Highly Stable COF-Supported Co/Co(OH)2 Nanoparticles Heterogeneous Catalyst for Reduction of Nitrile/Nitro Compounds under Mild Conditions

Article abstract of DOI:10.1002/smll.201801233

Ordered nanoporosity in covalent organic framework (COF) offers excellent opportunity for property development. Loading nanoparticles (nPs) onto them is one approach to introducing tailor-made properties into a COF. Here, a COF–Co/Co(OH)₂ composite containing about 16 wt% of <6 nm sized Co/Co(OH)₂ nPs is prepared on a N-rich COF support that catalyzes the release of theoretical equivalence of H₂ from readily available, safe, and cheap NaBH₄. Furthermore, the released H₂ is utilized for the hydrogenation of nitrile and nitro compounds to amines under ambient conditions in a facile one-pot reaction. The COF “by choice” is built from “methoxy” functionalized dialdehydes which is crucial in enabling the complete retention of the COF structure under the conditions of the catalysis, where the regular Schiff bonds would have hydrolyzed. The N-rich binding pockets in the COF ensure strong nP–COF interactions, which provides stability and enables catalyst recycling. Modeling studies reveal the crucial role played by the COF in exposing the active facets and thereby in controlling the activation of the reducing agent. Additionally, via density functional theory, we provide a rational explanation for how these COFs can stabilize nanoparticles which grow beyond the limiting pore size of the COF and yet result in a truly stable heterogeneous catalyst – a ubiquitous observation. The study underscores the versatility of COF as a heterogeneous support for developing cheap and highly active nonnoble metal catalysts.

Full text of DOI:10.1002/smll.201801233

FULL PAPER
Heterogenous Catalysts
www.small-journal.com
Highly Stable COF-Supported Co/Co(OH) Nanoparticles
2
Heterogeneous Catalyst for Reduction of Nitrile/Nitro
Compounds under Mild Conditions
Dinesh Mullangi, Debanjan Chakraborty, Anu Pradeep, Vijay Koshti, Chathakudath P. Vinod,
Soumendranath Panja, Sunil Nair, and Ramanathan Vaidhyanathan*
1
. Introduction
Ordered nanoporosity in covalent organic framework (COF) offers excellent
opportunity for property development. Loading nanoparticles (nPs) onto
them is one approach to introducing tailor-made properties into a COF.
Covalent organic framework (COF) owing
to their highly ordered structure, chemical
tunability, and porosity can serve as excel-
lent supports for metal nanoparticles
Here, a COF–Co/Co(OH) composite containing about 16 wt% of <6 nm
2
sized Co/Co(OH) nPs is prepared on a N-rich COF support that catalyzes
nPs).[1–6] Recently, we showed that COFs
2
(
the release of theoretical equivalence of H from readily available, safe, and
2
can serve as porous host for growing
small uncapped nPs. The resulting com-
posite showed exceptional stability and
high catalytic activity for multifold Heck
couplings, CO reduction, and oxygen evo-
cheap NaBH . Furthermore, the released H is utilized for the hydrogenation
4
2
of nitrile and nitro compounds to amines under ambient conditions in a facile
one-pot reaction. The COF “by choice” is built from “methoxy” functionalized
dialdehydes which is crucial in enabling the complete retention of the COF
structure under the conditions of the catalysis, where the regular Schiff bonds
would have hydrolyzed. The N-rich binding pockets in the COF ensure strong
nP–COF interactions, which provides stability and enables catalyst recycling.
Modeling studies reveal the crucial role played by the COF in exposing the
active facets and thereby in controlling the activation of the reducing agent.
Additionally, via density functional theory, we provide a rational explanation
for how these COFs can stabilize nanoparticles which grow beyond the
limiting pore size of the COF and yet result in a truly stable heterogeneous
catalyst – a ubiquitous observation. The study underscores the versatility
of COF as a heterogeneous support for developing cheap and highly active
nonnoble metal catalysts.
[
4,7,8]
lution from alkaline solutions.
On a
different note, recent works highlight the
potential interplay between magnetism
and catalysis in hydrogenation and CH
[
9,10]
bond activation reactions.
Encouraged
by these findings, we realized, if magnetic
nPs can be grown in the nanoconfine-
ments of a COF it might yield catalyti-
cally active “nanomagnets.” The modular
construct of a COF enables controlled
infusion of heteroatoms, which is a key to
functionalizing the COF for binding with
first- and second-row transition metals.
While there are reports on the use of
COF-supported noble metal (Pd, Au) as
[
3–6]
heterogeneous catalyst,
utilizing N-rich
D. Mullangi, D. Chakraborty, A. Pradeep, Prof. R. Vaidhyanathan
COF as support for nonnoble metal catalyst (Co, Ni) is fairly
Department of Chemistry
[7,8,11]
recent.
Indian Institute of Science Education and Research
Dr. Homi Bhabha Road, Pune 411008, India
E-mail: vaidhya@iiserpune.ac.in
Here, we demonstrate the effectiveness of this approach
(nP@COF) in designing catalyst for on-demand release of H2
from a very safe and cheap hydrogen source, namely, NaBH4
(high gravimetric (>4.0 wt%) and volumetric (30 wt% solu-
tion of NaBH ≈63 g H /L with no by-products such as CO
Dr. V. Koshti
Polyolefin Lab
Polymer Science and Engineering Division
CSIR-National Chemical Laboratory
Dr. Homi Bhabha Road, Pune 411008, India
4
2
[
12,13]
and CO2).
NaBH is stabler and cheaper than the well-
4
[
14,15]
investigated hydrogen storage material, the NH :BH .
Also, NaBH can release 4H per formula unit, compared with
3H by NH :BH . Usually, the hydrogen release from NaBH
is catalyzed by noble metals
cobalt as a hydrogen-releasing catalyst. Thus, there is a clear
need for efficient heterogeneous catalysts for H evolution from
NaBH . Developing such catalysts from Earth-abundant and
environmentally benign materials adds substantial value. The
activity of first-row transition metals in these types of redox reac-
tions is quite sluggish. To enhance their catalytic activity, they
3
3
Prof. C. P. Vinod
CSIR-NCL Catalysis and Inorganic Chemistry Division
Dr. Homi Bhabha Road, Pune 411008, India
4
2
2
3
3
4
[
16,17]
S. Panja, Prof. S. Nair
Department of Physics
and very little is known about
[18]
Indian Institute of Science Education and Research
Dr. Homi Bhabha Road, Pune 411008, India
2
4
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/smll.201801233.
DOI: 10.1002/smll.201801233
Small 2018, 1801233
1801233 (1 of 8)
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
are either engineered into high surface area nanostructures or
role of the alumina support. One drawback, in both MOF and
alumina catalyzed nitrile to amine conversions, is the need to
[19]
grown as small nPs using capping agents. But such capping
is known to inhibit access to the nP surfaces. In this regard,
COF can play an excellent role of being a high surface area
support for such catalytically active nPs.
employ 40 bar pressurized H as the reducing medium and
2
temperatures in the range of 85–110 °C. In comparison to these
inorganic supports, carbonaceous porous materials are advanta-
geous owing to their ability to absorb small organic molecules
The catalytic hydrogenation of nitriles is an atom-economic
route to amines.[ Nitriles are typically reduced to amines
20]
and can thereby boost catalytic activity. In case of COF, they
[13]
[
21]
using metal hydrides, hydrosilanes, or organocatalysts. But,
these methods have a poor catalyst economy and low functional
group tolerance. The most atom economy is realized from using
H . Undesirably, reductions using H are typically carried out
can be chemically tuned to mimic the solvent environments of
[
4]
a homogeneous catalyst. Hence, developing a heterogeneous
reduction catalyst with an atomically tunable organic support to
minimize the catalytic sluggishness of typical first-row transi-
tion metals under ambient conditions is still rewarding.
In this backdrop, here we present a cobalt-COF catalyst (Co@
2
2
using noble metal catalyst and under elevated temperatures and
pressures, which can prove unsafe (Table S1, Supporting Infor-
mation). While there are many reports on the reduction of nitro
compounds using NaBH or NH :BH type of solids, very few
COF) which releases H from an aqueous/methanolic solution
2
of NaBH , and the released H is used in one-pot fashion to
4
3
3
4
2
reports are available on their use in nitrile reduction (Table S1,
carry out these industrially valuable nitro/nitrile hydrogenations.
Supporting Information). Few cobalt-based heterogeneous cata-
lysts have been used for various reduction reactions.[
22–25]
How-
ever, the reductions using NaBH are usually achieved by use
4
2
. Results and Discussion
of Pd, Pt, and Au catalysts.[
26–28]
These are sensitive to moisture
and can be ignited, particularly under elevated pressure and
temperature (Tables S1 and S2, Supporting Information).
The most recent report comes from Lin and co-workers, who
demonstrated the use of a Co–Zr–MOF as a single-site cata-
lyst for the nitrile to amine reduction.[ While Beller and
co-workers found a high activity for this reaction in a catalyst
IISERP-COF5 is synthesized via solvothermal methods employing
a triazine triamine and 2,5-dimethoxy terephthaldehyde as shown
in Figure 1A. The nPs were loaded into the IISERP-COF5 using
a double solvent method involving hexane and a methanolic solu-
21]
[15]
tion of CoCl2. The addition resulted in a sharp color change
from yellow to pinkish-orange, which is retained even when the
solid is washed and dried, thus indicating an intimate loading
formed by supporting 1,10-phenanthroline decorated Co(OAc)
2
[29]
2+
on α-Al O .
In their work, they emphasize the crucial
of Co ions onto the solid. This solid, when reduced with
2
3
Figure 1. A) Schematic illustration of the synthesis of IISERP-COF5 and proposed topology structures. B) A perspective view of the 1D channels in the
most-probable structure of IISERP-COF5, the hexagonal layers stack in an eclipsed fashion with an interlayer separation of 3.5 Å. C) Fit from Pawley
refinement; unit cell of IISERP-COF5 (Formula: C H O N ): Sp.gr. P6, a =b = 37.21; c = 3.54 Å.
7
2
48
6
12
Small 2018, 1801233
1801233 (2 of 8)
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
NaBH (aq) yielded the composite (Co@COF) as a blackish-brown
colored fine powder (Figure S1, Supporting Information).
and triazine carbons (Figure S5, Supporting Information).
The signal observed at ≈159 ppm indicates the formation of
the imine(CN) bond which propagates to form the conju-
gated COF. These features corroborate well with the functional
groups finger-printed from the Fourier transform infrared
spectra: parent COF shows characteristic stretching bands
4
2
.1. Structure Solution
−
1
A 2D structure made up of hexagonal layers that is consistent
with the experimentally observed powder X-ray pattern was mod-
eled using the Material Studio program. For the structure solu-
for CN Schiff bond (1615 cm ) and aromatic triazine CN
−1
(1598 cm ); these bands are observed for the nP-loaded COF
suggesting stability of the COF toward the reduction procedure
(Figure S6, Supporting Information). Methoxy groups were
introduced in the COF owing to its ability to increase the sta-
[4]
tion, we adopted the same routine as in our earlier work. The
XCELL program was used to index the experimental powder
X-ray diffraction (PXRD) pattern and to identify the best possible
space group as being P6, for the structure including the hydrogen
atoms. The geometry optimized and energy minimized final
structure, with an eclipsed configuration, was obtained using
tight binding density functional theory (DFT; Figure 1B). The
experimental PXRD pattern could be refined against this model
using a Pawley routine. An excellent fit was obtained (Figure 1C).
Also, the pore dimensions observed for this eclipsed model agree
well with the experimentally determined pore sizes.
[
31]
bility at the Schiff links via its electron-donating character.
The COF poses remarkable stability to harsh treatments:
soaking in highly polar DMF, DMSO, MeOH, nonpolar ben-
zene, and acid (3 m HCl) and base (3 m NaOH) for 7 d and even
in boiling water (Figure 2C). This is crucial when it comes to
applying these COFs as a catalyst support for organic reactions
carried out in solution medium. The cobalt loading onto this
2
+
COF was achieved via an aqueous phase reduction of Co sup-
ported on the COF. The X-ray diffraction (XRD) pattern of Co-
loaded COF matches well with as made COF (Figure 2A); the
peaks corresponding to Co metal at higher 2θ were not seen,
[
7]
2
.2. COF and Co@COF Characterizations
suggesting many of them forming as small nPs.
A 77 K N2 adsorption isotherm of IISERP-COF5 shows
completely reversible type-2 profile (Figure 2). We attribute this
high uptake to the presence of highly ordered structure reflected
in the high degree of crystallinity observed in the PXRD.
Though they display a fluffy cotton-like morphology under the
SEM, they appear to be aggregates of flaky crystallites under
the TEM (Figures S8–S10 Supporting Information). At higher
resolutions, the small micropores are seen. A model-inde-
pendent Barrett–Joyner–Halenda (BJH) fit to the desorption
The IISERP-COF5 could be made in bulk quantities (grams)
with high crystallinity and purity as indicated by the PXRD
30]
(
Figure 2A).[
Thermogravimetric analysis (TGA) and the
variable temperature PXRD (VT-PXRD) of IISERP-COF5 dis-
play their exceptional thermal stability up to temperatures as
high as 400 °C. Solid-state 13C-NMR (500 MHz) spectra of
IISERP-COF5 show eight different peaks between δ = 114.57
and 159.90 ppm, which were assigned to the aromatic benzene
Figure 2. A) A comparison of the PXRDs of the COF and the Co@COF. Inset shows the thermal stability of the COF till 400 °C from the TGA and VT-PXRD.
B) N isotherms (77 K) of the COF and the Co@COF. Inset shows the PSD from a BJH fit. C) Stability of the COF toward acids, bases, and solvents from
2
PXRD studies. D) HR-TEM image of the Co@COF showing the uniform nPs of cobalt in the hexagonal COF layers. Lattice fringes indexing shows metallic
cobalt. Particle size mapping gives average sizes in the range of 3–6 nm. E) A room-temperature magnetism data showing ferromagnetic character of the
Co@COF. Inset shows the hysteresis from a 5 K measurement. F) H evolution from NaBH using Co@COF as catalyst. Inset shows the cycling stability.
2
4
Small 2018, 1801233
1801233 (3 of 8)
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
branch exhibits the presence of uniform ≈36 Å pores with
metal hydroxylated on the surface. Amorphous nPs of Co(OH)2
with unusual ferromagnetic character are reported, but they show
characteristic hysteresis only below 10 K. But here we have a
a pore volume of 1.154 cc g−1 (Figure 2B). The Brunauer–
2
−1
[35]
Emmet–Teller surface area was estimated to be 1036 m g .
A DFT model (cylindrical carbon @ 77 K N ) yielded an excel-
room-temperature ferromagnet, which suggests that majority
2
2
−1
lent fit with a surface area of 1305 m g and a pore volume of
is cobalt and not Co(OH) . However, magnetization does not
2
−
1
1
.37 cc g . This high porosity is drastically reduced (up to 85%)
saturate even at higher fields implying the presence of small
amounts of some paramagnetic or antiferromagnetic impu-
by the loading of nP (≈16 wt%) as evidenced from the 77 K N2
isotherm (Figure 2B), thus suggesting that the nPs predomi-
nantly reside in the pores of the COF. A particle size estimate
from TEM suggests the majority of the nPs being in the range
of 3–5 nm. It is important to realize that significant proportions
are above the limiting pore size of the COF (40 Å).
rity. Also, when cobalt salt is reacted with NaBH , Co–B alloy
4
[
36,37]
or boride can also form as an amorphous phase,
and such
[38]
borides can be ferromagnetic.
However, energy dispersive
X-ray analysis (EDAX)-elemental mapping clearly indicated the
lack of any boron in the Co@COF, and this was further con-
firmed from the inductively coupled plasma (ICP) analysis (see
the Supporting Information). Moreover, the XRD pattern of the
Co@COF annealed at 800 °C (Figure S31, Supporting Infor-
2
.3. Rationalization of the nP Structure and Composition
[39]
mation) matches well with the Co metal and not with Co B.
2
The formation of the cobalt nPs happens during the aqueous
phase reduction of this Co2 dispersed in COF using NaBH4
in the open air. But this can give rise to Co(OH) , Co O , CoO,
or Co. Hence, the product was analyzed thoroughly using a
range of techniques. The PXRD did not display the peaks due
to cobalt as they grow as small-sized nPs (3–6 nm). The lat-
tice fringes observed in the HR-TEM indexed to a d-spacing of
The XPS did not show any signal corresponding to boron
(B.E. ≈192 eV). To get further confirmation, we carried out a VT-
+
[36]
PXRD. Due to the small size of the nPs, we could not see any peak
due to Co metal even at 400 °C (Figures S29 and S30, Supporting
Information); however at 500 °C, a peak due to Co O (JCPDS:
2
3
4
3
4
781970) appears and this most probably is due to the air oxida-
[33]
tion of the cobalt. To obtain further clarity, we heated the Co@
0
.21 nm (Figure 2D), which corresponds to cubic cobalt metal
COF under a rich flow of N in a TGA till 800 °C and the PXRD
2
(
Co metal; inorganic crystal structure database (ICSD):α-Co
of the resulting black powder showed highly crystalline form of Co
(metal), 05–0727). However, the XPS measurements displayed
metal (face centered cubic (FCC),α-Co (metal), Figure S31, Sup-
[33,40]
the Co-2p3/2 levels with a binding energy (BE) of 781.1 eV which
corresponds to Co(OH) (Figure S25, Supporting Information).
This indicates that the surface of the Co metal is hydroxylated
under the aqueous phase. To gain further evidence, in an XPS
porting Information).
Thus, the lack of peaks in the VT-PXRD
[7]
below 400 °C confirms that the low-temperature hexagonal close
packed (HCP) phase does not grow into large particles, while the
2
results from the TGA under N suggest that majority of the nPs
2
experiment, the top layers of the Co/Co(OH) nPs in the COF
were peeled using sputtering and the resulting XPS was com-
pared with the XPS recorded on a fresh sample. The fresh
correspond to Co metal. This further supports the ability to grow
metallic nPs on the COF support.
2
sample is only Co(OH) at ≈781.2 eV (2P3/2) and after sputtering
2
you see an asymmetry toward lower binding energy with metallic
peak at 778.2 eV (Figure S25, Supporting Information). The
best distinction among Co(OH) , Co O , CoO, and Co comes
2.4. Role of COF as a Porous Support
To demonstrate the key role played by the COF in stabilizing
2
3
4
from their magnetic properties. In bulk, both the hydroxide and
the oxide are antiferromagnetic at room temperature,[ while
the Co metal is ferromagnetic even at a temperature as high
as 350 °C.[ Interestingly, Co@COF is a room-temperature
ferromagnet and the entire bulk of the sample is homogene-
ously attracted by permanent magnets (Figure 2E), confirming
that this is not a mere heterogeneous mixture of COF and
ferromagnetic nPs. In fact, this is an excellent proof for the
formation of true composite between COF and metal derived
nPs.[ Co O when grown as small nPs is known to exhibit
the uncapped catalytic Co/Co(OH) nPs, we carried out the syn-
2
32]
thesis without any COF. This yielded black-colored sample amor-
phous to X-ray (Figure S30, Supporting Information). Under
field emission scanning electron microscopy (FE-SEM) they
appear as aggregated fine flakes (Figure S32, Supporting Infor-
mation). From HRTEM, they were found to be stacked aggre-
gates of large nanosheets (Figure S34, Supporting Information).
They were mostly crystalline and the lattice fringes indexed to
neat Co metal 0.19 nm (101 facet) and 0.205 nm (002 facet)
33]
7,8]
and few to Co(OH) (0.27 nm). The selected area electron dif-
3
4
2
[34]
weak ferromagnetism at room temperature. Hence, the lat-
tice fringes from several HR-TEM images were indexed and
everyone indexed to 0.19 or 0.205 nm (α-Co (metal); Figure 2D
and Figure S26, Supporting Information), and none had the
fraction (SAED) pattern indicated polycrystallinity (Figure S34,
Supporting Information), wherein the external rings from gra-
phitic support are discernible, while no additional ring at 2.46 Å
is observed, indicating the absence of cobalt oxide. The lattice
fringes were present all across the sample, making their par-
ticle size determination impossible. However, a comparison
of the solid-state UV of this neat nP versus Co@COF showed
a blueshift for the latter (Figure S35, Supporting Information;
peak at 400 nm), which can be associated with smaller particle
characteristic d-spacing expected for Co O4 (0.46 nm), CoO
3
(
0.24 nm), or Co(OH) (0.27 nm). And this is well supported by
2
the lack of peaks corresponding to Co O or CoO in the XPS.
3
4
However, in the XPS, a feature corresponding to cobalt metal
Co-2p3/2 = 778.1 eV) was absent and only the peak due to
(
[41]
hydroxide was observed. This is in contrast with the observations
from HRTEM lattice imaging. Coupling the observations from
XPS, HRTEM, and the magnetism, we most likely have a cobalt
size. Surprisingly, they had some mesoporosity with a BJH
2
−1
pore radius of 15.3 Å and a surface area of 53 m g (Figure S36,
Supporting Information). However, the small intrinsic pore
Small 2018, 1801233
1801233 (4 of 8)
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
−1
volume (0.108 cc g ) indicates that the surface area and the pore
size come more from the interparticle spaces. A particle size
determination using diffused light scattering also confirms that
the unsupported or naked nPs were relatively smaller, which
would be expected to pack efficiently giving rise to mesopores
This immediately sets off the release of H visualized from con-
tinuous bubbling. The released hydrogen was measured using
the water-displacement method at different temperatures.
2
Quantity of H evolved as a function of time reveals the abrupt
2
change of kinetics. As expected, the total H volume reached
2
(
Figure S39, Supporting Information). Importantly, they also
the theoretical value of 98.8 mL in <6 min at room tempera-
ture. The same could be achieved in <2 min at 50 °C. This is
exhibit room-temperature ferromagnetism (Figures S40 and S41,
Supporting Information). However, a mass-balance comparison
visibly suggests that the particles of Co@COF are 6.25 times
lighter than the unsupported or naked Co/Co(OH) (Figure S43,
Supporting Information). Thus, the Co@COF represents a
lightweight room-temperature ferro-nanomagnet.
much faster compared with the Pt-LiCoO (15 min) reported
2
by Inokawa et al., which is among the best-performing cata-
[
12]
lyst. The activation energy calculated from Arrhenius rela-
2
−
1
tion to be 48 kJ mol , which is comparable to the lowest value
−
1 [42]
reported for any Co-based catalysts (less than 50 kJ mol ).
Moreover, the productivity of H remained same even after five
2
runs indicating the high durability of catalysts in NaBH hydrol-
4
2
.5. Co@COF-Catalyzed Hydrogen Evolution and
ysis. The H evolved from the NaBH was used in situ for the
2
4
One-Pot Organic Reductions
hydrogenation of nitro and cyano compounds to corresponding
primary amines. The catalysis was carried out by introducing
Hydrogen evolution from NaBH using Co@COF was moni-
Co@COF catalyst and NaBH into a methanolic solution of the
4
4
tored using a simple gas displacement cell. In a typical reac-
tion, into a flask containing 10 mg of preactivated Co@COF
substrate at room temperature for 3–12 h. High catalytic activity
was characterized by high conversion (>93–99% for the aro-
matic substrates and 82–86% for the aliphatic ones (Figure 3))
catalyst, 1 mL of an alkaline solution of NaBH was injected.
4
Figure 3. A schematic illustration of the catalytic nitrile to amine conversion. Reaction conditions: 0.0075 mmol of Co@COF catalyst (1.5 mol%), 0.5 mmol of
nitrile, 2.5 mmol of NaBH , and 3 mL of MeOH. The different substrates have been shown with their isolated % yields b (% GC conversion shown in brackets).
4
Small 2018, 1801233
1801233 (5 of 8)
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
Figure 4. Top: optimized structure of Co/Co(OH) clusters in COF. The clusters are grown from different basal planes. The relative energies are repre-
2
−
1
−
sented as “E” in units of kcal mol . Bottom: the [BH ] ions geometry optimized at the different energetic positions of the Co/Co(OH) cluster grown
from the (101) basal plane. The electrostatic potentials calculated from DMol showing the activated [BH ] ions at these different sites on the cluster.
4
2
[3]
−
4
The hydrogens of the capping hydroxyls have not been shown.
and selectivity toward primary amines. Similarly, the nitro com-
pounds were hydrogenated to corresponding amines in high
yields (Tables S3–S5, Supporting Information). The turnover
2.6. Computational Modeling of the NaBH Activation on the
Co@COF Surface
4
−
1
frequency (TOF) for these nitrile reductions were 25–30 h ,
The advantage of this nanoconfinement in exposing the nPs
which is quite reasonable for a nonnoble metal-based hetero-
facets and its potential impact on the activation of NaBH is
4
geneous catalyst.[
43]
This necessitates the catalyst to show
modeled via computational methods. For understanding the
energy associated with the growth of the Co/Co(OH)2 clus-
ters within the nanoconfinement of the COFs, we developed
“frustum” nanoclusters from the typical hexagonal close
long cycle life (Figure S44, Supporting Information). After
reaction, the Co@COF catalyst was easily recovered by filtra-
tion and even after five runs, it was showing good structural
integrity without any noticeable aggregation or growth of nPs
packed cobalt lattice (Sp.gr.: P6 /mmc; a = 2.77; c = 4.10 Å).
3
(Figure S45, Supporting Information).
The cleaved bonds at the shape boundaries of the cluster were
capped by hydroxyl oxygen atoms, giving it a Co/Co(OH)2
structure which corresponds to what is observed from TEM
and XPS. The frustum clusters were grown along five different
facets: (002), (101), (100), (010), and (111) (Figure S49, Sup-
porting Information). Note: The (002) and (101) are the facets
predominantly observed under the HRTEM. Each cluster was
dispersed in the COF lattice using a simulated annealing rou-
tine. Following this, a 4 × 4 × 4 cell of the Co@COF was geom-
etry optimized using the Forcite package. The relative energies
of these different Co@COF configurations are presented in
Figure 4. The cluster grown along the (101) facet had the lowest
relative energy.
To further evidence the role of the COF in the catalysis, we
synthesized the unsupported catalyst Co/Co(OH)2 nPs and
investigated its activity toward the nitro/nitrile reductions. The
hydrogenation of benzonitrile occurred with only about 30%
conversion, in <20% yield, even after 48 h. Detailed characteri-
zation of this control material is presented in the Supporting
Information. Importantly, the cyclic voltammogram (CV) of
the Co@COF matched well with the control (Figure S47, Sup-
porting Information). The CV suggested the catalyst being able
2
+
3+
to act as a Co /Co redox couple and a comparison of the sep-
aration of the oxidation–reduction peaks between the catalyst
and the unsupported Co/Co(OH) connotes the relatively faster
2
redox kinetics of the Co@COF.[ This suggests the catalyst
operates via a redox mechanism similar to the control but with
much higher activity, which can be attributed to cobalt catalyst
being present as highly dispersed nPs in the COF support. We
notice that the reduction reactions could be carried out by using
the Co@COF in neat methanol, even in the absence of hydroxyl
species (from aq. alkaline solution). However, the kinetics was
relatively slower. This suggests the presence of inherent redox
activity from the catalyst for the evolution of H from NaBH ,
44]
The activation of the [BH ] ions, from the NaBH , by the
−
4
4
catalyst surface directly decides the efficiency of the H evolu-
2
tion and in turn the tandem reduction reactions. To elucidate
how these nanoclusters can serve as a high activity catalyst in
−
activating the [BH ] ions, we calculated the energy associated
4
with the activation of the reactant and substrates on the sur-
−
face of these clusters. For this, we bonded the [BH ] ions on
4
the “plateau,” “step,” and the “edge” of the cluster grown along
the [101] facet and minimized its geometry and energy using
2
4
[
3]
but the highest activity is realized when the catalyst works in a
basic medium. This role of the catalyst is schematically repre-
sented in Figure S48 in the Supporting Information.
DMol, which yielded some fascinating results. The borohy-
dride gets activated and the Co…HB bond attached to the
cobalt surface gets elongated, this is manifested by the marked
Small 2018, 1801233
1801233 (6 of 8)
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
Figure 5. DFT optimized Co@COFs with increasing sizes of Co/Co(OH) clusters. A) Small cluster fitting the pore. B) Large cluster with a cone shape.
2
The narrow apex is buried in the pores, while the wider base is projecting out of the COF pore. The cluster is stabilized within the pores by the narrow
apex. C,D) COF with both small and large clusters (diameter > the limiting pore size of the COF).
change in BH bond lengths and the change in electron distri-
bution which is visualized from the potential maps (Figure 4).
From figure S50 in the Supporting Information, it can be seen
that this energy associated with the activation is slightly dif-
ferent for each facet. We believe that this is where the role of
COF in exposing the specific facets becomes crucial. Impor-
tantly, all these clusters seem to stabilize to a small size in the
nanoconfinement of the COF; however, the cluster exposing the
In fact, this is an ubiquitous observation in supported cata-
[
4,5a,6b,8,45,46]
lyst systems.
Thus, providing a plausible explanation
for this would be of immense value. We realize that in these
porous supports, the clusters could be growing from within the
pores and the growth can extend beyond the pore dimension
(Figure 5 and Figure S51, Supporting Information). And in
such scenario, the particles found under the TEM can be larger
than the COF’s limiting pore size. Since they are anchored to
the support by the part of the cluster residing within the pores,
they do not leach out. A DFT geometry optimization using a
2 × 2 × 5 cell reveals that such clusters can be stabilized within
the COF matrix, but the larger clusters tend to have higher
relative energy. Importantly, excellent convergence with accept-
able geometries and energies were obtained for all three config-
urations, suggesting they all can be expected to form under the
experimental conditions (Figure 5). However, when a cluster
of very large size (Dia. 6 nm; Height- 5 layers) was attempted,
a stable configuration for the Co@COF could not be obtained
using the DFT minimizations and the energies were unaccept-
ably large.
(101) facet has the lowest relative energy, which corroborates
well with it being the most frequently observed facet under the
HRTEM (Figure 2E and Figure S26, Supporting Information).
It is important to underscore the fact that the crystalline struc-
ture of the COF enables these atomic-level computations.
2
.7. Elucidating the Ability of COF to Stabilize Clusters Larger
than its Limiting Pore Size—An Ubiquitous Observation
As mentioned earlier, significant proportion of the nPs has
sizes larger than the limiting pore size of the COF. Rationally,
such large particles would be expected to leach from the surface
of the support, particularly under the harsh conditions of the
catalysis. For assessing the potential leaching of the nP from
the COF surface, a postcatalysis supernatant solution was ana-
lyzed under EDAX and no trace of cobalt was observed. Also,
neither the hydrogen evolution nor the nitrile reduction was
observed when this supernatant was employed in place of the
catalyst indicating the lack of np leaching and also suggesting
strong binding of the nP with the COF. How do such particles
exceeding the limiting pore size of the COF manage to adhere
so strongly?
3
. Conclusions
In summary, we have developed a nonnoble metal heterogeneous
catalyst comprising of surface hydroxylated cobalt nP supported
on a methoxy functionalized highly stable COF. It catalyzes the
hydrogenation of nitro/nitrile compounds in the presence of
readily available and safe NaBH under ambient conditions with
4
good yield, selectivity, and recyclability. The electron-donating
methoxy groups positioned strategically with respect to the Schiff
Small 2018, 1801233
1801233 (7 of 8)
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
bonds provide exceptional stability to the COF, while the N-rich
sites favor strong interaction with the cobalt nPs and prevent
[11] T. Banerjee, F. Haase, G. Savasci, K. Gottschling, C. Ochsenfeld,
B. V. Lotsch. J. Am. Chem. Soc. 2017, 139, 16228.
[
12] H. Inokawa, H. Driss, F. Trovela, H. Miyaoka, T. Ichikawa, Y. Kojima,
their leaching. The geometry-optimized Co@COF structures
−
S. F. Zaman, A. Al-Zahrani, Y. Alhamed, L. Petrov, Int. J. Energy Res.
and the electrostatic potentials for the [BH ] ions activated on
4
2
016, 40, 2078.
the nP surface display how the COF modulates the catalytic
activity by exposing specific facets of the cluster. While the
cyclic voltammetry experiments reveal the redox activity of the
Co@COF. Further studies are in progress to investigate the role
of magnetism in these heterogeneous ferromagnetic catalysts.
[
13] J. Mao, D. Gregory, Energies 2015, 8, 430.
[
14] H. Göksu, S. F. Ho, Ö. Metin, K. Korkmaz, A. Mendoza Garcia,
M. S. Gültekin, S. Sun, ACS Catal. 2014, 4, 1777.
[
15] Q. L. Zhu, J. Li, Q. Xu, J. Am. Chem. Soc. 2013, 135, 10210.
[16] M. Yadav, Q. Xu, Energy Environ. Sci. 2012, 5, 9698.
17] P. Brack, S. E. Dann, K. G. U. Wijayantha, Energy Sci. Eng. 2015, 3, 174.
[
[
18] J. Mahmood, S.-M. Jung, S.-J. Kim, J. Park, J.-W. Yoo, J.-B. Baek,
Chem. Mater. 2015, 27, 4860.
Supporting Information
[
[
19] C. J. Jia, F. Schuth, Phys. Chem. Chem. Phys. 2011, 13, 2457.
20] S. Enthaler, K. Junge, D. Addis, G. Erre, M. Beller, ChemSusChem
Supporting Information is available from the Wiley Online Library or
from the author.
2
008, 1, 1006.
[
21] P. Ji, K. Manna, Z. Lin, X. Feng, A. Urban, Y. Song, W. Lin,
J. Am. Chem. Soc. 2017, 139, 7004.
Acknowledgements
[22] F. A. Westerhaus, R. V. Jagadeesh, G. Wienhofer, M. M. Pohl,
J. Radnik, A. E. Surkus, J. Rabeah, K. Junge, H. Junge, M. Nielsen,
A. Bruckner, M. Beller, Nat. Chem. 2013, 5, 537.
23] J. Long, K. Shen, Y. Li, ACS Catal. 2017, 7, 275.
24] T. Zhang, K. Manna, W. Lin, J. Am. Chem. Soc. 2016, 138,
The authors thank IISER-Pune and the MHRD-FAST program for the
necessary funding. The authors thank “Funding by DST-Nanomission
under the Thematic Unit Program” (EMR/2016/003553) for the support.
The authors thank Dr. Santosh K. Singh (CSIR-NCL, Pune) for the cyclic
voltammetry measurements.
[
[
3
241.
[
25] H. Yang, S. J. Bradley, A. Chan, G. I. Waterhouse, T. Nann,
P. E. Kruger, S. G. Telfer, J. Am. Chem. Soc. 2016, 138, 11872.
26] M. Li, G. Chen, Nanoscale 2013, 5, 11919.
27] J. H. Lee, S. K. Hong, W. B. Ko, Asian J. Chem. 2011, 23, 2347.
28] J. Feng, S. Handa, F. Gallou, B. H. Lipshutz, Angew. Chem., Int. Ed.
[
[
[
Conflict of Interest
The authors declare no conflict of interest.
2
016, 55, 8979.
[
[
29] F. Chen, C. Topf, J. Radnik, C. Kreyenschulte, H. Lund, M. Schneider,
A. E. Surkus, L. He, K. Junge, M. Beller, J. Am. Chem. Soc. 2016,
138, 8781.
30] Y. Zhi, Z. Li, X. Feng, H. Xia, Y. Zhang, Z. Shi, Y. Mua, X. Liu,
J. Mater. Chem. A 2017, 5, 22933.
31] H. Xu, J. Gao, D. Jiang, Nat. Chem. 2015, 7, 905.
32] T. Mousavand, T. Naka, K. Sato, S. Ohara, M. Umetsu, S. Takami,
T. Nakane, A. Matsushita, T. Adschiri, Phys. Rev. B 2009, 79, 144411.
33] V. c. A. de la Peña O′Shea, P. R. r. de la Piscina, N. Homs, G. Aromí,
J. L. G. Fierro, Chem. Mater. 2009, 21, 5637.
Keywords
cobalt nanoparticles, covalent organic framework, heterogeneous
catalysis, hydrogen evolution, nitrile reduction
[
[
Received: March 30, 2018
Revised: May 18, 2018
Published online:
[
[
34] S. Farhadi, K. Pourzare, S. Sadeghinejad, J. Nanostruct. Chem. 2013,
3, 16.
[
[
[
[
[
1] P. J. Waller, F. Gándara, O. M. Yaghi, Acc. Chem. Res. 2015, 48,
053.
2] J. L. Segura, M. J. Mancheno, F. Zamora, Chem. Soc. Rev. 2016, 45,
635.
3] S. Y. Ding, J. Gao, Q. Wang, Y. Zhang, W. G. Song, C. Y. Su,
W. Wang, J. Am. Chem. Soc. 2011, 133, 19816.
3
[35] H. B. Li, P. Liu, Y. Liang, J. Xiao, G. W. Yang, RSC Adv. 2013, 3,
26412.
5
[36] J. Masa, P. Weide, D. Peeters, I. Sinev, W. Xia, Z. Sun, C. Somsen,
M. Muhler, W. Schuhmann, Adv. Energy Mater. 2016, 6, 1502313.
[37] S. W. Heinzman, B. Ganem, J. Am. Chem. Soc. 1982, 104, 6801.
[38] (Ed.: K. H. Buschow, V. I. Matkovich), Boron and Refractory Borides,
Springer-verlag, Berlin, Heidelberg 1977.
[39] P.-J. Cai, H. Wang, L.-h. Liu, L. Zhang, J. Ceram. Soc. Jpn. 2010, 118,
1102.
[40] Q. Meng, S. Guo, X. Zhao, S. Veintemillas-Verdaguer, J. Alloys
Compd. 2013, 580, 187.
[41] E. Gharibshahi, E. Saion, Int. J. Mol. Sci. 2012, 13, 14723.
[42] H. Zhang, T. Ling, X.-W. Du, Chem. Mater. 2015, 27, 352.
[43] X. Sun, A. I. Olivos-Suarez, L. Oar-Arteta, E. Rozhko, D. Osadchii,
A. Bavykina, F. Kapteijn, J. Gascon, ChemCatChem 2017, 9, 1854.
[44] M. C. D. Cooray, X. Zhang, Y. Zhang, S. J. Langford, A. M. Bond,
J. Zhang, J. Mater. Chem. A 2017, 5, 19289.
4] D. Mullangi, S. Nandi, S. Shalini, S. Sreedhala, C. P. Vinod,
R. Vaidhyanathan, Sci. Rep. 2015, 5, 10876.
5] a) P. Pachfule, M. K. Panda, S. Kandambeth, S. M. Shivaprasad,
D. D. Díaz, R. Banerjee, J. Mater. Chem. A 2014, 2, 7944; b) Z.-L. Du,
Q.-Q. Dang, X.-M. Zhang, Ind. Eng. Chem. Res. 2017, 56, 4275.
6] a) X. Shi, Y. Yao, Y. Xu, K. Liu, G. Zhu, L. Chi, G. Lu, ACS
Appl. Mater. Interfaces 2017, 9, 7481; b) P. Pachfule, S. Kandambeth,
D. D. Dıaz, R. Banerjee, Chem. Commun. 2014, 50, 3169.
[
[
7] D. Mullangi, V. Dhavale, S. Shalini, S. Nandi, S. Collins, T. Woo,
S. Kurungot, R. Vaidhyanathan, Adv. Energy Mater. 2016, 6, 1600110.
8] S. Nandi, S. K. Singh, D. Mullangi, R. Illathvalappil, L. George,
C. P. Vinod, S. Kurungot, R. Vaidhyanathan, Adv. Energy Mater.
[
2
016, 6, 1601189.
9] A. Ansari, P. Jayapal, G. Rajaraman, Angew. Chem., Int. Ed. 2015, 54,
64.
10] P. J. Chirik, Acc. Chem. Res. 2015, 48, 1687.
[45] W. Ma, P. Yu, T. Ohsaka, L. Mao, Electrochem. Commun. 2015, 52,
53.
[46] D. Kaleeswaran, R. Antony, A. Sharma, A. Malani, R. Murugavel,
ChemPlusChem 2017, 82, 1253.
[
5
[
Small 2018, 1801233
1801233 (8 of 8)
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim