Journal of

Materials Chemistry C
Published on 22 August 2013. Downloaded by Université de Strasbourg, Service Commun de la Documentation on 10/02/2016 16:36:25.

View Article Online

PAPER View Journal | View Issue

Efficient surfactant-free and chemical reductant-free

solvothermal deoxidation of solution-processable
Cite this: J. Mater. Chem. C, 2013, 1,
7246 sub-stoichiometric graphene oxide
Zhi-Li Chen,†b Fong-Yu Kam,†a Venu Keerthi,a Jie Song,a Chen Hu,a
Loke-Yuen Wong,b Geok-Kieng Limbc and Lay-Lay Chua*ab

We report an efficient solvothermal process to achieve deep deoxidation of octadecylamine functionalized

Received 11th June 2013
Accepted 22nd August 2013
DOI: 10.1039/c3tc31115e
www.rsc.org/MaterialsC
sub-stoichiometric graphene oxides (sub-GOx) in organic solvents. An initial average carbon oxidation state
of ca. 0.6 (i.e., 0.6 OH per basal-plane C) could be reduced to ca. 0.15, while retaining a critical density of the
alkyl chains for solvent processability, ca. 0.02 chains per basal-plane C. The products can thus be
characterized as single-sheet dispersions of alkyl-functionalized disordered “graphenes”. The oxidation
state was characterized by X-ray photoelectron spectroscopy, Fourier-transform infrared spectroscopy
and Raman spectroscopy. We found a strong solvent effect in the relative rates of alkyl chain degrafting
and/or chain scission versus deoxidation and/or regraphenization. The desired deoxidation and/or
regraphenization are promoted at higher temperatures and by the use of aprotic amide solvents.
Furthermore we found that carbon monoxide (CO) was a remarkably efficient fugitive deoxidizer. A key
advantage of this solvothermal deoxidation process is that the resultant dispersion of conductive
graphenes can be directly used for printing, coating or formulating into nanocomposites, without
further purification or heat treatment. Thin films of the deoxidized sub-GOx give dc conductivities of up
to 40 S cm
1, which is the highest reported to date for solvent-processable graphene derivatives. These
films show temperature independent conductivity, which suggests that its conductivity is largely limited
by tunneling between the alkyl chain spacers.

Introduction deoxidized to give the requisite percolating p-electron

The intractability of 2D graphene sheets poses a severe chal-
lenge for their potential use in numerous thin-lm or nano-
composite technologies. For applications that can tolerate some
disorder in graphene sheets, one way around this is to use
solvent-processable alkyl-functionalized graphene oxides
(GOx).1,2 The disorder in GOx sheets can be controlled by
choosing milder oxidative or quenching conditions to give sub-
stoichiometric graphene oxides (sub-GOx). The sub-GOx has
one-third to half of the carbon atoms oxidized to the oxygen-
bonded sp3 state with 10 nm diameter nanographene domains
embedded in an sp3 carbon network.3–5 These materials are
suitable for a variety of applications, such as broadband optical
absorption and optical-limiting.5 However for applications
requiring dc electrical conductivity, they still need to be further
a
Department of Chemistry, National University of Singapore, Lower Kent Ridge Road,
Singapore, S117543, Singapore. E-mail: chmcll@nus.edu.sg
b
Department of Physics, National University of Singapore, Lower Kent Ridge Road,
Singapore, S117542, Singapore
c
DSO National Laboratories, 20 Science Park Drive, Science Park 1, Singapore 118230,
Singapore
† These authors contributed equally to this article.
networks. One way to achieve this is through a post-deposition
heat treatment.3,4,6–9 Sub-GOx lms can be deoxidized10–12 at a
mild temperature of 150–300  C to give quality trap-free and
band-like transport of eld-induced charge carriers of both
signs.3 This suggests that the 2D system is inherently tolerant of
disorder. We use “deoxidation” to refer to what is in essence a
decomposition of the oxide to graphene and oxygen (and a
small amount of carbon dioxide). Sub-GOx can be deoxidized to
give better quality transport3,4,13 than heavily oxidized GOx,9,14–19
which typically gives variable-range-type hopping trans-
port.8,16,20–22 A high-resolution scanning tunneling microscopy
study has suggested that the deoxidation proceeds with some
regioregularity to generate the nal graphenite network.4
However, the required heat treatment temperature for deoxi-
dation is still too high for some applications, such as on plastic
foils or if an organic semiconductor or other temperature-sensi-
tive structures are present. In order to overcome this, the GOx can
be deoxidized while in the dispersed state in a solvent. The
chemical reduction of aqueous slurries of unfunctionalized GOx
and sub-GOx in hydrazine is well-known.17,19,23–26 These slurries
can also be reduced using other reductants, such as hydroqui-
none,27 sodium borohydride,8,20,28,29 or hydriodic acid.30 However,

7246 | J. Mater. Chem. C, 2013, 1, 7246–7254 This journal is ª The Royal Society of Chemistry 2013

Paper
the GOx sheets tend to restack as reduction progresses, causing an
Published on 22 August 2013. Downloaded by Université de Strasbourg, Service Commun de la Documentation on 10/02/2016 16:36:25.
undesired precipitation of partially reduced and disordered
aggregates. To circumvent this, dispersant aids, such as poly-
mers,23 pH modiers17,31 and surfactants,32 are oen used to
stabilize the suspension, which can then be deposited by inkjet
printing,33 or isolated by vacuum ltration.9,16,34 Recently, the
deoxidation of GOx slurries by heating in water has also been
reported,35 but it faces the same challenges. These reduced
materials give typically lms with dc electrical conductivity of the
order of 1–10 S cm
1.8,16,20 However, the granularity of the lms
and their contamination with dispersant aids present challenges
for some applications. Furthermore, these reduced GOx cannot be
transferred into organic solvents, which greatly limits their proc-
essability and applications. Therefore there is a need to nd a
route that produces stable dispersions of highly conductive
deoxidized GOx that can be used directly for printing, coating and
formulating into nanocomposites, without further purication or
heat treatment.
In this report, we demonstrate that octadecylamine func-
tionalized sub-GOx can be deeply deoxidized to a solvent-
processable and electrically conductive form, simply by heating
in a suitable organic solvent medium, without the use of
chemical reductants or dispersion stabilizers that leave side-
products. This solvothermal deoxidation process is remarkable,
because deoxidation can be promoted over the undesirable
alkyl chain degraing and/or scission, particularly by using
amide solvents, which may be further assisted by CO, a fugitive
deoxidizer. This agent is fugitive, because the product CO2 is a
gas. As a consequence, deoxidized sub-GOx can be produced
whilst retaining solvent-processability, which confers a unique
processing advantage central to numerous potential thin-lm
and nanocomposite applications. Furthermore, the sol-
vothermal deoxidation in organic media here is more effective
than those conducted in aqueous media.
Experimental section
Octadecylamine functionalized sub-GOx
Sub-GOx was prepared by a modied Staudenmaier oxidation. A
high-quality synthetic graphite (1.0 g; 496596, Sigma Aldrich)13
was added into a mixture of vigorously stirred concentrated
sulfuric acid (17.5 mL; 95–97%, Romil) and fuming nitric acid
(9.0 mL; 90%, Romil) and cooled to 0–5  C. Potassium chlorate
(11 g; >99%, Sigma-Aldrich) was then added slowly over 15 min
at 0–5  C, and the mixture was then warmed naturally to room
temperature where it remained for 7 days36 It is important to
maintain vigorous stirring to obtain homogenous oxidation.
The reaction mixture was then quenched with ice-water at
0–5  C, ltered through a 0.20 mm PTFE membrane, and puri-
ed by exhaustive washing with Millipore water until the pH is
neutral before vacuum drying at room temperature and storing
in the N2 glovebox.36 Finally, this puried sub-GOx was dried in
dynamic vacuum (10
2 mbar, 24 h), and transferred for storage
in the N2 glovebox (pO2, H2O < 1 ppm). We did not apply any
heating during the drying step to avoid any possible thermal
decomposition of the sub-GOx.37 Elemental analysis gave an
empirical formula: C2.00O0.82H0.80.
This journal is ª The Royal Society of Chemistry 2013
View Article Online
Journal of Materials Chemistry C
Octadecylamine (ODA) functionalized sub-GOx was then
prepared by adding sub-GOx (500 mg) to ODA (2.5 g; 97%,
Sigma-Aldrich) in 10.0 mL of dichlorobenzene (99.8%, Romil)
with intermittent sonication and continual stirring at 80  C for
4 h in a N2-purged glovebag. The functionalized material was
then precipitated with 50 mL of EtOH (>99.8%, HPLC grade,
Sigma-Aldrich), and then recovered by centrifugation at 3200g
(3100 rpm) for 30 min. This was redispersed in THF (1.0 mL;
>99.9%, HPLC grade, Sigma-Aldrich) and precipitated with
EtOH, collected using a centrifuge, and repeated until the
material was completely free of ODA. The functionalized sub-
GOx was then obtained by dispersing the solid into THF (50 mL)
with brief sonication followed by centrifugation at 3200g for 30
min to remove the insufficiently functionalized sub-GOx frac-
tion. The functionalized sub-GOx in the supernatant was then
precipitated with ethanol and then collected by centrifugation
at 5600g for 30 min, and vacuum dried to give a solid with an
empirical formula: C2O0.8H0.79–(C18H37NH)0.091.
Solvothermal deoxidation of octadecylamine functionalized
sub-GOx
Octadecylamine functionalized sub-GOx (0.2 mg mL
1) was
dispersed in N-methyl-2-pyrrolidone (NMP), N,N-dime-
thylformamide (DMF), N,N-dimethylacetamide (DMAc), 1,2-
dichlorobenzene (DCB), and 1,2,4-trichlorobenzene (TCB) in
7 mL vials with Teon-faced silicone rubber screw caps. These
vials were heated to 150  C with magnetic stirring, inside
isothermal wells drilled into an Al block and placed on a hot-
plate inside a N2 glovebox (pO2, H2O < 1 ppm). The deoxidized
materials were collected by precipitation with EtOH (6 mL per
mL of graphene dispersion), centrifugation at 5600g for 30 min
and vacuum dried to give the solid product.
Solid-state visible-infrared (Vis-IR) spectroscopy
These spectra were collected on samples dispersed into high
optical-clarity KBr pellets at known concentrations. In order to
collect over a wide energy and dynamic range, several tech-
niques were employed, and the spectra stitched together. To
make these pellets, 0.3 mg of the solid was accurately weighed
and dispersed into 200 mg of spectroscopic-grade KBr (Sigma
Aldrich) by grinding in a N2-purged glovebag to avoid moisture
absorption, and then compacted under vacuum to give optically
clear pellets at 10 bar pressure. The visible spectra were
collected using a modulation technique, in which the output of
a 50 W quartz-tungsten-halogen light source was mono-
chromatized (Newport CS260), modulated by an optical chopper
(130 Hz) and focused onto the KBr pellet sample. The trans-
mitted radiation was detected using a Si photodiode. The
photocurrent output was amplied using a transimpedance
amplier, demodulated using a lock-in amplier (SRS 830)
phase-locked to the optical chopper, and then read using a
source-measure unit (Keithley 2400) to give the usual trans-
mittance spectrum as a function of wavelength. The near-
infrared spectra were collected in the same way, but with an
InGaAs detector. The mid-infrared spectra were collected by the
usual Fourier-transform infrared (FTIR) technique using an
J. Mater. Chem. C, 2013, 1, 7246–7254 | 7247
 View Article Online

Journal of Materials Chemistry C Paper

LN2-cooled mercury cadmium tellurium detector in a N2-purged

Published on 22 August 2013. Downloaded by Université de Strasbourg, Service Commun de la Documentation on 10/02/2016 16:36:25.

Nicolet 8700-FT-IR spectrometer. The blank KBr pellet

produced under identical conditions gives an apparent absor-
bance < 0.2 (i.e., transmittance > 65%) over the entire spectral
range, which attests to the high optical clarity that has been
achieved in these compressed pellets.

Solution-state UV-vis spectroscopy

These spectra were collected using a deuterium tungsten
halogen light source (DH-2000, Ocean Optics), and a spectro-
graph (S1024DW, Ocean Optics) connected by optical bers
(P400-2-UV/VIS). The dispersions were placed in a 2.0 mm path
length quartz cuvette, with brief sonication to ensure homoge-
neous dispersion.

Solid-state Raman spectroscopy

Solid lms were drop-cast from THF onto oxygen-plasma-
cleaned Si(100) wafers and encapsulated using glass windows
with Paralm M spacers in the glovebox to avoid possible
photooxidation effects. The Raman spectra were collected using
a scanning confocal Raman microscope (Renishaw In-Via
Raman 2000) with 514 nm excitation wavelength at laser power
0.6 mW and 50 s exposure. No change in the spectra occurred
between repeat measurements at the same spot.
X-ray photoemission spectroscopy
C1s, O1s, N1s and Si2p core-level spectra were collected on ca.
10 nm thick lms that were drop-cast or spin-cast on H-termi-
nated Si(100) wafers. These wafers were rst prepared by stan-
dard buffered oxide etch of the native Si(100) wafers just before
lm deposition. Aer deposition, the lms were briey heated
to remove the solvent and quickly transferred to the ESCALAB
MKII spectrometer for measurements. The lms were excited
with MgKa1,2 X-ray photons (1253.6 eV) with the sample
grounded. The photo-emitted electrons normal to the lm were
analyzed at a constant pass energy of 20 eV for core-level spectra
to give an energy resolution of 0.7 eV. Thick lms and powdered
samples lead to sample charging that complicates analysis.
Electrical measurement
The functionalized graphene dispersion in TCB (1 mL; 0.3 mg
mL
1) was diluted with THF (9 mL) and spin-cast to give the
desired lm thicknesses on 300 nm thick SiO2/Si substrates. Au
electrodes (7 nm Cr/50 nm Au, 30–40 mm channel length,
1–1.2 mm channel width) were then thermally evaporated
through a shadow mask. Electrical conductivity was measured
using a semiconductor parameter analyzer (Keithley 4200-SCS
SCS) in the N2 glovebox.
Fig. 1 Time-dependent solution-state Vis-NIR spectra of solvothermal deoxi-
dation of alkyl-functionalized sub-GOx in DMF. Dispersion concentration, 0.2 mg
mL
1. Cell path length, 2 mm. Solvothermal deoxidation temperature, 150  C.
Inset: UV-Vis-NIR of pristine alkyl-functionalized sub-GOx. Dispersions of pristine
alkyl-functionalized sub-GOx and after 18 h reduction.
The spectrum of the pristine functionalized sub-GOx comprises
a broad electronic band that rises towards shorter wavelengths,
due to p–p* transitions in the nanographene domains.3,5 The
discrete transitions at the shorter wavelengths (230 nm and
295 nm) indicate that small isolated p-electron systems are
present. As time increases, the absorbance generally increases,
but more so over the longer wavelengths. This indicates deoxi-
dation of the carbon atoms in the basal plane, and an increase
in connectivity of the nanographene domains, similar to the
heat treatment of sub-GOx in the solid state.3,5 The spectral
evolution practically stops aer ca. 12 h.
The same experiment was repeated with DMAc, NMP, TCB
and DCB (for denition, please see the Experimental section),
all at 150  C. We chose these solvents because they are relatively
good media to disperse alkyl-functionalized sub-GOx and gra-
phenes,5 and the temperature as it is just below the boiling
point of most of these solvents. In all cases, no further change in
the absorption spectra was found aer 12 h, which indicates
that a limiting state is reached by this time. We found similar
spectral changes for DMAc and NMP, except that NMP produces
an even darker dispersion, but DMAc produces a lighter
dispersion. However both DCB and TCB produce light grey
dispersions with visible particles. Therefore there is a
pronounced solvent effect that inuences the course of sol-
vothermal deoxidation.

Estimation of oxidation state and alkyl chain fraction

Results and discussion
Solvothermal deoxidation of sub-GOx
Fig. 1(a) shows an example of the time evolution of the Vis-NIR
absorption spectra of the octadecylamine functionalized sub-
GOx dispersed in DMF and heated to 150  C in the N2 glovebox.
To quantify these changes, it is useful to estimate the change in
the average oxidation state (OS) of the interior carbon atoms on
the basal plane of these materials. This OS can naturally be
dened as the ratio of the number of sp3 carbon atoms (fourth
bond predominantly to oxygen) to the number of sp2 carbon

7248 | J. Mater. Chem. C, 2013, 1, 7246–7254 This journal is ª The Royal Society of Chemistry 2013
 View Article Online

Paper Journal of Materials Chemistry C

atoms in the plane, i.e., OS ¼ f(sp3 C) ¼ 1
f(sp2 C), where f basal plane, which are multiply bonded to oxygen atoms (C]O
denotes fraction. OS ¼ 0 for graphene and 1 for fully oxidized
Published on 22 August 2013. Downloaded by Université de Strasbourg, Service Commun de la Documentation on 10/02/2016 16:36:25.

and C(O)]O respectively), and also includes a variable contri-

graphene oxide. In principle, these two hybridization states can
be resolved through their chemical shis and quantied by C1s
core-level X-ray photoelectron spectroscopy (XPS), and the esti-
mated OS value then checked against the measured oxygen to
carbon ratio. However, despite the importance of knowledge of
the OS, this has previously been little investigated, because of
experimental difficulties (e.g., XPS charging shis, substrate
oxygen contamination), XPS chemical state broadening, and
uncertainties in the types of oxygen species present.
To overcome these challenges, we deposited ultrathin lms
of sub-GOx (ca. 10 nm thick) onto substantially oxygen-free
H-terminated Si(100) wafers. This avoids sample charging and
the attendant uncertainties in binding-energy corrections. We
devised the following two methods to estimate the OS and check
for self-consistency: (i) curve-tting of the XPS C1s spectral
envelop to obtain the fraction of oxidized carbon atoms in the
basal plane. (ii) Evaluation of the ratio of oxygen to basal-plane
carbon atoms. Even though the native SiO2 was etched away just
prior to lm deposition, small regrowth of oxide was sometimes
found, but this can be accurately subtracted through Si2p core-
level analysis. In addition, for the alkyl-functionalized sub-GOx,
the alkyl chain contribution has also to be taken into consid-
eration in the evaluation of the basal plane carbon intensity.
Fig. 2(a) and (b) show the C1s and N1s core-level spectra
respectively of these samples, together with pristine alkyl-
functionalized sub-GOx before and aer heat treatment at
150  C as references. Aer linear background subtraction, the
C1s core-level spectra were tted by sets of Gaussian compo-
nents: Ca, centered at 284.4–285.2 eV; Cb, at 286.0–287.0 eV; Cg,
288.1 eV, and Cd, 289.9 eV. Ca arises from carbon atoms that are
not bonded to oxygen atoms. These include both basal-plane
and alkyl chain carbon atoms.1,3,38,39 Cb arises from carbon
atoms on the basal plane that are bonded to oxygen
atoms.1,3,38,39 Cg and Cd arise from edge carbon atoms on the
bution from photoelectron shake-up effects. For sub-GOx, Cg is
typically 5–6%, and Cd even smaller, and so can be neglected.
Thus keto and carboxyl defects make only a small contribu-
tion,3,5 in contrast to heavily oxidized GOx.3,14,20,40 The sizeable
alkyl chain contribution however falls into Ca. To separate this
out, we independently evaluated the alkyl chain density from
calibrated FTIR measurements of the CH2 nas intensity in KBr
pellets, and subtracted its contribution from Ca to obtain Cb
only due to the basal plane carbon atoms.
The N1s spectra are typically weak due to small nitrogen
contribution from the alkylamine chains, and asymmetric with
tail towards higher binding energy. Because of low intensities,
no curve-tting was attempted. Instead the data were tted with
smoothed sum-of-Gaussian functions aer linear background
correction as guide to the eye. It is clear that the dominant
component lies at 399–400 eV, which is quite typical of alkyla-
mino nitrogens.38 The O1s spectra (not shown) show a peak
centered at 532.5 eV, which gives the total detected oxygen
intensities consistent with epoxy or hydroxyl oxygen.38
The results of the analyses are given by Cb/Cb and O/Cb for OS
in Table 1. The N/Cb and alkyl chain/Cb are also shown. The
pristine octadecylamine functionalized sub-GOx has an initial
OS z 0.57 (based on the mean of the two measurements). This
indicates that f(sp2 C) z 0. 43. The excellent agreement between
the two measures suggests that the oxygen is predominantly
OH. The pristine material has alkyl/Cb z 0.06, in good agree-
ment with CHNS analysis which gives alkyl/Cb z 0.05. Upon
heat treatment of thin lms to 150  C and 200  C, the OS
decreases to 0.3 and 0.2 respectively, which will provide a useful
reference. Upon heating the dispersed sub-GOx in solvents to
150  C, the OS also decreases to z0.3 for TCB, DCB and DMAc,
but further to z0.23 for DMF and 0.18 for NMP. These appear
to be more deeply deoxidized based on the C1s core-level shape

Table 1 Effectiveness of regraphenization

Oxidation state

Sample conditions Cb/Cb O/Cb N/Cb Alkyl/Cb

Reference
Graphene 0.00 0.0 0.002a —
Octadecylamine-sub-GOx 0.57 0.57 0.061 0.061b

Thermally deoxidized
Solid-state 200  C 90 min 0.20 0.14 0.025 NAc
Solid-state 150  C 12 h 0.30 0.33 0.04 NAc

Solvothermally deoxidized
TCB 150  C 12 h 0.27 0.30 0.017 0.006
DCB 150  C 12 h 0.26 0.32 0.017 0.004
DMAc 150  C 12 h 0.28 0.27 0.029 0.012
DMF 150  C 12 h 0.23 NAc 0.027 0.028
NMP 150  C 12 h 0.18 0.18 0.025 0.024
Fig. 2 X-ray photoelectron spectra of the deoxidized sub-GOx obtained in
TCB 200  C 90 min 0.35 0.35 0.033 0.030
various solvents. (a) C1s core-level spectra. (b) N1s core-level spectra. A pristine
TCB+CO 200  C 90 min 0.19 0.13 0.02 0.02
film of octadecylamine functionalized sub-GOx before and after heat treatment
a
at 150  C is shown for comparison. Noise at detection limit. b Pegged to N/Cb ratio. c Not available.

This journal is ª The Royal Society of Chemistry 2013 J. Mater. Chem. C, 2013, 1, 7246–7254 | 7249

Journal of Materials Chemistry C
than those obtained by typical reduction in aqueous
Published on 22 August 2013. Downloaded by Université de Strasbourg, Service Commun de la Documentation on 10/02/2016 16:36:25.
dispersions.41
Another feature that has emerged and is central to our work
is that alkyl chains are also eliminated with deoxidation.
However there is a strong solvent effect. When the functional-
ized sub-GOx was heated to 150  C in DCB or TCB, the N/Cb ratio
falls to 0.017 and alkyl/Cb to z0.005. Severe alkyl chain elimi-
nation and possibly scission occurs. However, when heated in
DMF or NMP, both ratios fall only to z0.025. Therefore the
alkyl chains have a higher survival probability in these solvents.
As a consequence, deeply deoxidized “graphene” single sheets
can be produced with the desired solvent processability.
p–p* absorption and vibrational spectroscopy
To characterize the products over a wide spectroscopic window,
we dispersed them into high optical-clarity KBr pellets and
collected their transmission spectra from the Vis to IR region.
Fig. 3(a) shows that the intensity of the absorption broadband
increases as: DCB z TCB z DMAc < DMF < NMP, which
correlates broadly with the OS. This absorption with shoulder at
ca. 0.5 eV (4000 cm
1) arises from the p–p* electronic transition
of the nanographene domains.
This trend is also supported by Raman spectroscopy of thin
lms cast onto Si(100) wafers. Fig. 3(b) indicates a more advanced
deoxidation in DMF and NMP than the other solvents. Deoxida-
tion of GOx causes a red-shi of the initial band at 1600 cm
1 to
1585 cm
1 and the emergence of a distinct defect band at 1620
Fig. 3 Spectroscopic characterization of the octadecylamine functionalized sub-GOx solvothermal deoxidized in various solvents. Heat treatment at 150  C 12 h (a)
Vis-IR spectra of pellets with 0.3 mg in 200 mg KBr. (b) Raman spectra of films. Excitation, 514 nm. Back-scattering configuration. (c) Expanded FTIR spectra of films in
the alkyl CH vibration region.
7250 | J. Mater. Chem. C, 2013, 1, 7246–7254
View Article Online
Paper
cm
1 together with intensication of the 1345 cm
1 band. This
evolution of the Raman spectra which conrmed that deoxidation
has taken place was also reported by other studies.8,26,42 Although
the Raman spectra may be inuenced also by edge structures of
the nanographenes, the results appear in line with increasing
degree of deoxidation across the samples.
In addition, narrow lines at 2800–3000 cm
1 due to the alkyl
chains (CH2 ns and na), and a broader band at 3400 cm
1 due to
OH and H2O, are also found in the FTIR spectra. The analysis of
these features reveals changes in the alkyl chain density. Fig. 3(c)
shows the expanded alkyl CH stretching vibration region. The CH2
ns and nas occur at 2920 and 2848 cm
1 respectively, while the CH3
ns and nas occur at ca. 2870 and 2955 cm
1.43,44 Upon heating to
150  C in various media, a loss in the CH2 band intensities occur.
The loss was more severe in chlorinated benzenes than in the
amide solvents. Furthermore, the CH3 band intensities increase
relative to the CH2 bands. This suggests that alkyl chain scission
also occurs on heating. Although the mechanism is not under-
stood, perhaps related to hydrogen-bond stabilization of the
amine or the slightly basic condition that arises from slight
solvent decomposition, the apparent stabilization of the alkyl-
amine chains in amide solvents enables deeply deoxidized sub-
GOx to remain solvent processable.
An effective fugitive deoxidizer for sub-GOx: carbon monoxide
Although DMF and DMAc are practically identical in terms of
solvent properties (dielectric constant, bond polarity and
This journal is ª The Royal Society of Chemistry 2013

Paper
Published on 22 August 2013. Downloaded by Université de Strasbourg, Service Commun de la Documentation on 10/02/2016 16:36:25.
Fig. 4 Effect of CO on deoxidation of octadecylamine functionalized sub-GOx in 1,2,4-trichlorobenzene, measured by absorbance of the dispersion at 532 nm
wavelength. (a) Time dependence with and without CO bubbled through the solvent at 1 bar. (b) Temperature–time–transformation plot in saturated CO, showing
temperature-dependent limiting behavior. Without CO, the absorbance after 12 h is ca. 0.4 at 200  C, 0.3 at 150  C, and 0.23 (unchanged) at room temperature. Cell
path length, 2.0 mm. Dispersion concentration, 0.2 mg mL
1.
Hansen solubility parameters), the greater efficiency of deoxi-
dation in DMF, and even more so in another closely related
amide solvent NMP raises the question whether an unexpected
agent is contributing to the chemistry. It is well known that
DMF decomposes slowly at its boiling point to CO and dime-
thylamine,45 which can be exploited as a nascent CO source for
chemical reactions.46–49 NMP also decomposes very slowly at
150  C, but to a complex mixture of CO, NOx and hydrocarbons.
This is conrmed under our solvothermal conditions by the
brown color that emerges (due to evolved NO2) in hot NMP. In
contrast DMAc is relatively more stable.45 The question arises
whether CO acts as a deoxidizer for sub-GOx.
To address this question, we introduced CO by bubbling at
1.0 atm into TCB which does not particularly promote deoxi-
dation of sub-GOx to check for acceleration. We conducted the
experiment at 150  C as before, monitoring the Vis-IR absorp-
tion intensity as a function of time. Fig. 4(a) shows a plot of the
absorbance of the dispersion at 532 nm as a function of time at
150  C, with or without 1 atm of CO. The results clearly show
that CO does promote deoxidation. Fig. 4(b) shows the plot of
absorbance for different times as a function of reaction
temperature. The data are then interpolated with isochronic
lines to construct an approximate temperature–time-trans-
formation diagram. It becomes clear that for each temperature
there is a limiting state characterized by no further increase in
absorbance with time. This is related to the situation discussed
for solvothermal deoxidation at 150  C in Fig. 1, and is thus a
generalization of this phenomenon over a wider temperature
range. We nd that the limiting state shows a step change at ca.
140  C where a large increase in absorbance occurs. Thus the
deoxidation proceeds strongly only beyond 150  C. Increasing
the temperature to 200  C sharply increases the rate of deoxi-
dation, but not the apparent nal extent.
XPS conrms that CO promotes the reduction of the OS of
the sub-GOx (Table 1). While solvothermal deoxidation in TCB
at 200  C for 90 min decreases the OS to 0.35, the introduction
of 1 atm of CO decreases this to 0.16, while retaining the
This journal is ª The Royal Society of Chemistry 2013
View Article Online
Journal of Materials Chemistry C
alkyl/Cb ratio at z0.02. This is even more deoxidized than what
was obtained with NMP, but potentially easier to control, and
without the evolution of corrosive NOx decomposition products.
Therefore we have established that CO is a surprisingly effective
deoxidizer for sub-GOx at elevated temperatures, and this
contributes to a mechanism for the observed solvent effect. This
nding is remarkable because CO was not previously known to
be able to deoxidize epoxides, except styrene oxide in the pres-
ence of the Au nanoparticle catalyst.50
Properties of the deoxidized sub-GOx
Next we characterize the dc conductivity of thin spin-cast lms
of the deeply deoxidized functionalized “graphene” obtained by
solvothermal deoxidation in TCB at 200  C for 90 min. Its
oxidation state is depicted in Table 1. No further purication
was necessary. The functionalized graphene dispersion in TCB
(1 mL; 0.3 mg mL
1) was diluted with THF (9 mL) and spin-cast
to give controlled thicknesses of at-lying sheets for in-plane dc
conductivity measurements. The lm thickness was determined
by the number of spin casts. The systematic change in the
interference color from colorless (5 nm) through dark-blue to
bluish-green (40 nm), as the lm thickness increases, shows
that the lm quality is high (Fig. 5(a), inset). The lms were then
evaporated with Au electrode arrays for measuring the dc elec-
trical conductivity in the N2 glovebox. The current–voltage
characteristics were linear, which indicates that the contacts
were Ohmic. No post-deposition heat treatment was necessary.
Fig. 5(a) shows the measured sheet resistance as a function
of lm thickness. The sheet resistance decreases from 40 kU
,
1 for the 5 nm thick lm to 6 kU ,
1 for the 40 nm thick
lm. This corresponds to the in-plane dc conductivity sdc
increasing from 0.3 S cm
1 for the 5 nm thick lm to 40 S cm
1
for the 40 nm thick lm. We believe that the rapid increase in
sdc with thickness in this ultrathin regime is related to the
development of percolation in the thicker lms, which has also
been observed in spray-coated lms. This is the highest
J. Mater. Chem. C, 2013, 1, 7246–7254 | 7251
 View Article Online

Journal of Materials Chemistry C Paper

Published on 22 August 2013. Downloaded by Université de Strasbourg, Service Commun de la Documentation on 10/02/2016 16:36:25.

Fig. 5 dc Electrical conductivity of ultrathin spin-cast film coatings of octadecylamine functionalized graphene. (a) Measured sheet resistance as a function of film
coating thickness. Inset: optical images of the coating in the channel between Au source and drain electrodes, and an atomic force microscopy image of a typical
deoxidized graphene single sheet. Channel length, 35 mm. Channel width, 1 mm. Deoxidation solvent, temperature and time are TCB, 200  C and 90 min. (b) A typical
temperature dependence of electrical conductivity in the Arrhenius representation. Channel length, 30 mm. Channel width, 1.2 mm.

reported conductivity to our knowledge based on deoxidized (1) Superior substrate adhesion. The alkyl-functionalized
single-sheet dispersions (not aggregates) that can be processed graphene materials adhere well to a variety of substrates,
from solvents. It is about one order of magnitude higher than including glass and plastic foils without any polymer binder or
the deoxidized sub-GOx obtained by solvothermal deoxidation adhesion promoter. This is shown clearly by the photographic
at 150  C in DMF (6 S cm
1) or NMP, and 2–3 times higher than results of the tape peel tests in Fig. 6. The top image shows that
that obtained by solid-state 200  C heat treatment (12 S cm
1). It a coating of unfunctionalized sub-GOx on polyethylene tere-
is also higher than the reduced GOx typically obtained by phthalate (PET) can be removed simply using a pressure-
chemical reduction in aqueous hydrazine hydrate or sodium sensitive adhesive tape. The middle and bottom images show
borohydride.8,16,19,20,41 Although the conductivity achieved is only
one-sixth of the c-axis conductivity (240 S cm
1) and one-twen-
tieth of the basal-plane conductivity (1000 S cm
1) of single
graphite crystals, it is already sufficient for a variety of appli-
cations in electromagnetic interference shielding, electrostatic
discharge protection, touch screens, voltage-supply electrodes
and fuel cells.
Fig. 5(b) shows a typical log sdc plot against inverse
temperature obtained from a deeply deoxidized alkyl-function-
alized sub-GOx. The apparent activation energy Ea for conduc-
tivity over 180–77 K is only 1.1 meV, rising to a few meV at room
temperature. Typically it is lower than a few meV. Since sdc is
given by the product of carrier density and mobility, and the
carrier density in sub-GOx is very weakly thermally activated,36
the low Ea here indicates that the mobility is also very weakly
thermally activated. Therefore bandlike transport is also ach-
ieved in these solvothermal-deoxidized materials, and their
ultimate conductivity is thus likely limited by tunneling across
the alkyl spacer (ca. 1.5 nm wide, cf. 0.34 nm between graphene
planes in graphite).

Excellent processability of the deoxidized sub-GOx

Finally we will demonstrate that the organic solvent processable
alkyl-functionalized sub-GOx as a class, including the alkyl-
functionalized graphenes, is vastly superior in terms of their
substrate adhesion and solvent processability than the usual
unfunctionalized GOx and their chemically reduced counter-
parts that are processed as aqueous slurries. Other advantages
have already been discussed previously.51–53
Fig. 6 Excellent adhesion of octadecylamine functionalized sub-GOx and gra-
phene film coatings on polyethylene terephthalate (PET) substrates. Left object in
each photograph is the film coating remaining on a 1.2  1.2 cm PET substrate,
after contacting the right half with pressure-sensitive adhesive tape and pull-off.
Right object is the adhesive tape in which the left half was previously contacted
with the film coating. Top: unfunctionalized sub-GOx. Middle: octadecylamine
functionalized sub-GOx. Bottom: octadecylamine functionalized graphene. All
spin-coated from tetrahydrofuran dispersion.

7252 | J. Mater. Chem. C, 2013, 1, 7246–7254 This journal is ª The Royal Society of Chemistry 2013
 View Article Online

Paper Journal of Materials Chemistry C

Published on 22 August 2013. Downloaded by Université de Strasbourg, Service Commun de la Documentation on 10/02/2016 16:36:25.

Fig. 7 Excellent processability of alkyl-functionalized sub-GOx. (a) Uniform jetting through print head nozzles. Inset: the photograph of 0.3 mg mL
1 octadecylamine
functionalized sub-GOx in chlorobenzene used for printing. (b) Printed arrays on hexamethyldisilazane-treated 300 nm thick SiO2/Si wafers. Inset: the zoom-in image of
a single printed droplet film, showing uniform interference color. (c) MicroRaman image of a printed droplet film, showing uniform G band intensity. (d) Printed arrays
on source–drain electrode patterns. Inset: zoom-in image of a single droplet film on an electrode pattern. (e) Printed lines on a commercial plastic foil, with increasing
graphene density from left to right. (f) Spray-coated film on an A4-size commercial plastic foil, shown here over a light background with “NUS ONDL” wordings to show
film uniformity.

respectively that the functionalized sub-GOx and solvothermal-
deoxidized graphenes here, on the other hand, pass the adhe-
sive-tape peel test. Therefore the alkyl-functionalized materials
provide superior adhesion to substrates. We attribute the
difference to the excellent sheet dispersion in the functional-
ized materials, but aggregation in the unfunctionalized
materials.
(2) Superior solvent processability. The alkyl-functional-
ized sub-GOx is also processable by inkjet printing from chlo-
robenzene and other solvent media. They do not clog the
nozzles of the print head. The experiments here were conducted
using a research-grade 16-jet shear-mode piezo-lm print head
(Fuji Dimatix DMP 2831) with a nozzle orice diameter of
20 mm.54 Stable jetting characteristics were obtained (Fig. 7(a)).
We have deposited arrays of sub-GOx lms (Fig. 7(b)) with
thicknesses of only 4–8 nm (i.e., 3–5 ODA-functionalized sub-
GOx monolayers), and good uniformity, as evidenced by the
uniform interference color. This is conrmed by microRaman
mapping (G band intensity) with micrometer resolution, which
reveals a thickness uniformity of 20% (Fig. 7(c)) and no “coffee-
stain” effect. The alkyl-functionalized sub-GOx can be readily
printed on a variety of substrates, including SiO2 with Au elec-
trode arrays (Fig. 7(d)) and on a commercial PET foil (Fig. 7(e)).
Finally, the alkyl-functionalized sub-GOx can be deposited by
spray coating onto commercial plastic foils (Fig. 7(f)). These
methods of processing are not readily available to aqueous
slurries due to dewetting and aggregation challenges. The alkyl-
functionalized graphene is also stable for a few hours. However,
if they precipitated out aer a long period, they can still be
easily redispersed back upon brief sonication.
In conclusion, alkyl-functionalized sub-GOx can be
dispersed in various organic solvents with high solution-proc-
essability. Their corresponding solvothermal-deoxidized forms
have slightly lower solution stability but they are still solution-
processable. We found a strong solvent effect in the relative
rates of alkyl chain degraing and/or chain scission versus
deoxidation and/or regraphenization in organic solvents.
Furthermore we found that carbon monoxide (CO) was a
remarkably efficient fugitive deoxidizer that can be generated in
situ during solvent decomposition or at situ by injecting CO gas.
Acknowledgements
This work was partially supported by the MOE ARF grant (R-143-
000-387-646), DSO National Laboratories grant (R-143-000-465-
592) and NRF-CRP grant (R-143-000-360-281). We thank Peter
Ho for detailed discussions.
References
1 S. Stankovich, R. D. Piner, S. T. Nguyen and R. S. Ruoff,
Carbon, 2006, 44, 3342.
2 S. Stankovich, D. A. Dikin, G. H. B. Dommett,
K. M. Kohlhaas, E. J. Zimney, E. A. Stach, R. D. Piner,
S. T. Nguyen and R. S. Ruoff, Nature, 2006, 442, 282.
3 S. Wang, P. J. Chia, L. L. Chua, L. H. Zhao, R. Q. Png,
S. Sivaramakrishnan, M. Zhou, R. G. S. Goh, R. H. Friend,
A. T. S. Wee and P. K. H. Ho, Adv. Mater., 2008, 20, 3440.
4 L. L. Chua, S. Wang, P. J. Chia, L. Chen, L. H. Zhao, W. Chen,
A. T. S. Wee and P. K. H. Ho, J. Chem. Phys., 2008, 129, 114702.

This journal is ª The Royal Society of Chemistry 2013 J. Mater. Chem. C, 2013, 1, 7246–7254 | 7253

Journal of Materials Chemistry C
5 G. K. Lim, Z. L. Chen, J. Clark, R. G. S. Goh, W. H. Ng,
Published on 22 August 2013. Downloaded by Université de Strasbourg, Service Commun de la Documentation on 10/02/2016 16:36:25.
H. W. Tan, R. H. Friend, P. K. H. Ho and L. L. Chua, Nat.
Photonics, 2011, 554–560.
6 I. Jung, D. A. Dikin, R. D. Piner and R. S. Ruoff, Nano Lett.,
2008, 8, 4283.
7 X. L. Li, G. Y. Zhang, X. D. Bai, S. M. Sun, X. R. Wang,
E. G. Wang and H. J. Dai, Nat. Nanotechnol., 2008, 3, 538.
8 W. Gao, L. B. Alemany, L. Ci and P. M. Ajayan, Nat. Chem.,
2009, 1, 403.
9 C. Mattevi, G. Eda, S. Agnoli, S. Miller, K. A. Mkhoyan,
O. Celik, D. Mastrogiovanni, G. Granozzi, E. Garfunkel and
M. Chhowalla, Adv. Funct. Mater., 2009, 19, 2577.
10 P. S. V. Jimenez, Mater. Res. Bull., 1987, 22, 601.
11 A. M. Rodrı́guez and R. V. Jiménez, Thermochim. Acta, 1984,
78, 113.
12 M. J. McAllister, J. L. Li, D. H. Adamson, H. C. Schniepp,
A. A. Abdala, J. Liu, M. Herrera-Alonso, D. L. Milius,
R. Car, R. K. Prud'homme and I. A. Aksay, Chem. Mater.,
2007, 19, 4396.
13 Z. L. Chen, F. Y. Kam, G. S. R. Goh, J. Song, G. K. Lim and
L. L. Chua, Chem. Mater., 2013, 25, 2944.
14 T. Szabo, O. Berkesi, P. Forgo, K. Josepovits, Y. Sanakis,
D. Petridis and I. Dekany, Chem. Mater., 2006, 18, 2740.
15 S. Stankovich, R. D. Piner, X. Chen, N. Wu, S. T. Nguyen and
R. S. Ruoff, J. Mater. Chem., 2006, 16, 155.
16 G. Eda, G. Fanchini and M. Chhowalla, Nat. Nanotechnol.,
2008, 3, 270.
17 D. Li, M. B. Müller, S. Gilje, R. B. Kaner and G. G. Wallace,
Nat. Nanotechnol., 2008, 3, 101.
18 Z. T. Luo, Y. Lu, L. A. Somers and A. T. Charlie Johnson,
J. Am. Chem. Soc., 2009, 131, 898.
19 S. J. Park, J. H. An, I. Jung, R. D. Piner, S. J. An, X. S. Li,
A. Velamakanni and R. S. Ruoff, Nano Lett., 2009, 9, 1593.
20 H. J. Shin, K. K. Kim, A. Benayad, S. M. Yoon, H. K. Park,
I. S. Jung, M. H. Jin, H. K. Jeong, J. M. Kim, J. Y. Chioi and
Y. H. Lee, Adv. Funct. Mater., 2009, 19, 1987.
21 A. Sinitskii, A. A. Fursina, D. V. Kosynkin,
A. L. Higginbotham, D. Natelson and J. M. Tour, Appl.
Phys. Lett., 2009, 95, 253108.
22 A. B. Kaiser, C. Gómez-Navarro, R. S. Sundaram,
M. Burghard and K. Kern, Nano Lett., 2009, 9, 1787.
23 S. Stankovich, D. A. Dikin, R. D. Piner, K. A. Kohlhaas,
A. Kleinhammes, Y. Jia, Y. Wu, S. T. Nguyen and
R. S. Ruoff, Carbon, 2007, 45, 1558.
24 C. Gómez-Navarro, R. T. Weitz, A. M. Bittner, M. Scolari,
A. Mews, M. Burghard and K. Kern, Nano Lett., 2007, 7, 3499.
25 S. Gilje, S. Han, M. Wang, K.-L. Wang and R. B. Kaner, Nano
Lett., 2007, 7, 3394.
26 H. L. Wang, J. T. Robinson, X. L. Li and H. J. Dai, J. Am. Chem.
Soc., 2009, 131, 9910.
27 A. B. Bourlinos, D. Gournis, D. Petridis, T. Szabó, A. Szeri and
I. Dékány, Langmuir, 2003, 19, 5984.
28 A. B. Bourlinos, D. Gournis, D. Petridis, T. Szabó, A. Szeri and
I. Dékány, Langmuir, 2003, 19, 6050.
29 Y. C. Si and E. T. Samulski, Nano Lett., 2008, 8, 1679.
30 I. K. Moon, J. H. Lee and H. Lee, Chem. Commun., 2011, 47,
9681.
7254 | J. Mater. Chem. C, 2013, 1, 7246–7254
View Article Online
Paper
31 X. B. Fan, W. C. Peng, Y. Li, X. Y. Li, S. L. Wang, G. L. Zhang
and F. B. Zhang, Adv. Mater., 2008, 20, 4490.
32 Y. Xu, H. Bai, G. Lu, C. Li and G. Shi, J. Am. Chem. Soc., 2008,
130, 5856.
33 V. Dua, S. P. Surwade, S. Ammu, S. R. Agnihotra, S. Jain,
K. E. Roberts, S. Park, R. S. Ruoff and S. K. Manohar,
Angew. Chem., Int. Ed., 2010, 49, 21154.
34 D. A. Dikin, S. Stankovich, E. J. Zimney, R. D. Piner,
G. H. B. Dommett, G. Evmenenko, S. T. Nguyen and
R. S. Ruoff, Nature, 2007, 448, 457.
35 Y. Zhou, Q. L. Bao, L. A. L. Tang, Y. L. Zhong and K. P. Loh,
Chem. Mater., 2009, 21, 2950.
36 S. Wang, P. J. Chia, L. L. Chua, L. H. Zhao, R. Q. Png,
S. Sivaramakrishnan, M. Zhou, G. G. S. Goh,
R. H. Friend, A. T. S. Wee and P. K. H. Ho, Adv. Mater.,
2008, 20, 3440.
37 M. A. Pimenta, G. Dresselhaus, M. S. Dresselhaus,
L. G. Cançado, A. Jorio and R. Saito, Phys. Chem. Chem.
Phys., 2007, 9, 1276.
38 G. Beamson and D. Briggs, High resolution XPS of organic
polymers, John Wiley & Sons Ltd, Chichester, 1992.
39 W. Gao, L. B. Alemany, L. Ci and P. M. Ajayan, Nat. Chem.,
2009, 1, 403.
40 S. J. Park, J. H. An, R. D. Piner, R. D. Piner, I. H. Jung,
D. X. Yang, A. Velamakanni, S. B. T. Nguyen and
R. S. Ruoff, Chem. Mater., 2008, 20, 6592.
41 S. Stankovich, D. A. Dikin, R. D. Piner, K. A. Kohlhass,
A. Kleinhammes, Y.-Y. Jia, Y. Wu, S. T. Nguyen and
R. S. Ruoff, Carbon, 2007, 45, 1558.
42 D. X. Yang, A. Velamakanni, G. Bozoklu, S. Park, M. Stoller,
R. D. Piner, S. Stankovich, I. Jung, D. A. Field, C. A. Ventrice
Jr and R. S. Ruoff, Carbon, 2009, 47, 145.
43 G. Zerbi, R. Magni, M. Gussoni, K. H. Moritz, A. Bigotto and
S. Dirlikov, J. Chem. Phys., 1981, 75, 3175.
44 C. Almirante, G. Minoni and G. Zerbi, J. Phys. Chem., 1986,
90, 852.
45 D. D. Perrin and W. L. F. Armarego, Purication of laboratory
chemicals, Permagon Press, New York, 3rd edn, 1988.
46 P. Serp, M. Hernandez, B. Richard and P. Kalck, Eur. J. Inorg.
Chem., 2001, 9, 2327.
47 Y. Wan, M. Alterman, M. Larheel and A. Hallberg, J. Org.
Chem., 2002, 67, 6232.
48 J. Muzart, Tetrahedron, 2009, 65, 8313.
49 W. L. F. Armarego and C. L. L. Chai, Purication of laboratory
chemicals, Elsevier/Butterworth-Heinemann, Oxford, 5th
edn, 2009.
50 Y. Mikami, A. Noujima, T. Mitsudome, T. Mizugaki,
K. Jisukawa and K. Keneda, Tetrahedron, 2010, 51,
5466.
51 S. J. Park and R. S. Ruoff, Nat. Nanotechnol., 2009, 4, 217.
52 K. P. Loh, Q. L. Bao, G. Eda and M. Chhowalla, Nat. Chem.,
2010, 2, 1015.
53 D. R. Dreyer, S. J. Park, C. W. Bielawski and R. S. Ruoff, Chem.
Soc. Rev., 2010, 39, 228.
54 L. Y. Wong, R. Q. Png, F. B. S. Silva, L. L. Chua,
D. V. M. Repaka, S. Chen, X. Y. Gao, L. Ke, S. J. Chua,
A. T. S. Wee and P. K. H. Ho, Langmuir, 2010, 26, 15494.
This journal is ª The Royal Society of Chemistry 2013