Facile photopatterning of polyfluorene for patterned neuronal networks

Article abstract of DOI:10.1039/c1sm05894k

In this paper, we demonstrated a facile photopatterning method that uses photocrosslinkable polyfluorene to fabricate micro-sized photopatterns on transparent indium tin oxide substrate for neuronal patterning. The modified poly(ethyleneimine) (m-PEI) with trimethoxysilane moiety was chemically attached to the hydroxyl group-terminated ITO surface and then the photopatternable polyfluorene derivative was spin coated as a cell-repellent layer onto the m-PEI-coated surface. The well-defined micropatterns were easily created over an entire surface by photocrosslinking of bromoalkyl-substituted polyfluorene (Br-PF) via the radical coupling reaction of a C-Br bond under UV irradiation without an initiator. UV-Vis absorbance, photoluminescence, ATR-FTIR and X-ray photoelectron spectroscopy were used to confirm the photocrosslinking process and the surface composition before and after the photocrosslinking of polyfluorene. The pairing of adhesive m-PEI and repulsive Br-PF effectively guided the neurite outgrowth and controlled neurite extension from individual neurons to the pre-patterned direction with excellent pattern fidelity. Guided neuronal cells were maintained for at least 25 days in vitro without any detachment of neuronal cells during cell culture. A photopatternable polyfluorene derivative in combination with cell-adhesive m-PEI is proved to be an effective way to modify the electrode surface to achieve single cell level neuronal networks.

Full text of DOI:10.1039/c1sm05894k

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
Soft Matter
Cite this: Soft Matter, 2011, 7, 10025
www.rsc.org/softmatter
PAPER
Facile photopatterning of polyfluorene for patterned neuronal networks†
a
Nam Seob Baek, Yong Hee Kim, Young Hwan Han, Bong Joon Lee, Tae-Dong Kim, Sin-Tae Kim,
Young-Seok Choi, Gook Hwa Kim, Myung-Ae Chung and Sang-Don Jung*
a
a
a
b
b
a
a
a
a
Received 15th May 2011, Accepted 28th July 2011
DOI: 10.1039/c1sm05894k
In this paper, we demonstrated a facile photopatterning method that uses photocrosslinkable
polyfluorene to fabricate micro-sized photopatterns on transparent indium tin oxide substrate for
neuronal patterning. The modified poly(ethyleneimine) (m-PEI) with trimethoxysilane moiety was
chemically attached to the hydroxyl group-terminated ITO surface and then the photopatternable
polyfluorene derivative was spin coated as a cell-repellent layer onto the m-PEI-coated surface. The
well-defined micropatterns were easily created over an entire surface by photocrosslinking of
bromoalkyl-substituted polyfluorene (Br-PF) via the radical coupling reaction of a C–Br bond under
UV irradiation without an initiator. UV-Vis absorbance, photoluminescence, ATR-FTIR and X-ray
photoelectron spectroscopy were used to confirm the photocrosslinking process and the surface
composition before and after the photocrosslinking of polyfluorene. The pairing of adhesive m-PEI and
repulsive Br-PF effectively guided the neurite outgrowth and controlled neurite extension from
individual neurons to the pre-patterned direction with excellent pattern fidelity. Guided neuronal cells
were maintained for at least 25 days in vitro without any detachment of neuronal cells during cell
culture. A photopatternable polyfluorene derivative in combination with cell-adhesive m-PEI is proved
to be an effective way to modify the electrode surface to achieve single cell level neuronal networks.
1
0
Introduction
CAM by means of photolithography, soft lithography tech-
11,12
niques like micro-contact printing,
13
or deep UV lithography
Artificial neuronal networks in vitro are a useful tool in studying
brain diseases and examining the communication between
neuronal cells and electronic devices. In order to achieve the
desired neuronal network in vitro, the specific ability of cell-
adhesive molecules (CAM) to create finely patterned substrates,
via some appropriate surface modifications, is required. The
appropriate surface modification that can promote cellular
adhesion and growth in bio-active devices plays a critical role in
techniques in combination with surface chemistry to chemically
14
immobilize CAM on the substrate. Although the soft lithog-
raphy protocol is a well-established method and is commercially
used to fabricate pre-patterned surfaces, expensive instruments
and multistep processes were necessary and it is hard to produce
narrow micropatterns (<10 mm) for a single-cell-based neuronal
network.
In our previous work, we fabricated micron-scale patterns
from the photodegradation of biocompatible modified-poly
1,2
the development of multi-electrode arrays (MEAs) and neu-
3,4
5
rochips or biosensors. Controlling the interface between
patterned electrodes coated with CAM and neurons plays an
important role in the interpretation of the cellular behavior of
(
ethyleneimine) (m-PEI) that was chemically bound to the
15
surface of indium tin oxide (ITO). The photopatterned m-PEI
layer facilitated the uptake of neuronal cells in the designed
pattern and was capable of controlling the neurite outgrowth and
extension along the patterned direction. However, it took
a relatively long exposure time in fabricating the desired patterns
via photodegradation of m-PEI by deep UV (254 nm) irradia-
tion. Even though we can save the exposure time by adopting
a strong intensity UV source, it costs too much. The other
drawback of the fabrication of patterns from the photo-
degradation of m-PEI is that it is hardly possible to get neuronal
networks free from neurons and their connections presenting in
the space areas in the grid-space type pattern. This drawback
results from the exposure of hydrophilic ITO or silicon oxide
surface after developing of photodegraded m-PEI. Although
neuronal cells are not spontaneously adhered to the surface of
6
neuronal networks in MEAs and can significantly improve the
7
device performance. Conventionally, the patterned substrate for
a neuronal network consists of the cell-adhesive region coated
9
8
with a positively charged polymer and the cell-repellent region.
Up to now many researchers have fabricated the patterns of
a
IT Convergence Technology Research Laboratory, Electronics and
Telecommunications Research Institute, Daejeon, 305-700, Republic of
Korea. E-mail: jungpol@etri.re.kr; Fax: +82 42 860 6564; Tel: +82 42
8
60 5557
b
Department of Advanced Materials, Hannam University, Daejeon, 305-
811, Republic of Korea
† Electronic supplementary information (ESI) available: Fig. S1, S2, S3,
S4, S5 and S6. See DOI: 10.1039/c1sm05894k
This journal is ª The Royal Society of Chemistry 2011
Soft Matter, 2011, 7, 10025–10031 | 10025

View Article Online
ITO and silicon oxide they do not completely lose the chance
because of the biocompatible and hydrophilic nature of ITO and
silicon oxide.
monomers are 3.3 ppm and 1.39 ppm, respectively. After
coupling of the dibromo-functionlized monomer (M1) with
dioxoborolan-functionlized monomer (M2), yielding the photo-
crosslinkable polyfluorene, the terminal dioxoborolan proton
In order to save fabrication time and get neuronal patterns
devoid of neurons presenting in the space area we adopted
a photocrosslinking system and a hydrophobic surface instead of
the time-consuming photodegradation and a hydrophilic surface.
For this we have selected polyfluorene (PF) as the photo-
crosslinking system. Owing to the fluorene group, it is hydro-
phobic so that it can repel neuronal cells, resulting in the
confinement of neuronal cells to the neuronal cell-adhesive
region. Biocompatible and water-soluble polyfluorene deriva-
1
had clearly disappeared in the H-NMR. The thermal stability of
the resulting polymer was evaluated by differential scanning
calorimetry (DSC) and thermogravimetric analysis (TGA)
ꢁ1
ꢀ
g
measurements at a heating rate of 10 C min . The T of
ꢀ
ꢀ
synthesized PF is 72 C and its degradation onset is ꢂ260 C. The
first decomposed region is attributed to the degradation of long
alkyl chains and the second one is due to the decomposition of
the polymer backbone. The synthesized Br-PF seems to be
thermally stable and is a suitable material to fabricate photo-
patterned substrate as a cell-repellent layer.
1
6
tives have been used as a probe for DNA sequence detection
17
and a singlet oxygen generator. Small features sized less than
1
mm are routinely obtained from photocrosslinkable poly-
However, many photocrosslinkable polymers need
Scheme 2 shows a schematic diagram of the photopatterning
procedure for the fabrication of micro-sized patterns. Details of
photopatterning procedures and patterned neuronal cell culture
18,19
mers.
a photo initiator, like diphenyiodonium hexafluoroarsenate, for
initiation of photocrosslinking. Because of the potential toxicity
the use of a conventional photo initiator is not a proper solution
for biomedical applications. To solve this we have introduced
bromoalkyl groups into PF. The bromoalkyl-substituted poly-
mers can be sufficiently crosslinked via a radical coupling reac-
tion initiated by the cleavage of the C–Br bond in the absence of
are described as following. The ITO substrate treated with
ꢀ
a mixture of NH
for 1 h was immersed in a 2% m-PEI solution under N
sphere overnight. The positively charged m-PEI are covalently
3
2
: H O
2
: H
2
O (1 : 4 : 20 v/v) solution at 70 C
2
atmo-
23
linked to the pre-treated ITO surface via sol–gel process. 1 wt%
Br-PF solution was then spin-coated (1500 rpm for 30 s) on the
m-PEI surface, resulting in the formation of a pinhole free thin
film. In order to get the designed micropattern the Br-PF coated
substrate was placed in a mask aligner equipped with a constant
20
a photo initiator. Furthermore, the biocompatible crosslinking
of Br-PF occurs using UV light, but not thermal treatment.
Therefore, the nontoxicity and low temperature processing for
the fabrication of patterned substrates are suitable for bio
applications.
ꢁ2
power (P325nm ¼ ꢂ60 mW cm ) supply and then exposed for
360 s without a photo initiator. After UV exposure, the substrate
For these reasons, we have introduced Br-PF layer onto the m-
PEI layer followed by area-selective photoexposure. The photo-
exposed Br-PF is crosslinked and acts as a neuronal cell repellent,
whereas the intact Br-PF is stripped and the naked positively
charged m-PEI attracts negatively charged neuronal cells. Here,
we describe the facile photopatterning characteristics of photo-
crosslinkable Br-PF and neuronal cell-repellent performance in
detail.
3
was rinsed with CHCl solution for 5 min followed by washing
with acetone, resulting in a specific micropattern with a thickness
of ca. 26 ꢃ 1 nm measured by alpha step profilometer for
neuronal pattering.
To confirm the photocrosslinking of polyfluorene, we moni-
tored the changes in ATR-FTIR, UV-Vis absorption and emis-
sion spectra with respect to the exposure time. UV-Vis
absorption and photoluminescence (PL) spectra of polyfluorene
bearing photocrosslinkable substituents are shown in Fig. 1. The
absorbance of the uncrosslinked Br-PF film showed a maximum
at 383 nm, similar to that of a water-soluble polyfluorene
Results and discussion
16,24
derivative
and the fluorescence maximum appeared at
The synthesis of Br-PF is shown in Scheme 1. 2,7-dibromo-
0
a wavelength of 441 nm with well-resolved structures when
excited at 370 nm. After photoexposure, there are almost no
changes in UV-vis absorption behavior. However, the emission
00
0
9
,9 -bis(6 -bromohexyl)fluorene and 2,2 -(9,9-dioctyl-9H-fluo-
rene-2,7-diyl)bis(4,4,5,5-tetramethyl-1,3,2- dioxaborolane) were
21,22
prepared according to the methods described in the literature.
The polyfluorene derivative (Br-PF) was prepared via Suzuki
cross coupling reaction catalyzed by Pd(PPh ) in toluene to give
3
4
photopatternable polymer as a yellowish solid. The chemical
1
structures involved in this study were identified by FT-IR, H-
NMR, GPC, UV-Vis absorption and fluorescence spectroscopy.
1
H-NMR spectra of materials used in this paper are presented in
Fig. S1.†
In the NMR spectrum, the observed d values for bromomethyl
2 3
protons (–CH Br, 2H) and for dioxaborolane (–CH , 12H) in
Scheme 2 A diagram of the photopatterning processes for patterned
Scheme 1 The synthesis of photocrosslinkable polyfluorene derivative.
0026 | Soft Matter, 2011, 7, 10025–10031
neuronal cell growth.
1
This journal is ª The Royal Society of Chemistry 2011

View Article Online
Fig. 1 Normalized UV-Vis absorption and photoluminescence spectra
ex ¼ 370 nm) of Br-PF film before and after photoexposure (360 s).
(
l
Fig. 2 Changes in UV-Vis absorption and PL intensities of crosslinked
spectrum of the photoexposed Br-PF film differs from that of the
unexposed film in that the intensity of green emission peak
increases with increasing exposure time. The rather broad green
emission has been reported for many conjugated polymers and
Br-PF film with respect to exposure time (measured after rinse with
3
CHCl ).
25
attributed to the formation of excimers and/or aggreagates. The
photocrosslinking should reduce the intermolecular distances
between emissive chromophores and thus can increase the
probability to form excimers and/or aggregates through inter-
molecular interaction. The degree of aggregation of polyfluorene
chains is known to be strongly affected by the molecular weight
the progress of photocrosslinking. Since the exposure time
dependence of absorbance, broad green emission and the pattern
thickness are closely linked with each other, we can predict the
pattern thickness from absorbance or emission spectra. For
example, we can get a pattern with the desired thickness through
the in situ monitoring of the broad green emission intensity while
photocrosslinking.
26
27
of the polymer, the type of side-chain, and the end group of
28
the polymer chain. Even though defective ketones generated
Notwithstanding its simplicity, contact angle measurement
enables us to qualitatively and quickly check the many processes
involved in the surface modification. Table 1 summarizes the
contact angles of bare ITO, pre-treated ITO, m-PEI layer and the
photopatternable polyfluorene layer before and after photo-
exposure (Fig. S5 in the ESI†). The contact angle of bare ITO
during the thermal annealing in air can produce green emis-
2
9,30
sions,
the generation of ketones can be avoided by perform-
30
ing the baking and photoexposure under nitrogen atmosphere.
For example, we could not find any significant evidence for the
oxidation of the substituent hydrogens in the 9-position of the
fluorene unit from the ATR-FTIR spectrum before and after
the photocrosslinking process under nitrogen environment
ꢀ
ꢀ
decreased from 58 ꢃ 4 to 35 ꢃ 2 by means of chemical treat-
ment, resulting in an increase of oxygen groups on the ITO
surface. The pre-treated ITO substrate was immersed in 2% m-
PEI solution. The contact angle of the m-PEI layer was slightly
increased. After spin coating of the Br-PF solution onto a m-PEI
(
Fig. S2 in the ESI†). In FT-IR, characteristic bands of m-PEI
coated on ITO substrate, corresponding to N–H stretching and
ꢁ1
ꢁ1
scissoring (broad 3300 cm and 1673 cm ), alkyl C–H (2947–
ꢁ1
ꢁ
1
ꢀ
2
830 cm ), and Si–O–Si (1132–1050 cm ) stretching were
layer, contact angle for unexposed Br-PF film is 98 ꢃ 2 , which
1
5
observed. In the crosslinked Br-PF film, the vibrational peaks
closely matched the linear alkyl-substituted polyfluorene spun
31
ꢁ1
ꢀ
at 2939 and 2861 cm for the C–H stretch vibrations of the
casted on quartz or silicon substrate (100–103 ). Once photo-
crosslinked, the contact angle of exposed Br-PF film did not
ꢁ1
aliphatic side-chain in the fluorene unit and 1612 cm for the
aromatic C]C stretch vibration were observed before and after
photocrosslinking.
significantly change after washing with CHCl solution. Owing
3
to its hydrophobic nature, the patterned Br-PF film can act as
a cell-repellent layer.
In order to see the effect of exposure time on the formation of
patterns and to determine the proper exposure time, Br-PF was
spin-coated onto the quartz plate and exposed for 30 s to 600 s
under nitrogen atmosphere (Fig. S3 in the ESI†). The exposure
was performed with and without a Cr photomask. After exposed
samples were rinsed with chloroform to strip the uncrosslinked
fraction of Br-P, thickness measurements were done with the
samples exposed with the Cr photomask. As shown in Fig. 2(a)
the absorbance and the photocrosslinked film thickness increased
with increasing exposure time, indicating that the photo-
crosslinking progresses gradually. The increase in intensity of the
broad emission band with increasing exposure time, shown in
Fig. 2(b), can be explained in terms of the formation of excimers
and/or aggregates with respect to the exposure time and confirms
To check the chemical compositions of the film, each layer was
evaluated by XPS measurement. XPS spectra of the m-PEI layer
and the Br-PF layer were depicted in Fig. 3. On the chemical
binding of m-PEI to ITO surface, the binding energies of C 1s, N
1s and Si 2p atoms in the m-PEI moiety were observed at 284.5
eV (C–C/C–H) and 286.4 eV (C–N), 398.9 eV and 101.9 eV,
15
respectively. To quantitatively analyze the chemical composi-
tions and demonstrate the cleavage of C–Br bond during the
photocrosslinking process, 1wt% Br-PF solution was spin-coated
onto the quartz substrate and anlyzed by XPS. The photo-
crosslinking process was monitored by the decrease in the
intensity of Br 3d peak under the same experimental conditions
(Fig. 3). In unexposed Br-PF film, peaks of 284.5 eV for C 1s,
This journal is ª The Royal Society of Chemistry 2011
Soft Matter, 2011, 7, 10025–10031 | 10027

View Article Online
Table 1 Contact angles of bare ITO, pre-treated ITO, m-PEI layer and Br-PF layers before and after photoexposure
a
b
c
d
Substrate
Bare ITO
Pre-treated ITO
m-PEI
Unexposed Br-PF
Exposed Br-PF
ꢀ
Contact angle ( )
58 ꢃ 4
chemical treatment. surface modification by sol–gel reaction. spin-coated film. after photocrosslinking.
35 ꢃ 2
44 ꢃ 1
98 ꢃ 2
93 ꢃ 2
a
b
c
d
with various features to see the photopatterning fidelity of Br-
PF.
Fig. 4 shows a fluorescence image of photocrosslinked patterns
and intensity profiles of the patterns. In particular, the pattern of
Fig. 4(d) is designed to fit to a conventional 64-channel MEA
feature with 10 mm line width and 200 mm spacing. Even though,
due to lack of a photomask, we obtained patterns with the
minimum feature size of 4 mm, this result does not limit pattern
fidelity of Br-PF to 4 mm and it seems possible to get even
submicron patterns. In patterned neuronal growth, single neuron
level connections are required so that a pattern line width of 6–10
mm is a reasonable window considering the size of a single cell,
around 10 mm in diameter. This consideration leads to the fact
that the photocrosslinkable Br-PF, in combination with its
Fig. 3 (top) XPS spectra of m-PEI layer on ITO substrate and (bottom)
Br-PF layer cast on quartz substrate (relative ratio of C 1s to Br 3d
species as a function of exposure time).
17
biocompatibility, has sufficient pattern resolution suitable for
use in patterned neuronal growth.
To verify the neuronal cell confining ability of hydrophobic
Br-PF, we have cultured hippocampal neuronal cells on the grid-
type micropattern, (d) in Fig. 5. The outmost surface of the grid
and the space is consisted of m-PEI and Br-PF, respectively. A
feature spacing of 200 mm was maintained for potential appli-
cation of this method to commercially available MEAs.
Compared to straight lanes, the grid-space type micropattern
provides substrates for long-term cell growth so that neuronal
cells were maintained for longer than 25 days in vitro (DIV).
5
32.2 eV for O 1s and 101.8 eV for Si 2p were observed. Also,
characteristic peaks at 71 eV for Br 3d, 185 eV and 192 eV for Br
p and 259 eV for Br 3s were detected, corresponding to carbon-
bound bromine in the Br-PF (Wide-scan XPS spectra in
3
32
Fig. S6†).
Analysis of the C 1s : Br 3d ratio as a function of exposure
time was carried out. Apparently, the Br 3d intensity becomes
gradually decreased with increasing exposure time from pristine
film (2.29%) to 600 s dose (1.86%). Additionally, the intensity of
the C 1s component is relatively increased from 0 s (97.71%) to
6
00 s dose (98.14%), suggesting that photocrosslinking of Br-PF
film occurs due to the radical coupling reaction initiated by the
cleavage of the C–Br bond. These results are consistent with the
degrees of crosslinking of the Br-PF film. As a result, it is likely
that photopatternable Br-PF is easily formed into a well-defined
lattice with a relatively short exposure time. We have arbitrarily
fixed 360 s as a proper exposure time and fabricated patterns
Fig. 5 (a) Fluorescence micrographs of the grid micropattern (360 s UV
3
exposed and rinsed with CHCl solution), (b) cells grown on patterned
ITO substrate (ICR mouse 15–16 days), (c) and (d) fluorescence images of
stained hippocampal neuronal cells (blue ¼ DAPI and green ¼ b-III
tubulin). The phase-contrast micrograph and fluorescent micrographs
4
ꢁ2
Fig. 4 Fluorescence images and intensity profiles of photocrosslinked
Br-PF pattern on quartz substrate. All scale bars are 100 mm.
were taken 4 days after cell seeding (5.0 ꢄ 10 cells cm ). All scale bars
are 100 mm (Line width ¼ 40 mm and spacing ¼ 200 mm).
1
0028 | Soft Matter, 2011, 7, 10025–10031
This journal is ª The Royal Society of Chemistry 2011

View Article Online
We have conducted immunofluorescent staining of cortical
neurons using DAPI and b-tubulin as a specific antibody for
clear observation. As expected, we could rarely find neuronal
cells adhered to the space area. In our previous report, we were
not able to completely confine neuronal cells in the grid area and
some neuronal cells were frequently resident in the space area,
materials useful for confining neuronal cells within a desired site,
such as an electrode in MEAs.
Experimental
Materials and instruments
15
which consisted of ITO. This dramatic enhancement of cell
confinement is mainly caused by the cell-repelling role of
hydrophobic Br-PF. Hydrophobic Br-PF repels not only the
neuronal cell body but also the neurites so that the extension of
neurites crossing the space zone is completely restricted. The
combination of neuronal cell adhesive m-PEI and repelling Br-
PF would be effective in acquiring the single neuronal cell signal
from the interface between neuronal tissue and electrode by
increasing the probability of the electrode to face a single
neuronal cell. Even though this approach is not limited to the
combination of m-PEI and Br-PF, Br-PF is advantageous in that
facile processing is possible.
1,6-Dibromohexane, 2,7-dibromofluorene, tetrabutylammonium
bromide (TBAB), 1-bromooctane, tetrakis(triphenylphosphine)
palladium(0) (Pd(PPh ) ), 2-isopropoxy-4,4,5,5-tetramethyl-
3
4
1,3,2-dioxaborolane, n-butyllithium (2.5 M in hexane), 20% tet-
rabutylammonium hydroxide (TBAH) and anhydrous toluene
were purchased from Aldrich Co., and used without further
purification. Tetrahydrofuran (THF) was distilled over sodium
benzophenone ketyl under nitrogen prior to use. Trimethoxy-
silylpropyl modified poly(ethyleneimine) (M 1500–1800, 50% in
w
1
13
isopropanol) was purchased from Gelest Inc. H-NMR and C-
NMR spectra were recorded with the use of Varian Oxford 300
MHz spectrometers; chemical shifts were reported in ppm units
with tetramethylsilane as an internal standard and the mass
spectra were taken by a JEOL JMS-AX505WA mass spectrom-
eter. The attenuated total reflection-Fourier transform infrared
(ATR-FTIR) spectrum was taken on a Nicolet iN10 spectrom-
eter with Ge crystal. X-ray photoelectron spectroscopy (XPS)
were performed by ESCALAB 200R instrument (VG scientific)
using monochromized Al-Ka operated at 1486.6 eV, 12.5 kV and
250 W X-ray radiation. The base pressure of the system was 5.0 ꢄ
We have also tried the single cell level neuronal network by
reducing the line width of the grid from 40 mm to 10 mm. As can
be seen in Fig. 6, a clear single cell level neuronal network was
achieved with excellent pattern fidelity. This result would, when
coupled with MEA, enable us to better understand the funda-
mental meaning of activity-dependent neuronal networks and
neuron–neuron interaction as well as neuron–surface interaction.
The directional control of patterned growth via electric field or
gradient profile of adhesive molecules should reinforce the
precise control of neuronal networks both at the cellular and
neurite level.
ꢁ10
ꢀ
Torr. XPS spectra acquired at 90 takeoff angle (TOA)
10
relative to the surface. The molecular weight and polydispersity
were analyzed by a Waters 1515 gel permeation chromatograph
(
(
GPC) with a refractive index detector at room temperature
THF as the eluent). Steady-state absorption spectra were
Conclusions
We have synthesized photocrosslinkable Br-PF, spectroscopi-
cally analyzed the photocrosslinking process, and fabricated
various patterns via photocrosslinking. The photocrosslinking
process is facile enough that the typical photoexposure time
required is a few minutes which is much shorter than the pho-
tobleaching time of 90 min for m-PEI. From the grid-space type
pattern with the outmost surface of the grid and the space con-
sisting of m-PEI and Br-PF, respectively, we could confirm the
neuronal cell-repellant property of Br-PF and hence obtain
excellent pattern fidelity of primarily cultured neuronal cells. We
have also achieved a clear single cell level neuronal network from
the grid with a line width of 10 mm. Finally, we add the
biocompatible, hydrophobic, photocrosslinkable without an
initiator, and fluorescent Br-PF as a member of the patternable
recorded by a Shimadzu UV-2401PC spectrophotometer and
photoluminescence spectra were measured by steady-state fluo-
rimeter (Edinburgh FS920) with 450 W Xe-lamp.
0
00
21
(M1) .
2
,7-dibromo-9,9 -bis(6 -bromohexyl)fluorene
To
mixture of 2,7-dibromofluorene (1.0 g, 3.09 mol), 1,6-dibromo-
hexane (6.12 g, 40.12 mol) and catalytic amounts of tetrabuty-
lammonium bromide in DMSO, 3 mL of 50% aqueous NaOH
was added. The reaction mixture was stirred for 24 h at r.t. and
then poured into dilute hydrochloric acid. The aqueous layer was
extracted by ethyl acetate. The combined extracts were washed
with water and NaHCO solution and then dried over MgSO .
3
4
Removal of the solvent under vacuum followed by flash chro-
matography using hexane/dichloromethane produced a white
ꢀ
solid (1.41 g, 70%). mp ¼ 70–72 C; d
H 3 4
(300MHz; CDCl ; Me Si)
7
1
.50 (2H, d), 7.43 (2H, d), 7.41 (2H, s), 3.28 (4H, t), 1.90 (4H, t),
.65 (4H, m), 1.18 (4H, m), 1.06 (4H, m), 0.56 (4H, m); d
; Me
5,5, 39.9, 33.6, 32.5, 28.9, 27.5, 23.4.
C
(
75MHz; CDCl
3
4
Si) 152.1, 139.0, 130.4, 126.3, 121.6, 121.2,
5
2
2
2
,7-Dibromo-9,9-dioctyl-9H-fluorene .
,7-dibromofluorene (1 g, 3.09 mol), 1-bromooctane (1.1 mL,
.48 mol) and catalyst amounts of tetrabutylammonium bromide
To
mixture
of
2
6
in 10 mL of DMSO, 3 mL of 50% aqueous NaOH was added.
ꢀ
The reaction mixture was stirred and heated at 100 C for 10 h
Fig. 6 A single cell level neuronal network formed on a pattern with
4
ꢁ2
a grid line width of 10 mm and spacing of 200 mm (1.0 ꢄ 10 cells cm , 3
DIV). Scale bars are 100 mm.
and then poured into dilute hydrochloric acid. The aqueous layer
was extracted by dichloromethane. The combined extracts were
This journal is ª The Royal Society of Chemistry 2011
Soft Matter, 2011, 7, 10025–10031 | 10029

View Article Online
washed with water and NaHCO solution and then dried over
3
m-PEI. For spin coating 10 mg of photopatternable Br-PF was
MgSO . The mixture was purified by flash chromatography with
4
dissolved in 1 mL CHCl and filtered through a 0.2 mm PTFE
3
an eluent of hexane to give a white solid (1.4 g, 83%). mp 45–
filter prior to spin coating. The dissolved polymer solution was
spin-coated onto the m-PEI-coated ITO substrate at 1500 rpm
ꢀ
4
7
7 C (from hexane); d
H
(300MHz; CDCl
3
; Me
4
Si) 7.52 (2H, d),
ꢀ
.44 (2H, d), 7.43 (2H, s), 1.90 (4H, m), 1.27–1.02 (20H, m), 0.82
(75MHz; CDCl ; Me Si) 152.5, 139.0,
for 30 s followed by baking at 80 C on a hotplate for 70 s to
(
6H, t), 0.57 (4H, m); d
C
3
4
remove solvent. Photocrosslinking was carried out using Cr
1
1
30.2, 126.1, 121.3, 121.1, 55.5, 40.2, 31.8, 29.7, 29.1, 23.5, 22.7,
4.2.
mask on a mask aligner (MIDAS System Co., MDA-400M,
ꢁ2
l
325nm ¼ 60 mW cm ) as shown in Scheme 2.
0
2
,2 -(9,9-dioctyl-9H-fluorene-2,7-diyl)bis(4,4,5,5-tetramethyl-
33
Neuronal cell culture and immunofluorescent staining
1
,3,2-dioxaborolane) (M2) . n-Butyllithium (2.5 M in hexane; 2.4
mL, 6.0 mmol) was added slowly over 30 min to a stirred solution
0
A primary neuronal cell suspension was prepared from the
cerebral cortex of fetal ICR mice (15–16 days of gestation) as
34
of 2,7-dibromo-9,9 -dioctylfluorene (1.5 g, 2.7 mmol) in dry
ꢀ
tetrahydrofuran (70 mL) at ꢁ78 C. The mixture was stirred for
previously described with minor modifications. The cell
1
h at the same temperature and then 2-isopropoxy-4,4,5,5-tet-
suspension was then plated on photopatterned substrate in a 12
4
ramethyl-1,3,2-dioxaborolane (1.7 mL, 8.2 mmol) was added
slowly. The mixture was allowed to warm up slowly to room
temperature and stirred vigorously overnight. The reaction was
quenched with water and the residue was extracted with
dichloromethane. The combined organic phase was dried over
well culture plate at a cell density of 5 ꢄ 10 . The culture medium
of neurobasal medium (Invitrogen) was supplemented with a 2%
2
B27 supplement and 1% N supplement, 2.0 mM glutamine with
antibiotics. To prevent the proliferation of non-neuronal cells,
cytosine arabinoside (3 mM) was added to the cultures 24 h after
plating. The culture medium was replaced with fresh medium
4
MgSO and evaporated under reduced pressure. The crude
product was purified by flash chromatography on silica gel and
recrystallised from acetone to give the title compound as col-
ourless crystals (1.3 g, 74%). d (300MHz; CDCl ; Me Si) 7.83
without cytosine arabinoside every 3 days. The cells were
ꢀ
cultured at 37 C in a humidified atmosphere of 5% CO and 95%
2
H
3
4
0
air. Anti-b-tubulin (Covance, Emeryville, CA) and 4 ,6-dia-
(
(
2H, d), 7.74–7.72 (4H, m), 1.96 (4H, m), 1.42 (24H, s), 1.26–1.05
20H, m), 0.81 (6H, t), 0.62 (4H, m); d (75MHz; CDCl ; Me Si)
50.5, 143.5, 133.3, 128.6, 119.4, 119.3, 83.7, 55.5, 40.1, 31.8,
midino-2-phenylindole dihydrochloride (Invitrogen, Carlsbad,
CA) were used for identification of neurons by immunofluores-
15
C
3
4
1
2
cent staining and the detailed procedure is described elsewhere.
9.8, 29.1, 29.0, 24.9, 23.7, 22.7, 14.1.
0
0
Poly[(9,9 -dioctylfluorene)-co-alt-(9,9 -bis(6-bromohexyl)fluo-
Acknowledgements
rine)] (Br-PF). Compound M1 (200 mg, 0.31 mmol), compound
M2 (198 mg, 0.31 mmol), and tetrakis(triphenylphosphine)
palladium(0) (18 mg, 0.02 mmol) were dissolved in 10 mL of
anhydrous toluene at room temperature under a nitrogen
The authors greatly acknowledge the financial support from
Ministry of Education, Science and Technology (Global Part-
nership Program, Project No. 2009-00503) and Electronics and
Telecommunications Research Institute (Project No. 11ZC1130).
T.-D. Kim acknowledges the Korean Research Council Indus-
trial Science and Technology (SK-0903-01).
atmosphere. 3 mL of 20% Bu NOH were added and stirred for
4
3
0 min. The mixture was stirred vigorously at reflux for 2 days.
After reaction, the mixture was poured into methanol. The crude
polymeric product was filtered off and dissolved in dichloro-
methane. The solution was filtered through a m-membrane filter
Notes and references
(0.25 mm) to remove residual catalyst particles and reprecipitated
in methanol. Finally the polymer was purified by Soxhlet
1 I. H. Lelong, V. Petegnief and G. Rebel, J. Neurosci. Res., 1992, 32,
5
62–568; A. I. Belenkov, V. Y. Alakhov, A. V. Kabanov,
S. V. Vinogradov, L. C. Panasci, B. P. Monia and T. Y. K. Chow,
Gene Ther., 2004, 11, 1665–1672.
extraction with acetone and methanol. The polymer was dried in
ꢀ
a vacuum oven at 50 C for 1 day, yielding 230 mg (69%)
yellowish solid. GPC (THF, polystyrene standard) d (300MHz;
H
2 B. Liu, J. Ma, E. Gao, Y. He, F. Cui and Q. Xu, Biosens. Bioelectron.,
2008, 23, 1221–1228.
CDCl ; Me Si) 7.85–7.62 (12H, m), 3.15 (4H, m), 2.2–2.02 (4H,
3
4
3
4
5
6
M. Maher, J. Pine, J. Wright and Y.-C. Tai, J. Neurosci. Methods,
999, 87, 45–56.
Q. Liu, H. Cai, Y. Xu, Y. Li, R. Li and P. Wang, Biosens. Bioelectron.,
006, 22, 318–322.
J. K. Chapin, Nat. Neurosci., 2004, 7, 452–455; C. T. Moritz,
S. I. Perlmutter and E. E. Fetz, Nature, 2008, 456, 639–642.
J. C. Chang, G. J. Brewer and B. C. Wheeler, Biosens. Bioelectron.,
m), 1.72 (4H, m), 1.26–1.05 (32H, m), 0.85 (14H, m); M ¼ 1.2 ꢄ
n
1
4
ꢁ1
1
0 g mol , PDI ¼ 2.3.
2
Surface modification and photopatterning
Before surface modification, ITO substrates were cut into 1 cm ꢄ
2001, 16, 527–533; J. C. Chang and G. J. Wheeler, J. Neural Eng.,
2006, 3, 217–226.
1
1
cm and sonicated in acetone, methanol and pure water for
0 min each, followed by immersing into NH : H O : H O
7
8
V. S. Polikov, P. A. Tresco and W. M. Reichert, J. Neurosci. Methods,
2
3
2
2
2
005, 148, 1–18.
ꢀ
mixture solution (1 : 4 : 20 v/v) at 70 C for 1 h. After which the
ITO surface was silanized with m-PEI by immersing in 2% m-PEI
in anhydrous EtOH for 12 h at room temperature followed by
J. C. Chang, G. J. Brewer and B. C. Wheeler, Biomaterials, 2003, 24,
2863–2870; F. Patolsky, B. P. Timko, G. Yu, Y. Fang, A. B. Greytak,
G. Zhang and C. M. Lieber, Science, 2006, 313, 1100–1104.
ꢀ
9 W. C. Chang and D. W. Sretavan, Langmuir, 2008, 24, 13048–13057;
K. Kang, G. Kang, B. S. Lee, I. S. Choi and Y. Nam, Chem.–Asian J.,
baking at 120 C for 10 min to reinforce covalent binding of
silanol moieties. The m-PEI-coated substrate was sonicated in
a bath of ethanol for 5 min for removal of non-specifically bound
2
010, 5, 1804–1809; A. M. Leclair, S. S. G. Ferguso and F. Lagugn ꢀe -
Labarthet, Biomaterials, 2011, 32, 1351–1360.
1
0030 | Soft Matter, 2011, 7, 10025–10031
This journal is ª The Royal Society of Chemistry 2011

View Article Online
1
1
0 T. G. Ruardij, M. H. Goedbloed and W. L. C. Rutten, IEEE Trans.
Biomed. Eng., 2000, 47, 1593–1599; H. Sorribas, C. Padeste and
L. Tiefenauer, Biomaterials, 2002, 23, 893–900.
1 C. S. Chen, M. Mrksich, S. Huang, G. M. Whitesides and
D. E. Ingber, Biotechnol. Prog., 1998, 14, 356–363; L. Yan,
W. T. S. Huck, X.-M. Zhao and G. M. Whitesides, Langmuir, 1999,
21 X.-H. Zhou, J.-C. Yan and J. Pei, Macromolecules, 2004, 37, 7078–
7080; H.-H. Lu, C.-Y. Liu, T.-H. Jen, J.-L. Liao, H.-E. Tseng,
C.-W. Huang, M.-C. Hung and S.-A. Chen, Macromolecules, 2005,
38, 10829–10835.
22 M.-C. Hung, J.-L. Liao, S.-A. Chen, S.-H. Chen and A.-C. Su, J. Am.
Chem. Soc., 2005, 127, 14576–14577.
1
5, 1208–1214; L. Lauer, C. Klein and A. Offenh €a usser,
Biomaterials, 2001, 22, 1925–1932.
2 L. Lauer, C. Klein and A. Offenh €a usser, Biomaterials, 2001, 22, 1925–
1932; C. G. Specht, O. A. Williams, R. B. Jackman and R. Schoepfer,
Biomaterials, 2004, 25, 4073–4078; H. D. Hwang, G. Kang,
J. H. Yeon, Y. Nam and J.-K. Park, Lab Chip, 2009, 9, 167–
23 C. J. Brinker and G. W. Scherer, in Sol–Gel Science: The Physics and
Chemistry of Sol–Gel Processing, Elsevier, Oxford, UK, 1990, Ch. 2,
pp. 97–228.
24 Q.-H. Xu, B. S. Gaylord, S. Wang, G. C. Bazan, D. Moses and
A. J. Heeger, Proc. Natl. Acad. Sci. USA, 2004, 101, 11634–11639.
25 J. Teetsov and M. A. Fox, J. Mater. Chem., 1999, 9, 2117–2122;
G. Zeng, W.-L. Yu, S.-J. Chua and W. Huang, Macromolecules,
2002, 35, 6907–6914.
26 G. Klaerner and R. D. Miller, Macromolecules, 1998, 31, 2007–2009.
27 C.-H. Chou, S.-L. Hsu, K. Dinakaran, M.-Y. Chiu and K.-H. Wei,
Macromolecules, 2005, 38, 745–751; H.-H. Lu, C.-Y. Liu, T.-H. Jen,
J.-L. Liao, H.-E. Tseng, C.-H. Huang, M.-C. Hung and S.-A. Chen,
Macromolecules, 2005, 38, 10829–10835.
28 J.-I. Lee, V. Y. Lee and R. D. Miller, ETRI J., 2002, 24, 409–414.
29 L. Romaner, A. Pogantsch, P. S. De Freitas, U. Scherf, M. Gaal,
E. Zojer and E. J. W. List, Adv. Fuct. Mater., 2003, 13, 597–601.
30 V. N. Bliznyuk, S. A. Carter, J. S. Scott, G. Kl €a rner, R. D. Miller and
D. C. Miller, Macromolecules, 1999, 32, 361–369; X. Gong, P. K. Iyer,
D. Moses, G. C. Bazan, A. J. Heeger and S. S. Xiao, Adv. Funct.
Mater., 2003, 13, 325–330.
1
1
70.
3 B. Lom, K. E. Healy and P. E. Hockberger, J. Neurosci. Methods,
993, 50, 385–397.
1
1
1
1
1
4 A. Reska, P. Gasteier, P. Schulte, M. Moeller, A. Offenh €a usser and
J. Groll, Adv. Mater., 2008, 20, 2751–2755.
5 N. S. Baek, J.-H. Lee, Y. H. Kim, B. J. Lee, G. H. Kim, I.-K. Kim,
M.-A. Chung and S.-D. Jung, Langmuir, 2011, 27, 2717–2722.
6 K.-Y. Pu and B. Liu, Biosens. Bioelectron., 2009, 24, 1067–1073;
X. Duan, L. Liu, F. Feng and S. Wang, Acc. Chem. Res., 2010, 43,
260–270.
1
1
1
7 Y. Tain, C.-Y. Chen, H.-L. Yip, W.-C. Wu, W.-C. Chen and
A. K.-Y. Jen, Macromolecules, 2010, 43, 282–291.
8 Q. Ji, X. Jiang and J. Yin, Langmuir, 2007, 23, 12663–12668;
E. Scheler and P. Strohriegl, J. Mater. Chem., 2009, 19, 3207–3212.
9 J. Y. Park, J. H. Lee and J.-B. Kim, Eur. Polym. J., 2008, 44, 3981–
31 M. Beinhoff, A. T. Appapillai, L. D. Underwood, J. E. Frommer and
K. R. Carter, Langmuir, 2006, 22, 2411–2414.
3986; P.-H. Wang, M.-S. Ho, S.-H. Yang, K.-B. Chen and
C.-S. Hsu, J. Polym. Sci., Part A: Polym. Chem., 2010, 48, 516–524.
0 Y. Tang, L. Ji, R. S. Zhu, Z. R. Wei and B. Zhang, J. Phys. Chem. A,
32 Y. Ofir, N. Zenou, I. Goykman and S. Yitzchaik, J. Phys. Chem. B,
2006, 110, 8002–8009; M. R. Lockett and L. M. Smith, Langmuir,
2009, 25, 3340–3343.
2
2005, 109, 11123–11126; K. S. Lee, K. Y. Yeon, K. H. Jung and
S. K. Kim, J. Phys. Chem. A, 2008, 112, 9312–9317; B. J. Kim,
Y. Miyamoto, B. Ma and J. M. J. Fr ꢀe echt, Adv. Funct. Mater.,
33 B. Liu and G. C. Bazan, J. Am. Chem. Soc., 2006, 128, 1188–1196;
B. Liu and S. K. Dishari, Chem.–Eur. J., 2008, 14, 7366–7375.
34 X. Q. Wang and S. P. Yu, J. Neurochem., 2005, 93, 1515–1523.
2009, 19, 2273–2281.
This journal is ª The Royal Society of Chemistry 2011
Soft Matter, 2011, 7, 10025–10031 | 10031