L. Liu et al.
Molecular Catalysis 505 (2021) 111504
indicated that most of the metallic Co NPs are well embedded in the
graphite layer. Besides, some defects were observed on the carbon layer
at the periphery of the Co active component (Fig. 3e, f), and these defects
can serve as reaction channels, thereby promoting the catalytic reaction
[21].
The XPS survey of Co@CN-x confirmed the existence of Co, C, N, and
O elements in these catalysts (Fig. S1). The content of Co from XPS is
lower than that from by ICP-AES, and the content of C and N obtained by
XPS is higher than that obtained by an organic element analyzer
(Table S2). This further supported the results obtained by TEM, that is,
Co species were coated inside the nitrogen-doped carbon material. The
C1s spectra of Co@CN-x (Fig. S2) showed that, the C1s peak becomes
more and more sharp with the increase of pyrolysis temperature. The
peak splitting results further confirmed that the composition of gra-
phene carbon gradually increases (Table S3). This shows that higher
temperature is conducive to the formation of graphene carbon, which is
consistent with the previous results [30]. The N 1s spectra of Co@CN-x
(Fig. 4) can be fitted into pyridinic-N, Co-Nx, pyrrolic-N, graphitic-N,
and oxidized-N, respectively [32,33]. It was well accepted that the
pyridine-type N species could function as the anchor to stabilize the
metal NPs [34]. As the pyrolysis temperature increases, the proportion
of graphitic-N and oxidized-N in these Co@CN-x samples increase
significantly, but even the catalysts pyrolyzed at 900 ℃ still have a
relatively high proportion of pyridinic-N (Table S4). This is probably the
reason why the catalyst in this study has a smaller particle size of the Co
NPs as shown in Fig. 3a–c.
Fig. 1. N2 sorption isotherms of (a) Co@CN-700, (b) Co@CN-800, (c)
Co@CN-900.
affords sufficient diffusion space for reactants during the reaction and
thus promoting significantly its catalytic performance.
As shown in Fig. 2, all the catalysts exhibited a broad diffraction peak
of (002) reflection of carbon at around 26.2◦. No obvious Co diffraction
lines were collected on the XRD spectra of Co@CN-700, which indicated
that at low pyrolysis temperatures, the catalyst precursor likely not be
efficiently pyrolyzed. This is also consistent with the previous N2 sorp-
tion characterization results. While at higher pyrolysis temperature, new
diffraction peaks at 44.1◦, 51.4◦, and 75.8◦ were observed, which are
corresponded to (111), (200), and (220) lattice planes of metallic Co,
respectively [28,29]. These results demonstrated that higher pyrolysis
temperature is beneficial to the crystallization and reduction of Co
precursor, which is also consistent with previous literature [30].
As shown in Fig. 3a–c, the mean size of Co NPs varies with the in-
creases of pyrolysis temperature from 700 to 900 ◦C, and is consistent
with the result obtained by XRD characterization before (ESI, Table S1).
The obvious agglomeration of Co NPs was observed on Co@CN-900,
which indicated that too high pyrolysis temperature was not beneficial
to control the particle size of Co species. According to the HRTEM image
of Co@CN-800 (Fig. 4d), the lattice distances of cores and shells are
0.203 nm and 0.34 nm, respectively, which are attributed to the (111)
planes of metallic Co and the (002) plane of graphite [31]. These results
As depicted in Fig. 5, the Co 2p3/2 spectrum can be deconvolved into
four peaks, such as metallic Co, Co-Ox, Co-Nx, and the “shake up” peak,
respectively (Fig. 5) [6]. The proportions of metallic Co increased
significantly with the increases of pyrolysis temperature, while the
proportions of Co-Ox and Co-Nx decreased, accordingly (Table S4).
Considering both the quantitative results and the catalyst preparation
process, it suggested that Co-Ox species were first formed, followed by
Co-Nx upon substitution O with N, and finally, the special Co NPs
encapsulated in N-doped carbon layers catalyst was obtained [35,36]. It
needs to be pointed out that the Co-Nx species still has a high proportion
in the catalyst (about 22.7 % for Co@CN-800), and some studies also
suggested that Co-Nx species often play a key role in the catalytic hy-
drogenation [37].
As manifested in Raman spectra of these catalysts (Fig. 6), the G and
D bonds were identified for all catalysts, which can be assigned to the
–
tangential stretching of C C bonds and graphitic defects, respectively
[38]. The ID/IG value increases with the increase of pyrolysis tempera-
ture, which indicates the increase of in the graphitization degree of the
catalyst (Table S1). Also, the typical characteristic peaks of Co-Ox at
469, 513, 617, and 675 cmꢀ 1 as shown in Fig. 6(a, b), and there is no
obvious peak in the Raman spectrum of Co@CN-900, which may be due
to the decomposition of Co-Ox and to form metallic Co at higher py-
rolysis temperature [39]. These results were consistent with the results
from XPS characterizations (Fig. 5 and Table S5).
3.2. Catalytic activity
The one-pot reductive amination of nitrobenzene with benzaldehyde
was investigated as a model reaction for catalyst screening. As shown in
Table 1, this reaction cannot be carried out normally in absence of a
catalyst (Entry 1 in Table 1). CN-800 was not active for the model re-
action also, indicating that Co is essential for the construction of
reductive amination catalyst (Entry 2 in Table 1). Although the con-
version of nitrobenzene reached 1.0 % under the same reaction condi-
tions, the selectivity of the target product given by the catalysts obtained
at different pyrolysis temperatures was quite different (Entry 3–5 in
Table 1). Co@CN-700 only gave a 36.7 % selectivity to N-benzylaniline,
while the intermediate N-benzylidene aniline was the main product. As
the catalyst pyrolysis temperature increased to above 800 ℃, the
selectivity of N-benzylaniline increased to 53.4 %, and then down to
Fig. 2. XRD patterns of (a) Co@CN-700, (b) Co@CN-800, (c) Co@CN-900.
3