572
Y. Feng et al. / Journal of Catalysis 381 (2020) 570–578
was used to describe the exchange and correlation effect [34]. In all
the cases, the cut-off energy was set to be 450 eV. The Monkhorst-
Pack grids [35] were set to be 3 ꢂ 3 ꢂ 1 for the all the surface cal-
culations. At least 16 Å vacuum layer was applied in z-direction of
the slab models, preventing the vertical interactions between slabs.
Spin-polarization was contained in all the cases. The Co@NC model
was constructed as one mono-layer N doped graphene staying on
the top of the (111) surface of Co.
carbon, respectively (Fig. 2b) [37]. The IG/ID band intensity ratios
of Co@C-N(600), Co@C-N(700), Co@C-N(800) and Co@C-N(900)
materials are 0.86, 0.97, 1.29 and 1.97, respectively, revealing that
the crystallization degree of graphitic carbon becomes better with
temperature, which was consistent with the XRD observations.
SEM images indicated that the pyrolyzed nanoparticles roughly
retained the polyhedral shape of parent ZIF-67, and shrunk slightly
(Fig. 3b). As shown in Fig. 3(d) and Figure S1, the presence of Co
nanoparticles encapsulated by a few layered carbon shells and
the particle size of Co nanoparticles increased with the pyrolysis
temperature. Furthermore, the highly dispersed Co nanoparticles
of materials were also observed, which was mainly due to the
ordered organic ligands of ZIF-67. Fig. 3(e, f) revealed that the car-
bon matrices and Co nanoparticles of Co@C-N(800) were crys-
talline and the lattice fringes with an inter-planar distance of
0.334 nm and 0.204 nm correspond to the C(002) plane and Co
(111) [38]. The diffraction rings in the SAED image (Fig. 3c) can
be attributed to Co, in good agreement with XRD results.
The adsorption energy of O2 was defined as
D
EB ¼ Eads ꢁ Eslab ꢁ EO2
where Eads is the electronic energy of the slab with an adsorbed O2,
Eslab is the electronic energy of the clean slab, and EO2 is the elec-
tronic energy of gaseous oxygen molecule. Under this definition, a
more negative value indicates a stronger binding system.
2.5. Catalytic oxidation esterification of HMF
In a typical run, HMF (5.0 mmol), catalyst (100 mg), Na2CO3
(30% mol relative to HMF), and methanol (5 mL) were added in a
stainless steel reactor equipped with a magnetic stirrer. The reac-
tion mixture was stirred at 100 °C and at 2 MPa O2. After comple-
tion of the reaction, the catalyst was separated and a sample of the
liquid mixture was subjected to GC analysis. The conversion of 5-
HMF and the yield of products were calculated according to the fol-
lowing equations:
3.2. Catalyst activity and stability
The selective oxidative esterification of HMF to DMFDCA was
performed in the presence of sodium carbonate using methanol
as the solvent at 100 °C under 2 MPa oxygen. The results of
exploratory catalytic experiments with the different catalytic
materials were summarized in Table 1. The parent ZIF-67 showed
no activity (entry 1). Pleasingly, the ZIF-67 pyrolysed materials
were effective and the yields of DMFDCA were disproportionately
affected by the pyrolysis temperature of the ZIF-67 precursors
(entries 2–5). Among the four Co@C-N(T) samples, Co@C-N(800)
was the most active for the HMF conversion, affording DMFDCA
in 99% conversion and 91% yield (entry 4).
ꢀ
ꢁ
Moles of HMF
Moles of HMF loaded
HMF conversion ¼ 1 ꢁ
ꢂ 100%
ꢂ 100%
ꢀ
ꢁ
Moles of product
Product yield ¼ 1 ꢁ
Moles of HMF converted
With the optimum Co@C-N(800) catalyst in hand, the reaction
conditions for the conversion of HMF into DMFDCA were subse-
quently investigated. It was found that the reaction temperature
of 100 °C and oxygen pressure of 2 MPa was optimum in terms
of both HMF conversion and DMFDCA yield (Table 1. entries 6–11).
Next, the effect of different strength bases on the Co@C-N(800)-
catalyzed oxidation of HMF to DMFDCA in methanol was investi-
gated (Fig. 4a). The introduction of weak bases (i.e., NaHCO3,
KHCO3) resulted in a slight increase in DMFDCA yield, while a sig-
nificant increase in yield from 83% to 91% was obtained when using
Na2CO3 and K2CO3. However, stronger bases (i.e., NaOH, KOH,
NaOMe and NaOEt) led to a high HMF conversion but a decrease
in DMFDCA yield, indicating that some undesired byproducts were
formed. The Co@C-N(800) material can efficiently catalyze the
oxidative esterification of HMF, even in the absence of base, to
afford DMFDCA in 83% yield (Figure 4 and S2), which can be attrib-
uted to the strong basic sites of Co@C-N(800) as evidenced with
CO2-TPD (Figure S3). The DMFDCA yield firstly increased and then
decreased with the increasing base concentration, and the
optimized content of Na2CO3 was 30 mol% with respect to HMF
(Figure S4). As shown in Fig. 4b, the yield of DMFDCA reached
83% in only one hour, and such high efficiency have not been
reported to date. A 95% yield of DMFDCA was obtained after a pro-
longed reaction time (12 h). These high conversion and selectivity
achieved by Co@C-N can compare well with those reported noble
metal based catalysts [4,6–13], even in a lower O2 pressure
(Table 1, entry 7). Furthermore, the non-noble metal catalysts are
cost-effective and feasible, especially for large-scale production.
The recyclability of the catalyst was also investigated. After the
reaction, the catalyst was dispersed in ethanol solution and was
easily separated with an external magnet given the magnetic prop-
erties of Co-based material (Fig. 4c). ICP-OES analysis of the reac-
tion solution conformed that the content of Co was below the
detection limit, suggesting no obvious Co leaching during the
3. Results and discussion
3.1. Catalyst preparation and characterization
The two-step synthesis of Co@C-N is shown in Fig. 1a. ZIF-67
was firstly synthesized by the modified room-temperature precip-
itation method [36]. The powder X-ray diffraction (XRD) patterns
of synthetic ZIF-67 (Fig. 1b) matched well with a simulated ana-
logue, confirming the formation of pure ZIF-67 crystals. As shown
in the SEM image (Fig. 3a), synthetic ZIF-67 was composed of
microcrystals with a typical rhombic dodecahedral shape, uniform
morphology and high crystallinity.
The Co@C-N materials were formed by a simple thermal decom-
position of ZIF-67 under an argon atmosphere and metal ions were
reduced in situ. Thermogravimetric (TG) analysis indicated that the
organic ligands in ZIF-67 began to decompose at ca. 500 °C
(Fig. 1c). Therefore, the final set of pyrolyzed temperatures chosen
were 600, 700, 800 and 900 °C, resulting in Co@C-N(T), where T
represents the thermolysis temperature. The corresponding mass
loss is about 28.3–46.6 wt%, which is significantly lower than the
theoretical values in the transformation from ZIF-67 to metallic
Co (73.3%), implying that carbon-containing materials are gener-
ated during the pyrolysis process.
XRD patterns of the Co@C-N materials were shown in Fig. 2a,
the peak at about 26.3° corresponds to a typical (0 0 2) interlayer
of graphite-type carbon sheets, and other peaks at about 44.2°,
51.5°, and 75.8° are attributed to metallic Co (PDF#. 15-0806).
The enhanced peak intensities of the Co diffraction peaks for the
Co@C-N(T) at higher calcination temperatures suggests the forma-
tion of a Co phase with a higher crystallization degree. The Raman
spectra of the samples also revealed characteristic carbon G and D
bands, corresponding to the graphitic sp2-carbon and disoriented