Macromolecules
Article
entry 5), which suggested that Tx-COF-2 has a significant part
as a photocatalyst. Similarly, the transformation of the reaction
was hard to observe when conducted in the absence of visible
light and the nitrogen atmosphere (Table 1, entries 3 and 4). It
could be concluded from all these control experiments that
light source, oxygen, and photocatalyst are necessary for this
reaction.
The superiority of Tx-COF-2 as a photocatalyst was further
examined by recycling experiments. Because of the insoluble
nature of Tx-COF-2 in all organic solvents, it was simply
removed from the reaction mixture by centrifugation after the
completion of the reaction. The experiment was performed on
4-carboxyphenylboronic acid. As shown in Figure 3a, without
any reactivation procedure or specific treatment, Tx-COF-2
gave excellent results for the conversion of 4-carboxy-
phenylboronic acid to 4-hydroxybenzoic acid even after 10
runs as a photocatalyst. The crystallinity of recycled Tx-COF-2
catalyst was fully comparable with that of the as-synthesized
Tx-COF-2, as observed from PXRD, with a small decrease in
the intensity69,70 (Figure S12). Furthermore, after the 10th
catalytic run, Tx-COF-2 retained its microporous nature with a
slight drop in the BET surface area, as shown by the nitrogen
adsorption and desorption isotherms (Figure 3b, Figures S13
conversion is shown in Figure 5. Initially, an excited
intermediate Tx-COF-2* was formed, upon irradiation of
visible light, followed by an electron extraction from DIPEA by
single-electron transfer (SET) pathway to generate a radical
anion Tx-COF-2•− which left a radical cation DIPEA•+. Then
the Tx-COF-2•− was oxidized by molecular oxygen to restore
•−
Tx-COF-2, used in upcoming cycles. In addition, O2 was
filled in the vacant p-orbital of boron atom, leading to the
generation of A. A hydrogen atom was pulled out from
DIPEA•+ to produce intermediate B. Next, loss of a −OH− ion
and rearrangement formed intermediate C. Finally, intermedi-
ate C hydrolyzed to afford the desired phenolic product D.
As mentioned earlier, the Tx-COF-2 was very stable in
contact with 12 M HCl, which also caused a visual change of
the color from yellow to orange. This phenomenon motivated
us to examine the sensing properties of Tx-COF-2 toward
HCl. Upon exposure to HCl vapor, the suspension (0.5 mg
mL−1 in 1,4-dioxane) of Tx-COF-2 went through a naked-eye
color change from light yellow to orange. When the HCl
treated suspension was subjected to light at 365 nm of
wavelength, an orange fluorescence emission color was also
observed (Figure 6a). This color change was completely
reversible when HCl treated Tx-COF-2 exposed to the NH3
vapor. The color changes occurred in fractions of a second
faster than those previously reported COF-based HCl
sensors.95,96 Notably, as-synthesized Tx-COF-2 was very
weakly fluorescent, but the HCl-treated suspension of Tx-
COF-2 showed remarkable enhancement in the fluorescence
intensity. Thus, Tx-COF-2 could be used as a “turn-on” sensor
because it offered a high increment in the fluorescence
intensity after treating with HCl vapor and reverted to its
pristine weakly fluorescent nature after exposure to NH3 vapor.
Besides, no degradation in the fluorescence intensity as well as
in naked-eye color changes up to 10 alternating cycles of
exposure to HCl and NH3 vapor confirmed the outstanding
reversibility and structural integrity of Tx-COF-2 as HCl
sensor. We then investigated the relationship between the
fluorescent intensity of 1,4-dioxane suspended Tx-COF-2 (0.5
mg mL−1) and the HCl concentration (diluted in 1,4-dioxane).
We observed that upon the addition of a very little amount of
HCl concentration (8 μmol L−1) to Tx-COF-2 suspension the
weak-intensity peaks at 495 and 515 nm started disappearing,
and a new peak appeared at 558 nm (λex = 478 nm) (Figure
6b). The fluorescence intensity enhanced gradually upon
increasing the HCl concentration from 8 to 65.3 μmol L−1 and
remained unchanged after that. The lowest limit of detection
was found to be 4.5 nmol L−1 from the calibration curve in the
range of HCl concentration from 1 to 55 μmol L−1, which
showed linear correlation (R2 = 0.9864) with detection at 558
nm (Figure 6e). The fluorescence lifetime decay of Tx-COF-2
was found to be 0.49 ns at 65.3 μmol L−1 HCl concentration
(Figure S18). A possible mechanism of sensing toward HCl is
shown in Figure 6f. The Tx-COF-2 presumably undergoes
protonation at the imine nitrogen atom, which takes part in the
conjugation of the COF and results in an apparent change in
fluorescence emission as well as in color.97 Previously reported
COFs support the fact of protonation at the nitrogen atom
present in the COFs.98 Next, the reusability and recyclability of
Tx-COF-2 as HCl sensor were investigated and shown in
Figure S16. Tx-COF-2 was recovered by the addition of a
solution of triethylamine (1.6 mmol L−1) from the HCl-treated
Tx-COF-2 at a concentration (1.6 mmol L−1) until the pH is
equal to 7. Moreover, the selectivity of the Tx-COF-2 toward
We then tested the versatility of the Tx-COF-2 as a
photocatalyst for the conversion of different derivatives of
arylboronic acid to the corresponding phenols. As shown in
Table 2, Tx-COF-2, as a photocatalyst, can easily convert all
the substituted arylboronic acids to the corresponding phenols.
Usually, the reaction rate is faster with the substrates bearing
electron-deficient groups (Table 2, entries 1−5) than that of
the substrates bearing electron-rich groups (Table 2, entries 6
and 7). In reported literature, this selectivity has been
explained according to the fact that the empty p-orbital of
boron atoms present in electron-deficient arylboronic acid
substrates are more available for the O2 radical anion.94
•−
To get a deeper understanding of the photocatalytic
procedure, various electronic properties of Tx-COF-2 were
studied. Initially, the UV−vis absorption spectrum of Tx-COF-
2 showed a broad light absorption edge in the range 300−700
nm, indicating that Tx-COF-2 can absorb light in the visible
region (Figure 4a). The optical band gap (Eg) of Tx-COF-2
was determined by the Kubelka−Munk plot and estimated to
be 2.65 eV, indicating its semiconducting nature (Figure 4a,
inset). Cyclic voltammetry (CV) was further recorded to
determine the oxidative potential of Tx-COF-2. The CV
experiment displayed an oxidation peak with an onset value of
+0.93 V versus the saturated calomel electrode (SCE), which
corresponds to the valence band potential (EVB) (Figure
S10).70 The approximate conduction band potential (ECB) was
evaluated by using the formula ECB = EVB − Eg. Thus, the ECB
value of Tx-COF-2 was estimated to be −1.72 V. The
reduction potential (Ered) of O2 and the oxidation potential
(Eox) of DIPEA versus SCE are −0.86 V and +0.90 V,
respectively. For oxidation of DIPEA and reduction of O2, a
photocatalyst is expected to possess a EVB value greater than
Eox of DIPEA and a ECB value less than Ered of O2.94 The ECB
and EVB of Tx-COF-2 entirely meet the above-mentioned
criteria (Figure 4b). In addition, electron paramagnetic
resonance (EPR) spectroscopy was performed to confirm the
•−
generation of O2 upon addition of the superoxide radical
scavenger 5,5-dimethyl-1-pyrroline N-oxide (DMPO), shown
in Figure S11. A probable reaction mechanism for the
6600
Macromolecules 2021, 54, 6595−6604