HAL Id: hal-02081087
https://hal.archives-ouvertes.fr/hal-02081087
Submitted on 27 Mar 2019
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entic research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diusion de documents
scientiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
A Fully Noble Metal-Free Photosystem Based on
Cobalt- Polyoxometalates Immobilized in a Porphyrinic
Metal- Organic-Framework for Water Oxidation
Grégoire Paille, Maria Gomez-Mingot, Catherine Roch-Marchal, Benedikt
Lassalle-Kaiser, Pierre Mialane, Marc Fontecave, Caroline Mellot-Draznieks,
Anne Dolbecq
To cite this version:
Grégoire Paille, Maria Gomez-Mingot, Catherine Roch-Marchal, Benedikt Lassalle-Kaiser, Pierre Mi-
alane, et al.. A Fully Noble Metal-Free Photosystem Based on Cobalt- Polyoxometalates Immobilized
in a Porphyrinic Metal- Organic-Framework for Water Oxidation. Journal of the American Chemical
Society, American Chemical Society, 2018, �10.1021/jacs.7b11788�. �hal-02081087�
A Fully Noble Metal-Free Photosystem Based on Cobalt-
Polyoxometalates Immobilized in a Porphyrinic Metal-
Organic-Framework for Water Oxidation
Grégoire Paille,
§,‡
Maria Gomez-Mingot,
‡
Catherine Roch-Marchal,
§
Benedikt Lassalle-Kaiser,† Pierre
Mialane,
§
Marc Fontecave,
*,‡
Caroline Mellot-Draznieks,
*,‡
and Anne Dolbecq
*,§
§Institut Lavoisier de Versailles, UMR CNRS 8180, Université de Versailles Saint-Quentin en Yvelines, Université Paris-
Saclay, 45 Avenue des Etats-Unis, 78035 Versailles Cedex, France
‡
Laboratoire de Chimie des Processus Biologiques, UMR CNRS 8229, Collège de France, Université Pierre et Marie Curie,
PSL Research University, 11 Place Marcelin Berthelot, 75231 Paris Cedex 05, France
†Synchrotron Soleil, l’Orme des Merisiers, Saint-Aubin, 91192 Gif-sur-Yvette Cedex, France
ABSTRACT: The sandwich-type polyoxometalate (POM) [(PW
9
O
34
)
2
Co
4
(H
2
O)
2
]
10-
was immobilized in the hexagonal channels of
the Zr(IV) porphyrinic MOF-545 hybrid framework. The resulting composite was fully characterized by a panel of physicochemical
techniques. Calculations allowed identifying the localization of the POM in the vicinity of the Zr
6
clusters and porphyrin linkers
constituting the MOF. The material exhibits a high photocatalytic activity and good stability for visible-light-driven water oxida-
tion. It thus represents a rare example of an all-in-one fully noble metal-free supramolecular heterogeneous photocatalytic system,
with the catalyst and the photosensitizer within the same porous solid material.
INTRODUCTION
In the current energetic transition, one major challenge con-
cerns the development of noble metal-free, selective, efficient,
and recyclable heterogeneous photocatalysts. In this respect, the
design of new multifunctional porous and robust hybrid solids
bringing together i) catalysis, ii) light capture and iii) porosity is a
particularly attractive alternative to create novel “three-in-one”
photoactive catalysts. Due to their porosity and the functional
tunability of their organic linkers,
1
metal-organic frameworks
(MOFs) represent an ideal platform. MOFs may incorporate redox
catalytic centers while organic linkers may behave as light-
harvesting units capable of transferring electrons to neighboring
catalytic centers.
2
Furthermore, synthetic or post-synthetic strate-
gies allow grafting or encapsulating additional functional units
within the MOF’s pores.
3
For example some of us have recently
reported the heterogenization of a Rh-based molecular catalyst as
the constitutive linker of a mixed-linker UiO-67, providing an
active and recyclable solid catalyst for CO
2
photoreduction.
4
Polyoxometalates (POMs) can also play the role of functional
catalytic units. POMs are soluble anionic metal oxide clusters of
d-block transition metals in high oxidation states (usually W
VI
,
Mo
V,VI
, V
IV,V
) exhibiting properties that can be exploited in many
fields.
5
In particular, POMs can undergo multielectron redox
transformations conferring them catalytic redox activity,
6
while
being also known as proton and electron relays.
7
Several studies
have reported the successful incorporation of POMs into MOFs
cavities leading to the so-called POM@MOFs.
8
In that context,
one of the most studied MOFs for encapsulation is the highly
porous MIL-101,
9
but HKUST-1,
10
and more recently ZIF
11
and
Zr-based MOFs
12
have also been investigated as robust host struc-
tures. Besides, photocatalytic properties of POMs in the presence
of molecular porphyrins were investigated in homogeneous condi-
tions for various reactions such as reduction of silver cations,
13
hydrogen
14
and oxygen
15
evolution reactions. These studies evi-
denced that porphyrins may be efficiently used for visible light
sensitization of POMs.
In the present work, we report the design of the first noble
metal-free heterogeneous photosystem using a POM as a catalyst
immobilized in the pores of a porphyrinic MOF. We selected
MOF-545,
16
also known as PCN-222
17
and MMPF-6,
18
for its
unique properties: i) a high surface area thanks to hexagonal
channels large enough to accommodate POMs (Figure 1), ii) an
excellent chemical and thermal stability, iii) the ability to capture
visible light due to the porphyrin linker. It has also proved being
an efficient host for biomimetic iron complexes.
19
We selected the
[(PW
9
O
34
)
2
Co
4
(H
2
O)
2
]
10-
POM (named P
2
W
18
Co
4
, Figure 1) as
the catalytically active guest. This POM exhibits a tetracobalt
oxide core sandwiched between two [PW
9
O
34
]
9-
polyoxotungstate
cages and is known for its homogenous photocatalytic activity for
oxygen evolution reaction (OER) in the presence of a ruthenium-
based molecular photosensitizer.
20
The new POM@MOF photo-
system reported here was fully characterized and evaluated for its
photocatalytic performances for water oxidation. Besides, density
functional theory calculations provided insights into the unique
structural features of the POM-MOF interface.
RESULTS AND DISCUSSION
Synthesis. The encapsulation of the POMs was performed by
mild aqueous impregnation of MOF-545 with an excess of the
alkaline salt of the P
2
W
18
Co
4
POM, monitored by UV-Vis spec-
troscopy of the supernatant solution. Once the MOF was added to
the solution for impregnation, the intensity of the Co-based d-d
absorption at 566 nm gradually decreased and stabilized after 6
hours (Figure S1a). The amount of POMs deduced from the dif-
ference in absorbance of the solution before and after impregna-
tion is ~ 1 POM per unit of MOF and represents the amount of
POMs encapsulated in the MOF pores plus the amount of POMs
adsorbed at the surface of the MOF particles.
Figure 1. POM@MOF-545 components. (a) P
2
W
18
Co
4
POM;
(b) TCPP-H
2
linker; (c) Zr-based unit; (d) P
2
W
18
Co
4
@MOF-545.
The position of the POM is obtained from computations (see text).
WO
6
, green polyhedra; ZrO
8
, blue polyhedral or spheres; Co,
cyan spheres; O, red spheres; C, H, grey; N, dark blue.
The POM@MOF is then filtrated and washed with water. The
absorbance of the first washing solution indicates that ~0.45 POM
per unit of MOF is released during washing (Figure S1c). The
following washing solutions no longer contain POMs. These
experiments show that the amount of encapsulated POMs is esti-
mated ~0.55 POM per unit of MOF i.e. 0.18 POM per {Zr
6
} units.
The composite material, named P
2
W
18
Co
4
@MOF-545, was then
synthesized in large quantities, carefully washed with water and
analyzed by various techniques.
Characterizations. Elemental analysis (Table S1) allows
proposing the formula [Zr
6
O
16
H
18
][TCPP-
H
2
]
2
[P
2
W
18
Co
4
]
0.2
•26H
2
O and confirms the amount of encapsulat-
ed POM determined by UV-Vis spectroscopy. SEM-EDS ele-
mental mapping shows a uniform distribution of the Zr, Co and W
elements in the bulk material of impregnated MOF (Figure S2)
and indicate average Zr/Co and Zr/W ratios consistent with the
results of elemental analysis (Table S2). STEM-HAADF images
coupled to EDS mapping of the various elements were also rec-
orded (Figure 2). On the one hand, they show that the solid mate-
rial consists of rod-like shape crystals of 2.4 x 0.4 x 0.4 m aver-
age dimensions and nicely confirm the localization of the POM-
specific W and Co elements within the MOF crystals, while the
MOF host is characterized by Zr, N and O mapping. On the other
hand, they also show that the POM species exhibit higher concen-
trations at both extremities of the crystal rods, consistent with the
alignment of the channels of the MOF along the c axis, i.e. the
longest dimension of the rods. Thermogravimetric analysis (TGA)
measurements show mass losses of ca. 53% for the
P
2
W
18
Co
4
@MOF-545 composite and 65% for the bare MOF-545
(Figure S3). These weight losses are assigned to water removal,
linker decomposition and formation of inorganic oxides. The
lower weight loss of P
2
W
18
Co
4
@MOF-545 with respect to the
bare MOF-545 is in agreement with the presence of POMs in the
POM@MOF material (Table S3). While this adventitious popula-
tion is observed during encapsulation by UV-vis spectroscopy, it
is absent in samples characterized by elemental analysis, carried
out after the washing steps. In the proposed formula, the negative
charge of the POMs is likely compensated by the protonation of
the hybrid framework as confirmed by the absence of alkaline
cations shown by EDS analysis (Table S2). Consistently, the
isoelectric point of MOF-545 is at pH 8,
21
indicating that the
MOF is indeed cationic under the pH synthetic conditions. Rely-
ing on IR spectroscopy
22
and DFT
23
studies on the related NU-
1000 material, we thus propose the following proton localization
on the charged Zr
6
-clusters, [Zr
(
3
-O)
4
(
3
-OH)
4
(OH)
2
(H
2
O)
6
]
10+
.
Figure 2. STEM-HAADF images of POM@MOF-545 and EDS
mapping of the various elements contained in the POM (W, Co,
O) and MOF (Zr, O, N).
Brunauer–Emmett–Teller (BET) surface area measurements
(Figure 3a) calculated from the N
2
adsorption/desorption iso-
therms show the expected correlation between decreased surface
areas and POM encapsulation, from 2080 m
2
g
-1
for the bare MOF
to 1180 m
2
g
-1
for P
2
W
18
Co
4
@MOF-545. Furthermore, the pore
distribution is also found to be strongly modified upon POMs’
encapsulation (Figure S4). These measurements indicate that the
POMs are indeed located in the MOF’s largest pores, i.e. the
hexagonal channels. Indeed a decrease in intensity of the peak
attributed to the hexagonal pores is observed while the peak at-
tributed to the triangular pores remains unchanged. This is not
surprising considering their large diameter (~36 Å) compatible
with the adsorption of the bulky POMs species (~16 Å) unlike the
triangular channels which have too small diameters (~16 Å) (Fig-
ure 1).
Figure 3. (a) BET N
2
adsorption/desorption isotherms (77 K,
P/P
0
= 1 atm.) of MOF-545 and P
2
W
8
Co
4
@MOF-545. (b)
XANES spectra of P
2
W
18
Co
4
, P
2
W
18
Co
4
@MOF-545 before and
after catalysis, compared to related oxides.
The powder X-ray diffraction patterns of the bare MOF-545
and of the POM@MOF-545 composite (Figure S5) confirm that
the crystallinity of the MOF host is maintained upon incorporation
of the POM. The UV-Vis spectrum of P
2
W
18
Co
4
@MOF-545
merges the characteristic bands of the MOF-545
at 390 and 400-
700 nm, and those of the P
2
W
18
Co
4
POMs around 300 and 560
nm (Figure S6). Co K-edge X-ray absorption near edge spectra
(XANES) of P
2
W
18
Co
4
and P
2
W
18
Co
4
@MOF-545 were also
[Zr
6
(
3
-O)
4
(
3
-OH)
4
(OH)
2
(H
2
O)
6
]
10+
[TCPP-H
2
]
4-
[(PW
9
O
34
)
2
Co
4
(H
2
O)
2
]
10-
a)
b)
c)
d)
P
2
W
18
Co
4
@MOF-545
~16 Å
~36 Å
~15 Å
1 m
POM
POM@MOF
before
POM@MOF
after
CoO
Co
3
O
4
CoOOH
b)
a)
POM
POM@MOF
before
POM@MOF
after
CoO
Co
3
O
4
CoOOH
b)
a)
.
Figure 4. Computed position of the P
2
W
18
Co
4
POM in MOF-545. (a) “Side” and (b) “top” views of the POM positioned between two Zr
6
-
clusters and two porphyrins. (c) Detailed lateral view of the POM-MOF interface and the hydrogen-bond network. One porphyrin is not
represented for more clarity. Distances are given in Å. The external Co centers with water molecules ligands are highlighted in bright cyan
collected (Figure 3b). The similarity of the two spectra shows that
the POM is intact after its immobilization within the MOF. The
slight difference can be attributed to the electrostatic interaction
between the POMs and the MOF, as observed for
P
2
W
18
Co
4
@MIL-101(Cr).
9c
Also, the comparison with reference
cobalt oxides indicates the absence of degradation into the typical
cobalt oxides known to be active for OER activity (Co
3
O
4
or
CoOOH).
DFT Calculations. To probe the {POM, MOF} potential ener-
gy surface and the most likely positions of the POM within the
pores, we applied a combination of simulated annealing (SA)
calculations and DFT-D3 level geometry optimizations (see text
and Figures S7-S10 in SI for details). SA calculations show that
the insertion of POMs occurs exclusively in the hexagonal chan-
nels of MOF-545 in line with the above findings (Figure S8).
Figures 4a and 4b illustrate the position of the POM in MOF-
545’s channels as obtained from DFT-D3 level geometry optimi-
zation of the lowest energy position extracted from SA calcula-
tions, in a “side” view and a “top” view, respectively. This com-
puted position reflects the chemical environment that the POM
may adopt when adsorbed at the MOF’s internal surface before
catalysis. The POM is located in the vicinity of two Zr
-clusters
connected through two porphyrinic linkers. It is stabilized by a
particularly dense network of hydrogen bonds that involve –OH
groups and H
2
O molecules belonging either to the POM or the
MOF. As illustrated in Figure 4c, these H-bonds are concentrated
at the POM-MOF interface mainly around one of two external
Co-OH
2
centers of the tetracobalt oxide core. One Zr
6
-cluster
allows the anchoring of the POM to the MOF’s surface through
strong H-bonds: i) between the terminal water of the Co1 center,
H
2
O
Co1
, and an –OH
Zr1
group of the MOF,
(O(H
2
O
Co1
)
…
H(OH
Zr1
)=1.7 Å); ii) between oxygen atoms of the
WO
6
moieties and hydrogen atoms of the MOF, the H atoms
belonging to a
3
-OH group (O
W
…
H(
3
- OH
Zr
)=1.5 Å) and a
terminal water molecule of the Zr-cluster (O
W
…H(H
2
O
Zr1
)=1.5
Å). The other Zr
6
-cluster provides further stabilization of the
POM thanks to a more peripheral H-bond (with respect to the Co-
core) between terminal hydroxyl groups at the POM-MOF inter-
face (HO
W
…
OH
Zr2
=1.7 Å). DFT calculations indicate that the
host-guest interactions (mainly hydrogen bonds and electrostatic
interactions) are very strong (~ 176 kcal mol
-1
). Interestingly, the
computed structure reveals a shuttling of protons at the POM-
MOF interface from the MOF to the POM (see SI for details). The
basic character of the POM is thus apparent with the formation of
the two terminal –OH groups at its surface, labeled OH
Co-W
and
OH
W
(see Figure 4c). They both result from the transfer of H
atoms from terminal water molecules coordinated to the Zr atoms
(Zr1 and Zr2, respectively, in dark blue in Figure 4c) to oxygen
atoms of the WO
6
moieties, leaving OH
Zr1
and OH
Zr2
hydroxyls at
the MOF’s surface. The protonation of the bridging oxygen Co-
O-W found here is in line with previous computational studies,
24
showing that Co-O-W bridges are the most basic ones. Overall,
DFT calculations reveal that the POM-MOF interface strongly
affects the local arrangement of H
2
O molecules and –OH groups
towards optimized H-bond type host-guest interactions, when
compared to the isolated POM and MOF counterparts.
The recent theoretical study by Hill and Poblet
24
establishes
Co-OH
2
units as the reactive site. Interestingly, our DFT results
suggest that the Co-OH
2
catalytic site exposed at the MOF-POM
interface is hosted within a hydrophilic (water and -OH rich)
catalytic pocket which may provide ideal shuttling of protons and
water molecules transiting from the solvent.
Photocatalytic activities. The photocatalytic OER activity of
P
2
W
18
Co
4
@MOF-545 was studied under visible light irradiation
in pH 8 borate buffer and with Na
2
S
2
O
8
as the electron acceptor.
As shown in Figure 5, O
2
was formed immediately upon exposure
to light and increased linearly with time (TOF = 40 10
-3
s
-1
calcu-
lated for the first 15 min) before reaching a plateau after 1 h of
reaction. Addition of a fresh solution of the sacrificial acceptor,
Na
2
S
2
O
8
, resulted in a new cycle of O
2
production with the same
initial TOF (Figure S11). These results clearly prove that the
system is mainly limited by the consumption of the electron ac-
ceptor. The effect of pH was studied, showing that the optimal pH
value is ~8 (Figure S12). The results show indeed that the
photocatalytic OER activity of the POM@MOF is lower at pH 7.5
and 8.5 than at pH 8. This may be due to the chemical instability
of the POM at pH > 8 and the too acidic conditions for doing the
OER when pH < 8.
25a
P
2
W
18
Co
4
@MOF-545 showed a good
activity during OER with a turnover number (per POM) of 70
after 1 hour reaction. Control experiments with i) no catalyst, ii)
no irradiation, iii) MOF-545 with no encapsulated POM, or iv) a
a) b) c)
solution containing the TCCP-H
2
linker and P
2
W
18
Co
4
, did not
show any significant O
2
evolution (Figure 5 and Figure S13).
Figure 5. Kinetics of visible-light-driven O
2
production meas-
ured by GC analysis over 0.5 mg of P
2
W
18
Co
4
@MOF-545 (blue
square), P
2
W
18
Co
4
@MOF-545 recycled once (red triangle), twice
(pink stars), 131 M TCPP-H
2
and 13 M P
2
W
18
Co
4
in solution
(green circle).
25
Reaction conditions: 5 mM Na
2
S
2
O
8
in 2 mL of
80 mM borate buffer solution, pH 8, visible light (> 420 nm,
280 W).
The fact that TCCP-H
2
is unable to photosensitize P
2
W
18
Co
4
seems to contradict its ability to do it in the context of the
POM@MOF. This difference may be understood in light of the
recent results from Xu et al.
26
who reported that the incorporation
of the TCPP-H
2
linker into the MOF-545 results in its valence
band (HOMO) shifting from 1.24 to 1.35 V vs. NHE, thus increas-
ing the driving force for water oxidation. The reaction mechanism
should thus imply the following steps: (i) light capture by the
porphyrin; (ii) one-electron oxidation of the excited state by the
sacrificial electron acceptor; (iii) one-electron oxidation of the
POM; (iv) after accumulation of 4 oxidizing equivalents on the
POM, oxidation of water into O
2
(Figure 6).
Figure 6. Schematic representation of the proposed mechanism
for the light-driven OER by P
2
W
18
Co
4
@MOF-545.
In order to assess the recyclability of the photocatalytic material,
the reaction was performed using the POM@MOF recovered after
3 h, and assayed in an additional photocatalytic run. A decrease of
the TON of ~11 % was then observed (Figure 5). A third recov-
ery-catalytic cycle was performed, showing an even smaller loss
of TON (<5 %). Even though the final TONs are lower, we note
that the initial rates remain similar. Moreover, TGA (Figure S3),
XRD (Figure S5) and EDS analyses (Table S2) did not show any
difference between the composite before and after the reaction,
confirming its stability upon photocatalysis. The XANES spec-
trum (Figure 3b) recorded on P
2
W
18
Co
4
@MOF-545 after a 3 h
photocatalytic experiment shows a slight shift of the main edge
position towards higher energies (0.4 eV) with respect to the
initial material. Given that the pre-edge region did not present any
change in intensity or any new peak, we exclude a net change in
the cobalt oxidation state. We rather attribute this shift to changes
in the local environment of the cobalt and exclude any drastic
modification in its structure. These results are in line with the
studies of Schiwon et al.
27
who demonstrated the stability of the
same POM under chemically induced OER.
CONCLUSIONS
In summary, this is the first time that a porphyrinic
POM@MOF system, devoid of any noble metal, is used for visi-
ble-light water oxidation in aqueous solution. While very few
studies have achieved the stable incorporation of the two key
components (the photosensitizer and the catalysts) of a photosys-
tem within a solid MOF, for CO
2
or proton photoreduction,
19,26,31
this is also the first such system developed for water
photooxidation. Calculations provided valuable information on
the localization of the POMs in the pores, especially a detailed
view of the POM-MOF interface that reveals strong host-guest
interactions. Such a computational approach has never been re-
ported so far for POM@MOF materials. The above results show
that the unique activity of this POM@MOF photosystem benefits
from two main factors: i) immobilization of the porphyrin as a
ligand in the MOF increases its oxidizing power and ii) the con-
finement of POMs inside the pores of the MOF plays a key role in
the stabilization of the cobalt POM’s catalytic site while the
POM-MOF interface provides key components (-OH, labile water
molecules) relevant to the OER mechanism. Furthermore this new
POM@MOF photosystem takes full advantage of the already
known stabilization of the porphyrin excited state in this kind of
hybrid framework.
26
The POM@MOF composite is stable and
easily reusable. Among others, work is in progress in order to
study the influence of the metalation of the porphyrin on the
catalytic activity.
32
The screening of various POMs as guest spe-
cies is also under study. Overall, this work opens the way to a
whole family of porphyrin-based MOFs as light-sensitive hosts
for the elaboration of noble metal-free heterogeneous and recy-
clable photosystems.
EXPERIMENTAL SECTION
Na
10
[(PW
9
O
34
)
2
Co
4
(H
2
O)
2
] (Na
10
P
2
W
18
Co
4
)
28
and tetrakis(4-
carboxyphenyl)porphyrin TCPP-H
2
29
were synthesized according
to reported procedures. All of the other reagents were purchased
from commercial sources and used as received.
Synthesis of MOF-545. MOF-545 was synthesized according
to a slightly modified procedure.
30
ZrOCl
2
•8H
2
O (325 mg, 1.0
mmol) and TCPP-H
2
(65 mg, 0.086 mmol) were dissolved in 80
mL of DMF and 2.5 mL of dichloroacetic acid in a 100 mL
round-bottomed flask. All reactants were stirred briefly before
heating. The mixture was heated to 130 °C for 15 h, and allowed
to cool down to room temperature. The solid was recovered by
centrifugation, washed with DMF, and acetone. The resulting
powder was dispersed in 25 mL of DMF and 2.5 mL of 1 M HCl
and refluxed for 2 h. After centrifugation, the solid was washed
with DMF and acetone and then soaked into acetone overnight.
The powder was washed with acetone and diethyl ether and dried
overnight in a 90°C oven. 85 mg of a purple powder were collect-
ed (yield 83 % based on TCPP-H
2
). EDS analysis have shown the
presence of chlorine with a Zr/Cl ratio equal to ~4 therefore the
formula of the MOF was assumed to be
[Zr
6
O
16
H
17.5
][C
48
H
26
N
4
O
8
]
2
Cl
1.5
•14H
2
O. Anal. Calc. (found)
(2700 g mol
-1
): C 42.71 (42.05), H 3.64 (2.76), N 4.15 (3.99).
Synthesis of P
2
W
18
Co
4
@MOF-545. 100 mg of MOF-545
(3.7×10
-5
mol) was dispersed in 20 mL of a 5 mM solution of
Na
10
P
2
W
18
Co
4
. The suspension was stirred for 6 h. POM@MOF-
{TCPP-MOF}
hν
{TCPP-MOF} *
{TCPP-MOF}
+
S
2
O
8
2-
SO
4
2-
+ SO
4
- •
POM-red
POM-ox
H
2
O
O
2
