Z.-H. He et al.
Molecular Catalysis 496 (2020) 111192
O2 and heated to the desired temperature. Upon reaction completion,
the reactor was cooled to room temperature and the gas in the autoclave
was vented slowly. The liquid was analyzed by GC (GC9720, Zhejiang
Fuli Analytical Instruments Co., Ltd., China) and GC–MS (Agilent
6890N-5975). The procedures for hydrogenation reaction were similar
to those of dehydrogenation.
Co3O4 (ICDD PDF Card No. 00-042-1467). The result indicated that
Co3O4 is successfully loaded on AlN support. The XRD patterns of
Co/AlN catalysts which were reduced from Co3O4/AlN with different
loadings presented well defined peaks of AlN (Fig. 1b). No Co◦ peak was
obviously found, albeit with a high Co loading (8%). The reason for this
phenomenon is probably that the particle size of Co3O4 NPs dwindled
upon reduction, which is consistent with the following TEM results.
The XRD results of other catalysts including Co3O4 and Co supported
on CeO2, SiO2, and BN are given in Fig. S3, and the Co loadings of all
these catalysts were 4%. A peak at 37.1◦ could be found in these three
unreduced catalysts, which could be assigned to (311) plane of Co3O4.
However, no Co peak could be detected after reduction, indicating that
the Co NPs size decreases in the reduced catalysts, and this is similar to
the Co3O4/AlN and Co/AlN catalysts.
Reuse and hot filtration experiments
Upon reaction completion, the solid catalyst was filtrated and
washed by ethanol (5 × 3 mL) and dried at 60 ◦C for 3 h under vacuum.
The catalyst was then used directly in the next run, and the other op-
erations were similar to those of standard run.
The hot filtration experiments were accomplished as followings.
After the dehydrogenation or hydrogenation reaction was performed for
6 h, the solid catalyst was filtrated and the filtrate continued being
reacted for another 6 or 8 h. Upon reaction completion, the liquid was
analyzed to detect whether the yield of product was increased.
Fig. 2 shows the TEM images of the Co3O4/AlN and Co/AlN catalysts.
The image of Co3O4/AlN catalyst showed that Co3O4 NPs were loaded
on AlN in the size range of 15–30 nm, which was calculated from more
than 150 NPs (Fig. 2a and b). The HRTEM images of the Co3O4 NPs
exhibited the lattice spacing of 2.44 and 4.68 Å, which were contributed
to the (311) and (111) planes of Co3O4 (Fig. 2c and d). The results
confirmed that Co3O4 exists in Co3O4/AlN, and this is in line with the
XRD results.
Results and discussion
Catalyst characterizations
After reduction in H2 atmosphere, the morphology of Co/AlN cata-
lyst was changed remarkably. The AlN support showed the nanosheet
morphology with a thickness range of 3–5 nm. It is worth noting that
although the morphology changed, the crystal of AlN retained according
to the former XRD results. Besides, the particle size of Co◦ NPs was
centered in the range of 6–9 nm, indicating that the size was dwindled
after reduction [42]. The HRTEM images of the Co/AlN showed the
lattice fringes of 2.47 and 2.03 Å in Fig. 2g and h, which were assigned to
CoO (111) and Co◦ (002) planes, respectively. The results indicated that
in Co3O4/AlN, the Co metal is in the form of Co3O4, while in Co/AlN, the
Co species are composed of Co◦ and CoO.
The catalyst textural structures were analyzed by N2 adsorption/
desorption tests at ꢀ 196 ◦C, and the results were collected in Table 1 and
the adsorption/desorption isothermal were given in Fig. S1. All the
samples showed low BET surface areas due to the poor pore structure of
AlN support. The different values of the tested samples were probably
due to the error of measurement especially such a low BET surface area.
Other carriers such as CeO2 and BN supported catalysts were also tested,
and the results are given in Fig. S2 and Table S1. All the sample showed
low BET surface areas, indicating the following different catalytic per-
formances related less with the BET surface areas.
To analyze the surface composition of the catalyst, XPS was used to
investigate the surface information, which is given in Fig. 3. In Co 2p
XPS spectrum of Co3O4/AlN catalyst, the peaks appeared at 780.4,
781.9, 796.0, and 797.9 eV could be assigned to be Co3+ 2p3/2, Co2+
2p3/2, Co3+ 2p1/2, and Co2+ 2p1/2, respectively, while the peaks at 786.3
and 802.6 eV were attributed to the satellite peaks (Fig. 3a) [43,44]. The
Al 2p XPS spectrum of Co3O4/AlN was deconvoluted into two peaks at
The XRD patterns of AlN supported Co3O4 and Co catalysts are
represented in Fig. 1. In Fig. 1a, the profile of AlN showed the peaks at
33.3, 36.0, 37.9, 49.8, 59.4, 66.2, 69.8, 71.4, 72.6, and 81.0◦, which
could be attributed to the (100), (002), (101), (102), (110), (103), (200),
(112), (201), and (202) planes of AlN hexagonal wurtzite phase (ICDD
PDF Card No. 00-003-1144) [33,39–41]. No Co3O4 peak could be found
in Co3O4/AlN catalysts with the Co loadings lower than 4%, indicating
that Co3O4 was dispersed well on AlN support. After increasing the Co
loading to 6 or 8%, some new peaks appeared at 31.1, 36.8, 44.8, and
65.2◦ were assigned to the (220), (311), (400), and (440) planes of
–
–
74.1 and 74.8 eV, which were attributed to Al N and Al O species
(Fig. 3c) [45]. The N species appeared at 403 and 398.8 eV, and they
–
–
could be assigned to the N N or unbonded N and Al N species (Fig. 3e)
[46,47]. The O 1s XPS spectrum of Co3O4/AlN catalyst was also
collected, which could be deconvoluted into three peaks at 532.6, 531.6,
and 530.6 eV. These peaks were attributed to OH–, oxygen atoms in the
Table 1
The results of N2 adsorption/desorption of various catalysts.
–
neighbourhood of oxygen vacancies (VO), and M O especially the
Entry
Samples
BET suface areaa
Pore sizesb
(nm)
Total volumec
(cm3/g)
– –
O Al species, respectively [48,49].
Co
(m2/g)
The Co 2p XPS spectrum of Co/AlN showed the peaks at 779.8,
781.4, 795.1, and 797.2 eV, which were assigned to Co3+ 2p3/2, Co2+
2p3/2, Co3+ 2p1/2, and Co2+ 2p1/2, respectively. No Co◦ species was
detected from the XPS spectra because the surface Co◦ species was
partial oxidized in the air. However, the Co3+/Co2+ ratio was decreased
remarkably, indicating that after reduction Co3+ was reduced to Co2+
and/or Co◦. The Al 2p and N 1s XPS spectra of Co/AlN were almost the
same as those of Co3O4/AlN, indicating that AlN was very stable during
the reduction (Fig. 3c vs d, and e vs f). The O 1s XPS spectrum of Co/AlN
were fitted into three peaks at 532.8, 531.4, and 530.2 eV, which were
attributed to OHꢀ , oxygen atoms in the neighbourhood of oxygen va-
1
2
AlN
6
5
11.6
4.9
0.01
0.02
1% Co3O4/
AlN
3
4
5
6
2% Co3O4/
AlN
10
10
10
8
7.1
6.4
4.9
4.9
0.04
0.02
0.03
0.03
4% Co3O4/
AlN
6% Co3O4/
AlN
8% Co3O4/
AlN
7
1% Co/AlN
2% Co/AlN
4% Co/AlN
6% Co/AlN
8% Co/AlN
2
6.8
8.2
3.6
9.2
1.8
0.01
0.02
0.02
0.03
0.04
8
8
–
–
cancies (VO), Co O and/or Al O species [48,49]. Interestingly, the
9
13
20
21
ꢀ
–
–
peaks of Co O, Al O, and OH were reduced, implying that after
10
11
–
reduction, the Co O species were reduced and the oxygen vacancies
a
were generated.
Specific surface area based on the Brunauer-Emmett-Teller equation.
The pore size distribution was obtained by using the Barrett-Joyner-Halenda
b
To discover the interaction between Co species and AlN support, the
Co XPS spectra of Co3O4/AlN and Co/AlN were comparatively studied,
and the results are given in Fig. S4. Compared with the Co XPS in pure
(BJH) method.
c
The total pore volume.
3