VOx molecular level grafted g-C3N4 for highly selective oxidation of methanol to dimethoxymethane

Article abstract of DOI:10.1016/j.mcat.2019.03.008

A composite material VO_x/g-C₃N₄ was prepared and its composition and structure was characterized by FT-IR, XRD, SEM, TEM and XPS. The results showed that VO_x was grafted on g-C₃N₄ surface with molecularlevel dispersion via V or O atoms interacting with the characteristic groups on g-C₃N₄ surface. The highly dispersed VO_x/g-C₃N₄ exhibited effectively catalytic activity and high selectivity for oxidation of methanol into methoxymethane (DMM), giving 44.9 mol/(mol_V·h) TOF and 95.2% selectivity of DMM. The dependence of catalytic activity on the surface properties and preparation procedure of the catalyst, reaction conditions and kinetics, and the stability of the catalyst were explored. Also, the in situ DRIFTS spectra of methanol adsorption on the VO_x/g-C₃N₄ surface and effect of O₂ on the adsorption, as well as the reaction mechanism were discussed.

Full text of DOI:10.1016/j.mcat.2019.03.008

MolecularCatalysis469(2019)48–56
Contents lists available at ScienceDirect
Molecular Catalysis
journal homepage: www.elsevier.com/locate/mcat
VOx molecular level grafted g-C3N4 for highly selective oxidation of
methanol to dimethoxymethane
Hongrui Ma, Hongxia Wang, Bin Lu⁎, Jingxiang Zhao, Qinghai Cai⁎
Key Lab for Photonic and Electronic Bandgap Materials, Ministry of Education, School of Chemistry and Chemical Engineering, Harbin Normal University, No. 1 Shida
Road Limin Development Zone, Harbin, 150025, PR China
A R T I C L E I N F O
A B S T R A C T
Keywords:
g-C3N4 supported VOx catalyst
Methanol
Oxidation
Dimethoxymethane
A composite material VOx/g-C3N4 was prepared and its composition and structure was characterized by FT-IR,
XRD, SEM, TEM and XPS. The results showed that VOx was grafted on g-C3N4 surface with molecularlevel
dispersion via V or O atoms interacting with the characteristic groups on g-C3N4 surface. The highly dispersed
VOx/g-C3N4 exhibited effectively catalytic activity and high selectivity for oxidation of methanol into methox-
ymethane (DMM), giving 44.9 mol/(molV·h) TOF and 95.2% selectivity of DMM. The dependence of catalytic
activity on the surface properties and preparation procedure of the catalyst, reaction conditions and kinetics, and
the stability of the catalyst were explored. Also, the in situ DRIFTS spectra of methanol adsorption on the VOx/g-
C3N4 surface and effect of O2 on the adsorption, as well as the reaction mechanism were discussed.
1. Introduction
polyoxometallates [10], ReOx/TiO2 [11], RuOx/SiO2 [12] and VOx/TS-
1 [13] etc. Their catalytic activity for methanol conversion is all at-
Selective oxidation of alcohols is an attractive process for synthesis
of organic compounds. As a simplest alcohol, selective oxidation of
methanol into dimethoxymethane (DMM) attracts much attention due
toits environmentally-benigncharacter [1]. So, DMM iswidely applied
as a raw material in industries for producing the pharmaceuticals,
perfumes, fine chemicals and excellent solvents [2]. It is also used as a
potential diesel additive which can decrease the soot emission [3]. A
traditional method for synthesizing DMM in industry is by condensation
of formaldehyde (FA) with methanol using mineral acid catalysts [4].
This catalytic process encounters some problems, such as intricate post-
processing of the reaction products, costly technology and equipment
corrosion and environment pollution caused by the acidic catalysts.
Hence, a simpler and environment-benign technology is required for
DMM synthesis. In recent years, one-step selective oxidation of me-
thanol or DME [5] into DMM is developed, which possesses economical
and environmental advantages as compared with the traditional pro-
cesses. The potential protocol via oxidation of methanol includes two
continuous processes, the first is carried out by the oxidation of CH3OH
to formaldehyde by O2 gas, and the followed one is by condensation of
formaldehyde with CH3OH.This process is highly atom-economic as its
only by-product is water. Recently, many heterogeneous metal oxides
or their supported catalysts have been developed and used in this one-
step procedure, involving SbRe2O6 [6,7], Fe-Mo-O [8], Re/γ-Fe2O3 [9],
tributed to their bifunctionalized active sites, redox and acidic sites.
Very recently, V2O5/TiO2 and its acid-modified materials have been
paid more attentions [14–17]. However, the latter ones with stronger
acidity usually cause direct etherification of methanol to generate di-
methyl ether. Meanwhile, the supportsTiO2 and ZrO2 that can catalyze
dehydrogenation reactions usually caused the selective formation of
methyl formate [10], which revealed that the selectivity strongly de-
pends on the active components and the oxide supports with the acid-
base and redox properties on the catalyst surface [18]. On the other
hand, the exploration on influence of the surface properties and inter-
action between VOx species and the supports on the catalytic activity
has been deeply detailed although the roles of different VOx species to
the catalytic activity have been in the debate for some time [19].
Graphic carbon nitride (g-C3N4) is a polymer composed from C, N
and some impurity H, connected via tris-triazine-based patterns. g-C3N4
possesses electron-rich character and it has plently basic functional
groups and hydrogen-bonding interaction on the surface due to the
existence of N and H atoms. The electron-rich character provides ben-
efits to oxidation of organic compounds and basic functional groups like
-NH2 can be applied as mediate species for grafting metal or metal
compounds to highly dispersewith thedecreased particle size on its
surface [20]. In view of the basic nature on g-C3N4 surface that can
limit methanol being converted into dimethyl ether, an easy and
⁎ Corresponding authors.
E-mail addresses: lupin630411@163.com (B. Lu), caiqinghai@yahoo.com (Q. Cai).
https://doi.org/10.1016/j.mcat.2019.03.008
Received 24 January 2019; Received in revised form 4 March 2019; Accepted 7 March 2019
2468-8231/©2019ElsevierB.V.Allrightsreserved.

H. Ma, et al.
MolecularCatalysis469(2019)48–56
effective preparation of catalyst by VOx grafted on g-C3N4 support was
developed. This highly-dispersed composite catalyst presented excellent
catalytic performance for selective oxidation of methanol into DMM.
The effect of the surface structure and composition of the catalyst and
the interaction between VOx species and the g-C3N4 support on me-
thanol oxidation by O2 in a fix-bed reactor were explored in details, and
the results were addressed in this paper.
2. Experimental
2.1. Preparation of the catalyst
5 g of melamine (> 99.5%) was placed in a porcelain boat, which
was then placed in a tubular furnace purged by high-purity nitrogen
gas. The samples were calcined from room temperature to 500 °Cusing a
heating rate of 5 °C/min and then calcined at 500 °C for 4 h, affording
2.37 g of g-C3N4 with 47.5% yield after cool down.
V/g-C3N4 was prepared by the impregnation method. 0.8–1.6g of
NH4VO3 (> 99.0%) and 100 mL of deionized water were loaded into a
250 mL beaker fitted with a magnetic stirrer. The beaker was heated to
90 °C in a water bath for dissolution. When the solution was clarified
with pale yellow color, 2.0 g of g-C3N4 was added to the solution and
the mixture was violently stirred at this temperature for 6 h. After the
impregnation,a solid material was obtained by filtering the mixture,
followed by dried in an oven at 90 °C for 1 h. The dried solid was placed
in a furnace,then directly heated and calcined at 400, 450, 500, 550 and
600 °C for 2 h to obtain V/g-C3N4-t (denoted asV/g-C3N4-40, V/g-C3N4-
45, V/g-C3N4-50, V/g-C3N4-55 and V/g-C3N4-60, respectively).
Fig. 1. XRD patterns of g-C3N4 and V/g-C3N4-t samples (V loading 7.14%).
2.2. Characterization
X-ray powder diffraction (XRD) of the samples was measured by a
Bruker AXS D8 advanced X-ray diffractometer (Cu Kα,
λ
=0.15406 nm). Morphology of the samples was measured by a Hitachi
S-4800 scanning electron microscopy (SEM).FT-IR spectrum was per-
formed on a Vertax-80 spectrometer. The recording of transmission
electron microscopy (TEM) was carried out using a FEI Tecnai F20
equipment. X-ray photoelectron spectroscopy (XPS) was conducted by a
Kratos-AXIS ULTRA DLDX X-ray photoelectron spectrometer.
Vanadium loading on V/g-C3N4-50 was analyzed by 7500CE ICP-MS
(Agilent company).
FT-IR spectrophtometer (Brucker TENSOR II) fitted with a diffuse
reflectance optics accessory (Harrick Scientific products Inc.) was used
for recording in situ DRIFTS spectra. Before measurement, V/g-C3N4-50
was heated to 200 °C for 2 h in vacuo. Afterward, introducing the me-
thanol molecules onto the sample surface was conducted through high
pure N2 (purity 99.99%), and then evacuating the system to about
9.8 × 10−4 Pa. The adsorption of methanol and O2 gas were measured
at 200 °C from 5 to 40 min.
Fig. 2. TEM images of g-C3N4(a) and V/g-C3N4-50 (b, V loading 7.14%).
2.3. Catalytic test and analytic procedure
3. Results and discussion
Evaluation for catalytic activity of V/g-C3N4-t was conducted in a
fixed bed tubular reactor with6 mm internal diameter at atmospheric
pressure. 0.5 g catalyst (40 mesh) was put in the constant temperature
region of the reactor. The raw materials and N2 gas were fed into the
reactor at molar ratio of CH3OH (> 99.5%)/O2 (high purity)/N2 (high
purity) = 1/100/300, corresponding to the space velocity (VHSV) of
2406 mL g−1 h−1. The reactant and product mixture was detected on
online GC (Linghua 9890A, Shanghai) equipped with a FID detector and
HP-5 a 30 m ×0.32 mm ×0.25 μm column and GC–MS (Agilent GC
7890A-MS5975C).
3.1. Characterization
Crystal phase and composition of the prepared g-C3N4 and V/g-
C3N4-t samples were determined by X-ray powder diffraction (XRD) and
their XRD patterns were depicted in Fig. 1. As seen from the figure,
whether g-C3N4 or V/g-C3N4-t samples have almost the same XRD
images. The strong peaks observed at 2θ = 27.9° for all the samples
were assigned to the diffraction plane (002) of g-C3N4 [21]. There were
no any characteristic peaks assigned to vanadium compounds found in
49

H. Ma, et al.
MolecularCatalysis469(2019)48–56
Fig. 3. EDX elemental mapping and high-resolution TEM images of V/g-C3N4-50(V loading 7.14%).
a-SEM, b-carbon, C-Nitrogen, d-oxygen, e-vanadium, f-high-resolution TEM.
these XRD patterns, suggesting that the particle
and V elements on V/g-C3N4-50 surface. The mapping results as shown
in Fig. 3 represented that the distribution of VOx species was very
uniform by highly-dispersed dots (Fig. 3e), which further confirmed
that VOx may exist by the form of molecular level on the surface. This
was in agreement with the results obtained from SEM, TEM and XRD
characterization.
FT-IR spectra of the g-C3N4 and V/g-C3N4-t samples were depicted
in Fig. 4. A set of absorption bands observed at 3182.9-3420.0 cm−1 in
Fig. 4a were ascribed to stretching vibration of NeH bonds and the
bands at 1247.1–1632.4 cm−1 were related to stretching vibrations of
CeN heterocycles [22]. Meanwhile, the peak at 815.3 cm−1was due to
the typical wending vibration of the triazine unit [23]. The FT-IR
spectra in Fig. 4b exhibited that some new absorption bands at 884-
885 cm−1for the five V/g-C3N4-t samples, 915-927 cm−1for V/g-C3N4-
40, V/g-C3N4-45, V/g-C3N4-50 and V/g-C3N4-60, 933 and 945 cm−1 for
V/g-C3N4-55 sample were likely assigned to the VeOeV linkage for
VOx species as compared with g-C3N4 [24]. The above shifted bands
were related to the interaction between VOx species and the support.
Besides the sample V/g-C3N4-60 possessing a weak absorption band at
size of vanadium compounds dispersed on g-C3N4 was too small to
be checked by XRD detector. This deduction was verified by TEM, SEM
and EDX measurements. The SEM images of g-C3N4 and V/g-C3N4-50
(Fig. SI1) exhibited that g-C3N4 and V/g-C3N4-50 samples were very
similar and all are irregular plate-like structure (Fig. SI1a, SI1b). No any
containing V compound particles were found through SEM scanning, so
was for TEM detection (Fig. 2a, b). Even the microstructure of V/g-
C3N4-50 conducted by HRTEM revealed that no any crystal lattice for V
compound was detected in Fig. 3f. To further explore the distribution of
V compounds on the surface, two different areas, area 1 for small flakes
and area 2 for big flake, were designated as shown in Figs. SI2a and
SI2b, and then the energy-dispersive X-ray spectroscopy (SEM-EDX)
was conducted on the two areas. The presence of C, N, O and V lines
was confirmed on these two areas (Fig. SI3), which suggested that VOx
groups would be uniformly distributed on the V/g-C3N4-50 surface,
whether on the small flakes or big ones, implying that the size of grafted
VOx groups were too tiny to be seen by SEM and TEM. Besides, the EDX
elemental mapping was applied to elucidate the dispersion of C, N, O
50

H. Ma, et al.
MolecularCatalysis469(2019)48–56
3.2. Catalytic activity
The catalytic performance of V/g-C3N4-t prepared at various calci-
nation temperatures was estimated and the results were shown in Fig. 6.
The conversion was basically increment as the calcined temperature
raising from 400 to 500 °C for V/g-C3N4-40, V/g-C3N4-45 and V/g-
C3N4-50 at 7 h. Unexpectedly, as the calcined temperature was further
increased from 550 to 600 °C, the conversion was gradually lowed for
V/g-C3N4-55 and V/g-C3N4-60. The selectivity of DMM held high level
with more than 91.0% as the reaction time from 5 to 8 h with the only
by-product dimethyl ether approaching to 9.0% selectivity. The high
selectivity of DMM was corresponding to the used basic support g-C3N4
and the higher activity of V/g-C3N4-50 was likely ascribed to its surface
properties. As FT-IR spectra depicted in Fig. 4, stronger absorption
bands at 1026.0–1075.5 cm−1 attributed to stretching vibrations of V]
O or interacted V]O bands byg-C3N4support was found for V/g-C3N4-
45,
V/g-C3N4-55 and V/g-C3N4-60 samples, which was likely related to
the lower activity of these three catalysts. On the contrary, experiment
results indicated that V/g-C3N4-40 and V/g-C3N4-50 gave higher cata-
lytic activity, maybe related to the disappearance of the bands at
1026.0–1075.5 cm−1. Also, only the weak absorption bands around
884.0 and 920–927 cm−1 assigned to the VeOeV linkage for VOx
species were observed for V/g-C3N4-40 and V/g-C3N4-50 samples.
These findings represented that the catalytic activity of VOx supported
catalysts was independent on V]O bands and the VeOeV linkage for
VOx species [27]. It was likely related to the strong interaction between
VOx species and the g-C3N4 surface enough to cause almost dis-
appearance of V]O and VeOeV bond characters. This was very dif-
ferent from our previously reported work [25], in which we found that
crystal V2O5 with stronger V]O vibration band represented high ac-
tivity for methanol oxidation when using SO42-/V2O5-ZrO2 catalyst.
2-
This difference in the catalytic activity between V/g-C3N4-t and SO4
/
V2O5-ZrO2 likely depended on the interaction between VOx active sites
and the supports, g-C3N4 and ZrO2.
The effect of V loadings on the reaction was conducted and the
results were shown in Fig. 7. The increment of V loadings form 0 (g-
C3N4) to 7.14% on g-C3N4 led to the conversion remarkably increasing
from near 0 to 38.5% at reaction time of 6 h. When further increasing
the loadings to 9.74, 13.36 and 100% (V2O5), the conversion of me-
thanol was decreased to 27.2, 14.2 and 18.7%, respectively. However,
the selectivity of DMM was slightly changed as the loadings were in-
creased at 6 h, providing above 90% selectivity. However, it exhibited
95.2% selectivity for the catalyst V/g-C3N4-50 with loading 7.14%. The
lower catalytic activity of the catalyst with high V loading was likely
attributed to abundant V]O bonds on the surface, so was for pure
V2O5. Hence, it also represented that the interaction between the VOx
active sites and the support g-C3N4 was the reason that V/g-C3N4 pos-
sesses effective activity and high selectivity for oxidation of methanol.
The dependence of reaction temperatures on the oxidation of me-
thanol using V/g-C3N4-50 as catalyst was alsoexplored and the results
were shown in Fig. 8. As seen in the figure, the conversion was slightly
increased as reaction temperature raising from 140 to 180 °C; and then
it sharply increased to 38.5% at 200 °C and 6 h. Meanwhile, the se-
lectivity of DMM was almost unchanged in whole temperature range at
reaction time 6 to 8 h, exhibiting above 90% selectivity. Furthermore,
when the temperature continued toraising from 200 to 220 °C, there
was a sharp drop for the conversion from 5 to 8 h, which was attributed
to the result of reaction thermodynamic constraint for DMM synthesis.
This phenomenon was very similar to our previously reported results
using SO42−/V2O5-ZrO2 catalyst [25]. Besides, according to the ex-
perimental results at reaction temperatures 160 and 200 °C, kinetic
plots of ln (Ct/C0) versus reaction time weremade as shown in Fig. 9,
where Ct and C0 are the concentration of methanol at certain time t in
the reaction process and at the initial, respectively. The obtained ki-
netics plots for the reaction at 160 and 200 °C are all approaching to
Fig. 4. FT-IR spectra of g-C3N4 and V/g-C3N4-t samples (V loading 7.14%).
1026 cm−1 that was assigned to stretching vibration of V]O bonds in
V2O5 molecule [25], there were obviously three shifting bands at
1075.5, 1058.9 and 1042.6 cm−1corresponding to V]O bond in VOx
species for the samples V/g-C3N4-45, V/g-C3N4-55 and V/g-C3N4-60,
respectively. V/g-C3N4-40 and V/g-C3N4-50 samples did not exhibit
absorption band in this region. The shift to the direction of high wa-
venumber for former three samples or almost no absorption for the
latter two noes was also due to the interaction between V]O bonds and
the functionalized groups on the surface, implying that supported va-
nadium compounds mainly existed in the form of VOx on the g-C3N4
surface.
X-ray photoelectron spectroscopy (XPS) of g-C3N4 and V/g-C3N4-50
was shown in Fig. 5. The survey spectrum indicated that C, N and O
elements existed in the g-C3N4 and V/g-C3N4-50 was composed of C, N,
O and V elements, which displayed the formation of V/g-C3N4-50
combined with XRD and FT-IR analysis. The detailed analysis for high-
resolution C1s, N1s, O1s and V2p was conducted. In C1s spectrum of V/
g-C3N4-50 (Fig. 5b), the two lines at 284.9 and 288.3 eV were assigned
to graphic carbon and sp2 C atoms (NeC]N) [26], respectively. Likely,
the binding energies of N1s for V/g-C3N4-50 at 399.0 and 400.8 eV and
that for g-C3N4at 398.5 and 400.5 eV were observed (Fig. 5c). In ad-
dition, the characteristic peaks at 517.4, 533.0 and 532.5 eV were as-
signed to V2p and O1s (O element in VOx and g-C3N4, Fig. 5d and 5e).
Therefore, VOx molecular layer dispersed V/g-C3N4 materials were
formed by interaction of VOx species with functionalized groups on g-
C3N4 surface. The VOx species grafted nature was likely advantageous
to the catalytic performance.
51

H. Ma, et al.
MolecularCatalysis469(2019)48–56
Fig. 5. XPS spectra of g-C3N4 and V/g-C3N4-50 (V loading 7.14%).
linear, revealing that the oxidation reaction of methanol is quasi-first
order in methanol whether at 160 or 200 °C. The first order rate con-
stants calculated from the plots are k433 = 0.0497 h−1 at 160 °C and
stage, the conversion of methanol was obviously increased from the
reaction time of 1 h to 6 h, giving 38.5% conversion at 6 h. Subse-
quently, the higher activity was held during the period of the time from
6 to 8 h. Over the reaction time of 8 h, the activity was gradually lowed
to 16.7% at 12 h reaction time, followed by almost constant conversion
observed from 12 to 20 h. The selectivity of DMM held almost constant
during the period of 4–20 h, giving above 90% selectivity as a whole.
According to the characterization of the used catalyst, it was considered
that deactivation of the catalyst was attributed to changes of the surface
properties. XRD pattern of the used catalyst, as well as that of fresh one
was depicted in Fig. 11. New diffraction peaks of V2O5 crystalline at
26.9 and 40.4° assigned to 110 and 311 planes and
NH4V3O8crystallineat 20.9°were obviously
k
473 = 0.0945 h−1 at 200 °C, respectively. Remarkably, the reaction
rate at 200 °C approaches to twice that at 160 °C. The active energy for
the oxidation was estimated to be 27.41 kJ/mol based on the rate
constants.
3.3. Reusability of the catalyst
To evaluate application prospect of the developed catalyst, the test
for the working life of the catalyst V/g-C3N4-50 with V loading of
7.14% was performed and the result was shown in Fig. 10. On the initial
52

H. Ma, et al.
MolecularCatalysis469(2019)48–56
Fig. 6. The catalytic activity of V/g-C3N4-t(V loading 7.14%).
Reaction conditions: methanol 0.05 mL/min, O2 5 mL/min, N215 mL/min,
catalyst 0.5 g, reaction temperature 200℃.
Fig. 9. Kinetic curves for methanol oxidation.
Fig. 10. Stability of the catalystV/g-C3N4-50(V loading 7.14%).
Reaction conditions: methanol 0.05 mL/min, O2 5 mL/min, N2 15 mL/min,
temperature 200 °C and catalyst 0.5 g.
Fig. 7. The catalytic activity of V/g-C3N4-50 at various V loadings.
Reaction conditions: methanol 0.05 mL/min, O2 5 mL/min,N215 mL/min, cat-
alyst 0.5 g and temperature 200 °C.
Fig. 11. XRD patterns of the fresh and used catalystV/g-C3N4-50(V loading
7.14%).
Fig. 8. The catalytic activity of V/g-C3N4-50(V loading 7.14%) at various re-
actiontemperatures.
Reaction conditions: methanol 0.05 mL/min, O2 5 mL/min, N2 15 mL/min and
catalyst 0.5 g.
53

H. Ma, et al.
MolecularCatalysis469(2019)48–56
Fig. 12. FT-IR spectra of the fresh and used catalystV/g-C3N4-50(V loading
7.14%).
Fig. 14. In situ DRIFTS spectra of methanol on V/g-C3N4-50 (V loading 7.14%).
Fig. 13. TEM photographs of the used catalystV/g-C3N4-50(V loading 7.14%).
observed. However, the band at 1057.2 cm−1 assigned to the
stretching vibration of interacted V]O bond was remarkably increased
as shown in Fig. 12. Besides, three new bands were observed at 2718.2,
2460.5 and 2114.7 cm−1, which were likely assigned to stretching vi-
bration of eNeH+ cations and V5+]O overtone bands [28], respec-
tively. In addition, the V2O5 crystalline particles were observed in TEM
image of the used catalyst with about 15 nm diameter (Fig. 13a). A
high-resolution TEM image showed that two lattice fringes were found
for the V2O5 crystalline particles with 0.3408 (110 plane) and
0.2243 nm (311 plane, Fig. 13b), respectively. As a result, the formation
of V2O5 and/or NH4V3O8 crystalline by interaction of VOx with eNH2
Fig. 15. In situ DRIFTS spectra of methanol and O2 on V/g-C3N4-50 (V loading
7.14%).
54

H. Ma, et al.
MolecularCatalysis469(2019)48–56
Scheme 1. The proposed reaction mechanism.
on the surface and H2O produced in the presence of the methanol
oxidation were surely caused the catalyst deactivation.
molecularly dispersed on g-C3N4 surface, which displayed highly cat-
alytic activity and selectivity for oxidation of methanol into di-
methoxymethane (DMM) using O2 as oxidant, giving 44.9 mol/(molV·h)
TOF and 95.2% selectivity of DMM. The working lifetime for the cat-
alyst was evaluated in fix-bed reactor and deactivation was found as the
reaction time reached to above 8 h. To explore the reason for the cat-
alyst deactivation, the surface composition and structure of the used
catalyst were determined by XRD, FT-IR and TEM. The results exhibited
that the formation of V2O5 crystalline and NH4V3O8crystalline by in-
teraction of VOx with -NH2 on the surface and H2O produced during the
methanol oxidation were the reason that the catalyst deactivated.
Besides, the in situ DRIFTS spectra of methanol adsorption and the
effect of O2 gas on the adsorption of methanol were explored. Based on
the spectra, the reaction mechanism for methanol oxidation was pro-
posed.
3.4. Adsorption of methanol and reaction mechanism
Adsorption behavior of methanol on V/g-C3N4-50 at various ad-
sorption time from 5 to 40 min. was measured by in-situ DRIFTS and
the spectra were shown in Fig. 14. As seen from the Fig. 14a, the bands
at 3708.2 and 3678.7 cm−1 assigned to stretching vibration of hydroxyl
groups OeH and a set of bands at 2800–3000 cm−1 attributed to
stretching vibration of CeH in methanol molecules were remarkably
decreased as the adsorption time prolonged from 5 to 40 min. These
bands almost vanished at the adsorption time 40 min. Besides, the
bands at 1731.7, 1716.5 and 1699.5 cm−1 ascribed to carbonyl C]O
were observed and a set of bands at 1000–1055.5 cm−1 corresponding
to bending vibration of CeH bonds were also obviously reduced. These
findings implied that the adsorbed methanol molecules rapidly con-
verted into formaldehyde via dehydrogenation on the catalyst surface,
which almost completed within 40 min. When O2 was led into the
system for methanol adsorption, the surface conversion of methanol
into HCHO was sharply enhanced. As shown in Fig. 15, the dis-
appearance of the characteristic bands at 2964.6 and 2920.7 cm−1 as-
signed to stretching vibration of CeH, as well as the bands around
1032.3 cm−1, was so rapid that the surface conversion completed
within 20 min, which indicated that O2 could promote the dehy-
drogenation of methanol by means of reacting with produced H linked
to V/g-C3N4-50 surface to form H2O. On the other hand, GC–MS mea-
surement exhibited that the intermediate formaldehyde was not found
in the reaction mixture, which showed that the produced HCHO im-
mediately reacted with methanol to generate DMM once the HCHO
molecules produced and desorbed on the surface, and DMM entered
into the solution. This finding meant that the desorption of HCHO was
likely the rate determined step. As a result, the reaction mechanism was
depicted in Scheme 1.
Acknowledgement
This work was financially supported by the National Natural Science
Foundation of China (No. 21671050).
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.mcat.2019.03.008.
References
[1] Q. Zhang, H. Zhao, B. Lu, J. Zhao, Q. Cai, J. Mol. Catal. A: Chem. 421 (2016)
117–121.
[2] K. Fuji, S. Nakano, E. Fujita, Synthesis 4 (1975) 276–277.
[3] Q.D. Zhang, Y.S. Tang, C.H. Yang, Y.Z. Hang, J. Mol. Catal. A 263 (2007) 149–155.
[4] S. Satoh, Y. Tanigawa, US Pat., 6379507 (2002).
[5] H. Liu, E. Iglesia, J. Phys. Chem. B 107 (2003) 10840–10847.
[6] S. Chen, Y.L. Meng, Y.J. Zhao, X.B. Ma, J.L. Gong, AICHE J. 59 (2013) 2587–2593.
[7] Y.Z. Yuan, H.C. Liu, H. Imoto, T. Shido, Y. Iwasawa, Chem. Lett. 29 (2000)
674–675.
[8] K.A. Thavornprasert, M. Capron, L. Jalowiecki-Duhamel, O. Gardoll, M. Trentesaux,
A.S. Mamede, G. Fang, J. Faye, N. Touati, H. Vezin, J.L. Dubois, J.L. Couturier,
F. Dumeignil, Appl. Catal. B: Environ. 145 (2014) 126–135.
[9] Y.Z. Yuan, T. Shido, Y. Iwasawa, Chem. Commun. (2000) 1421–1422.
[10] H. Liu, E. Iglesia, J. Catal. 223 (2004) 161–169.
[11] O.A. Nikonova, M. Capron, G. Fang, J. Faye, A.S. Mamede, L. Jalowiecki-Duhamel,
F. Dumeignil, G.A. Seisenbaeva, J. Catal. 279 (2011) 310–318.
[12] H. Yu, K. Zeng, X.B. Fu, Y. Zhang, F. Peng, H.J. Wang, J. Yang, J. Phys. Chem. C 112
(2008) 11875–11880.
[13] Q.D. Zhang, Y.S. Tan, C.H. Yang, Y.Z. Han, J. Mol. Catal. A Chem 263 (2007)
149–155.
Firstly, CH3OH molecules interacted with the surface V5+ or N as
acidic and basic site respectively, and then it dehydrogenated to form
CH3OeVe and eNeH adsorbed on the surface. Afterwards, hydrogen
in CH3OeVe species was captured by the basic site to generate HCHO
and eNeH adsorbed on the surface. The HCHO and -N-H further re-
acted with O2 gas and methanol molecules on the surface or in the bulk
to produce DMM and water. At the same time, the catalyst had realized
a catalytic cycle.
[14] Y. Fu, J. Shen, Chem. Commun. (2007) 2172–2173.
[15] S. Chen, S. Wang, X. Ma, J. Gong, Chem. Commun. 47 (2011) 9345–9346.
[16] X. Lu, Z. Qin, M. Dong, H. Zhu, G. Wang, Y. Zhao, W. Fan, J. Wang, Fuel 90 (2011)
1335–1339.
4. Conclusions
V/g-C3N4 was successfully prepared by NH4VO3 grafted on g-C3N4
surface. The characterization results showed that VOx species were
[17] A. Gervasini, P. Carniti, J. Keränen, L. Niinistö, A. Auroux, Catal. Today 96 (2004)
55

H. Ma, et al.
MolecularCatalysis469(2019)48–56
[23] Z. Zhu, Z.Y. Lu, D.D. Wang, X. Tang, Y.S. Yan, W.D. Shi, Y.S. Wang, N.L. Gao,
X. Yao, H.J. Dong, Appl. Catal. B: Environ. 182 (2016) 115–122.
[24] H. Zhao, S. Bennici, J. Shen, A. Auroux, J. Catal. 272 (2010) 176–189.
[25] M. Tao, H. Wang, B. Lu, J. Zhao, Q. Cai, New J. Chem. 41 (2017) 8370–8376.
[26] Y.F. Li, L. Fang, R.X. Jin, Y. Yang, X. Fang, Y. Xing, S.Y. Song, Nanoscale 7 (2015)
758–764.
187–194.
[18] J.M. Tatibouet, Appl. Catal. 148 (1997) 213–252.
[19] G. Centi, Appl. Catal. A Gen. 147 (1996) 267–298.
[20] J. Zhu, P. Xiao, H. Li, A.A.C. Carabineiro, ACS Appl. Mater. Interface 6 (2014)
16449–16465.
[21] Y.M. He, L.H. Zhang, B.T. Teng, M.H. Fan, Environ. Sci. Technol. 49 (2015)
649–656.
[27] B.M. Weckhuysen, D.E. Keller, Catal. Today 78 (2003) 25–46.
[28] M.A. Centeno, I. Carrizosa, J.A. Odriozola, Appl. Catal. B: Environ. 29 (2001)
307–314.
[22] K.X. Li, Z.X. Zeng, L.S. Yana, S.L. Luo, X.B. Luo, M.X. Huo, Y.H. Guo, Appl. Catal. B:
Environ. 165 (2015) 428–437.
56