Nitrogen-doped carbon materials prepared by ammoxidation as solid base catalysts for knoevenagel condensation and transesterification reactions

Article abstract of DOI:10.1002/adsc.201000029

Nitrogen-doped carbon materials were prepared by ammoxidation of commercial carbon sources (carbon black and activated carbon) and applied as base catalysts for Knoevenagel and transesterification reactions. It was shown that these carbon materials were active and the activities were different depending on the ammoxidation conditions (temperature and ammonia concentration in air) and carbon sources used. The bulk, textural, and surface properties of the nitrogen-doped carbon materials were examined by several methods to clarify possible factors determining their final catalytic activities. The activated carbon-derived catalysts were more active than the carbon black-derived ones. The surface area and porosity were not responsible for this difference between the two carbon sources but the difference in the reactivity with oxygen was important. The reactivity of carbon sources with oxygen should influence the doping of nitrogen onto their surfaces by ammoxidation with ammonia and air and the resulting activities as base catalysts. The catalytic activity increases with the amount of nitrogen doped and, therefore, the nitrogen doped should be responsible for the catalytic activities. In addition, the activities are maximal at a ratio of nitrogen to oxygen of around 1, suggesting the importance of cooperative functions of nitrogen and oxygen on the surface of carbons.

Full text of DOI:10.1002/adsc.201000029

FULL PAPERS
DOI: 10.1002/adsc.201000029
Nitrogen-Doped Carbon Materials Prepared by Ammoxidation
as Solid Base Catalysts for Knoevenagel Condensation and
Transesterification Reactions
Naokatsu Kan-nari,a Shuhei Okamura,b Shin-ichiro Fujita,b Jun-ichi Ozaki,a
and Masahiko Araib,*
a
Department of Chemical and Environmental Engineering, Graduate School of Engineering, Gunma University, Kiryu
376-8515, Japan
b
Division of Chemical Process Engineering, Faculty of Engineering, Hokkaido University, Sapporo 060-8628, Japan
Fax : (+81)-11-706-6594; e-mail: marai@eng.hokudai.ac.jp
Received: January 13, 2010; Revised: May 12, 2010; Published online: June 8, 2010
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/adsc.201000029.
Abstract: Nitrogen-doped carbon materials were difference in the reactivity with oxygen was impor-
prepared by ammoxidation of commercial carbon tant. The reactivity of carbon sources with oxygen
sources (carbon black and activated carbon) and ap- should influence the doping of nitrogen onto their
plied as base catalysts for Knoevenagel and transes- surfaces by ammoxidation with ammonia and air and
terification reactions. It was shown that these carbon the resulting activities as base catalysts. The catalytic
materials were active and the activities were differ- activity increases with the amount of nitrogen doped
ent depending on the ammoxidation conditions (tem- and, therefore, the nitrogen doped should be respon-
perature and ammonia concentration in air) and sible for the catalytic activities. In addition, the activ-
carbon sources used. The bulk, textural, and surface ities are maximal at a ratio of nitrogen to oxygen of
properties of the nitrogen-doped carbon materials around 1, suggesting the importance of cooperative
were examined by several methods to clarify possible functions of nitrogen and oxygen on the surface of
factors determining their final catalytic activities. carbons.
The activated carbon-derived catalysts were more
active than the carbon black-derived ones. The sur- Keywords: ammoxidation; base catalysts; carbon;
face area and porosity were not responsible for this Knoevenagel condensation; nitrogen doping; transes-
difference between the two carbon sources but the terification
Introduction
materials so prepared for Knoevenagel reactions.
Those results indicate the attractive potential of the
Recently several authors have reported the prepara- preparation of carbon materials having various chemi-
tion and design of interesting functional carbon-based cal and electrochemical functions. A recent communi-
materials. Toda et al. successfully made a solid acid cation has discussed possible functions of nitrogen-
catalyst of amorphous carbon materials in which a containing carbon alloys,[5] in which different modes
sulfonite group was introduced.[1] Ozaki et al. tailored of carbon bonding coexist in the absence and pres-
nitrogen- and/or boron-doped carbon materials via ence of other foreign atoms like nitrogen. In addition,
controlled carbonization of polymers, which were ef- previous XPS studies showed the presence of various
fective as the electrode in proton exchange membrane types of nitrogen species in nitrogen-doped carbon
fuel cells.[2] Nitrogen-doped carbon materials were materials, which could have different chemical func-
also prepared by Favia et al.[3] using cold plasma treat- tions.[6] Thomas et al. prepared graphitic carbon ni-
ment with ammonia and oxygen and by Bitter et al.[4] tride materials using a well-designed synthetic chemis-
using chemical vapor deposition of compounds such try, in which the materials were made by polymeri-
as acetonitrile and pyridine containing both carbon zation of cyanamide, dicyanamide, or melamine.[7]
and nitrogen on supported metal particles. The latter Various materials with different structures and prop-
authors used their nitrogen-doped carbon nanotube erties can be obtained depending on reaction condi-
1476
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2010, 352, 1476 – 1484

Nitrogen-Doped Carbon Materials Prepared by Ammoxidation as Solid Base Catalysts
tions. These carbon nitride materials are active cata- gen adsorption/desorption, XPS, XRD, and so on,
lysts for Friedel–Crafts and cyclization reactions. Very from which the quantity and chemical state of the ni-
recently, Jin et al. have prepared carbon nitride nano- trogen doped and the catalytically active working
particles using mesoporous silica nanoparticles as a sites have been discussed.
template and indicated that these are active catalysts
for transesterification reactions.[8]
Those previous results demonstrate that nitrogen- Results and Discussion
containing carbon materials can be metal-free active
catalysts for electrochemical and organic synthetic re- Catalytic Activities
actions. The synthesis and catalytic application of ni-
trogen-modified carbon materials is a challenging task The nitrogen-doped carbon materials prepared were
with scientific and industrial significance. As known applied for base-catalyzed Knoevenagel and transes-
from the above-mentioned works, there are three ap- terification reactions (Scheme 1). The influence of
proaches to the preparation of nitrogen-containing
carbon materials, which are (i) synthetic chemistry
using small building-unit molecules, (ii) carbonization
of nitrogen-containing macromolecules, and (iii)
doping of nitrogen to bulk carbon materials. In the
present work, our attention is paid to the last method
in which the doping procedures are easy. A useful
method of nitrogen doping is amination or ammoxi-
dation, in which a carbon sample is thermally treated
with ammonia or a mixture of ammonia and air.[9] It
was reported quite early by Rideal and Wright that
nitrogen-containing carbon materials so prepared
were active for the oxidation of aqueous oxalic
acid.[9a] Stçhr et al. reported that the formation of
H2O from H2 and O2 was catalyzed by carbon materi-
als activated by thermal treatment with ammonia.[9b]
Scheme 1. Base-catalyzed reactions studied. ECC=ethyl cy-
anocinnamate, MA=methyl acetate.
Other authors also indicated previously the catalytic
potentials of nitrogen-doped carbon materials treated
with ammonia as fuel cell cathodes and catalysts for
the oxidation of oxalic acid and others.[9c–e] To the ammoxidation conditions such as treatment gas and
best of our knowledge, however, only a few works[4,7,8] heating schedule was first examined using CB as a
have been reported so far to apply such nitrogen- carbon source and Knoevenagel condensation as a
doped carbon materials prepared via ammoxidation test reaction. Table 1 presents the catalytic activities
and other methods for organic synthetic reactions.
of CB-based materials prepared under different con-
Considering the previous results and the theoretical ditions. The raw CB material was heat-treated with
expectation of interesting chemical functions of nitro- gas (1) (NH3 +air, NH3, or N2) at increasing tempera-
gen-doped carbon materials, the present authors have tures up to 6008C, kept at this temperature for 1 h,
prepared nitrogen-doped carbon materials via am- cooled to a certain switching temperature at which
moxidation of active carbon and carbon black under the gas was changed from gas (1) to gas (2) (N2 or
different conditions and examined their potential as air), and further cooled down to room temperature.
solid base catalysts, as a part of a study on the prepa- When the whole treatment was made with pure N2
ration of new solid base catalysts.[10] Base-catalyzed alone, a low ethyl cyanocinnamate (ECC) yield of 8%
chemical reactions are of practical importance to the was obtained (entry 1). The reaction was found to
production of drugs, fragrances, and chemical inter- proceed in the absence of catalyst, giving a very small
mediates.[11] The use of solid base catalysts is currently ECC yield of 5%. So, the raw CB was less active for
recommended, as compared with homogeneous ones, the present Knoevenagel reaction. When pure NH3
because of their easy separation and recycling proce- was used for gas (1), the catalytic activity was en-
dures. The safe nature and low cost of carbon materi- hanced and the ECC yield was larger for a lower
als are also practically important. The ammoxidized switching temperature (entries 2–5). The lowest tem-
carbon materials have been shown to be active for perature of 258C gave the largest yield of 50%,
such reactions as Knoevenagel condensation and which, however, should be ascribable to the presence
transesterification reactions.[12] The structural features of NH3 itself physisorbed on the carbon. When air
of the nitrogen-doped carbon catalysts prepared have was used during the cooling from 6008C to room tem-
been examined by different methods including nitro- perature, an ECC yield of 14% was obtained
Adv. Synth. Catal. 2010, 352, 1476 – 1484
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1477

Naokatsu Kan-nari et al.
FULL PAPERS
Table 1. Influence of ammoxidation conditions on the activi-
ty of CB-based catalysts for Knoevenagel condensation.[a]
Entry
Ammoxidation conditions[b]
Gas (1)
ECC
Temp. [8C] Gas (2) yield [%]
1
2
3
4
N2
600
600
300
100
25
N2
N2
N2
N2
N2
air
air
air
air
air
8
NH3
NH3
NH3
NH3
NH3
NH3
12
13
19
50
14
21
24
27
46
5
6
600
600
7[c]
8
NH3 (90%)+air 600
NH3 (90%)+air 300
NH3 (90%)+air 100
9
10
[a]
For reaction conditions, see the text.
[b]
A carbon sample was heat-treated with a stream of gas
(1) at 100 cm3 minÀ1 and at increasing temperatures up to
6008C, kept at 6008C for 1 h, and cooled to the tempera-
ture given. Then, the treatment gas was switched with a
stream of gas (2) and cooled down to room temperature.
A carbon sample was heat-treated in air at increasing
temperatures up to 6008C, with NH3 at this temperature
for 1 h, and then cooled in air to room temperature.
[c]
Figure 1. Results of (a) Knoevenagel condensation and (b)
transesterification reaction using nitrogen-doped AC- and
CB-derived carbon catalysts prepared via ammoxidation at
different temperatures (NH3 concentration 90%, ammoxida-
tion time 1 h).
(entry 6), which was comparable to that obtained with
N2 for gas (2) (entry 2). When the CB was treated
with air instead of NH3 at increasing temperatures up
to 6008C, the ECC yield increased from 12% to 21%
(entries 2, 7). When NH3 (90%)+air was used for gas
(1) and air for gas (2), the ECC yield was significantly
enhanced (entries 2/8, 3/9, 4/10). These results indi- yield appeared at 7008C for the former samples but
cate an important fact that the coexistence of O2 (in at a lower temperature of 6008C for the latter ones.
air) with NH3 is required to prepare more active The maximum yield was two-fold larger than that
carbon-based catalysts. The authors speculate that the with the sample prepared at 4008C for either the AC
À
À
O2 molecules break chemical bonds such as C C, C
O, and/or C H at the surface of carbon and assist the reaction giving a methyl acetate (MA) yield of 10%
doping of nitrogen from NH3.
or the CB carbon source. For the transesterification
À
without catalyst, the product yield simply increased
The influence of ammoxidation conditions was fur- with temperature less significantly for the AC-derived
ther examined with the two carbon sources CB and samples compared with the results of Knoevenagel
AC, which were treated with NH3 and/or air at differ- condensation. The CB-derived sample prepared at
ent temperatures. The treatment gas mixture was 4008C was inactive for transesterification but those
changed to air at a switching temperature of 3008C. prepared at higher temperatures had similar catalytic
This switching temperature was selected to avoid the activities. The CB and AC carbon materials took cata-
influence of physisorbed NH3 residues on the carbon lytic activities for the base-catalyzed reactions by am-
materials. When the samples prepared with switching moxidation and the effects of ammoxidation tempera-
temperatures of 1008C and 3008C were heat-treated ture were different depending on the starting carbon
at a rate of 158C minÀ1 in an argon stream, the de- sources and the chemical reactions. The difference in
sorption of NH3 was measured by a mass spectrome- the reactivity to O2 between the two raw carbon ma-
ter and observed to occur at temperatures of 100– terials, which will be described later, may be responsi-
4008C for the former sample but not for the latter. ble; this difference should be crucial for the doping of
Figure 1 gives the results of Knoevenagel and transes- nitrogen and the generation of catalytically active
terification reactions with the carbon materials am- sites. In addition, the difference in the temperature ef-
moxidized at different temperatures. For the former fects between Knoevenagel and transesterification re-
reaction for which a 5% ECC yield was obtained in actions suggests that the structural features of work-
the absence of catalyst, the AC-derived samples were ing active sites are different between the two reac-
more active than the CB-derived ones. A maximum tions.
1478
asc.wiley-vch.de
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2010, 352, 1476 – 1484

Nitrogen-Doped Carbon Materials Prepared by Ammoxidation as Solid Base Catalysts
Figure 2 shows the influence of NH3 concentration
at a treatment temperature of 6008C. When the NH3 dized with 90% NH3 at 6008C for different periods of
concentration was 90% or lower, the product yield time and longer treatments were observed to be less
effective (see Supporting Information, Figure S1).
The amoxidation treatments for 2 h and 1 h were suf-
ficient to activate the CB and AC samples, respective-
ly. This difference would be explained by a difference
in the reactivity with O2 between the two carbon sour-
ces, as will be discussed later with thermogravimetric
results.
The CB and AC carbon materials were ammoxi-
One of merits of using heterogeneous catalysts is
their possibility of easy separation and recycling. So,
the recycling test was conducted with a CB-derived
carbon sample prepared by ammoxidation at 6008C.
After the first reaction run, the catalyst was separated
by suction filtration, dried in air at 1208C overnight,
and used for the next run. For transesterification reac-
tion, the CB-based carbon catalyst was able to be re-
cycled without significant loss of activity for three re-
peated runs (see Supporting Information, Figure S2).
For Knoevenagel condensations, unfortunately, the
catalyst was observed to lose its activity for the
second run. When the catalyst was further treated in
an N2 stream at 3008C for 1 h after the drying at
1208C, its initial activity was recovered by 50% but
not completely. The change of product yield with re-
action time indicated the loss of activity (see Support-
ing Information, Figure S3). A possible factor for this
catalyst deactivation was water formed during the re-
Figure 2. Results of (a) Knoevenagel condensation and (b)
transesterification reaction using nitrogen-doped AC- and
action (Scheme 1). However, this was unlikely be-
cause the activity was not completely recovered by
CB-derived carbon catalysts prepared via ammoxidation
heat treatment at a high temperature of 3008C as
with different NH3 concentrations (ammoxidation tempera-
mentioned above and the activity was not found to be
ture 6008C, time 1 h).
influenced by addition of a porous material like zeo-
lite as a water catcher in the reaction mixture. Anoth-
er possible factor was benzoic acid, which would form
did not change with the concentration so much for by oxidation of benzaldehyde and poison the active
the two reactions. Namely, the ammoxidized carbon base sites. This was also unlikely because no positive
materials had similar catalytic activities irrespective effects appeared even when the liquid reaction mix-
of the NH3 concentration used. For Knoevenagel con- ture was pretreated by bubbling with an argon stream
densation with the catalyst samples prepared with to remove the dissolved O2 and the reaction was run
60% and 70% NH3, a small amount of benzoic acid in an argon atmosphere using a glove-box. At present,
(<2.5%) was detected to form along with a trace the catalyst deactivation during the Knoevenagel re-
amount of benzaldehyde dibutyl acetal. It was likely action cannot be explained satisfactorily. It is specu-
that the presence of 40–30% O2 caused the produc- lated that a small quantity of the product ECC is ad-
tion of O-containing surface functional groups on the sorbed on and poisons the active sites of the nitrogen-
carbon materials facilitating the oxidation of benzal- doped carbon catalysts. Another possibility is the
dehyde to benzoic acid and this acid catalyzed the for- base-catalyzed hydrolysis of the ester yielding an acid,
mation of acetal from benzaldehyde and 1-butanol which could then deactivate the active basic sites.
(solvent). At a 100% of NH3, however, the product
yields were decreased, indicating again the impor-
tance of the coexistence of O2. An interesting excep- Bulk and Textural Properties
tion was the AC-based sample treated with 100%
NH3 for the transesterification, which indicated the The changes of bulk and textural properties of CB
highest activity. These results also suggest that Knoe- and AC carbon materials with the ammoxidation
venagel and transesterification reactions are catalyzed were examined. The weight loss and the surface area
by different active sites.
of the carbon samples were measured after the am-
Adv. Synth. Catal. 2010, 352, 1476 – 1484
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1479

Naokatsu Kan-nari et al.
FULL PAPERS
moxidation with 90% NH3 at different temperatures
for 1 h (see Supporting Information, Figure S4). The
ammoxidation caused the weight loss of the CB- and
AC-derived samples, which were about 5% and 15%,
at 6008C. The weight loss was larger for the latter
than the former. The weight decreased by more than
20% on the ammoxidation at 8008C. The surface
areas of the raw CB and AC samples were 1400 m2 gÀ1
and 1200 m2 gÀ1, respectively, and did not change so
much on ammoxidation at 6008C and 8008C. The sur-
face areas of the AC-derived samples were smaller
than those of the CB-derived ones although the
former were more active than the latter for Knoeve-
nagel and transesterification reactions. The total pore
volume of the raw AC material was 0.58 cm3 gÀ1,
which was found to increase to 0.62 and 0.70 cm3 gÀ1
on the ammoxidation at 6008C and 8008C, respective-
ly. The raw CB material had a larger pore volume of
3.26 cm3 gÀ1, changing to 3.28 and 3.57 cm3 gÀ1 for the
samples prepared at 6008C and 8008C, respectively.
Furthermore, it was found that the pore size distribu-
tion did not change so much with the ammoxidation
for either CB or AC carbon material. Similar results
were reported by Jansen and van Bekkum[9g]; these
Figure 3. XPS spectra of AC- (a) and CB (b)-based catalysts
prepared by ammoxidation with 90% NH3 for 1 h at the dif-
ferent temperatures given.
authors modified carbon materials by amination and for the carbon materials ammoxidized with 90% NH3
ammoxidation at temperatures between 2008C and at different temperatures. The amount of nitrogen
4208C and observed that their pore structures were doped was estimated by measuring the total area of
virtually unchanged by these treatments. Those tex- the N 1s spectrum obtained. In Figure 4 the product
tural features are unable to explain the significant dif- yield of either Knoevenagel or transesterification re-
ferences in the catalytic activity among the various ni- actions is plotted against the amount of N relative to
trogen-doped carbon catalysts prepared, which C. It is clear that a positive correlation exists between
depend on the ammoxidation conditions and carbon the N/C ratio and the product yield, that is, the cata-
sources used (Figure 1 and Figure 2). The chemical lytic activity for either reaction. The doping of nitro-
properties of their surfaces should be responsible for gen is responsible for the catalytic activity of ammoxi-
their base catalysis, as will be discussed later.
dized CB and AC materials. A wide range XPS analy-
The reactivity of the two CB and AC carbon sour- sis was also made for the CB and AC carbon samples
ces with O2 was examined by thermogravimetric anal- but we did not detect the presence of metal species
ysis in which the sample was heated in air at a rate of such as Fe, which had a possibility of playing some
108C minÀ1 up to 8008C. The weight loss was ob- role in the nitrogen doping through oxidation of
served to start at 5508C for the former sample but at carbon.
a lower temperature of 4508C for the latter (not
The XPS data (Figure 3) indicated the presence of
shown). That is, the AC was more reactive with O2 several nitrogen species of different electronic states
than the CB, which should be crucial for determining (Scheme 2). The relative amounts of these different
the easiness of nitrogen doping on ammoxidation, the nitrogen species were estimated by deconvolution of
structure of nitrogen-containing active sites formed, the spectra. Figure 5 shows the results for the AC-
and thus the final catalytic activities. The amount of and CB-derived samples ammoxidized with 90% NH3
nitrogen doped was observed to be larger for the AC at different temperatures. The assignment of the bind-
compared to the CB, as described later (see Figure 4). ing energies to the corresponding chemical states of
nitrogen was made according to previous work.[6a,13]
Although the activities of those nitrogen-doped
Surface Chemical Properties
carbon catalysts were different as mentioned above,
they were very similar in the distribution of the differ-
The ammoxidized carbon samples were examined by ent nitrogen species. The fraction of pyridine-type ni-
XPS and it was confirmed that nitrogen was doped trogen (398.5 eV) was comparable to that of pyrrole-/
onto the CB and AC materials and the extent of ni- pyridone-type nitrogen (400.5 eV) each in about 40%
trogen doping depended on ammoxidation conditions with small amounts of quaternary (401.5 eV) and oxi-
and carbon sources used. Figure 3 gives XPS traces dized (402.5 eV)
N
species and adsorbed NOx
1480
asc.wiley-vch.de
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2010, 352, 1476 – 1484

Nitrogen-Doped Carbon Materials Prepared by Ammoxidation as Solid Base Catalysts
Figure 4. Plot of the product yield in (top) Knoevenagel con-
densation and (bottom) transesterification reaction against
the N content for nitrogen-doped AC- (open circle) and
Figure 5. Distributions of
N species of (top) AC- and
CB-derived (closed circle) carbon catalysts.
(bottom) CB-derived carbon catalysts prepared by ammoxi-
dation with 90% NH3 at the different temperatures given.
1=pyridine-type; 2=pyrrole/pyridone-type; 3=quaternary-
type, 4=oxidized-type; 5=adsorbed NOx.
(405.0 eV). The different states of nitrogen doped
should be unlikely to indicate the same catalytic activ-
ity.[4,5,14] At present, unfortunately, it was impossible
to specify the type of nitrogen doped that is involved gen is lower than those of the other types of nitrogen,
in the catalytically active sites. Previously Bitter et al. implying that the density of electrons is smaller and
reported, for nitrogen-containing carbon nanotubes the Lewis basicity would be weaker. We cannot deny
prepared by chemical vapor deposition of C- and N- a possibility that the other types of nitrogen than the
containing precursors (acetonitrile, pyridine), that the pyridine-type one may also contribute to the basic
initial activity in Knoevenagel condensations in- active sites for Knoevenagel and transesterification
creased with the amount of pyridine-type nitrogen in- reactions. Previously the present authors studied the
cluded.[4] The binding energy of pyridine-type nitro- same transesterification reaction with mesoporous
smectite-like catalysts and reported that the reaction
was catalyzed by basic sites of medium strength,
which was examined by temperature programmed de-
sorption of CO2.[10] It was further found that the nitro-
gen-doped carbon materials were inactive for Knoe-
venagel condensation between benzaldehyde and di-
ethyl malonate CH2ACHTUNGTRNE(NUG CO2C2H5)2 instead of ethyl cya-
noacetate CH2(CN)CO2C2H5 under the present reac-
tion conditions. The abstraction of a proton from the
methylenic group is the first and rate-determining
step[12a] and it may be more difficult for diethyl malo-
nate than for ethyl cyanoacetate. This indicates that
Scheme 2. Possible structures of nitrogen doped to carbon.
Adv. Synth. Catal. 2010, 352, 1476 – 1484
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1481

Naokatsu Kan-nari et al.
FULL PAPERS
such strong basic sites are absent in the nitrogen- nitrogen-doped catalyst at an N/O ratio of around 1
doped carbon catalysts prepared.
(Figure 6).
An interesting relationship was found when the ac-
The above discussion is based on a possible struc-
tivity for either Knoevenagel condensation or transes- ture of carbon surfaces as depicted in Scheme 2, in
terification reaction was plotted against the N/O which different states of doped nitrogen species are
atomic ratio of the nitrogen-doped carbon catalysts distinguished by XPS. It was previously reported that
(Figure 6). For both reactions, the maximum activities there existed various functional groups on the surface
of carbon materials such as carbonyl, carboxylic anhy-
dride, lactonic, quinoline-like, xanthene, and ether
groups.[15] These functional groups would exist on the
present starting carbon materials and be formed and
changed by the ammoxidation. These oxygen-contain-
ing groups should be considered for more a detailed
discussion on the mechanisms of nitrogen doping and
the catalytic actions of doped nitrogen species and
neighbor functional groups in organic synthetic reac-
tions.
To further give light on the features of our nitro-
gen-doped carbon catalysts, the same reactions were
conducted with pyrrole, pyridine and piperidine simi-
lar to their structural moieties (Scheme 2). The active
AC-derived sample ammoxidized with 90% NH3 at
6008C for 1 h was selected, which was doped with a
larger amount of nitrogen. This carbon sample con-
tained nitrogen in 6.6 mol% as estimated from XPS
and so 0.55 mmol nitrogen was included in a 0.1 g of
the sample. Similar molar amounts of pyrrole, pyri-
dine and piperidine were used for the reaction runs
under the identical conditions. Table 2 shows that pyr-
role and pyridine were almost inactive for the Knoe-
venagel condensation and transesterification while pi-
peridine was active for both reactions. The activity of
the nitrogen-doped carbon catalyst was comparable to
Figure 6. Plots of the product yield in Knoevenagel conden-
sation (a) and transesterfication reaction (b) against the N/
O ratio for the AC- (open circles) and CB- (closed circles)
derived nitrogen-doped carbon catalysts
Table 2. Comparison of a nitrogen-doped carbon catalyst
with homogeneous catalysts of pyridine and piperidine.
Entry Catalyst
ECC yield MA yield
appeared at an N/O ratio of 0.8–1.0, indicating the im-
portance of the presence of the same amounts of sur-
face N and O species. The authors believe that there
operates some cooperation of N- and O-containing
sites, which promotes these two reactions. A possible
model for such cooperation is the pyridone-type of ni-
trogen depicted in Scheme 2, in which OH group is
attached to the C atom next to the N atom of the pyr-
idone ring. The H of the OH group may form hydro-
gen bonding with the negatively polarized O atom of
the carbonyl group of one substrate (benzaldehyde,
ethyl acetate) and the hydrogen bonding is likely to
increase the positive nature of the C atom. A carban-
ion would be produced from the counterpart substrate
(ethyl cyanoacetate, methanol) by the catalytic action
of basic N sites. The cooperation of the hydrogen
bonding and the basic N site would promote the nu-
cleophilic attack of the carbanion to the carbonyl sub-
strate. This could explain the maximum activity of the
[%][a]
[%][b]
1
2
3
4
N-doped carbon[c]
pyrrole[d] (pKa 0.4)[e]
50
9
49
–
[i]
pyridine[f] (pKa 5.25-5.5)[e] 15
8
50
piperidine[g] (pKa 11)[e]
89[h]
[a]
[b]
[c]
Yield of ethyl cyanocinnamate in Knoevenagel condensa-
tion at 808C for 1 h.
Yield of methyl acetate in transesterification at 1408C
for 4 h.
AC-derived material ammoxidized with 90% NH3 at
6008C. 0.1 g.
[d]
[e]
[f]
0.038 cm3.
Refs.[12b,16]
0.044 cm3.
0.054 cm3.
[g]
[h]
The substrate ethyl cyanoacetate was completely con-
sumed but a by-product of butyl cyanocinnamate was
also formed.
[i]
Not measured.
1482
asc.wiley-vch.de
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2010, 352, 1476 – 1484

Nitrogen-Doped Carbon Materials Prepared by Ammoxidation as Solid Base Catalysts
Table 3. Results of Knoevenagel condensation reactions of ethyl cyanoacetate with different carbonyl substrates.[a]
Substrate
Benzaldehyde
60
Butyraldehyde
55
Acetophenone
Benzophenone
Yield[b] [%]
–
–
[c]
[c]
[a]
AC derived catalyst prepared via ammoxidation with 90% NH3 at 7008C for 1 h was used. Reaction conditions: catalyst
100 mg; ethyl cyanoacetate 9.4 mmol; carbonyl substrate 9.9 mmol; solvent 1-butanol 4 cm3; temperature 808C; time 1 h.
Knoevenagel coupling products.
[b]
[c]
No reaction occurred.
that of the piperidine for transesterification. It was moxidation conditions (temperature and NH3 concen-
also active for Knoevenagel condensation but less tration) and carbon sources used. The doping of nitro-
active than the piperidine. It is important, however, gen is promoted by the coexistence of oxygen with
that such an undesired by-product as butyl cyanocin- NH3 and the higher reactivity of parent carbon mate-
namate was not formed with the nitrogen-doped rials to oxygen.
carbon catalyst. Thus, one can say that the nitrogen-
The textural properties (surface area and porosity)
doped carbon materials will function as effective basic are unimportant for determining the activities of the
catalysts comparable or similar to the molecular basic ammoxidized carbon catalysts. The ratio of surface N/
catalyst of piperidine.
O species is important and the maximum activities
Finally, the performance of nitrogen-doped carbon appear at the ratio of 0.8–1.0, suggesting cooperation
catalyst was further tested for Knoevenagel condensa- of N- and O-containing sites for the two reactions.
tion reactions of ethyl cyanoacetate with different al-
The ammoxidation of carbon materials is an inter-
dehydes and ketones under the same conditions. esting route for the production of metal-free carbon-
Table 3 presents the results with benzaldehyde, butyr- based catalysts for organic synthetic reactions.
aldehyde, acetophenone, and benzophenone using an
AC-derived catalyst prepared through ammoxidation
with NH3 90% at 7008C for 1 h. The nitrogen-doped Experimental Section
catalyst was also active for butyraldehyde, similar to
benzaldehyde, in which no by-product was observed. Nitrogen Doping to Carbon Materials
However, it was inactive for acetophenone and ben-
Two different raw carbon materials, carbon black (CB,
zophenone. It is known that ketones are generally less
Ketjen Black EC600 JD) and activated carbon (AC, GL Sci-
reactive than aldehydes.[12a]
ence 60/80 mesh), were used. A weighed (about 150 mg)
The present results demonstrate that the ammoxi-
dation of carbon materials is an attractive method for
the production of metal-free carbon-based catalysts
for organic synthetic reactions. The use of carbon ma-
terials has several merits for green chemistry due to
their chemical/physical natures. Further studies are in
progress in our laboratories to apply the ammoxidized
carbon catalysts for other reactions (oxidation, for ex-
ample) and to examine nitrogen-doped catalysts pre-
pared from other carbon sources different in the
structure compared to AC and CB used in the present
work.
carbon sample was placed in a quartz reactor and heated to
400–8008C (mainly 6008C) in a stream of gas (1) (NH3 +air,
NH3, or N2) at a flow rate of 100 cm3 minÀ1. The sample was
treated at this temperature for 1 h and cooled to a certain
temperature, at which the treatment gas was changed to gas
(2) (N2 or air), and further cooled down to room tempera-
ture. The doping of nitrogen to the carbon materials was
confirmed by X-ray photoelectron spectroscopy.
Catalytic Activity Measurements
The catalytic activity of those nitrogen-doped carbon sam-
ples prepared was tested for Knoevenagel condensation and
transesterification reactions using a Teflon-lined autoclave
(50 cm3) under conditions: for the former, benzaldehyde
9.9 mmol, ethyl cyanoacetate 9.4 mmol, solvent 1-butanol
4 cm3, catalyst 100 mg, temperature 808C, time 1 h; for the
latter, ethyl acetate 10 mmol, ethanol 200 mmol, catalyst
100 mg, temperature 1408C, time 4 h. After the reaction
run, the reaction mixture was suction filtered, toluene was
added to the separated liquid mixture as an internal stan-
dard, and the liquid mixture was analyzed by gas chroma-
tography (Shimadzu GC14-B with Zebron ZB-50 column)
and mass spectrometry (Shimadzu QP5050 A with GL Sci-
ence TC-17 column). To examine the catalyst recycling, the
separated solid carbon catalyst was dried in air at 1208C
Conclusions
The following conclusions are drawn from the present
work on the carbon materials that were chemically
modified by ammoxidation with NH3 and air.
The ammoxidized carbon materials are active cata-
lysts for base-catalyzed Knoevenagel condensation
and transesterification reaction.
The doping of nitrogen onto the carbon surfaces is
responsible for the catalytic activities for these reac-
tions. The amount of nitrogen doped depends on am- overnight and then used for the next reaction run.
Adv. Synth. Catal. 2010, 352, 1476 – 1484
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1483

Naokatsu Kan-nari et al.
FULL PAPERS
[7] For a review: A. Thomas, A. Fischer, F. Goettmann, M.
Antonietti, J.-O. Mꢂller, R. Schlçgl, J. M. Carlsson, J.
Mater. Chem. 2008, 18, 4893, and references cited
therein, including the original works of these authors.
[8] X. Jin, V. V. Balasubramanian, S. T. Selvan, D. P.
Sawant, M. A. Chari, G. Q. Lu, A. Vinu, Angew. Chem.
2009, 121, 8024; Angew. Chem. Int. Ed. 2009, 48, 7884.
[9] a) E. K. Rideal, W. M. Wright, J. Chem. Soc. (London)
1926, 1813; b) B. Stçhr, H. P. Boehm, R. Schlçgl,
Carbon 1991, 29, 707, and references cited therein; c) J.
Mrha, Collect. Czech. Chem. Commun. 1996, 31, 715;
d) H. P. Boehm, G. Mair, T. Stçhr, A. R. de Rincꢁn, B.
Tereczki, Fuel 1984, 63, 1061; e) H. P. Boehm, A. R. de
Rincꢁn, T. Stçhr, B. Tereczki, A. Vass, J. Phys. Chim.
Biol. Paris 1987, 84, 1449; f) P. Vinke, M. van der Etjk,
M. verbree, A. F. Volskamp, H. van Bekkum, Carbon
1994, 32, 675; g) P. J. J. Jansen, H. van Bekkum, Carbon
1994, 32, 1507; h) P. Pietrzak, H. Wachowska, P. Now-
icki, K. Babel, Fuel Process. Technol. 2007, 88, 409.
[10] S. Fujita, B. M. Bhanage, D. Aoki, Y. Ochiai, N. Iwasa,
M. Arai, Appl. Catal. A: Gen. 2006, 313, 151.
[11] a) H. Hattori, Chem. Rev. 1995, 95, 537; b) K. Tanabe,
W. Hçlderich, Appl. Catal. A: Gen. 1999, 181, 399;
c) G. Busca, Ind. Eng. Chem. Res. 2009, 48, 6486.
[12] a) R. A. Sheldon, H. van Bekkum, Fine Chemicals
through Heterogeneous Catalysis, Wiley-VCH, Wein-
heim, 2001; b) J. Clayden, N. Greeves, S. Warren, P.
Wothers, Organic Chemistry, Oxford University,
Oxford, 2001; c) A. Corma, S. Iborra, Adv. Catal. 2006,
49, 239; d) J. J. Li, (Ed.), Name Reactions for Homolo-
gation, Part 1, Wiley, Hoboken, New Jersey, 2009; e) R.
Ballini (Ed.), Eco-Friendly Synthesis of Fine Chemicals,
Royal Society of Chemistry, Cambridge, 2009.
Characterization of Carbon-Based Catalysts
The textural properties of the carbon samples so prepared
were measured by nitrogen adsorption/desorption (Quan-
tachrome NOVA 1000). X-ray photoelectron spectroscopy
(XPS) (Perkin–Elmer EACA-5600) was used to examine
C1s, N1s, and O1s core levels for the nitrogen-doped carbon
samples. The charge-up shift correction of binding energy
was made by setting the C1s binding energy of the carbon-
ized sample at 284.5 eV. No sample pretreatment was made
before the XPS measurements.
Acknowledgements
The authors are thankful to Mr. Katsuaki Yamada of Hok-
kaido University for his kind assistance in collecting chemical
reaction data.
References
[1] M. Toda, A. Takagaki, M. Okamura, J. N. Kondo, S.
Hayashi, K. Domen, M. Hara, Nature 2005, 438, 178.
[2] a) J. Ozaki, S. Tanifuji, N. Kimura, A. Furuichi, A. Oya,
Carbon 2006, 44, 1298; b) J. Ozaki, T. Anahara, N.
Kimura, A. Oya, Carbon 2006, 44, 3348; c) J. Ozaki, N.
Kimura, T. Anahara, A. Oya, Carbon 2007, 45, 1847;
d) S. M. Lyth, Y. Nabae, S. Moriya, S. Kuroki, M. Kaki-
moto, J. Ozaki, S, Miyata, J. Phys. Chem. C 2009, 113,
20148.
[3] P. Favia, N. De Vietro, R. Di Mundo, F. Fracassi, R.
d’Agostino, Plasma Processes Polym. 2006, 3, 66.
[4] S. van Dommele, K. P. de Jong, J. H. Bitter, Chem.
Commun. 2006, 4859.
[13] S. Maldonado, S. Morin, K. J. Stevenson, Carbon 2006,
44, 1429.
[14] J. de D. Lopez-Gonzalez, A. Lopez-Peinado, R. M.
Martin-Aranda, M. L. Rojas-Cervantes, Carbon 1993,
31, 1231.
[5] T. Ikeda, M. Boero, S. Huang, K. Terakura, M. Oshima,
J. Ozaki, J. Phys. Chem. C 2008, 112, 14706.
[6] a) E. Raymundo-PiÇero, D. Cazorla-Amorꢁs, A. Li-
nares-Solano, J. Find, U. Wild, R. Schlçgl, Carbon
2002, 40, 597; b) M. Kawaguchi, T. Kawashima, T. Na-
kajima, Chem. Mater. 1996, 8, 1197; c) M. N. Oliveira,
A. M. Botelho do Rego, O. Conde, Surf. Coat. Technol.
1998, 100–101, 398.
[15] a) H. P. Boehm, Carbon 1994, 32, 759; b) B. K. Prad-
han, N. K. Sandle, Carbon 1999, 37, 1323; c) I. I.
Salame, T. J. Bandosz, J. Colloid Interface Sci. 2001,
240, 252.
[16] J. McMurry, Organic Chemistry, 5th edn., Brooks/Cole,
Pacific Grove, 2000.
1484
asc.wiley-vch.de
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2010, 352, 1476 – 1484