Inorganic Chemistry
Article
framework, which are responsible for the catalytic activity,
while the framework is exclusively used as innocent porous
support to prevent the aggregation and deactivation of active
sites.
centrifugation and employed in subsequent reaction cycles
without apparent loss in activity: kZnBDP = 9.3 × 10−3, 8.8 ×
10−3, 8.2 × 10−3, and 7.8 × 10−3 h−1 for the four subsequent
Therefore, the cumulative rate constant per Zn active site
after three cycles is 3.8 times higher when using the ZnBDP
MOF (0.15 h−1) with respect to the homogeneous Zn(NO3)2
catalyst (0.04 h−1). The high stability of both MOFs is proven
due to the maintenance of XRD patterns after 72 h of reaction,
with only minor changes attributed to changes in pore
occupancy (see Figure 3).38 A hot filtration test of ZnBDP
Here we study for the first time the activity of unmodified
ZnBDP as a heterogeneous catalyst, based on the reported
Zn(NO3)2 homogeneous C−Cl amination catalyst.4 This
framework consists of tetrahedrally coordinated N4Zn metal
centers interconnected by BDP linkers to generate 3D square
channels with edges of 13.2 Å, large pores of close to 11.1 Å
(see Figure 1a).36−38 In order to account for the changes of
porosity over substrate diffusion and performance, the zinc-
pyrazolate ZnBDP was compared with that of zinc-imidazolate
ZIF-8. This choice is based on the similar environment of the
ZnN4 units, but featuring a different porous architecture with
large 11.6 Å pores connected through small apertures of 3.4 Å
(see Figure 1b).39
RESULTS AND DISCUSSION
■
We tested both Zn reticular catalysts in the proof-of-concept
amination of the sp2 C4 carbon of 4-chloropyridine (Cl-py)
with morpholine. The yield of the 4-aminated pyridine product
after 72 h of reaction increases in the order ZIF-8 (26%) <
ZnBDP (51%) < Zn(NO3)2 (82%), for the same catalyst mass
(see Figure 2a). The kinetic rate constants were obtained from
Figure 3. Powder XRD patters of ZnBDP (a) with respect to ZIF-8
(b) catalysts before and after the amination of Cl-py with morpholine.
demonstrated the heterogeneous nature of the reaction (see
filtrate results in 0.21 wt % Zn leaching from ZnBDP. Taking
into account that 13 wt % of Zn in the MOF structure is used
with respect to 4-chloropyridine, only 0.03 wt % Zn will be in
the form of leached homogeneous Zn2+. This amount is way
below the 1.4 wt % Zn used in the reported homogeneous
catalyzed process,4 confirming the results of the hot filtration
tests. Furthermore, no evidence of 4-pyridination of 1,4-
bis(pyrazol-4-yl)benzene ligand was found according to the
NMR spectra of the digested MOF before and after the
reaction (see Figure S7). Therefore, the changes in catalytic
performance between single-site Zn(II) frameworks cannot be
ascribed to the potential degradation of ZnBDP but to
differences imposed by the dimensions of their pores and
channels.
Figure 2. Kinetic profiles (a) and activity (TONZn) expressed in mol
of Cl-py converted divided per mol of Zn (b), for the amination of 10
mg of Cl-py catalyzed by 10 mg of ZnBDP (red) or ZIF-8 (blue) with
respect to Zn(NO3)2 (black).
To further confirm the key role played by the diffusion of
the substrates into the ZnN4-MOF active sites, solid−liquid
adsorption experiments of a solution of 4-chloropyridine (Cl-
py) in CH3CN (2.5 mg·mL−1) were performed. The ZnBDP
the linear fit of the logarithm of the 4-chloropyridine
conversion vs time, assuming a pseudo-first-order reaction,
which resulted in different slopes (see Figure S5 in the
kZIF‑8 (0.004 h−1) < kZnBDP (0.009 h−1) < kZnNO3 (0.024 h−1).
For the same Zn-catalyst mass, the Cl-py/Zn molar ratio
increases in the order Zn(NO3)2 < ZIF-8 < ZnBDP (being 1.8
< 2.3 < 4.5, respectively). When the kinetic rate constants are
multiplied by the corresponding substrate/catalyst ratio, a
normalized rate constant per Zn active site is obtained. Its
value is similar for the heterogeneous ZnBDP and the
homogeneous Zn(NO3)2 (ca. 0.04 h−1) and four times higher
than in the case of microporous ZIF-8 (ca. 0.01 h−1).
The number of product molecules per zinc site (TONZn)
during the course of the reaction is represented in Figure 2b. It
indicates the good catalytic performance of ZnBDP with
respect to ZIF-8 and the Zn(NO3)2 homogeneous catalyst.4 In
contrast to the soluble Zn2+ species, which are not possible to
recover and recycle, ZnBDP can be isolated by simple
shows a higher uptake with respect to ZIF-8, resulting in 1.0 vs
−1
0.3 molCl‑py·molZn
with both reticular Zn catalysts (i.e., Zn12(C4N2H5)24 and
Zn4(C12N4H8)4 modeled crystalline cells of ZIF-8 and ZnBDP,
respectively) in order to understand the experimental results.
The calculated values are in line with the experimental uptake
of Cl-py molecules by the MOFs, corresponding to 0.4 and 1.0
−1
molCl‑py or morpholine·molZn docked in ZIF-8 and ZnBDP,
respectively. More importantly, the molecular dynamics results
indicate a higher diffusivity in the ZnBDP with respect to ZIF-
8, especially in the case of the amine nucleophile (2 × 10−5 vs
1 × 10−7 cm2·s−1), as shown in Figure 4 (molecule A). The
adsorption and diffusion results further confirm that the
changes in the pore window of ZnBDP and ZIF-8 may be the
origin of their different catalytic performance. On the one
hand, the 1.3 nm edges of the square channels from ZnBDP
favor the infiltration and docking of Cl-py molecules inside the
B
Inorg. Chem. XXXX, XXX, XXX−XXX