nmat 2006

ARTICLES
Robust free-standing nanomembranes of
organic/inorganic interpenetrating
networks
RICHARD VENDAMME1
, SHIN-YA ONOUE1
, AIKO NAKAO2
AND TOYOKI KUNITAKE1
*
1
Topo Chemical Design Laboratory, Frontier Research System (FRS), The Institute of Physical and Chemical Research (RIKEN), Hirosawa 2-1, Wako-shi,
Saitama 351-0198, Japan
2
Surface Science Division, The Institute of Physical and Chemical Research (RIKEN), Hirosawa 2-1, Wako-shi, Saitama 351-0198, Japan
*e-mail: kunitake@ruby.ocn.ne.jp
Published online: 21 May 2006; doi:10.1038/nmat1655
Hybrid sol–gel materials have been a subject of
intensive research during the past decades because
these nanocomposites combine the versatility of organic
polymers with the superior physical properties of glass.
Here, we report the synthesis, by spin coating, of hybrid
interpenetrating networks in the form of free-standing
nanomembrane (around 35-nm thick) with unprecedented
macroscopic size and characteristics. The quasi-2D
interpenetration of the organic and inorganic networks
brings to these materials a unique combination of
properties that are not usually compatible within the
same film: macroscopic robustness and homogeneity,
nanoscale thickness, mechanical strength, high flexibility
and optical transparency. Interestingly, such free-standing
nanofilms of macroscopic size can seal large openings,
are strong enough to hold amounts of liquid 70,000 times
heavier than their own weight, and are flexible enough to
reversibly pass through holes 30,000 times smaller than
their own size.
F
reely suspended ultrathin (thickness < 100 nm) films have
been a theoretical and experimental curiosity for several
decades because, with macroscopic sizes and molecular-
scale thickness, they combine at the same time the properties
of macroscopic materials and colloids along with individual
molecules1
. Such films may be used as permselective membranes or
as scaffolds for the organization of nanoparticles. They are also of
great interest for use in microelectromechanical devices in the form
of sensors or actuators that endure repeated elastic deformation2
.
The recent miniaturization of these devices has created a demand
for thinner membranes with a broad range of properties, such
as defect-free uniformness, macroscopic stability and elasticity
together with nanometre thickness3
. However, the synthesis of
materials with such conflicting features is still a challenge and
largely remains to be explored. To date, several approaches
have been implemented for the fabrication of free-standing
nanoscale films from polymers and/or from inorganic materials:
cast films4
, layer-by-layer (LbL) assembly of polyelectrolyte
multilayers5–9
, crosslinking of amphiphilic Langmuir–Blodgett10
and self-assembled11,12
monolayers, and assembly of triblock
copolymers13
. However, attempts to fabricate highly compliant
nanomembranes have not been very successful, producing instead
free-standing films with lateral dimensions restricted to the
micrometre range and that are too fragile to sustain significant
mechanical stress or to endure changing environments. To avoid
such problems, relatively thick (a fraction of a micrometre)
membranes are usually fabricated. Recently, Jiang and Tsukruk
made significant progress in this area by preparing freely suspended
nanomembranes (with a size up to 1.2 mm2
) from polyelectrolyte
multilayers by means of a spin-assisted LbL process2,14,15
. They
demonstrated the superior mechanical properties of these materials
and showed that organized arrays of nanostructures, such as gold
nanoparticles16
or carbon nanotubes17
can be embedded in the
films, leading to anisotropic optical or mechanical properties.
The LbL technique is versatile18
and can include a wide variety
of molecules, including neutral elements19
, but it is often time
494 nature materials VOL 5 JUNE 2006 www.nature.com/naturematerials
NaturePublishing Group©2006

ARTICLES
(i) Spin-coating of underlayers
(iii) Polymerization/crosslinking (iv) Nanofilm recovery in ethanol
(ii) Spin-coating of precursors
Silicon wafer or glass plate
Sacrificial layer (100 nm)
Ultraviolet + moisture
Interpenetrated networks
Liquid solution of monomers
Air flow
Air shear force
PVA (5 nm)
35 nm
5 cm
a
b
3 cm
Figure 1 Preparative procedure for self-supporting hybrid IPN nanofilm. a, Spin-coating of underlayers (i); spin-coating of the reactive formulation containing both
organic and inorganic precursors (ii); formation of the two networks by different mechanisms (iii); recovery of the ultrathin film by dissolution of the sacrificial underlayer in
ethanol (iv). b, Micrograph of an IPN hybrid nanofilm detached from the substrate and floating in ethanol.
consuming. As a result, efficient alternative methods for the
preparation of free-standing nanofilms are highly desired.
In our laboratory, we developed a facile spin-coating process for
the preparation of self-supporting ultrathin films of titania (TiO2)
and other metal oxides. Using this technique, macroscopically
uniform and self-supporting metal-oxide-gel films with thicknesses
of several decades of nanometres and lateral dimensions of
several centimetres were obtained20,21
. Unfortunately, the inorganic
networks are rather brittle, and the gel films are readily breakable
after dissolution of the sacrificial layer. To overcome the lack
of flexibility and robustness of the inorganic matrix, in this
paper we introduce a new method for the synthesis of ultrathin
organic–inorganic hybrid films with an interpenetrating network
(IPN) structure. Recently, hybrid IPNs in the bulk state have
received increasing attention because these materials combine the
advantages of polymers (versatility, flexibility and light weight)
with the physical properties of glass (heat and mechanical
resistance)22–25
. They are known to have distinct morphology
and to have properties different from simple sums of those of
the original polymers and gels. Typically, the inorganic phase
is formed in situ in a polymer matrix by the sol–gel process
and, owing to this procedure the growing inorganic structures
are homogeneously distributed within the organic polymer on
a molecular scale26
. However, although blending two different
components to combine their advantages is an attractive idea, it is
not always an easy matter, because two components with different
physical properties generally have poor compatibility27
. In this
paper, we prepare self-supporting ultrathin IPN hybrids and show
that these quasi-2D objects enable the preparation of robust free-
standing nanomembranes with dimensions and properties that
have not been reported to date.
Our method, as shown schematically in Fig. 1a, is rather
simple. In a preliminary step, reactive formulations containing
both the organic and inorganic precursors were prepared and
diluted in chloroform. The inorganic precursor consists of
zirconium tetrabutoxide Zr(BuOn
)4 and the organic precursors
are a combination of a monofunctional monomer, 4-hydroxy
butylacrylate (HOBuA), a bifunctional crosslinking agent,
hexanediol diacrylate (HDODA), and the photo-initiator Darocur
4165. Three hybrid formulations with varying amounts of
zirconium tetrabutoxide (denoted as Hybrid 1, Hybrid 2 and
Hybrid 3) were prepared in addition to a pure organic and a pure
nature materials VOL 5 JUNE 2006 www.nature.com/naturematerials 495

ARTICLES
Table 1 Composition of the reactive formulations and of the corresponding cured hybrid nanofilms.
Sample Molar ratios (in mol%) of the organic and inorganic precursors Elemental composition (in mol%) of the IPN as obtained by XPS
contained in the reactive formulations before spin-coating (and as calculated from molar ratios of reactive formulations)*
Zr(BuOn
)4 HOBuA HDODA Initiator Carbon Oxygen Zirconium ZrO2
Organic — 95.6 2.45 1.96 — — — —
Hybrid 1 26.8 69.9 1.79 1.43 63.9 (64.2) 32.4 (32.6) 3.75 (3.20) 11.2 (9.6)
Hybrid 2 42.3 55.2 1.41 1.13 58.1 (58.5) 36.0 (35.6) 5.94 (5.85) 17.8 (17.6)
Hybrid 3 52.4 45.5 1.17 0.93 56.6 (53.8) 35.9 (38.1) 7.49 (8.06) 22.5 (24.2)
Zirconia 100 — — — — — — —
* The elemental compositions of the hybrid nanofilms were determined by XPS, using 40-nm-thick IPN nanofilms directly synthesized on silicon wafers. The IPN compositions were calculated from the precursor molar ratios in the initial
solution, assuming no preferential loss of a monomer during the spin-coating process and that conversions to organic and inorganic networks were complete.
Zr(BuOn
)4 formulation as shown in Table 1. All of the reactive
formulations were homogeneous, and could be stored for a few
days without hydrolysis of the inorganic precursor.
Preparation of self-supporting hybrid nanomembranes was
carried out as follows. First, a 100-nm-thick ‘sacrificial’ layer
of poly(vinylphenol) was spin-coated onto a clean silicon wafer,
followed by a 5-nm-thick layer of poly(vinyl alcohol) (PVA).
The PVA layer provides a stable basement for the preparation
of the IPN hybrid, by preventing the chloroform solution of
monomers from partially dissolving the sacrificial layer, and enables
a smooth detachment of the IPN nanofilm. Then, the reactive
formulation was spin-coated in a nitrogen atmosphere under
ultraviolet light. The ultraviolet radiation induced decomposition
of the photo-initiator and the subsequent radical polymerization
to the poly(HOBuA-co-HDODA) network. Simultaneously, the
residual humidity in the PVA underlayer induced hydrolysis
of Zr(BuOn
)4 and its condensation into the ZrO2 network by
means of a sol–gel process. The combination of the ultraviolet
photopolymerization and the sol–gel process allows simultaneous
growth of the two different networks during the spin-coating
process at room temperature. Then, the edges of the sample
were scratched with a needle to facilitate the film detachment,
and the substrate was immersed in ethanol, where the sacrificial
underlayer was dissolved. The ultrathin film was not released by
simple immersion in the solvent, but gentle solvent flow around
the substrate with a micropipette made separation of the film rapid
and easy. As shown in Fig. 1b, the size of the self-supporting film
floating in ethanol is similar to the size of the substrate, and no
cracks can be observed (see Supplementary Information, Movie
S1). The film was then washed in a bath of fresh ethanol. The
nanofilm can be transferred onto a wide variety of substrates,
such as alumina membrane, glass plate and transmission electron
microscopy (TEM) grid. Detachment behaviour differed between
nanofilms with comparable thickness (approximately 40 nm).
Organic, Hybrid 1 and Hybrid 2 ultrathin films were easy to detach
from the substrate. The size of the detached self-supporting films
was only limited by the wafer size. In contrast, Hybrid 3 could not
be released as one piece and was broken into pieces of around 1 cm2
during the detachment process. The inorganic ZrO2 film was very
difficult to detach and was broken into small pieces (several mm2
)
during detachment. Apparently, films with high ZrO2 contents are
too fragile to survive the detachment process.
The development of the organic network was confirmed using
IPN samples directly prepared on a gold-coated glass substrate
and by monitoring the decrease of the IR band at 810 cm−1
characteristic of the acrylic unit (Fig. 2a). A conversion of more
than 95% of the acrylate double bond was achieved after 2 min of
irradiation. Elemental analysis of the hybrid nanofilms was carried
out by X-ray photoelectron spectroscopy (XPS). A typical XPS
spectrum (Fig. 2b) of an IPN hybrid layer has three characteristic
(ii)
(i)
Acrylic unit
(810 cm–1)
(i) Without ultraviolet
(ii) With 2 min ultraviolet
Absorbance
Wavenumber (cm–1)
2,000 1,600 1,200 800
0.020
0.025
0.030
0.035
0.040
Zr 3d
PE(kc.p.s.)
O 1s
PE(kc.p.s.)
Binding energy (eV)
C 1s
PE(kc.p.s.)
Binding energy (eV)
O 1s
C 1s
Zr 3d
PE(kc.p.s.)
Binding energy (eV)
0
40
80
120
160
200
800 600 400 200 0
186
Binding energy (eV)
189 183
0
2
4
6
8
10
12
291 288 285
0
5
10
15
20
25
30
537 534 531
0
10
20
30
40
a
b
Figure 2 Characterization of a 40-nm-thick hybrid IPN layer synthesized by
spin-coating. a, Influence of ultraviolet irradiation on the infrared spectra of the
sample Hybrid 1 synthesized on a gold-coated glass plate. b, XPS spectrum of
Hybrid 1 prepared on a silicon wafer showing the three characteristic peaks of
carbon, oxygen and zirconium (PE = photoelectron emission.).
peaks of carbon, oxygen and zirconium. As can be see in Table 1,
the increase in the fraction of the inorganic precursor in the
reactive formulation leads to an increased zirconium content in
the nanofilm. Interestingly, the content of carbon, oxygen and
zirconium in the three different hybrids, as determined by XPS,
essentially agreed with those calculated from the molar fraction

ARTICLES
Scratched
domain
80
40
0
1,000
500
0
60.26 nm
0 nm
0 40 80 0 500 1,000
51.0
45.8
30.5
15.3
0
0 16 32 48
Distance (nm)Distance (μm)
64
Zdata(nm)
80
3.10
2.33
1.55
0.78
0
0 199.6 399.2 598.8 798.4
Zdata(nm)
998
28 nm
3.04 nm
0 nm
a b
μm
nm
nmμm
Figure 3 AFM observations of a Hybrid 2 nanofilm deposited on a silicon wafer. a, Large-scale (scanning range 80 μm×80 μm) topographic image of the film surface
and the corresponding height profile. Part of the film was scratched with a needle on the centre of the image allowing easy measurement of the film thickness.
b, Topographic AFM image (scanning range 1 μm×1 μm) of the film surface and the corresponding height profile. The surface roughness remains within 2.5 nm.
of the different precursors in the initial formulation by assuming
quantitative conversion of all of the monomers. In all cases, the
agreement is excellent. It is clear that Zr(BuOn
)4 precursor and the
organic monomers were completely incorporated into the film.
To examine morphological details, ‘self-floating’
nanomembranes were first deposited onto a silicon wafer and
examined by atomic force microscopy (AFM). Figure 3a shows a
large-scale (80 μm×80 μm) topographic AFM image of a Hybrid 2
nanofilm. Film thickness was measured at different places of the
substrate by locally scratching the specimen with a needle, and was
found to be constant with an accuracy of 10% over the whole film.
The surface of the nanofilm seems as flat as the silicon wafer surface
at a microscopic scale. For observations carried out at a smaller
scale (1 μm ×1 μm), any phase separation was not detected, and
the surface roughness remained within 2.5 nm (Fig. 3b).
IPN nanofilms were then transferred onto a porous alumina
membrane (Anodisc). The cross-sectional scanning electron
microscopy (SEM) photograph of the nanofilm on porous alumina
allows direct measurement of the membrane thickness, as given
in Fig. 4a. The thickness depends both on the concentration and
composition of the reactive formulation, and is in the range
of 10–70 nm under the current preparative conditions (see the
Supplementary Information). For similar specimens, the thickness
measured by SEM is slightly larger but essentially agrees with the
AFM observations. It should be noted that the thickness measured
by SEM includes a few nanometres of platinum used for coating the
sample and is a maximum value of the thickness.
The SEM top-view image (Fig. 4b) shows that the pores of
Anodisc (200 nm at maximum) are fully covered with the nanofilm
without any defects or cracks. TEM observation was carried
out on IPN nanofilms transferred onto a copper grid. Figure 4c
demonstrates that a smooth amorphous surface is formed for the
hybrid nanofilm with the lowest ZrO2 fraction (Hybrid 1). The
organic and inorganic components seem to penetrate each other to
form molecular IPN. In contrast, when the ZrO2 concentration is
increased, a regular lattice with domain sizes of 5–10 nm is formed
(Fig. 4d). The majority of the lattice fringe reveals a periodicity of
0.29 nm. However, XRD spectra did not show any characteristic
peak of the tetragonal or cubic phase of ZrO2, and we could not
determine the exact structure of the regular lattice. This observation
was unexpected because the sol–gel synthesis of bulk ZrO2 below
350◦
C is known to give an amorphous phase28
. However, formation
of tetragonal ZrO2 at low temperatures has been reported for
thin films produced through monolayer-mediated deposition from
aqueous dispersions29
. As the surface roughness remained within
2.5 nm, 3D crystal growth cannot exist in the IPN ultrathin film.
Model microstructures of the free-standing hybrid
nanomembrane are presented in Fig. 5. In the case of an ideal
IPN structure (Fig. 5a), the two networks are fully developed and
interpenetrated. However, this picture may not be the case for our
nanomembranes because the content of the inorganic network
is not sufficient to allow complete interconnection of ZrO2 gel
over the whole film. In a more realistic picture of Fig. 5b, the
inorganic network forms domains that are evenly dispersed in the
soft organic matrix.
The self-supporting nanofilms floating in ethanol (Fig. 1b) are
very flexible, robust, and can easily be folded into small shapes.
Figure 6a and b show optical micrographs of the aspiration process
of a 16 cm2
nanofilm into a micropipette with a tip diameter of
approximately 320 μm. The size of the micropipette hole is 30,000
times smaller than that of the film. The film can pass completely
through the micropipette due to its extreme thinness and flexibility,

ARTICLES
35-nm-thick
nanofilm
Anodisc
300 nm
10 nm
3 nm
10 nm
0.297 nm
ZrO2
nanocrystal
a
500 nm
Anodisc
b
c d
Figure 4 Microscopic characterization of free-standing hybrid nanofilms. a, SEM side-view image of a Hybrid 1 nanomembrane on an Anodisc. b, SEM top-view image
of a Hybrid 1 nanomembrane. c, TEM image of a free-standing Hybrid 1 nanomembrane on a copper grid. d, TEM image of a Hybrid 3 nanomembrane showing the presence
of a nanocrystalline domain dispersed in an amorphous matrix.
and can be released into the solvent. Soon after release, the hybrid
ultrathin film adopts a folded appearance, but rapidly regains its
original film shape by gentle manipulation with a spatula. The
aspiration/release/shape recovery cycle can be repeated many times
without damaging or cracking the material (see Supplementary
Information, Movie S2). The ability of the nanofilm to ‘flow’ into
such tiny orifices is an indication of its extreme flexibility. Such
an experiment could not be achieved for the pure inorganic film
because of the rapid damage to the film.
It is possible to take the hybrid nanofilms, Hybrid 1 and
Hybrid 2, out of the solvent into the air using a suitable frame,
such as a wire loop (Fig. 6c). The films did not rupture upon
drying, and remained stable for several months. To the best of our
knowledge, free-standing membranes with comparable thickness
and dimensions have not been reported previously. The freely
suspended films are basically highly transparent but can reflect the
light of a lamp. In contrast, although the pure organic nanofilm
can be detached and remain macroscopically self-supporting in
ethanol, it cannot be obtained as free-standing film over the wire
loop due to breakage during the drying process. Therefore, the ZrO2
gel in the rubbery polymer matrix must act as a reinforcing agent. It
is known that novel properties can be obtained by the combination
of organic and inorganic elements when they are mixed at the
molecular level30
. In particular, the mechanical strength of organic
polymers and rubbers can be improved by the incorporation of
inorganic segments31,32
. The toughness of such hybrid materials
arises from chain extension in the organic amorphous region
between relatively rigid inorganic blocks33,34
.
Permselectivity is an important feature of ultrathin films35,36
. To
confirm such potential, a free-standing nanomembrane, Hybrid 1,
was placed on the top of a glass tube (with an inner diameter
of 6 mm) and dried in air. The dried nanofilm sticks to the
glass tube, ensuring the robust attachment of the freely suspended
nanomembrane. The tube was positioned vertically with the
nanofilm at the bottom. Ethanol was slowly introduced from the
upper side and filled the tube (see Fig. 6e). It is remarkable that
a film as thin as 35 nm can hold a column of ethanol as high as
4 cm without rupturing. In fact, the weight of the liquid is 70,000
times greater than the weight of the film that holds it. Ethanol
permeated slowly through the film with a speed of approximately
420 μl min−1
cm−2
. One drop of ethanol came out of the glass tube
every 10 s (see Supplementary Information, Movie S3). Pore-size
control of the ultrathin polymer network, which is essential for use
as separation membranes, may be achieved by adjusting the degree
of crosslinking of the organic network, and also by modifying the
initial organic/inorganic precursor ratio.
The outstanding macroscopic flexibility and robustness of
the hybrid IPN nanofilm will come mainly from the intimate
mixing of the soft organic phase and rigid inorganic phase at
the molecular scale, as confirmed by TEM, and the synergistic
reinforcing effect22,30,33
of the two different networks. The chemical
bonding of the organic matrix with the inorganic network
through the hydroxyl group on the organic polymer backbone
will suppress formation of phase-separated domains. Moreover,
the simultaneous formation of the organic and inorganic networks
ensures a homogeneous dispersion of the zirconia domains in the
organic network. The physical and chemical crosslinking of the
two phases will also give a greater macroscopic stability to the
ultrathin films. For instance, ultrathin Langmuir–Blodgett36
and
multilayered films6,37,38
that are stronger and more robust after

ARTICLES
a
b
Figure 5 Model nanostructures of free-standing hybrid nanomembranes of
interpenetrating networks. a, Ideal IPN structure where the two networks are fully
developed and interpenetrated. b, More realistic structure of a hybrid IPN, in which
the hard inorganic network domains are not necessarily connected but rather
dispersed in the soft organic polymer matrix. The organic network is represented in
black. The ZrO2 network is represented in blue.
chemical crosslinking have been reported. Another possibility that
can explain the observed macroscopic robustness is the unique
conformation of the crosslinked organic polymers prepared under
shear stress of centrifugal forces, where the organic network tends
to adopt a much more spread conformation, and a high level of
in-plane, biaxial orientation of polymer chains2,39,40
.
Preliminary work was undertaken to determine the modulus
of the film by macroscopic dynamic viscoelastic measurements.
Early results indicate that the specimen Hybrid 1 has a Young’s
modulus of around 150 MPa, demonstrating that this IPN hybrid
nanofilm is in the rubbery state at room temperature. This
modulus is much lower than the value recently reported for
glassy polymer multilayer films containing gold nanoparticles2,15
.
Therefore, the outstanding macroscopic stability of the IPN hybrid
films will not come from a high modulus. The free-standing
hybrid IPN nanofilms are capable of sustaining significant, repeated
mechanical deformations as probed with a bulging test described
in detail in the literature15,41
. This test was carried out with free-
standing films covering a 1-mm-diameter hole by applying pressure
from one side of the film and detecting its deflection with an
optical microscope. Figure 6d shows the side-view optical images of
a deflected free-standing nanofilm with different pressures. From
this experiment, we found that the Hybrid 1 specimen has an
ultimate tensile strength σ = 105 MPa and an ultimate tensile
elongation ε = 2.6% (average values on 15 samples). These results
qualitatively agree with those reported for hybrid polymer films of
comparable thickness2,15
, and clearly demonstrate the robustness
of the hybrid IPN nanofilms. We speculate that the combination
of elasticity and toughness may be essential features for obtaining
robust and large-scale free-standing nanomembranes. For instance,
spider silk is an inspiring example of free-standing biomaterial with
unsurpassed properties. Although its Young’s modulus is rather
small, it is known that the extraordinary toughness of this material
comes from the subtle interplay between elastic protein strands and
crystalline nanoblocks42
.
In summary, the hybrid nanofilm of an organic acrylate
network and an inorganic zirconia network was found to combine
often-incompatible properties of extreme thinness, robustness and
flexibility. The pure organic nanofilm was not robust enough to
be obtained as large free-standing membrane and a pure zirconia
nanofilm was broken into small pieces soon after detachment from
the substrate because of its brittleness. The combination of soft and
hard networks in a single material is an efficient method to reinforce
the mechanical properties of polymers. Recently, Gong et al.
reported the preparation of double-network gels with extremely
high mechanical strength43
. The double network gels comprise
a rigid polyelectrolyte network and a flexible neutral polymer
network, and their optimal combination was essential to obtain
the desired property. The highly crosslinked network component
has a high Young’s modulus but is quite brittle on its own, and the
dramatically enhanced mechanical strength observed44
is attributed
to the effective relaxation of stress by the loosely crosslinked
networks, which dissipates the fracture energy and prevents crack
development. Another reinforcement technique is to disperse a
small amount of a hard inorganic network in a soft polymer
matrix. For instance, the ultimate strength of poly(dimetylsiloxane)
increased 10 times following dispersion of 10 wt% silica, whereas
the elongation at rupture remained quasi-constant31
.
Such synergy effects must be operating in the unprecedented
combination of nanometre thickness and macroscopic robustness
in the current study. Scientifically, it is important to study
how (almost) two-dimensionally extended networks are made of
individual polymeric structures as well as the molecular mechanism
for the origin of macroscopic mechanical strength within film
thickness of 10–30 nm. On the practical side, the significance
of robust nanofilms with macroscopic dimensions is extensive.
These features provide fundamental advantages in the design of
separation membranes in general. The molecular function of
proteins and organic compounds can be readily incorporated to the
nanofilm, because film thickness and molecular sizes of the latter
component are close enough. Controlled, efficient ion transport
across self-supporting nanofilms should have enormous industrial
impact in many practical applications, such as fuel cells.
METHODS
MATERIALS
PVA (98 mol% hydrolysed, Mw ≈ 78,000 g mol−1
) was purchased from
Polysciences. Poly(4-vinylphenol) (Mw ≈ 8,000 g mol−1
) was obtained from
Aldrich. 4-hydroxybutyl acrylate and 1,6-hexanediol diacrylate were purchased
from Acros Organics and Alfa Aesar, respectively. The photo-initiator Darocur
4165 was a donation from Kyoritsu Chemicals. Zirconium tetra-n-butoxide was
obtained from Kanto Chemicals. All chemicals were used as received without
further purification. A porous alumina membrane (Anodisc, pore size
0.2 μm, diameter 25 mm, thickness 60 μm) was purchased from
Whatman International.
PREPARATION OF FREE-STANDING HYBRID NANOMEMBRANES
The fabrication of a 35-nm-thick free-standing nanofilm, Hybrid 1, was carried
out as follows. First, a solution of the organic precursors was prepared by

ARTICLES
a b
c d
e
Flow
Flow
1 cm 500 μm
350 μm
0.5 kPa
3.0 kPa
7.0 kPa
(i)
(ii)
(iii)
35-nmmembrane
Ethanol(3.2cm)
Figure 6 Manipulation and properties of a 35-nm-thick Hybrid 1 self-supporting nanomembrane. a, Optical micrograph showing the aspiration process of a 16 cm2
nanofilm into a micropipette with a tip diameter of 320 μm. b, Close-up micrograph around the tip during the aspiration process of the nanofilm. It is remarkable that, owing
to its flexibility and extreme thinness, the nanofilm can reversibly pass through a hole 30,000 times smaller than its own area. c, Micrographs of a large free-standing
nanofilm in the air supported by a wire loop. The nanofilm is transparent and can reflect the light. d, Side-view optical images of a free-standing nanofilm that deformed by
different pressures applied from below (bulge test). e, Nanofilm supporting a column of ethanol 70,000 times heavier than its own weight. The ethanol slowly permeates
through the film, as denoted by the formation of a solvent droplet at the bottom of the tube (see Supplementary Information, Movie S3).
mixing 700 μl (5,060 μmol) of 4-hydroxybutyl acrylate, 29 μl (130 μmol) of
1,6-hexanediol diacrylate and 17.7 μl (130 μmol) of photo-initiator Darocur
4165 in a brown glass beaker. Then 75 μl of this organic solution was mixed
with 75 μl of zirconium tetra-n-butoxide in 5 ml of chloroform. Spin-coating
was conducted with a MIKASA spincoater 1H-D7. An ethanol solution of
poly(vinylphenol) (20 mg ml−1
) was first spin-coated on a clean silicon wafer at
a speed of 3,000 r.p.m. for two minutes. Then a PVA solution in water
(5 mg ml−1
) was spin-coated at 3,000 r.p.m. for 2 min. The chloroform
solution containing IPN precursors was finally spin-coated at 4,000 r.p.m. for
2 min under a nitrogen atmosphere. The ultraviolet light was switched on after
10 s of spin-coating and left on until the end of the preparation.
INSTRUMENTS AND METHODS
Macroscopic images of self-supporting ultrathin films were photographed by a
digital camera RICOH RDC-7, with 640×480 pixels. Fourier transform
infrared spectroscopy measurements were carried out using a Thermonicolet
Nexus 870 FT-IR spectrometer. Irradiation of the sample during the
spin-coating was carried out with a Lightningcure LC5 (Hamamatsu). The
irradiation system was composed of a static ultraviolet lamp together with a
light filter, allowing irradiation of the sample with a wavelength of 365 nm and
an intensity of 23 mW cm−2
during the spin-coating process. AFM
measurements were carried out by non-contact mode on an explorer scanning
probe microscope TMX2100 (TopoMetrix). SEM observations were carried out
on a Hitachi S-5200 field-emission microscope. Specimens for the SEM
experiments were coated with a 2-nm-thick platinum layer using an
ion-sputtering coater (Hitachi; E-1030, 15 mA, 30 s). TEM observations were
carried out using a JEOL JEM 2100 F/SP transmission electron microscope at
200 kV. XPS measurements were carried out on an ESCALAB 250 (VG) using
Al Kα (1486.6 eV) radiation. Preliminary dynamic viscoelastic macroscopic
measurements were carried out in collaboration with Professors Tanaka and
Nagamura’s group at the Department of Applied Chemistry, Kyushu University,
and will be reported in a separate publication. The detailed procedure for the
bulge test is given in the Supplementary Information.
Received 13 January 2006; accepted 12 April 2006; published 21 May 2006.
References
1. Stroock, A., Kane, R. S., Weck, M., Metallo, S. J. & Whitesides, G. M. Synthesis of free-standing
quasi-two-dimensional polymers. Langmuir 19, 2466–2472 (2003).
2. Jiang, C., Markutsya, S., Pikus, Y. & Tsukruk, V. Freely suspended nanocomposite membranes as
highly sensitive sensors. Nature Mater. 3, 721–727 (2004).
3. O’Connel, P. A. & McKenna, G. B. Rheological measurements of thermoviscoelastic response of
ultrathin polymer films. Science 307, 1760–1763 (2005).
4. Mattsson, J., Forrest, J. A. & Borjesson, L. Quantifying glass transition behaviour in ultrathin
free-standing polymer films. Phys. Rev. E 62, 5187–5200 (2000).
5. Tang, Z., Kotov, N. A., Magonov, S. & Ozturk, B. Nanostructured artificial nacre. Nature Mater. 2,
413–418 (2003).
6. Mallwitz, F. & Laschewsky, A. Direct access to stable, freestanding polymer membranes by
layer-by-layer assembly of polyelectrolytes. Adv. Mater. 17, 1296–1299 (2005).
7. Mamedov, A. et al. Molecular design of strong single-wall carbon nanotube/polyelectrolyte
multiplayer composites. Nature Mater. 1, 190–194 (2002).
8. Huck, W. T., Stroock, A. D. & Whitesides, G. M. Synthesis of geometrically well-defined, molecularly
thin polymer films. Angew. Chem. Int. Edn 39, 1058–1061 (2000).
9. Mamedov, A. & Kotov, N. Free-standing layer-by-layer assembled films of magnetite nanoparticles.
Langmuir 16, 5530–5533 (2000).
10. Mallwitz, F. & Goedel, W. A. Physically cross-linked ultrathin elastomeric membranes. Angew. Chem.
Int. Edn 40, 2645–2647 (2001).
11. Eck, W., Küller, A., Grunze, M., Völkel, B. & Gölzhäuser, A. Freestanding nanosheets from
crosslinked biphenyl self-assembled monolayers. Adv. Mater. 17, 2583–2587 (2005).
12. Xu, H. & Goedel, W. A. Polymer-silica hybrid monolayers as precursors for ultrathin free-standing
porous membranes. Langmuir 18, 2363–2367 (2002).
13. Nardin, C., Winterhalter, M. & Meier, W. Giant free-standing ABA triblock copolymer membranes.
Langmuir 16, 7708–7712 (2000).

ARTICLES
14. Jiang, C., Markutsya, C. & Tsukruk, V. Compliant, robust, and truly nanoscale free-standing
multilayer films fabricated using spin-assisted layer-by-layer assembly. Adv. Mater. 16,
157–161 (2004).
15. Markutsya, S., Jiang, C., Pikus, Y. & Tsukruk, V. Freely suspended layer-by-layer nanomembranes:
testing micromechanical properties. Adv. Funct. Mater. 15, 771–780 (2005).
16. Jiang, C., Markutsya, S., Shulda, H. & Tsukruk, V. Freely suspended gold nanoparticle arrays. Adv.
Mater. 17, 1669–1673 (2005).
17. Ko, H., Jiang, C., Shulda, H. & Tsukruk, V. Carbon nanotube arrays encapsulated into freely
suspended flexible films. Chem. Mater. 17, 2490–2493 (2005).
18. Decher, G. Fuzzy nanoassemblies toward layered polymeric multicomposites. Science 277,
1232–1237 (1997).
19. Johal, M. S. et al. Polyelectrolyte trilayer combinations using spin-assembly and ionic self-assembly.
Langmuir 19, 8876–8881 (2003).
20. Hashizume, M. & Kunitake, T. Preparation of self-supporting ultrathin films of titania by
spin-coating. Langmuir 19, 10172–10178 (2003).
21. Hashizume, M. & Kunitake, T. Preparation and functionalization of self-supporting (polymer/metal
oxide) composite ultrathin films. RIKEN Rev. 38, 36–39 (2001).
22. Sharp, K. G. Inorganic/organic hybrid materials. Adv. Mater. 10, 1243–1248 (1998).
23. Matejka, L. & Dukh, O. Organic-inorganic hybrids networks. Macromol. Symp. 171, 181–188 (2001).
24. Saegusa, T. Organic-inorganic polymers hybrids. Pure Appl. Chem. 67, 1965–1970 (1995).
25. Imai, Y., Naka, K. & Chujo, Y. Reversible formation of interpenatrating polymer network structure in
organic-inorganic polymer hybrids. Polym. J. 30, 990–995 (1998).
26. Tamaki, R., Naka, K. & Chujo, Y. Synthesis of polystyrene/silica gel polymer hybrids by in situ
polymerisation method. Polym. Bull. 39, 303–310 (1997).
27. Sperling, L. H. & Mishra, V. The current status of interpenetrating polymer networks. Polym. Adv.
Technol. 7, 197–208 (1996).
28. Chang, S. & Doong, R. ZrO2 thin films with controllable morphology and thickness by spin-coated
sol-gel method. Thin Solid Films 489, 17–22 (2005).
29. Polli, A. D. et al. Structural characterisation of ZrO2 thin films produced via self-assembled
monolayer-mediated deposition from aqueous dispersions. Thin Solid Films 379, 122–127 (2000).
30. Sanchez, C. et al. Designed hybrid organic-inorganic nanocomposites from functional nanobuilding
blocks. Chem. Mater. 13, 3061–3083 (2001).
31. Rajan, G. S., Sur, G. S., Mark, J. E., Sharfer, D. W. & Beaucage, G. Preparation and characterisation of
some unusually transparent poly(dimethylsiloxane) nanocomposites. J. Polym. Sci. B 41,
1897–1901 (2003).
32. Sur, G. S. & Mark, J. E. Reinforcing effects from silica-type fillers containing hydrocarbon groups.
Polym. Bull. 14, 325–329 (1985).
33. Mammeri, F., Le Bourhis, E., Rozes, L. & Sanchez, C. Mechanical properties of hybrid
organic-inorganic materials. J. Mater. Chem. 15, 3787–3811 (2005).
34. Sur, G. S. & Mark, J. E. Elastomeric networks cross-linked by silica or titania fillers. Eur. Polym. J. 21,
1051–1052 (1985).
35. Merkel, T. C. et al. Ultrapermeable, reverse-selective nanocomposite membranes. Science 296,
519–522 (2002).
36. Kato, D. et al. Permselective monolayer membrane based on two-dimensional cross-linked
polysiloxane LB films for hydrogen peroxide detecting glucose sensors. Chem. Commun.
2616–2617 (2002).
37. Ichinose, I., Mizuki, S., Ohno, S., Shiraishi, H. & Kunitake, T. Preparation of cross-linked films based
on layer-by-layer assembly of polymers. Polym. J. 31, 1065–1070 (1999).
38. Lutkenhaus, J. L., Hrabak, K. D., McEnnis, K. & Hammond, P. T. Elastomeric flexible free-standing
hydrogen-bonded nanoscale assemblies. J. Am. Chem. Soc. J. 127, 17228–17234 (2005).
39. Cho, J., Char, K., Hong, J. D. & Lee, K. B. Fabrication of highly ordered multilayer films using a spin
self-assembly method. Adv. Mater. 13, 1076–1078 (2001).
40. Chiarelli, P. A. et al. Controlled fabrication of polyelectrolyte multilayer thin films using
spin-assembly. Adv. Mater. 15, 1167–1171 (2001).
41. Goedel, W. A. & Heger, R. Elastomeric suspended membranes generated via Langmuir-Blodgett
transfer. Langmuir 14, 3470–3474 (1998).
42. Becker, N. et al. Molecular nanosprings in spider capture-silk threads. Nature Mater. 2,
278–283 (2003).
43. Gong, J. P., Katsuyama, Y., Kurokawa, T. & Osada, Y. Double networks hydrogels with extremely high
mechanical strength. Adv. Mater. 15, 1155–1158 (2003).
44. Na, Y. H. et al. Structural characteristics of double networks gels with extremely high mechanical
strength. Macromolecules 37, 5370–5374 (2004).
Acknowledgements
This work was supported by the postdoctoral program for foreign researchers of the Japan Society for
the Promotion of Science (JSPS) through a fellowship awarded to R.V.
Correspondence and requests for materials should be addressed to T.K.
Supplementary Information accompanies this paper on www.nature.com/naturematerials.
Competing financial interests
The authors declare that they have no competing financial interests.
Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Vendamme et.al. Supplementary Information 1/4
Supplementary information
Movie S1: Manipulation of a 16 cm2
self-supporting 35 nm thick Hybrid_1 nanofilm floating
in ethanol.
Movie S2: Aspiration and release of a 16 cm2
Hybrid_1 nanofilm through a micropipette with
a tip diameter of 320 µm.
Movie S3: Permeation of ethanol through a freely suspended Hybrid_1 nanofilm (35 nm
thick) attached at the bottom of a glass tube.
0 100 200 300 400 500
0
10
20
30
40
50
60
70
Filmthickness(nm)
Precursor concentration (mM)
Figure S1. Thickness of the free-standing Hybrid_1 nanofilms as a function of the precursor
concentration in chloroform.
© 2006 Nature Publishing Group

a) b)
c) d)
Figure S2. Additional SEM pictures of Hybrid_2 and Hybrid_3 nanofilms. a, SEM side-
view image of a Hybrid_2 nanofilm (42 nm thick) on an ANODISC. b, SEM top-view image
a Hybrid_2 nanofilm. c, SEM side-view image of a Hybrid_3 nanofilm (29 nm thick) on
porous alumina. d, SEM top-view image the Hybrid_3 nanofilm.

The bulging test was conducted in accordance with the known routine (see references 2, 15
and 41). The tensile stress σ and tensile strain ε of ultrathin films can be measured by
applying an overpressure to one side of a freely suspended film that covers a metal plate with
a circular hole and measuring the resulting deflection of the film. In the present study, the
applied pressure was controlled with a digital manometer and the membrane deflection was
monitored with an optical microscope. A scheme of the set-up is presented in figure S3.
Figure S3. Scheme of the bulging experiment. The pressure, P, vertical displacement of the
film centre, d, radius of the opening, a, and film thickness, h, are indicated in the figure.
From such an experiment, it is possible to determine the ultimate tensile stress σ and the
ultimate tensile strain ε (respectively defined as the strength and elongation at membrane
rupture) using the following formulas (explained in Ref. 15):
σ = (P × a2
) / (4 × h × d)
ε = (2 × d2
) / (3 × a2
)

Where P is the pressure required to break the membrane, a is the radius of the opening (0,5
mm in our set-up), h is the membrane thickness, and d is the deflection of the membrane
centre at rupture. Using this method, the average ultimate mechanical strength σ of the
Hybrid_1 specimen was determined as 105 Mpa (average value from 15 samples). The lowest
value is 93 MPa and the highest 136 MPa. The average ultimate elongation ε is 2,6 %. These
results qualitatively agree with those reported in the literature for organic/inorganic hybrid
nanofilm (Ref 2 and 15) and clearly demonstrate the robustness of the hybrid IPN
nanomembranes.
A detailed analysis of the ultimate mechanical properties of hybrid nanofilms as a function of
the composition (organic/inorganic molar ratio) and the organic network crosslinking density
is currently under investigation and will be reported elsewhere.

NEWS & VIEWS NATURE|Vol 441|25 May 2006
418
species9
. This leaves the extent of overlap
between statistical and ecological significance
as an interesting and open question.
Wecangofurther:onwhatbasisdidDarwin
make his assertion about the discreteness of
species? This question is distinct from debates
aboutthedefinitionofspeciesinnature.Black-
berriesreproduceasexually,anditisimpossible
to agree on how many ‘species’ there are; but,
nonetheless, we all know a ‘blackberry’ when
we see one and do not wonder if it is actually a
raspberry.Greattits,bluetitsandcoaltitsareall
quite distinct when considered as a set, but are
surelyjustmore-or-lesscontinuousvariantson
a tit theme when compared with flamingos.
Bacteria that are vastly different genetically are
all called Legionella because they clump along
the single niche axis that matters to us: they all
cause Legionnaire’s disease.
So what is the correct or meaningful frame
of reference when thinking about the ecologi-
cal nature of species? As well as providing
stimulating theoretical results, Scheffer and
van Nes1
have revitalized the fundamental
question of how we should look at the ecologi-
cal identity of species. ■
Sean Nee and Nick Colegrave are at the Institute
of Evolutionary Biology, School of Biological
Sciences, University of Edinburgh,
West Mains Road, Edinburgh EH9 3JT, UK.
e-mails: sean.nee@ed.ac.uk;
n.colegrave@ed.ac.uk
1. Scheffer,M.&vanNes,E.H.Proc.NatlAcad.Sci.USA103,
6230–6235(2006).
2. Roughgarden,J.TheoryofPopulationGeneticsand
EvolutionaryEcology(Macmillan,NewYork,1979).
3. May,R.M.StabilityandComplexityinModelEcosystems
(PrincetonUniv.Press,1974).
4. Sasaki,A.J.Theor.Biol.186,415–430(1997).
5. Hastings,A.TrendsEcol.Evol.19,39–45(2004).
6. Hubbell,S.P.TheUnifiedNeutralTheoryofBiodiversity
andBiogeography(PrincetonUniv.Press,2001).
7. Colinvaux,P.Ecology2(Wiley,NewYork,1993).
8. Nee,S.Funct.Ecol.19,173–176(2005).
9. Holling,C.S.Ecol.Monogr.62,447–502(1992).
Free-standing nanofilms are a
wonder of membrane technology.
Although it’s no easy matter to
produce them, once made these
quasi-two-dimensional objects
display fascinating behaviour,
combining macroscopic surface
area with nanoscopic depth.
A remarkable example is reported
by Toyoki Kunitake and colleagues in
Nature Materials (R. Vendamme et al.
doi:10.1038/nmat1655; 2006). They
have prepared an ultrathin film that
is barely visible to the naked eye, but
is so flexible it can be drawn through
a micropipette hole 30,000 times
smaller than its width (pictured).
Despite its flimsy appearance, the
film can support a liquid body
70,000 times heavier than its own
weight, and withstand significant
deformation. It is also stable to
various environmental and
mechanical stresses. Even more
impressively, the film breaks
records for size in being several
centimetres across, yet only around
35 nanometres thick.
This apparently incompatible
combination of strength and
thinness is a result of the film’s
hybrid composition. It consists of an
organic polymeric network, which
makes it pliable and deformable,
interpenetrated by zirconia
(zirconium dioxide), which confers
strength and stability. To prepare
the nanofilm, the two materials are
generated simultaneously from their
precursors on a spin-coating plate.
The chemical processes involved are
quite different: the polymer forms
by light-induced crosslinking of its
monomers, whereas the zirconium
precursor reacts with residual traces
ofwaterinthefilm’spolyvinylalcohol
substrate. Nevertheless, the
components intertwine to give
nanofilms with properties that make
them useful as sensors, actuators
and separation membranes.
Maria Bellantone
MATERIALS SCIENCE
Filmreview
STEM CELLS
Good,badandreformable
Viktor Janzen and David T. Scadden
The ability of stem cells to continuously supply vast numbers of cells is
magnificent, but it can be devastating if it runs amok, as in some tumours.
So what makes a normal stem cell turn bad, and can it be redeemed?
The stem cell is a bit like the griffin of mythol-
ogy—halflion,halfeagle;grandandpowerful,
but potentially monstrous in effect. These
essentially unspecialized cells can renew their
own population while supplying cells that
mature(differentiate)intothespecializedcells
necessary for all tissues. Although this ability
to reproduce and self-renew is sublime when
functioning properly, its disorder creates
masses of dysfunctional replicating cells.
Indeed, stem-cell-like cells have been found in
a range of human tumours. Not all cancer is
due to a stem cell gone bad, but some cancer-
initiating cells are probably stem cells, and
the rest acquire the stem-cell feature of self-
renewal. This raises the troubling spectre that
normal stem cells and cancer stem cells might
share the molecular features essential to their
nature. So attempting to treat cancer by
disrupting the functions of the cancer stem
cells might also disturb normal stem cells —
potentially fatally.
In this issue, however, Yilmaz et al. (page
475)1
and Zhang et al. (page 518)2
report that
there may be key molecular distinctions
between the normal and malignant stem cell
that might be of use in designing therapies
that target malignant stem cells, while sparing
normal stem cells.
The investigations centred on a protein
called PTEN (for ‘phosphatase and tensin
homologue’), a known tumour suppressor and
anintracellularmodulatorofseveralmajorcell-
signalling pathways. Notably, PTEN inhibits
signalling through the AKT pathway that
responds to growth factors (Fig. 1a). Growth
factors bind to specific receptors on the cell
surface and induce a cascade of cellular modi-
fications in which phosphate groups are added
to a series of proteins. Essentially, the activa-
tion signal is passed along the pathway like a
baton in a relay race until it reaches the final
‘effector’ proteins that carry out the pathway
response: for example, changing the expres-
sion of particular genes or halting the cell
cycle. When the growth factor binds to its
receptor, the enzyme PI3K is activated, and it
is this step that PTEN inhibits. Activation of
PI3K leads to phosphorylation and activation
of the AKT protein, which in turn can poten-
tially phosphorylate more than 9,000 proteins.
Two key downstream AKT effectors, called
mTOR and FOXO, are implicated in cancer
development.
Yilmaz et al.1
and Zhang et al.2
used PTEN-
deficient mice to examine how a lack of this
protein affects cell proliferation, programmed
cell death and cell localization in haemato-
poietic stem cells (which produce blood and
immune cells) (Fig. 1b). Previous work had
shown that PTEN deficiency increases the
proliferation of stem or progenitor cells (a
slightly more differentiated cell type) in the
fetal mouse brain. It also increases self-renewal

nmat 2006

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Viewers also liked

Viewers also liked (20)

Similar to nmat 2006

Similar to nmat 2006 (20)

nmat 2006