Oxygen insertion catalysis by sp2 carbon

Article abstract of DOI:10.1002/anie.201103340

Black matter in catalysis: Graphitic carbon catalyzes the insertion of O atoms into acrolein. Such complex multistep atom rearrangements were believed to be the exclusive domain of metal (oxide) catalysis. In the C-catalyzed process, the nucleophilic O atoms terminating the graphite (0001) surface abstract the formyl H atom and the activated aldehyde is oxidized by a mobile epoxide O atom. Thus, the sp² carbon acts as a bifunctional catalyst.

Full text of DOI:10.1002/anie.201103340

Communications
DOI: 10.1002/anie.201103340
Metal-Free Catalysis
Oxygen Insertion Catalysis by sp2 Carbon**
Benjamin Frank, Raoul Blume, Ali Rinaldi, Annette Trunschke,* and Robert Schlçgl
Dedicated to the Fritz Haber Institute, Berlin, on the occasion of its 100th anniversary
Elemental carbon with sp2 hybridization is omnipresent in our
lives, and in its natural form, graphite, an integral part of
pencils, batteries, lubricants, steel, and electric motor brushes.
High-tech electronic devices and the neutron moderators in
nuclear power plants are made of synthetic graphite with a
low density of defects (the number of carbon atom disloca-
tions in the hexagonal lattice). Nanostructured sp2 carbon is
used as a pigment (carbon black) and as a polymer filler. The
discovery and synthesis of nonplanar carbon allotropes, such
as carbon nanotubes (CNTs) and fullerenes,[1] ushered in a
new era of cutting-edge applications for elemental carbon.
The basic structural unit is graphene,[2] with edge defects and a
curvature induced by non-six-membered carbon rings.
possibility to modify and optimize the chemical environment
of the active sites, and thus to obtain well-defined structure–
reactivity correlations. The absence of strongly Lewis-acidic
metal cations minimizes the deposition of coke and thus
deactivation of the catalyst. This in turn can reduce process
costs by not necessitating the addition of steam or periodic
regeneration of the catalyst by coke burning.[9] Such progress
in ODH and DH catalysis encouraged us to test nano-
structured carbon catalysts for oxygen insertion in the
selective gas-phase oxidation of acrolein to acrylic acid
(AA) as a model reaction [Eq. 1].
1
ð1Þ
CH2¼CHCHO þ = O2 ! CH2¼CHCOOH
2
In chemistry, graphitic forms of carbon have an intriguing
potential for catalysis, with a broad spectrum of application
covering hydrogenation, oxidation, polymerization, and
chlorination reactions.[3,4] The increased reaction scope has
recently been reviewed by Dreyer and Bielawski,[5] who
themselves investigated the catalytic activity of graphene
oxide in several reactions under mild conditions in the liquid
phase.[6] The most prominent example of heterogeneous gas-
phase catalysis by carbon materials is the selective oxidative
dehydrogenation (ODH) of ethylbenzene to styrene[7,8]—a
reaction of high industrial relevance—by nucleophilic oxygen
atoms located at the prismatic edges of stacked graphene
sheets or at surface defects in the (0001) graphitic surface. For
the oxygen-free pathway (DH), the activity of nanocrystalline
diamonds coated with defective graphene shells, so-called
“bucky diamonds”, even exceed the industrial potassium-
promoted iron catalyst.[9] It is not surprising that the ODH of
light alkanes suffers from a much lower selectivity because
Here, hydrogen abstraction from the formyl group is
followed by oxygenation to form the carboxy group. This
industrial process is carried out over optimized Mo/V mixed
metal oxide catalysts to provide AA in yields of greater than
95% at temperatures as low as 200–3008C.[12] The addition of
steam to the feed increases the performance and lifetime of
the catalysts.[12–14] Graphitic carbon allotropes can resist the
oxidative stress under these conditions.[15] However, the use of
natural graphite flakes as the heterogeneous catalyst results in
a poor C3H4O conversion of 0.4% at an AA selectivity of only
37%. A comparison of graphite flakes of different diameter
discloses the location of the active sites: decreasing the size of
the flakes by using ball-milled synthetic graphite results in an
increase in the conversion to 8.8%, while the selectivity is
maintained at a moderate level of 66%; this finding indicates
that the prismatic edges of stacked graphene layers host the
active moieties for this reaction. A pronounced evolution of
the catalytic performance is observed, and the initial decay of
activity accompanied by an increasing selectivity during the
initial 12 h time-on-stream can be related to the (trans)-
formation of oxygen species on the carbon surface (see
Figure S1 in the Supporting Information) and the healing of
point defects. Similar to (bulk) metal oxides, the switch
between a reduced and oxidized state during the reaction is
realized by reoxidation with gas-phase oxygen. However, in
contrast to metal oxides and the related Mars-van Krevelen
redox mechanism, the bulk of the graphite catalyst cannot
serve as a reservoir for oxygen atoms. Instead, the (0001)
graphitic surface, which mediates the dissociative adsorption
of oxygen in the combustion of carbon,[16] could take on this
role as well as serve as an electron buffer to facilitate the
redox process at the active sites. However, the rate of
formation of AA of 86.5 mmol gꢀ1 hꢀ1 on natural graphite is
unquestionably poor compared to the industrial and academic
state of the art.[12,17]
ꢀ
the C H bond in the product molecule is weaker than in the
substrate[10]—a common problem in selective oxidation
reactions.[11] Point defects filled with electrophilic oxygen
likely initiate nonselective combustion pathways. The low-
dimensionality of nanostructured carbon materials offers the
[*] Dr. B. Frank, Dr. R. Blume, Dr. A. Rinaldi, Dr. A. Trunschke,
Prof. Dr. R. Schlçgl
Department of Inorganic Chemistry
Fritz Haber Institute of the Max Planck Society
Faradayweg 4-6, 14195 Berlin (Germany)
E-mail: trunschke@fhi-berlin.mpg.de
Homepage: http://www.fhi-berlin.mpg.de/ac/
[**] This work was supported financially by the EnerChem project of the
Max Planck Society. We thank D. S. Su and O. V. Khavryuchenko for
discussions, and L. I. Csepei for assistance with the experiments.
The chemical structures shown in Figures 1, 3, and S3 were created
using Jmol open-source software.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201103340.
10226
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 10226 –10230

The screening of low-dimensional carbon allotropes[18]
gives insight into basic structure–activity relationships
(Table 1 and see S1 in the Supporting Information). The
superior performance is observed for the allotropes with bent
on the curved basal plane because of pronounced charge
localization.
O2ðgÞ ! O2ðsÞ ! O2ꢀðsÞ ! O22ꢀðsÞ ! OꢀðsÞ ! O2ꢀðsÞ
ð2Þ
The interaction and activation of small molecules such as
O2, H2O, CO, or CO2 by the defective (0001) surface has been
described theoretically.[19] However, a higher degree of sp3
hybridization, as present in the active carbon and nano-
diamond samples, is detrimental to the selectivity of forma-
tion of AA. The moderate performance of fishbone-like
carbon nanofibers (CNFs), which solely expose the prismatic
edges to the outer tubular surface, further highlights the
importance of extended areas of the (0001) basal plane for the
reaction. The relatively poor performance of the quinone
model catalyst (MCT, macrocyclic trimer) compared to its
superiority over other nanocarbon catalysts in the ODH of
ethylbenzene[7] is in line with these results and gives rise to the
question of whether the presence of nucleophilic quinine
groups is also sufficient for the selective oxidation of C3H4O
to AA. The stability of the catalytic system is confirmed by
the long-term reaction over 120 h. High-resolution trans-
mission electron microscopy (HRTEM) and Raman spec-
troscopy analyses of the catalysts (see Figure S4 in the
Supporting Information) show the structural integrity of the
most promising MWCNT catalyst as well as the absence of
severe surface damage. In accordance with the conditions of
the industrial process, we added 5 vol% H2O to the reaction
stream. In the case of the MWCNTs, the conversion of C3H4O
increased from 14% without addition of H2O up to 19% in
the presence of H2O. In parallel, the selectivity for AA also
increased from 85 to 87%. A further increase in the steam
content up to 40% drives the C3H4O conversion and AA
selectivity up to 24% and 90%, respectively. The improve-
ment of the catalytic performance is related to the modified
surface properties of the carbon surface under wet conditions.
TPD analysis of the catalysts clearly reveals an increased
number of carboxy species on switching from a dry to a wet
feed (Figure 1A,B), whereas the amount of other oxygen
surface species remains fairly constant. A similar result is
obtained by quasi in situ XPS analysis (Figure 1C–E), where
according to band assignments in the literature,[15] the carboxy
band is located at around 533.0–533.5 eV. Accordingly, the
C1s range (see Figure S5 in the Supporting Information) is
characterized by bands at 286 and 288.5 eV, thus indicating
Table 1: Structural data and catalytic performance of various carbon
allotropes in the selective oxidation of acrolein to acrylic acid.
Nanocarbon
SBET
d[a]
[nm]
X[C3H4O][b]
[%]
S[AA][b]
[%]
[m2 gꢀ1
]
natural graphite
synthetic graphite
MWCNT
fishbone CNF
OLC
nanodiamond
fullerenes
MCT
11.0
349
541
50
315
1.7
320
5.0
835
123
6
0.4
8.8
14
37
66
85
35
9.9
15–20
5–15
5–15
0.7
–
1.8
4.5
3
75
51
31[c]
5
41[c]
12
activated carbon
–
26[c]
51[c]
[a] Characteristic diameter of the nanostructured carbon as determined
by HRTEM, XRD for natural graphite (La), and Raman analysis for
synthetic graphite (La). [b] Determined after 15 h time-on-stream;
5 vol% C3H4O/10 vol% O2/He, 3008C, 3000 hꢀ1; side products are CO,
CO2, and trace amounts of acetic acid. [c] No stable performance
because of severe oxidative degradation (C balance >100%).
graphene sheets, namely, multiwalled carbon nanotubes
(MWCNTs) and onionlike carbon (OLC), whereas the sp3-
hybridized nanodiamonds result in a low selectivity for the
formation of AA. Disordered forms of carbon, such as
activated carbon, cannot coordinate the selective reaction
and, in addition, rapidly deactivate because of oxidative
degradation.[15] The rapid loss of activity is also observed with
the C60 fullerenes, probably with an open-cage structure in the
oxidative atmosphere, thus showing that single-shell graphitic
allotropes such as fullerenes, graphene, and single-walled
CNTs (SWCNTs) cannot resist the oxidative stress at 3008C
and combust within hours. A heterogeneous model catalyst
that provides solely the nucleophilic diketonic carbonyl
groups[7] (see Figure S3 in the Supporting Information) also
shows a low selectivity in the formation of the desired acid.
When the number of ketonic and phenolic oxygenated sites,
which temperature-programmed desorption (TPD) show
decompose to CO at 650–8508C (see Table S2 in the
Supporting Information), are considered, reaction rates of
6.8 ꢀ 10ꢀ5 sꢀ1 and 1.9 ꢀ 10ꢀ4 sꢀ1 can be estimated for the
reaction with the graphite and the MWCNT catalysts,
respectively. The oxygen functionalities located at the edges
and defects of the curved graphene layer have a higher
intrinsic activity than those terminating the flat (0001)
surface. Furthermore, the bent structure induces a remark-
able rise in the selectivity for the formation of AA. A
pronounced charge location as a consequence of the curva-
ture benefits the adsorption and activation of dioxygen and
may control the type of surface species formed to favor the
selective turnover. On the planar basal plane, the unselective
peroxo species, as the intermediate of O2 adsorption with
stepwise charge transfer [Eq. 2], will be more stable than the
epoxy group, which is the product of dissociation, as favored
ꢀ
=
C O and C O bonds, respectively, as part of the surface
functional groups. Their intensity is higher in the MWCNT
catalysts and lower in the graphite catalyst.
This finding conclusively proves the effect of surface
acidity, which has been similarly discussed for metal oxide
catalysts.[13,20] Water molecules transform Lewis acid sites into
Brønsted acids, thus ultimately blocking the centers to total
oxidation. The surface protonation further enables the
formation of AA from the acrylate surface complex and
additional steam favors its desorption by competitive adsorp-
tion on the catalyst surface. A shift of the band from 725 to
6908C, which corresponds to phenol decomposition, is the
clearest change in the CO desorption profile (Figure 1A).
The destabilization of these species, which likely represent the
Angew. Chem. Int. Ed. 2011, 50, 10226 –10230
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10227

Communications
analysis (7.8 mmolgꢀ1). This observation proves that the
highly stable ketones (1.4 mmolgꢀ1) will not transfer oxygen
to the substrate molecule. Instead, labile electrophilic oxygen
ꢀ
atoms likely insert into the formyl C H bond after its
activation by nucleophilic oxygen species. Regarding the fact
that high activity/selectivity of carbon allotropes is strictly
correlated to the exposition of the (0001) basal plane to the
ꢀ ꢀ
outer surface, we suggest the mobile epoxide C O C species,
as the electrophilic product of dissociative O2 adsorption,[16,23]
to be responsible for selective turnover. These species are
equipped with a high reactivity as a result of severe strain,
because the two incorporated carbon atoms change their
configuration from a planar sp2- to a distorted sp3-hybridized
geometry. Indeed, the graphitic epoxide can migrate to the
prismatic edges,[16,23] where catalytic transformation takes
place, or instead transform to more stable groups (including
CO/CO2 formation) even at temperatures as low as 2008C.[24]
The insertion of epoxide oxygen, which is present in graphene
oxide, into organic molecules was indicated in a recent
report[25] that demonstrates the removal of typical epoxide IR
bands on oxidation of benzyl alcohol in the absence of O2. The
removal of such species to leave the intact (0001) basal plane
has further been predicted for the reaction of epoxides with
adsorbed hydrogen to hydroxyl radicals and/or water.[26]
To localize the active sites at the prismatic edges, we
modified the surface of MWCNTs with either 1 wt% B2O3 or
P2O5. It is reported, that these heteroatoms selectively block
the zigzag and armchair terminations of graphene sheets,
respectively.[27] The decreased C3H4O conversions of 6.5% for
the B2O3/MWCNTs and 5.2% for the P2O5/MWCNTs thus
provide evidence for the catalytic activity of both config-
urations. However, the P modification results in an increased
AA selectivity (90.9%) compared to the pristine MWCNTs,
whereas the B-modified sample shows a lower AA selectivity
of 83.9%. This observation suggests that the quinone group
located at the zigzag termination of the graphene layer acts
more selectively in the C3H4O oxidation than the one in the
armchair configuration, even though this effect is only weakly
pronounced. It should also be considered that this kind of
modification also affects the surface acidity and blocks point
defects. The robustness of the catalytic system is unique:
neither (short-term) harsh oxidative stress nor high-temper-
ature calcination, which substantially defunctionalizes the
carbon surface, can induce persistent damage to its catalytic
function. The MWCNT catalyst rapidly recovers after treat-
ment in air at 5008C, as well as after TPD in He up to 8508C,
and the catalyst approaches its initial activity and selectivity
within a few hours (see Figure S8 in the Supporting Informa-
tion). As compared to metal oxide catalysts, which often
comprise several more- or less-active and selective crystallo-
graphic phases that may interchange during the course of a
reaction and cause irreversible deactivation, the graphitic
carbon is the thermodynamically most favorable form of
carbon, and thus provides no possibility for structural
rearrangement.[10b]
Figure 1. Surface analysis of the catalysts used. A) CO desorption from
MWCNTs, indicating the presence of anhydrides (500ꢁ1208C), phe-
nols (610ꢁ1208C), and ketones/quinones (830ꢁ1508C). B) CO2
desorption from MWCNTs, indicating the presence of carboxylic acids
(270ꢁ1708C), anhydrides (490ꢁ1208C), and lactones
(700ꢁ1108C).[3] The MWCNTs were pretreated for 15 h at 3008C in a
stream of 5 vol% C3H4O/10 vol% O2/He, then cooled in He to 1008C.
The TPD profile (filled symbols) was recorded at a heating rate of
10 Kminꢀ1 (red lines). The experiment was repeated under wet
conditions (open symbols), i.e., 5 vol% H2O was added to the
reactant/carrier gas during the pretreatment, cooling, and TPD. The
functionalized surface termination is illustrated below the TPD profiles,
respectively. C gray, O red, H white. C–E) Synchrotron-excited quasi in
situ XP spectra (O1s range) of (C) synthetic graphite and MWCNTs
[D) dry, E) wet].
reduced state of the active site, will likely affect their
reactivity with oxygen. The positive effect of water on the
catalyst reoxidation[15,21] is in line with a number of similarities
between metal oxide and carbon-based catalysts.
Steady-state isotopic tracer kinetic analysis (SSITKA; see
Figure S6 in the Supporting Information) on switching the
16O2 to 18O2 reveals the vital exchange of oxygen in the
acrolein molecule. This analysis suggests the reversible
formation of an acetal-like adsorbate on the carbon surface,
whose participation as an intermediate in the selective
oxidation remains rather unclear. However, this process
results in the formation of singly and doubly labeled AA.
Again, this effect is well-known to occur over Mo/V
catalysts.[14] Subsequent TPD analysis of the 18O2-treated
MWCNT catalyst (see Figure S6c,d in the Supporting Infor-
mation) leads to the identification of O atoms as being
responsible for insertion into the substrate. The 18O/16O
exchange is observed in the low-temperature range of CO and
CO2 desorption, whereas the high-temperature range is
dominated by C16Ox. This finding suggests that the nucleo-
philic ketones and quinones, as well as their reduced counter-
parts phenols, are not removed during the redox cycle of
selective C3H4O oxidation, in contrast to ODH catalysis
which operates at somewhat higher temperature.[22] Indeed,
the temperature-programmed reaction of C3H4O with the
oxygen-treated MWCNTs yields only a small amount of AA
(7.3 mmolgꢀ1, see Figure S7 in the Supporting Information),
which is in the range of the amount released during the TPD
The variation of the gas-hourly space velocity (GHSV)
reveals a weakly decreasing AA selectivity as the C3H4O
conversion increases up to 15% (see Figure S9 in the
Supporting Information), which suggests that the formation
10228 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 10226 –10230

of COx, the only by-products observed, occurs by direct
C3H4O combustion, whereas AA is more resistant to total
oxidation. In contrast to O2, the H2O concentration has a
major effect on the product distribution. Thus, the effect of
H2O on the reaction was also investigated by kinetic analysis.
The preexponential factor remains fairly constant at around
(3 ꢁ 1) ꢀ 1024 moleculesgꢀ1 sꢀ1, which indicates that the
number of active sites will not change when switching from
the dry to the wet feed. However, the addition of 5% H2O to
the reactant gas mixture significantly lowers the apparent
activation energy from 85 to 80 kJmolꢀ1 (Figure 2A). The
lowering of the barrier must be attributed to the surface
rate orders agree with the multistep reaction mechanism and
indicate that the carbon surface is almost saturated with
C3H4O and/or AA. The reaction rate controls the balance
between the surface reaction of C3H4O with the oxidized
carbon surface, the catalyst reoxidation, and the desorption of
AA from the active site. The conjugated p electrons of the
substrate and reaction product molecules can favor their
adsorption on the (0001) basal plane through van der Waals
forces. The reaction order of C3H4O remains constant in the
ꢀ
presence of 5% H2O, which suggests that C H activation is
not affected. The decrease in the rate order of O2 to 0.29 can
likely be interpreted by the accelerated reoxidation of the
carbon surface by the addition of steam, which has also been
observed for the combustion of graphitic carbon as well as for
the oxidation of VOx clusters supported on g-Al2O3.[15,21]
Taking all these experimental results and theoretic back-
ground into account, the pathway shown in Figure 3 for
C3H4O oxidation is suggested.
Figure 3. Suggested reaction pathway for the oxidation of C3H4O at the
graphitic carbon surface. The active domain is illustrated as a
rectangular section of a planar graphene sheet with a hole defect,
which is terminated by arbitrarily positioned oxygen functionalities. O2
adsorbs dissociatively at the (0001) surface to form mobile epoxy
groups, which migrate to the prismatic edge sites. The adsorption of
C3H4O at the nucleophilic oxygen sites, i.e., the ketones/quinones,
initiates its oxygenation by epoxy oxygen atoms to form acrylic acid.
Figure 2. Reaction kinetics of C3H4O oxidation. A) Arrhenius plot of
&
the reaction rates under dry ( ) and wet (&) conditions at 280–3208C
with a gas feed of 5 vol% C3H4O/10 vol% O2/He. B) Rate of C3H4O
oxidation as a function of the partial pressure p of O2 (circles) or
C3H4O (triangles) over MWCNTs under dry (filled symbols) and wet
(open symbols) conditions. The reaction rates were measured at
3008C. Wet reaction conditions were obtained by adding 5 vol% H2O
to the reaction gas mixture.
In conclusion, the element carbon has been demonstrated
to be a highly robust and selective catalyst for a model
chemical reaction involving the insertion of oxygen into an
organic molecule. In the selective oxidation of acrolein, a
productivity of 26.5 mmolgꢀ1 hꢀ1 could be achieved for acrylic
acid, which is almost half as high as for the industrial doped
MoV mixed oxide (about 60 mmolgꢀ1 hꢀ1).[12] Fundamental
structure–activity relationships enabled the catalytic perfor-
mance to be optimized with respect to the carbon micro-
structure and reaction conditions. Such a morphological
control on a macroscopic scale is difficult to achieve for
metal (oxide) catalysts,[28] thus giving elemental carbon a
unique position in catalysis research. In general, carbon
catalysis in the absence of polyvalent metal sites with complex
electronic and spin structures allows for facile and in-depth
theoretical analysis, and thus is an ideal model for under-
standing reaction mechanisms. In the current study, the
reaction mechanism of selective C3H4O oxidation could be
acidity induced by surface carboxy groups, which likely affect
the adsorption of the reactant and/or the stability of the
transition state. The activation energy under dry conditions
corresponds to the energy barrier of epoxy group hopping
(0.9 eV) as determined by first principles calculations,[23] thus
suggesting that oxygen mobility on the (0001) graphitic
surface might be the rate-determining factor of the reaction.
This is plausible if we consider the reaction to occur at the
edge of the basal planes of CNTs. A moderate influence of
catalyst reoxidation on the overall rate is confirmed by a
kinetic isotope effect (KIE) of k(16O2)/k(18O2) = 1.06 (see
Figure S6b in the Supporting Information). The reaction
orders for C3H4O and O2 over the MWCNTs under dry
conditions are 0.47 and 0.32 (Figure 2B), respectively. Broken
Angew. Chem. Int. Ed. 2011, 50, 10226 –10230
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10229

Communications
[4] K. P. De Jong, J. W. Geus, Catal. Rev. Sci. Eng. 2000, 42, 481.
[5] D. R. Dreyer, C. W. Bielawski, Chem. Sci. 2011, 2, 1233.
[6] a) H.-P. Jia, D. R. Dreyer, C. W. Bielawski, Tetrahedron 2011, 67,
4431 – 4434; b) H.-P. Jia, D. R. Dreyer, C. W. Bielawski, Adv.
Synth. Catal. 2011, 353, 528 – 532; c) D. R. Dreyer, S. Murali, Y.
Zhu, R. S. Ruoff, C. W. Bielawski, J. Mater. Chem. 2011, 21,
3443; d) D. R. Dreyer, S. Park, C. W. Bielawski, R. S. Ruoff,
Chem. Soc. Rev. 2010, 39, 228.
[7] J. Zhang, X. Wang, Q. Su, L. Zhi, A. Thomas, X. Feng, D. S. Su,
R. Schlçgl, K. Mꢁllen, J. Am. Chem. Soc. 2009, 131, 11296 –
11297.
[8] J. Zhang, D. S. Su, A. Zhang, D. Wang, R. Schlçgl, C. Hꢂbert,
Angew. Chem. 2007, 119, 7460 – 7464; Angew. Chem. Int. Ed.
2007, 46, 7319 – 7323.
clarified. The importance of nucleophilic and electrophilic
oxygen species, the prismatic edges as well as the basal planes
of graphite, and the abundance of protons being delivered
either by carboxy groups or by water were highlighted.
Important contributions from ab initio calculations available
for the graphene model is an independent basis for interpre-
tation of the experimental data. We could identify substantial
similarities between the metal oxide and the carbon-catalyzed
reaction, which might help to bridge the gap between these
apparently incompatible catalytic systems.
Experimental Section
[9] J. Zhang, D. S. Su, R. Blume, R. Schlçgl, R. Wang, X. Yang, A.
Catalytic testing: Sample amounts of 500 mg (100–300 mm) were
tested in a fixed-bed quartz reactor under atmospheric pressure. The
reactor was heated at 5 Kminꢀ1 up to 3008C and the feed comprising
5% C3H4O, 10 % O2, and 0–5% H2O in He was dosed by mass-flow
controllers (O2, He) and saturators (H2O, C3H4O) at a total flow of
25 mLminꢀ1 (GHSV of 3000 hꢀ1). Kinetic measurements were con-
ducted at 275–3258C with feed streams of 2–10% C3H4O and 5–20%
O2 in He. Differential conditions were ensured by an increased total
flow of 100 mLminꢀ1, which resulted in C3H4O conversions of < 5%.
The reactants and products in the inlet and outlet streams were
quantified by an on-line gas chromatograph (Varian CP-4900).
Structural characterization: HRTEM analyses were recorded on
a Philips CM 200 LaB6 microscope at an acceleration electron
voltage of 200 kV. The specimens were prepared by suspension of the
nanocarbon powder in ethanol. Drops of the suspensions were
deposited on carbon-enhanced C grids and dried in air. Laser Raman
spectroscopy was performed on powder samples by using an ISA
LabRam instrument equipped with an Olympus BX40 microscope.
The excitation wavelength was 632.8 nm and a resolution of 0.9 cmꢀ1
was used. Specific surface areas were determined by N2 physisorption
at ꢀ1968C and calculated from the adsorption branch in the range of
p/p0 = 0.05–0.3 by the method of Brunauer, Emmet, and Teller (BET).
For XPS and TPD analyses, the samples were cooled down to ambient
temperature in the dry and wet feed, respectively. For TPD, the
sample was directly subjected to a temperature ramp (10 Kminꢀ1) in
25 mLminꢀ1 of He. The molecules released from the surface by
desorption and/or decomposition of surface species up to 8508C were
quantified by an on-line mass spectrometer and an on-line gas
chromatograph, respectively. Synchroton-excited XPS analyses were
performed at the ISISS beamline at BESSY II. The compressed
sample tablets were analyzed under quasi in situ conditions in an
ultrahigh vacuum at 3008C.
´
Gajovic, Angew. Chem. 2010, 122, 8822 – 8826; Angew. Chem.
Int. Ed. 2010, 49, 8640 – 8644.
[10] a) B. Frank, J. Zhang, R. Blume, R. Schlçgl, D. S. Su, Angew.
Chem. 2009, 121, 7046 – 7051; Angew. Chem. Int. Ed. 2009, 48,
6913 – 6917; b) X. Liu, B. Frank, W. Zhang, T. P. Cotter, R.
Schlçgl, D. S. Su, Angew. Chem. 2011, 123, 3376 – 3380; Angew.
Chem. Int. Ed. 2011, 50, 3318 – 3322; c) B. Frank, M. Morassutto,
R. Schomꢃcker, R. Schlçgl, D. S. Su, ChemCatChem 2010, 2,
644 – 648.
[11] B. K. Hodnett in Supported Catalysts and Their Applications
(Eds.: D. C. Sherrington, A. P. Kybett), RSC, 2001, pp. 1 – 8.
[12] T. Ohara, T. Sato, N. Shimizu, G. Prescher, H. Schwind, O.
Weiberg, K. Marten, H. Greim in Ullmann’s Encyclopedia of
Industrial Chemistry, Wiley-VCH, Weinheim, Germany, 2003.
´
[13] J. Tichy, Appl. Catal. A 1997, 157, 363 – 385.
[14] P. Kampe, L. Giebeler, D. Samuelis, J. Kunert, A. Drochner, F.
Haa, A. H. Adams, J. Ott, S. Endres, G. Schimanke, T.
Buhrmester, M. Martin, H. Fuess, H. Vogel, Phys. Chem.
Chem. Phys. 2007, 9, 3577.
¨
[15] B. Frank, A. Rinaldi, R. Blume, R. Schlogl, D. S. Su, Chem.
Mater. 2010, 22, 4462 – 4470.
[16] R. T. Yang, C. Wong, Science 1981, 214, 437 – 438.
[17] M. Sadakane, N. Watanabe, T. Katou, Y. Nodasaka, W. Ueda,
Angew. Chem. 2007, 119, 1515 – 1518; Angew. Chem. Int. Ed.
2007, 46, 1493 – 1496.
[18] O. Shenderova, V. Zhirnov, D. Brenner, Crit. Rev. Solid State
Mater. Sci. 2002, 27, 227 – 356.
[19] a) M. K. Kostov, E. E. Santiso, A. M. George, K. E. Gubbins,
M. B. Nardelli, Phys. Rev. Lett. 2005, 95, 136105; b) S. M. Lee,
Y. H. Lee, Y. G. Hwang, J. R. Hahn, H. Kang, Phys. Rev. Lett.
1999, 82, 217; c) S. C. Xu, S. Irle, D. G. Musaev, M. C. Lin,
J. Phys. Chem. C 2009, 113, 18772 – 18777.
[20] T. V. Andrushkevich, Catal. Rev. - Sci. Eng. 1993, 35, 213 – 259.
[21] B. Frank, R. Fortrie, C. Hess, R. Schlçgl, R. Schomꢃcker, Appl.
Catal. A 2009, 353, 288 – 295.
Received: May 16, 2011
Revised: June 9, 2011
[22] J. Zhang, X. Liu, R. Blume, A. Zhang, R. Schlçgl, D. S. Su,
Science 2008, 322, 73 – 77.
Published online: August 16, 2011
[23] J.-L. Li, K. N. Kudin, M. J. McAllister, R. K. Prud’homme, I. A.
Aksay, R. Car, Phys. Rev. Lett. 2006, 96, 176101.
[24] H.-K. Jeong, Y. P. Lee, M. H. Jin, E. S. Kim, J. J. Bae, Y. H. Lee,
Chem. Phys. Lett. 2009, 470, 255 – 258.
[25] D. R. Dreyer, H.-P. Jia, C. W. Bielawski, Angew. Chem. 2010,
122, 6965 – 6968; Angew. Chem. Int. Ed. 2010, 49, 6813 – 6816.
[26] A. Jelea, F. Marinelli, Y. Ferro, A. Allouche, C. Brosset, Carbon
2004, 42, 3189 – 3198.
[27] C. Park, R. T. K. Baker, J. Phys. Chem. B 1999, 103, 2453 – 2459.
[28] a) X. Xie, Y. Li, Z.-Q. Liu, M. Haruta, W. Shen, Nature 2009, 458,
746 – 749; b) A. Celaya Sanfiz, T. W. Hansen, A. Sakthivel, A.
Trunschke, R. Schlçgl, A. Knoester, H. H. Brongersma, M. H.
Looi, S. B. A. Hamid, J. Catal. 2008, 258, 35 – 43.
ꢀ
Keywords: aldehydes · carbocatalysis · C H activation ·
graphite · oxygen
.
[1] a) S. Iijima, Nature 1991, 354, 56 – 58; b) H. W. Kroto, J. R.
Heath, S. C. O’Brien, R. F. Curl, R. E. Smalley, Nature 1985, 318,
162 – 163.
[2] K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang,
S. V. Dubonos, I. V. Grigorieva, A. A. Firsov, Science 2004, 306,
666 – 669; see also A. K. Geim, Angew. Chem. 2011, 123, 7100 –
7122; Angew. Chem. Int. Ed. 2011, 50, 6966 – 6985; K. S.
Novoselov, Angew. Chem. 2011, 123, 7123 – 7141; Angew.
Chem. Int. Ed. 2011, 50, 6986 – 7002.
[3] P. Serp, J. L. Figueiredo, Carbon Materials for Catalysis, John
Wiley & Sons, 2009.
10230 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 10226 –10230