F. Li et al.
Journal of Solid State Chemistry 301 (2021) 122337
Zn-ADBA shows broad absorption bands range from the ultraviolet to
visible light regions with a maximum absorption peak of 406 nm in the
solid state, which maybe correspond to the n–π* and π–π* transition of
Table 1
Optimization of the reaction conditions and control catalytic experiments for the
Zn-ADBA catalyzed CDC reaction.a
H2ADBA ligands [45]. Based on the Kubelka-Munk function [46], the
optical band gap of Zn-ADBA was determined to be 2.651 eV (Fig. 4a
inset), exhibiting great photocatalytic potential. Fig. 4b shows the tran-
sient photocurrent intensities of Zn-ADBA with several repeating cycles
of intermittent on-off irradiation. It is clear that the Zn-ADBA catalyst
displays a powerful transient photocurrent response to visible light,
which could be attributed to the effective generation, separation, and
migration of photoinduced charge carriers in Zn-ADBA [47]. Therefore,
the visible-light driven aerobic cross dehydrogenative coupling (CDC)
reaction of N-phenyl-tetrahydroisoquinoline and nitromethane was
selected as a model reaction to check the photocatalytic performances of
Zn-ADBA. As shown in Table 1, upon irradiation (30 W fluorescent lamp)
at 25 ꢁC in the presence of air as oxygen source over 4 h, the reaction
yields quickly increased with the amount of catalyst increasing from 1 to
5 mol%, and no obvious yield increase was found when the amount
further increases to 7 mol% (Fig. S7). It was also found that the yield of
the extended reaction did not increase significantly after 4 h, suggesting
that a 5 mol% loading of catalyst and a reaction time of 4 h are the
optimal reaction conditions, and the reaction system giving a high yield
of 97.6% (Fig. S8). The catalyst was filtered out after 2 h of reaction, the
filtrate afforded only 8% additional yield for another 2 h at the same
reaction conditions, confirming the typical inhomogeneity of the pho-
tocatalytic reaction.
Control experiment revealed that almost no product was observed
when the reaction was carried out in the dark, and only a 4.3% of yield
was detected when there was no catalyst, which demonstrated that both
the photocatalyst and the light are indispensable for the efficient progress
of the coupling reaction. When the same equivalent of metal or ligand is
used, the yields are 17.8% and 19.3%, respectively, and the yield of the
mixture system of metal and ligand (1:2) is only 26.6% (Table 1). In
addition, the stability of Zn-ADBA for photocatalytic conversion can be
confirmed by recycling experiments (Fig. S9). The solid of Zn-ADBA
separated from the reaction suspension by simple filtration can be reused
at least five times with insignificant loss of activity (84.3% of yield after
five cycles). The PXRD patterns of Zn-ADBA isolated from the reaction
mixture further confirmed the well-retained skeleton (Fig. S10). It has
been reported that both the superoxide radical anion (Oꢃ2-) and singlet
oxygen (1O2) may be the key active species in the coupling reaction. As a
result, the electron paramagnetic resonance (EPR) tests were performed
to confirm the active oxygen species in the photocatalytic reaction by
employing 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) and 2,2,6,6-tetra-
methyl-1-piperidine (TEMP) as catching agents. As shown in Fig. 4c,
no signal was observed upon irradiation of the mixture containing TEMP
and Zn-ADBA in the absence or presence of N-phenyl-tetrahy-
droisoquinoline, indicating that no 1O2 was generated in the reaction
process. There was also no signal when DMPO was added to a CH3NO2
solution of Zn-ADBA, whereas the signal of Oꢃ2- captured by using DMPO
as a radical scavenger was clearly detected in the presence of N-phenyl-
tetrahydroisoquinoline. The results demonstrate that Oꢃ2- is the active
species during the catalytic process.
Entry
Catalyst
Zn-ADBA
Zn-ADBA
Zn-ADBA
Zn-ADBA
Zn-ADBA (in dark)
None
Cat (mol %)
Yield (%)
47.9
79.8
97.6
98.1
trace
4.3
17.8
19.3
26.6
1
2
3
4
5
6
7
8
9
1
3
5
7
5
_
5
5
5
H2ADBA
Zn(Ac)2⋅2H2O
Zn(Ac)2⋅2H2O þ H2ADBA
1H NMR analysis.
a
The yields were determined by.
there was no further conversion in the filtrate for another 10 h at the
same conditions (Fig. 2a). The result indicated that there were no active
species leaching into the catalytic system and Zn-ADBA was a true het-
erogeneous catalyst. In addition, the recyclability of the MOF catalyst
was demonstrated by five cycles of experiments. Upon completion of the
reaction, the catalyst could be easily isolated from the reaction suspen-
sion by simple centrifugation and washing, and directly reused in a new
round of reaction. As depicted in Fig. 2b, Zn-ADBA solids can be reused
at least five times without a significant loss of reactivity (88.6–82.7%
yield). The PXRD patterns of the reused catalyst matched well with the
simulated patterns from the single crystal X-ray diffraction data of Zn-
ADBA, showing that the framework and crystallinity of Zn-ADBA were
well maintained after the catalytic reaction (Fig. 2c). Control experi-
ments revealed that a very low background reaction (<5% yield) was
observed in the absence of catalyst, and the only use of same equiv. of
ligand H2ADBA or Zn(Ac)2⋅2H2O led to a yield of 18.8% or 20.9%,
respectively. Even when a mixture of metal and ligand (1:2) was added,
only a small increase to 26.6% yield was obtained (entries 16–18,
Table S2). These results suggested that the special coordination between
ADBA2- and Zn(II) played a critical role in the catalytic reaction, which
greatly activated the aldehyde group of substrates as Lewis acid catalytic
site. Such carbonyl activation was further confirmed by the IR spectrum
of Zn-ADBA immersed in a DCM solution containing benzaldehyde
–
–
(Fig. 2d). The shift of the characteristic C O stretching vibration peak
from 1690 cmꢀ1 (free benzaldehyde) to 1680 cmꢀ1 indicated that the
catalyst has an effective activation effect on the substrate benzaldehyde
When para-substituted aromatic aldehydes were used as substrates
(1a–1e), a broad yield range of 33.6–95.6% was achieved in the Zn-
ADBA reaction systems, suggesting a significant electron effect on the
reaction yield. The aryl aldehydes with electron withdrawing nitro- and
brom-species at the para position gave high reaction yields of 92.3% and
95.6%, respectively, while those with electron donating dimethylamino-
and hydroxyl-groups at the para position respectively exhibited very low
yields of 33.6% and 35.9% (Fig. 3). The result may be because that the
electron cloud is closer to the electron-withdrawing species to expose the
lone pair of electrons of carbonyl oxygen when electron-withdrawing
groups substituted benzaldehydes used, leading to high activity of the
aldehyde group [42–44]. Additionally, for the nitro-substituted benzal-
dehydes, the effects of steric hindrance are also investigated (1e–1g). The
reaction yield decreased with the increase of the steric hindrance, and the
o-nitrobenzaldehyde system gave a lowest yield of 59.6%.
Based on the experimental results and previous reports [48,49], a
possible reaction mechanism is proposed (Fig. S11): upon illumination,
Zn-ADBA changes to its excited state through the electronic transition,
and the electrons in the excited state lack electrons to capture the elec-
trons on the N atom in N-aryl-tetrahydroisoquinoline to form ammoni-
ated cationic groups. The electron-derived Zn-ADBA reacts with oxygen
to obtain a superoxide anion radical (Oꢃ2-), and O2 oxidizes the sp [3]
C–H bond of the nitrogen atom in the adjacent amine to obtain an imine
ion intermediate. On this basis, the imine ion is captured by a nucleophile
to obtain the corresponding C–C coupling product. Herein, the proba-
bility of electron transfer between Zn-ADBA and N-phenyl--
tetrahydroisoquinoline was calculated by density functional theory
(DFT) to verify the reaction possibility. DFT calculations (Fig. 4d) show
3.3. Photocatalytic C–C coupling reaction of N-aryl-tetrahydroisoquinoline
with nitromethane
Before testing the photooxidation activity of Zn-ADBA, the photo-
response capacity of Zn-ADBA was investigated by UV–Vis diffuse
reflectance spectroscopy and photocurrent analyses. As shown in Fig. 4a,
5