Alternative ball-milling synthesis of vanadium-substituted polyoxometalates as catalysts for the aerobic cleavage of C-C and C-O bonds

Article abstract of DOI:10.1039/d1dt01585k

Vanadium-substituted phosphomolybdic acids (H3+x[PMo12-xVxO40], denoted as Vx) are well-known oxidation catalysts that are generally prepared by the hydrothermal treatment of MoO3 and V2O5 in the presence of H3PO4. This synthesis procedure is highly energy consuming and the Vx yields are not always acceptable. In the present work, an alternative hybrid mechanochemical/hydrothermal synthesis of Vx is proposed, comprising the ball-milling of MoO3 and V2O5, followed by a hydrothermal attack. The resulting materials, with 2 ≤ x ≤ 3, obtained from this new route were compared, in terms of yield, energy consumption and catalytic activity, with a reference V3 sample prepared through a conventional hydrothermal treatment. The ball-milling step proved to lead not only to a shorter and far more energy-saving synthesis procedure, but also to high yields of Vx. Moreover, Vx from this alternative route proved to be generally more active than the conventionally prepared V3 in the aerobic oxidative cleavage of C-O and C-C bonds in 2-phenoxyacetophenone, used herein as a lignin model compound.

Full text of DOI:10.1039/d1dt01585k

Dalton
Transactions
View Article Online
View Journal
PAPER
Alternative ball-milling synthesis of vanadium-
substituted polyoxometalates as catalysts for the
aerobic cleavage of C–C and C–O bonds†
Cite this: DOI: 10.1039/d1dt01585k
Louay Al-Hussaini,a,b Sabine Valange,c Maria Elena Gálvez *b and
a
Franck Launay
*
Vanadium-substituted phosphomolybdic acids (H3+x[PMo12−xVxO40], denoted as Vx) are well-known oxi-
dation catalysts that are generally prepared by the hydrothermal treatment of MoO3 and V2O5 in the pres-
ence of H3PO4. This synthesis procedure is highly energy consuming and the Vx yields are not always
acceptable. In the present work, an alternative hybrid mechanochemical/hydrothermal synthesis of Vx is
proposed, comprising the ball-milling of MoO3 and V2O5, followed by a hydrothermal attack. The result-
ing materials, with 2 ≤ x ≤ 3, obtained from this new route were compared, in terms of yield, energy con-
sumption and catalytic activity, with a reference V3 sample prepared through a conventional hydrothermal
treatment. The ball-milling step proved to lead not only to a shorter and far more energy-saving synthesis
procedure, but also to high yields of Vx. Moreover, Vx from this alternative route proved to be generally
more active than the conventionally prepared V3 in the aerobic oxidative cleavage of C–O and C–C
bonds in 2-phenoxyacetophenone, used herein as a lignin model compound.
Received 16th May 2021,
Accepted 30th July 2021
DOI: 10.1039/d1dt01585k
rsc.li/dalton
acids, as well as diethyl ether, a flammable, harmful and envir-
onmentally hazardous solvent. The “oxo-peroxo” route, invol-
Introduction
Vanadium-substituted
phosphomolybdic
acids, ving the attack of the starting oxides by hydrogen peroxide, is
H3+x[PMo12−xVxO40] – denoted as Vx – have been widely used less developed, since it generally leads to the formation of
as catalysts in different aerobic oxidation reactions.1 Among rather unstable vanadium complexes. On the other hand, the
them, the cleavage of C–C and C–O bonds2 has several interest- hydrothermal route has been often preferred, because it is a
ing applications, such as in the aerobic delignification of relatively simple procedure that consists of the acidic attack of
wood pulp3 or lignin depolymerisation4 leading to the pro- the starting oxides in refluxing water.7,8 However, this route
duction of aromatics from biomass.5 Usually, Vx catalysts are still requires long reaction times and works under very diluted
synthesized either from their corresponding metal salts conditions, especially when high vanadium contents are tar-
through the “etherate” route,6 or from their metal oxides, fol- geted. The rate-limiting step seems to be the acidic attack of
lowing a hydrothermal7,8 or an “oxo-peroxo”9 route (Fig. 1). the starting MoO3 and V2O5 oxides.
Though widely used, the “etherate” pathway involves strong
Ball-milling has attracted growing interest in recent years10
for applications such as solvent-free reactions,11–15 and the
activation of oxides,16,17 carbonates18 and dioxygen,19 as well
aSorbonne Université, CNRS, UMR 7197, Laboratoire de Réactivité de Surface (LRS),
F-75005 Paris, France. E-mail: franck.launay@sorbonne-universite.fr
bSorbonne Université, CNRS, UMR 7190, Institut Jean le Rond d’Alembert, F-75005
Paris, France. E-mail: elena.galvez_parruca@sorbonne-universite.fr
cInstitut de Chimie des Milieux et Matériaux de Poitiers (IC2MP), Université de
Poitiers, CNRS, ENSI Poitiers, B1, 1 rue Marcel Doré, F-86073 Poitiers Cedex 9,
France
†Electronic supplementary information (ESI) available: Weights engaged and
elemental analyses of the different Vx prepared; Comparison of the XRD profiles
of V3-HT with a JCPDS reference and Rietveld refinement results; the TGA pro-
files of all Vx catalysts; the proportion of the different isomers in V2 measured by
31P NMR, XRD and 31P NMR analysis of V2-BM20 materials; details on energy
consumption; the XRD profiles of all the precursors of V2-BM50 and V2-BM20; the
estimation of the crystallite size; the synthetic procedure of K; an example of
HPLC profile; calibration curves. See DOI: 10.1039/d1dt01585k
Fig. 1 Main conventional procedures for the synthesis of vanadium-
substituted phosphomolybdic acids (Vx).
This journal is © The Royal Society of Chemistry 2021
Dalton Trans.

View Article Online
Paper
Dalton Transactions
as for the synthesis of a variety of composites,20 organometal- polyoxometalates produced. We further assessed their catalytic
lics,21 MOFs22 and metallic compounds,17,23 i.e. amorphous activity and selectivity in the aerobic cleavage of a 2-phenoxyace-
and nano-structured materials22,24,25 for catalytic22 and energy- tophenone (K), used as a lignin model compound. Special
related applications.24–26 In the field of catalysis, Zazhigalov emphasis was also put on the determination of the energy con-
et al. reported that a mixed oxide of vanadium and molyb- sumption of both processes. Vx with x = 2 or 3 were targeted,
denum (MoV2O8) synthesized by milling MoO3 and V2O5 since with a higher vanadium content the PMoVx compounds
single oxides was a much more active catalyst for benzene oxi- are known to be more unstable especially at high temperatures.1
dation into maleic anhydride than the parent oxides.27 To the In the next step, the physicochemical features of the oxide mix-
best of our knowledge, there is only one example of Vx syn- tures obtained upon ball-milling of the starting V2O5 and MoO3
thesis using a ball-milling step reported in the literature.28 oxides were also investigated, with the aim of evaluating the
Molchanov et al. showed that Vx may be prepared through the relevance of the mechanochemical step in the synthesis
hydrothermal attack of a ball milled mixture of V2O5 and process. Gaining further knowledge on this hybrid mechano-
MoO3 with H3PO4 at 80 °C.28 The authors studied the influ- chemical method can provide valuable information about the
ence of the duration of the milling step for x = 4, the pro- feasibility of this route for the synthesis of Vx materials.
portions of the starting oxides and the H3PO4 attack con-
ditions. It has also to be noted that the authors also tried to
carry out the whole synthesis of V3 in a miller using a two-step
procedure (first MoO3 and V2O5, followed by the addition of
H3PO4). The synthesized Vx compounds were characterized in
Experimental
Material synthesis
the solid state by FT-IR, 31P and 51V MAS NMR, but no infor-
mation was reported on their speciation in solution. Moreover,
the Vx materials synthesized through this route were never
tested as catalysts, not even characterized in view of their appli-
cation in catalysis. Other recent studies on polyoxometalates
and mechanochemistry have also been reported in the
literature29,30 by Xu et al.29 and Leng et al.,30 but the synthesis
of polyoxometalates was carried out using conventional pro-
cedures and not by ball-milling.
The literature survey reveals that the hybrid synthesis pro-
cedure (ball-milling and subsequent hydrothermal treatment)
has scarcely been explored for the preparation of
V-polyoxometalates. Such a hybrid pathway can however rep-
resent a more sustainable approach for Vx synthesis than the
100% hydrothermal route, requiring less energy and no
additional reactant/solvent. In view of this, the present work
aims at studying various hybrid synthesis procedures of Vx
involving a mechanochemical step (BM–HT) and comparing
them with the 100% hydrothermal route (HT), with a focus on
the molecular structure and yield of the vanadium-substituted
A reference V3 material, denoted as V3-HT, was synthesized4
through the attack of MoO3 (Acros Organics) and V2O5 (Sigma-
Aldrich) by H3PO4 (purity 85%, Carlo Erba) in refluxed water,
following a procedure adapted from that described by Kern
et al.7 In this synthesis, 0.39 g (3.3 mmol) of 85% H3PO4 was
dissolved in 200 mL of deionized water, followed by the
addition of 3.93 g of MoO3 (27.3 mmol) and 0.92 g (5.1 mmol)
of V2O5 (Tables 1 and S1†). Compared to the procedure
reported earlier by Kern and co-workers and dedicated to the
synthesis of V1-HT only,7 a longer hydrothermal attack (6.5 h
instead of 3 h) and more diluted conditions have been used in
our work due to the higher value of x. After cooling, the
residual oxides were removed by filtration and water was then
evaporated from the filtrate, yielding an orange powder.
In the hybrid BM–HT route, the ball milling step was
carried out in a Fritsch Pulverisette 7 planetary miller ( jar dia-
meter: 140 mm) using zirconium oxide balls (∅ = 10 mm,
weight: 68 g) placed inside a zirconium oxide coated autoclave.
The rotation speed was set to 700 rpm and the ZrO2 ball/oxide
weight ratio, denoted as r, was either 20 or 50. For r = 50, the
Table 1 Summary of the synthesized Vx materials, conditions with the values of xexp and the yieldsa
e
Synthesis procedure
HT
Vx
rb
tBM c (h)
tHT d (h)
T (°C)
xexp
Yieldf (%)
79
V3-HTd
—
0
6.5
100
2.7
BM50
V3-BM50-1BM-1.5HT
V2-BM50-1BM-1.5HT
50
20
1
1.5
80
2.8
1.9
97
95
BM20
V2-BM20′-1BM-1.5HT
V2-BM20-1BM-3HT
V2-BM20-2BM-1.5HT
V2-BM20-2BM-3HT
V2-BM20-4BM-1.5HT
V2-BM20-4BM-3HT
1
1
2
2
4
4
1.5
3
1.5
3
1.5
3
100
80
80
2.1
2.3
2.0
2.3
2.3
2.2
96
97
96
95
96
98
a The weights of V2O5, MoO3 and H3PO4 85 wt% employed for each synthesis are detailed in Table S1† and the elemental analyses of the resulting
c
materials. b Ball/oxide weight ratio (r) with zirconium oxide balls (∅ = 10 mm, weight: 68 g). The rotation speed was set to 700 rpm. tBM = ball
milling duration. d tHT = hydrothermal treatment duration. e Determined from ICP, relative error = 10%. f Calculated from 31P NMR ( 10%).
Dalton Trans.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
milling duration (tBM) was fixed to 1 h and the ball milled Catalytic experiments and monitoring
mixture with nominal theoretical x values (xth) equal to 2 or 3
In the aerobic cleavage catalytic experiments, 2-phenoxyaceto-
phenone (0.32 g, 1.5 mmol) and Vx (Mo + V, 18 mol%) were
dissolved in 15 mL of MeCN/AcOH 90–10 in a Schlenk tube
connected to a gas burette system to monitor the dioxygen
consumption. The Schlenk tube was purged and then heated
during 24 h at 82 °C. After cooling the reaction mixture, an
aliquot of 0.5 mL was withdrawn and dropped in a 20 mL volu-
metric flask and diluted with H2O–MeOH–AcOH 49 : 49 : 2 for
was subsequently attacked by phosphoric acid in water at
80 °C during tHT = 1.5 h, affording the series of Vx-BM50-1BM
-
1.5HT materials (see Tables 1 and S1†). Another series of
materials (V2-BM20-tBM-tHT) were prepared for a nominal value
of xth equal to 2 and with r = 20 (weight of balls kept constant)
but tBM was varied from 1 to 4 h and the ball-milled mixture
was attacked during either tHT = 1.5 h or 3 h at 80 °C (unless
otherwise specified) (see Tables 1 and S1†).
HPLC analysis (Shimadzu LC-20AD) using
a Shim-pack
A preliminary one-pot 100% ball-milling procedure where
MoO3, V2O5 and H3PO4 were introduced together was
attempted (1 h, 700 rpm) for both ball/oxide weight ratios (see
Table S1†). The synthesis was successful for r = 50. However,
the zirconia balls were severely damaged, thereby resulting in
a high contamination of the final product with ZrO2 as shown
by XRD. Such eroded ZrO2 had to be eliminated by centrifu-
gation in water (7500 rpm, 10 min).
Generally, the mixture of the starting oxides, initially clear
brown, turned dark brown after milling. Upon the addition of
the solid to the aqueous solution containing a stoichiometric
amount of H3PO4, the mixture turned red on the first few
seconds of the acid attack. At the end, the suspension was fil-
tered to eliminate the residual traces of zirconium oxide orig-
inating from the abrasion of the balls, and then water was
evaporated to get the Vx solids.
XR-ODS column (4.6 × 100 mm, 2.2 µm) heated to 40 °C in a
CT0-10AS oven. The mobile phase was composed of aqueous
AcOH (0.5 vol%) and methanol (HPLC grade, VWR Chemicals)
used at a flow rate of 0.4 mL min−1 (0–13 min: 60–40 (MeOH),
13–17 min: gradient from 40 to 60% of methanol, 17–33 min:
40–60 (MeOH), 33–37 min: gradient from 60 to 40% of metha-
nol and 37–50 min: 60–40 (MeOH)). The compounds were
detected by using an SPD-M20A UV detector (210 and 220 nm).
Benzaldehyde, benzoic acid and phenol were quantified
according to the calibration curves (ESI†).
Results and discussion
Influence of the synthesis conditions on the Vx materials
The first series of Vx compounds were synthesized using the
hybrid ball milling-hydrothermal procedure (BM–HT) with a
ZrO2 ball/oxide weight ratio r = 50, a milling duration of 1 h
and an acid attack duration of the ball-milled mixture of 1.5 h.
Physicochemical characterization
X-ray powder diffraction analyses (XRD) were performed on a
Bruker Advance D8 diffractometer, using a Cu Kα radiation source
(λ = 1.5406 Å) without a monochromator. The diffraction patterns
Fig. 2a shows the powder XRD patterns of V2-BM50-1BM
-
1.5HT and V3-BM50-1BM-1.5HT together with the pattern
,
were collected from 5° to 50°, at a scanning rate of 0.34° min−1
.
acquired for conventionally synthesized V3-HT. These patterns
were compared to the JCPDS 00-043-0317 reference for
The crystalline phases were identified by Rietveld analysis of the
diffraction patterns (factors RP and Chi2), using the Fullprof soft-
ware. Liquid 31P nuclear magnetic resonance (NMR) analyses were
performed on a 400 MHz Bruker apparatus. 30 mg of the solid was
dissolved in 250 µL of D2O (Euriso-top) and 250 µL of H2O. Then
7.5 µL of dioxane (SDS, Carlo-Erba) was added to the solution. For
each analysis, 16 scans were recorded with a relaxation delay of 32
s. The P, Mo and V contents for all samples were determined by
inductive coupled plasma analyses (ICP) from Crealins (Lyon,
Villeurbanne France) on an ICP Thermo-Fischer iCAM.
Thermogravimetric analyses (TGA) were performed on a Cp
SDTQ600 system, in order to determine the hydration index of Vx,
denoted as n. The temperature was increased up to 600 °C at a
heating rate of 10 °C min−1 under an air flow of 100 mL min−1
.
Scanning electron microscopy analyses (SEM) were carried out
using a Hitachi SU-70 microscope. The sample was introduced at
3.2 mm from the electron beam source, the accelerating voltage
was set to 1 kV and the intensity of the emission current was set to
47 µA. Particle size (granulometry) measurements were performed
on a Fristch Analysette 22 Nano Tec. Raman spectroscopy analyses
were performed using a RamanRXN spectrometer equipped with a
laser working at λ = 785 nm, with a power of 10–12 mW and a
resolution of 4 cm−1. The duration of each scan was 10 s and the
number of scans was set to 30.
Fig. 2 (a) Powder-XRD patterns and (b) liquid 31P NMR spectra
(D2O/H2O/dioxane, pH = 1) of Vx-BM50-1BM-1.5HT and V3-HT materials.
This journal is © The Royal Society of Chemistry 2021
Dalton Trans.

View Article Online
Paper
Dalton Transactions
H3PMo12O40·13H2O (Fig. S1 in the ESI†). Both Vx-BM50-1BM
-
possible impurity, a synthesis yield was calculated, i.e. defined
1.5HT materials showed a typical Keggin structure, as evi- as the extent of H3PO4 incorporation into the Vx structures (cal-
denced by the presence of diffraction peaks at ca. 2Θ = 8°, at culated from the areas of the H3PO4 peaks in the 31P NMR
15–18° and at ca. 28° in all XRD patterns. A closer look at the spectra, see details in the ESI†). The values of this synthesis
diffractograms indicated that the crystalline structure was yield were much higher for V3-BM50-1BM-1.5HT (97%) than
more impacted by the hydration index n than by the vanadium those for V3-HT (79%), pointing to a more effective incorpor-
content.31–33 In V3-BM50-1BM-1.5HT
, one single crystalline ation of vanadium into H3+x[PMo12−xVxO40] for the hybrid ball-
phase could be identified, similarly to V3-HT, but yielding milled/hydrothermal material in comparison with the conven-
slightly different Rietveld refinement parameters (RP = 10 and tionally synthesized one (Table 1).
Chi2 = 2.4 for V3-BM50-1BM-1.5HT, RP = 15 and Chi2 = 13 for V3-
HT, see Table S2 in the ESI†). On the other hand, V2-BM50-1BM
For the lowest V-content (xth = 2), the 31P NMR spectrum
obtained was relatively more simple, allowing us to get valu-
-
1.5HT appeared to be constituted of at least two Vx crystalline able information about the different isomers and Vx species
phases differing in their hydration index, n. The main phase present in V2-BM50 (Table S4, see the ESI† for further details
corresponds to n = 13, also present in V3-HT, whereas the sec- about these calculations). According to Pettersson,34 six types
ondary phase corresponds to n = 30 (Fig. 2a, JCPDS 00-043- of isomers could be identified in the 31P NMR spectrum of V2
0316).
polyoxometalates. The individual contributions of α-1,4
The synthesis of vanadium-substituted phosphomolybic (−4.21 ppm, V atoms on the same triad), α-1,2 and α-1,5
acids generally leads to mixtures of different Vx stoichi- (−4.03 ppm, vicinal V atoms on different triads), α-1,6 and
ometries, with x representing the average number of vanadium α-1,11 (−3.94 ppm, no vicinity of V atoms) as well as β isomers
equivalents. A non-total incorporation of vanadium into the (−3.77 ppm, most probably not vicinal β-4,10 and β-4,11
H3+x[PMo12−xVxO40] structure results in lower values of x, vis-à- isomers,34 Fig. S3 in the ESI†) with respect to the total amount
vis the nominal V-loading (xth), together with the presence of of V2 obtained are presented in Table S4.† In V2-BM50-1BM
-
water-insoluble V2O5 residues at the end of the synthesis. ICP 1.5HT, the α-1,4 isomer was found to be present to a much
analyses allowed us to determine the experimental values of x greater extent than the other five possible isomers, i.e. α-1,4
in this series of materials. The theoretical (xth) and experi- amounts to 58.5% vs. 12.6% for α-1,2 and α-1,5, 16.3% for
mental (xexp) Vx stoichiometries, as well as the elemental com- α-1,6 and α-1,11, and 12.7% for the β isomers. In fact, 71% of
position in wt% of the polyoxometalates synthesized, are given all the V2 isomers in V2-BM50-1BM-1.5HT were shown to bear
in Table 1 and Table S1 in ESI†. For the Vx-BM50-1BM-1.5HT vicinal V atoms (α-1,4, α-1,2 and α-1,5, Table S4†). This infor-
materials, the x values obtained are very close to the theore- mation can be highly relevant for the application of these
tical values (less than 10% difference which can be related to materials in catalysis, since previous studies evidenced that
the experimental error). In contrast, for the conventionally syn- their activity in oxidation reactions can be directly linked to
thesized V3-HT material, the difference between xexp and xth is the presence of isomers bearing vicinal V atoms.36
more pronounced (>10%). Indeed, during the conventional
The results presented so far prove that V3 can be success-
hydrothermal synthesis of V3-HT, a solid residue enriched in fully obtained a with higher yield and with a much shorter syn-
V2O5 had to be filtered before water evaporation, resulting una- thesis time than conventionally prepared V3-HT, when using r
voidably in a lower value of xexp. These observations point to = 50 during the ball-milling step. Lowering the ball to oxide
an improvement of the hydrothermal attack of the oxides weight ratio will produce higher amounts of materials per syn-
favored by the ball-milling step.
Whatever the synthesis route, the TGA weight-losses for Vx using r = 20 instead of 50, for xth = 2, keeping the ball-
hydration (T1 = 215–222 °C) and constitutive water (T2 milling and the hydrothermal attack (80 °C) duration constant.
thesis batch. Therefore, we further considered the synthesis of
=
416–421 °C) took place at relatively similar temperatures for The xth value of 2 was chosen due to the superior stability of
V3-BM50-1BM-1.5HT and V3-HT (Fig. S2 and Table S3 in ESI†), V2 in solution, vis-à-vis V3, and for their easier characterization,
showing that their vanadium contents are quite comparable. especially by 31P NMR. However, the resulting material,
The hydration indexes (n) calculated from the TGA plots range denoted as V2-BM20-1BM-1.5HT, did not exhibit the desired
between 12 and 15, and are in agreement with the results Keggin structure and an important H3PO4 signal was observed
derived from the XRD analysis (Fig. 2a). Further insights into in its 31P NMR spectrum (see Fig. S4 in the ESI†). Evidently,
the purity and composition of the synthesized Vx were gained the 1 h milling time (tBM) and the subsequent 1.5 h acid attack
through liquid analyses. The 31P NMR spectra acquired for V2- at 80 °C did not lead to a successful synthesis of V2. Previous
BM50-1BM-1.5HT, V3-BM50-1BM-1.5HT and V3-HT are presented in studies showed that an increased amount of powder in the
Fig. 2b. These spectra are rather complex due to the presence ball-milling jar resulted in a thicker powder layer between the
of various isomers of V3 in solution, the partial protonation of colliding entities,37,38 thereby leading to a dramatic decrease
V3 at pH ∼ 1 and the co-existence in solution of different Vx of the local temperature during milling,39,40 as a consequence
compounds (such as V2) as previously shown in the literature of hindered heat transfer and more elastic shocks.39–44 These
by 31P NMR.34,35 First, based on the very similar values of xexp experimental observations were even predicted mathemat-
and xth obtained for V2-BM50-1BM-1.5HT and V3-BM50-1BM
-
ically, using the energy equation defined by Magini et al.42 The
1.5HT, respectively, and considering that H3PO4 is the only duration of the ball-milling step (tBM) was therefore increased
Dalton Trans.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
Table 2 Energy saving (in %) for different Vx synthesized including the
ball-milling step vs. the one prepared through conventional hydro-
thermal synthesis
and the influence of both the duration (tHT) and the tempera-
ture (THT) of the H3PO4 hydrothermal attack was further inves-
tigated. These experiments led to the formation of the second
series of materials denoted as V2-BM20-tBM-tHT. First, the temp-
Vx
Energy saving (%)
erature of the acid attack was set to 100 °C (V2-BM20′-1BM
-
a
V3-HT
—
1.5HT), keeping the milling and attack durations unchanged.
Then, while performing the H3PO4 attack at 80 °C, its duration
V3-BM50-1BM-1.5HT
30
31
was extended up to 3 h without modifying the duration of the V2-BM50-1BM-1.5HT
milling step (1 h), affording the V2-BM20-1BM-3HT compound. In
both cases, the recovered Vx materials showed the desired Keggin
V2-BM20-1BM-3HT
structure, as evidenced by their corresponding XRD profiles V2-BM20-2BM-1.5HT
V2-BM20′-1BM-1.5HT
68
62
54
43
19
10
V2-BM20-2BM-3HT
V2-BM20-4BM-1.5HT
V2-BM20-4BM-3HT
(Fig. S5 in the ESI†). Moreover, the absence of the signal of H3PO4
in the 31P NMR spectra of V2-BM20′-1BM-1.5HT and V2-BM20-1BM
-
3HT proved that both syntheses were quantitative (Fig. S6 in the
ESI†). Longer milling steps (tBM = 2 or 4 h) were also assessed,
a Energy consumption: 627 kW h mol−1. Details in Tables S5 and S6
(ESI†).
affording the V2-BM20-2BM-1.5HT, V2-BM20-2BM-3HT, V2-BM20-4BM
-
1.5HT and V2-BM20-4BM-3HT materials. Very similar results in terms
of structure and V2 yields were obtained. For this V2-BM20 series,
the xexp values determined from the ICP analyses and xth values
were at the best different from 1% (V2-BM20-2BM-1.5HT) and at
milled mixtures, V2-BM50-1mix (tBM = 1 h, first series) and V2-
BM20-tBMmix (tBM = 1, 2 or 4 h, second series) were recovered
after the milling step (prior to the hydrothermal acid attack).
They were subsequently characterized by XRD (Fig. 3a and
Fig. S8, ESI†), Raman spectroscopy (Fig. 3b), SEM (Fig. 4a and
b) and laser granulometry (Fig. 5).
worst 13% (V2-BM20-2BM-3HT, V2-BM20-4BM-1.5HT, and V2-BM20-1BM
-
3HT, see Table 1). TGA-DSC analyses of the V2-BM20 materials
(Fig. S7 in the ESI†) evidenced that the loss of constitutive water
occurred almost at the same temperature (ca. T2 = 428 °C, see
Table S3†), corresponding to the hydration indexes, n, between 12
and 15. For the V2-BM20 series, the yields ranged from 92 to 97%
with the predominance of the α-1,4 isomer (37–48%), while the
contribution of the isomers containing vicinal vanadium atoms
remained quite high, between 61 and 65% (see Table S4†).
The XRD patterns acquired for the ball-milled mixtures V2-
BM50-1mix and V2-BM20-tBMmix are shown in Fig. 3a and
Energy consumption was further determined for both the
hybrid synthesis route (BM–HT) and the 100% hydrothermal
procedure (HT), using the methodology proposed by Hao and
co-workers.45 Briefly, considering that the main energy inputs
are the ball-milling step itself and the heating process during
the acid attack, the global energy consumption Et (in kW h)
can be calculated through the simple relation Et = PBM BM
t
+
P
HTtHT, where PBM and PHT stand, respectively, for the power
consumed during the milling step and during the acid attack
(see Tables S5, 6 and the ESI† for further details). Based on the
results obtained, the energy saving (in %) was calculated for
each material, in comparison with the conventional hydro-
thermal route, i.e. V3-HT (Table 2).
These values prove that the energy consumption during the
hybrid synthesis is substantially reduced when compared to
that involved in the conventional hydrothermal pathway for V3-
HT (see Table S5 in the ESI†). Energy savings up to almost
70% can be reached with the hybrid pathway (i.e. for V2-BM20′-
1
BM-1.5HT synthesis). This fact and the overall short synthesis
time indicate that the hybrid milling-hydrothermal route can
be a very prospective pathway for the preparation of Vx, with a
high yield and purity.
Importance of the mechanochemical step: characterization of
the ball-milled oxides
Fig. 3 (a) Powder-XRD patterns and (b) Raman spectra of the starting
In order to gain deeper insight into the positive influence of
oxides (V2O5 and MoO3), the hand-milled mixture V2
and the
hand mix
the mechanochemical step in the overall procedure, the ball- ball-milled material V2-BM20-2mix
.
This journal is © The Royal Society of Chemistry 2021
Dalton Trans.

View Article Online
Paper
Dalton Transactions
the resulting mixtures. Moreover, the XRD reflections observed
for the V2-BM50-1mix and V2-BM20-tBMmix ball-milled materials
appear to be shifted to lower 2θ values, in comparison with
those of pure MoO3, pointing to the formation of a mixed-
oxide phase. Molchanov et al.28 suggested the formation of a
mixed oxide upon ball-milling of the single starting V2O5 and
MoO3 oxides, although any experimental evidence supported
their hypothesis.
Fig. 3b displays the Raman spectrum acquired for V2-BM20-
mix, in comparison with the spectra registered for the starting
Fig. 4 SEM images (magnification ×5k) for (a) the hand-milled mixture,
2
V2 hand mix; (b) the ball-milled mixture, V2-BM20-2mix
.
oxides and for their hand-milled mixture. The Raman bands
for pure MoO3 46,47 and pure V2O5 48,49 can be seen in the spec-
trum of V2-BM20-2mix. However, these bands appear broader,
probably as a consequence of some structural disorder caused
upon ball-milling.50 In addition, the Raman spectrum of V2-
BM20-2mix shows several bands appearing between 200 and
600 cm−1, which cannot be observed either for the hand-
milled mixture or in the spectra of MoO3 and V2O5. The pres-
ence of these bands further suggests the formation of a mixed
oxide phase.27 However, no significant shift in the position of
the Raman bands corresponding to MoO3 was observed,
meaning that, in spite of the partial substitution of Mo by V
within the crystalline network, the original structure of this
oxide is well preserved upon ball-milling.
The representative SEM micrographs acquired for the hand-
milled mixture V2 hand mix and for the ball-milled mixture V2-
BM20-2mix are respectively presented in Fig. 4. The SEM image
for the hand-mixed solid, Fig. 4a, evidences the presence of
smooth particles having sizes ca. 2 × 0.6 µm, with quite
regular elongated shapes that stuck together forming a rela-
tively compact surface morphology.
Fig. 5 Laser granulometry histograms for the starting oxides (V2O5 and
MoO3), the hand-milled mixture V2 hand mix, and the ball-milled mixtures
V2-BM20-1mix and V2-BM20-4mix
.
Granular but more irregular structures are present in the
SEM micrograph acquired for the ball-milled mixture V2-BM20-
2
mix, Fig. 4b, pointing to some sintering occurring during the
Fig. S8.† These diffractogramms are presented together with ball-milling step. The particles appear to be more agglomer-
the patterns of the starting V2O5 and MoO3 oxides, as well as ated than those in the hand-milled mixture, and bigger par-
for a hand-milled mixture prepared with the same Mo/V ratio ticles (sizes up to 5 µm) can be easily observed. Sintering
(V2 hand mix). Both V2O5 (JCPDS 9-387, space group Pmmn) and seemed to take place to a higher extent upon ball-milling at r =
MoO3 (JCPDS 05-0508, space group Pnma) show an ortho- 50 (not shown). As previously commented, lower r ratios may
rhombic crystalline structure. Well-defined diffraction lines result in lower local temperatures attained during milling,42,43
are observed for the hand-milled mixture V2
mix. As disfavoring sintering and particle agglomeration.
hand
expected, its XRD pattern is the result of the combination of
In a further step, the grain size distribution in the different
the diffractograms of the parent MoO3 and V2O5 oxides at a 1/ materials was assessed by laser granulometry (histograms
8 V2O5/MoO3 weight ratio (the reflections corresponding to shown in Fig. 5). The grain size distribution observed for the
V2O5 are scarcely perceptible). In contrast, the patterns hand-milled material looks very similar to that of MoO3, point-
acquired for the ball-milled oxides are characterized by the ing to very similar grain sizes. In the case of the ball-milled
presence of wide diffraction signals and a less-well defined oxides, V2-BM20-1mix and V2-BM20-4mix, the grain size distri-
baseline. Crystallite sizes of 2.8 Å for V2-BM50-1mix and around bution considerably changes with respect to the starting
1.4–1.7 Å for V2-BM20-tBMmix were calculated from the widths at oxides. A bimodal grain size distribution is observed, com-
half peak height of the (110) diffraction (the Scherrer equation, posed of a main peak centered around 20 µm and a second
shape factor K = 0.888, see details in the ESI†). These crystallite population of smaller particle sizes around 2–5 µm. The grain
sizes are considerably smaller than those calculated for the distribution does not seem to change for longer ball-milling
hand-milled mixture of MoO3 and V2O5 (Table S7, ESI†). This times, i.e. similar curves were obtained for V2-BM20-1mix and
decrease in the crystallite size becomes more pronounced for for V2-BM20-4mix. The ball-milling step alters not only the crys-
the solids ball-milled at r = 20, whereas the duration of the tallinity of the solid but also its grain size, in complete agree-
milling step did not seem to affect the crystalline structure of ment with the SEM observation of these materials. Sintering
Dalton Trans.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
occurs as a consequence of an increase of temperature in the
The cleavage experiments were carried out in MeCN/AcOH,
ball-mill jar, yielding a solid consisting of particle aggregates at 82 °C under an atmospheric pressure of O2 during 24 h
with bigger sizes than those of the starting materials (mainly (optimal reaction conditions, identified in our previous
MoO3).
work4). The catalyst loading was fixed at 18 mol% with respect
Therefore, the XRD and Raman spectra are consistent with to Mo and V, according to the available literature on aerobic
the formation of a Mo–V mixed oxide phase, probably due to delignification involving H3PMo12O40 as a catalyst.55 Under
the insertion of V cations into the original MoO3 lattice. This such reaction conditions the oxidation of K affords mainly
does not lead to any dramatic change in the crystalline struc- phenol (PhOH) from C–O cleavage, benzaldehyde (PhCHO)
ture. However, higher crystallinity and smaller crystallite sizes and benzoic acid (PhCOOH) from the C–C cleavage (see Fig. 7,
were observed in the ball-milled solids, which may be a conse- as well as the HPLC analyses presented in Fig. S9, ESI†).
quence of some sintering occurring during the ball-milling
Table 3 shows the conversion, product yield and carbon
step. In spite of that, the surface of the ball-milled materials is balance (CB) values determined during the aerobic cleavage of
rougher than that of MoO3 or the hand-milled oxides (SEM K in the presence of the different synthesized Vx materials.
analyses), which looks to be highly beneficial for the further
For V contents corresponding to x = 3, the material V3-
acid attack with H3PO4, affording an easier synthesis of the BM50-1BM-1.5HT (entry 2) evidenced catalytic performance
V-substituted phosphomolybdic acids.
similar to that of conventionally synthesized V3-HT (entry 1).
However, a slightly higher K conversion was observed in the
presence of the catalyst obtained through the hybrid ball-
milling/hydrothermal procedure (78% vs. 72% for V3-HT),
together with lower carbon balance values, probably due to
phenol degradation.
Activity and selectivity towards the C–C and C–O cleavage of
the synthesized Vx materials
Simple lignin models have been often used in the screening of
the catalysts and reaction conditions for lignin oxidative depo-
lymerisation.51 Among the very few reported contributions
involving Vx, Evtuguin et al. studied the activity of V5 as the
catalyst for the aerobic cleavage of 1-(3-methoxy-4-hydroxyphe-
nyl)-2-(2-methoxyphenoxy)ethanol (1) and 1-(3,4-dimethoxy-
phenyl)-2-(2-methoxyphenoxy)ethanol (2),52,53 see Fig. 6.
Ketone-type substrates such as benzoin (3), and, more interest-
ingly, dimeric compounds such as 2-phenoxyacetophenone
(K), have been considered as well,4 since benzylic hydroxyl oxi-
dation may occur in situ under aerobic oxidative conditions.54
We have recently proved that V3-HT can be a very efficient
catalyst for the aerobic cleavage of 2-phenoxyacetophenone K4.
Here, we have evaluated the ability of the whole Vx series (1.9 <
xexp < 2.8) prepared through the hybrid ball-milling/hydro-
thermal route to catalyze the aerobic cleavage of C–C and C–O
bonds in K. The activity and selectivity of the PMoVx-BM–HT
materials were compared to those of conventionally syn-
thesized V3-HT, with emphasis on possible correlations with
their respective physicochemical properties.
Within the first series of hybrid BM–HT materials, V2-BM50-
BM-1.5HT (entry 3) led to the highest K conversion (84% vs.
1
78% for V3-BM50-1BM-1.5HT), along with high yields of phenol
(54%), benzaldehyde (19%) and benzoic acid (49%), while
maintaining a correct carbon balance. Even with a lower V
content (x = 2 instead of 3), V2-BM50-1BM-1.5HT evidenced
improved catalytic performance compared to its counterpart
and V3-HT, the conventionally synthesized material which is
rather unusual, since the oxidation potential is expected to
increase with x but it is also known that V3 materials are less
stable in solution than their V2 analogs and are found to be de-
activated earlier.56
For the series of V2-BM20-tBM-tHT materials (entries 4–9),
conversions between 70 and 81% were obtained. As previously
reported,4 the carbon balance values decrease with increasing
conversion, due to a higher extent of oxidation that may some-
times be not very selective. At x = 2, this lack of selectivity
seems to affect mostly the yield of benzaldehyde, as observed
for V2-BM20-4BM-3HT, entry 9 (7% PhCHO yield and a carbon
balance of 67% which is the lowest CB value of this series).
However, no direct relationship can be established between
the synthesis parameters (duration of the milling step and
hydrothermal attack) and the catalytic performance of these
materials. The average values were calculated, indicating that
the six samples of the V2-BM20-tBM-tHT series afforded similar
conversions and yields. For these V2-BM20-tBM-tHT, no evident
correlation can be drawn between the relative presence of the
α-1,4 isomer and their catalytic activity and selectivity.
Fig. 6 Dimeric aromatic molecules studied as substrates in the phos-
phomolybdic acid catalyzed aerobic cleavage reaction.
Fig. 7 Aerobic cleavage of K.
This journal is © The Royal Society of Chemistry 2021
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Table 3 Aerobic cleavage of K in the presence of different synthesized Vx: conversion, product yield and carbon balance. The amounts (in percen-
tage) of the α-1,4 isomer, calculated from the liquid 31P NMR spectra, are also shown for each catalyst
HPLC yield, η (%)
C–C cleavage
Entry
1
Catalyst
V3-HT
α-1,4 Isomer (%)
Conversion (%)
72
PhOH
52
PhCHO
12
PhCOOHa (Σ)
CBb (%)
77
n.d.
42 (54)
2
3
V3-BM50-1BM-1.5HT
V2-BM50-1BM-1.5HT
n.d.
58.5
78
84
38
54
16
19
32 (48)
49 (68)
62
78
4
5
6
7
8
9
V2-BM20′-1BM-1.5HT
V2-BM20-1BM-3HT
V2-BM20-2BM-1.5HT
V2-BM20-2BM-3HT
V2-BM20-4BM-1.5HT
V2-BM20-4BM-3HT
47.9
42.2
49.1
46.0
45.2
37.4
77
81
70
75
79
81
48
42
57
37
52
59
14
17
23
15
12
7
38 (52)
47 (64)
31 (54)
38 (53)
41 (53)
39 (46)
70
69
81
67
70
67
c
Average V2-BM20
45
3
77
3
49
5
15
3
39
3
71
3
K 100 mM, atm. O2, Mo + V 18 mol%, MeCN + 10 vol% AcOH, 82 °C, 24 h. a Σ = ηPhCHO + ηPhCOOH b Calculated as CB = (6ηPhOH + 7(ηPhCHO
. +
ηPhCOOH) + 14(100 − conv.))/14 < 100%, since C1 products from the oxidation of the CH2 group in K were not quantified. c Average values for the
six V2-BM20 materials. Confidence limit = 80%; n.d.: not determined.
However, V2-BM50-1BM-1.5HT for which the highest conversion mation of a highly crystalline Mo–V mixed-oxide in the milling
was observed among all the materials, shows the highest rela- step, exhibiting surface rugosity that facilitated the acid attack.
tive amount of the α-1,4 isomer (58.5%), which seems to lead
The different Vx materials synthesized were used as catalysts
to a somehow enhanced selectivity towards C–C cleavage (Σ = for the aerobic cleavage of C–O and C–C bonds in a dimeric
ηPhCHO + ηPhCOOH amounts to 68%, whereas the C–C global lignin model (K). Generally, the materials prepared through the
yield determined for V2-BM20-4BM-3HT was only 46% for a α-1,4 hybrid mechanochemical/hydrothermal route, with 2 < xexp < 3,
isomer content of 37.4%). In this sense, a shorter ball-milling exhibited slightly improved catalytic performance compared to
step at r = 50 seems to be more beneficial in terms of catalytic the conventionally synthesized V3-HT. Among them, V2-BM50-
activity and C–C– cleavage selectivity than that at r = 20 at
longer milling and attack durations, since the presence of the milling step performed with a ball/oxide weight ratio r = 50 for
α-1,4 isomer seems to be moreover favoured. 1 h and a 1.5 h-acid attack of the ball-milled mixture, afforded
1BM-1.5HT, obtained through the hybrid route comprising a ball-
Let us note here that the spent V3-HT catalyst was pre- higher K conversion than V3-HT, i.e. 84% vs. 72%. Moreover,
viously characterized by 31P NMR.4 In brief, the Keggin struc- this material showed improved selectivity towards C–C cleavage,
ture was shown to be unstable upon reduction by the sub- which was related to an increased amount of the α-1,4 isomer,
strate. As a consequence, the Vx compound dissociates into as determined from the analysis of the liquid 31P NMR spectra.
+
pervanadyl cations VO2 and into other unstable lacunary
Therefore, the hybrid mechanochemical/hydrothermal
species, which are most probably the catalytically active route explored in the present work allowed us to obtain Vx with
species in oxidative cleavage reactions. In the case of V2, the high yields and a short synthesis time, with respect to a conven-
release of pervanadyl cations is favored by the presence of tional hydrothermal procedure. This route can provide up to
vicinal vanadium atoms as in the α-1,4 isomer,57 which may 68% energy saving, yielding materials with very promising appli-
therefore explain their improved catalytic activity, particularly cations in lignin oxidative depolymerization reactions.
towards C–C cleavage.
Author contributions
LAH, MEG and FL have equally contributed to the research
presented herein. All the authors have equally contributed to
Conclusions
A series of vanadium-substituted phosphomolybdic acids, Vx,
were successfully synthesized through a hybrid mechanochem-
ical/hydrothermal route, involving a ball-milling step, followed
by subsequent H3PO4 attack. The resulting solids showed the
typical Keggin structure of Vx materials. Their hydration
indexes were very similar to those observed in the reference
material prepared through a conventional hydrothermal route.
the writing of the manuscript. All the authors have agreed to
its submission for publication and accepted the responsibility
for having properly included all (and only) co-authors.
Conflicts of interest
High Vx yields were obtained, as a consequence of the for- There are no conflicts to declare.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
21 N. R. Rightmire and T. P. Hanusa, Dalton Trans., 2016, 45,
2352–2362.
Acknowledgements
We gratefully acknowledge the Program “Génie des Procédés” 22 S. Leukel, M. Panthöfer, M. Mondeshki, G. Kieslich, Y. Wu,
at Sorbonne Université for the Ph.D fellowship (2016-2019) of
L. Al-Hussaini, as well as Sorbonne Université, the European
N. Krautwurst and W. Tremel, Chem. Mater., 2018, 30(17),
6040–6052.
Union (ERDF) and the CNRS for their financial support. We 23 L. Takacs, Prog. Mater. Sci., 2002, 47, 355–414.
would like to thank Mohamed Selmane, Claire Troufflard, and 24 C. Xu, S. De, A. M. Balu, M. Ojeda and R. Luque, Chem.
Aurélie Bernard, Jean-Marc Krafft and Sandra Casale for XRD,
Commun., 2015, 51, 6698–6713.
NMR, Raman and SEM characterization.
25 H. Lyu, B. Gao, F. He, C. Ding, J. Tang and J. C. Crittenden,
ACS Sustainable Chem. Eng., 2017, 5, 9568–9585.
˜
26 M. J. Munoz-Batista, D. Rodriguez-Padron, A. R. Puente-
Santiago and R. Luque, ACS Sustainable Chem. Eng., 2018,
6, 9530–9544.
Notes and references
1 I. A. Weinstock, R. E. Schreiber and R. Neumann, Chem. 27 V. A. Zazhigalov, S. V. Khalameida, N. S. Litvin,
Rev., 2018, 118, 2680–2717.
2 L. El Aakel, F. Launay, J.-M. Brégeault and A. Atlamsani,
Chem. Commun., 2001, 2218–2219.
3 A. R. Gaspar, J. A. F. Gamelas, D. V. Evtuguin and
C. P. Neto, Green Chem., 2007, 9, 717–730.
I. V. Bacherikova, J. Stoch and L. Depero, Kinet. Catal.,
2008, 49, 692–701.
28 V. V. Molchanov, G. M. Maksimov, R. I. Maksimovskaya,
V. V. Goidin and R. A. Buyanov, Inorg. Mater., 2003, 39,
687–693.
4 L. Al-Hussaini, F. Launay and E. Gálvez, Materials, 2020, 13, 29 L. Xu, Y. Wang, T. Xu, S. Liu, J. Tong, R. Chu, X. Hou and
812–833. B. Liu, ChemCatChem, 2018, 10, 5386–5390.
5 C. Liu, S. Wu, H. Zhang and R. Xiao, Fuel Process. Technol., 30 Y. Leng, Y. Jiang, H. Peng, Z. Zhang, M. Liu, K. Jie,
2019, 191, 181–201.
6 G. A. Tsigdinos and J. Hallada, Inorg. Chem., 1968, 7, 437–
P. Zhang and S. Dai, Catal. Sci. Technol., 2019, 9, 2173–
2179.
441.
31 J. She, Z. Fu, J. Li, B. Zeng, S. Tang, W. Wu, H. Zhao, D. Yin
and S. R. Kirk, Appl. Catal., B, 2016, 182, 392–404.
32 P. Mothé-Esteves, M. Maciel Pereira, J. Arichi and B. Louis,
Cryst. Growth Des., 2010, 10, 371–378.
7 F. Kern, S. Ruf and G. Emig, Appl. Catal., A, 1997, 150, 143–
151.
8 T. Onoda and T. Otake, Pat. US4146574, 1979.
9 V. F. Odyakov and E. G. Zhizhina, Russ. J. Inorg. Chem., 33 D. Barats-Damatov, L. J. W. Shimon, Y. Feldman,
2009, 54, 361–367.
T. Bendikov and R. Neumann, Inorg. Chem., 2015, 54, 628–
10 J. Qu, L. Sha, C. Wu and Q. Zhang, Nanomaterials, 2019, 9,
80–94.
634.
34 L. Pettersson, Mol. Eng., 1993, 3, 29–42.
11 J. G. Hernández and T. Frišić, Tetrahedron Lett., 2015, 56, 35 L. Pettersson, I. Andersson, A. Selling and J. H. Grate,
4253–4265. Inorg. Chem., 1994, 33, 982–993.
12 Z. Wang, Z. Li, M. Ng and P. J. Milner, Dalton Trans., 2020, 36 R. Neumann and A. Khenkin, Chem. Commun., 2006, 2529–
49, 16238–16244. 2538.
13 F. Preishuber-Pflügl and M. Wilkening, Dalton Trans., 2016, 37 D. R. Maurice and T. H. Courtney, Metall. Mater. Trans. A,
45, 8675–8687. 1990, 21, 289–303.
14 N. Pétry, H. Benakki, E. Clot, P. Retailleau, F. Guenoun, 38 S.-Y. Lu, M. Q.-J. Mao, Z. Peng, X.-D. Li and J.-H. Yan, Chin.
F. Asserar, C. Sekkat, T.-X. Métro, J. Martinez and
F. Lamaty, Beilstein J. Org. Chem., 2017, 13, 2169–2178.
15 A. Beillard, X. Bantreil, T.-X. Métro, J. Martinez and
F. Lamaty, Chem. Rev., 2019, 119, 7529–7609.
16 V. Šepelák, S. Bégin-Colin and G. Le Caër, Dalton Trans.,
2012, 41, 11927–11948.
Phys. B, 2012, 21, 078201–078209.
39 Y.-S. Kwon, K. B. Gerasimov and S.-K. Yoon, J. Alloys
Compd., 2002, 346, 276–281.
40 R. Schmidt, H. M. Scholze and A. Stolle, Int. J. Ind. Chem.,
2016, 7, 181–186.
41 G. Mulas, L. Schiffini and G. Cocco, Mater. Sci. Forum,
1996, 225, 237–242.
17 E. Gil-González, A. Perejón, P. E. Sánchez-Jiménez,
M. A. Hayward and L. A. Pérez-Maqueda, J. Eur. Ceram. 42 M. Magini, A. Iasonna and F. Padella, Scr. Mater., 1996, 34,
Soc., 2017, 37, 945–954. 13–19.
18 P. Opitz, M. P. Asta, A. Fernandez-Martinez, M. Panthöfer, 43 L. Takacs and J. S. McHenry, J. Mater. Sci., 2006, 41, 5246–
A. Kabelitz, F. Emmerling, M. Mondeshki and W. Tremel,
Cryst. Growth Des., 2020, 20, 6831–6846.
19 M. Liu, L. Wang, K. Zhao, S. Shi, Q. Shao, L. Zhang, X. Sun,
5249.
44 P. P. Chattopadhyay, I. Manna, S. Talapatra and S. K. Pabi,
Mater. Chem. Phys., 2001, 68, 85–94.
Y. Zhao and J. Zhang, Energy Environ. Sci., 2019, 12, 2890– 45 L. Hao, Y. Lu, H. Sato, H. Asanuma and J. Luo, Int. J. Miner.
2923. Process., 2013, 121, 51–58.
20 J. Li, C. Lai, X. Jia, L. Wang, X. Xiang, C.-L. Ho, H. Li and 46 G. Mestl, P. Ruiz, B. Delmon and H. Knozinger, J. Phys.
W.-Y. Wong, Composites, Part B, 2015, 77, 248–256.
Chem., 1994, 98, 11269–11275.
This journal is © The Royal Society of Chemistry 2021
Dalton Trans.

View Article Online
Paper
Dalton Transactions
47 D. Liu, W. W. Lei, J. Hao, D. D. Liu, B. B. Liu, X. Wang, 52 D. V. Evtuguin, A. I. D. Daniel, A. J. D. Silvestre,
X. H. Chen, Q. L. Cui, G. T. Zou, J. Liu and S. Jiang, J. Appl.
Phys., 2009, 105, 023513.
F. M. L. Amado and C. P. Neto, J. Mol. Catal. A: Chem.,
2000, 154, 217–224.
48 L. Abello, E. Husson, Y. Repelin and G. Lucazeau, 53 M. Wang, J. Ma, H. Liu, N. Luo, Z. Zhao and F. Wang, ACS
Spectrochim. Acta, Part A, 1983, 39A, 64l–651. Catal., 2018, 8, 2129–2165.
49 F. D. Hardcastle and I. E. Wachs, J. Phys. Chem., 1991, 95, 54 Y. Ma, Z. Du, J. Liu, F. Xia and J. Xu, Green Chem., 2015, 17,
5031–5041. 4968–4973.
50 E. Skwarek, S. Khalameida, W. Janusz, V. Sydorchuk, 55 T. Voitl and P. R. von Rohr, Ind. Eng. Chem. Res., 2010, 64,
N. Konovalova, V. Zazhigalov, J. Skubiszewska-Zięba and
R. Leboda, J. Therm. Anal. Calorim., 2011, 106, 881–894.
520–525.
56 I. V. Kozhevnikov, Chem. Rev., 1998, 98, 171–198.
51 J. Zakzeski, P. C. A. Bruijnincx, A. L. Jongerius and 57 M. J. Janik, B. B. Bardin, R. J. Davis and M. Neurock,
B. M. Weckhuysen, Chem. Rev., 2010, 110, 3552–3599. J. Phys. Chem. B, 2006, 110, 4170–4178.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2021