fig. 4). The PR refractive index modulation Dn was
calculated from DFWM measurements since h exclusively
depended on Dn (eqn (2)). Buildups of Dn could be fitted by bi-
exponential functions, with a fast and a slow time constants (sfast
and sslow, respectively). The resulting Dn and best-fit time
Table 1 Composition, glass-transition temperature (Tg) and absorption coefficient (a) of the PR blends characterized in this work. Weight percentages
of chromophore (calculated excluding the hexamethylenic connecting bridge) and total carbazole (calculated excluding the hexamethylenic connecting
bridge in 2 and the ethyl group in ECZ) units in each composite are also shown
Composition (wt%)
Functional components (wt%)
2
3
PVK
ECZ
TNF
Chromophore
Carbazole
Tg [ꢀC]
a [cmꢁ1
]
M2-71
M2-64
M2-54
M2-50
M3-42
71.0
64.0
54.0
50.0
—
—
—
—
—
—
—
—
—
7.6
28.0
35.0
45.0
49.0
49.0
1.0
1.0
1.0
1.0
1.0
38.2
34.4
29.0
26.9
29.9
32.1
37.3
44.6
47.6
48.1
31
23
6
110
105
93
82
115
ꢁ1
42.4
8
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 12220–12228 | 12225

View Article Online
Experimental section), the sample conditions prior to recording
can strongly vary.29 Fig. 6, left, compares these two writing
schemes in both the relatively hard M2-71 and the soft M2-54
composites. In the harder material, sample pre-poling (WB1-
switch method) allowed a much faster response as chromophores
only needed to slightly reorient under the effect of the developing
ESC (instead of reorient from a random distribution), which
mitigated the handicapped chromophore mobility. By contrast,
pre-poling barely improved the recording in the soft M2-54,
corroborating that the PR response was not orientation-limited
when the Tg was low enough. The distinct limiting factors
depending on the material’s Tg was further evidenced by the
dependence of the PR buildup speed on the total recording
intensity I (insets in Fig. 6). Indeed, the speed sfastꢁ1 was weakly I-
dependent in the harder material, indicating orientation-limited
dynamics,7,13 but nearly proportional to I in the soft composite,
as the speed became rather photoconductivity-limited.28 It is
worth noting that the little influence of sample pre-illumination
in the soft material demonstrates that the photogeneration of
uniform charge density prior to measurement did not enhance
the PR speed, which would occur, by contrast, in case of
a charge-generation-limited response.5,30 This suggests that our
material was efficiently sensitized by carbazole/TNF so that the
grating speed was rather limited by charge redistribution, which
is consistent with the typically low hole mobility in carbazole-
based systems. Thus, the clear intensity-dependence of the
buildup speed (inset in Fig. 6b) indicates that higher irradiance
led to improved charge transport, maybe due to modification of
the trap landscape, although this aspect is unclear at this point.
On the other hand, pre-illumination did not induce speed slow-
down, suggesting that no optical trap activation31 occurred.
The grating decay gave further information as the erasure
conditions strongly varied in each scheme, having a distinct effect
depending on the composite’s Tg (Fig. 6, right). On the one hand,
in the E-switch method the BR modulation relaxed as the aligned
Fig. 5 Dependence on the M2 composites’ Tg of (a) the PR refractive
index modulation Dn, (b) time constants sfast and sslow (both from Dn
fitting) and sBR (from ellipsometry fitting) and (c) conductivities slight and
sdark (left axis) and conductivity contrast M (right axis). PR gratings were
recorded using the E-switch method (applied field of 60 V mmꢁ1).
Ellipsometry was performed at 60 V mmꢁ1, while conductivities were
measured at 22 V mmꢁ1. Lines connecting points are to guide the eye.
In particular, sslow decreased by lowering the Tg more strongly
than sfast. This would agree with the usual assumption that the
slower component is mainly related to the chromophore orien-
tation (BR modulation), while the fast component is due to the
formation of the space charge field upon charge generation and
transport (EO modulation). This is, however, an oversimplified
description and both fast and slow components are also affected
by the different trapping sites, especially with Tg below RT.25
Indeed, sslow drastically dropped from Tg ¼ 31 to 23 ꢀC (in more
than one order of magnitude), while it decreased moderately for
Tg’s below RT. On the other hand, sfast was also influenced to
some extent.26 The worsening of Dn at the lowest Tg is explained
by the conductivity behavior (Fig. 5c), as slight barely improved
by decreasing Tg’s, while sdark increased over more than 2 orders
of magnitude. This led to a drastic reduction of the conductivity
contrast M (down to 0.68) in the softest composite, causing the
lowering of Dn. Similar deterioration of both M and photo-
refractive Dn has previously been observed by operating at
temperatures well above the material’s Tg27 and, in particular, in
DCDHF glasses.28 In general, DCDHF-based materials with
a Tg below room temperature exhibited inferior PR perfor-
mance,13,14,18 despite an improved orientational rate. In this
ꢀ
work, we achieved the best performance at Tg ¼ 6 C (M2-54)
Fig. 6 A comparison of h transient behaviors upon two different
recording/erasure schemes (E- and WB1-switch methods) performed in
(a) hard M2-71 and (b) soft M2-54 composites. The recording intensity I
was 320 mW cmꢁ2. Arrows indicate the beginning of grating buildup/
erasure (E or WB1 on/off). Insets: I-dependence of the experimental speed
sfastꢁ1 (symbols) and fit functions Ib with b ¼ 0.20 (a) and 0.89 (b) (lines).
The external field was 60 V mmꢁ1 in all cases.
and profited from the enhanced speed, keeping good material
processability.
Further insights into the physical processes governing the
material PR response can be gained from the dependence of the
grating dynamics on the recording/erasure scheme. By writing
the grating with either the E-switch or WB1-switch method (see
12226 | J. Mater. Chem., 2012, 22, 12220–12228
This journal is ª The Royal Society of Chemistry 2012

View Article Online
chromophore spontaneously randomized in absence of an
external field while the EO component decreased parallel to ESC
120 V mmꢁ1, although some saturation was evident above 50 V
mmꢁ1. Such deterioration of TBC has been typically caused by
strong beam fanning in other DCDHF-based materials.7,13,16
However, this detrimental effect was not significant in our
composites, as the total transmitted WB intensity decreased only
slightly at high voltages (inset in Fig. 7b). Thus, G saturation was
probably caused by loss of grating contrast (due to unequal WB
intensities within the sample resulting from the high TBC).
Regarding the dynamics, the buildup response speeded up for
increasing fields (Fig. 7c), as expected due to enhanced charge
photogeneration/transport and chromophore alignment, and
buildup times (sfast) as low as 35 ms at 120 V mmꢁ1 were achieved.
,
mainly due to recombination of the present charge carriers. This
erasure scheme was much slower in the hard material M2-71,
strongly limited by the inefficient chromophore reorientation.
Besides, the relatively fast decay in the soft material M2-54
indicated rather fast charge recombination in the material. On
the other hand, the WB1-switch method consists in optical
erasure, in which ESC decreased mainly through uniform charge
photogeneration upon the homogeneous illumination (opposite
to dark decay) followed by slight chromophore re-orientation in
the progressively uniform total field. The decay in this case,
which was increasingly effective at higher E due to the field-
dependent photogeneration in organic materials,32 achieved
clearly faster rates with this scheme. Once again, the decay was
slower in the hard material due to inefficient relaxation of the BR
modulation. In any case, optical erasure exhibited faster
dynamics than writing, which is an important technical feature
for real applications. Interestingly, resulting Dn decay curves
(calculated from eqn (2)) could be fitted by mono-exponential
functions in the soft composite, but bi-exponential in the hard
one, which manifests the existence of one predominant process
and two competing ones, respectively.
Erasure times (WB1-switch) were as low as 20 ms at 120 V mmꢁ1
.
The high electrical breakdown of the soft composite M2-54 is
remarkable (and even higher, up to 170 V mmꢁ1, in the harder
M2-71), which reveals the high structural stability of our multi-
functional polymers. These figures of merit are superior to most
of those reported in previous DCDHF-based PR materials and
are among the best data found in multifunctional polymers and
even standard composites.
Several composites based on homopolymer 3 were also
prepared and their PR properties were characterized, which also
allowed us to evaluate the convenience of anchoring the carba-
zole to the copolymer 2 together with the chromophore. Quali-
tatively, PR performance in M3 materials exhibited same
dependence on Tg and recording/erasure schemes as that in M2,
which manifests analogous physical properties in both polymers.
This suggests that the attached carbazole groups did not steri-
cally hindered the chromophore orientability in copolymer 2.
The best M3 composite (M3-42, Tg ¼ 8 ꢀC) is directly compared
The PR grating field-dependence, governed by the raise of ESC
and Dn with increasing E, was also characterized (Fig. 7, dis-
playing M2-54 data). h exhibited a sinusoidal dependence (eqn
(2)), with near total diffraction at 65 V mmꢁ1 in the relatively thin
37 mm samples (maximum Dn of 0.024 was achieved at 120 V
mmꢁ1). G monotonously increased (eqn (3)), up to 350 cmꢁ1 at
ꢀ
in Fig. 7 with the best M2 material (M2-54, Tg ¼ 6 C). M3-42
showed very similar field dependence as that of M2-54 but an
inferior PR effect, which was more evident in G than in h,
implying a larger phase-shift Q in the M2 material (according to
eqn (3)). Again, significant beam fanning was ruled out, as shown
in the inset of Fig. 7b. Concerning the dynamics, grating buildup
rates were lower (about two-fold) in the M3 material (Fig. 7c), as
well as the erasure rates. Assuming no orientation-limited
performance, those facts suggest a somewhat more effective
charge photogeneration and transport in copolymer 2, favoring
charge migration (large Q) and fast ESC formation. This is
further supported by the smaller photoconductivity measured in
the M3 composites (sph ¼ 2.4 and 3.1 pS cmꢁ1 in M3-42 and M2-
54, respectively, measured at 22 V mmꢁ1). A possible explanation
could be that the well-distributed carbazole units attached to the
copolymer 2, in which the NLO units act as spacers, would help
the charge transfer with TNF molecules and the subsequent hole
transport. The totally random distribution of carbazole in M3
materials may also lead to increased disorder of the density of
states (more dispersive charge transport32) or some p-stacking of
the planar carbazole groups,33 in both cases resulting in
decreasing photoconductivity.
Conclusions
Photorefractive low-Tg DCDHF glasses and fully functionalized
copolymers have previously been reported, generally showing
very slow PR response due to hindered chromophore orient-
ability. Unfortunately, further lowering of the Tg led to a faster
Fig. 7 Field-dependence of steady-state h and G and buildup time
constant sfast, measured in M2-54 and M3-42. Inset in (b): normalized
total transmitted WB intensity (I1 + I2) against the applied field. Lines
connecting points are to guide the eye.
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem., 2012, 22, 12220–12228 | 12227

View Article Online
ꢀ
€
5 E. Mecher, F. Gallego-Gomez, H. Tillmann, H. H. Horhold,
J. C. Hummelen and K. Meerholz, Nature, 2002, 418, 959.
6 (a) Z. Peng, A. R. Gharavi and L. Yu, J. Am. Chem. Soc., 1997, 119,
4622; (b) Q. Wang, L. Wang and L. Yu, Macromol. Rapid Commun.,
2000, 21, 723; (c) Q. Wang, L. Wang, J. Yu and L. Yu, Adv. Mater.,
2000, 12, 974.
but smaller effect, or to material instability. Here, we synthesized
two new multifunctional DCDHF-containing copolymers, with
attached carbazole units, one of them (2) being suitable for the
preparation of efficient low-Tg PR blends. Composites’ Tg,
progressively lowered by adding plasticizer, governed the mate-
rial PR response, which turned from orientation-limited to
photoconductivity-limited, and finally suffered strong deterio-
ration due to increased dark conductivity. Nevertheless, an
optimum Tg was found at 6 ꢀC that enabled a fast response
before significant PR strength reduction occurred. Furthermore,
the composite exhibited higher structural stability than prior
DCDHF-based materials, preventing electrical breakdown or
7 (a) P. M. Lundquist, R. Wortmann, C. Geletneky, R. J. Twieg,
M. Jurich, V. Y. Lee, C. R. Moylan and D. M. Burland, Science,
1996, 274, 1182; (b) M. Grasruck, A. Schreiber, U. Hofmann,
S. Zilker, A. Leopold, S. Schloter, C. Hohle, P. Strohriegl and
D. Haarer, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 60,
16543.
8 U. Gubler, M. He, D. Wright, Y. Roh, R. Twieg and W. E. Moerner,
Adv. Mater., 2002, 14, 313.
9 C. Engels, D. Van Steenwinckel, E. Hendrickx, M. Schaerlaekens,
A. Persoons and C. Samyn, J. Mater. Chem., 2002, 12, 951.
10 (a) W. You, L. Wang, Q. Wang and L. Yu, Macromolecules, 2002, 35,
4636; (b) J. Hwang, J. Sohn and S. Y. Park, Macromolecules, 2003, 36,
7970; (c) H. Li, R. Termine, L. Angiolini, L. Giorgini, F. Mauriello
and A. Golemme, Chem. Mater., 2009, 21, 2403.
11 Y. Zhang, T. Wada and H. Sasabe, J. Mater. Chem., 1998, 8, 809.
12 (a) M. J. Cho, D. H. Choi, P. A. Sullivan, A. J. P. Akelaitis and
L. R. Dalton, Prog. Polym. Sci., 2008, 33, 1013; (b) L. R. Dalton,
S. J. Benight, L. E. Johnson, D. B. KnorrJr, I. Kosilkin,
B. E. Eichinger, B. H. Robinson, A. K.-Y. Jen and R. M. Overney,
Chem. Mater., 2011, 23, 430.
beam fanning even at applied fields above 100 V mmꢁ1
.
Outstanding PR performance was achieved (Dn ¼ 0.024, G ¼
350 cmꢁ1 and buildup (erase) times of 35 (20) ms at 120 V mmꢁ1
and low intensity 320 mW cmꢁ2), which is promising for practical
applications. In addition, a NLO homopolymer (3), identical to
copolymer 2 but without carbazole, was synthesized and char-
acterized in PR composites. Although their qualitative behavior
was similar, the homopolymer exhibited inferior PR perfor-
mance, which suggests a favorable role of the carbazole units
attached to the copolymer, contributing to better charge gener-
ation/transport.
13 O. Ostroverkhova, D. Wright, U. Gubler, W. E. Moerner, M. He,
A. Sastre-Santos and R. J. Twieg, Adv. Funct. Mater., 2002, 12,
621.
14 Z. Hou, W. You and L. Yu, Appl. Phys. Lett., 2003, 82, 3385.
15 W. You, Z. Hou and L. Yu, Adv. Mater., 2004, 16, 356.
16 D. Wright, U. Gubler, Y. Roh, W. E. Moerner, M. He and
R. J. Twieg, Appl. Phys. Lett., 2001, 79, 4274.
17 W. You, Z. Hou and L. Yu, Appl. Phys. Lett., 2005, 86, 151906.
18 W. You, S. Cao, Z. Hou and L. Yu, Macromolecules, 2003, 36, 7014.
19 Y. Chen, Y. He, F. Wang, H. Chen and Q. Gong, Polymer, 2001, 42,
1101.
Acknowledgements
We thank support from the Spanish Government MEC and the
European Community (FEDER) (grants MAT2008-06648-C01,
MAT2011-28167-C02-01, CTQ2008-05901/BQU and Con-
solider-Ingenio 2010 project HOPE CSD2007-00007). FGG
thanks the Juan de la Cierva program of the Spanish Govern-
ment MEC (grant JCI-2006-3029-2615) for support.
ꢀ
20 E. Font-Sanchis, F. J. Cespedes-Guirao, A. Sastre-Santos,
B. Villacampa, J. Orduna, R. Alicante and F. Fernandez-Lazaro,
ꢀ
ꢀ
Tetrahedron, 2009, 65, 4513.
21 E. Font-Sanchis, R. E. Galian, F. J. Cespedes-Guirao, A. Sastre-
ꢀ
ꢀ
ꢀ
ꢀ
Santos, L. R. Domingo, F. Fernandez-Lazaro and J. Perez-Prieto,
Phys. Chem. Chem. Phys., 2010, 12, 7768.
Notes and references
22 Z. Li, J. Luo, J. Li, C. Zhan and J. Qin, Polym. Bull., 2000, 45, 105.
23 (a) C. C. Teng and H. T. Man, Appl. Phys. Lett., 1990, 56, 1734; (b)
J. S. Schildkraut, Appl. Opt., 1990, 19, 2839.
1 (a) P. Yeh, Introduction to Photorefractive Nonlinear Optics, Wiley,
New York, 1993; (b) Photorefractive Materials and Their
€
Applications, I and II, ed. P. Gunter and J. P. Huignard, Springer
Series in Optical Science, Springer, Berlin, 2006, vol. 114.
2 (a) O. Ostroverkhova and W. E. Moerner, Chem. Rev., 2004, 104,
3267; (b) J. Thomas, C. W. Christenson, P.-A. Blanche,
M. Yamamoto, R. A. Norwood and N. Peyghambarian, Chem.
24 H. Kogelnik, Bell Syst. Tech. J., 1969, 48, 2909.
25 O. Ostroverkhova and K. D. Singer, J. Appl. Phys., 2002, 91, 9481.
26 A very similar behavior was observed in DCDHF-containing FF
polymers with Tg’s ranging between 26 and 9 ꢀC (ref. 14) or
between 20 and 6 ꢀC (ref. 18).
€
Mater., 2011, 23, 416; (c) S. Kober, M. Salvador and K. Meerholz,
Adv. Mater., 2011, 23, 4725; (d) A. Sastre-Santos, L. Martın-Gomis
€
ꢀ
27 (a) T. K. Daubler, R. Bittner, K. Meerholz, V. Cimrova and
D. Neher, Phys. Rev. B: Condens. Matter Mater. Phys., 2000, 61,
13515; (b) G. B. Jung, T. Mutai, R. Fujimura, S. Ashiara,
T. Shimura, K. Araki and K. Kuroda, Jpn. J. Appl. Phys., Part 1,
2004, 43, 8316; (c) J.-W. Oh, W.-J. Joo, I. K. Moon, C.-S. Choi and
N. Kim, J. Phys. Chem. B, 2009, 113, 1592.
ꢀ
ꢀ
ꢀ
and F. Fernandez-Lazaro, Fullerenes as Photosensitizers in
Photorefractive Materials, in Handbook of Carbon Nano Materials,
ed. F. D’Souza and K. M. Kadish, World Scientific Publishing Co.,
Singapore, 2011.
3 W. E. Moerner, S. M. Silence, F. Hache and G. C. Bjorklund, J. Opt.
Soc. Am. B, 1994, 11, 320.
28 O. Ostroverkhova, M. He, R. Twieg and W. E. Moerner,
ChemPhysChem, 2003, 4, 732.
4 (a) B. Kippelen, S. R. Marder, E. Hendrickx, J. L. Maldonado,
G. Guillemet, B. L. Volodin, D. D. Steele, Y. Enami,
€
29 E. Mecher, R. Bittner, C. Brauchle and K. Meerholz, Synth. Met.,
1999, 102, 993.
30 E. Mecher, F. Gallego-Gomez, K. Meerholz, H. Tillmann,
€
ꢀ
H. Horhold and J. C. Hummelen, ChemPhysChem, 2004, 5, 277.
J. F. Sandalphon, Y. J. Yao, J. F. Wang, H. Rockel, L. Erskine
€
and N. Peyghambarian, Science, 1998, 279, 54; (b) K. Roy
Choudhury, Y. Sahoo and P. N. Prasad, Adv. Mater., 2005, 17,
31 S. M. Silence, G. C. Bjorklund and W. E. Moerner, Opt. Lett., 1994,
19, 1822.
ꢀ
2877; (c) F. Gallego-Gomez, J. A. Quintana, J. M. Villalvilla,
M. A. Dıaz-Garcıa, L. Martın-Gomis, F. Fernandez-Lazaro and
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
€
32 D. Hertel and H. Bassler, ChemPhysChem, 2008, 9, 666.
33 F. Gallego-Gomez, E. M. Garcıa-Frutos, J. M. Villalvilla,
€
A. Sastre-Santos, Chem. Mater., 2009, 21, 2714; (d) S. Kober,
J. Prauzner, M. Salvador, F. B. Kooistra, J. C. Hummelen and
ꢀ
ꢀ
ꢀ
J. A. Quintana, E. Gutierrez-Puebla, A. Monge, M. A. Dıaz-Garcıa
ꢀ
ꢀ
and B. Gomez-Lor, Adv. Funct. Mater., 2011, 21, 738.
K. Meerholz, Adv. Mater., 2010, 22, 1383.
12228 | J. Mater. Chem., 2012, 22, 12220–12228
This journal is ª The Royal Society of Chemistry 2012