Full Papers
can give more active catalysts than AC. Our previous study
also showed that the same N species were formed on AC by
the ammoxidation, but their amounts were smaller than those
of the present N-carbon catalysts prepared from PAN.[6] This
might be the reason for the lower catalytic activity of N-carbon
prepared from AC compared to that of the N-carbons prepared
from PAN. The catalytic activity of N-doped carbon nano-
tube[2,3] was also lower than that of PAN-C500-AO400 (Table 3,
entries 3, 4). Other types of solid-base catalysts such as smec-
tite, MgO, and a Mg–Al mixed oxide were also used for the
same reaction.[11,12] They had lower activity than PAN-C500-
AO400, even if the difference of the reaction temperature was
taken into account (entries 1, 5–7). Thus, PAN-C500-AO400 is
the most active among the solid-based catalysts listed in
Table 3.
the graphite structure. X-ray photoelectron spectra indicated
the presence of pyridine-type and pyrrole-/pyridone-type N
species. The catalytic activity of the ammoxidized N-carbon
was correlated with the amount of pyridine type N species,
suggesting that they are involved in the catalytically active
sites; however, such N species on the calcined PAN samples
were practically inactive. On the basis of these observations, it
has been suggested that the presence of pyridine-type N
atoms in the large graphene structure is significant for the
emergence of the catalytic activity. The most active N carbon
prepared from PAN was much more active than other solid-
base catalysts reported so far including N carbon derived from
activated carbon, N-doped carbon nanotubes, and a few inor-
ganic solid-base catalysts.
As listed in Table 2, the surface concentration of N(1) over
PAN-C200-AO600 was higher than that over PAN-C200-C400.
This would be a possible reason for the higher activity of the
former N-carbon catalyst. For N doping on carbon materials,
the dissociation of CÀC bond is considered to be required.[13]
The rate of the dissociation would be faster at higher tempera-
tures, which would promote the N doping. Table 2 also shows
that the amount of N(1) species tended to decrease with in-
creasing calcination temperature and no N species was detect-
ed by XPS over PAN-C500. The calcination at high tempera-
tures would cause the evolution of N atoms, probably as
HCN.[7] Therefore, N atoms originating from the source PAN
polymer would disappear by the calcination at 5008C (PAN-
C500). Despite of this, PAN-C500-AO400 contained the largest
amount of N(1) species and, hence, it was the most active
(Tables 1 and 2). When PAN was calcined at 5008C, approxi-
mately 90% of the PAN sample was burned off. Such a signifi-
cant burn-off did not occur for the calcination of PAN at 200
and at 4008C. Hence, PAN-C500 would contain a larger
amount of deficient sites than PAN-C200 and PAN-C400.
Jansen and Bekkum propose N-doping mechanisms in which
carboxylic groups contribute to the doping[13] and our previous
work showed that O2 molecules promote N-doping of AC
using NH3.[6] Probably, the deficient sites over the calcined PAN
involve O species and some kind of those O species that can
promote N-doping. It is highly probable that the larger
amount of those deficient sites over PAN-C500 resulted in the
larger amount of doped N species. It is difficult, at present, to
precisely depict what reactions occur in the course of the am-
moxidation of the calcined PAN. The ladder structure of the
calcined PAN and the amount and nature of N and O atoms in
it may depend on the calcination conditions. Detailed study on
this issue may give more effective N-carbon catalysts.
Experimental Section
Raw starting carbon materials were prepared from PAN powder
(copolymer containing 3 mol% ethyl vinylacetate) supplied by Mit-
subishi Rayon by the calcination in air at 200, 400, or 5008C for
3 h. Those carbon materials were named as PAN-C200, PAN-C400,
and PAN-C500, respectively. Nitrogen doping on these samples
was performed by ammoxidation. A weighed carbon sample
(ꢀ150 mg) was placed in a quartz reactor and heated to 400 or
6008C in a stream of 90 vol.% NH3–air at 100 cm3 minÀ1. The
carbon sample was treated at this temperature for 1 h, cooled to
3008C, at which the treatment gas was then changed to air, and
further cooled to room temperature. N-carbon samples thus pre-
pared were designated as PAN-C200-AO600, PAN-C500-AO400 etc.,
for which the first and the second numbers represent the tempera-
tures for the calcination and for the ammoxidation, respectively.
The catalytic activity of those N-carbon materials prepared was
tested for Knoevenagel condensation of benzaldehyde with ethyl
cyanoacetate. The reaction was performed in a Teflon-lined auto-
clave (100 cm3) at 808C for 1 h using the catalyst (100 mg), benzal-
dehyde (9.9 mmol), ethyl cyanoacetate (9.4 mmol), and 1-butanol
solvent (4 cm3). After the reaction run, the reaction mixture was
suction-filtered, ethyl benzene was added to the separated liquid
mixture as an internal standard, and the liquid mixture was diluted
with 1-butanol to 25 cm3 and analyzed by gas chromatography
(Shimadzu GC14-B with Zebron ZB-50 column) and mass spectrom-
etry (Shimadzu QP5050A with GL Science TC-17 column). A few re-
action runs were also performed with 30 mg of the catalyst.
The textural properties of the carbon samples prepared were mea-
sured by nitrogen adsorption/desorption (Quantachrome NOVA
1000). The surface area was determined by the BET equation. XPS
measurements of the samples were conducted on JEOL JPS-9200
using monochromatic AlKa radiation. The charge-up shift correc-
tion of the binding energy was done by using C1s binding energy
at 284.5 eV. Diffuse-reflectance FTIR spectra of PAN and PAN-C200
were measured on a JASCO FTIR-620. A spectrum of KBr was used
as the background. The numbers of basic sites on the carbon sam-
ples were determined by acid–base titration. An amount of ap-
proximately 0.1 g of the carbon sample was dispersed in 0.01m
HCl solution (15 cm3) under stirring overnight. Then, the sample
was separated by filtration and the concentration of HCl in the fil-
trate was determined by titration with 0.01m NaOH solution. The
amount of basic sites was determined from the decrease of the
HCl concentration.
Conclusions
N-carbon catalysts were prepared from polyacrylonitrile (PAN)
by calcination and subsequent ammoxidation. The catalytic ac-
tivity of the calcined PAN for the Knoevenagel condensation
reaction was greatly enhanced by the ammoxidation depend-
ing on both the calcination and ammoxidation temperatures.
XRD measurements suggested the occurrence of the growth in
ChemCatChem 2015, 7, 2965 – 2970
2969
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim