Hierarchical Porous Organometallic Polymers Fabricated by Direct Knitting: Recyclable Single-Site Catalysts with Enhanced Activity

Article abstract of DOI:10.1002/adma.201905950

Porous organometallic polymers (POMPs) with hierarchical pore structures, high specific surface areas, and atomically dispersed metal (Ir, Pd, Ru) centers are successfully fabricated by a facile one-pot method through direct knitting of diverse N-heterocyclic carbene metal (NHC-M) complexes. These polymers can function as recyclable solid single-site catalysts and exhibit excellent catalytic activity and selectivity in both dehydrogenation and hydrogenation reactions even at ppm-level catalyst loadings. Remarkably, a record turnover number (TON) of 1.01 × 10⁶ is achieved in the hydrogenation of levulinic acid to γ-valerolactone, which is 750 times higher than that attained with corresponding bis-NHC-Ir complex.

Full text of DOI:10.1002/adma.201905950

CommuniCation
www.advmat.de
Hierarchical Porous Organometallic Polymers Fabricated
by Direct Knitting: Recyclable Single-Site Catalysts
with Enhanced Activity
Yajing Shen, Qingshu Zheng, Haibo Zhu, and Tao Tu*
Dedicated to Prof. Li-Xin Dai on the occasion of his 95th birthday
supported catalysts by simply knitting rigid
Porous organometallic polymers (POMPs) with hierarchical pore structures,
high specific surface areas, and atomically dispersed metal (Ir, Pd, Ru) centers
are successfully fabricated by a facile one-pot method through direct knitting
of diverse N-heterocyclic carbene metal (NHC-M) complexes. These polymers
can function as recyclable solid single-site catalysts and exhibit excellent
catalytic activity and selectivity in both dehydrogenation and hydrogenation
reactions even at ppm-level catalyst loadings. Remarkably, a record turnover
number (TON) of 1.01 × 10 is achieved in the hydrogenation of levulinic acid
to γ-valerolactone, which is 750 times higher than that attained with corre-
sponding bis-NHC-Ir complex.
aromatic ligands via hypercross-linking
(Friedel-Crafts reactions) has recently been
[
13]
proposed.
tion/activation, several MCPPs have been
After further postmodifica-
[14–17]
successfully developed.
However,
with postmodification, avoiding the prob-
lems, like low density of catalytic sites due
to incomplete modification and random/
nonselective metal anchoring, can be dif-
ficult. These issues may still lead to low
catalytic activity and selectivity, especially
6
[18]
in some challenging transformations.
In light of N-heterocyclic carbene-metal
(NHC-M) complexes can act as robust and viable catalysts in
A crucial bottleneck hampering the application of immobi-
lized molecular catalysts in industry is their reduced cata-
lytic activity, probably due to altered chemical environment
[
19–28]
a broad range of reactions,
developed for their immobilization.
supporting strategy may represent one of the most practical and
various strategies have been
[
29,30]
Among them, self-
[1–3]
and decreased accessibility of the substrates to active sites.
[31–34]
Porous materials have been widely recognized as a promising
scaffold for catalyst immobilization with high efficiency and
efficient approaches.
It has to be noted that NHC ligands
still have to be carefully designed and modified before self-
supporting. Alternatively, the high catalytic activities and robust-
ness of NHC-M complexes make them perfect candidates for
direct knitting. To the best of our knowledge, there are only two
reported examples on MCPPs in which the NHC ligands are
reusability owing to their intrinsic porosity and high surface
areas.[
4–6]
They can function as either heterogeneous catalysts
themselves with built-in active sites or hosts with catalytic spe-
[7]
cies inside their pores/channels. For most metal-containing
porous polymers (MCPPs),[
8,9]
ligands/complexes have to be
firstly knitted and then coordinated to active metal ions (post-
carefully designed and modified,[ which not only increase
10]
modification approach).
[14,15]
Direct knitting of NHC-M com-
the cost and difficulty of the immobilization but may also com-
plexes to fabricate porous organometallic polymers (POMPs)
that do not require postmodification/activation is still unknown.
Herein, we have designed and prepared a series of unprec-
edented porous organometallic polymers POMPs-NHC-M 1–3
via direct knitting three types of bis-chelating NHC-M com-
plexes (M = Ir, Pd, and Ru, Figure 1a). Using this facile strategy,
which does not require postmodification/activation, 1) the
metal active sites are atomically dispersed in the hierarchical
porous matrix, and 2) the intrinsic well-defined bis-chelating
mode and their molecular catalytic properties are maximally
retained, which allows 3) the precise design and control of
the immobilized catalysts at the molecular level. The advan-
tages of direct knitting allow the resulting POMPs-NHC-M 1–3
to behave as recyclable solid single-site catalysts and exhibit
extremely high catalytic activities in both hydrogenation and
dehydrogenation reactions.
[11,12]
promise the intrinsic activity of the molecular catalysts.
To
overcome these disadvantages, a facile approach for accessing
Y. Shen, Dr. Q. Zheng, Dr. H. Zhu, Prof. T. Tu
Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials
Department of Chemistry
Fudan University
2
005 Songhu Road, Shanghai 200438, China
E-mail: taotu@fudan.edu.cn
Prof. T. Tu
State Key Laboratory of Organometallic Chemistry
Shanghai Institute of Organic Chemistry
Chinese Academy of Sciences
Shanghai 200032, China
Prof. T. Tu
College of Chemistry and Molecular Engineering
Zhengzhou University
In light of their robustness toward Friedel-Crafts reaction
conditions, bis-chelating NHC-M (Ir, Pd and Ru) complexes
1
00 Kexue Avenue, Zhengzhou 450001, China
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/adma.201905950.
4
, 5 and 6, derived from corresponding benzimidazolium
[
24,25,27]
salts,
were selected to fabricate POMPs via direct knit-
DOI: 10.1002/adma.201905950
ting with benzene (Figure 1a). Concerning the plausible impact
Adv. Mater. 2019, 1905950
1905950 (1 of 9)
© 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
Figure 1. a) Fabrication of hierarchical porous organometallic polymers (POMPs-NHC-M 1–3, M: Ir, Pd, or Ru) via direct knitting of bis-NHC-M
complexes 4, 5, or 6 by using Friedel-Crafts reaction in the presence of FeCl , formaldehyde dimethyl acetal in dichloroethane with or without benzene
3
[14]
addition. b) Fabrication of POMP-NHC-Ir 8 via postmodification of poly-NHC 7 with Ir(acac)(CO)2.
of the ratio between the crosslinking reagent and the Lewis acid
on the porous structure, surface area and catalytic activity,
entry 9), which may benefit for its further characterization by
using F NMR technique. In order to verify the effect of knit-
[
17]
19
with bis-NHC-Ir complex 4a, the amounts of FeCl , formalde-
hyde dimethyl acetal (FDA) and benzene in the knitting process
ting linkers, we also fabricate POMP-NHC-Ir 1j by using bis-
NHC-Ir complex 4a with 1,3,5,7-tetraphenyladamantane (tAd,
3
were investigated for the generation of POMP-NHC-Ir 1a-h
instead of benzene). After optimization, 20 equiv. FeCl and
3
(
Table 1, entries 1–8). In addition, fluorinated bis-NHC-Ir com-
20 equiv. FDA were selected to prepare POMP-NHC-Pd 2 and
POMP-NHC-Ru 3 from corresponding molecular complexes
5 and 6, respectively (Table 1, entries 10–11). For comparison,
plex 4b was also involved in the direct knitting to generate corre-
sponding POMP-NHC-Ir 1i without benzene addition (Table 1,
Table 1. Fabrication of POMP-NHC-M by direct knitting with different ratios of FDA and FeCl , and of POMP-NHC-Ir 8 by postmodification.
3
Entrya)
POMPs
Benzene [equiv.]
FeCl3 [equiv.]
FDA [equiv.]
M
Ir
Metal content [%]b)
2
−1 c)
SBET [m g
]
1
2
3
4
5
6
7
8
9
POMP-NHC-Ir 1a
POMP-NHC-Ir 1b
POMP-NHC-Ir 1c
POMP-NHC-Ir 1d
POMP-NHC-Ir 1e
POMP-NHC-Ir 1f
POMP-NHC-Ir 1g
POMP-NHC-Ir 1h
POMP-NHC-Ir 1i
POMP-NHC-Pd 2
POMP-NHC-Ru 3
POMP-NHC-Ir 8
3
3
9
9
5.69
6.05
6.34
6.46
12.9
14.5
14.9
17.3
13.3
6.97
3.93
5.37
226
357
440
450
33
12
15
20
20
20
20
20
20
20
20
20
12
15
20
20
20
20
20
20
20
20
20
Ir
3
Ir
3
Ir
1
Ir
0.5
0.2
0
Ir
26
Ir
24
Ir
42
0
Ir
52
10
11
12
3
Pd
Ru
Ir
381
906
417
3
3
a)
Reaction conditions: a mixture containing 0.1 mmol of bis-chelating NHC-M complex 4a, 4b, 5, or 6, benzene, FeCl3, and FDA in 1 mL of dry dichloroethane were heated
b)
c)
in a sealed Schlenk tube at 80 °C for 24 h under a N2 atmosphere (M for Ir, Pd or Ru); The percentage of metal atoms were determined by ICP-AES; BET specific surface
area calculated from the N2 adsorption isotherms at 77 K.
Adv. Mater. 2019, 1905950
1905950 (2 of 9)
© 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
Figure 2. TEM and HAADF images (with EDX mapping as insets) of a,d) POMP-NHC-Ir 1d, b,e) POMP-NHC-Pd 2, and c,f)POMP-NHC-Ru 3. Nitrogen
adsorption and desorption isotherms of POMPs g) 1d, h) 2, and i) 3, measured at 77 K with the corresponding pore width distribution of POMPs
1
d, 2, and 3 determined by quenched solid density functional theory as insets.
POMP-NHC-Ir 8 was accessed via a postmodification approach
The accurate metal contents of POMPs 1–3 and 8 were
determined by inductively coupled plasma atomic emission
spectroscopy (ICP-AES, Table 1), which clearly indicated metal
from known Poly-NHC 7[ and Ir(acac)(CO) (acac = acetyl
14]
2
acetonate, Figure 1b) as a control under the similar reaction
conditions (Table 1, entry 12).
contents were strongly affected by the amounts of FeCl , FDA,
3
With these POMPs in hand, the morphologies were
then investigated. Transmission electron microscopy (TEM)
studies revealed that POMPs-NHC-Ir 1 and 8 adopted sim-
ilar morphologies with porous structures (Figure 2a, and
Figures S8–S11 and S17, Supporting Information). Upon
increasing the amounts of FeCl3 and FDA, the number of
macropores increased, especially within the matrix of POMP-
NHC-Ir 1d. Similar results were also observed by scanning
electron microscopy (SEM, Figures S1–S7, Supporting Infor-
mation). Other POMPs based on bis-NHC-Pd 5 and NHC-Ru
and benzene as well as corresponding metal complexes. Sub-
sequently, high-angle annular dark-field scanning transmis-
sion electron microscopy (HAADF-STEM) and energy disper-
sive X-ray (EDX) spectroscopy mapping indicated that metal
centers were all uniformly dispersed in the matrices of all
POMPs (Figure 2d–f), and no obvious iridium nanoparticles
were observed even in the postmodification case (Figure S17,
Supporting Information). Notably, no Fe residue was found
in all solid POMPs 1–3 after Soxhlet extraction and repeated
rinse with water and methanol, further confirmed by EDX and
X-ray photoelectron spectroscopy (XPS) analysis (Figures S24
and S38, Supporting Information). To our surprise, Cl ele-
ment was found in the EDX and XPS spectra with low inten-
sity, which may be attributed to the partial ligand exchange of
weak coordinated CO ligands with the Cl elements presented
6
exhibited similar lamellar morphologies with obvious
macropores (Figures 2b–c). In contrast, dense lamellar mor-
phology without obvious macropores was observed with POMP-
NHC-Ir 1j accessed by using tAd instead of benzene (Figure
S18, Supporting Information). The powder X-ray diffraction
(
PXRD) patterns further confirmed that all obtained POMPs 1–3
in the chlorinated solvent DCE and FeCl during the knitting
process.
3
[
35]
and 8 were amorphous (Figure S36, Supporting Information).
Adv. Mater. 2019, 1905950
1905950 (3 of 9)
© 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
The porosities of the POMPs 1–3 and 8 were then investi-
gated by nitrogen sorption measurements. The adsorption
and desorption isotherms of POMP-NHC-Ir 1d show steep
increases in the N uptake in the P/P < 0.05 region, which is
Information). Extremely, in the case of POMP NHC-Ir 1j
formed with tAd, mainly micropores along with a little content
of mesopores were observed (Figure S35, Supporting Informa-
tion). Surprisingly, the BET specific surface areas of Poly-NHC
7, fabricated under the standard knitting conditions with 1d,
2
0
a typical behavior of microporous materials (Figure 2g). The
2
−1
capillary condensation hysteresis between P/P 0.45 and 1.00
was very high (1068 m g , Figure S34, Supporting Informa-
tion). However, after postmodification with iridium species,
the specific surface area of POMP-NHC-Ir 8 was drastically
0
indicates also the presence of mesopores. Further analysis of
the pore size distribution using quenched solid density func-
tional theory (QSDFT)[ reveals a dominant pore width at
36]
2
−1
decreased to 417 m g , which might be caused by the occupa-
tion of iridium species to the existed pores within the POMP
matrices.
0
3
.9–1.5 nm, a small peak at 1.6–2.1 nm and a broad peak at
.0–8.0 nm (Figure 2g). Similar porosities were observed with
POMPs 1a–c (Table 1, entries 1–3, Figure S25a,b, Supporting
Information). Therefore, combining with the macropores
observed in TEM images (Figure 2a), POMP-NHC-Ir 1d actually
possesses hierarchical pores, including micropores, mesopores,
and macropores with the highest Brunauer–Emmett–Teller
The successful immobilization of bis-chelating NHC-M
complexes within the polymer matrices was further confirmed
by Fourier transform infrared (FT-IR) spectroscopy, solid-state
13
C NMR spectroscopy and XPS (Figure 3a,b and Figure S45,
Supporting Information). In the case of POMP-NHC-Ir 1d,
2
−1
−1
(BET) specific surface areas (450 m g , Table 1, entry 4) in
the presence of the absorption band at 2088 cm in its FT-IR
all fabricated POMPS, suggesting the higher amount of FDA
spectra, corresponding to the CO ligand attached to the iridium
center, indicated the successful knitting of bis-NHC-Ir com-
[25]
and FeCl , the larger the BET and the portion of micropores.
3
Other POMPs 2 and 3 also showed hierarchical porous struc-
tures (Table 1, entries 10–11, Figure 2h,i), further indicated
the generality of our direct knitting strategy. Decreasing the
plex 4a (Figure S45, Supporting Information). In the solid-state
C NMR spectra (Figure 3a, blue line, Figure S50, Supporting
13
Information), peaks for CAr (135 ppm, aromatic rings) and CMe
(35 ppm, N-methyl groups and bridging methylene) in the
POMPs correspond well with those observed in bis-NHC-Ir 4a.
However, the broad signals for CNHC and CCO (≈190 ppm) are
uncertainly assigned, even with more than 100 mg samples, due
to the extremely low NHC content (0.8 % based on Ir) in the
equivalents of benzene had great impact on the N sorption
2
properties, not only much lower BET specific surface areas
were obtained, but also the pore structures were different, the
percentage of mesopores and macropores were significantly
enhanced (Table 1, entries 5–8, Figures S27–S30, Supporting
1
3
Figure 3. a) Solid-state C NMR spectra of POMP-NHC-Ir 1h (red line), 1d (blue line), and bis-NHC-Ir 4a (black line). Note: asterisks denote spinning
[
14,15,44]
19
sidebands.
b) Solid-state F NMR spectra of POMP-NHC-Ir 1i (red line) and bis-NHC-Ir 4b (black line). c) X-ray photoelectron spectroscopy
(XPS) spectra of Ir(I) in molecular complex 4a, freshly prepared solid 1d, and recovered 1d from the reaction mixture after 15 runs of the hydrogenation
of LA to GVL. d) Normalized XANES spectra at the L edge of solid 1d, the standard Ir(0) powder, and NHC-Ir(I) complex 4a were added for comparison.
3
e) Magnitude of Fourier transformed Ir L-edge EXAFS spectra of solid 1d and references. All the spectra were determined via least-squares EXAFS curve
fitting analysis. f) Magnified HAADF-STEM images of solid 1d, the well dispersed white dots represented iridium atoms.
Adv. Mater. 2019, 1905950
1905950 (4 of 9)
© 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
POMP matrix of solid 4a.[37] To solve this ambiguous problem,
POMP-NHC-Ir 1h fabricated via direct knitting with molecular
precursor 4a without benzene addition was also investigated
by solid-state C NMR spectroscopy. Fortunately, signals for
CNHC and CCO (193 and 183 ppm) were distinguishable with
To obtain the quantitative structural parameters involving the
single iridium atoms, least-squares EXAFS curve fitting anal-
ysis was performed. The Fourier transform of the k -weighted
EXAFS fitted curves were consistent with the experimental
results (Figure S54, Supporting Information). The absence of
IrIr bonds in POMP-NHC-Ir 1d (Figure 3e, Table S1, Sup-
porting Information) implied that the Ir(I) centers are atomically
well dispersed in the polymer matrix. Considering that chlorine
was found in XPS and EDS studies (Figures S24 and S38, Sup-
porting Information) due to partial exchange of CO ligands
2
13
200 mg solid samples loading, although the intensity was still
low (Figure 3a). The similar chemical shifts of CNHC observed
with solids 1h and 4a (193 vs 188 ppm) indicate the complex
structure should be well preserved after direct knitting. To make
this conclusion more convincing, fluorinated POMP-NHC-Ir 1i,
obtained from fluoro-containing bis-NHC-Ir complex 4b after
knitting, was analyzed by both solid-state 13C NMR and F NMR.
with Cl elements of chlorinated solvent and FeCl during the
3
19
[37]
knitting process, IrCl bond was implemented for EXAFS
19
Broad but almost identical signals were observed in the F NMR
data fitting. The obtained good R-factor of 0.012 suggested the
existence of IrCl bond with small quantity within the polymer
matrix. Furthermore, the IrC interaction at 2.03 Å with a coor-
dination number of 4.1 observed with solid 1d is smaller than
that at 2.08 Å with a coordination number of 5.6 found with
spectra with solid samples of POMP-NHC-Ir 1i and complex 4b
(Figures 3b and Figure S51, Supporting Information), further
emphasizing the structure preservation of the robust bis-NHC-
Ir complex during direct knitting. Besides, the thermostability
of solid 1d was studied by thermogravimetric analysis (TGA).
The decomposition temperature is relatively high (300 °C,
Figure S37, Supporting Information), confirming its robustness.
Furthermore, to our delight, almost identical binding ener-
gies of 65.0 eV (Ir 4f5/2) and 62.0 eV (Ir 4f7/2) for POMP-
NHC-Ir 1d and bis-NHC-Ir 4a were found in the XPS spectra
[
38]
complex bis-NHC-Ir 4a (Figure 3e). Despite the slight change
of coordination environment around Ir center, the bis-chelating
backbone of complex 4a were well preserved. Remarkably, mag-
nified HAADF-STEM images showed well dispersed white dots,
indicating that individual Ir atoms were atomically dispersed in
the matrix of solid 1d. Similarly, the atomically dispersed Pd
and Ru dots can also be found in the corresponding HAADF-
STEM images (Figures S14 and S16, Supporting Information).
These outcomes clearly indicated that POMPs with single-site
metal atoms were readily achieved via direct knitting.
(Figure 3c), indicating that the oxidation state of the Ir(I) in
POMP-NHC-Ir 1d was well retained after direct knitting.
Similar results were obtained with the POMP-NHC-Pd 2 and
POMP-NHC-Ru 3. The binding energies of Pd(II) in POMP-
NHC-Pd 2 were 342.9 eV (3d3/2) and 337.7 eV (3d5/2), which
were quite close to the 342.8 eV (3d3/2) and 337.6 eV (3d5/2) in
bis-NHC-Pd complex 5, indicating the Pd maintained the oxi-
dation state of +2 after direct knitting (Figure S42, Supporting
Information). In the case of POMP-NHC-Ru 3, only 3d5/2 signal
was selected to study the coordination environment of Ru in
the POMP matrix owing to the spectra interference with the 1s
signal of C. As shown in Figure S41 (Supporting Information),
identical binding energy was observed with POMP-NHC-Ru 3
and NHC-Ru complex 6, further confirmed the oxidation state
of Ru center in the POMP matrix was +2. In contrast, signals
at 65.6 eV (4f5/2) and 62.6 eV (4f7/2) were observed in the post-
modification case, which suggested the coexistence of mono-
NHC-Ir and bis-NHC-Ir species in the matrix of POMP-NHC-Ir
In the case of postmodification, the XANES curve of POMP-
NHC-Ir 8 displayed a higher intensity of the Ir-L white line
3
than that observed with solid 1d and complex 4a (Figure S55a,
Supporting Information), which was consistent with the XPS
results. The EXAFS spectra also revealed the absence of IrIr
bonds in the solid 8, and the IrC interaction at 2.08 Å with
a coordination number of 10.0, which was greater than that
observed with molecular complex 4a (2.08 Å, coordination
number of 5.6, Figure S55b, Supporting Information), which
might confirm the coexistence of mono-NHC-Ir and bis-NHC-
Ir species in the matrix of POMP-NHC-Ir 8.
Having these POMPs with hierarchical porosity and atomi-
cally dispersed active sites in hand, their activities function
[
39–43]
as single-site catalysts
were then investigated. As an
8
(Figures S39 and S40, Supporting Information). All these out-
important platform molecule in laboratory and industrial appli-
comes clearly indicated the well-defined molecular structures of
cations, γ-valerolactone (GVL) has attracted considerable atten-
[
31,45–50]
NHC-M (+1, +2) complexes were well preserved in the POMPs
tion.
Therefore, the hydrogenation of biomass levulinic
1
–3 matrices fabricated by direct knitting strategy.
acid (LA) to GVL was selected as the first model reaction. After
careful catalyst screening and optimization of the reaction
conditions (Table S2, Supporting Information), a quantitative
yield was finally achieved in the presence of 130 ppm POMP-
NHC-Ir 1d under 30 atm hydrogen at 100 °C within 4 h. Other
POMPs 1a-c all resulted in lower yields (49–88%, Figure 4a),
which might be ascribed to the lesser content of active Ir(I)
and the smaller BET surfaces areas (Table 1, entries 1–3 and
Subsequently, the chemical oxidation state and coordi-
nation environment of iridium in the matrix of POMP-
NHC-Ir 1d were further investigated by X-ray absorption
near-edge structure (XANES) and extended X-ray absorp-
tion fine structure (EXAFS) analyses. The normalized Ir-L3
edge in the XANES spectra of solid 1d, standard iridium
powder and structurally defined bis-NHC-Ir 4a are shown in
Figure 3d. The XANES curve of solid 1d is quite similar to
12 vs 4). Therefore, the amounts of FeCl and FDA used for
3
that of molecular complex 4a. The intensity of the Ir-L white
the direct knitting not only affect the morphologies, but also
exhibit obvious impacts on their catalytic activities: higher
3
line increased along the sequence Ir powder < bis-NHC-Ir
4
a ≈ solid 1d. Based on the results of Ir(0) in iridium powder,
amount of FeCl and FDA, higher yields. As a comparison, the
3
Ir(I) in complex 4a, the stable valence state of the Ir centers
in the 1d matrix should be ≈ +1, which is consistent with the
XPS results (Figure 3c).
POMP-NHC-Ir 8 prepared via postmodification using the same
amounts of FeCl and FDA for solid 1d fabrication, only gave
3
a yield of 68%. The existence of less-active mono-coordination
Adv. Mater. 2019, 1905950
1905950 (5 of 9)
© 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
Figure 4. a) Catalyst optimization in the hydrogenation of LA to GVL. Reaction condition: 15 mmol Levulinic acid, 1.1 equiv. KOH and 5 mL iPrOH were
added into the autoclave, then 130 ppm catalysts were added and the reaction mixture was reacted under 30 bar H pressure at 100 °C for 4 h. Yield
2
1
was determined by H NMR. b) Turnover numbers (TONs) obtained by POMPs and other selected viable catalysts. Reusability of c) POMP-NHC-Ir
1
d and d) POMP-NHC-Ir 8 (carried out on a 15 mmol scale with 270 ppm catalyst under the optimized reaction conditions) for 4 h (light blue) or for
2
4 h (dark blue).
NHC-Ir motif as well as the entrapped inactive iridium species
in the matrix pores of POMP-NHC-Ir 8 might be responsible
for this disappointed outcome.
Compared with the homogenous analogues, including
complexes bis-NHC-Ir 4a, bis-NHC-Pd 5, and NHC-Ru 6,
POMP-NHC-Ir 1d, POMP-NHC-Pd 2, and POMP-NHC-Ru
13 (1.2 × 105),[31] and at least three orders of magnitude higher
than those achieved by previously reported Ir, Ru, Zr, and Cu
3
inorganic nanoparticles (up to 8.2 × 10 , Table S4, Supporting
Information).
After the hydrogenation reaction, solid 1d was readily
recovered by simple centrifugation, filtration, and wash with
additional MeOH. To our delight, the recovered solid 1d can
be directly reused in the second run of reaction simply by
recharging the vessel with LA, KOH, and i-PrOH, and no sig-
nificant loss in the yield of GVL was observed. At 270 ppm cat-
alyst loading, solid 1d could be reused for 13 runs under the
optimized reaction conditions (4 h, Figure 4c, light blue). By
slightly extending the reaction time to 24 h, quantitative yields
could still be achieved for two more runs (Figure 4c, dark blue).
Additionally, the reusability of POMP-NHC-Ir 8 was also car-
ried out under the otherwise identical reaction conditions. Poor
yield was found even at the first run, and continuously declined
yields were observed in the subsequent cycles, probably due to
random anchoring and existence of mono-NHC-Ir species in
the solid matrix of 8 (Figure 4d).
3
exhibited much higher catalytic activities under otherwise
identical reaction conditions (17% vs 99%, 18% vs 82%, and
6% vs 81%, Figure 4a). It has to be pointed out, there are few
reports on Pd-catalyzed hydrogenation of LA to GVL under
5
hydrogen atmosphere,[
51,52]
and good result and enhanced
catalytic activity could be achieved by POMP-NHC-Pd 2 even
at 130 ppm catalyst loadings. Furthermore, other privileged
and commercially available catalysts (10-12)[
46–50]
were also
employed in this transformation, only trace to inferior yields
were obtained under optimal reaction conditions (Figure 4a).
Although our previously reported self-supported catalyst 13
also gave a quantitative yield under the identical reaction condi-
tions,[ a record turnover number (TON) of 1.01 × 10 could be
achieved by further decreasing the catalyst loading of POMP-
NHC-Ir 1d to 0.678 ppm (Figure 4b). This TON value is one
order of magnitude higher than that of self-supported catalyst
31]
6
The properties of solids 1d and 8 during the hydrogenation
process were characterized by ICP-AES, TEM, XPS, and FT-IR.
Adv. Mater. 2019, 1905950
1905950 (6 of 9)
© 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
ICP-AES revealed that after the first run with solid 1d, negligible
iridium leaching occurred in subsequent runs (Table S5, Sup-
porting Information). The metal leaching (1.6% of the original
Ir content in solid 1d) observed in the first run could be attrib-
uted to decomposed iridium species that were trapped in the
pores of the matrix during the knitting process. TEM images
revealed that the freshly prepared and recovered 1d had similar
morphologies (Figure 1a vs Figure S19, Supporting Informa-
tion). And XPS analysis gave almost identical peaks for the
fresh and recovered 1d (Figure 3c). These results clearly dem-
onstrate that no iridium agglomeration or cluster formation
occurred during recycle of the solid 1d, and thus the polymer
matrix is helpful for avoiding dimerization and aggregation
were also involved. Compared with their molecular precursors,
higher catalytic activities were observed with the corresponding
POMPs (17% vs 99%, and 4% vs 60%, respectively, Table S2,
entries 12–13, Supporting Information). In this context, we
believed the formation of iridium dimer species with molecular
bis-NHC-Ir complex 4a in the LA to GVL transformations was
the only reason leading to the lower activity. In addition, when
the POMP-NHC-Ir 1j with mainly micropores was employed
in the hydrogenation of LA to GVL under standard reaction
conditions, a decreased yield (87%) was obtained, suggesting
macropores may indeed accelerate the mass and heat transfer
to increase the transformation efficiency (Table S2, entry 21,
Supporting Information).
which are often observed in homogenous bis-NHC-Ir com-
With all these control experiments, the ultrahigh activity of
solid 1d can be mainly attributed to i) isolation effect endowed
by the linkers within the matrix, which efficiently prevented the
formation of possible inactive dimers, oligomers, or nanoparti-
cles from bis-NHC-Ir motifs during the reactions; ii) the struc-
turally preserved bis-chelating modes and atomically dispersed
metal centers with similar retained chemical environment to its
molecular catalyst; and iii) its unique hierarchical porosity, in
which the micropores potentially enhanced its catalytic activity
plexes.[
53–56]
In contrast, high iridium leaching was observed
with solid 8 accessed by postmodification (9.4% lost after three
runs, Table S6, Supporting Information), and Ir(0) species were
observed in the recovered solid 8, confirmed by XPS analysis
(Figure S43, Supporting Information). Further TEM study
confirmed the formation of iridium nanoparticles during the
hydrogenation reaction catalyzed by solid 8 (Figure S21, Sup-
porting Information).
Noticeably, an absorption band at 2088 cm− (attributed to
CO ligands) was not visible in the FT-IR spectrum of recovered
solid 1d (Figure S45 in the Supporting Information, blue line vs
red line). This result implied that CO ligands were weakly coor-
dinated to the Ir center and might dissociate during the reac-
tion process. To confirm this assumption, a control experiment
was performed in which CO was bubbled through a suspension
of recovered solid 1d, and this process was conducted according
to a literature procedure for preparation of homogeneous bis-
1
due to pore confinement effects, and the mesopores and
[7]
[
58]
the macropores allowed rapid mass and heat transfer.
All
of these issues are critical and helpful for the reaction activity
improvements with POMPs.
Because bis-NHC-Ir complexes can serve as bifunctional
[
25,26]
catalysts in various transformations,
after investigation of
their applicability in the hydrogenation reaction, we further
explored the bifunctional catalytic applicability of solid 1d in
oxidative reactions. Considering the increasing global demand
for lactic acid, which is a raw material in the syntheses of
biodegradable polylactide materials, pharmaceuticals, foods
[25]
NHC-Ir-(CO) complexes. To our delight, the CO signal re-
2
appeared after CO bubbling (Figure S45, black line, Supporting
Information), confirming that CO was weakly bound to the Ir(I)
center. This observation revealed that POMPs may behave like
homogeneous molecular catalysts, allowing precise design and
control of immobilized catalysts at the molecular level. And
the slight alternation of coordination environment around the
Ir center in the solid 1d did not affect its catalytic activity, as
supported by the unchanged high catalytic efficiency during the
catalyst recycling experiments (Figure 4c).
[
59]
and green solvents,
its synthesis via the dehydrogenation
of glycerol, a main waste product of the biodiesel and soap
[
60]
industries,
was selected as a model reaction. As expected,
solid 1d exhibited excellent activity and selectivity (99%) in this
challenging transformation even at 340 ppm catalyst loading
(Table S7, Supporting Information). In contrast, POMP-NHC-
Ir 8, prepared by the postmodification approach, hardly acceler-
ated the transformation, and almost no product was detected
under otherwise identical reaction conditions, further high-
lighting the advantage of the direct knitting approach. Solid
1d could also be reused for six times. After the sixth run, over
90% yield was still achieved (Figure S56, Supporting Informa-
tion). ICP-AES analysis indicated that > 99% of the Ir(I) spe-
cies were still immobilized in the solid polymer matrix after
six runs at 115 °C for 36 h (Table S8, Supporting Information).
Similar TEM morphologies, XPS images and FT-IR spectra
were observed for the freshly prepared and recovered solid 1d
(Figures S20, S44, and S46). All these results clearly demon-
strated that solid 1d could function as a highly efficient, bifunc-
tional, single-site solid catalyst.
To further probe why the performance of POMPs 1d, 2
and 3 exhibited much higher catalytic activity than the cor-
responding molecular complexes 4–6, several control experi-
ments were carried out. In the case of bis-NHC-Pd(II) 5 and
NHC-Ru(II) 6, Pd and Ru nanoparticles were found in the TEM
images after reactions (Figures S22 and S23, Supporting Infor-
mation) and also could be successfully isolated by centrifuga-
tion. Mercury drop test[ (Table S3, Supporting Information)
with no obvious decrease in yield suggested the metal nano-
particles formed during the process hardly catalyzed reaction.
Hence, the inferior yields obtained by complexes 5 and 6 might
be attributed to the formation of inactive metal nanoparticles
during the hydrogenation. However, when two drops of Hg
were also added into the reaction mixture with bis-NHC-Ir 4a
as catalyst, neither decrease in yield nor iridium nanoparticles
57]
In summary, we have successfully fabricated a series of
robust porous organometallic polymers (POMPs) with hierar-
chical pore structures, high specific surface areas and atomi-
cally dispersed metal sites (Ir, Pd, and Ru) via a direct knitting
approach. The intrinsic properties of the active and structure-
defined bis-chelating NHC-metal complexes with different
were found. Taking account of the possible dimer formation
with bis-NHC-Ir complexes,[
53–56]
POMPs-NHC-Ir 1h-i by
direct knitting of complexes 4a-b without benzene addition
Adv. Mater. 2019, 1905950
1905950 (7 of 9)
© 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
valence states (+1 to +2) are highly preserved during immo- Conflict of Interest
bilization, enabling these solid POMPs function as single-site
The authors declare no conflict of interest.
bifunctional catalysts with excellent activity and selectivity in
both dehydrogenation and hydrogenation reactions. Remark-
6
ably, a record turnover number (TON: 1.01 × 10 ) was achieved
in the hydrogenation of levulinic acid to γ-valerolactone, and Keywords
the POMP-NHC-Ir could be reused for up to 15 runs without
biomass transformation, direct knitting, N-heterocyclic carbene metal
complexes, porous organometallic polymers, single-site catalysts
obvious losses of catalytic activity or selectivity even at ppm-
level catalyst loadings. This study demonstrates not only that
hierarchical porous single-site solid catalysts can be readily
fabricated by direct knitting of privileged organometallics but
also that these immobilized catalysts can be precisely designed
and controlled at the molecular level. All these outcomes may
pave the way for direct accessing and tuning recyclable hetero-
geneous catalysts based on privileged homogeneous catalysts
while maintaining or even enhancing the catalytic activities.
Following work of extending the direct knitting approach to
other privileged homogeneous catalysts, such as metalloporphy-
rins and metal-pincer complexes, are ongoing in our laboratory.
Received: September 11, 2019
Revised: November 13, 2019
Published online:
[
[
1] Z. Wang, G. Chen, K. Ding, Chem. Rev. 2009, 109, 322.
2] Handbook of Asymmetric Heterogeneous Catalysis (Eds: K. Ding,
Y. Uozumi), Wiley-VCH, Weinheim 2008.
[
3] J. A. Gladysz, Chem. Rev. 2002, 102, 3215.
[4] A. G. Slater, A. I. Cooper, Science 2015, 348, aaa8075.
[
[
[
5] J. Liu, N. P. Wickramaratne, S. Z. Qiao, M. Jaroniec, Nat. Mater.
015, 14, 763.
6] J. Wu, F. Xu, S. Li, P. Ma, X. Zhang, Q. Liu, R. Fu, D. Wu, Adv.
Mater. 2019, 31, 1802922.
2
Experimental Section
7] Q. Sun, Z. Dai, X. Meng, F.-S. Xiao, Chem. Soc. Rev. 2015, 44, 6018.
Synthesis of POMPs-NHC-M 1–3: In the case of POMP-NHC-Ir 1d,
[25]
bis-NHC-Ir complex 4a (0.1 mmol, 63 mg) and benzene (or 0.3 mmol,
3 mg) were dissolved in 1 mL of dry anaerobic 1,2-dichloroethane (DCE)
under N atmosphere. After stirring, anhydrous FeCl (2.0 mmol, 324 mg)
[8] G. Li, S. Zhao, Y. Zhang, Z. Tang, Adv. Mater. 2018, 30, 1800702.
[9] S. Kramer, N. R. Bennedsen, S. Kegnæs, ACS Catal. 2018, 8, 6961.
[10] D. Zhao, D. J. Timmons, D. Yuan, H.-C. Zhou, Acc. Chem. Res.
2011, 44, 123.
[11] W. Wang, A. Zheng, P. Zhao, C. Xia, F. Li, ACS Catal. 2014, 4, 321.
[12] Z. Xie, C. Wang, K. E. deKrafft, W. Lin, J. Am. Chem. Soc. 2011, 133,
2
2
3
and formaldehyde dimethyl acetal (FDA, 2.0 mmol, 152 mg) were added.
The resulting mixture was heated to 80 °C and stirred for 24 h under a
nitrogen atmosphere. After cooling to room temperature, the precipitate
was collected and washed by methanol, distilled water, dichloromethane,
and acetone successively. Further purification of the polymer was carried
out by Soxhlet extraction from methanol for 48 h. The product was dried
in vacuum for 24 h at 60 °C to give a dark brown powder (Yield: 90%).
The preparation of POMPs-NHC-Ir 1a-c and 1e-g was similar to that of
2
056.
[
[
[
[
13] L. Tan, B. Tan, Chem. Soc. Rev. 2017, 46, 3322.
14] S. Xu, K. Song, T. Li, B. Tan, J. Mater. Chem. A 2015, 3, 1272.
15] Z. Jia, K. Wang, T. Li, B. Tan, Y. Gu, Catal. Sci. Technol. 2016, 6, 4345.
16] B. Li, Z. Guan, W. Wang, X. Yang, J. Hu, B. Tan, T. Li, Adv. Mater.
POMP-NHC-Ir 1d, except the amount of FeCl , FDA and benzene were
3
2
012, 24, 3390.
varied (see the Supporting Information). In the case of solids 1h and 1i,
bis-NHC-Ir complexes 4a and 4b were directly knitted without benzene.
The POMP 1j was prepared by using similar procedure to that of
POMP-NHC-Ir 1d, except that 1,3,5,7-tetraphenyladamantane was used
instead of benzene. Following the same synthetic procedure of solid 1d,
POMPs-NHC-Pd 2 and NHC-Ru 3 were readily obtained from molecular
complexes bis-NHC-Pd(II) 5 and NHC-Ru(II) 6, respectively. In addition,
after tableting the resulted POMPs powder into sheet, the conductivity of
these polymer materials was also investigated by multimeter and four-
point probe method, no obvious conductivity was found.
[
17] B. Li, R. Gong, W. Wang, X. Huang, W. Zhang, H. Li, C. Hu, B. Tan,
Macromolecules 2011, 44, 2410.
[18] Y. Zhang, S. N. Riduan, Chem. Soc. Rev. 2012, 41, 2083.
[19] A. Doddi, M. Peters, M. Tamm, Chem. Rev. 2019, 119, 6994.
[20] C. Fliedel, A. Labande, E. Manoury, R. Poli, Coord. Chem. Rev. 2019,
3
94, 65.
[
21] D. A. Hey, R. M. Reich, W. Baratta, F. E. Kühn, Coord. Chem. Rev.
2018, 374, 114.
[22] S. Shi, S. P. Nolan, M. Szostak, Acc. Chem. Res. 2018, 51, 2589.
[
23] D. Janssen-Müller, C. Schlepphorst, F. Glorius, Chem. Soc. Rev.
017, 46, 4845.
2
[
24] S. Movassaghi, S. Singh, A. Mansur, K. K. H. Tong, M. Hanif,
H. U. Holtkamp, T. Söhnel, S. M. F. Jamieson, C. G. Hartinger,
Organometallics 2018, 37, 1575.
Supporting Information
Supporting Information is available from the Wiley Online Library or
from the author.
[
[
[
[
[
25] M. G. Manas, J. Campos, L. S. Sharninghausen, E. Lin,
R. H. Crabtree, Green Chem. 2015, 17, 594.
26] J. Campos, U. Hintermair, T. P. Brewster, M. K. Takase,
R. H. Crabtree, ACS Catal. 2014, 4, 973.
27] G. M. Roberts, P. J. Pierce, L. K. Woo, Organometallics 2013, 32,
Acknowledgements
2
033.
28] C. F. Rentzsch, E. Tosh, W. A. Herrmann, F. E. Kühn, Green Chem.
009, 11, 1610.
29] R. Zhong, A. C. Lindhorst, F. J. Groche, F. E. Kühn, Chem. Rev.
017, 117, 1970.
Financial support from the Ministry of Science and Technology of China
2
(No. 2016YFA0202902), the National Natural Science Foundation
of China (Nos. 21861132002, 21871059, and 21572036), and the
Department of Chemistry, Fudan University is gratefully acknowledged.
The authors also thank Prof. Yongning Zhou, Dr. Jiong Li, and the
BL14W1 at the Shanghai Synchrotron Radiation Facility for XAS
experiment.
2
[30] W. Wang, L. Cui, P. Sun, L. Shi, C. Yue, F. Li, Chem. Rev. 2018, 118,
9843.
[31] Y. Liu, Z. Sun, C. Huang, T. Tu, Chem. - Asian J. 2017, 12, 355.
Adv. Mater. 2019, 1905950
1905950 (8 of 9)
© 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
[
[
[
32] Z. Sun, J. Chen, T. Tu, Green Chem. 2017, 19, 789.
33] Z. Sun, Y. Liu, J. Chen, C. Huang, T. Tu, ACS Catal. 2015, 5, 6573.
34] S. M. Sarkar, Y. Uozumi, Y. M. Yamada, Angew. Chem. 2011, 123,
[46] X. Zhang, P. Zhang, C. Chen, J. Zhang, G. Yang, L. Zheng, J. Zhang,
B. Han, Green Chem. 2019, 21, 54.
[47] S. Li, Y. Wang, Y. Yang, B. Chen, J. Tai, H. Liu, B. Han, Green Chem.
2019, 21, 770.
9
609; Angew. Chem., Int. Ed. 2011, 50, 9437.
[
[
[
35] K. H. G. Mak, P. K. Chan, W. Y. Fan, W. K. Leong, Y. Li, J. Organomet.
Chem. 2013, 741-742, 176.
36] A. V. Neimark, Y. Lin, P. I. Ravikovitch, M. Thommes, Carbon 2009,
[48] D. M. Alonso, S. G. Wettstein, J. A. Dumesic, Green Chem. 2013, 15,
584.
[49] J. Q. Bond, D. M. Alonso, D. Wang, R. M. West, J. A. Dumesic,
Science 2010, 327, 1110.
[50] L. Deng, J. Li, D. M. Lai, Y. Fu, Q. X. Guo, Angew. Chem. 2009, 121,
6651; Angew. Chem., Int. Ed. 2009, 48, 6529.
[51] B. Banerjee, R. Singuru, S. K. Kundu, K. Dhanalaxmi, L. Bai,
Y. Zhao, B. M. Reddy, A. Bhaumik, J. Mondal, Catal. Sci. Technol.
2016, 6, 5102.
[52] C. Michel, J. Zaffran, A. M. Ruppert, J. Matras-Michalska,
M. Jedrzejczyk, J. Grams, P. Sautet, Chem. Commun. 2014, 50,
12450.
[53] J. Campos, L. S. Sharninghausen, R. H. Crabtree, D. Balcells, Angew.
Chem., Int. Ed. 2014, 53, 12808; Angew. Chem. 2014, 126, 13022.
[54] L. S. Sharninghausen, J. Campos, M. G. Manas, R. H. Crabtree,
Nat. Commun. 2014, 5, 5084.
[55] S. P. Smidt, A. Pfaltz, E. Martínez-Viviente, P. S. Pregosin,
A. Albinati, Organometallics 2003, 22, 1000.
[56] R. Crabtree, Acc. Chem. Res. 1979, 12, 331.
[57] K. Yu, W. Sommer, J. M. Richardson, M. Weck, C. W. Jones, Adv.
Synth. Catal. 2005, 347, 161.
[58] B. Fang, J. H. Kim, M.-S. Kim, J.-S. Yu, Acc. Chem. Res. 2013, 46,
1397.
[59] M. Dusselier, P. Van Wouwe, A. Dewaele, E. Makshina, B. F. Sels,
Energy Environ. Sci. 2013, 6, 1415.
[60] M. R. Monteiroa, C. L. Kugelmeierb, R. S. Pinheiroc, M. O. Batalhad,
A. da Silva Césare, Renewable Sustainable Energy Rev. 2018, 88, 109.
4
7, 1617.
37] M. Rodriguez-Castillo, G. Lugo-Preciado, D. Laurencin, F. Tielens,
A. van der Lee, S. Clement, Y. Guari, J. M. Lopez-de-Luzuriaga,
M. Monge, F. Remacle, S. Richeter, Chem. - Eur. J. 2016, 22, 10446.
38] B. J. van Weerdenburg, A. H. Engwerda, N. Eshuis, A. Longo,
D. Banerjee, M. Tessari, C. F. Guerra, F. P. Rutjes, F. M. Bickelhaupt,
M. C. Feiters, Chem. - Eur. J. 2015, 21, 10482.
[
[
[
39] X. Su, X. F. Yang, Y. Huang, B. Liu, T. Zhang, Acc. Chem. Res. 2019,
5
2, 656.
40] S. Wei, A. Li, J. C. Liu, Z. Li, W. Chen, Y. Gong, Q. Zhang,
W. C. Cheong, Y. Wang, L. Zheng, H. Xiao, C. Chen, D. Wang,
Q. Peng, L. Gu, X. Han, J. Li, Y. Li, Nat. Nanotechnol. 2018, 13, 856.
41] A. Han, W. Chen, S. Zhang, M. Zhang, Y. Han, J. Zhang, S. Ji,
L. Zheng, Y. Wang, L. Gu, C. Chen, Q. Peng, D. Wang, Y. Li, Adv.
Mater. 2018, 30, 1706508.
[
[
42] P. Liu, Y. Zhao, R. Qin, S. Mo, G. Chen, L. Gu, D. M. Chevrier,
P. Zhang, Q. Guo, D. Zang, B. Wu, G. Fu, N. Zheng, Science 2016,
3
52, 797.
43] X.-F. Yang, A. Wang, B. Qiao, J. Li, J. Liu, T. Zhang, Acc. Chem. Res.
013, 46, 1740.
[
[
[
2
44] P.-F. Wei, M.-Z. Qi, Z.-P. Wang, S.-Y. Ding, W. Yu, Q. Liu, L.-K. Wang,
H.-Z. Wang, W.-K. An, W. Wang, J. Am. Chem. Soc. 2018, 140, 4623.
45] W. Cao, L. Lin, H. Qi, Q. He, Z. Wu, A. Wang, W. Luo, T. Zhang,
J. Catal. 2019, 373, 161.
Adv. Mater. 2019, 1905950
1905950 (9 of 9)
© 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim