Onion-Like Graphene Carbon Nanospheres as Stable Catalysts for Carbon Monoxide and Methane Chlorination

Article abstract of DOI:10.1002/cctc.201500662

Thermal treatment induces a modification in the nanostructure of carbon nanospheres that generates ordered hemi-fullerene-type graphene shells arranged in a concentric onion-type structure. The catalytic reactivity of these structures is studied in comparison with that of the parent carbon material. The change in the surface reactivity induced by the transformation of the nanostructure, characterized by TEM, XRD, X-ray photoelectron spectroscopy (XPS), Raman, and porosity measurements, is investigated by multipulses of Cl₂ in inert gas or in the presence of CH₄ or CO. The strained C-C bonds (sp²-type) in the hemi-fullerene-type graphene shells induce unusually strong, but reversible, chemisorption of Cl₂ in molecular form. The active species in CH₄ and CO chlorination is probably in the radical-like form. Highly strained C-C bonds in the parent carbon materials react irreversibly with Cl₂, inhibiting further reaction with CO. In addition, the higher presence of sp³-type defect sites promotes the formation of HCl with deactivation of the reactive C-C sites. The nano-ordering of the hemi-fullerene-type graphene thus reduces the presence of defects and transforms strained C-C bonds, resulting in irreversible chemisorption of Cl₂ to catalytic sites able to perform selective chlorination. Tidy up the carbon! CO and CH₄ chlorination over hemi-fullerene-type graphene is described. The surface nano-ordering, induced by thermal treatment, transforms strained C-C bond sites resulting in irreversible Cl₂ chemisorption to catalytic sites that are able to selectively chlorinate CO and CH₄.

Full text of DOI:10.1002/cctc.201500662

DOI: 10.1002/cctc.201500662
Full Papers
Onion-Like Graphene Carbon Nanospheres as Stable
Catalysts for Carbon Monoxide and Methane Chlorination
Gabriele Centi,*[a] Katia Barbera,[a] Siglinda Perathoner,*[a] Navneet K. Gupta,[b]
Erika E. Ember,[b] and Johannes A. Lercher*[b]
Thermal treatment induces a modification in the nanostructure
of carbon nanospheres that generates ordered hemi-fullerene-
type graphene shells arranged in a concentric onion-type
structure. The catalytic reactivity of these structures is studied
in comparison with that of the parent carbon material. The
change in the surface reactivity induced by the transformation
of the nanostructure, characterized by TEM, XRD, X-ray photo-
electron spectroscopy (XPS), Raman, and porosity measure-
ments, is investigated by multipulses of Cl2 in inert gas or in
the presence of CH4 or CO. The strained CÀC bonds (sp2-type)
in the hemi-fullerene-type graphene shells induce unusually
strong, but reversible, chemisorption of Cl2 in molecular form.
The active species in CH4 and CO chlorination is probably in
the radical-like form. Highly strained CÀC bonds in the parent
carbon materials react irreversibly with Cl2, inhibiting further
reaction with CO. In addition, the higher presence of sp3-type
defect sites promotes the formation of HCl with deactivation
of the reactive CÀC sites. The nano-ordering of the hemi-fuller-
ene-type graphene thus reduces the presence of defects and
transforms strained CÀC bonds, resulting in irreversible chemi-
sorption of Cl2 to catalytic sites able to perform selective
chlorination.
Introduction
Carbon has been widely used as a support material for metal
nanoparticles, particularly for hydrogenation reactions,[1] but
recently the possibility to use carbon directly as a catalyst has
emerged as an interesting option.[2–4] The catalytic perfor-
mances of carbon materials depend on their nanoscale struc-
ture[5–7] as well as on the presence of functional groups, both
of these aspects are reliant on preparation, doping, and func-
tionalization.[2,3] Significant advances have been made recently
in this direction for nanocarbon materials such as carbon nano-
tubes (CNTs), graphene, fullerene, and other nanocarbon cata-
lysts (including fibers, coils, and horns, among others).[8,9] How-
ever, onion-type carbon materials, owing to their intrinsic
more complex structure and the presence of different types of
sites, remain less understood, although there is some relevant
progress in this area.[10] Critical aspects to be further analyzed
include the role of nano-ordering and the formation of fuller-
ene-type nanostructures. Thermal annealing allows an increase
in the degree of ordering (graphitization) and it may be possi-
ble to create fullerene-type nanostructures, depending on the
nature of the starting carbon.
Although research on the catalytic properties of nanocarbon
materials, such as CNTs and graphene, is increasing exponen-
tially,[11–16] the catalytic use of onion-type carbon materials is
largely unexplored. These materials may be visualized as hemi-
fullerene-type graphene shells arranged in a concentric onion-
type layered structure.[11] It is known that the strain of CÀC
(sp2-type) bonds may activate the reactivity owing to the cur-
vature present in these nanostructures, enabling coordination
and activation of reactants ranging from O2 to Cl2.[3,11] There-
fore, these structures have, in principle, interesting catalytic
properties, which may be tuned by nano-ordering and suitable
control of the parent amorphous carbon material, which in
contrast, does frequently not show such catalytic properties.[17]
In general, the unique catalytic features of nanocarbons are
derived from their atomic arrangement, that is, from their
linear, trigonal, planar, or tetrahedral geometries,[18] and the
presence of sp., sp2-, and sp3-hybridized carbon bonds.[19,20] In
addition, the presence of surface strains, point defects, dan-
gling bonds, and curvatures further enhances the catalytic per-
formance.[21,22] Onion-like structures combines several of these
features,[23–26] making it difficult to identify and characterize the
specific sites responsible for a particular catalytic behavior. It
has been suggested that the “surface roughness” influences
the reactivity of onion-like carbon materials, facilitating redox
reaction and desorption of reaction products, but more precise
indications have not been provided.[24,27,28]
[a] Prof. G. Centi, Dr. K. Barbera, Prof. S. Perathoner
Department of Industrial Chemistry
and Engineering of Materials and CASPE-INSTM
University of Messina
98166 Messina (Italy)
E-mail: centi@unime.it
perathoner@unime.it
[b] Dr. N. K. Gupta, Dr. E. E. Ember, Prof. J. A. Lercher
Technische Universität München
Department Chemie
Lichtenbergstr. 4, 85747 Garching (Germany)
E-mail: johannes.lercher@mytum.de
Supporting Information and ORCID(s) from the author(s) for this article
are available on the WWW under http://dx.doi.org/10.1002/
cctc.201500662.
This publication is part of a Special Issue on Carbon in the Catalysis
Community. Once the full issue has been assembled, a link to its Table of
Contents will appear here.
ChemCatChem 2015, 7, 3036 – 3046
3036
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
The presence of curvature and strains in CÀC bonds may ac-
tivate adsorbed reactants and decrease the activation energy,
but it is not known whether the activation leads to irreversible
chemisorption or is effective in promoting catalytic reactions.
It may be expected that the nano-ordering induced by thermal
treatment of carbon materials, which are precursors to surface
nano-onion-type structures, can thus provide a useful ap-
proach to understand this problem, and at the same time pro-
vide a way to create specific catalytic sites. In the presence of
disordered nanostructures, carbon is more reactive at the
edges, which have lower bond energies (ꢀ7 kcalmolÀ1), and
usually result in irreversible reactions with adsorbed reactants.
More ordered carbon nanostructures have a larger energy bar-
rier for chemisorption, but would allow, in principle, the trans-
formation of stoichiometric chemisorption sites into catalytic
sites.[2]
The selective synthesis of halogenated compounds from CH4
over carbon catalysts helps to understand the radical chemistry
of Cl2. It should be noted that there is great interest in the con-
version of methane to olefins and gasoline by mono-halogen
alkanes, as an alternative route to methanol conversion.[41–45]
Carbon materials offer an interesting alternative to metal
oxides, which may form volatile halogenides[46] and have prob-
lems with long-term stability related to the volatility of MeClx
species.
In previous work[47,48] we have shown a method to obtain
carbon nanospheres characterized by stacks of sp2-graphene-
like nanosheets starting from an amorphous carbon obtained
by the controlled pyrolysis of glucose. These materials showed
an onion-like carbon structure, that is, spherical closed carbon
graphene shells arranged in a concentric layered structure, re-
sembling that of an onion.[49–51] Following the acronym sug-
gested by Zhang et al.[52] we indicated these materials as
onion-like carbon nanospheres (OCNSs). The resulting material
shows high surface area and a structural order in the stacking
of the graphitic layers. These results evidence the possibility to
prepare different types of nanostructure sp2–sp3-mixed-carbon
systems, with different degrees of nano-ordering. The aim of
this work is to compare two carbon nanomaterials with an
analogous nanostructure, but a different degree of ordering, in
the reaction with Cl2, CO, and CH4. The combined experiments
help to understand how the local nanostructures influence the
surface reactivity.
It has to be remarked that the catalytic behavior of strained
CÀC (sp2-type) bonds is still largely unexplored. In nanocarbon
materials, such as graphene and CNTs, which have been exten-
sively studied from the catalytic perspective,[3] the reactive
sites are associated essentially to edges or defects. Fullerenes,
owing to the curvature of CÀC sp2-type bonds, are known to
readily react with various bi-atomic molecules, but in a stoichio-
metric way only.[11] The possibility of the utilization of these re-
active sites for the catalytic activation of small bi-atomic mole-
cules is thus highly challenging. Nano-onion-type carbon mate-
rials have a high presence of strained CÀC sites, and the
tuning of their characteristics is also possible. From a catalytic
perspective, therefore, these materials have the potential of
being of high interest, especially when combined with a low-
cost preparation method, as is described herein.
Results and Discussion
As a reaction to probe the surface chemistry, the catalytic ac-
tivation of Cl2 is used. This reaction is relevant to understand
the utilization of these catalysts in chlorination reactions, for
example, the synthesis of chloroform from methane or the syn-
thesis of phosgene from CO; two structure-dependent reac-
tions realized on large scale (about 5–6 Mtons/y).[29,30] Phos-
gene is produced industrially by catalytic chlorination of CO,
a strongly exothermic reaction.[31–33] The formation of byprod-
ucts (hydrochloric acid, and carbon tetrachloride[34]) depends
strongly on the type of activated carbon catalysts used in the
process. In fact, the extensive chlorination of the carbon mate-
rial leads to weakening of the carbon–carbon bond and the
formation of gas-phase chlorination products, with a concomi-
tant loss of catalyst material and a decrease of the available
active sites. The gas-phase chlorinated hydrocarbons can fur-
ther polymerize in the gas phase and form high-molecular by-
products, which readsorb onto the surface and deactivate the
catalyst by blocking the active sites. Therefore, the carbon
properties have to be optimized to minimize both HCl and
CCl4 formation, as well as irreversible chlorine adsorption,
while maintaining as high a rate as possible for phosgene syn-
thesis.[35,36] Although the reaction of graphene-type materials
with chlorine atoms has been previously reported,[37–40] results
on the uptake, rather than on the catalytic behavior, have not
been shown. In addition, the dependence of these properties
on the nanostructure is unclear.
Physicochemical characterization
TEM micrographs of CHT and CHT-tr samples (see definition in
the Experimental Section, Synthesis) are shown in Figure 1. It
should be noted that we have previously studied similar sam-
ples.[47] Although the samples reported here were prepared
again choosing optimized conditions, the results are similar,
confirming the reproducibility of the preparation. The hydro-
Figure 1. TEM images of CHT and CHT-tr catalysts at different magnifications.
ChemCatChem 2015, 7, 3036 – 3046
www.chemcatchem.org
3037
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
thermal procedure used to prepare the CHT sample gives rise
to a uniform distribution of spherically shaped carbon nano-
particles, with an average main diameter in the 200–400 nm
range. After thermal annealing, the spherical shape and mor-
phology is almost retained, although a different degree of or-
dering is shown by TEM characterization (Figure 1). CHT-tr
shows the presence of a graphitic-like structure, with a charac-
teristic nano-onion rearrangement, initiating at the surface and
propagating inside the nanoparticles, forming concentric shells
of C sp2-bonded atoms from the spherical or cylindrical shape
of the original amorphous precursor.
after annealing is surprising, because typically microporosity
decreases. One explanation could be the presence of hydrocar-
bon impurities in the carbon micropores, filling the micropores
and making them inaccessible during the determination of the
microporosity. The desorption of these adsorbed organic spe-
cies during the annealing procedure would thus lead to the
apparent increase of the microporosity. However, thermogravi-
metric (coupled with mass spectrometer analysis of the gas
evolution) and FTIR data on similar samples[47] also shows an
increase of the microporosity after annealing, excluding that
the increase of microporosity is related to desorption of the or-
ganic species present in the carbon micropores.
Therefore, we may conclude that the effect is related
Table 1. Crystallographic parameters (from XRD measurements) of CHT and CHT-tr
(see text for description).
to a structural reorganization, which is consistent
with the TEM and XRD results and previous literature
indications.[53–56]
Sample
d-002
La
Lc
L
GI
relative GI
Raman spectroscopy illustrates further the differen-
ces between the CHT and CHT-tr samples (Figure 2).
CHT shows a broad fluorescence background, which
could be due to the presence of organic impurities
[nm]
[nm]
[nm]
(LiF as reference)
(100% for CHT-tr)
CHT
CHT-tr
0.376
0.349
–
4.5
0.79
1.04
–
0.34
5.5
20.1
27.3
100
The crystallographic parameters determined by XRD patterns
for CHT and CHT-tr samples are reported in Table 1. The calcu-
lated d002 spacing for CHT-tr was 0.349 nm, suggesting the for-
mation of a quasiordered graphitic layer, although with larger
interlayer spacing (d002 for graphite is 0.335).This d002 value is
significantly larger in the CHT sample, owing to the presence
of a highly disorganized situation. In CHT the (110) reflection
is so weak, that it does not allow an estimation of La (basal
plane length), and as consequence, L (effective dimension of
graphitic microcrystallites, see the Supporting Information).
The Lc (height of layered stacking, see the Supporting Informa-
tion) value indicates the presence of some ordering along this
crystallographic direction, but not in the orthogonal direction,
as expected from a disordered organization, characteristic for
the precursor of an onion-type nanostructure. This indication is
in agreement with the TEM data (Figure 1).
Figure 2. Raman spectra of CHT and CHT-tr.
and/or surface defects. The thermogravimetric data
cited above indicates the absence of significant or-
ganic adsorbed species, but does not allow the pres-
ence of traces of adsorbed species, enough to give
the fluorescence effect, to be excluded. However, it is
likely that the fluorescence is related to the strongly
defective situation, as indicated by TEM data. It is
known that strong fluorescence originates in nano-
carbon materials from the electronic transitions
among/between the nonoxidized carbon regions and
the boundary of oxidized carbon atom regions,
where different functionalized groups CÀO, C=O, and
Table 2. Textural properties of the samples. Surface area determined by BET and Lang-
muir methods, area of the micropores, total pore volume, and volume of micropores.
Sample
SABET–SALang
Micropore
Total pore
Micropore
À1 [a]
[m2 gcat
]
area[b]
volume
[cm3 gcat
volume[b]
[m2 gcat
]
]
[cm3 gcat
]
À1
À1
À1
CHT
CHT-tr
635–840
585–770
503
550
0.31
0.26
0.23
0.25
[a] SA=Surface area. [b] determined by t-plot method.
The textural characteristics of the CHT and CHT-tr samples
have been determined by nitrogen sorption isotherms
(Table 2). The CHT sample has a very high surface area
(ꢀ640 m2 gcatÀ1) with a major contribution of the micropore
fraction (ꢀ500 m2 gcatÀ1). Upon annealing in N2 at 10008C for
6 h, there is a decrease both in surface area and pore volume
(À20 and À16%, respectively), with a corresponding increase
in the microporosity (ꢀ10%). This increase in microporosity
O=CÀOH are present.[57] This is in agreement with the indica-
tion by EDS that the amount of oxygen in CHT is approximate-
ly 10%.
The Raman spectrum of CHT-tr exhibits two main bands at
1586 and 1344 cmÀ1. These bands are also present in the CHT
sample, but are broader and weaker. The G (1580–1600 cmÀ1)
and the D (1350 cmÀ1) peaks are usually assigned to zone
center phonons of E2g symmetry and K-point phonons of A1g
ChemCatChem 2015, 7, 3036 – 3046
www.chemcatchem.org
3038
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
symmetry, respectively. They can be attributed to many forms
of sp2-bonded carbons with various degrees of graphitic order-
ing, ranging from microcrystalline graphite to amorphous
carbon.[5] The G peak involves the in-plane bond-stretching
motion of pairs of all sp2 sites (not necessarily only those ar-
ranged in six fold rings). However, the D peak is forbidden in
perfect graphite and becomes active in the presence of
a more disordered structure.[54–56] The intensity ratio between
these two bands (ID/IG) is usually indicative of the relative
degree of disorder.[53] In the case of CHT-tr, the ID/IG is equal to
0.77. According to the Tuinstra and Koenig equation:[58]
nealing.[54] The transformation during the annealing process is
confirmed by the analysis of the O1s region (Figure 3), present-
ing a main peak centered at 530.03 eV with a right shoulder
centered at approximately 533 eV, which disappears after ther-
mal annealing. This shoulder position is consistent with that
expected for ÀCOOH-type surface species, whereas that
at around 530 eV is typical for quinones and C=O species,
which are known to be thermally more stable than
ÀCOOH-type oxygen functional groups present in carbon ma-
terials.[5]
ID CðlÞ
ð1Þ
¼
Catalytic reactivity
IG
La
Reversible and irreversible Cl2 uptake by CHT and CHT-tr
a correlation between the ID/IG ratio and the La value obtained
from XRD measurements exists. In our case, the values lie
within the range where ID/IG is proportional to 1/La, analogous
to the case of graphite or nanocrystalline graphite with uni-
form crystallization (see S3 and Table S1, in the Supporting In-
formation).[53,54] It should be noted that some authors have
argued that the absolute intensity of the G band may be
a more suitable parameter than the D/G ratio.[59,60] However,
this estimation requires a more complex procedure, for exam-
ple, measuring the G-band intensity of liquid suspensions of
single-walled CNT (SWCNT) samples as a function of solid con-
centration.[60] This procedure is necessary to quantify the purity
of SWCNT, but is not necessary to have an indication of the
degree of disorder, as in our case.
The surface reactivity of CHT and CHT-tr towards Cl2 was stud-
ied by a sequential multipulse method at 2008C. HCl was the
only product observed in the reactor outlet stream, in addition
to Cl2. From the difference of Cl2 content in the inlet and
outlet streams, the amount of reacted Cl2 can be determined.
By subtracting the HCl formed, it is possible to determine the
amount of irreversibly chemisorbed Cl2. In addition, from the
difference in the peak shape of the single Cl2 peaks, with re-
spect to those observed when the catalyst is absent, an indica-
tion of the type of reversible chemisorption of Cl2 can be de-
duced.
Typical profiles recorded during Cl2 sequential pulses on
CHT and CHT-tr are shown in Figure 4a and 4b, respectively.
HCl is the major product formed initially, up to approximately
3000 s for CHT and approximately 1500 s for CHT-tr. Cl2 starts
to be detected when the formation of HCl begins to end. How-
ever, the formation of HCl does not correspond to the total
amount of converted Cl2. From the amount of Cl2 converted
and that of HCl detected in the gas phase, the total irreversible
Cl2 uptake is calculated (Table 3). The amount of irreversible
The presence of oxygen functionalities in CHT and CHT-tr
was analyzed by X-ray photoelectron spectroscopy (XPS) meas-
urements of C1s and O1s signals. The C1s spectra of CHT and
CHT-tr are shown in Figure 3. CHT shows a right shoulder in
Table 3. Cl2 multipulse experiments on CHT and CHT-tr at 2008C in con-
tinuous He flow. Experimental error in the Æ5–10% range.
Sample
Irreversible Cl2 uptake
HCl formation
À1
À1
[mmolgcat
]
[mmolgcat
]
CHT
CHT-tr
7.36
2.16
9.70
3.90
uptake of Cl2 by CHT is three times higher than the amount
calculated for CHT-tr. Similarly, HCl formation is ꢀ2–3 times
higher in CHT with respect to that in CHT-tr. After approxi-
mately 6000 s, the Cl2 pulse intensity at the reactor outlet
reaches an intensity almost corresponding to the inlet value. It
should be noted that the formation of HCl requires the pres-
ence of H-donor species on carbon sites near to those of chlor-
ine activation, as discussed in a more detail below.
Figure 3. XPS of C1s and O1s for CHT and CHT-tr.
comparison with the sharp and symmetric graphitic C1s peak
present in CHT-tr, probably because of the removal of function-
al CÀOH, C=O, and COOH groups[61] during the annealing pro-
cess. The main C1s peak shifts from approximately 283.8 eV in
CHT to 282.7 eV in CHT-tr, which is consistent with the change
from a prevalent sp3 to a sp2 configuration after thermal an-
The difference in peak shape between the inlet and outlet
Cl2 pulses, as a consequence of the interaction of Cl2 with the
catalyst, provides further interesting indications. For the CHT
ChemCatChem 2015, 7, 3036 – 3046
www.chemcatchem.org
3039
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
and their energy of interaction
with the catalyst. Both CHT and
CHT-tr show a very broad, struc-
tured desorption profile of HCl
between 200 and 8008C. The
total
amount
of
HCl
at
(8.2 mmolgÀ1)
desorbing
three different temperatures
(300, 380, and 5458C) from CHT
was approximately 14 times
higher than in the case of CHT-tr
(0.6 mmolgÀ1). For both cata-
lysts, no other chlorinated spe-
cies were detected during the
TPD experiments. Also, the CHT-
tr sample shows three desorp-
tion peaks, but with maxima
shifted to lower temperatures
(285, 325, and 4408C) and the
peak relative to the intermediate
temperature becoming more in-
tense. These results confirm that
thermal treatment not only
induces a strong decrease of
the sites responsible for HCl
chemisorption, but also a slight
change in their nature.
Figure 4. Cl2 multipulse experiments on a) CHT and b) CHT-tr in He continuous flow. c) Comparison of the reactor
outlet Cl2 pulses for CHT and CHT-tr. Experimental conditions: 30–50 Cl2 pulses (230 s pulse length) (3.6 mmol Cl2/
pulse), He flow: 10 mLminÀ1, 10 mg of material, T=2008C.
However, it could be suggest-
ed that the three peaks corre-
spond to three types of chemisorbed HCl species with different
energy of interaction, with the surface sites, or maybe a differ-
ent interpretation is possible. Part of the discussion, which will
be more extensive later, is that HCl is derived from the reaction
of Cl atoms produced by dissociative chemisorption of Cl2 and
H-donor sites, related to defective sp3-type C atoms. When
these sites are located near to a Cl atom coordinated to
carbon, desorption of HCl may occur. Otherwise, a surface mi-
gration of hydrogen atoms to the carbon atom coordinating Cl
could occur, and HCl may form and desorb. The three ob-
served peaks for HCl desorption in the TPD experiments after
the multipulse of Cl2 (Figure 5) may be, therefore, alternatively
interpreted as deriving from three different energy paths nec-
essary for surface diffusion of the hydrogen to reach the
carbon atom coordinating Cl. It should be also recalled that
neither on CHT nor CHT-tr samples is desorption of Cl2 detect-
ed during TPD experiments. Therefore, all chemisorbed (intact)
Cl2 desorbs reversibly at 2008C (even if the peak-shape analysis
discussed above evidences the presence of a strong chemi-
sorption on CHT-tr), whereas dissociative chemisorbed Cl2 re-
quires higher temperatures for desorption. The ability to ex-
plain why this occurs could lead to progressive catalyst deacti-
vation.
Figure 5. Temperature-programmed desorption of CHT and CHT-tr after ex-
posure to multiple pulses of Cl2 (6000 s, experimental conditions as in
Figure 4).
catalyst the difference is minor and the peak remains sharp
(Figure 4c), however, a greater difference is noted for CHT-tr,
which shows a long tail. This finding indicates an unusually
strong, but reversible, interaction of Cl2 with the more ordered
catalyst. Therefore, CHT-tr is characterized by a high amount of
reversible adsorption of Cl2, whereas CHT shows a high irrever-
sible uptake of Cl2, as well as higher formation of HCl.
The structural changes on the active sites of CHT-tr, caused
by Cl2 treatment, were monitored by in situ Raman spectrosco-
py. Raman spectra recorded during the interaction of Cl2 with
CHT-tr (Figure 6) show that after a short time of Cl2 exposure
(3 min) both D and G bands decreased in intensity, indicating
After equilibration with Cl2 at 2008C (e.g. after 6000 s), the
temperature of the catalyst was increased linearly [tempera-
ture-programmed desorption (TPD) experiments] to analyze
the type of surface species present on the catalyst (Figure 5)
ChemCatChem 2015, 7, 3036 – 3046
www.chemcatchem.org
3040
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Figure 6. In situ Raman spectra of CHT-tr in contact with Cl2 at 408C.
Figure 7. Cl2 multipulse tests at 2008C on a) CHT and b) CHT-tr in He/CO
(5 vol.% CO) continuous flow. Experimental conditions as in Figure 4.
a marked interaction of sp3- and sp2-hybridized carbons with
adsorbed Cl2. On increasing the time of exposure, the D band
does not change further, whereas the G-band intensity contin-
ued to decrease. The D-band intensity was not affected when
the Cl2 stream was replaced with N2, whereas the G band re-
covered part of the intensity, indicating the reversibility of the
adsorption processes on Cl2 in CHR-tr, according to the results
reported in Figure 4. As commented before, the D band is
a second-order Raman scattering process that is enhanced by
the presence of defects, whereas the G band involves the in-
plane bond-stretching motion of pairs of all sp2 sites. These
in situ Raman results thus indicate that defect sites (sp3-like) ir-
reversibly interact with Cl2, whereas CÀC bonds (sp2-type) can
give reversible chemisorption of Cl2. However, pure graphene
does not interact with Cl2 (except at the edges), neither is it
active in phosgene generation, and the highly strained CÀC
bonds in fullerene give only irreversible dissociative chemisorp-
tion.[62,63]
In fact, the ratio of the irreversible Cl2 adsorption of CHT with
respect to that of CHT-tr remained almost constant at, for ex-
ample, approximately threefold. We may conclude that the
concentration of irreversibly adsorbed Cl2 remains almost iden-
tical in the presence and absence of CO, indicating that the re-
action of the irreversible chlorine chemisorbed species with CO
does not occur.
Although COCl2 does not form on the CHT catalyst, a signifi-
cant amount of COCl2 is detected with the CHT-tr catalyst. For
CHT-tr, there is an initial delay in COCl2 formation, which starts
when HCl formation ends. On the contrary, much more HCl for-
mation is detected for CHT. It is worth noting that the forma-
tion of HCl passes through a maximum for both catalysts, sug-
gesting that the mechanism of formation of HCl involves first
the dissociative chemisorption of Cl2 and then the reaction of
the formed species with the H-donor species, thus generating
HCl. Note also that TPD experiments allowed the exclusion of
chemisorbed water or hydrocarbons over the catalysts surface
and, therefore, these H-donor species are those associated to
sites present on the catalyst.
Reactivity of the surface chlorine species
The reactivity of surface chlorine species in the two catalysts
was studied by multipulse Cl2 experiments in the presence of
a continuous flow of He–CO (5 vol% CO) or He–CH4 (5 vol%
CH4).
The conversion of Cl2 and the formation of COCl2 on the
CHT and CHT-tr catalysts and the relative graphitization index
(GI), that is, the graphitic degree in respect to the CHT-tr
sample (set equal to 100%), are summarized in Figure 8. GI
CO as a probe molecule
The reactivity of chemisorbed chlorine species towards CO was
investigated by sending pulses of Cl2 while continuously feed-
ing a mixture of 5 vol% CO in He over the catalyst. The results
for the CHT and CHT-tr catalysts are reported in Figure 7. The
presence of CO in the feed (Figure 7) modifies significantly the
behavior with respect to the similar tests, where instead an
inert gas only was fed together with Cl2 (Figure 4). Cl2 starts to
be detected in the reactor stream after approximately 2000 s
for CHT, but is not detected for CHT-tr, which is different to the
case without CO. The irreversible Cl2 uptake was estimated to
be 10.5 and 3.7 mmolgÀ1 for CHT and CHT-tr, respectively.
These values are slightly higher than those measured in the
absence of CO (Table 3), but the difference is within experi-
mental error, which can be estimated in the range of Æ5–10%.
Figure 8. Relationship between graphitization degree, Cl2 conversion, and
COCl2 formation on CHT and CHT-tr.
ChemCatChem 2015, 7, 3036 – 3046
www.chemcatchem.org
3041
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
characterizes the transition from sp3 to sp2 carbon and the for-
mation of the onion-like structure.
a unique carbonaceous species, a nanoscopically defined
hemi-fullerene-type structure. The combination of the layered
carbon structure and the specific bending of the hemisphere
create highly active sp2-hybridized carbon atoms, which are
proposed to be the site responsible for Cl2 reversible activa-
tion. Carbon-terminating defects (sp3 hybridized) are responsi-
ble for irreversible chemisorption, similar to highly strained
CÀC bonds. Raman in situ experiments (Figure 6) support this
conclusion, and tests with methane as a probe molecule sug-
gest a radical-like character of this reversibly chemisorbed (un-
dissociated) Cl2 species.
CH4 as a probe molecule
To explore the radical nature of the chemisorbed chlorine spe-
cies, their reactivity with CH4 was investigated over both
a CHT-tr sample and in an empty reactor, as reference. The re-
sults obtained are summarized in Figure 9. The reaction,
The results reported in Figure 4 and 7 show that CHT and
CHT-tr differ in terms of reversible and irreversible uptake of
chlorine species on the surface. Cl2 uptake was mainly irreversi-
ble in CHT, whereas CHT-tr shows less irreversible chemisorp-
tion (as well as HCl formation) and the presence of a reversible
chemisoption, although with strong interactions with carbon
surface sites. Based on previous comments, Raman and other
characterization results, it is reasonable to attribute the sites
able to reversibly activate Cl2 to moderately strained CÀC
bonds (sp2-type). Nano-ordering during thermal treatment in-
duces the conversion of sites with irreversible (dissociative)
chemisorption of Cl2 to sites that chemisorb Cl2 strongly, but in
a molecular (reversible) form. Dissociatively chemisorbed Cl2
may then react with surface H-donor sites to generate HCl. The
formation of the HCl, therefore, requires the presence of H-
donor sites. TPD data excludes the presence of possible hydro-
gen donor species, such as water or hydrocarbons adsorbed
on the carbon surface, therefore, these H-donor species are as-
sociated to hydrogen atoms saturating the carbon defects, or
are present as ÀOH or ÀCOOH surface groups.
Figure 9. Total amount of CH3Cl, CH2Cl2, CHCl3, and CCl4 formed in 6000 s
during Cl2 multipulse experiments at 2008C in an empty reactor and in the
presence of CHT-tr under continuous He/CH4 (5 vol% CH4) flow. Experimen-
tal conditions as in Figure 4.
Catalyst characterization data show that these H-donor sites
may be associated to sp3 carbons at defect sites (carbon-termi-
nating sites). When these H-donor sites are no longer available,
CO reacts with molecular chemisorbed Cl2, leading to COCl2.
However, although the HCl maximum formation in CHT is de-
tected at approximately 1000 s, COCl2 formation is not detect-
ed over this catalyst, at least up to 6000 s. This finding sug-
gests that desorption of HCl does not allow the restoration of
the original active site and thus surface reactivity toward Cl2
decreases progressively (Figure 7a). Reasonably, this is due to
the fact that the other Cl atoms derived from dissociative
chemisorption of Cl2 remain chemisorbed on the catalyst,
blocking the active sites. Only at higher temperatures does the
mobility of hydrogen on the surface allow the complete de-
sorption of HCl, as shown from the TPD results (Figure 5). Data
in Figure 4c shows that sites that strongly chemisorb Cl2, al-
though still in a reversible molecular form, are present in CHT-
tr, but not in CHT.
known to occur with radical-type chlorine species,[64] was con-
ducted at 2008C pulsing Cl2 in a continuous flow of He and
CH4 (5 vol%), and under dark conditions (light may promote
the halogenation). Halogenated methane species, CH3Cl,
CH2Cl2, CHCl3, and CCl4, are the products of the reaction, with
a distribution consistent with the presence of radical-type
chlorine species on the surface of the catalyst. These products
were also observed without the catalyst (in the empty reactor),
albeit in minor amounts, and were considered in the evalua-
tion of the catalyzed reaction. The fact that the distribution of
the products, however, does not significantly change in the
tests with and without the catalyst is in accordance with the
radical-type mechanism and proves that the onion-type cata-
lyst was active in producing chlorine-radicals, which reacted
with methane to give chloromethanes with a distribution in
agreement with that expected for radical-type reactions.
These observations may be rationalized by considering that
different types of CÀC (sp2-type) bonds are present in the cata-
lysts:
1) Highly strained CÀC bonds, derived from disordered nano-
onion-type structures that can irreversibly react with Cl2,
giving rise to homolytic dissociation of the ClÀCl bond;
these sites are similar to those present in fullerene-type
Nature of the reactive sites
Overall, spectroscopic, structural characterization, and kinetic
measurements show that the thermal treatment creates
ChemCatChem 2015, 7, 3036 – 3046
www.chemcatchem.org
3042
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
carbon materials and have a high bending (pyramidaliza-
tion) degree and a similar irreversible uptake of Cl2.[11,65–67]
2) Strained CÀC bonds, derived from the nano-ordering pro-
cess, are able to strongly chemisorb Cl2, but in a molecular
reversible form; these sites are responsible for COCl2 forma-
tion, in presence of CO, and also promote CH4 chlorination.
3) CÀC bonds with moderate to weak strain do not interact,
or have a weak interaction, with Cl2.
Therefore, it is reasonable that HCl forms when H-donor
sites, associated to defective sp3-carbon sites, are present near
to the first two types of sites. In such configurations, the other
Cl atom (derived from Cl2 dissociative chemisorption) remains
strongly chemisorbed, but not in a position near to a H-donor
site, thus resulting in the inhibition of that site. Nano-ordering
that occurs during the thermal treatment leads to the elimina-
tion of these defect sites, in agreement with the significantly
lower formation of HCl in CHT-tr compared to that in CHT. The
larger formation of COCl2 in CHT-tr originates from both the
modification of the characteristics of CÀC bond (sp2-type) and
the elimination of carbon defects (sp3-type).
Figure 10. Relationship between the graphitization index (GI) of carbon ma-
terials and the catalytic Cl2 activation.
This indication is further supported from the data reported
in Figure 8, in which Cl2 conversion and COCl2 formation on
the two catalysts are compared with respect to the relative
graphitization index. This index characterizes the transition
from sp3 to sp2 carbon and thus the formation of the onion-
like structure. This result gives further support to the above in-
terpretation. However, further experiments are necessary for
a better understanding of the unusually high strength of the
interaction between Cl2 and CHT-tr and the exact nature of
these sites in the onion-type nanostructure. The relationship
between HCl formation and the concentration/nature of defect
sites in carbon also requires further clarification.
The irreversibly chemisorbed Cl2 species, owing to their
strong interaction, are not reactive towards CO, and are not in
equilibrium with the reversibly chemisorbed species. This find-
ing is consistent with the hypothesis that irreversibly adsorbed
species are those that are homolytically dissociated (e.g. ClCÀ
CCl), whereas those that are reversibly chemisorbed are molec-
ular species. Thermal treatment, by inducing nano-ordering,
minimizes the concentration of defects (sp3-type carbon sites)
and the amount of strongly bent CÀC bonds (sp2-type), limit-
ing irreversible Cl2 chemisorption and HCl formation. To further
support this indication, Figure 10 reports the inverse correla-
tion between the graphitization index (GI) of the carbon
atoms, the total amount of HCl formed, and the concentration
of irreversibly chemisorbed Cl2. Note that CHT-tr contains only
one fourth of the sp3 carbon atoms compared to the parent
CHT. This is paralleled by an approximately four times lower Cl2
uptake and HCl formation.
Scheme 1. Proposed COCl2 synthesis mechanism on CHT-tr carbon materials.
the discussed experimental results. During Cl2 interaction with
sp2-carbon sites, labile charge-transfer complexes [*C···Cl2] are
formed. Raman in situ spectra confirm this type of interaction,
and reactivity data shows that when there is a relatively strong
interaction, although reversible and involving undissociated Cl2
molecules, a subsequent reaction with CO to form phosgene
may occur. Following the subsequent charge transfer from
·
carbon material to Cl2, a new chlorine radical [*CÀCl2 ] species
is generated that reacts with CO, forming COCl2 via an acyl-
chloride radical [·COCl]. The presence of such a radical-like
chlorine chemisorbed species is confirmed by the tests with
CH4 as probe molecule. In fact, the product distribution of
methylchlorides is in agreement with the presence of radical-
like species at the active carbon surface.
Reaction mechanism
The tentative catalytic cycle for COCl2 formation, which de-
scribes the experimental evidence discussed above, is present-
ed in Scheme 1. Although in part speculative and deserving of
further investigations, this proposal may be useful to analyze
The key in the reaction mechanism is the type of carbon
sites responsible for the first step of labile charge-transfer com-
plex formation. According to the discussion in the previous
section, CÀC sp2-type bonds are not active enough for this ac-
ChemCatChem 2015, 7, 3036 – 3046
www.chemcatchem.org
3043
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
tivation of Cl2. Bending of the bond (strained CÀC bonds) re-
duces the activation energy, allowing coordination and rela-
tively strong chemisorption, although still in a reversible mo-
lecular form [Eq. (2a)]. A higher bending degree (highly strain-
ed CÀC sites), as present in more disordered onion-type mate-
rials and fullerene-type carbons, further decreases the activa-
tion energy, leading to dissociative chemisorption [Eq. (2b)].
plays a key role in the concentric growth of the spherical
carbon.[48] As for the macroshape, the nano-order in this mate-
rial resembles that of nano-onion, but with a great disorder.
The presence of sp3–sp2 mixed coordination, with the curva-
ture process leading to nanospheres with a size of ꢀ200 nm,
favors the formation of disordered stacking planes with various
terminations and defects (sp3-type carbon atoms). Thermal an-
nealing, however, rearranges these layers into materials or-
dered at the nanoscopic level, according to the description
outlined above.
*
*
*
*
*
C¼C þ Cl2 ! CÀC ðÁ Á ÁCl2Þ
ð2aÞ
ð2bÞ
*
*
*
C¼C þ Cl2 ! ðClÞ CÀC ðClÞ
Formation of HCl occurs either consecutively to the forma-
Conclusions
tion of dissociatively chemisorbed Cl2 [Eq. (2b)] by reaction
with carbon sites with terminating defects (*CÀH) [Eq. (3a)] or
alternatively by direct heterolytic chemisorption on these sites
[Eq. (3b)]. We believe that Equation (3a) is more likely and is in
better agreement with the experimental results, in particular
with Raman data (Figure 6), however, this is an aspect deserv-
ing of further studies.
A novel nanostructured carbon was synthesized by a combina-
tion of thermal decomposition in an aqueous environment
and subsequent thermal annealing. This carbon material shows
a unique strain on the graphene sheets, inducing carbon sites
for strong chemisorption of Cl2.
In the parent carbon material, the high degree of bending in
CÀC sp2-type bonds induces irreversible dissociative chemi-
sorption of Cl2, and the high concentration of defect carbon
atoms (sp3-type, to which H atoms are coordinated to saturate
the valence) favors further reaction of chemisorbed Cl species
to generate HCl. Thermal annealing forms nano-ordered hemi-
spherical graphene shell structures (hemi-fullerene type) with
a decrease of the strain in CÀC sp2-type bonds, as well as a re-
duction of defect (sp3-type) carbon atoms. The reorganization
transforms sites able to irreversibly (dissociatively) chemisorb
Cl2 to sites able to strongly, but reversibly and in a molecular
form, coordinate Cl2. These sites react with CO to catalytically
generate COCl2, and also promote CH4 chlorination. Reorgani-
zation also reduces the concentration of carbon defects (sp3-
type), thus decreasing HCl formation and catalyst deactivation.
During interaction with these sites, the formation of [*CÀCl2]
adducts with radical character is enabled. By using CO and CH4
as reactive probe molecules, it was possible to show that this
is the only species able to contribute to COCl2 synthesis and
CH4ÀnCln formation. The properties and catalytic potential of
this new material are currently being further explored.
*
*
*
*
*
*
ðClÞ CÀC ðClÞ þ CÀH ! ðClÞ CÀC þ HCl þ C
ð3aÞ
ð3bÞ
*
*
CÀH þ Cl2 ! CÀCl þ HCl
Generation of the active sites
The results show how two carbon samples, with almost identi-
cal round-shaped macroscopic morphology and similar surface
area and porosity, change their reactivity towards Cl2 activation
and further catalytic CO and CH4 chlorination by variation of
their structural order.
By using thermal annealing, onion-type carbon nanospheres,
with hemi-fullerene-type ordered nanoshells on the surface,
have been prepared. The nanoshells are obtained by thermal
reordering of the precursor, highly defective, carbon material,
but possibly already have the suitable structure to form this
specific surface nanostructure upon annealing. The catalyst ob-
tained after the thermal treatment, with respect to the parent
carbon material, has sp2-carbon sites with unusually high
strength for the adsorption of Cl2 molecularly (in a reversible
manner), leading to high catalytic activity in phosgene genera-
tion, significantly higher than for many other type of carbon
materials (although not shown here). The annealing process
lowers the amount of sp3-type defect carbon sites, responsible
for Cl2 conversion to HCl. Possibly, the thermal ordering pro-
cess, indicated by the increase of GI, also causes modifications
of highly strained sp2-carbon sites responsible for irreversible
Cl2 chemisorption, although there is no definitive evidence on
this aspect.
Experimental Section
Experimental set-up and safety issues
One major advantage of the experimental setup designed for Cl2
activation studies is the possibility to use minimal amounts of toxic
and corrosive gases. A six-port valve connected to an online mass
spectrometer (MS) was used for the automated dosage of Cl2. All
the gaseous residues were neutralized in two interconnected con-
centrated KOH and Cu2+ solution-containing bottles. An active
carbon-containing filter accounts of any remaining toxic gaseous
intermediates.
This special carbon nanostructure requires, to be created by
thermal treatment, a suitable and specific carbon arrangement
in the (quasi-amorphous) precursor. The synthesis of these pre-
cursor nanospheres depends not only on the precursors and
the heat-treatment conditions, but also on the environment in
which the carbon spheres are generated. In the case of CHT
synthesis, the solid–liquid interface between the carbon source
and the surrounding water during the thermal decomposition
Two different types of Cl2 and CO detectors were placed under the
fume hood to constantly monitor the Cl2 and CO concentration at
the work station. To avoid contamination with CO, two additional
filters for CO entrapment were installed at the outlet of the six-
port valve and of the reactor. The maximum amount of COCl2 that
could theoretically be formed in the pulse reactor (0.025 ppm) is
ChemCatChem 2015, 7, 3036 – 3046
www.chemcatchem.org
3044
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
lower than the permissible exposure limit (0.1 ppm averaged over
a work shift of up to 10 h a day, 40 h per week with a ceiling level
of 0.2 ppm averaged over a 15 min period).
tion of the carbon materials under inert conditions at 1508C for
1 h. After Cl2 treatment further measurements under inert condi-
tions (in N2 flow) were performed. For in situ Raman measure-
ments a temperature-regulated quartz reactor (1.62 mm internal di-
ameter) was used.
Synthesis
Powder carbon synthesis
Activity tests
The sample indicated as CHT was obtained by dissolution of pure
glucose in distilled water (180 mL) to form a clear solution, and
placed in a Teflon-lined autoclave at 2008C for 20 h. The product
was then separated by filtration, washed several times with hot
water, acetone, and ethanol. It was dried at 1008C and further
treated at 2008C for 2 h. The quantitative yield was approximately
20 wt%.
Multipulse experiments were carried out to investigate the interac-
tions of the carbon materials with Cl2 and to identify the transient-
ly formed intermediates when reacting with CO and CH4. For each
experiment, the solid carbon material (10 mg, particle size
ꢁ80 mm) packed in a quartz reactor (0.4 cm internal diameter) was
activated in He flow (10 mLminÀ1) at 1508C for 1 h prior to cataly-
sis. The schematic representation of the experimental setup con-
structed for the Cl2 activation study is shown in the Supporting In-
formation, Figure S2. Multiple Cl2 pulses (3.6 mmol Cl2/pulse, pulse
length 230 s) were performed to understand the catalytic activa-
tion of Cl2 on the multiple active sites in the selected carbons. To
probe the nature of activated Cl2 species in separate experiments,
5% CO and CH4 in He carrier (flow rate 10 mLminÀ1) were exposed
to the Cl-containing carbon sample. All gaseous products were an-
alyzed by online mass spectrometry (PFEIFFER OMNIstar/TM mass-
spectrometer, QMS 200). For the identification of the surface-
bound species, temperature-programmed desorption (TPD) was
Thermal annealing procedure
For each thermal annealing procedure, an amount of sample
(ꢀ200 mg) was placed in a vertical fixed-bed reactor. The sample
underwent graphitization at an annealing temperature of 10008C
for 6 h under a bottom-up N2-gas flow (100 STPmLminÀ1). This
sample is named CHT-tr.
Characterization
performed up to 8008C with
a temperature increment of
58CminÀ1 in He flow (flow rate 10 mLminÀ1).
Carbon morphology structure was investigated by transmission
electron microscopy (TEM) by using a Philips CM12 microscope
(resolution 0.2 nm) equipped with a high-resolution camera, at an
accelerating voltage of 120 kV. Suitable specimens for TEM analyses
were prepared by ultrasonic dispersion in iso-propyl alcohol by
adding a drop of the resulting suspension onto a holey carbon-
supported grid.
Acknowledgements
The authors acknowledge support from the European Communi-
ty’s Seventh Framework Program [FP7/2007–2013] under grant
agreement no. NMP-LA-2010–245988 (INCAS).
Samples were characterized by X-ray powder diffraction (XRD) by
using a Philips X-Pert diffractometer with a monochromatic CuKa
(l=1.54056 Š) radiation at 40 kV and 30 mA. Data were collected
over a 2q range of 10–100 degrees, with a step size of 0.048 at
a time per step of 3 s. The graphitization degree of onion-like
carbon nanospheres was quantitatively determined by using LiF as
a standard material (in a ratio from 1 to 5). For quantification, the
peak area and the intensity were used and analyzed comparatively
to the standard sample. Results are presented in Table 1 and
shown in Figure S1.
Keywords: carbon · chlorination · graphene · nanostructures ·
strained molecules
[1] a) Y. Qin, L. He, J. Duan, P. Chen, H. Lou, X. Zheng, H. Hong, ChemCat-
Chem 2014, 6, 2698; b) V. Calvino-Casilda, A. J. Lopez-Peinado, C. J.
Duran-Valle, R. M. Martin-Aranda, Catal. Rev. Sci. Eng. 2010, 52, 325;
c) T. J. Bandosz in Carbon Materials for Catalysis (Eds.: P. Serp, J. L. Fig-
ueiredo), Wiley, Hoboken, 2008, p. 45.
[2] a) W. Qi, D. Su, ACS Catal. 2014, 4, 3212; b) D. C. Sorescu, K. D. Jordan, P.
Avouris, J. Phys. Chem. B 2001, 105, 11227; c) Z. Y. Zhu, S. M. Lee, Y. H.
Lee, T. Frauenheim, Phys. Rev. Lett. 2000, 85, 2757; d) J. Zhang, X. Liu, R.
Blume, Z. Zhang, R. Schlçgl, D. Su, Science 2008, 322, 73.
[3] a) D. S. Su, G. Centi, S. Perathoner, Chem. Rev. 2013, 113, 5782; b) G.
Centi, S. Perathoner, D. S. Su, Catal. Surv. Asia 2014, 18, 149; c) D. S. Su,
G. Centi, S. Perathoner, Catal. Today 2012, 186, 1.
[4] a) J. H. Bitter, J. Mater. Chem. 2010, 20, 7312; b) X. Pan, X. Bao, Acc.
Chem. Res. 2011, 44, 553.
[5] R. Arrigo, M. Hävecker, S. Wrabetz, R. Blume, M. Lerch, J. McGregor,
E. P. J. Parrott, J. A. Zeitler, L. F. Gladden, A. Knop-Gericke, R. Schlçgl,
D. S. Su, J. Am. Chem. Soc. 2010, 132, 9616–9630.
The specific surface areas were calculated by using the Brunauer
Emmett Teller (BET) equation from the adsorption branch of the
isotherms, obtained at À1968C on a Quantachrome sorption analy-
ser. Prior to the measurements, samples were heated in a N2 flow
at 3508C for 1 h. The micropore volume was evaluated by the t-
plot method.
X-ray Photoelectron Spectroscopy (XPS) data were measured by
using a Physical Electronics GMBH PHI 5800-01 spectrometer oper-
ating with a monochromatized AlKa radiation with a power beam
of 300 W. The pass energy for determination of the oxidation state
and concentration of surface species was 11.0 and 58.0 eV, respec-
tively. The BE regions of C1s (280–300 eV) and O1s (524–544 eV)
were investigated, taking the Al2p line (73.0 eV) of aluminum stan-
dard as a reference for signal calibration. Ar+ 2 kW voltage small
beam sputtering was performed to remove adventurous carbon.
[6] L. Mleczko, G. Lolli, Angew. Chem. Int. Ed. 2013, 52, 9372–9387; Angew.
Chem. 2013, 125, 9540–9557.
[7] J.-W. Shi, X. Zong, X. Wu, H.-J. Cui, B. Xu, L. Wang, M.-L. Fu, ChemCat-
Chem 2012, 4, 488–491.
[8] D. S. Su, J. Zhang, B. Frank, A. Thomas, X. Wang, J. Paraknowitsch, R.
Schlçgl, ChemSusChem 2010, 3, 136.
[9] R. Chakravarti, M. L. Kantam, H. Iwai, S. S. Al-deyab, K. Ariga, D.-H. Park,
J.-H. Choy, K. S. Lakhi, A. Vinu, ChemCatChem 2014, 6, 2872.
[10] D. S. Su, J.-O. Müller, R. E. Jentoft, D. Rothe, E. Jacob, R. Schlçgl, Top.
Catal. 2004, 30–31, 241.
Raman spectra at 514 nm (2.41 eV, Ar+ laser) were recorded with
a Renishaw (Series 1000) spectrometer equipped with a 50” objec-
tive. All Raman spectra were obtained following the thermal activa-
ChemCatChem 2015, 7, 3036 – 3046
www.chemcatchem.org
3045
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[11] V. Georgakilas, J. A. Perman, J. Tucek, R. Zboril, Chem. Rev. 2015, 115,
4744.
[39] K. Gopalakrishnan, K. S. Subrahmanyam, P. Kumar, A. Govindaraj, C. N. R.
Rao, RSC Adv. 2012, 2, 1605.
[40] M. Yang, L. Zhou, J. Wang, Z. Liu, Z. Liu, J. Phys. Chem. C 2012, 116, 844.
[41] S. Abate, G. Centi, S. Perathoner, Natl. Sci. Rev. 2015, 2, 143.
[42] I. M. Y. A. M. Rabiu, J. Power Energy Eng. 2013, 1, 41.
[43] C. Hammond, S. Conrad, I. Hermans, ChemSusChem 2012, 5, 1668.
[44] Y. Wei, D. Zhang, Z. Liu, B.-L. Su, Cuihua Xuebao 2012, 33, 11.
[45] Z. Guo, B. Liu, Q. Zhang, W. Deng, Y. Wang, Y. Yang, Chem. Soc. Rev.
2014, 43, 3480.
[46] J. He, T. Xu, Z. Wang, Q. Zhang, W. Deng, Y. Wang, Angew. Chem. Int. Ed.
2012, 51, 2438; Angew. Chem. 2012, 124, 2488.
[47] K. Barbera, L. Frusteri, G. Italiano, L. Spadaro, F. Frusteri, S. Perathoner,
G. Centi, Chin. J. Catal. 2014, 35, 869.
[12] T. J. Bandosz, M. Seredych, Chin. J. Catal. 2014, 35, 807.
[13] J. Matthiesen, T. Hoff, C. Liu, C. Pueschel, R. Rao, J.-P. Tessonnier, Chin. J.
Catal. 2014, 35, 842.
[14] L. R. Radovic, C. Mora-Vilches, A. J. A. Salgado-Casanova, Chin. J. Catal.
2014, 35, 792.
[15] G. Centi, S. Perathoner, ChemSusChem 2011, 4, 913.
[16] M. Godino-Ojer, A. J. López-Peinado, R. M. Martín-Aranda, J. Przepiórski,
E. PØrez-Mayoral, E. Soriano, ChemCatChem 2014, 6, 3440.
[17] W. Zhang, L. Theil Kuhn, ChemCatChem 2013, 5, 2591.
[18] Y. Zhu, Y. Lin, B. Zhang, J. Rong, B. Zong, D. S. Su, ChemCatChem 2015,
DOI: 10.1002/cctc.201402930.
[48] L. Frusteri, C. Cannilla, K. Barbera, S. Perathoner, G. Centi, F. Frusteri,
Carbon 2013, 59, 296.
[19] R. Wang, X. Sun, B. Zhang, X. Sun, D. Su, Chem. Eur. J. 2014, 20, 6324.
[20] J. Zhu, A. Holmen, D. Chen, ChemCatChem 2013, 5, 378.
[21] A. Ramos, I. Cameµn, A. B. García, Carbon 2013, 59, 2-32.
[22] D. Tasis, N. Tagmatarchis, A. Bianco, M. Prato, Chem. Rev. 2006, 106,
1105.
[49] X.-Y. Liu, J.-M. Zhang, K.-W. Xu, Physica B 2014, 436, 54.
[50] J. Zhang, D. S. Su, R. Blume, R. Schlçgl, R. Wang, X. Yang, A. Gajovic,
Angew. Chem. Int. Ed. 2010, 49, 8640; Angew. Chem. 2010, 122, 8822.
[51] Z. Wang, L. Yu, W. Zhang, Z. Zhu, G. He, Y. Chen, G. Hu, Phys. Lett. A
2003, 307, 249–252.
[23] C. T. J. Low, F. C. Walsh, M. H. Chakrabarti, M. A. Hashim, M. A. Hussain,
Carbon 2013, 54, 1.
[52] W. Zhang, M. Yao, X. Fan, S. Zhao, S. Chen, C. Gong, Y. Yuan, R. Liu, B.
Liu, J. Chem. Phys. 2015, 142, 034702.
[53] a) J. Robertson, Phys. Rev. B 2000, 61, 14095; b) S. E. Rodil, J. Robertson,
Phys. Rev. B 2003, 67, 155306.
[24] S. Stankovich, D. A. Dikin, R. D. Piner, K. A. Kohlhaas, A. Kleinhammes, Y.
Jia, Y. Wu, S. T. Nguyen, R. S. Ruoff, Carbon 2007, 45, 1558.
[25] S. Y. Sawant, R. S. Somani, A. B. Panda, H. C. Bajaj, Mater. Lett. 2013, 94,
132.
[54] P. K. Chu, L. Li, Mater. Chem. Phys. 2006, 96, 253.
[55] N. A. Marks, D. W. M. La, Phys. Rev. B 2009, 79, 075430.
[56] A. C. Ferrari, Solid State Commun. 2007, 143, 47.
[57] J. Shang, L. Ma, J. Li, W. Ai, T. Yu, G. G. Gurzadyan, Sci. Rep. 2012, 2, 792.
[58] F. Tuinstra, J. L. Koening, J. Chem. Phys. 1970, 53, 1126.
[59] D. Nishide, H. Kataura, S. Suzuki, K. Tsukagoshi, Y. Aoyagi, Y. Achiba,
Chem. Phys. Lett. 2003, 372, 45.
[26] D. R. Dreyer, H.-P. Jia, C. W. Bielawski, Angew. Chem. Int. Ed. 2010, 49,
6813; Angew. Chem. 2010, 122, 6965.
[27] Y. Z. Jin, C. Gao, W. K. Hsu, Y. Zhu, A. Huczko, M. Bystrzejewski, M. Roe,
C. Y. Lee, S. Acquah, H. Kroto, D. R. M. Walton, Carbon 2005, 43, 1944.
[28] C. Su, K. P. Loh, Acc. Chem. Res. 2013, 46, 2275.
[29] B. Keggenhoff, H. Lokum, J. Münnig, Process and apparatus for the pro-
duction of phosgene, Patent US7442835, 2008.
[60] V. M. Irurzun, M. Pilar Ruiz, D. E. Resasco, Carbon 2010, 48, 2873.
[61] S. Yumitori, J. Mater. Sci. 2000, 35, 139.
[30] C. J. Mitchell, K. S. van der Velde, M. Smit, R. Scheringa, K. Ahrika, D. H.
Jones, Catal. Sci. Technol. 2012, 2, 2109.
[62] Y. Ling, G. K. Koyanagi, D. Caraiman, V. Baranov, D. K. Bohme, Int. J. Mass
Spectrom. 1999, 182/183, 349.
[63] I.-Y. Jeon, H. M. Kim, I. T. Choi, K. Lim, J. Ko, J. C. Kim, H.-J. Choi, M. J. Ju,
J.-J. Lee, H. K. Kim, Nano Energy 2015, 13, 336.
[64] G. Czakó, J. M. Bowman, J. Chem. Phys. 2012, 136, 044307/1.
[65] Z. Chen, W. Thiel, A. Hirsch, ChemPhysChem 2003, 4, 93.
[66] I. H. Campbell, D. L. Smith, Solid State Physics 55 (Eds.: H. Ehrenreich,
F.Spaepen), Academic Press, New York, 2000.
[67] P. A. van Hal, M. P. T. Christiaans, M. M. Wienk, J. M. Kroon, R. A. J. Jans-
sen, Phys. Rev. B 1999, 103, 4352.
[31] J.-P. Senet, C. R. Acad. Sci. Ser. IIc: Chim. 2000, 3, 505.
[32] L. Cotarca, H. Eckert, Phosgenations—A Handbook, Wiley-VCH, Wein-
heim, 2005.
[33] V. A. Volodina, A. A. Kozlovskii, S. I. Kuzina, A. I. Mikhailov, High Energy
Chem. 2008, 42, 311.
[34] T. Zhang, C. Troll, B. Rieger, J. Kintrup, O. F. K. Schlüter, R. Weber, J.
Catal. 2010, 270, 76.
[35] J. Heuser, H. Kauth, C. Kords, Phosgene having low content of carbon tet-
rachloride, Patent US6930202, 2005.
[36] F. A. Via, G. L. Soloveichlk, P. G. Kosky, W. V. Cicha, Method for producing
phosgene, Patent US6500984, 2002.
[37] H. S¸ahin, S. Ciraci, J. Phys. Chem. C 2012, 116, 24075.
[38] B. Li, L. Zhou, D. Wu, H. Peng, K. Yan, Y. Zhou, Z. Liu, ACS Nano 2011, 5,
5957.
Received: June 10, 2015
Revised: July 29, 2015
Published online on August 31, 2015
ChemCatChem 2015, 7, 3036 – 3046
www.chemcatchem.org
3046
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim