New carbons with controlled nanoporosity obtained by nanocasting using a SBA-15 mesoporous silica host matrix and different preparation routes

Article abstract of DOI:10.1016/j.jpcs.2003.10.008

An ordered mesoporous silica matrix (SBA-15) was infiltrated with carbon using either a liquid or a gas route with various carbon precursors (sucrose, petroleum pitch and propylene). The resulting SiO₂/C materials were then chemically treated to dissolve selectively the silica matrix and lead to ordered mesoporous carbon replicas. The influence of the precursors on the physico-chemical properties of the SiO₂/C composites and the carbon replicas were studied by X-ray diffraction, nitrogen adsorption measurements, scanning and transmission electron microscopies. It was shown that the content, the route for introducing the carbon into the host matrix, the microporosity and the mesopores order of the carbon replicas are determined by the choice of the precursors. The thermal evolution of the carbon replicas is also studied.

Full text of DOI:10.1016/j.jpcs.2003.10.008

Journal of Physics and Chemistry of Solids 65 (2004) 139–146
www.elsevier.com/locate/jpcs
New carbons with controlled nanoporosity obtained by nanocasting using
a SBA-15 mesoporous silica host matrix and different preparation routes
a,
J. Parmentier , S. Saadhallah , M. Reda , P. Gibot , M. Roux , L. Vidal ,
b
a,b
a
a
b
*
b
C. Vix-Guterl , J. Patarin
a
a
Laboratoire de Mat e´ riaux Min e´ raux, Ecole Nationale Sup e´ rieure de Chimie de Mulhouse, UMR CNRS 7016, Universit e´ de Haute Alsace,
rue Alfred Werner, Mulhouse Cedex 68093, France
Institut de Chimie des Surfaces et Interfaces–UPR CNRS 9069, 15 rue Jean Starcky, B.P. 2488-Mulhouse Cedex 68057, France
3
b
Abstract
An ordered mesoporous silica matrix (SBA-15) was infiltrated with carbon using either a liquid or a gas route with various carbon
precursors (sucrose, petroleum pitch and propylene). The resulting SiO /C materials were then chemically treated to dissolve selectively the
silica matrix and lead to ordered mesoporous carbon replicas. The influence of the precursors on the physico-chemical properties of the
SiO /C composites and the carbon replicas were studied by X-ray diffraction, nitrogen adsorption measurements, scanning and transmission
2
2
electron microscopies. It was shown that the content, the route for introducing the carbon into the host matrix, the microporosity and the
mesopores order of the carbon replicas are determined by the choice of the precursors. The thermal evolution of the carbon replicas is also
studied.
q 2003 Elsevier Ltd. All rights reserved.
Keywords: A. Microporous materials; A. Nanostructures; B. Vapour deposition; D. Microstructure; D. Surface properties
1
. Introduction
well-tailored pore morphology by the porous structure of the
host matrix. The synthesis of a very large panel of carbon
materials can then be forecast due to the broad variety of
available host matrices (lamellar, micro, meso and macro-
porous). Some applications of these templates are given
below.
Nanostuctured materials have attracted a large interest
for their various potential applications in numerous fields
such as energy storage, electrochemistry, adsorption,
catalysis, etc. This goal can be achieved if tailored materials
with controlled properties can be synthesized. In the
particular case of nanoporous carbon materials, micro and
mesopores are required to allow the adsorption of large
amounts and varieties of chemicals from gases to liquids
and the easiness of their diffusion. The classically used
activated carbons suffer from an insufficient control of the
pore size and arrangement during their preparation pro-
cesses [1]. Therefore, a new strategy has been proposed
based on a host–guest chemistry of porous materials. This
concept can briefly be described as follows: a porous
material (also named template or mould) is filled with
carbon by means of various routes (liquid or gas). The
resulting host/carbon materials are then chemically treated
to selectively remove the template. A negative carbon
replica of the porous structure of the mould is obtained with
The carbonization of organic polymer between the layer
of clays performed by the group of Kyotani and Tomita led
to the formation of unique carbon materials, like ultra-thin
graphite film [2]. The authors have noticed that a typical
non-graphitized carbon precursor such as polyfurfuryl
alcohol can be well graphitized by this method. It means
that the structure of the carbon material at the nanometer
scale can be influenced and controlled by the size and the
shape of the nanopores of the templates. Pillared clays with
a higher interlayer spacing (2 nm) were also be used by
Bandosz et al. [3].
Microporous templates like zeolites were also tested.
Zeolites of structure type FAU (zeolite Y) [4–8], LTL
(zeolite L) [4], BEA (zeolite beta) [4], MOR (zeolite
Mordenite) [7] and MFI (zeolite silicalite-1) [7] were filled
with carbon using various carbon precursors such as
acrylonitrile, furfuryl alcohol, organic resin or propylene.
Under specific condition, Ma et al. [6] were able to obtained
*
Corresponding author. Fax: þ33-3-8933-6885.
E-mail address: j.parmentier@univ-mulhouse.fr (J. Parmentier).
0022-3697/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jpcs.2003.10.008

140
J. Parmentier et al. / Journal of Physics and Chemistry of Solids 65 (2004) 139–146
pure microporous carbon replica with a surface area as high
2.2. The carbon infiltration processes
2
3
as 3600 m /g and with a micropore volume of 1.5 cm /g.
This concept was applied by Ryoo et al. [9] for the
synthesis of a carbon replica (named CMK-1) with the use
of an ordered mesoporous silica like MCM-48 as template.
Since, numerous carbon replicas have been prepared by
using different mesoporous templates (SBA-15, HMS, etc.)
and carbon precursors. Various carbon-filling techniques
were also tested such as the sucrose solution impregnation
followed by carbonization [9], or the pyrolitic decompo-
sition of propylene gas [10]. To our knowledge, the study of
the characteristics of the carbon replica obtained, for a given
mesoporous template, using different carbon precursors and
preparation route, has not been investigated.
The sucrose process is performed as indicated by Ryoo
et al. [9] and is not reported in details here. Briefly, a liquid
solution of sucrose was mixed with the SBA-15 silica at
room temperature in the presence of H SO (catalyst) before
2 4
to be dried and subsequently impregnated with another
sucrose solution. The mixture was then heat-treated in
vacuum at 1173 K for 3 h. These two liquid-impregnation
steps allow a carbon content of 36 wt% to be reached for the
SiO /C material.
2
The chemical vapor deposition (CVD) process is based
on the pyrolytic decomposition of propylene (2.5 vol% in
Argon) at 1023 K on the SBA-15 silica matrix as reported
elsewhere [10]. The duration of the deposition allows a
In the present work, several carbon materials were
synthesized by means of two different impregnation routes
and three different carbon precursors: the liquid impreg-
nation (with sucrose and pitch) and the gas phase
impregnation with propylene. Pitch is an excellent source
of graphitizable [11,12] carbon, both readily and cheap
since the calcined mesophase formed during the pitch
carbonization generates a pre-graphitic structure that can be
developed into a graphitic one by a high-temperature
treatment. All these infiltrations were performed on the
same batch of SBA-15 silica, an ordered mesoporous silica
with an one dimensional hexagonal arrangement of
mesopores interconnected with micropores. Therefore a 3
D-porosity is present in that host matrix that allows the
cohesion of the carbon replica. The present paper deals with
the comparison of different carbon-infiltrated silicas (SiO2/
C materials) and their corresponding carbon replicas. The
influence of the carbon content was also investigated in
order to get a better insight of the mechanisms of the carbon
filling within the host matrix regarding to the precursors
used. The resulting carbons were characterized using
different techniques and their thermal stability (in inert
atmosphere) was studied. Moreover, it is worth noting that
pitch petroleum was never used as liquid carbon precursor
for the infiltration of ordered mesoporous silica materials.
control of the carbon content in the final SiO
this study, values in the range 22–52 wt% of carbon into the
SiO /C have been obtained, compared to the maximal
/C material. In
2
2
theoretical value of about 63 wt% for the material SBA-15/
C used for this study.
For the Pitch process, petroleum pitch named ‘Ashland
A240’ was used as the carbon precursor. A calculated
amount of pitch (density 1.24) was mixed with the SBA-15
template in order to fill all its mesoporosity (pore volume
0
3
.95 cm /g). The pitch impregnation was performed by
stirring this mixture for 4 h at 575 K (temperature which
corresponds to the lowest viscosity of the pitch). The SBA-
1
5/Pitch mixture was cooled down before to be heat-treated
in argon up to 1175 K (heating rate: 3 K/min) for the
conversion of pitch to carbon. With a pitch having a carbon
yield of 45 wt%, a carbon content of about 36 wt% was
obtained for the SiO /C sample after one-impregnation. A
2
second impregnation was performed on the SiO /C sample
2
assuming that all the pitch all the pitch of the first infiltration
was introduced into the porosity. It allows reaching a carbon
content of 48 wt%.
All the SBA-15/C samples were then treated with
hydrofluoric acid to selectively leach out the silica matrices
and recover the carbon replicas [9]. The host silica SBA-15
template as well as the SBA-15/C and the C samples were
characterized by different techniques as described below.
2
. Experimental
2
.3. Characterizations
2
.1. Synthesis of the mesoporous silica template: SBA-15
The different materials (SiO moulds, SiO /C composites
2
2
and carbon replicas) were characterized by powder X-ray
diffraction (XRD) on a Philips diffractometer (Cu Ka1
radiation (l ¼ 0:15406 nm)). Nitrogen adsorption/desorp-
tion measurements were performed on a Micromeritics
ASAP 2010 at 77 K. Prior to the measurements, the samples
were outgazed at 473 K overnight. The specific surface area
was determined using the BET method in the relative
pressure 0.02–0.2 [13]. The pore volume was calculated by
considering the volume of nitrogen adsorbed at the value of
The synthesis of SBA-15 silica was performed using 1 g
of Pluronic P123 (BASF), 65 g of H O, 9.7 g of 12 M HCl
2
and 2.1 g of tetraethylorthosilicate (TEOS, Fluka, 98%).
The mixture of P123/water/hydrochloric acid was stirred
vigorously before the TEOS addition. This mixture was
aged at 313 K for 24 h followed by a second treatment at
3
50 K for 24 h. After filtration and washing, the powder is
dried at room temperature and heat-treated at 873 K in
air for 6 h to remove the organic template before using it
as a mould.
P=P of 0.95. The mesopore-size distribution was evaluated
0
by the Barett–Joyner–Halenda method from the desorption

J. Parmentier et al. / Journal of Physics and Chemistry of Solids 65 (2004) 139–146
141
Table 1
by the infiltration process X (X ¼ CVD, Sucrose or Pitch)
with a carbon content of y%. The carbon replicas are named
Porous characteristics of the SBA-15/C composites
CX-y%
.
Sample
Vp
S
Pore diameter
(nm,BJH) [14]
BET
2
(m /g)
3
cm /g)
(
SBA-15 (host matrix)
SBA-15/CCVD-52%
SBA-15/CCVD-36%
SBA-15/CCVD-22%
SBA-15/CSucrose-37%
SBA-15/CPitch 2–48%
SBA-15/CPitch 1-36%
0.95
0.04
0.31
0.46
0.20
0.10
0.06
800
25
5.5
3. Results and discussion
Non-porous
330
425
350
240
60
4.5
4.0
4.5
3.1. Comparison of the SiO /C composites
2
a
Micropores
As shown in Table 1 and as expected, the carbon-filled
4.4
mesoporous silica samples exhibit a lower pore volume and
surface area than the starting SBA-15 template. Never-
theless, differences can be observed depending on the
carbon-filling content and the infiltration process used.
The samples having a similar high-carbon content, close
to 50 wt% compared to the maximum value of 63 wt%,
were obtained using the CVD and Pitch processes (samples
SBA-15/CCVD-52% and SBA-15/CPitch 2–48%). They both
show a similar low pore volume, whereas a strong difference
in their specific surface areas can be noticed (25 compared
Vp, total pore volume; SBET, specific surface area.
a
Pore diameter below 2 nm.
branch of the hysteresis loop of the isotherm and using a
cylindrical pore model [14].
The High-Resolution Transmission Electron Micro-
scopic (HRTEM) images were performed on a TOPCON
EM002B (200 keV, Cs ¼ 0.4 mm) electron microscope.
The morphology of the initial and final materials was
observed on a scanning electron microscope (Philip
XL 30).
2
to 240 m /g, respectively), suggesting the presence of
external carbon in the latter case and/or the formation of
micropores within the carbon material.
The SiO /C composites containing around 36 wt% of
The materials obtained after the different processes and
treatments are referred in the text as SBA-15 for the silica
host matrix, SBA-15/CX-y% for the SiO /C material obtained
2
carbon (SBA-15/CCVD-36%, SBA-15/CSucrose-37%, and SBA-
15/CPitch 1–36%) are nearly half-filled. Therefore, a residual
2
Fig. 1. Nitrogen adsorption/desorption isotherms of the SBA-15/C composite samples.

142
J. Parmentier et al. / Journal of Physics and Chemistry of Solids 65 (2004) 139–146
porosity should be present. A simple calculation based on
the pore volume of the silica matrix and the density of
carbon and silica gives for the mixed samples, a theoretical
surface area and a higher micropore volume compared to the
CVD sample; these could be related to the formation of
microporoses by the release of volatile species (H O, CO,
2
3
pore volume close to 0.44 cm /g, assuming a perfectly
CO ,…) during the conversion of sucrose into carbon.
2
homogeneous filling with no pore closure. The isotherms of
these three samples represented in Fig. 1 show for all
of them, a large step in the range P=P 0.4–0.7 characteristic
The decrease in the carbon infiltrated content to a
level of 20 wt% leads to an isotherm (Fig. 1) with a
typical triangular hysteresis shape (type H2) which is
already visible for the CVD sample containing 36 wt% of
carbon. This behavior is characteristic of the desorption
of nitrogen from ‘bottle shape’ pores. This pore shape
can easily be explained by considering that with the
CVD process, although propylene penetrates all the
mesoporosity, its pyrolytic decomposition at the end
takes place preferentially at the entrance of the pores
(where the vicinity of the propylene flow allows its
renewal) but without pore closure.
0
of a large mesopore size distribution with pores of smaller
size than the starting SBA-15 sample (steep step in the range
P=P 0.70–0.75) as reported in Table 1. These features
0
indicate that, whatever the carbon infiltration process used
for this carbon content, all the pores are partially filled but
not in a homogeneous way (large mesopore size distri-
bution). Moreover, one can notice that at the difference of
the SBA-15/CCVD-36% sample, the SBA-15/CSucrose-37% and
SBA-15/CPitch 1–36% samples display a lower mesopore
volume. This observation suggests the presence of a higher
volume of closed mesoporosity for the latter samples,
especially for the one obtained with the pitch precursor. It is
worth noting that the sucrose sample exhibits here a higher
The different characteristics of the composite SBA-15/
carbon samples can be explained by the filling process.
With the CVD process, a homogeneous carbon filling
seems to be obtained, the pyrolytic decomposition of
Fig. 2. SEM micrographs of the samples SBA-15 template (a); CCVD-52% (b); CPitch 1–36% (c); CSucrose-36% (d).

J. Parmentier et al. / Journal of Physics and Chemistry of Solids 65 (2004) 139–146
143
Fig. 3. Powder X-ray patterns of the different carbon replicas (Cu Ka1 radiation).
propylene taking place through the entire sample with a
carbon deposit at the surface of the pores and at the end
preferentially at the aperture of the pores. Nevertheless,
a high carbon content can be reached without pore
closure at the difference of the liquid processes (Pitch
and Sucrose) where a different mechanism takes place.
The liquid precursors are in a first step introduced
through all the mesoporosity before to be turned into
carbon. Therefore, carbon is present through all the silica
material but probably with a lower homogeneity than the
CVD process. Indeed, with liquid precursors, a phenom-
enon of mesopore closure seems to be present that
hinders a complete and homogeneous carbon-filling,
promoting the formation of external carbon especially
for high carbon contents (e.g. Pitch process). These
features are verified by SEM for the carbon replicas as
described bellow.
3.2. Study of the SBA-15 carbon replicas
The SEM micrographs of the SBA-15 silica template
and its carbon replicas are shown in Fig. 2. One can
clearly see, that the particle morphology of the template
is conserved for the carbon replicas even when a high-
carbon content has been impregnated into the host matrix
by the CVD process (CCVD-52% sample). Nevertheless,
with the Pitch process, agglomerates with a less well-
defined morphology can be noticed when 36 wt% of
Fig. 4. TEM micrographs of the different carbon replicas: (a) CCVD-52%; (b) CSucrose-36% and (c) CPitch 2–48%
.

144
J. Parmentier et al. / Journal of Physics and Chemistry of Solids 65 (2004) 139–146
Table 2
Some characteristics of the SBA-15 carbon replicas
the MCM-48 carbon replica but to our knowledge, no new
model was proposed for the SBA-15 replica. Therefore, the
interpretations of the carbon isotherms (Fig. 5) are at the
present time still unclear. The characteristics of these
b
Sample
Vp
S
A (nm) hexagonal unit-cell parameter
BET
3
cm /g) (m /g)
2
(
isotherms (pore volume V and specific surface area ðSBETÞ)
p
C
C
C
C
C
C
CVD-52%
0.55
1.05
1.03
1.11
0.58
750
910
10.7
10.4
10.0
9.4
are reported in Table 2.
CVD-36%
The pore volume differs depending on the amount of total
carbon introduced into the silica matrix and the proportion
of external carbon. As mentioned above, with the CVD
process, a homogeneous carbon filling is obtained and for a
high carbon content the carbon replica (CCVD-52% sample) is
CVD-22%
890
Sucrose-36%
Pitch 1–36%
1470
920
10.3
9.4
Pitch 2–48% 0.36
530
Vp, total pore volume; SBET, specific surface area.
3
a
characterized by a pore volume close to 0.55 cm /g,
For the SBA-15 silica matrix, the hexagonal unit-cell parameter is equal
whereas when the amount of carbon infiltrated into the
host matrix decreases, more void remains leading to a
carbon replica with a higher pore volume.
to 11.0 nm.
carbon is introduced into the silica-host matrix (CPitch 1–
sample); this observation indicates the presence of
For the pitch process, the sample with a high carbon
content (CPitch 2 – 48%) exhibits a lower pore volume
compared to the CVD sample in relation to its high external
content of carbon as assumed above. Moreover, this type of
carbon seems to be poorly porous as suggested by the low
surface area measured. For this process, a decrease in the
carbon content allows a better control of the carbon filling
and therefore an increase in the pore volume and surface
area (see Table 2, sample CPitch 1–36%). For similar carbon
contents, the CVD and Sucrose processes lead probably to a
lower proportion of external carbon as proved by the higher
pore volume obtained. It is worth noting that with the
sucrose process, an additional microporosity is created
during the carbonization step by the release of volatile
species as shown by the high pore volume and surface area
of the carbon replica.
3
6%
external carbon as clearly observed for the CPitch 1–48%
sample (not reported).
Whatever the process used, when the carbon content is
higher than 20 wt% [15], one can see on the X-ray patterns
(
Fig. 3) that the carbon replicas exhibit a long range
ordering (hexagonal symmetry) as confirmed by the TEM
experiments (Fig. 4). The decrease in the cell parameters
(
Table 2) can be attributed to a contraction of the silica
template during the heat-treatment performed for the carbon
infiltration process. The higher temperature used for the
Pitch and Sucrose processes (1173 K) compared to the CVD
process (1023 K) explains the higher contraction observed
for the formers samples.
The pores of the carbon replicas have the shape of the
walls of the SBA-15 template. It means that classical gas
adsorption/desorption models considering cylindrical pores
cannot be applied for these materials. A new model based
on rod-aligned slitlike pore [16] was developed for
The thermal stability of these carbon replicas and their
ability to graphitize were studied by XRD. One can see
on the corresponding patterns recorded at low angles
Fig. 5. Nitrogen adsorption/desorption isotherms of the carbon replicas CCVD-36%, CPitch 1–36%, CSucrose-36%
.

J. Parmentier et al. / Journal of Physics and Chemistry of Solids 65 (2004) 139–146
145
Fig. 6. Powder XRD patterns of the carbon replicas heat-treated at 1675 K (Cu Ka1): (a) low angles; (b) high angles. For comparison, the XRD pattern of a pure
carbonized pitch sample is also represented ( p : impurities).
(
Fig. 6(a)) that the relative intensity of the more intense
the carbon replica (the pore has the shape of the walls of
the host matrix), the nature of the carbon precursor
influences also directly the characteristics of the porous
structure of the material to which it is converted. Indeed,
for a mesoporous template, each carbon precursor has its
own maximal carbon content per infiltration step; then
the resulting carbons differ as regards porosity (presence
of an additional microporosity with sucrose), compo-
sitions, nanometric scale organization (graphitization).
These model carbon materials, with various
controlled characteristics, could also contribute to
determine the optimal properties of some of their
potential applications.
correlation peaks have decreased and that the two distinct
peaks in the 2u range of 2–38 have disappeared after the
heat-treatment at 1675 K. It means that the heat-treatment
alters slightly the long-range order of the carbon replicas
as confirmed by the TEM observations which pointed out
the existence of microdomains having different orien-
tations within the particles (not reported). Nevertheless,
the latter technique has shown that the local order at the
mesoscopic scale is preserved. XRD patterns recorded at
2
u angles above 108 are characteristic of a poorly
organized carbon structure at the atomic scale (Fig. 6(b)).
The presence of a broad peak around 258 (2u), observed
only for the CCVD-52%-1675 K and CPitch-36%-1675 K
samples, can be ascribed to the turbostratic carbon
structure. This feature is consistent with the fact that
the carbon materials obtained form pitch carbonization or
pyrolitic decomposition of propylene are well known to
lead to graphitizable carbons contrary to the one prepared
by infiltration of sucrose. For comparison, the XRD
pattern of a pure carbonized pitch sample calcined at
Acknowledgements
The French Minister of Research and the CNRS are also
gratefully acknowledged for their financial supports through
the ‘Nanosciences-Nanotechnologies’ and ‘Materials’
programs.
1
173 K is also reported in Fig. 6(b). Although, a higher
carbonization temperature is used, it is worth noting that
the intensity of this peak for the CPitch-36%-1675 K sample
is lower than the one observed for the pure carbonized
pitch sample, suggesting that the pitch carbonization
taking place in a mesoporous confined media may be
lead to a carbon material with different characteristics.
References
[
[
1] T. Kyotani, Carbon 38 (2000) 269–286.
2] (a) T. Kyotani, N. Sonobe, A. Tomita, Nature 331 (1988) 331–333.
(
b) N. Sonobe, T. Kyotani, A. Tomita, Carbon 29 (1991) 61–67. (c) T.
Kyotani, T. Mori, A. Tomita, Chem. Mater. 6 (1994) 2138–2142.
3] (a) T. Bandosz, K. Putyera, J. Jagiello, J.A. Schwarz, Carbon 32
[
(
1994) 659–664. (b) T. Bandosz, J. Jagiello, K. Putyera, J.A.
4
. Conclusion
Schwarz, Langmuir 11 (1995) 3964–3969. (c) T. Bandosz, J. Jagiello,
K. Putyera, J.A. Schwarz, Chem. Mater. 8 (1996) 2023–2029.
[
[
[
4] S.A. Johnson, E.S. Brigham, P.J. Olliver, T.E. Mallouk, Chem. Mater.
The comparison of the SiO /C materials and the
2
1
1 (1997) 2448–2458.
corresponding carbon replicas has shown that the
mechanism of the carbon infiltration differs according
to the carbon precursor used. Although a given silica
template allows a control of the pore morphology of
5] J.R. Odriguez-Mirasol, T. Cordero, L.R. Radovic, J.J. Rodriguez,
Chem. Mater. 10 (1998) 550–558.
6] Z. Ma, T. Kyotani, A. Tomita, Carbon 40 (2002) 2367–2374.
[7] P. Enzel, T. Bein, Chem. Mater. 4 (1992) 819–824.

146
J. Parmentier et al. / Journal of Physics and Chemistry of Solids 65 (2004) 139–146
[
8] T. Kyotani, T. Nagai, S. Inoue, A. Tomita, Chem. Mater. 9 (1997)
09–615.
9] R. Ryoo, S.H. Joo, S. Jun, J. Phys. Chem. B 103 (1999) 7743.
[13] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938)
6
309–319.
[
[14] P. Barrett, L.G. Joyner, P. Halenda, J. Am. Chem. Soc. 73 (1951)
373–385.
[
[
[
10] J. Parmentier, J. Patarin, J. Dentzer, C. Vix-Guterl, Ceram. Int. 28
2002) 1–7.
11] I. Mochida, Y. Korai, C. Ku, F. Watanabe, Y. Sakai, Carbon 38 (2000)
05.
(
[15] F. Ehrburger-Dolle, I. Morfin, E. Geissler, F. Bley, F. Livet, C. Vix-
Guterl, S. Saadallah, J. Parmentier, M. Reda, J. Patarin, M. Iliescu,
J. Werckmann. Langmuir 19 (2003) 4303–4308.
[16] T. Ohkubo, J. Miyawaki, K. Kaneko, R. Ryoo, N.A. Seaton, J. Phys.
Chem. B 106 (2002) 6523–6528.
3
12] B. Rand, Design and Control of Structure of Advanced Carbon
Materials for Enhanced Performance, Kluwer, The Netherlands, 2001.