Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
D. He, T. Wang, T. Li et al.
Journal of Catalysis 400 (2021) 397–406
was stirred at 120 °C for 12 h. Then, the autoclave was cooled to
room temperature and the pressure was carefully released. Subse-
quently, the reaction mixture was diluted with 10 mL toluene for
quantitative analysis by GC-FID (Agilent 7890B-5977A).
Procedure for recycling test. The used catalyst of the N-doped car-
bon/CuAlOx was separated from the reaction mixtures by cen-
trifuging, washed three times using methylbenzene. After being
dried in air at room temperature, it was recovered and directly
recharged into the autoclave for the next run.
It revealed that each ligand had quite different effects on the
adsorptions of reactant and product. As proposed in the Sabatier
principle [57], the binding strength between the reaction interme-
diates and catalyst is essential in determining the catalytic perfor-
mance of a catalyst. An optimal catalyst should have neither too
strong nor too weak binding strength of reaction species, and ide-
ally should be strong enough to activate the reactant but weak
enough to release the product smoothly. Based on this principle,
we can find that 1,10-phen modified Cu catalyst has such advan-
tages, i.e., this catalyst has much stronger ArNO2 adsorption
strength than the clean Cu catalyst, indicating a strong ability in
activating the reactant. Meanwhile, it has slight weaker adsorption
strength of ArNH2 than pure Cu, which indicates a more favorable
desorption of the product. Similarly, 4,7-phen also enhances the
adsorption strength of ArNO2 while weakens the ArNH2 adsorption
despite that the enhancement is smaller compared with 1,10-phen.
In contrary, ligand-2 weakens the adsorption of the ArNO2,
which will potentially reduce the ability in activating the reactant
and eventually affect the performance of Cu catalyst. Additionally,
we evaluated the effects of these ligands in H2 activation, 4,7-phen
and 1,10-phen have similar energy barriers with clean Cu catalyst
in H2 dissociation, whereas ligand-2 has much higher H2 dissocia-
tion energy barrier indicating its potentially lower hydrogenation
activity. Our simulations reveal that 1,10-phen has the most posi-
tive effect in ArNO2 hydrogenation to ArNH2 because of the
enhanced ability in both reactant activation and product desorp-
tion, followed by 4,7-phen while 1,7-phen shows a negative effect.
Indeed, all the simulated results are in reasonable agreement with
the experimentally detected yields as listed in Table 2, i.e., 1,10-
phen modified Cu has highest yield followed by ligand-3 and clean
Cu, while 1,7-phen modified Cu catalyst has lowest yield. In addi-
tion, we found that 2,20-Bipyridine has a quite different manners
compared with 1,10-phenathroline. As shown in entry 5 of Table 2,
the 2,20-Bipyridine ligand enhanced the adsorptions of both reac-
tant and product for nitrobenzene hydrogenation, which is differ-
ent from that for 1,10-phenathroline. This result further confirms
the special role of 1,10-phenathroline ligand in improving catalytic
activities. The above experiment and DFT characterization results
proved that our idea to manipulate the adsorption–desorption
behavior of reactants and products on CuAlOx surface through
introducing organic ligands was feasible.
3. Results and discussion
3.1. Screening of different ligands
At first, the hydrogenation of nitrobenzene was chosen as a
model reaction to study the effects of organic ligands on the cat-
alytic performance of supported Cu catalyst, i.e., CuAlOx (Table 1).
As shown in Table 1, we can see that only 27% yield of aniline was
obtained by CuAlOx lack of ligands and some ligands would inhibit
the catalytic hydrogenation performance of copper, like pyridine
and dipyridine, but some ligands, especially 1,10-phen, could sig-
nificantly improve the catalytic performance with 81% yield
(Table 1, Entry 1, 11). It should be noted that the catalytic perfor-
mance was possibly related to the position or distance of two N-
atoms in organic ligand.
3.2. DFT calculation
To validate our hypothesis further and get some insights into
the different catalytic performances between clean Cu and Cu dis-
posed by 1,10-phen, we performed DFT studies to compare the
adsorption energies of ArNO2 and ArNH2 on clean Cu and Cu dis-
posed by 1,10-phen.
It revealed that the adsorption of ArNO2 on 1,10-phen_Cu (111)
surface (ꢂ1.27 eV) was much stronger than on clean Cu (111) sur-
face (ꢂ0.72 eV), which indicated a higher concentration of reactant
on the surface and eventually enhanced activation of the reactant.
The adsorption energy of ArNO2 on 1,10-phen_Cu (111) surface
(ꢂ1.27 eV) is also much stronger than 1,10-phenathroline
adsorption energy (ꢂ1.01 eV). In this respect, full coverage of
1,10-phenathroline is not competitive at all in the presence of
nitrobenzene. Therefore, we concluded that the partial precovering
of 1,10-phenathroline on Cu surface is beneficial for nitrobenzene
adsorption. Close inspections showed that the enhanced adsorp-
tion strength came from the formation of H-bonds between ArNO2
and phen as shown in Fig. 2c. However, the adsorption of the
ArNH2 was slightly weaker on 1,10-phen_Cu (111) surface
(ꢂ0.89 eV) than the clean Cu (111) surface (ꢂ0.91 eV). Therefore,
we attributed the better performance of 1,10-phen_Cu catalyst in
this reaction partially to the enhanced adsorption strength of reac-
tant as well as desorption of product. Furthermore, we also evalu-
ated the abilities of these two catalysts in H2 activation, and the
potential energy diagram, as well as structures of H2 dissociation.
The results were shown in Figs. S1 and S2. The 1,10-phen_Cu
(111) surface has a higher H2 dissociation energy barrier than Cu
(111) surface, indicating the slightly weaker activity in H2
activation.
3.3. Optimization of the reaction
In subsequent studies, we conducted optimization of the reac-
tion conditions (Table 3). It was found that 99% nitrobenzene con-
version and 99% aniline yield were achieved by using CuAlOx in the
presence of 1,10-phen at 120 °C with 1.5 MPa H2 for 12 h (Table 3,
Entry 9). Noteworthy, the catalyst exhibited good reusability for at
least 5 runs at full conversions by simple filtration using solvent
without further addition of 1,10-phen (Fig. 4a). The recycling test
of the catalyst has also been performed at relatively low conver-
sions (17%–19%) within kinetic region, and it can be found that
the catalyst could be reused at least 5 times without obvious loss
in catalytic performance (Fig. 4b).
Moreover, we try to construct CuAlOx-M as catalyst for the
reduction of nitroarenes (Table S5). Clearly, the best catalytic per-
formance was obtained by treatment of CuAlOx with 1,10-phen in
the presence of nitrobenzene for 8 h (Table S5, entry 11). With the
macromolecular modification of supported Cu catalysts (CuAlOx-
M) in hand, the scope and limitation of reduction of nitroarenes
were investigated. As shown in Table 4, nitroarenes with both
electron-withdrawing and electron-donating groups on the aro-
matic ring were effectively reacted to afford the desired products
in 98–99% yields (2a–2i). It was found that the electronic and steric
properties of the aryl substituents on the aromatic ring have little
To get more insights into the roles of different ligands in affect-
ing the catalytic performances of Cu catalyst, we performed sys-
tematic DFT computations and all the computational details
could be found in supporting information. As a start, we compared
the adsorption energies of ArNO2 (reactant) and ArNH2 (product)
as well as H2 dissociation energy barriers on clean and different
ligands modified Cu (111) surfaces (Fig. 3).
399