Ceria-promoted Co@NC catalyst for biofuel upgrade: synergy between ceria and cobalt species

Article abstract of DOI:10.1039/d0ta11941e

Ceria-promoted Co@NC (NC, N doped carbon) catalysts are prepared by pyrolysis of biomass materials. Characterization results indicate that ceria and Co species facilitate the distribution of each other due to the formation of a Ce-O-Co solid solution. The specific surface area of the catalyst increased from 378.77 to 537.7 m²g^?1viathe introduction of ceria. The electron transfer from Co to Ce further enhanced their interaction, and Co species facilitate the formation of more defective oxygen vacancies on ceria, which are beneficial to the activities of catalytic hydrogenation and catalytic transfer hydrogenation (CTH), respectively. Thus, Co/Ce@NC (0.99% Co loading) pyrolyzed at 850 °C exhibits excellent performance in the hydrodeoxygenation (HDO) of vanillin with high metal utilization. Catalytic hydrogenation and CTH coexisted in the presence of H₂and ethanol, and >99% yield of creosol can be obtained in each of them. The reaction processes are monitored. No intermediate is found in aqueous media, while ethoxymethyl-4-methoxy-2-phenol is detected in ethanol. Moreover, Co/Ce@NC presents outstanding stability and general applicability. This work provides new insights into the construction of M@NC (M, metal) catalysts and the HDO process of biofuel upgrade.

Full text of DOI:10.1039/d0ta11941e

Journal of
Materials Chemistry A
View Article Online
View Journal | View Issue
PAPER
Ceria-promoted Co@NC catalyst for biofuel
upgrade: synergy between ceria and cobalt
species†
Cite this: J. Mater. Chem. A, 2021, 9,
541
8
abc
a
a
a
a
Bowei Wang,
Xilong Yan,
Ruixiao Gao, Dan Zhang, Yuyao Zeng, Fangying Zhang,
abc
abc
abc
Yang Li
*
and Ligong Chen
*
Ceria-promoted Co@NC (NC, N doped carbon) catalysts are prepared by pyrolysis of biomass materials.
Characterization results indicate that ceria and Co species facilitate the distribution of each other due to
the formation of a Ce–O–Co solid solution. The specific surface area of the catalyst increased from
2
ꢀ1
3
78.77 to 537.7 m
g
via the introduction of ceria. The electron transfer from Co to Ce further
enhanced their interaction, and Co species facilitate the formation of more defective oxygen vacancies
on ceria, which are beneficial to the activities of catalytic hydrogenation and catalytic transfer
ꢁ
hydrogenation (CTH), respectively. Thus, Co/Ce@NC (0.99% Co loading) pyrolyzed at 850 C exhibits
excellent performance in the hydrodeoxygenation (HDO) of vanillin with high metal utilization. Catalytic
2
hydrogenation and CTH coexisted in the presence of H and ethanol, and >99% yield of creosol can be
Received 9th December 2020
Accepted 26th February 2021
obtained in each of them. The reaction processes are monitored. No intermediate is found in aqueous
media, while ethoxymethyl-4-methoxy-2-phenol is detected in ethanol. Moreover, Co/Ce@NC presents
outstanding stability and general applicability. This work provides new insights into the construction of
M@NC (M, metal) catalysts and the HDO process of biofuel upgrade.
DOI: 10.1039/d0ta11941e
rsc.li/materials-a
price and scarce reserve have greatly limited their application in
industry. Hence, developing transition metal catalysts like Cu,
1
. Introduction
5,6
With increasing demand for energy for the constant develop- Ni, Co, and Fe has become one of the most popular research
ment of the economy, fossil fuels have become diminishing elds. Previous studies have proved that transition metal cata-
resources. The transformation of biomass into high value- lysts, especially Cu catalysts, showed satisfactory catalytic
added chemicals and biofuels is an attractive solution for the performance in HDO via regulating the support or adding
energy crisis, which meets the development requirements of a promoter. However, there are few reports about other transi-
7
energy and the environment. However, the oxygen contents in tion metal catalysts. What is worse, some inevitable troubles
biomass substances are always high and naturally become still hinder the practical application of these catalysts. For
a serious obstacle for their application. The high oxygen content example, their metal contents are much higher than those of
can bring about many disadvantages, such as high viscosity, noble metal catalysts, suffering from the aggregation or leach-
1,2
less volatility, and inferior compatibility and stability. There- ing of the metal component, which increase the risk of catalyst
fore, various approaches have been employed to solve this deactivation and heavy metal pollution.
problem. Among them, hydrodeoxygenation (HDO) is extremely
attractive because oxygen can be effectively removed as water.
In 2013, novel M@NC (M ¼ Co or Fe; NC ¼ N doped carbon)
3,4
catalysts were reported and they exhibited excellent perfor-
Undoubtedly, the catalyst plays an important role in the HDO mance in the selective hydrogenation of nitroarenes, which
reaction. Up to now, noble metal catalysts have been widely were usually synthesized via the pyrolysis of a metal complex
studied and have exhibited excellent performance, but the high and carbon source (commercial carbon black, biomass-derived
8,9
carbon source, MOF/COF materials, etc.). Since then, this
fashion of nanoscale catalysts has attracted more and more
a
School of Chemical Engineering and Technology, Tianjin University, Tianjin 300350, attention due to their merits, such as high metal utilization,
10–14
P. R. China. E-mail: lgchen@tju.edu.cn; liyang777@tju.edu.cn
b
good stability, and excellent catalytic performance.
Collaborative Innovation Center of Chemical Science and Engineering (Tianjin),
However, this fashion can only catalyze simple reactions like the
transformation of nitro compounds into amines. As for more
complicated processes like HDO, M@NC catalysts can no longer
Tianjin 300072, P. R. China
c
Tianjin Engineering Research Center of Functional Fine Chemicals, Tianjin, P. R.
China
15
take effect. Therefore, it is essential to enhance the activity of
†
Electronic supplementary information (ESI) available. See DOI:
0.1039/d0ta11941e
these novel catalysts and expand their applications, especially
1
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. A, 2021, 9, 8541–8553 | 8541

View Article Online
Journal of Materials Chemistry A
Paper
for HDO that is an important method for biofuel upgrade. (1.25 g) in 80 vol% ethanol (60 mL) was stirred under reux for
Among the few reports on HDO application of M@NC cata- 4 h. Subsequently, the solvent was slowly removed under
1
6–18
ꢁ
Fe–N–C@Al O was reported to catalyze the HDO of vacuum. The obtained solid was dried in a vacuum at 60 C and
2 3
lysts,
carboxylic acids resoundingly. Nonetheless, reaction condi- then ground into a ne powder. Finally, the powder was pyro-
17
ꢁ
ꢁ
tions were harsh (320 C, 4 MPa H ) and the catalyst dosage was lyzed at 850 C for 2 h under an argon atmosphere to yield the
2
extremely high (about 69 mol%, based on the active metal catalyst, denoted as Co/Ce@NC. The obtained catalyst could be
component), resulting in the poor stability. ZnO/Co@N-CNTs used without further activation. The samples pyrolyzed at
ꢁ
ꢁ
were synthesized from the pyrolysis of Zn/Co-ZIF and used for 700 C and 1000 C were designated as Co/Ce@NC-700 and Co/
18
the HDO of vanillin, whereas the catalyst dosage was larger Ce@NC-1000, respectively. A similar procedure was applied for
than 20 mol%. The preparation process of the catalyst was also the preparation of catalysts without vitamin B , Ce(NO ) -
1
2
3 3
tedious due to the use of the MOF material as the carbon source. $6H O, and chitosan, denoted as Ce@NC, Co@NC, and Co/Ce,
2
In the above samples, acid sites were introduced to assist in respectively. Co–CeO @NC was similarly prepared with stoi-
2
2 3 3 2
conducting the HDO process, but the hydrogenation activity chiometric commercial CeO instead of Ce(NO ) $6H O. Co/
was not enhanced directly. Furthermore, in order to conduct Ce@C was prepared from glucose instead of chitosan, while
HDO smoothly, the advantage of high metal utilization had to Co/Ce@NC (WP) was obtained from cobaltous acetate rather
be given up.
than vitamin B12. Co@C was prepared from glucose instead of
In this work, we attempted to construct a novel M@NC chitosan, and other conditions were the same as for Co@NC.
catalyst with excellent HDO activity. Ceria was introduced to The preparation details of all catalysts are summarized in the
promote the distribution of Co species and further enhance the ESI (Table S1).†
catalytic activity directly. Meanwhile, all typical advantages of
M@NC catalysts were reserved, especially for high active metal
utilization. Thus, ceria-promoted Co@NC catalysts were
2.3 Characterization
prepared. HDO of vanillin was chosen as the model reaction to The patterns of X-ray diffraction (XRD) were collected on D8-
evaluate the catalytic performance, because it is a crucial Focus (Bruker, Germany) using a Cu-Ka X-ray source in the
ꢁ
ꢁ
ꢁ
ꢀ1
component of lignin-derived pyrolysis oil and can provide liquid range of 5 –80 (4 min ). Raman spectra were collected using
biofuel via HDO. Characterization studies were employed to a DXR Microscope (Thermo Scientic, U.S.A.). SEM images were
study the structure–activity relationship of catalysts. The results recorded on an S-4800 microscope (Hitachi, Japan). TEM images
indicated that the in situ introduced ceria NPs and Co species were obtained with JEM-2100F and JEM-F200 multipurpose high
could facilitate the dispersion of each other. The formation of resolution transmission electron microscopes (JEOL, Japan) at an
Ce–O–Co solid solution promoted the activity of catalytic accelerating voltage of 200 kV. High angle annular dark eld
hydrogenation due to the enhanced interaction between (HAADF) micrographs and EDS mapping analyses were obtained
electron-decient Co and electron-rich doped N species. Mean- with a JEM-F200 high resolution transmission electron micro-
while, the activity of ceria NPs for catalytic transfer hydrogena- scope (JEOL, Japan). Powder samples were dispersed on a carbon-
tion (CTH) was enhanced by virtue of more defective oxygen coated copper grid for TEM and EDS analyses. Inductively
vacancies caused by doped Co species. The optimal Co/Ce@NC coupled plasma optical emission spectrometer (ICP-OES) anal-
catalyst exhibited excellent performance on HDO not only via ysis was conrmed using an Optima 5300DV (Perkin Elmer,
catalytic hydrogenation, but also through CTH with ethanol as U.S.A.). The sample was dispersed in a mixture of nitric acid and
the hydrogen donor. Moreover, the HDO process was discussed hydrochloric acid (volume ratio 4 : 1). The solution obtained aer
in detail to expound the catalytic mechanism. The stability and digestion and dilution was used for tests. Elemental analysis (EA)
general applicability of the catalyst were also investigated.
results were obtained using an Avorio MICRO cube element
analyzer (Elementar, Germany). Brunauer–Emmett–Teller (BET)
surface areas and pore volumes of the nanocomposites were
measured by nitrogen physisorption at liquid nitrogen temper-
ature on a NOVA 2000e analyzer (Quantachrome, U.S.A.). The
pore size distributions were obtained using the Barrett–Joyner–
Halenda (BJH) method. Magnetization curves were obtained on
a SQUID-VSM magnetic property measurement system at 300 K
2. Experimental section
2
.1 Materials
Ce(NO $6H O (AR), CeO
95%) were purchased from Sun Chemical Technology
Shanghai) Co., Ltd., Shanghai, China. Vitamin B12 (98%) was
3
)
3
2
2
, and chitosan (deacetylation degree
$
(
(
(
Quantum Design, U.S.A.). X-ray photoelectron spectroscopy
XPS) was recorded on an ESCALAB Xi+ X-ray photoelectron
provided by Meryer Chemical Technology Co., Ltd., Shanghai,
China. Commercially available reagents were used without
further purication.
3
spectrometer (Thermo Scientic, U.S.A.). NH -TPD was per-
formed on Auto Chem II 2920 equipment (Micromeritics, U.S.A).
ꢁ
The sample was preprocessed at 300 C for 2 h and subsequently
2.2 Preparation of catalysts
ꢁ
ꢀ1
cooled to 50 C under a helium ow (50 mL min ), saturated
ꢀ1
Vitamin B12 was chosen as the Co source. Chitosan was selected with dry gaseous NH (10% in helium, 30 mL min ) for about
3
ꢀ1
as the carbon source because it is an inexpensive and conve- 1 h. The sample was treated under a helium ow (50 mL min
nient biomass with a high content of N. Initially, the suspension for 1 h, and then, NH -TPD was performed at a rate of
O (0.63 g), and chitosan 10 C min to 700 C under a helium ow.
)
3
ꢁ
ꢀ1
ꢁ
3 3 2
of vitamin B12 (0.17 g), Ce(NO ) $6H
8542 | J. Mater. Chem. A, 2021, 9, 8541–8553
This journal is © The Royal Society of Chemistry 2021

View Article Online
Paper
.4 General procedure of catalyst evaluation
Journal of Materials Chemistry A
2
As shown in the XRD patterns (Fig. 1), a broad peak centered
ꢁ
ꢁ
15
at 20 –25 was ascribed to the presence of amorphous carbon.
All catalytic reactions were carried out in a 30 mL stainless-steel
autoclave equipped with a polytetrauoroethylene lining and
a magnetic stirring bar. Typically, the substrate (0.125 mmol),
solvent (2 mL), and catalyst (5 mmol%) were charged into the
reactor. The autoclave was pressurized to 1 MPa hydrogen or
nitrogen pressure and then was placed into an oil bath with
magnetic stirring. Aer the reaction completed, the catalyst was
magnetically separated from the reaction system, while the rest
of the reaction mixture was analyzed by gas chromatography
ꢁ
ꢁ
ꢁ
ꢁ
The evident diffraction peaks at about 28.5 , 33.0 , 47.5 , 56.3 ,
ꢁ
ꢁ
ꢁ
ꢁ
5
9.1 , 69.4 , 76.7 , and 79.1 in the samples except Co@NC were
indexed to the (111), (200), (220), (311), (222), (400), (331), and
420) crystal faces of CeO (JCPDS 34-0394), respectively, indi-
cating the formation of CeO
with Ce@NC, the intensity of CeO
(
2
19,20
2
during pyrolysis.
Compared
2
diffraction peaks in the
samples containing Co became much weaker, even in Co/Ce,
ꢁ
and the peaks at 2q > 65 almost disappeared. Meanwhile,
they shied to a lower 2q value (inset picture in Fig. 1). Both of
the above phenomena manifested the formation of the Ce–O–
(GC) with an OV-101 capillary column and an FID detector (GC-
4
1
2
000A, EWAI, China); “>99” is used to indicate the result of
3
+
ꢁ
Co solid solution architecture, in which Ce and defects of
oxygen vacancies could generate substoichiometric CeO2ꢀx
00%. The oven was set to a temperature program as 80 C for
ꢁ
ꢀ1
ꢁ
,
min and 20 C min up to 280 C for 5 min. The structure of
21
leading to the lattice distortion. As for Co/Ce@NC-700, Co/
the product was conrmed with a Bruker Avance III HD 400
MHz NMR spectrometer (Bruker, Germany). Furthermore, the
spent catalyst was washed with ethanol at least 3 times and
reused in the stability test.
Vanillin conversion and 2-methoxy-4-methylphenol (MMP)
selectivity are dened according to eqn (1) and (2), respectively.
Ce@NC, and Co/Ce@NC-1000, the intensity of CeO diffrac-
2
tion peaks strengthened with the increase of pyrolysis temper-
ature. Signicantly, there was no diffraction peak of Co species
in the catalysts except for Co@NC due to the low loading and/or
ꢀ
ꢁ½
ꢂꢃ
vanillinꢂ
out
Conversion ð%Þ ¼ 1 ꢀ
ꢃ 100
(1)
½
vanillinꢂ
in
½
MMPꢂ
out
Selectivity ð%Þ ¼ 1
ꢃ 100
(2)
ꢀ ½vanillinꢂ
out
[vanillin]in is the concentration of vanillin in the original solu-
tion. [vanillin]out and [MMP]out are the concentrations of
vanillin and MMP in the reaction mixture, respectively. The
calculation method for the selectivity of other products and
intermediates is the same as that of MMP.
3. Results and discussion
3.1 Catalyst characterization
All catalysts were synthesized via pyrolysis in an inert atmo-
sphere, and the details are described in Scheme 1. Then, the
catalysts were characterized to understand their structure and Fig. 1 XRD patterns of (a) Co/Ce@NC-700, (b) Co/Ce@NC, (c) Co/
composition.
Ce@NC-1000, (d) Co@NC, (e) Co/Ce, and (f) Ce@NC.
Scheme 1 Fabrication of Co/Ce@NC catalysts.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. A, 2021, 9, 8541–8553 | 8543

View Article Online
Journal of Materials Chemistry A
Paper
Fig. 2 Raman spectra of (a) Co/Ce@NC-700, (b) Co/Ce@NC, (c) Co/Ce@NC-1000, (d) Co@NC, (e) Co/Ce, and (f) Ce@NC.
2
6
good dispersion of Co species. In Co@NC, two inconspicuous framework. In Co/Ce@NC-700, CeO
2
species of weak contrast
ꢁ
ꢁ
peaks at 44.2 and 51.5 could be assigned to the (111) and (200) and a small number of well-dened CeO
planes of the face-centered cubic Co phase (JCPDS 15-0806), due to the low pyrolysis temperature. As for Co/Ce@NC, CeO
respectively, demonstrating the poor dispersion of Co NPs in NPs with a diameter of 2–5 nm were dispersed on the carbon
2
NPs were observed
2
22
the sample without CeO2.
support uniformly. With the increase of pyrolysis temperature,
NPs became larger in Co/Ce@NC-1000 (7–
Ce@NC, a sharp peak at 465 cm corresponded to the strong 19 nm). It was worth noting that the particle size of all Co/
The Raman spectra of the catalysts are presented in Fig. 2. In the diameter of CeO
2
ꢀ
1
23
F
2g vibration mode of the CeO
red shied and broadened slightly in Co/Ce@NC-700, Co/ indicating that Co species also beneted the dispersion of CeO
Ce@NC, and Co/Ce@NC-1000, indicating the formation of the NPs. The d-spacing values of NPs in Co/Ce@NC-700, Co/
2
structure. However, this peak Ce@NC samples was extremely smaller than that of Ce@NC,
2
21,24
Ce–O–Co solid solution.
Meanwhile, this phenomenon Ce@NC, and Co/Ce@NC-1000 were 0.312, 0.308, 0.303 nm,
became more remarkable with the increase of pyrolysis respectively, corresponding to the (111) plane of cubic CeO2.
ꢀ
1
temperature. Except for Co/Ce, the D band at about 1345 cm
Signicantly, they were smaller than the d-spacing value of
ꢀ1
and the G band at about 1602 cm were observed in the spectra 0.317 nm in Ce@NC, demonstrating the lattice distortion of
21,27
of other samples, and the ratios of I
D
/I
G
were almost identical CeO
2
due to the incorporation of Co species.
Moreover, the
(
1.01 ꢄ 0.02). Moreover, the hump-like Raman spectra reected lattice distortion could get serious with the increase of pyrolysis
15,25
the presence of amorphous carbon.
Raman spectra agreed well with XRD results.
The above analyses of temperature. EDS mapping analysis veried the presences of
C, N, O, P, Co, and Ce elements, in which the contents of P and
The morphologies of the obtained catalysts were character- Co were really low. It also further revealed that the Co species
ized by SEM and TEM. According to the SEM images (Fig. S1†), and CeO NPs were in intimate contact, indicating the interac-
2
Co/Ce@NC samples presented a rock-like structure with many tion between them. The above results were in accordance with
folds on the surface. Co@NC also exhibited a similar architec- XRD and Raman characterization.
ture, but its surface was smoother. However, as for Co/Ce,
Physical and textural properties of Co/Ce@NC-700, Co/
a different architecture was found due to the lack of carbon. Ce@NC, Co/Ce@NC-1000, Co@NC, and Co/Ce are summa-
For Ce@NC, there were many pores in the bulk. TEM images rized in Table 1. It was found that the Co content of all samples
were taken to further understand the distribution of metal was about 1 wt%, except for Co/Ce@NC-700. The Co content of
species. As shown in Fig. 3d, the size distribution of Co NPs in Co/Ce@NC-700 was much lower because the weight loss by
Co@NC was wide, ranging from a few nm to a few hundred nm. chitosan degradation decreased during the insufficient pyrol-
15
This morphological feature was consistent with the reported ysis process. As for Co/Ce@NC-700, Co/Ce@NC, and Co/
8,22
Co@NC catalysts. However, Co NPs could not be easily found Ce@NC-1000, the N content decreased with the increase of
15
in the samples containing Ce, probably because the Co content pyrolysis temperature. The N content of Co@NC was higher
was low and most of the Co was incorporated into the CeO
2
than that of other samples, and its BET specic surface area
8544 | J. Mater. Chem. A, 2021, 9, 8541–8553
This journal is © The Royal Society of Chemistry 2021

View Article Online
Paper
Journal of Materials Chemistry A
Fig. 3 TEM images of (a) Co/Ce@NC-700, (b) Co/Ce@NC, (c) Co/Ce@NC-1000, (d) Co@NC, (e) Co/Ce, and (f) Ce@NC; HAADF image of Co/
Ce@NC with EDS mapping (C, N, O, P, Co, and Ce).
Table 1 Physical and textural properties of Co/Ce@NC-700, Co/Ce@NC, Co/Ce@NC-1000, Co@NC, Co/Ce and Ce@NC
Elemental composition (wt%)
a
a
b
b
a
b
b
2
ꢀ1
3
ꢀ1
Sample
Co
Ce
C
N
P
H
O
S
BET (m g
)
V
total (cm g
)
Raverage (nm)
Co/Ce@NC-700
Co/Ce@NC
Co/Ce@NC-1000
Co@NC
Co/Ce
Ce@NC
0.47
0.99
1.16
1.03
1.16
—
36.54
42.81
45.70
—
75.38
42.17
48.04
41.60
32.48
77.20
2.09
2.08
1.55
1.07
2.86
0.72
1.63
0.31
0.49
0.56
0.42
0.52
—
1.86
1.33
0.86
2.61
0.16
1.57
10.70
11.23
18.17
15.88
19.97
14.59
498.39
537.70
289.72
378.77
150.19
234.97
0.40
0.60
0.29
0.11
0.13
0.13
16.92
19.11
18.97
19.04
16.97
18.99
40.04
a
b
Determined by ICP-OES. Calculated by EA.
2
ꢀ1
(
SSA) was 378.77 m g . When ceria was introduced into the benecial to Co distribution according to TEM images, on the
catalyst, SSAs of Co/Ce@NC-700 and Co/Ce@NC were higher one hand. On the other hand, small ceria distributed uniformly
28
than that of Co@NC. It was probably because ceria was could prevent the graphitic layers from aggregating.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. A, 2021, 9, 8541–8553 | 8545

View Article Online
Journal of Materials Chemistry A
Paper
High-resolution Ce 3d spectra of the samples are presented
in Fig. 5. In Co@NC, no signal was detected in this region due to
the absence of Ce. In the Ce 3d spectra of other samples, eight
separate peaks could be obtained. Among them, the peaks
labeled n
peaks denoted as m
characteristic peaks of Ce , in which n
0
, n
1
, and n
2
from the Ce3d5/2 orbital as well as the
, and m from the Ce3d3/2 orbital were
and m were ascribed to
1 1
and m were the results of Ce(3d 4f )O(2p ),
0
, m
1
2
4
+
0
0
9
2
4
9
1
5
Ce(3d 4f )O(2p ), n
and n and m were derived from Ce(3d 4f )O(2p ), respectively,
whereas the peaks labeled n and m were related to Ce , which
9
0
6
2
2
0
0
3+
9
1
5
30,31
4+
were attributed to Ce(3d 4f )O(2p ).
Therefore, Ce and
3
+
Ce coexisted on the surface of all samples containing Ce. The
content ratios of Ce and Ce (Ce /Ce ) were 0.22, 0.24, 0.26,
3
+
4+
3+
4+
0
1
.25, and 0.18 for Co/Ce@NC-700, Co/Ce@NC, Co/Ce@NC-
000, Co/Ce, and Ce@NC, respectively. There was more Ce
3
+
in Co-containing samples compared with Ce@NC, indicating
3+
30,32
that the doped Co facilitated the generation of Ce .
Mean-
Fig. 4 Magnetization curves of (a) Co/Ce@NC-700, (b) Co/Ce@NC,
while, Co species were benecial to the formation of oxygen
(
c) Co/Ce@NC-1000, (d) Co@NC, (e) Co/Ce, and (f) Ce@NC.
32,33
vacancies (,) on CeO2.
Moreover, the eight peaks of the Co-
containing samples shied to a low BE in contrast to Ce@NC.
Among them, the shi of the Ce BE peak position in Co/Ce was
slighter, possibly because the CeO content was much higher in
2
Nonetheless, the SSA of Co/Ce@NC-1000 was much lower than
that of Co@NC due to the higher pyrolysis temperature.
15
Compared with Co/Ce@NC, the SSA of Ce@NC was much lower,
indicating the importance of Co species in ceria NP dispersion.
Among all samples, the SSA of Co/Ce was the smallest,
demonstrating the contribution of the carbon matrix to the SSA.
The magnetic properties of Co/Ce@NC-700, Co/Ce@NC, Co/
Ce@NC-1000, Co@NC, Co/Ce and Ce@NC were measured at
ambient temperature (300 K), and the results are described in
Fig. 4. Except for Ce@NC, the other samples exhibited ferro-
magnetism, indicating the magnetism primarily induced by Co
species. According to the hysteresis loops of Co/Ce@NC-700,
Co/Ce@NC, and Co/Ce@NC-1000, it could be found that the
specic saturation magnetization (SSM) apparently increased
with increasing pyrolysis temperature. Meanwhile, SSM seemed
to be consistent with the Co content of the samples (Table 1).
However, the SSM of Co/Ce@NC was stronger than that of
Co@NC although their Co contents were similar. This
phenomenon veried the interaction between ceria and Co
species because the Co dopant could endow nonmagnetic ceria
29
with magnetic performance. This result was in accordance
with XRD, Raman, and TEM analyses.
To further conrm the chemical state of the elements, XPS
studies have been performed. There was no signal in the Co 2p
region of Ce@NC. However, in the Co 2p spectra of Co/Ce@NC-
700, Co/Ce@NC, Co/Ce@NC-1000, Co@NC, and Co/Ce, an
inconspicuous peak at about 777–784 eV was found, indicating
the presence of Co (Fig. S2†). This peak was too inferior to be
deconvoluted because of the low content. Compared with
Co@NC, the Co 2p peak of Ce-containing samples, especially
for Co/Ce@NC and Co/Ce@NC-1000, showed an inconspicuous
shi (ꢅ0.2 eV) in the direction of high BE. This result could be
explained by the electron transfer from Co to Ce, further
vindicating the interaction between ceria and Co species. The
electron transfer could lead to an electron-decient state of Co
Fig. 5 High-resolution XPS spectra of the Ce 3d region for (a) Co/
Ce@NC-700, (b) Co/Ce@NC, (c) Co/Ce@NC-1000, (d) Co@NC, (e)
Co/Ce, and (f) Ce@NC.
30
species.
8546 | J. Mater. Chem. A, 2021, 9, 8541–8553
This journal is © The Royal Society of Chemistry 2021

View Article Online
Paper
Journal of Materials Chemistry A
3
+
4+
Co/Ce than other samples. Together with the slightly high BE stronger than that in Ce@NC, because Ce /Ce ratios of Co/
shi of the Co 2p spectra of the Ce-containing samples, the Ce@NC-700, Co/Ce@NC, Co/Ce@NC-1000, and Co/Ce were
generation of the Ce–O–Co solid solution in the samples con- higher than that of Ce@NC, which could further cause the shi
3
+
30
taining Ce and Co was demonstrated. The interaction of Ce , to low BE of OL. The ratios between O and O (O /O ) for Co/
V
A
V
A
,
and transition metal species present in the solid solution Ce@NC and Ce@NC were 1.02 and 0.77, respectively, further
was proved to be effective in the adsorption of the H mole- conrming that Co could benet the generation of , on ceria.
2
30
cule. The above analyses agreed well with XRD, Raman, TEM,
and magnetism results.
According to high-resolution XPS spectra of the N 1s region
for the samples (Fig. S3 and Table S2†), it was found that the
Fig. 6 summarizes the high-resolution O 1s spectra of the surface N content was generally higher than the bulk N content
samples. Each O 1s spectrum of samples including ceria could (Table 1). In Co@NC, the primarily N doping mode was
be deconvoluted into three peaks. The low BE peak at about pyridinic N, while the main doping fashion of N was pyrrolic N
5
5
29 eV could be ascribed to lattice oxygen (O ), the peak at about in Ce@NC. As for Co/Ce@NC and Co/Ce, either surface or
30 eV was assigned to surface active oxygen adsorbed on bulk N content was lower than that of Co@NC or Ce@NC. The N
L
oxygen vacancies (O
the characteristic of adsorption oxygen (O
the O peak in Co@NC, demonstrating that O
L
from CeO in the samples with Ce. It was obvious that O peaks temperature. The N doping modes in these samples were
V
), and the high BE peak at about 532 eV was content in Co/Ce was the lowest among all samples. In Co/
34
A
). There was only Ce@NC-700, Co/Ce@NC, and Co/Ce@NC-1000, it was found
mainly came that the N content decreased with the increase of pyrolysis
A
L
15
2
in the O 1s spectra of the samples with Co exhibited a slight pyridinic N, Co–N, pyrrolic N, and graphitic N. With the
displacement in the direction of low BE compared with that of increase of pyrolysis temperature, the proportion of pyridinic N
2
ꢀ
4+
Ce@NC. The charge transfer from O to Ce can be hindered gradually decreased and pyrrolic and graphitic N demonstrated
3
+
by Ce because the electrons in O 1s are more tightly attached an inverse trend, while Co–N proportion changed little probably
3
+
to Ce . Therefore, this effect in the samples with Co was because the Co content was too low to construct more Co–N
sites. Phosphorus was introduced into the obtained catalysts
using vitamin B12. In each high-resolution XPS spectra of the P
2
1
1
p region for the samples except Ce@NC, the peak at about
32.1 eV was ascribed to the P–C bond, while the peak at about
35
33.1 eV could be assigned to the P–O bond (Fig. S4†).
3.2 Catalytic performance
As a typical component of bio-oil pyrolyzed from lignin, vanillin
is one of the representative compounds to study biofuel
upgrade. Thus, the performances of the prepared catalysts were

2
rst evaluated by the HDO of vanillin with H as the reductant,
and the results are summarized in Table 2. Ce@NC could not
catalyze the reaction, indicating the indispensability of Co
species (entry 14). All catalysts with Co could be used to conduct
the reaction in aqueous medium, but the performances were
different. Compared with Co@NC, ceria signicantly enhanced
the activity of the catalyst (entries 2 and 12). However, the
pyrolysis temperature presented a remarkable effect on the
performance of Co/Ce@NC catalysts. Due to insufficient pyrol-
ysis, Co/Ce@NC-700 exhibited the worst performance among
them (entries 1, 2 and 11). The activity of Co/Ce@NC was much
higher than that of Co/Ce@NC-1000, probably because high
pyrolysis temperature might cause the agglomeration of
9,15
metallic NPs and the loss of doped N species.
analysis agreed well with characterization results. It was worth
mentioning that the performance of Co–CeO @NC was much
The above
2
poorer than that of Co/Ce@NC (entry 15). Characterization
results illustrated that both ceria and Co-based NPs could be
found in Co–CeO @NC and the dispersion of ceria was much
2
poorer than that in Co/Ce@NC (Fig. S5†). Moreover, the Ce–O–
Co solid solution could not form easily in this preparation
method (Fig. S6†), indicating that it was more benecial to the
catalytic performance than Co-based NPs. These results
demonstrated that the in situ introduced ceria was better for
Fig. 6 High-resolution XPS spectra of the O 1s region for (a) Co/
Ce@NC-700, (b) Co/Ce@NC, (c) Co/Ce@NC-1000, (d) Co@NC, (e)
Co/Ce, and (f) Ce@NC.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. A, 2021, 9, 8541–8553 | 8547

View Article Online
Journal of Materials Chemistry A
Paper
a
Table 2 Hydrodeoxygenation of vanillin over several catalysts
b
c
Entry
Catalyst
Solvent
Conversion (%)
Selectivity (%)
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
Co/Ce@NC-700
Co/Ce@NC
Co/Ce@NC
Co/Ce@NC
Co/Ce@NC
Co/Ce@NC
Co/Ce@NC
Co/Ce@NC
Co/Ce@NC
Co/Ce@NC
Co/Ce@NC-1000
Co@NC
Water
Water
4.4 ꢄ 0.7
93.2 ꢄ 1.6
77.3 ꢄ 1.2
>99
81.5 ꢄ 1.4
79.5 ꢄ 1.6
53.4 ꢄ 1.3
16.1 ꢄ 0.9
12.0 ꢄ 0.7
11.0 ꢄ 0.4
34.7 ꢄ 1.4
20.6 ꢄ 1.2
4.5 ꢄ 0.7
—
>99
>99
>99
>99
94.0 ꢄ 1.6
95.4 ꢄ 1.4
>99
Methanol
Ethanol
n-Propanol
n-Butanol
Acetonitrile
1,4-Dioxane
THF
Toluene
Water
Water
Water
Water
Water
Water
Water
Water
85.0 ꢄ 1.7
>99
0
1
2
3
4
5
6
7
8
>99
>99
>99
>99
—
>99
>99
>99
Co/Ce
Ce@NC
2
Co–CeO @NC
46.8 ꢄ 1.9
40.0 ꢄ 2.0
89.8 ꢄ 1.2
4.6 ꢄ 0.8
Co/Ce@C
Co/Ce@NC (WP)
Co@C
>99
a
ꢁ
b
c
Reaction conditions: vanillin (0.13 mmol); solvent: 2 mL; 5 mol% catalyst; 170 C; 1 MPa H
2
; 6 h. Conversion of vanillin. Selectivity of 2-
methoxy-4-methylphenol.
catalytic performance due to the formation of the Ce–O–Co they can work with the H
2
molecule simultaneously and then
solid solution, which was different from the reported Co@NC activate the H–H bond (Fig. S8 and S9†). According to the Mott–
36,37
catalysts with CeO
2
.
As for Co/Ce, the catalytic activity was Schottky effect, the synergistic effect of metal and N species
much lower than that of Co/Ce@NC, manifesting that the N doped in carbon is the key factor for the activation of hydrogen
22,39
doped carbon matrix played an important role in the activity of molecules.
The increased electron deciency of Co species
the catalyst (entry 13). Moreover, compared with Co/Ce@NC strengthened the interaction between Co and electron-rich N
and Co@NC, respectively, the catalytic activities of Co/Ce@C species in the carbon matrix. This wonderful synergistic effect
and Co@C decreased dramatically, further indicating the would directly result in the enchantment of catalytic activity.
8–10,15
importance of doped N species (entry 16).
Although doped Moreover, the in situ introduced ceria improved their activity for
P species have been proved to be conducive to catalytic perfor- catalytic hydrogenation of nitrobenzene and styrene, further
mance of M@NC catalysts, it did not seem relevant to the verifying the above analysis (Table S3†).
38
catalytic activity in this work (entry 17). This was probably
It was worth noting that the best catalytic performance was
because the amount of doped P species was too small to affect achieved in ethanol, and the reaction took place more smoothly
the catalytic performance. Moreover, the catalytic performance in alcohols than other solvents except water. In order to
of Co/Ce@NC strongly depended on the solvent. It was found
that the catalytic activity was enhanced with the increase of
solvent polarity (entries 2–10), and the catalyst presented
excellent performance in aqueous or alcoholic medium.
Table 3 Catalytic transfer hydrogenation of vanillin over several
catalysts
15
a
Obviously, Co/Ce@NC exhibited the most excellent perfor-
mance among all catalysts in the HDO of vanillin. The HDO Entry Catalyst
b
c
Solvent
Conversion (%) Selectivity (%)
activity of Co/Ce@NC was signicantly promoted by the in situ
1
2
3
4
5
6
7
8
9
Co/Ce@NC-700 Ethanol
6.9 ꢄ 0.8
26.2 ꢄ 1.1
65.1 ꢄ 1.4
71.2 ꢄ 1.1
>99
52.2 ꢄ 1.2
33.2 ꢄ 1.3
36.4 ꢄ 1.5
50.2 ꢄ 1.1
9.9 ꢄ 0.9
54.0 ꢄ 1.9
46.1 ꢄ 0.6
5.7 ꢄ 1.1
68.2 ꢄ 1.2
introduced ceria. In previous reports, the enhancement of HDO
activity of M@NC catalysts was principally ascribed to the
accession of the acidic support that could provide sufficient
Co/Ce@NC
Co/Ce@NCd
Co/Ce@NC
Co/Ce@NC
Co/Ce@NC
Methanol 8.0 ꢄ 0.7
Ethanol
Ethanol
30.8 ꢄ 1.2
>99
17,18
acidic sites to catalyze the dehydration.
However, this was
n-Propanol 43.7 ꢄ 1.3
n-Butanol 43.0 ꢄ 0.5
not appropriate for this work, because ceria could not endow
more acid active sites to the catalyst (Fig. S7†). Due to the
formation of the Ce–O–Co solid solution, the dispersion of Co
species was effectively improved by ceria. Meanwhile, the elec- 10
tron transferred from Co to Ce, leading to the further depres- 11
Co/Ce@NC-1000 Ethanol
20.3 ꢄ 1.1
5.6 ꢄ 0.6
9.1 ꢄ 1.0
20.0 ꢄ 1.3
4.0 ꢄ 1.1
27.3 ꢄ 1.4
31.9 ꢄ 1.0
Co@NC
Co/Ce
Ce@NC
Co–CeO @NC
2
Ethanol
Ethanol
Ethanol
Ethanol
Ethanol
1
1
2
3
Co/Ce@C
sion of Co electron density in Co/Ce@NC. DFT calculation
results indicated that the mechanism of H activation was the
heterolytic splitting of H via the synergistic effect between Co
species and doped N species in the carbon material, in which
Co/Ce@NC (WP) Ethanol
2
a
Reaction conditions: vanillin (0.13 mmol); solvent: 2 mL; 5 mol%
b c
2
ꢁ
catalyst; 170 C; 1 MPa N
of 2-methoxy-4-methylphenol. 10 mol% catalyst; 220 C; 3 h.
2
; 6 dh. Conversion of vanillin. Selectivity
ꢁ
8548 | J. Mater. Chem. A, 2021, 9, 8541–8553
This journal is © The Royal Society of Chemistry 2021

View Article Online
Paper
Journal of Materials Chemistry A
understand the effect of the alcohol solvent, the HDO reaction emerged. For example, the precursors of Co/N–C-600 and ZnO/
was proceeded in an N atmosphere rather than an H atmo- Co–N–CNT are MOFs whose preparation is tedious. What is
2
2
sphere. The results are listed in Table 3. It was found that all noteworthy is that they have to sacrice the superiority of high
catalysts could catalyze the transfer hydrogenation of vanillin metal utilization to guarantee the catalytic activity. Although the
with alcohols as hydrogen donors. Similar to the results in Table reaction conditions are relatively mild with such a low metal
ꢁ
2
, Co/Ce@NC exhibited the best performance among them. utilization, 150 C and 0.75 MPa H
2
are still indispensable. Even
Various alcohols were selected as hydrogen donors in the CTH worse, M@NC catalysts with high metal utilization could only
of vanillin. The performance was the poorest when using catalyze the transfer hydrogenation of vanillin (Co@NC-700 and
methanol as the hydrogen donor (entry 2). As for ethanol, n- Co@NG-6). The introduction of acid sites is another idea to
propanol, and n-butanol, vanillin conversion was proportional design M@NC catalysts with high HDO activity, but acid sites
to the carbon chain length, while the selectivity of 2-methoxy-4- may cause the deactivation of the catalyst due to carbon depo-
methylphenol (creosol, MMP) showed an inverse trend (entries sition. In this work, Co/Ce@NC was prepared by pyrolysis of
3 3 2
3, 5 and 6). The byproduct was benzyl ether. Although similar biomass materials, and Ce(NO ) $6H O was added to the
results have been reported in a previous report, the solvent mixture of chitosan and vitamin B12 directly. The preparation
effect of alcohols in CTH is still too complicated to under- process is simple like typical M@NC catalysts. The synergy
40
stand. Many factors, such as solvent polarity, steric effect, between ceria and cobalt species is successfully utilized to
catalyst, and the stability of the product and/or intermediate, enhance the activity of M@NC catalysts. The performance is
may affect the CTH process and the distribution of prod- improved via promoting the activity for H2 activation and
16,40,41
ucts.
Therefore, further study is still needed. A plausible ethanol dehydrogenation, rather than the introduction of acid
explanation was that the dehydrogenation of alcohol was easier sites. The Co loading of the obtained catalyst is about 1 wt%,
40
2
as the carbon chain got longer, but the activity for catalytic and the catalyst dosage is the lowest for H activation (Table
hydrogenation was much weaker due to the decrease of solvent S4†). Furthermore, Co/Ce@NC can catalyze the HDO of vanillin
polarity. Based on the results in Table 3, ethanol was chosen as via transfer hydrogenation when only using the green solvent
the ideal solvent and hydrogen donor, because it is a green ethanol as the reductant.
42
solvent produced sustainably from biomass via fermentation.
The catalytic performance of Ce@NC was much better than
that of Co@NC (entries 8 and 10), indicating that the active sites
3.3 Possible catalytic mechanism
of CTH on Co/Ce@NC mainly came from ceria rather than Co In order to investigate the conversion of vanillin as well as the
species, which were different from those of Co/Ce@NC for distribution of product and intermediate over Co/Ce@NC, the
vanillin HDO in water. It is well known that the defective , is reaction processes in water (H atmosphere) and ethanol (H2
2
43
the active site for the dehydrogenation of alcohols. According and N atmosphere, respectively) were monitored respectively.
2
to characterization results, the introduced Co species could not As shown in Fig. 7, only the HDO product MMP was detected in
only improve the dispersion of ceria NPs, but also introduce aqueous medium (Fig. 7a). When ethanol was used as the
more defective , sites, which could enhance the activity of Co/ solvent, MMP and the intermediate ethoxymethyl-4-methoxy-2-
Ce@NC for ethanol activation signicantly. Furthermore, the N- phenol (EMP) were found whether under H or not (Fig. 7b and
2
doped carbon matrix also played an important role in CTH c). As for EMP, it was generated from the hydrogenation of the
(entries 3, 9 and 12), probably because heteroatom-doped acetal intermediate derived from vanillin and ethanol or the
carbon materials are always in favor of the dispersion of reaction between vanillyl alcohol with ethanol (or ethanal).
active sites, the adsorption of reactant, the distribution of metal However, it was speculated that EMP was not derived from the
2
2,28,44
species, and mass transfer.
Similar to catalytic hydroge- acetal intermediate in this work, because it could be detected in
45,46
nation, it seemed that doped P was still dispensable for the CTH the reaction medium easily according to previous reports.
process (entry 13). As mentioned above, ceria helped Co species The intermediate of vanillyl alcohol was also not detected in any
enhance the activity for HDO in aqueous medium, while Co of the above reaction systems. To further understand the reac-
helped ceria NPs improve the catalytic performance of CTH in tion process, vanillyl alcohol was chosen as the reactant under
ethanol. It was interesting that Co and ceria species could help the above reaction conditions (Fig. 7d–f). In every system, the
each other in Co/Ce@NC.
conversion rate of vanillyl alcohol was much faster than that of
Excellent non-noble metal-based catalysts reported in vanillin. This might be the reason why vanillyl alcohol could not
previous studies are summarized (Table S4†). All of them be detected. It was also found that the reaction between vanillyl
exhibited remarkable performance in the HDO of vanillin. alcohol and ethanol was much easier than the hydrogenolysis of
However, there are still some merits for our catalyst compared vanillyl alcohol under N (Fig. 7f). Therefore, vanillyl alcohol
2
with them. First of all, M@NC catalysts are generally better than was hard to be captured due to the high reaction rate. Mean-
other non-noble metal-based catalysts, because they do not while, EMP was also detected in ethanol, indicating that EMP
need another reduction process before use, whereas typical probably came from vanillyl alcohol and ethanol (or ethanal)
M@NC catalysts cannot catalyze the HDO of vanillin smoothly (Fig. 7e and f). Moreover, the selectivity of EMP was higher than
and it is hard to enhance their activity due to the pyrolysis that of MMP at the early stage of the reaction by employing
preparation process at high temperature. To solve this problem, vanillyl alcohol as the reactant in ethanol (Fig. 7e and f). When
ingenious but complicated designs for pyrolysis precursors vanillin was the reactant, the selectivity of EMP was lower than
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. A, 2021, 9, 8541–8553 | 8549

View Article Online
Journal of Materials Chemistry A
Paper
Fig. 7 Time course of vanillin (a–c) or vanillyl alcohol (d–f) conversion and product selectivity over Co/Ce@NC. Reaction conditions: (a) vanillin
ꢁ
ꢁ
(
(
0.13 mmol); 2 mL water; 5 mol% catalyst; 170 C; 1 MPa H
2
; (b) vanillin (0.13 mmol); 2 mL ethanol; 5 mol% catalyst; 170 C; 1 MPa H
2
; (c) vanillin
; (e)
ꢁ
ꢁ
0.13 mmol); 2 mL ethanol; 10 mol% catalyst; 220 C; 1 MPa N
vanillyl alcohol (0.13 mmol); 2 mL ethanol; 5 mol% catalyst; 170 C; 1 MPa H
2
; (d) vanillyl alcohol (0.13 mmol); 2 mL water; 5 mol% catalyst; 170 C; 1 MPa H
2
ꢁ
2
; (f) vanillyl alcohol (0.13 mmol); 2 mL ethanol; 10 mol% catalyst;
ꢁ
2
20 C; 1 MPa N
2
.
that of MMP, demonstrating the presence of vanillin (Fig. 7a and d). In ethanol under an N
2
atmosphere, the HDO of
hydrogenolysis. vanillin was conducted through CTH. The capture of EMP
The plausible catalytic processes over Co/Ce@NC are illus- indirectly proved the existence of vanillyl alcohol although it
trated in Fig. 8. The HDO of vanillin proceeded via catalytic was not found in the reaction system. EMP is the main
46
hydrogenation in water. Macroscopically, MMP was produced byproduct. Fortunately, it could transfer to MMP smoothly.
by the hydrogenolysis of vanillin. However, the presence of Although the etherication impeded the hydrogenolysis of
vanillyl alcohol could not be ruled out, because the reaction vanillyl alcohol, the HDO of vanillin as a whole was signicantly
speed of vanillyl alcohol was much faster than that of vanillin improved in the presence of both H and ethanol, due to the
2
8550 | J. Mater. Chem. A, 2021, 9, 8541–8553
This journal is © The Royal Society of Chemistry 2021

View Article Online
Paper
Journal of Materials Chemistry A
Fig. 8 Possible catalytic mechanism over Co/Ce@NC.
coexistence of catalytic hydrogenation and CTH. According to illustrated that ceria NPs dispersed uniformly on the N-doped
the above characterization and experimental results, Co/ carbon material, while Co-based NPs could not be found. In
ꢁ
Ce@NC exhibited activities of catalytic hydrogenation and Co/Ce@NC pyrolyzed at 850 C, the in situ introduced ceria NPs
CTH simultaneously. In catalytic hydrogenation, ceria improved and Co species could facilitate the dispersion of each other. EDS
the Co species activity for H
2
activation, while the introduction mapping and ICP-OES results conrmed the presence of Co.
of Co endowed more active sites on ceria for ethanol dehydro- Compared with Co@NC, the specic surface area of Co/Ce@NC
8,41,47,48
2
ꢀ1
genation.
The synergy between Co and Ce provided the increased from 378.77 to 537.7 m g by the introduction of
excellent performance of Co/Ce@NC.
ceria. Magnetic tests further manifested the interaction
between ceria and Co species. XPS characterization revealed the
electron transfer from Co to Ce, which could strengthen the
interaction between Co and doped N species. Coincidentally,
doped Co species endowed more defective ,. According to the
results of catalytic performance, it was found that the pyrolysis
temperature affected the catalytic activity, because insufficient
pyrolysis could not provide the active sites while the high
pyrolysis temperature caused the agglomeration of ceria NPs
and the loss of doped N species. Co/Ce@NC exhibited the most
excellent performance in the HDO of vanillin. The conversion of
3.4 Stability and applicability
Co/Ce@NC exhibited splendid stability, which could run 5
cycles without deactivation and could be easily isolated using an
external magnet (Fig. S10†). The Co content of the reaction
solution in the stability tests was below the detection limit of
ICP-OES, indicating that Co species did not leach during the
reaction. Moreover, Co/Ce@NC showed outstanding general
applicability (Table S5†). The HDO of many aldehydes similar to
vanillin proceeded smoothly. The HDO of vanillone also could
be catalyzed by Co/Ce@NC successfully, indicating the potential
application of Co/Ce@NC in ketone substrates. As for carboxylic
acids substrates like vanillic acid, the decarboxylic reaction
could be catalyzed via Co/Ce@NC and the corresponding
products were obtained.
ꢁ
vanillin could achieve 93.2% and >99% in water (170 C, 1 MPa
ꢁ
H
2
, 6 h, and 5 mol% catalyst) and ethanol (220 C, 1 MPa N
2
, 6 h,
and 10 mol% catalyst), respectively. Both of the MMP selectiv-
ities were >99%. In the presence of H2 and ethanol, both
ꢁ
vanillin conversion and MMP selectivity achieved >99% (170 C,
1
MPa H
catalytic hydrogenation and CTH processes. In Co/Ce@NC, the
interaction between Co and doped N species could activate H
2
, 6 h, and 5 mol% catalyst) due to the concurrence of
4
. Conclusions
2
for catalytic hydrogenation, while the defective , on ceria NPs
could catalyze the dehydrogenation of alcohols. Benzyl ethers
were the by-products in the alcohol medium, indicating the
presence of vanillyl alcohol indirectly. EMP was found to be the
main intermediate in ethanol, and it could transform into MMP
smoothly. Furthermore, Co/Ce@NC presented good stability
and general applicability. Obviously, this work affords a novel
The synergy between ceria and Co species on N doped carbon
was successfully constructed and utilized to enhance the activity
of M@NC catalysts. A series of Co/Ce@NC catalysts were
prepared, characterized, and applied in the HDO of vanillin.
XRD results indicated the presence of ceria in Co/Ce@NC
catalysts. The formation of the Ce–O–Co solid solution was
conrmed by XRD and Raman spectra. TEM characterization
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. A, 2021, 9, 8541–8553 | 8551

View Article Online
Journal of Materials Chemistry A
Paper
idea on the design of HDO catalysts for biofuel upgrade and 18 V. Ranaware, D. Verma, R. Insyani, A. Riaz, S. M. Kim and
provides a new strategy to construct M@NC catalysts with high
activity.
J. Kim, Green Chem., 2019, 21, 1021–1042.
19 J. Fan, X. Niu, W. Teng, P. Zhang, W.-x. Zhang and D. Zhao, J.
Mater. Chem. A, 2020, 8, 17174–17184.
2
0 S. Yang, H. Shen, F. Cheng, C. Wu, Y. Cao, S. Zhuo, Q. Zhang
and H. Zhang, J. Mater. Chem. A, 2020, 8, 14006–14014.
Conflicts of interest
There are no conicts to declare.
21 W. Shan, Z. Feng, Z. Li, J. Zhang, W. Shen and C. Li, J. Catal.,
004, 228, 206–217.
2
2
2 H. Guo, R. Gao, M. Sun, H. Guo, B. Wang and L. Chen,
ChemSusChem, 2019, 12, 487–494.
Acknowledgements
Project funded by the National Natural Science Foundation of 23 T.-y. Chen, J. Su, Z. Zhang, C. Cao, X. Wang, R. Si, X. Liu,
China (Grant No. 21808161 and 21978208) and the China
Postdoctoral Science Foundation (Grant No. 2020T130465).
B. Shi, J. Xu and Y.-F. Han, ACS Catal., 2018, 8, 8606–8617.
24 X. Wang, L. Li, T. Zhang, B. Lin, J. Ni, C.-T. Au and L. Jiang,
Chem. Commun., 2019, 55, 474–477.
2
5 M. C. Mangarella and K. S. Walton, Carbon, 2015, 95, 208–
References
219.
1
2
3
4
Z. Lv, Q. Sun, X. Meng and F.-S. Xiao, J. Mater. Chem. A, 2013, 26 F. Hu, Y. Peng, J. Chen, S. Liu, H. Song and J. Li, Appl. Catal.,
, 8630–8635.
B, 2019, 240, 329–336.
G.-H. Han, M. W. Lee, S. Park, H. J. Kim, A. J.-P. Ahn, 27 M. Chen, C. Wang, Y. Wang, Z. Tang, Z. Yang, H. Zhang and
M.-g. Seo and K.-Y. Lee, J. Catal., 2019, 377, 343–357.
J. Wang, Fuel, 2019, 247, 344–355.
F. Zhang, Y. Jin, Y. Fu, Y. Zhong, W. Zhu, A. A. Ibrahim and 28 B. Wang, L. Si, J. Geng, Y. Su, Y. Li, X. Yan and L. Chen, Appl.
M. S. El-Shall, J. Mater. Chem. A, 2015, 3, 17008–17015. Catal., B, 2017, 204, 316–323.
F. Yang, N. J. Libretto, M. R. Komarneni, W. Zhou, 29 B. Vodungbo, Y. Zheng, F. Vidal, D. Demaille and
1
J. T. Miller, X. Zhu and D. E. Resasco, ACS Catal., 2019, 9,
791–7800.
V. H. Etgens, Appl. Phys. Lett., 2007, 90, 062510.
30 G. Zhou, B. Dai, H. Xie, G. Zhang, K. Xiong and X. Zheng, J.
7
5
6
7
X. Kong, W. Li, X. Li, L. Liu, W. Geng and L. Liu, J. Energy
Inst., 2020, 93, 658–665.
X.-P. Yin, S.-F. Tang, C. Zhang, H.-J. Wang, R. Si, X.-L. Lu and
T.-B. Lu, J. Mater. Chem. A, 2020, 8, 20925–20930.
D. Verma, R. Insyani, H. S. Cahyadi, J. Park, S. M. Kim,
CO Util., 2017, 21, 292–301.
2
31 Z. Yang, H. Yi, X. Tang, S. Zhao, Y. Huang, X. Xie, L. Song and
Y. Zhang, J. Hazard. Mater., 2019, 373, 321–334.
32 M. Konsolakis, M. Sgourakis and S. A. C. Carabineiro, Appl.
Surf. Sci., 2015, 341, 48–54.
J. M. Cho, J. W. Bae and J. Kim, Green Chem., 2018, 20, 33 S. Xu, F. Xie, H. Xie, G. Zhou and X. Liu, Chem. Eng. J., 2019,
253–3270. 375, 122023.
F. A. Westerhaus, R. V. Jagadeesh, G. Wienhofer, M. M. Pohl, 34 X. Chen, X. Chen, E. Yu, S. Cai, H. Jia, J. Chen and P. Liang,
J. Radnik, A. E. Surkus, J. Rabeah, K. Junge, H. Junge,
Chem. Eng. J., 2018, 344, 469–479.
M. Nielsen, A. Bruckner and M. Beller, Nat. Chem., 2013, 5, 35 J. Zhang, Z. Zhao, Z. Xia and L. Dai, Nat. Nanotechnol., 2015,
37–543.
10, 444–452.
R. V. Jagadeesh, A.-E. Surkus, H. Junge, M.-M. Pohl, 36 T. Stemmler, F. Chen, S. Pisiewicz, A.-E. Surkus, M.-M. Pohl,
3
8
9
5
J. Radnik, J. Rabeah, H. Huan, V. Schuenemann,
C. Topf and M. Beller, J. Mater. Chem. A, 2015, 3, 17728–
A. Brueckner and M. Beller, Science, 2013, 342, 1073–1076.
17737.
1
1
1
1
0 P. Zhou, L. Jiang, F. Wang, K. Deng, K. Lv and Z. Zhang, Sci. 37 R. Ferraccioli, D. Borovika, A.-E. Surkus, C. Kreyenschulte,
Adv., 2017, 3, e1601945. C. Topf and M. Beller, Catal. Sci. Technol., 2018, 8, 499–507.
1 H. Guo, B. Wang, P. Qiu, R. Gao, M. Sun and L. Chen, ACS 38 Y. Duan, T. Song, X. Dong and Y. Yang, Green Chem., 2018,
Sustainable Chem. Eng., 2019, 7, 8876–8884.
20, 2821–2828.
2 G. Jaiswal, V. G. Landge, D. Jagadeesan and E. Balaraman, 39 H. Su, K.-X. Zhang, B. Zhang, H.-H. Wang, Q.-Y. Yu, X.-H. Li,
Nat. Commun., 2017, 8, 2147.
M. Antonietti and J.-S. Chen, J. Am. Chem. Soc., 2017, 139,
811–818.
3 W. Liu, L. Zhang, X. Liu, X. Liu, X. Yang, S. Miao, W. Wang,
A. Wang and T. Zhang, J. Am. Chem. Soc., 2017, 139, 10790– 40 R. Fan, C. Chen, M. Han, W. Gong, H. Zhang, Y. Zhang,
0798. H. Zhao and G. Wang, Small, 2018, 14, 1801953.
4 B. Sahoo, A.-E. Surkus, M.-M. Pohl, J. Radnik, M. Schneider, 41 P. Panagiotopoulou, N. Martin and D. G. Vlachos, J. Mol.
1
1
S. Bachmann, M. Scalone, K. Junge and M. Beller, Angew.
Chem., Int. Ed., 2017, 56, 11242–11247.
5 B. Wang, L. Liu, R. Gao, D. Zhang, L. Chen, Y. Li and
Y. Wang, Int. J. Hydrogen Energy, 2020, 45, 1649–1657.
6 L. Jiang, P. Zhou, C. Liao, Z. Zhang and S. Jin, ChemSusChem,
Catal. A: Chem., 2014, 392, 223–228.
42 P. Panagiotopoulou and D. G. Vlachos, Appl. Catal., A, 2014,
480, 17–24.
43 N. C. Nelson, B. W. Boote, P. Naik, A. J. Rossini, E. A. Smith
and I. I. Slowing, J. Catal., 2017, 346, 180–187.
44 X. Kong, Z. Sun, M. Chen, C. Chen and Q. Chen, Energy
Environ. Sci., 2013, 6, 3260–3266.
1
1
1
2018, 11, 959–964.
7 J. Li, J. Zhang, S. Wang, G. Xu, H. Wang and D. G. Vlachos,
ACS Catal., 2019, 9, 1564–1577.
8552 | J. Mater. Chem. A, 2021, 9, 8541–8553
This journal is © The Royal Society of Chemistry 2021

View Article Online
Paper
Journal of Materials Chemistry A
45 Y. Wang, Q. Cui, Y. Guan and P. Wu, Green Chem., 2018, 20, 47 D. Formenti, F. Ferretti, C. Topf, A.-E. Surkus, M.-M. Pohl,
2
110–2117.
6 X. Tong, P. Guo, S. Liao, S. Xue and H. Zhang, Green Chem.,
019, 21, 5828–5840.
J. Radnik, M. Schneider, K. Junge, M. Beller and
F. Ragaini, J. Catal., 2017, 351, 78–89.
48 Z. Wu, M. Li, D. R. Mullins and S. H. Overbury, ACS Catal.,
4
2
2012, 2, 2224–2234.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. A, 2021, 9, 8541–8553 | 8553