LETTERS
Longitudinal unzipping of carbon nanotubes to form
graphene nanoribbons
Dmitry V. Kosynkin
1
, Amanda L. Higginbotham
1
, Alexander Sinitskii
1
, Jay R. Lomeda
1
, Ayrat Dimiev
1
,
B. Katherine Price
1
& James M. Tour
1,2,3
Graphene, or single-layered graphite, with its high crystallinity and
interesting semimetal electronic properties, has emerged as an
exciting two-dimensional material showing great promise for the
fabrication of nanoscale devices
1–3
. Thin, elongated strips of
graphene that possess straight edges, termed graphene ribbons,
gradually transform from semiconductors to semimetals as their
width increases
4–7
, and represent a particularly versatile variety of
graphene. Several lithographic
7,8
, chemical
9–11
and synthetic
12
procedures are known to produce microscopic samples of graphene
nanoribbons, and one chemical vapour deposition process
13
has
successfully produced macroscopic quantities of nanoribbons at
950
6C. Here we describe a simple solution-based oxidative process
for producing a nearly 100% yield of nanoribbon structures by
lengthwise cutting and unravelling of multiwalled carbon nano-
tube (MWCNT) side walls. Although oxidative shortening of
MWCNTs has previously been achieved
14
, lengthwise cutting is
hitherto unreported. Ribbon structures with high water solubility
are obtained. Subsequent chemical reduction of the nanoribbons
from MWCNTs results in restoration of electrical conductivity.
These early results affording nanoribbons could eventually lead
to applications in fields of electronics and composite materials
where bulk quantities of nanoribbons are required
15–17
.
We obtained oxidized nanoribbons by suspending MWCNTs in
concentrated sulphuric acid followed by treatment with 500 wt%
KMnO
4
for 1 h at room temperature (22 uC) and 1 h at 55–70 uC
(Methods). After isolation, the resulting nanoribbons were highly
soluble in water (12 mg ml
21
), ethanol and other polar organic
solvents. The opening of the nanotubes appears to occur along a line,
similar to the ‘unzipping’ of graphite oxide
18,19
, affording straight-
edged ribbons. This could occur in a linear longitudinal cut (Fig. 1a)
or in a spiralling manner, depending upon the initial site of attack
and the chiral angle of the nanotube. Although depicted in Fig. 1a as
occurring on the mid-section of the nanotube rather than at one end,
the location of the initial attack is not known.
The mechanism of opening is based on previous work on the oxida-
tion of alkenes by permanganate in acid. The proposed first step in
the process is manganate ester formation (2, Fig. 1b) as the rate-
determining step, and further oxidation is possible to afford the
dione (3, Fig. 1b) in the dehydrating medium
20
. Juxtaposition of the
buttressing ketones distorts the b,c-alkenes (red in 3), making them
more prone to the next attack by permanganate. As the process
continues, the buttressing-induced strain on the b,c-alkenes lessens
because there is more space for carbonyl projection; however, the
bond-angle strain induced by the enlarging hole (or tear if originating
from the end of the nanotube) would make the b,c-alkenes (4, Fig. 1b)
increasingly reactive. Hence, once an opening has been initiated, its
further opening is enhanced relative to an unopened tube or to an
uninitiated site on the same tube. The ketones can be further
converted, through their O-protonated forms, to the carboxylic
acids
21
that will line the edges of the nanoribbons. Finally, relief of
the bond-angle strain when the nanotube opens to the graphene
ribbon (5, Fig. 1b) slows further dione formation and cutting
20
.
Thus, the preference for sequential bond cleavage over random
opening and subsequent cutting, as occurs with nitric acid oxidation,
can be explained by concerted attachment to neighbouring carbon
atoms by permanganate, contrasting with the random attack on non-
neighbouring carbon atoms by the nitronium species from nitric acid.
The surface of the now-less-strained nanoribbon remains prone to
1,2-diol formation, which leads to the overall highly oxidized ribbon,
but this is less likely to result in further oxidative cutting to the dione
owing to relief of the tubular strain on the double bonds.
We achieved the same unzipping process in single-walled carbon
nanotubes (SWCNTs), to produce narrow nanoribbons, but their
subsequent disentanglement is more difficult. See Supplementary
Figs 5 and 6 for images and analysis of those SWCNT-derived narrow
nanoribbons and their reduction products.
We used transmission electron microscopy (TEM), atomic force
microscopy (AFM) and scanning electron microscopy (SEM) to
image the ribbon structures. TEM analysis shows nanoribbons
(Fig. 1c) produced from MWCNTs with a starting diameter of 40–
80 nm and approximately 15–20 inner nanotube layers (additional
TEM images of untreated MWCNTs can be found in the
Supplementary Fig. 1a, b). After reaction, the width of the carbon
nanostructures increased to .100 nm and they had linear edges with
little pristine MWCNT side-wall structure remaining (see
Supplementary Figs 1 and 2 for more images). The MWCNTs used
were produced from a chemical vapour deposition process
22
;we
attempted the same H
2
SO
4
–KMnO
4
treatment on a single sample
of laser-oven-produced MWCNTs, but fewer nanoribbon-like struc-
tures were detected. AFM imaging (Fig. 1d) shows the presence of
single atomic layers after tip sonication of the solution for 30 min to
yield well-dispersed and sonication-shortened ribbons suitable for
imaging. SEM imaging (Fig. 1e) of nanoribbons on a silicon surface
shows that the ribbons remain long (,4 mm in this image) when not
cut by tip sonication; they can be dispersed as single or thin layers and
they display uniform widths and predominantly straight edges over
their entire length (see Supplementary Fig. 1c, d for other images).
The degree of consecutive tube opening in the MWCNTs can also
be controlled by adjusting the amount of oxidizing agent introduced
into the system; using TEM, we found that in 80–100% of the
MWCNTs present, the side walls completely unravelled to form
nanoribbons when 500 wt% KMnO
4
was used. The successive open-
ing reaction was demonstrated in five iterations, each containing a
stepwise increase in the amount of KMnO
4
: 100 wt% KMnO
4
in the
1
Department of Chemistry,
2
Department of Mechanical Engineering and Materials Science,
3
Smalley Institute for Nanoscale Science and Technology, Rice University, MS-222, 6100
Main Street, Houston, Texas 77005, USA.
Vol 458
|
16 April 2009
|
doi:10.1038/nature07872
872
Macmillan Publishers Limited. All rights reserved
©2009
first iteration (sample I), 200 wt% in the second iteration (sample II),
and so on until the final iteration, when we used 500 wt% (sample V).
This resulted in consecutive unencapsulation of the different layers
by unzipping of the successive MWCNTs (see Methods for details). It
is evident from TEM images (Fig. 2a–e) that the walls of the
MWCNTs open to a higher degree as the level of oxidation increases,
with less MWCNT inner tube remaining in successive iterations. This
is highlighted in a statistical plot (Fig. 2f) showing the decrease of the
average diameter of remaining MWCNTs from ,65 nm to ,20 nm
as the amount of KMnO
4
exposure is increased. The smaller-
diameter tubes that remained after treatment with 500 wt%
KMnO
4
were exposed to the reaction conditions for less time than
the larger-diameter tubes and, thus, may not have had the chance to
fully react; no difference in the rate of unzipping between smaller-
and larger-diameter nanotubes can be inferred from this data.
The degree of oxidation of the product formed (partly and/or
completely unravelled MWCNTs) from each of the five iterative
KMnO
4
treatment steps was monitored using attenuated-total-
reflection infrared (ATR–IR) spectroscopy and thermogravimetric
Tube unzipping
a
OO
MnO
2
–
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
1
2
3
4
5
KMnO
4
H
2
SO
4
2 KMnO
4
H
2
SO
4
b
KMnO
4
H
2
SO
4
22–70 °C
2 h
50 nm 50 nm
c
2 m
d
2.0 nm
(μm)
0.5 1.5
1.0 nm
500 nm
500 nm
Sonication cut
500 nm
e
Figure 1
|
Nanoribbon formation and imaging. a, Representation of the
gradual unzipping of one wall of a carbon nanotube to form a nanoribbon.
Oxygenated sites are not shown.
b, The proposed chemical mechanism of
nanotube unzipping. The manganate ester in
2 could also be protonated.
c, TEM images depicting the transformation of MWCNTs (left) into oxidized
nanoribbons (right). The right-hand side of the ribbon is partly folded onto
itself. The dark structures are part of the carbon imaging grid.
d, AFM images
of partly stacked multiple short fragments of nanoribbons that were
horizontally cut by tip-ultrasonic treatment of the original oxidationproduct
to facilitate spin-casting onto the mica surface. The height data (inset)
indicates that the ribbons are generally single layered. The two small images
on the right show some other characteristic nanoribbons.
e, SEM image of a
folded, 4-mm-long single-layer nanoribbon on a silicon surface.
20
30
40
50
60
70
80
Average MWCNT
diameter (nm)
Sample
100 nm 100 nm
100 nm
100 nm
100 nm
ab
c d
e
f
Increasing oxidation
IIIIIIIVV
III
III
IV
V
4,000 3,000 2,000 1,000
0.0
0.5
Absorbance
Wavenumber (cm
–1
)
V Most oxidized
0.0
0.5
IV
0.0
0.5
III
0.0
0.5
II
0.0
0.5
I Least oxidized
g
10 20 30
0
1
Intensity (a.u.)
2θ (°)
V Most oxidized
0
1
IV
0
1
III
0
1
II
0
1
I Least oxidized
h
Figure 2
|
Stepwise opening of MWCNTs to form nanoribbons. a–e, TEM
images of the stepwise opening of MWCNTs representing the incremental
exposure of the system to KMnO
4
: the least oxidized sample (sample I) is in
a and the most oxidized sample (sample V) is in e. f, Scatter plot showing how
the average MWCNT diameter (determined from studying 15–20 TEM
images per sample, each with ,5 MWCNTs per image) changes with
increasing exposure to KMnO
4
. Error bars indicate the standard deviation of
the average MWCNT diameter across the sample.
g, ATR–IR spectroscopy of
stepwise opening/oxidation of MWCNTs.
h, X-ray diffraction analysis of the
stepwise opening of the nanotube. h, diffraction angle; a.u., arbitrary units.
NATURE
|
Vol 458
|
16 April 2009 LETTERS
873
Macmillan Publishers Limited. All rights reserved
©2009
analysis (TGA). ATR–IR spectroscopy (Fig. 2g) reveals the appearance
of a C5O stretch (purple region, Fig. 2g) increasing from 1,690 cm
21
in sample III (green line) to 1,710 cm
21
in sample V (black line),
consistent with declining conjugation. The COO–H/O–H stretch
(,3,600–2,800 cm
21
; yellow region, Fig. 2g) appears with sample
III and continues to increase through the series, indicating an increase
in the number of carboxyl and hydroxyl functionalities as well as the
possible presence of trapped water. TGA shows an increase in the total
weight loss (20% and 49% in samples I and V, respectively) with
increasing exposure to KMnO
4
, implying an increase in the number
of volatile side-wall functionalities present, which corroborates there
being a higher degree of oxidation (Supplementary Fig. 3a).
Furthermore, Raman spectroscopy (Supplementary Fig. 3b) shows
an increasing level of disorder (appearance of a D band at 1,321–
1,328 cm
21
) with increasing oxidation, consistent with ATR–IR spec-
troscopy and TGA observations.
We also performed X-ray diffraction analysis (Fig. 2h), to investi-
gate further the structure of the partly and completely unzipped
MWCNT–nanoribbon structures. The graphite (002) spacing
increases with the level of oxidation. Samples I–III all have 2h values
of ,25.8u, corresponding to a d spacing of 3.4 A
˚
. Sample IV shows
two peaks, one at 10.8u and one at 25.4u, with d spacings of 8.2 A
˚
and
3.5 A
˚
, respectively. Sample V shows a predominant peak at 10.6u,
corresponding to a d spacing of 8.3 A
˚
, with minimal signal contributed
by MWCNTs (2h 5 25.8u); this spectrum is very similar to that of
graphite oxide (Supplementary Fig. 3c).
Both the nanoribbons and graphite oxidepossess oxygen-containing
functionalities such as carbonyls, carboxyls and hydroxyls
23
that have
been shown to existat the edges and the surface
24
. The surface oxidation
disrupts the p-conjugated network, rendering the nanoribbons and
graphite oxide poorly conductive. Hydrazine (N
2
H
4
) reduction of
graphite oxide
25
is known to provide a means of restoring conjugation
and, thus, some of the conductivity, to form chemically converted
graphene (CCG)
25–28
. The structure of CCG is thought to be a patch-
work of intact graphene islands interspersed with regions of tetrahedral
sp
3
-hybridized carbon atoms due to incomplete reduction and
incomplete re-aromatization; therefore, the electrical conductivity is
not as highas that found in the originalgraphite
26
. Thecarboxyl groups,
which are found predominately at the edges
25
, are not reduced by N
2
H
4
and remain in the product, further disrupting the p network
7,8
.
Furthermore, as the number of oxygen-containing functionalities
decreases during the reduction process, the tendency to aggregate as
aresultofp stacking increases.
The reduction of oxidized nanoribbons was carried out with aqueous
N
2
H
4
in the presence of ammonia. To prevent re-aggregation during
the reduction procedure, we first dispersed the nanoribbons in an
aqueous surfactant solution, sodium dodecyl sulphate (SDS), to
produce stable dispersions of reduced nanoribbons that retained their
straight-edged structure (Fig. 3a). The reaction progress was monitored
by ultraviolet absorption (Fig. 3b); the bathochromic shift of l
max
and
the hyperchromicity over the entire range (.230 nm) indicates that
electronic conjugation of the ribbons was restored
25
.
To provide further evidence that the reduction procedure
decreased the number of oxygen-containing functionalities from
the nanoribbon surface, we performed ATR–IR spectroscopy,
X-ray photoemission spectroscopy (XPS; Fig. 3c, d) and TGA. The
reduced nanoribbons show almost complete elimination of the
COO–H/O–H stretching region (,3,600–2,800 cm
21
; yellow region,
Fig. 3c) and a significant decrease in the C5O stretching region
(,1,710 cm
21
; purple region, Fig. 3c) in the ATR–IR spectrum (blue
line, Fig. 3c) in comparison with the intense COO–H/O–H and C5O
stretches observed for the oxidized nanoribbons (red line, Fig. 3c).
Edge carboxylic acids will remain.
In the XPS carbon 1s spectra of the oxidized and reduced nanoribbons
(Fig. 3d), the signals at 286 eV and 287 eV correspond to C–O and C5O,
respectively. The shoulder at 289 eV is assigned to carboxyl groups.
Upon reduction (blue line), the 286- and 287-eV peaks diminish to a
shoulder of the C–C peak (284.4 eV), indicating significant deoxygena-
tion of the nanoribbons by N
2
H
4
. As reported for CCG, the most
dominant peak after reduction is the C–C peak at 284.8 eV (ref. 29).
In addition, the XPS-determined atomic concentration of oxygen (com-
plete table found in Supplementary Fig. 4b) decreases from 42% to 16%
upon reduction, but is still higher than the oxygen content of MWCNTs
(2.1%), owing, in large part, to the edge carboxylic acid moieties.
The TGA weight loss of the reduced nanoribbons was ,33% less
than that of the oxidized starting material, which also indicates that
fewer oxygen-containing functionalities are presenton the nanoribbon
surface (see Supplementary Fig. 4d for TGA curves). The TEM image of
a reduced nanoribbon shows its straight edge and buckled appearance
(Fig. 3a). Nitrogen adsorption measurements of as-prepared and
reduced nanoribbons give surface areas, determined using
Brunauer–Emmett–Teller theory, of 445 and 436 m
2
g
21
,respectively,
after pre-outgassing at 400 uC for 12 h (ref. 13). The density of the
oxidized ribbons was found to be 2.0 g cm
23
using solution density
matching (bromotrichloromethane). When considering the overall
conversion of MWCNTs to reduced nanoribbons, the material weight
yield is 99% (Methods), underscoring the efficiency of the overall
process.
Recent interest in graphene nanoribbons has focused on the study of
the reactive edges having zigzag or armchair morphologies that
dominate their electronic and magnetic behaviour
5
. Although
zigzag-edged structures are presumed by the mechanism described
50 nm
295 290 285 280
0.0
0.2
0.4
0.6
0.8
1.0
Intensity (a.u.)
Binding energy (eV)
C–O,
C=O
COOH
C–C
4,000 3,000 2,000 1,000
0.0
0.1
0.0
0.1
0.0
0.1
Absorbance (a.u.)
200 400 600 800 1,000 1,200 1,400
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Absorbance (a.u.)
Wavelength (nm)
cba d
Wavenumber (cm
–1
)
Figure 3
|
Characterization of the oxidized and reduced nanoribbons
derived from MWCNTs. a
, TEM image of reduced nanoribbons obtained by
treatment of oxidized nanoribbons with N
2
H
4
. Detailed examination of the
image reveals that 2–3 ribbons are stacked with apparent buckling. The dark
structures are part of the carbon imaging grid.
b, Changes in the ultraviolet
spectrum of an aqueous solution of oxidized nanoribbons (red,
l
max
5 234 nm) after treatment with N
2
H
4
(blue, l
max
5 267 nm). c, ATR–IR
spectroscopy of nanoribbons before (red) and after reduction (blue),
compared with MWCNT starting material (black).
d, Normalized,
superimposed XPS carbon 1s spectra of the oxidized nanoribbons (red) and
the reduced nanoribbons (blue).
LETTERS NATURE
|
Vol 458
|
16 April 2009
874
Macmillan Publishers Limited. All rights reserved
©2009
here, we were unable to achieve the edge resolution needed to confirm
this. Thisis due, in part, to edgecurling, and could be further attributed
to the extensive edge oxidation; this may be removed only upon treat-
ment at $2,000 uC, which would result in reconstruction and modified
electronic properties.
In spite of the oxidized edges and planes of the nanoribbons derived
by this bulk process, the electronic properties of the nanoribbons before
and after chemical reduction were studied by building three-terminal
devices on Si–SiO
2
substrates. The long length of the nanoribbons
make them easily adapted structures for device fabrication (Fig. 4a);
electron-beam-patterned platinum electrodes were evaporated on
top of the nanoribbon stack. As-prepared nanoribbons are poor con-
ductors owing to the high number of oxygen-containing functionalities
present on the surface; however, their conductivity can be dramatically
increased either by chemical reduction using N
2
H
4
or by annealing in
H
2
(Fig. 4b). Thick nanoribbon stacks show little gate effect, which is in
accord with previously reported data
30
. Conversely, bilayers of these
reduced graphene nanoribbons have field-effect properties with a min-
imum conductivity at zero gate voltage, which is as expected for
undoped field-effect devices made from exfoliated graphene sheets
and is superior to CCGs (Fig. 4c, d)
2,30
. The conductivities obtained
from these wide nanoribbons are analogous to device properties
reported
11,13
for other wide nanoribbons either exfoliated or grown
by chemical vapour deposition. We have so far been unable to build
acceptable devices from narrow nanoribbons derived from SWCNTs,
owing to their extreme entanglement (Supplementary Fig. 5); more-
over, edge oxidation in those small structures may retard their elec-
tronic utility. Although the preparative route described here can have
the advantage of producing accessible nanoribbons on a large scale,
these unzipping-derived nanoribbons, with their residual oxidized
defect sites, have electronic characteristics inferior to those of wide,
mechanically peeled sheets of graphene
2,30
.
METHODS SUMMARY
Nanoribbon formation. MWCNTs were used as received from Mitsui & Co. (lot
no. 05072001K28). We suspended MWCNTs in concentrated sulphuric acid
(H
2
SO
4
) for a period of 1–12 h and then treated them with 500 wt% potassium
permanganate (KMnO
4
). The H
2
SO
4
conditions aid in exfoliating the nanotube
and the subsequent graphene structures. The reaction mixture was stirred at room
temperature for 1 h and then heated to 55–70 uC for an additional 1 h. When all of
the KMnO
4
had been consumed, we quenched the reaction mixture by pouring
over ice containing a small amount of hydrogen peroxide (H
2
O
2
). The solution
was filtered over a polytetrafluoroethylene (PTFE) membrane, and the remaining
solid was washed with acidic water followed by ethanol/ether.
Stepwise oxidation of MWCNTs to nanoribbons. We followed the above reac-
tion procedure, except that 100 wt% KMnO
4
was added in portionsuntil 500 wt%
was achieved. When the KMnO
4
had been consumed at every step, a portionof the
reaction solution was removed and worked up for analysis as described above.
Nanoribbon reduction. We treated a water solution (200 mg l
21
) of the above-
isolated nanoribbons (with or without 1 wt% SDS surfactant) with 1 vol% con-
centrated ammonium hydroxide (NH
4
OH) and 1 vol% hydrazine monohydrate
(N
2
H
4
?H
2
O). Before being heating to 95 uC for 1 h, the solution was covered
with a thin layer of silicon oil.
Full Methods and any associated references are available in the online version of
the paper at www.nature.com/nature.
Received 1 October 2008; accepted 11 February 2009.
1. Geim, A. K. & Novoselov, K. S. The rise of graphene. Nature Mater. 6, 183
–
191
(2007).
2. Novoselov, K. S. et al. Two-dimensional gas of massless Dirac fermions in
graphene. Nature 438, 197
–
200 (2005).
3. 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).
4. Areshkin, D. A., Gunlycke, D. & White, C. T. Ballistic transport in graphene
nanostrips in the presence of disorder: importance of edge effects. Nano Lett. 7,
204
–
210 (2007).
5. Nakada, K., Fujita, M., Dresselhaus, G. & Dresselhaus, M. S. Edge state in
graphene ribbons: nanometer size effect and edge shape dependence. Phys. Rev. B
54, 17954
–
17961 (1996).
6. Son, Y.-W., Cohen, M. L. & Louie, S. G. Energy gaps in graphene nanoribbons. Phys.
Rev. Lett. 97, 216803 (2006).
7. Han, M. Y., Oezyilmaz, B., Zhang, Y. & Kim, P. Energy band-gap engineering of
graphene nanoribbons. Phys. Rev. Lett. 98, 206805 (2007).
8. Chen, Z., Lin, Y.-M., Rooks, M. J. & Avouris, P. Graphene nano-ribbon electronics.
Physica E 40, 228
–
232 (2007).
9. Schniepp, H. C. et al. functionalized single graphene sheets derived from splitting
graphite oxide. J. Phys. Chem. B 110, 8535
–
8539 (2006).
10. Rollings, E. et al. Synthesis and characterization of atomically thin graphite films
on a silicon carbide substrate. J. Phys. Chem. Solids 67, 2172
–
2177 (2006).
11. Li, X., Wang, X., Zhang, L., Lee, S. & Dai, H. Chemically derived, ultrasmooth
graphene nanoribbon semiconductors. Science 319, 1229
–
1232 (2008).
12. Yang, X. et al. Two-dimensional graphene nanoribbons. J. Am. Chem. Soc. 130,
4216
–
4217 (2008).
13. Campos-Delgado, J. et al. Bulk production of a new form of sp
2
carbon: crystalline
graphene nanoribbons. Nano Lett. 8, 2773
–
2778 (2008).
14. Saito, T., Matsushige, K. & Tanaka, K. Chemical treatment and modification of
multi-walled carbon nanotubes. Physica B 323, 280
–
283 (2002).
15. Son, Y.-W., Cohen, M. L. & Louie, S. G. Half-metallic graphene nanoribbons.
Nature 444, 347
–
349 (2006).
16. Liang, G., Neophytou, N., Nikonov, D. E. & Lundstrom, M. S. Performance
projections for ballistic graphene nanoribbon field-effect transistors. IEEE Trans.
Electron. Dev. 54, 677
–
682 (2007).
17. Wang, X. et al. Room-temperature all-semiconducting sub-10-nm graphene
nanoribbon field-effect transistors. Phys. Rev. Lett. 100, 206803 (2008).
18. Li, J.-L. et al. Oxygen-driven unzipping of graphitic materials. Phys. Rev. Lett. 96,
176101 (2006).
19. Ajayan, P. M. & Yakobson, B. I. Oxygen breaks into carbon world. Nature 441,
818
–
819 (2006).
20. Wolfe, S., Ingold, C. F. & Lemieux, R. U. Oxidation of olefins by potassium per-
manganate: mechanism of a-ketol formation. J. Am.Chem. Soc.103,938
–
939(1981).
21. Banoo, F. & Stewart, R. Mechanisms of permanganate oxidation. IX.
Permanganate oxidation of aromatic alcohols in acid solution. Can. J. Chem. 47,
3199
–
3205 (1969).
22. Endo, M. Grow carbon fibers in the vapor phase. Chemtech 18, 568
–
576 (1988).
23. Hummers, W. S. Jr & Offeman, R. E. Preparation of graphitic oxide. J. Am. Chem.
Soc. 80, 1339 (1958).
24. Lerf, A., He, H., Forster, M. & Klinowski, J. Structure of graphite oxide revisited. J.
Phys. Chem. B 102, 4477
–
4482 (1998).
15.0
–1.0 –0.5 0.0 0.5 1.0
–3
–2
–1
0
1
2
3
I
sd
(10
–7
A)
V
ds
(V)
V
g
= –40 V
V
g
= –20 V
V
g
= 0 V
V
g
= 20 V
V
g
= 40 V
Pt
SiO
2
Pt
500 nm
0 0.2 0.4 0.6 0.8 1.0
0.0
5.0
10.0
15.0
20.0
Current (µA)
Voltage (V)
b
d
c
7.5
(nm)
(µm)
0.5 1.0
1.6 nm
0.75 nm
–40 0 40
1.8
2.0
2.2
2.4
2.6
2.8
V
g
(V)
V
ds
= 1 V
a
1 µm
10 µm
Figure 4
|
Device fabrication and electrical properties of graphene
nanoribbons on SiO
2
–
Si. a, SEM image of a multi-terminal device based on a
multilayer (,10-nm-thick) stack of graphene nanoribbons with platinum
electrodes. Inset, larger image of a similar device.
b, Current–voltage curves for
three different types of device: as-prepared (red), N
2
H
4
-reduced (blue) and
H
2
-annealed (green) nanoribbons (,300 nm wide, 10 nm thick (AFM) with a
channel length of ,500 nm; characteristic of the .10 devices measured for
each of the three states).
c, AFM image of another device based on a N
2
H
4
-
reduced and annealed (H
2
/Ar at 300 uC for 10 min) bilayer nanoribbon
showing that the ribbon consists of two overlapping nanoribbons in the SiO
2
channel region. Typical height profile (inset) across this nanoribbon shows
steps of about 0.75 nm, which correspond to individualgraphene sheets. These
sheets overlap in the middle, resulting in a height of ,1.6 nm.
d,Source–drain
current (I
sd
), source–drain voltage (V
ds
) and gate voltage (V
g
) dependencies
for the device shown in
c; p-doped silicon was used as a back gate.
NATURE
|
Vol 458
|
16 April 2009 LETTERS
875
Macmillan Publishers Limited. All rights reserved
©2009
25. Li, D., Mueller, M. B., Gilje, S., Kaner, R. B. & Wallace, G. G. Processable
aqueous dispersions of graphene nanosheets. Nature Nanotechnol. 3, 101
–
105
(2008).
26. Stankovich, S. et al. Synthesis of graphene-based nanosheets via chemical
reduction of exfoliated graphite oxide. Carbon 45, 1558
–
1565 (2007).
27. Bourlinos, A. B. et al. Graphite oxide: chemical reduction to graphite and surface
modification with primary aliphatic amines and amino acids. Langmuir 19,
6050
–
6055 (2003).
28. Eda, G., Fanchini, G. & Chhowalla, M. Large-area ultrathin films of reduced
graphene oxide as a transparent and flexible electronic material. Nature
Nanotechnol. 3, 270
–
274 (2008).
29. 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).
30. Novoselov, K. S. et al. Electric field effect in atomically thin carbon films. Science
306, 666
–
669 (2004).
Supplementary Information is linked to the online version of the paper at
www.nature.com/nature.
Acknowledgements The authors thank P. M. Ajayan, W. Guo, J. Duque, Z. Sun, and
Z. Jin for technical assistance and discussions. Mitsui & Co. generously donated the
MWCNTs. The work was funded by the US Defense Advanced Research Projects
Agency, the US Federal Aviation Administration, Department of Energy
(DE-FC-36-05GO15073) and Wright Patterson Air Force Laboratory through the
US Air Force Office of Scientific Research.
Author Contributions D.V.K. discovered the unzipping reaction and made most of
the analysed ribbons. A.L.H. obtained and analysed most of the analysis data
including the TEM, AFM, ultraviolet, XPS, TGA, Raman and infrared data; she also
made some of the ribbons and wrote the majority of the manuscript. A.S. fabricated
and tested the electronic devices. J.R.L performed some of the spectral analysis
including the X-ray diffraction. A.D. prepared and studied the nanorib bons on
silicon surfaces for electrical analysis. B.K.P. performed some of the AFM analyses.
J.M.T. oversaw and directed all aspects of the syntheses, data analysis and
manuscript correction and finalization.
Author Information Reprints and permissions information is available at
www.nature.com/reprints. Correspondence and requests for materials should be
addressed to J.M.T. (tour@rice.edu).
LETTERS NATURE
|
Vol 458
|
16 April 2009
876
Macmillan Publishers Limited. All rights reserved
©2009