© 2006 Nature Publishing Group

Graphene-based composite materials
Sasha Stankovich1*, Dmitriy A. Dikin1*, Geoffrey H. B. Dommett1, Kevin M. Kohlhaas1, Eric J. Zimney1,
Eric A. Stach3, Richard D. Piner1, SonBinh T. Nguyen2 & Rodney S. Ruoff1
Graphene sheets—one-atom-thick two-dimensional layers of
sp2-bonded carbon—are predicted to have a range of unusual
properties. Their thermal conductivity and mechanical stiffness
may rival the remarkable in-plane values for graphite
(,3,000 Wm21 K21 and 1,060 GPa, respectively); their fracture
strength should be comparable to that of carbon nanotubes for
similar types of defects1–3; and recent studies have shown that
individual graphene sheets have extraordinary electronic trans-
port properties4–8. One possible route to harnessing these proper-
ties for applications would be to incorporate graphene sheets in a
composite material. The manufacturing of such composites
requires not only that graphene sheets be produced on a sufficient
scale but that they also be incorporated, and homogeneously
distributed, into various matrices. Graphite, inexpensive and
available in large quantity, unfortunately does not readily exfoliate
to yield individual graphene sheets. Here we present a general
approach for the preparation of graphene-polymer composites via
complete exfoliation of graphite9 and molecular-level dispersion
of individual, chemically modified graphene sheets within poly-
mer hosts. A polystyrene–graphene composite formed by this
route exhibits a percolation threshold10 of ,0.1 volume per cent
for room-temperature electrical conductivity, the lowest reported
value for any carbon-based composite except for those involving
carbon nanotubes11; at only 1 volume per cent, this composite has a
conductivity of ,0.1 S m21, sufficient for many electrical applica-
tions12. Our bottom-up chemical approach of tuning the graphene
sheet properties provides a path to a broad new class of graphene-
based materials and their use in a variety of applications.
Graphite oxide is a layered material produced by the oxidation of
graphite (Fig. 1a). In contrast to pristine graphite, the graphene-
derived sheets in graphite oxide (graphene oxide sheets) are heavily
oxygenated, bearing hydroxyl and epoxide functional groups on their
basal planes, in addition to carbonyl and carboxyl groups located at
the sheet edges13,14. The presence of these functional groups makes
graphene oxide sheets strongly hydrophilic, which allows graphite
oxide to readily swell and disperse in water2,15,16. Previous studies by
our team have shown that a mild ultrasonic treatment of graphite
oxide in water results in its exfoliation to form stable aqueous
dispersions that consist almost entirely of 1-nm-thick sheets
(Fig. 1b)9. Given the uniformity of the observed thicknesses and that
platelets one-half of the observed minimum thickness are never
detected (or any other inverse integer value, such as one-third, and
so on), we believe that these represent fully exfoliated graphene oxide
sheets. In fact, at present, exfoliation of graphite oxide is the only way
to produce stable suspensions of quasi-two-dimensional carbon
sheets, making this a strategic starting point for large-scale synthesis
of graphene sheets. As such, graphite oxide has recently attracted
attention as a filler for polymer nanocomposites17–19.
Nanocomposites have been heralded20 as a ‘radical alternative to
conventional filled polymers or polymer blends,’ especially in the
fields of transportation and electronics. Unfortunately, owing to their
hydrophilic nature, graphene oxide sheets can only be dispersed in
aqueous media that are incompatible withmost organic polymers. In
addition, graphite oxide is electrically insulating, unlike graphite,
which limits its usefulness for the synthesis of conductive nanocom-
posites. It has been demonstrated9,18,21,22, however, that the electrical
conductivity of graphite oxide can be significantly increased by
chemical reduction, presumably owing to the restoration of a
graphitic network of sp2 bonds. But reduction of exfoliated graphene
oxide nanoplatelets in water results in their irreversible coagulation9,
which then makes dispersion within a polymer matrix at the
individual sheet level impossible.
We recently demonstrated that the exfoliation behaviour of
graphite oxide can be altered by changing the surface properties of
graphene oxide sheets by way of chemical functionalization. A
number of chemically modified graphite oxides were prepared by
treating graphite oxide with organic isocyanates23. The isocyanate
treatment reduces the hydrophilic character of graphene oxide sheets
by forming amide and carbamate ester bonds to the carboxyl and
hydroxyl groups of graphite oxide, respectively. As a result, such
isocyanate-derivatized graphite oxides no longer exfoliate in water
but readily form stable dispersions in polar aprotic solvents (such as
N,N-dimethylformamide (DMF)), consisting of completely exfoliated,
functionalized individual graphene oxide sheets with thickness
,1 nm, as determined by atomic force microscopy, AFM (Fig. 1c).
These dispersions of isocyanate-derivatized graphite oxide allow
graphene oxide sheets to be intimately mixed with many organic
polymers, facilitating synthesis of graphene–polymer composites.
Coupled with the possibility that graphite oxide can be chemically
reduced (see above), we were hopeful that nanocomposites of
polymer and isocyanate-derivatized graphite oxide could be rendered
electrically conductive.
Hence, we were gratified to find that electrically conductive
graphene–polymer nanocomposites can be prepared by solution-
phase mixing of the exfoliated phenyl isocyanate-treated graphite
oxide sheets with polystyrene, followed by their chemical reduction
(Fig. 1d). These composites feature individual graphene sheets well
dispersed throughout the polymer matrix (Fig. 1g). Similar disper-
sions have been achieved with other styrenic polymers such as
acrylonitrile-butadiene-styrene and styrene-butadiene rubbers.
Chemical reduction was essential for inducing electrical conduc-
tivity, as composite samples with un-reduced phenyl isocyanate-
treated graphite oxide sheets were insulating. In addition, the
presence of the polymer in solution during the reduction step was
key to preventing the agglomeration of the sheets. As the reduction
proceeds, the sheets become coated with the polymer and remain
individually dispersed (Fig. 1d).When the reduction step preceded the
introduction of the polymer, significant agglomeration of graphene
LETTERS
1Department of Mechanical Engineering, 2Department of Chemistry, Northwestern University, 2145 Sheridan Road, Evanston, Illinois 60208-3111, USA. 3School of Materials
Engineering and Birck Nanotechnology Center, Purdue University, 501 Northwestern Avenue, West Lafayette, Indiana 47907, USA.
*These authors contributed equally to this work.
Vol 442|20 July 2006|doi:10.1038/nature04969
282

© 2006 Nature Publishing Group

sheets occurred. Further details of the preparation and fabrication of
composite samples are described in Methods.
In a concurrent research direction, we explored a rapid
thermal expansion of graphite oxide (,1,000 8C under inert gas;
R. K. Prud’homme et al., manuscript in preparation). However, the
thermally expanded graphite oxide platelets so produced are not fully
exfoliated, consisting of several-nanometre-thick multilayer stacks,
and their dispersion in a polymer matrix is a challenging step of
manufacture. For the purpose of comparison, polystyrene compo-
sites prepared with such thermally expanded graphite oxide, by the
same procedure as was used for phenyl isocyanate-treated graphite
oxide, are less homogeneous and have higher electrical percolation
thresholds (see Supplementary Information). For this and other
reasons, we have focused on the bottom-up approach that yields
molecular-level dispersion of individual graphene sheets as observed
by scanning electron microscopy, SEM (Fig. 1g).
The quality of nanofiller dispersion in the polymer matrix directly
correlates with its effectiveness for improving mechanical, electrical,
thermal, impermeability and other properties. The properties of a
composite are also intimately linked to the aspect ratio and surface-to-
volume ratio of the filler. The potential properties of our graphene-
sheet-based composites thus appear promising owing to the extremely
high aspect ratios of the sheets as determined from SEM images
(Fig. 1g), in which the average lateral dimension was estimated to
be ,1mm, similar to the values found by AFM (Fig. 1b, c). However,
in contrast to the typically flat sheets seen by AFM when deposited
onto atomically flat substrates, the sheets in the composites are
crumpled and wrinkled, and at times folded.
At 2.4 volume per cent (vol.%) loading, the composite appears in
the SEM images to be almost entirely filled with the graphene sheets,
even though 97.6 vol.% is still filled by the polymer (Fig. 1g, right
panel). This ‘visual’ effect is due to the enormous surface area of the
sheets. The specific surface area of an individual graphene sheet is
,2,600m2 g21, so 1mm3 of a composite with 1 vol.% graphene sheets
would have ,50 mm2 of sheet surface area. SEM image analysis was
used to estimate the apparent graphene sheet surface area at different
concentrations (Fig. 2a–d). It revealed a surface area lower by roughly
a factor of two than the calculated values for each concentration. The
nanoscale corrugation of the graphene sheets (not evident in SEM
imaging) may contribute to the discrepancy. The intrinsic mechanics
of the sheets, their crumpling, wrinkling and folding, and whether
they can be processed to be non-crumpled (for example, so that
Figure 1 | Process flow of graphene–polymer composite fabrication.
a, SEM and digital image (inset) of natural graphite. b, A typical AFM
non-contact-mode image of graphite oxide sheets deposited onto a mica
substrate from an aqueous dispersion (inset) with superimposed cross-
sectionmeasurements taken along the red line indicating a sheet thickness of
,1 nm. c, AFM image of phenyl isocyanate-treated graphite oxide sheets on
mica and profile plot showing the,1 nm thickness. d, Suspension of phenyl
isocyanate-treated graphite oxide (1mgml21) and dissolved polystyrene in
DMF before (left) and after (right) reduction by N,N-dimethylhydrazine.
e, Composite powder as obtained after coagulation in methanol.
f, Hot-pressed composite (0.12 vol.% of graphene) and pure polystyrene of
the same 0.4-mm thickness and processed in the same way. g, Low (top row)
and high (bottom row) magnification SEM images obtained from a fracture
surface of composite samples of 0.48 vol.% (left) and 2.4 vol.% (right)
graphene in polystyrene.
NATURE|Vol 442|20 July 2006 LETTERS
283

© 2006 Nature Publishing Group

layering could be more effectively achieved) are issues for further
study. Heat treatment of such composites to 300 8C, well above the
glass transition temperature of polystyrene (100 8C), does not result
in agglomeration of the graphene sheets (as assessed from extensive
SEM imaging), suggesting their thermal stability over a broad
temperature range.
As SEM cannot spatially resolve the thickness of an individual
graphene-based sheet, transmission electron microscopy (TEM) was
used to determine if the graphene-based sheets were indeed present
in the composites as single, exfoliated sheets or as multi-layered
platelets. Extensive high-resolution phase-contrast imaging showed
no evidence of multi-layer stacks, even though SEMof the same slices
clearly indicated the presence of nanoplatelets. In contrast, selected-
area electron diffraction (SAED) of composite samples yielded spot
patterns matching those expected for individual graphene-based
sheets (Fig. 2e, f, and Supplementary Information). SAED of pure
polystyrene samples fabricated in an identical manner (see Methods)
showed the ring pattern expected for an amorphous material.
The conductive nature of the graphene filler motivated our study
of the percolation behaviour through electrical measurements. As a
control, we prepared polystyrene composites containing phenyl
isocyanate-treated graphite oxide sheets that were not chemically
reduced. At similar graphene concentrations, these samples were
greyish in colour compared to the almost black composites filled with
the reduced material, and they were not electrically conductive.
To obtain the true value of the conductivity of the composites, we
made separate measurements of in-plane and transverse resistances
(Fig. 3, inset) and developed a numerical simulation programbased on
FEMLAB(version3.1,ComsolAB),which takes into account geometry
of the samples and electrical leads. All measurements and subsequent
analysis revealed minor anisotropy, with the in-plane conductivity,
jk, always slightly higher than the transverse conductivity, j’.
A rapid increase in the direct current electrical conductivity of
composite materials takes place when the conductive filler forms an
infinite network of connected paths through the insulating matrix.
When the filler particles are rigid bodies, the conductivity of such
media is typically described with a bond percolation model10. The
conductivity of the composite, j c, above the percolation threshold
is then treated with a power law: j c ¼ j f[(f 2 f c)/(1 2 f c)]t
(ref. 24), where j f is the conductivity of the filler, f the filler volume
fraction, f c the percolation threshold (the onset of the transition),
and t is the ‘universal critical exponent’.
Percolation in polystyrene–graphene composites occurs when the
filler concentration, f c, is near 0.1 vol.% (Fig. 3). This percolation
threshold is, to the best of our knowledge, about three times lower
than reported for any other two-dimensional filler25. Such a low
percolation threshold for a three-dimensional isotropic case is
evidently due to the extremely high aspect ratio of the graphene
sheets and their excellent homogeneous dispersion in these compo-
sites. Randomly oriented oblate ellipsoids (disks) with an aspect ratio
of 1,000 are predicted to have a geometric percolation threshold of
,0.1 vol.% (ref. 26). At a loading of about 0.15 vol.%, the conduc-
tivity of our composites already satisfies the antistatic criterion
(1026 Sm21) for thin films12 and rapidly rises over a 0.4 vol.%
range. An increase in graphene sheet loading above 0.5 vol.% yields
a more gradual increase in electrical conductivity, with values of
,0.1 Sm21 at 1 vol.% and ,1 Sm21 at 2.5 vol.%.
At present, research in nanocomposite materials using carbon-
based nanofillers is dominated by carbon nanotubes. The electrical
properties of our composites compare well with the best values
reported in the literature for nanotube–polymer composites. The
Figure 2 | SEM and TEM images of graphene–polystyrene composite.
a–d, SEM images of the microtomed composites reveal different
morphologies of the graphene sheets, including their packing, at different
concentrations (vol.%): a, 0.24; b, 0.96; c, 1.44; and d, 2.4. Scale bar, shown
in a, applies to a–d. e, f, High-resolution phase contrast images and SAED
patterns (inset) of e, cast film made from powder composite, and f,
microtomed composite sample. The SAED patterns show the six-fold
rotational symmetry expected for diffraction with the beam incident along
[0001]. Our experimentally obtained patterns show all of the reflections
expected for such single graphene-based sheets. The d spacings associated
with the spots were determined using both gold and graphite calibrants, and
the first five sets of spots correspond to d-spacings (A˚) of 4.23, 2.45, 2.12,
1.42 and 1.23. More details are presented in Supplementary Information.
Additionally, these high-resolution images show regions where fringes are
observed and regions where they are not, which indicates that there is
significant local curvature in the sheets.
Figure 3 | Electrical conductivity of the polystyrene–graphene composites
as a function of filler volume fraction. Main figure, composite conductivity,
j c, plotted against filler volume fraction,f. Right inset, logj c plotted against
log(f 2 f c), where f c is the percolation threshold (see text). Solid lines in
both graphs are calculated conductivities based on the fitting (inset, log–log
plot) of the experimental data to the effective conductivity equation
described in the text. Fitted parameters are: t ¼ 2.74 ^ 0.20,
j f ¼ 104.92^0.52 Sm21 and f c ¼ 0.1 vol.%. Left inset: top and middle
diagrams show the four-probe setup for in-plane and transverse
measurements, respectively; bottom diagram, one of the computed
distributions of the current density (contour lines) with local directions and
magnitude (shown by arrows) in a specimen for the following conditions—
the sample thickness is twice the electrode width and the gap between them,
and the in-plane resistivity is 10 times lower than the transverse resistivity.
LETTERS NATURE|Vol 442|20 July 2006
284

© 2006 Nature Publishing Group

percolation threshold of around 0.1 vol.% is almost the same as was
achieved for single-wall carbon nanotube (SWNT)–polystyrene
composites made by latex technology27, and just twice and four
times higher than those measured for SWNT–polyimide11,24 and
poly(phenyleneethynylene)-functionalized SWNT–polystyrene
composites28, respectively. Absolute conductivities of our gra-
phene–polystyrene composites are also essentially the same as the
values reported for SWNT-filled polystyrene composites27. However,
in contrast to these reports, which feature thin film composites, our
samples are made by a quasi-industrial method (hot-pressing; we
have also made samples by injection moulding that show similar
values of electrical conductivity). Furthermore, graphene sheets have
higher surface-to-volume ratios than SWNTs owing to the inaccessi-
bility of the inner nanotube surface to polymer molecules. This
makes graphene sheets potentially more favourable for altering all
matrix properties—such as the mechanical, rheological and per-
meability properties, and degradation stability. Also, SWNTs are still
much more expensive than graphite, which is a commodity material
with about 1,000,000 t used annually, and which is sold for a few US
dollars per kilogram29. In contrast, graphite oxide can be readily
prepared from a vast array of graphites, and its industrial production
has also recently become a topic of interest30.
We therefore expect that our method for preparation and incorpo-
ration of individual graphene sheets into polymer matrices will lead
to the further development of a broad new class of materials with
enhanced properties and even introduce new functionalities to
polymer composites.
METHODS
Fabrication of composites. Graphite oxide was prepared by the Hummers
method from SP-1 graphite (Bay Carbon), and dried for a week over phosphorus
pentoxide in a vacuum desiccator. Dried graphite oxide (50mg) was suspended
in anhydrous DMF (5ml, Dow-Grubbs solvent system), treated with phenyl
isocyanate (2mmol, Sigma-Aldrich) for 24 h, and recovered by filtration
through a sintered glass funnel (50ml, medium porosity). Stable dispersions
of the resulting phenyl isocyanate-treated graphite oxidematerials were prepared
by ultrasonic exfoliation (Fisher Scientific FS60, 150W, 1 h) in DMF
(1mgml21). Polystyrene (Scientific Polymer Products, approximate
Mw ¼ 280 kD, PDI ¼ 3.0) was added to these dispersions and dissolved with
stirring (Fig. 1d, left). Reduction of the dispersed material (Fig. 1d, right) was
carried out with dimethylhydrazine (0.1ml in 10ml of DMF, Sigma-Aldrich) at
80 8C for 24 h. Upon completion, the coagulation of the polymer composites was
accomplished by adding the DMF solutions dropwise into a large volume of
vigorously stirred methanol (10:1 with respect to the volume of DMF used). The
coagulated composite powder (Fig. 1e) was isolated via filtration; washed with
methanol (200ml); dried at 130 8C under vacuum for ,10 h to remove residual
solvent, anti-solvent, and moisture; crushed into a fine powder with a mortar
and pestle, and then pressed (Fig. 1f) in a hydraulic hot press (Model 0230C-X1,
PHI-Tulip) at 18 kN with a temperature of 210 8C. Further description of
the composite samples’ preparation is given in Supplementary Information.
To convert wt% loading of graphene sheets in the composite samples to vol.%
(as used in the text), a density for the phenyl isocyanate-treated graphite oxide
sheets of 2.2 g cm23 was assumed along with the known density of polystyrene,
1.05 g cm23. Composites of acrylonitrile-butadiene-styrene and styrene-
butadiene rubbers (Sigma-Aldrich) were similarly prepared using the appropriate
copolymers.
Microscopy. AFM images were taken on an AutoProbe CP/MT scanning probe
microscope (MultiTask; Veeco Instruments). Imaging was done in non-contact
mode using a V-shaped ‘Ultralever’ probe B (Park Scientific Instruments, boron-
doped Si with frequency f c ¼ 78.6 kHz, spring constants k ¼ 2.0–3.8Nm21,
and nominal tip radius 10 nm). All images were collected under ambient
conditions at 50% relative humidity and 23 8C with a scanning raster rate of
1Hz. Samples for AFM images were prepared by depositing dispersions of
functionalized graphite oxide inDMFon a freshly cleavedmica surface (Ted Pella
Inc.) and allowing them to dry in air.
Electron micrographs were acquired using a field emission SEM (LEO 1525)
and a TEM (JEM-3010, JEOL, operated at 300 kV). Two SEM imaging con-
ditions were used: ‘surface imaging’ operating at #1 kV, and ‘sub-surface
imaging’ at $5 kV. Microtomed and cast film samples were prepared for
TEM. Pressed composites were microtomed to slices of 40–60 nm thickness
(Leica Ultracut UCT, Reichert Inc.) and dropped onto standard TEM grids.
Samples investigated in this way were 0.24, 1.44 and 2.4 vol.%. A polystyrene
control sample was also prepared and imaged in this way. Before imaging with
TEM, SEM was used to verify that a sufficient amount of highly dispersed
platelets were present. Cast-film samples were prepared to image graphene sheets
‘lying flat’ on the carbon support film, with the beam incident along the [0001]
direction of the graphene sheet. A small amount (,0.01mg) of un-pressed
powder composite was dissolved in toluene (10ml) together with additional
polystyrene (,0.02mg) as diluent. The solutionwas sonicated (Crest Ultrasonic
Tru-Sweep 175HT, 45W, 1 h) before casting, producing a highly dilute solution
that, when cast on a mica substrate, formed a film approximately 40 nm thick, as
measured by AFM.
Electrical measurements. Hot-pressed composite samples having 0.3–0.5mm
thickness were cut into strips (1–2mm wide and 15–20mm long). A thin ‘skin’
layer deficient in the phenyl isocyanate-treated graphite oxide platelets was
removed by oxygen plasma etching (Plasma-Preen II-862, Plasmatic Systems;
2 torr O2, 1.5min, 350W) to reduce the contact resistance between the metal
leads and sample tested. After etching, a thin gold film (,25 nm) was thermally
deposited (BOC Edwards Auto 306 Evaporation System) on one side of the
sample for ‘longitudinal’ measurements (gap between contacts is 0.15mm), and
on the top and bottom sides for ‘transverse’measurements. The d.c. resistance of
the composites was measured with a d.c. power supply (HP6612C), digital
multimeter (HP34401A) and picoammeter (Keithley 6485) using a standard
four-probe technique with a threshold detection limit of 0.1GQ.
Received 18 January; accepted 5 June 2006.
1. Dresselhaus, M. S. & Dresselhaus, G. Intercalation compounds of graphite. Adv.
Phys. 51, 1–-186 (2002).
2. Hirata, M., Gotou, T., Horiuchi, S., Fujiwara, M. & Ohba, M. Thin-film particles
of graphite oxide 1: High-yield synthesis and flexibility of the particles. Carbon
42, 2929–-2937 (2004).
3. Yu, M. F., Lourie, O., Moloni, K., Kelly, T. F. & Ruoff, R. S. Strength and breaking
mechanism of multiwalled carbon nanotubes under tensile load. Science 287,
637–-640 (2000).
4. Novoselov, K. S. et al. Electric field effect in atomically thin carbon films.
Science 306, 666–-669 (2004).
5. Novoselov, K. S. et al. Two-dimensional gas of massless Dirac fermions in
graphene. Nature 438, 197–-200 (2005).
6. Zhang, Y., Tan, Y.-W., Stormer, H. L. & Kim, P. Experimental observation of the
quantum Hall effect and Berry’s phase in graphene. Nature 438, 201–-204
(2005).
7. Zhang, Y., Small, J. P., Amori, M. E. S. & Kim, P. Electric field modulation of
galvanomagnetic properties of mesoscopic graphite. Phys. Rev. Lett. 94, 176803
(2005).
8. Berger, C. et al. Ultrathin epitaxial graphite: two-dimensional electron gas
properties and a route toward graphene-based nanoelectronics. J. Phys. Chem.
B 108, 19912–-19916 (2004).
9. Stankovich, S. et al. Stable aqueous dispersions of graphitic nanoplatelets via
the reduction of exfoliated graphite oxide in the presence of poly(sodium 4-
styrenesulfonate). J. Mater. Chem. 16, 155–-158 (2006).
10. Stauffer, D. & Aharnoy, A. Introduction to Percolation Theory (Taylor & Francis,
London, 1991).
11. Ounaies, Z., Park, C., Wise, K. E., Siochi, E. J. & Harrison, J. S. Electrical
properties of single wall carbon nanotube reinforced polyimide composites.
Compos. Sci. Technol. 63, 1637–-1646 (2003).
12. Chung, D. D. L. Electrical applications of carbon materials. J. Mater. Sci. 39,
2645–-2661 (2004).
13. He, H., Klinowski, J., Forster, M. & Lerf, A. A new structural model for graphite
oxide. Chem. Phys. Lett. 287, 53–-56 (1998).
14. Lerf, A., He, H., Forster, M. & Klinowski, J. Structure of graphite oxide revisited.
J. Phys. Chem. B 102, 4477–-4482 (1998).
15. Hirata, M., Gotou, T. & Ohba, M. Thin-film particles of graphite oxide 2:
Preliminary studies for internal micro fabrication of single particle and
carbonaceous electronic circuits. Carbon 43, 503–-510 (2005).
16. Szabo, T., Szeri, A. & Dekany, I. Composite graphitic nanolayers prepared by
self-assembly between finely dispersed graphite oxide and a cationic polymer.
Carbon 43, 87–-94 (2005).
17. Kovtyukhova, N. I. et al. Layer-by-layer assembly of ultrathin composite films
from micron-sized graphite oxide sheets and polycations. Chem. Mater. 11,
771–-778 (1999).
18. Kotov, N. A., Dekany, I. & Fendler, J. H. Ultrathin graphite oxide-polyelectrolyte
composites prepared by self-assembly. Transition between conductive and
non-conductive states. Adv. Mater. 8, 637–-641 (1996).
19. Cassagneau, T., Guerin, F. & Fendler, J. H. Preparation and characterization of
ultrathin films layer-by-layer self-assembled from graphite oxide nanoplatelets
and polymers. Langmuir 16, 7318–-7324 (2000).
20. Vaia, R. A. & Wagner, H. D. Framework for nanocomposites. Mater. Today 7,
32–-37 (2004).
21. Du, X. S., Xiao, M., Meng, Y. Z. & Hay, A. S. Novel synthesis of conductive
poly(arylene disulfide)/graphite nanocomposite. Synth. Met. 143, 129–-132 (2004).
NATURE|Vol 442|20 July 2006 LETTERS
285

© 2006 Nature Publishing Group

22. Liu, P. & Gong, K. Synthesis of polyaniline-intercalated graphite oxide by an in
situ oxidative polymerization reaction. Carbon 37, 706–-707 (1999).
23. Stankovich, S., Piner, R. D., Nguyen, S. T. & Ruoff, R. S. Synthesis and
exfoliation of isocyanate-treated graphene oxide nanoplatelets. Carbon (in the
press).
24. McLachlan, D. S. et al. AC and DC percolative conductivity of single wall
carbon nanotube polymer composites. J. Polym. Sci. B 43, 3273–-3287 (2005).
25. Chen, G., Weng, W., Wu, D. & Wu, C. PMMA/graphite nanosheets composite
and its conducting properties. Eur. Polym. J. 39, 2329–-2335 (2003).
26. Garboczi, E. J., Snyder, K. A., Douglas, J. F. & Thorpe, M. F. Geometrical
percolation threshold of overlapping ellipsoids. Phys. Rev. E 52, 819–-828
(1995).
27. Grossiord, N., Loos, J. & Koning, C. E. Strategies for dispersing carbon
nanotubes in highly viscous polymers. J. Mater. Chem. 15, 2349–-2352 (2005).
28. Ramasubramaniam, R., Chen, J. & Liu, H. Homogeneous carbon nanotube
polymer composites for electrical applications. Appl. Phys. Lett. 83, 2928–-2930
(2003).
29. Olson, D. W. USGS Mineral Commodity Summaries. khttp://minerals.usgs.gov/
minerals/pubs/commodity/graphite/graphmcs06.pdfl (2006).
30. Titelman, G. I. & Bron, S. Method of Manufacturing of Graphite Oxide (GO)
P1391–-P1392 (Research Disclosure 500, IMI TAMI Institute for Research &
Development Ltd., Haifa, 2005).
Supplementary Information is linked to the online version of the paper at
www.nature.com/nature.
Acknowledgements SEM was done at the Electron Probe Instrumentation
Centre at Northwestern University. This work was funded by the NASA
University Research, Engineering and Technology Institute on Bio-Inspired
Materials (BIMat) and by the NSF NIRT grant ‘Nanostructured carbons from
self-assembled block copolymer precursors: From synthesis and
characterization to devices’. We appreciate J. A. Ibers and A. L. Ruoff for
critically reading an early version of this manuscript.
Author Information Reprints and permissions information is available at
npg.nature.com/reprintsandpermissions. The authors declare no competing
financial interests. Correspondence and requests for materials should be
addressed to R.S.R. (r-ruoff@northwestern.edu) or S.T.N.
(stn@northwestern.edu).
LETTERS NATURE|Vol 442|20 July 2006
286