Adsorption and desorption characteristics of 3-dimensional networks of fused graphene

Article abstract of DOI:10.1016/j.susc.2011.08.016

Here we explore the exceptional structural characteristics of a set of graphene-related materials prepared by a wet chemical approach. We present a comprehensive study of the effects of morphology, sonication, temperature, probe species, and stacking behaviour on the measurement of graphene surface area. Nitrogen gas was used in the solid state gas adsorption measurements and methylene blue dye for adsorption measurements on aqueous dispersions of graphene. The surface area values obtained are among the highest reported for synthetic graphenes: 1700 m² g^{- 1} in aqueous dispersions and 612 m² g^{- 1} in the solid state. Microscopy revealed the graphene used in the study was present in large part as free sheets and electron diffraction confirmed the successful synthesis of high quality graphene with a regular C-C bond length of 1.41 ± 0.02 A?.

Full text of DOI:10.1016/j.susc.2011.08.016

Surface Science 606 (2012) 34–39
Contents lists available at SciVerse ScienceDirect
Surface Science
journal homepage: www.elsevier.com/ locate/susc
Adsorption and desorption characteristics of 3-dimensional networks of
fused graphene
a,
a
b
a,c
Mohammad Choucair ⁎, Nicholas M.K. Tse , Matthew R. Hill , John A. Stride
a
School of Chemistry, University of New South Wales, Sydney 2052, Australia
CSIRO Division of Materials Science and Engineering, Private Bag 33, Clayton South MDC, Victoria 3169, Australia
Bragg Institute, Australian Nuclear Science and Technology Organisation, PMB 1, Menai, NSW 2234, Australia
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 7 July 2011
Accepted 22 August 2011
Available online 28 August 2011
Here we explore the exceptional structural characteristics of a set of graphene-related materials prepared by
a wet chemical approach. We present a comprehensive study of the effects of morphology, sonication, tem-
perature, probe species, and stacking behaviour on the measurement of graphene surface area. Nitrogen gas
was used in the solid state gas adsorption measurements and methylene blue dye for adsorption measure-
ments on aqueous dispersions of graphene. The surface area values obtained are among the highest reported
Keywords:
Graphene
Nitrogen adsorption
Methylene blue
Defects
for synthetic graphenes: 1700 m2
g
−1
in aqueous dispersions and 612 m2
g
−1
in the solid state. Microscopy
revealed the graphene used in the study was present in large part as free sheets and electron diffraction con-
firmed the successful synthesis of high quality graphene with a regular C–C bond length of 1.41± 0.02 Å.
© 2011 Elsevier B.V. All rights reserved.
1
. Introduction
Since the isolation of free graphene sheets reported in 2004 by
graphene by nitrogen gas adsorption have indicated that values are
much lower than the theoretical maximum largely due to the ag-
glomeration of individual graphene sheets [12,13]. Our recent devel-
opment of a chemical synthesis pathway to gram-scale quantities of
graphene [16] has allowed for the thorough elucidation of the various
parameters that affect the measurement of surface areas of porous
fused graphene structures and free sheets, in the solid state and as
aqueous dispersions, respectively.
Novoselov et al. [1], it has developed into a major focus of study in
the science of carbon-based materials, due in part to its exceptional
properties at the atomic scale and the very nature of this most funda-
mental of materials [1–3]. Until recently, the piece-meal synthesis of
graphene has limited the study of the bulk physical properties, how-
ever various remarkable properties have been attributed to it in some
truly exceptional experiments performed on isolated sheets [3–6].
Graphene is a material which is essentially all surface. If both sides
of a graphene sheet are considered, then the effective surface area is
extremely large, 2630 m g . This has led many researchers to con-
sider it as a potential candidate material in applications benefiting
from extreme surface areas [7,8]. However, the functionality and per-
formance of emerging graphene-based technologies will be heavily
dependent on the ability to control the extent of surface availability
2
. Materials and methods
All of the graphene samples studied herein were obtained by the
solvothermal reaction of sodium and ethanol [16]. Reactions were
performed in either a 25 ml Teflon-lined Parr Instrument Company
2
−1
4
749 reactor or 125 ml Titanium Parr Instrument Company 4750 re-
actor. Typical synthesis consists of heating a 1:1 molar ratio of sodium
ca. 2 g) and alcohol precursor (e.g. 5 ml of ethanol) in a sealed reac-
(
[9–11]. Different synthesis methods yielding bulk quantities of gra-
tion vessel at 220 °C for 72 h to yield the solvothermal product. This
material is then rapidly pyrolysed in air and the resulting material pu-
rified in 1 g batches. Acid washed decolourising charcoal (BDH Che-
micals), and 98–99% natural graphite (Hopkins and Williams) were
used for comparative purposes. Impurity levels in all samples were
obtained by X-ray photoelectron spectroscopy (XPS), unless other-
wise stated. Characterisation measurements are described in the Sup-
porting Information.
phene ultimately converge on the same issue: identifying the most
appropriate post-synthesis fabrication method while retaining a
set of favoured physical properties necessary for an envisioned
technology.
To date, bulk physical properties have been investigated on
graphene-related materials [12], chemically modified graphene [13],
graphene oxide paper [14], and composites incorporating graphene
[15]. The reported Brunauer–Emmett–Teller (BET) surface areas for
2.1. Sample preparation
⁎
Corresponding author. Tel./fax: +61 2 93854672.
E-mail address: m.choucair@unsw.edu.au (M. Choucair).
For the summary of sample preparation, see Table 1.
0039-6028/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.susc.2011.08.016

M. Choucair et al. / Surface Science 606 (2012) 34–39
35
Table 1
Table 2
Summary of sample preparation methods.
Summary of the gas adsorption results and pore data obtained.
Sample Precursor Purification method
Solid state surface area analysis of the different graphene samples
Sample 1 Sample 2 Sample 3 Sample 4 Sample 5
11.3 477 153 21.7 484
1
1
Ethanol
Stirring in 100 ml of milliQ water for 96 h, then washed by
vigorous shaking in 300 ml of milliQ water 5 times with
vacuum filtration between each wash. The product was
collected and dried under vacuum for 24 h at 130 °C and
contains less than 7 wt.% impurities.
Sample 1 purified further by stirring in 100 ml of slightly
acidified ethanol for 24 h. The product was washed with an
ethanol–water mixture upon filtration, collected and dried
under vacuum for 1 h at 130 °C. The final sample contains less
than 1 wt. % impurities.
SABET (m2 g−1
Total pore volume
)
0.039 0.32 0.17 0.075 0.62
3
−1 a
(cm
g
)
Micropore volume
0.00052
0.1
0.17
397
598
0.048
100
195
0.0056
11.9
29.9
0.15
325
612
b
Ethanol
3
−1 b
(cm
g
)
Micropore surface area
2
−1 b
(
m
g
)
SALangmuir (m g−1
2
)
15.3
=0.99.
a
2
3
Ethanol
Ethanol
Stirring in 100 ml of milliQ water for 1 h, vacuum filtered, and
stirred in 100 ml of slightly acidified ethanol for 1 h. The
product was washed with an ethanol-water mixture upon
filtration, collected and dried under vacuum for 1 h at 130 °C.
Stirring in 100 ml of milliQ water for 24 h, vacuum filtered,
and stirred in 25 ml of slightly acidified ethanol for 24 h. The
product was washed with an ethanol–water mixture upon
filtration, collected and dried under vacuum for 24 h at
Calculated at P/P
Calculated using t-plot analysis.
o
b
size constraint, adsorption is most likely limited to monolayers.
Thus the Langmuir equation is more appropriate when calculating
surface areas of highly porous materials as it assumes monolayer ad-
sorption rather than the multilayer process taken into account by
the BET equation [17]. In this regard, Langmuir surface areas for
130 °C; the final sample contains less than 1 wt.% impurities.
4
5
Ethanol
Sample 2 sonicated for 6 h in ethanol, collected by rotary
evaporation of ethanol, and drying under vacuum for 24 h at
2
−1
2
−1
samples 2 and 5 were calculated to be 598 m g
and 612 m g
130 °C.
respectively.
Methanol Stirring in 100 ml of milliQ water for 1 h, vacuum filtered, and
stirred in 100 ml of slightly acidified ethanol for 1 h. The
product was washed with an ethanol–water mixture upon
filtration, collected and dried under vacuum for 1 h at 130 °C.
As the samples contained low levels of impurities of up to 7 wt.%
(NaOH, Na O, TiO , see Table S2 Supporting Information), the effect
2
2
of particle coverage on the measurement of surface area was consid-
ered. However, the contributions by impurities that affected the solid
state measurements were found to be far less than that of porosity.
This can be most easily seen with samples 2 and 5. Both have similar
pre-treatments and no impurities with similar elemental composition
and correspondingly similar BET surface areas, Table 2. Although,
when a large number of impurities were present on the surface, ca.
20wt.%, they were found to block adsorption sites that then prohibit
maximal gas loading. Removal of these impurities allowed the full
carbon surface area to be exposed, resulting in an adsorption surface
area more in accord with the true absolute area (Figure S3, Support-
ing Information).
The same samples analyzed by gas adsorption were also investi-
gated by MB in order to compare the surface areas obtained between
2
the two techniques. Surface areas observed by N adsorption were
consistently lower than those recorded in solution experiments
with MB, Fig. 3. This indicated that a smaller fraction of the sample
3
. Results and discussion
Contrary to initial expectations, the lowest BET surface area was
obtained from the sample believed to have the greatest quantity of
free sheets, sample 1, Table 2. The pore size distribution for this sam-
ple, (Figure S1 and Table S1, Supporting Information), indicates most
pores are in the 1–10 μm range, with a median pore size of 4.50 μm,
whilst the sample displayed a porosity of 91%. The difference between
3
−1
the total pore volume (BET, 0.039 cm g ) and the total intrusion
3
−1
volume (Hg, 4.69 cm g ) is due to the fact that the mercury ad-
sorption technique takes into account much larger pore sizes, includ-
ing voids due to inter-particle packing, in contrast to the BET method.
This result indicates that the sample is non-uniform, with a range of
length scales.
The apparent density of sample 1 as determined by mercury porosi-
2
surface was accessible under low partial pressures of N at 77 K,
−
3
metry (2.18 g cm ) (Table S1, Supporting Information) lies well within
−
3 1
the commonly observed values for graphite (2.09–2.23 g cm ), whilst
2
−1
the relatively low BET surface area of 11.3 m g confirms the absence
of significant mesoporosity (2–50 nm), suggesting that the graphene is
present as large discrete sheets, in accord with the Hg intrusion data
(Fig. 1).
The BET data for the sample that underwent only mild agitation, 3,
2
−1
showed more than an order of magnitude increase (153 m g
)
over sample 1. It is therefore essential to consider washing time
when measuring surface area. Furthermore, increased specific surface
areas were generally observed for samples having higher porosity,
with the methanol-sourced sample, 5, displaying the largest BET sur-
2
−1
face area at 484 m g , comparable to the maximum surface area
2
−1
obtained for an ethanol based graphene, 2, of 477 m g
.
The large hysteresis observed in the gas desorption branches sug-
gests the adsorbed gas molecules are partially trapped within the
highly porous materials, Fig. 2. Notably, the trapping effect was not
observed for samples exhibiting low BET surface areas (samples 1
and 4). The trapping results from the sample pores being of similar
size to the kinetic diameter of the probe gas, nitrogen. Given this
3.30 µm
Fig. 1. Scanning electron microscope image (SEM) of a large sheet of graphene in sam-
ple 1. The graphene sheet is almost fully transparent under the beam. Further SEM im-
ages can be found in Supporting Information Figures S2 and S3.
1
Graphite Mineral Data, viewed 16 December 2009, (http://webmineral.com/data/
Graphite.shtml).

36
M. Choucair et al. / Surface Science 606 (2012) 34–39
Desorption
Adsorption
Fig. 2. Typical nitrogen adsorption–desorption isotherms for synthesized graphene
samples at 77 K.
Fig. 4. Plot of surface area obtained by the MB technique as a function of sonication
time for sample 1b, with the increase in temperature of the sonication bath overlaid
to highlight the effect of temperature. The uncertainty values for the surface area
values obtained for sample 1b were typically less than 1% of the reported surface
areas at a 95% level of confidence, and were calculated according to Hibbert [24].
when compared to ambient conditions in solution. A corollary to this
is that there must also be many surfaces that are effectively not ex-
posed to gas adsorption, i.e. the space between graphene layers ap-
proaches the kinetic diameter of the gas probe species.
The degree of mechanical perturbation of the samples prior to
MB adsorption determines the extent of free sheets, existing as frag-
mentation products of the 3-D porous network. The diffusion kinetics
favour the adsorption of MB onto free sheets over the porous struc-
ture as the dye molecules need not permeate through open pores. A
greater quantity of free graphene sheets therefore corresponds to a
greater surface area, meaning that the surface area should increase
with sonication time, Fig. 4 (Supp. Info S4).
MB adsorption has been widely used as a measure of the surface
areas of graphitic materials. The accepted calibration being that
each milligram of adsorbed MB represents 2.45 m2 of specific surface
area [18]. The values for samples sonicated for 1 h are summarized
below together with the gas adsorption values, Fig. 3. The MB surface
2
−1
area values for natural graphite (9.70 m g
(
)
and charcoal
2
−1
521.14 m g ) tend to agree with literature values reported by
both the MB technique and gas adsorption (ca. 7–15 m g
2
−1
for
The sample containing the greatest number of free sheets corre-
2
−1
2
−1
graphite, and ca. 300–1000 m g
for charcoal), with the range in
sponds to the highest MB surface area, 1b (1700 m g ), a trend
continued with samples of decreasing surface area and increased
porosity, Fig. 3. The plots showing the effect of sonication time on
surface area for samples 1b and 3 (Figure S4 Supp. Info.) show a
maximum surface area at similar periods of sonication, before de-
creasing slightly with additional sonication. This apparent decrease
in surface area is explained by the temperature of the water in the
sonication bath over the sonication period of 6 h, reaching ca.
319 K after 2 h of sonication. This increase of more than 20 K over
the initial temperature is simply a result of the heating effects of
reported values largely due to sample pre-treatment [19–23]. The
values obtained for the graphene samples were much greater than
the other carbons investigated and the MB analysis of the samples
tend to complement the BET and Langmuir surface areas. Whilst the
2
−1
MB results for samples 2 and 5 (810 and 936 m g ) are similar
and in accord with the BET and Langmuir data, those samples having
undergone greater mechanical perturbation prior to analysis (sam-
2
−1
ples 1b and 3), showed much larger surface areas (1462 m g
2
−1
and 1240 m g
respectively).
Graphite
Charcoal
Langmuir
BET
Methylene Blue
Sample 5
Sample 2
Sample 3
Sample 1b
1
10
100
1000
10000
surface area / m2g-1
Fig. 3. Comparison of the surface areas obtained by gas (from BET and Langmuir
isotherms) and 1 h sonication MB adsorptions for different graphene samples. Also
shown are the MB values for charcoal and graphite. The uncertainty values upon repro-
ducibility for the MB results were typically less than 2% of the reported surface areas at
a 95% level of confidence with the exception of the graphite sample (U=27.94%), and
were calculated according to Hibbert [24].
Fig. 5. Plot showing the surface area values obtained for non-sonicated sample 1b at
constant temperatures of 17 °C (290 K), 23 °C (296 K), and sonicated samples of 1b
at a constant temperature of 17 °C (290 K). The uncertainty values for the MB results
were typically less than 5% of the reported surface areas at a 95% level of confidence,
with the exception of the surface area of 180 min non-sonicated sample at 17 °C,
(U=5.76%), and were calculated according to Hibbert [24].

M. Choucair et al. / Surface Science 606 (2012) 34–39
37
Fig. 6. Typical TEM images of (left) ethanol derived and (right) methanol derived samples. Scale bars (left) 500 nm and (right) 200 nm. Regions of relative opacity indicate folds or
overlap within the sheets. The ethanol samples appear as large sheets with crimping and layering evident whereas the methanol samples show a more rigid structure.
sonication, Fig. 4. These non-isothermal conditions mean that the
equilibrium dynamics of MB adsorption and desorption naturally
result in a maximum in the measured surface area.
during the first 2 h. which was more than double the average surface
2
−1
area obtained at 290 K over the same period, 632 m g . After
30 min. of sonication and at 290 K the sample required 50% more
sonication to approach the surface area observed at 296 K, ca.
In order to further investigate the effects of sonication and tem-
perature on the measured values, experiments were conducted on
sample 1b, under isothermal conditions at 290 K and 296 K without
sonication and at 290 K with sonication, Fig. 5. The optimisation of
temperature was found to be essential, as a significant difference in
the surface area was obtained with just a 6 K change in temperature.
2
−1
1182 m g
.
Whilst increased molecular mobility may lead to increased adsorp-
tion, at some temperature the increased mobility acts to desorb MB
molecules. This is due to an increase in collisions and the Boltzmann
factor, which favours increased desorption. When the sample was
maintained at a constant temperature of 290 K the surface areas
2
−1
An average surface area of 1305 m g
was measured at 296 K
300 nm
Fig. 7. AFM images (top) of a methanol derived sample, 5, and (bottom) an ethanol derived sample, 1b. Both show the tendency for the sheets to form layers.

38
M. Choucair et al. / Surface Science 606 (2012) 34–39
with and without sonication clearly display the effects of physical
agitation. The kinetics of the MB solution and the effects on the in-
teraction of the dye with the graphene play a major role in the de-
termination of the surface area by this method.
Agglomeration of graphene sheets is commonly referred to in liter-
ature to explain the apparent decrease in measured surface area of
bulk graphene from the theoretical value [12]. Layering, crimpling,
and folding is indeed observed in our samples under TEM, Fig. 6.
In addition, the porosity of the samples is believed to affect the
measured MB surface area. For example, the measured MB surface
prone to fragmentation. This is supported by the gas adsorption
data, Table 2.
AFM topographies of samples 5 and 1b showed porous sheets with
a tendency to layer one on top of another, Fig. 7. The ethanol derived
sample showed step heights of 5 Å and 10 Å indicating fragments of
single and bi-layered sheets. The methanol derived sample showed
much greater abilities to layer with a larger fragment of greater
than 25 Å in thickness having layers of ca. 5 Å. Single area electron
diffraction (SAED) patterns also showed the sheets form few layers
(Fig. 8). This layering is believed to be related to offset overlapped
π–π interactions caused by defects (Figure S5, Supporting Informa-
tion). The more defective the graphene, the greater the offset.
Hence, the reduced ability of the sheets to tend towards a 3-
dimensional graphitic structure, i.e. having a higher degree of disper-
sion. This in turn affects the measurable MB surface area.
area of the methanol-derived sample 5 was constant around ca.
− 1
9
51 m2
g
rather than displaying a peak value (Figure S4, Sup-
porting Information). This is believed to be due to sample 5 not
readily breaking up under sonication as the high number of small
pores act to increase the rigidity of the structure, rendering it less
10 mm
10 mm
Fig. 8. Single area electron diffraction (SAED) patterns of (left) an ethanol derived sample (focal length 900 mm, 10° offset), and (right) a methanol derived sample (focal length
010 mm, 2° offset). The spots are indicative of free graphene sheets, having a regular C–C bond length of 1.41± 0.02 Å layered on top of one another, with small angular offsets.
1
The off-set is determined from the angle between the two sets of spots. This is obtained by drawing hexagons connecting the points in each pattern (left), and calculating the angle
one hexagon would need to be rotated by to get to the other—with the maximum being 60° as this translates one point on the hexagon to an equivalent point. Lower plots are
√3
background-corrected plots of the respective samples, integrated along the (hk0) directions, having d-spacings of
n
/ .

M. Choucair et al. / Surface Science 606 (2012) 34–39
39
[
[
[
2] Y. Zhang, Y.W. Tan, H.L. Stormer, P. Kim, Nature 438 (2005) 201.
3] A.K. Geim, K.S. Novoselov, Nat. Mater. 6 (2007) 183.
4] X. Li, X. Wang, L. Zhang, S. Lee, H. Dai, Science 319 (2008) 1229.
4
. Conclusions
An integrated set of characterisation techniques have been utilized
[5] Y. Hernandez, V. Nicolosi, M. Lotya, F.M. Blighe, Z. Sun, S. De, et al., Nat. Nanotechnol.
(2008) 563.
[
[
[8] S. Patchkovskii, J.S. Tse, S.N. Yurchenko, L. Zhechkov, T. Heine, G. Seifert, Proc.
Natl. Acad. Sci. U. S. A. 102 (30) (2005) 10439.
[
3
to probe the structure of a novel series of graphene related materials. It
was shown that the experimental conditions required by the charac-
terisation technique including temperature, sample morphology,
washing time, and sonication strongly affected the surface accessible
to the probe species. Nitrogen adsorption was affected by the consol-
idation of the graphene layers at 77 K, whereas kinetic effects influ-
enced the outcome of MB adsorption. Nevertheless, the surface area
values obtained from these samples were among the highest reported
6] V.C. Tung, M.J. Allen, Y. Yang, R.B. Kaner, Nat. Nanotechnol. 4 (2008) 25.
7] L. Schlapbach, A. Züttel, Nature 414 (2006) 353.
9] S.-L. Chou, J.-Z. Wang, M. Choucair, H.-K. Liu, J.A. Stride, S.-X. Dou, Electrochem.
Commun. 12 (2) (2010) 303.
[10] H. Wang, Y. Yang, Y. Liang, J.T. Robinson, Y. Li, A. Jackson, et al., Nano Lett. 11
(2011) 2644–2647.
[
[
11] Z.-S. Wu, W. Ren, L. Xu, F. Li, H.-M. Cheng, ACS Nano 5 (2011) 5463–5471.
12] S. Stankovich, D.A. Dikin, R. Piner, K.A. Kolhaas, A. Kleinhammes, Y. Jia, et al., Car-
bon 45 (2007) 1558.
2
−1
2
−1
2
for synthetic graphenes (MB 1700 m g , N 612 m g ), all the
more remarkable given the facile synthetic procedure.
[13] H.C. Schniepp, J.L. Li, M.J. McAllister, H. Sai, M. Herrera-Alonso, D.H. Adamson,
et al., J. Phys. Chem. B 110 (2006) 8535.
[
14] D.A. Dikin, S. Stankovich, E.J. Zimney, R.D. Piner, G.H.B. Dommett, G. Evmenenko,
et al., Nature 448 (2007) 457.
Acknowledgements
[
15] S. Stankovich, D.A. Dikin, G.H.B. Dommett, K.M. Kohlhaas, E.J. Zimney, E.A. Stach,
et al., Nature 442 (2006) 282.
This research was supported by an Australian Postgraduate Award
to M.C. Microscopy facilities courtesy of the electron microscope units
at the University of New South Wales and the University of Sydney.
M.R.H acknowledges funding from the CSIRO Hierarchical Materials
Emerging Science Initiative.
[
[
16] M. Choucair, P. Thordarson, J.A. Stride, Nat. Nanotechnol. 4 (2009) 30.
17] S.J. Gregg, K.S.W. Sing, Adsorption, Surface Area and Porosity, Academic Press
London, London, 1982.
18] H.P. Boehm, A. Clauss, G.O. Fischer, U. Hofmann, Anorg. Allg. Chem. 316 (1962) 119.
19] C. Hontoria-Lucas, A.J. Lopez-Peinado, JdD Lopez-Gonzalez, M.L. Rojas-Cervantes,
R.M. Martin-Aranda, Carbon 43 (1995) 1585.
[
[
[
[
[
20] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, et al.,
Pure Appl. Chem. 57 (4) (1985) 603.
21] S. Wang, Z.H. Zhua, A. Coomes, F. Haghseresht, G.Q. Luc, J. Colloid Interface Sci.
Appendix A. Supplementary data
284 (2005) 440.
Supplementary data to this article can be found online at doi:10.
22] F. Marken, M.L. Gerrarda, I.M. Mellora, R.J. Mortimera, C.E. Maddena, S. Fletcera,
et al., Electrochem. Commun. 3 (2001) 177.
23] H. Hermanna, T. Schubertb, W. Grunera, N. Matterna, Nanostruct. Mater. 8 (1997) 215.
24] D.B. Hibbert, Analyst 131 (2006) 1273.
1
016/j.susc.2011.08.016.
[
[
References
[1] K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang, Y. Zhang, S.V. Dubonos, et al.,
Science 306 (2004) 666.