Heterogeneous selective oxidation catalysts based on coordination polymer MIL-101 and transition metal-substituted polyoxometalates

Article abstract of DOI:10.1016/j.jcat.2008.05.014

Titanium- and cobalt-monosubstituted Keggin heteropolyanions, [PW₁₁CoO₃₉]^5- and [PW₁₁TiO₄₀]^5-, were electrostatically bound to the chromium terephthalate polymer matrix MIL-101. The MIL-supported polyoxometalate (POM) catalysts were characterized by elemental analysis, XRD, N₂ adsorption, and FT-IR-spectroscopy. The catalytic performance of both MIL-101 and the novel composite materials M-POM/MIL-101 was assessed in the oxidation of three representative alkenes-α-pinene, caryophyllene, and cyclohexene-using molecular oxygen and aqueous hydrogen peroxide as oxidants. Ti-POM/MIL-101 demonstrated fairly good catalytic activity and selectivity in α-pinene allylic oxidation and caryophyllene epoxidation with hydrogen peroxide, whereas Co-POM/MIL-101 catalyzed α-pinene allylic oxidation by molecular oxygen. Both composite materials are stable to POM leaching, behave as true heterogeneous catalysts, and can be used repeatedly without sustaining a loss of activity and selectivity in oxidations with O₂ and H₂O₂, provided that rather mild reaction conditions (T < 50 °C, [H₂O₂] < 0.2 M) are used with the latter oxidant.

Full text of DOI:10.1016/j.jcat.2008.05.014

Journal of Catalysis 257 (2008) 315–323
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Heterogeneous selective oxidation catalysts based on coordination polymer
MIL-101 and transition metal-substituted polyoxometalates
N.V. Maksimchuk a, M.N. Timofeeva a, M.S. Melgunov a, A.N. Shmakov a, Yu.A. Chesalov a, D.N. Dybtsev b, V.P. Fedin b,
O.A. Kholdeeva a,∗
a
Boreskov Institute of Catalysis, Prospekt Akademika Lavrentieva 5, Novosibirsk 630090, Russia
Nikolaev Institute of Inorganic Chemistry, Prospekt Akademika Lavrentieva 3, Novosibirsk 630090, Russia
b
a r t i c l e i n f o
a b s t r a c t
5−
5−
,
Article history:
Titanium- and cobalt-monosubstituted Keggin heteropolyanions, [PW11 CoO39
]
and [PW11 TiO40
]
Received 4 October 2007
Revised 23 March 2008
Accepted 11 May 2008
Available online 12 June 2008
were electrostatically bound to the chromium terephthalate polymer matrix MIL-101. The MIL-supported
polyoxometalate (POM) catalysts were characterized by elemental analysis, XRD, N2 adsorption, and
FT-IR-spectroscopy. The catalytic performance of both MIL-101 and the novel composite materials M-
POM/MIL-101 was assessed in the oxidation of three representative alkenes—α-pinene, caryophyllene,
and cyclohexene—using molecular oxygen and aqueous hydrogen peroxide as oxidants. Ti-POM/MIL-101
demonstrated fairly good catalytic activity and selectivity in α-pinene allylic oxidation and caryophyllene
epoxidation with hydrogen peroxide, whereas Co-POM/MIL-101 catalyzed α-pinene allylic oxidation by
molecular oxygen. Both composite materials are stable to POM leaching, behave as true heterogeneous
catalysts, and can be used repeatedly without sustaining a loss of activity and selectivity in oxidations
Keywords:
Metal-organic framework
MIL-101
Supported polyoxometalate
Composite materials
Co(II)
Ti(IV)
Selective oxidation
Alkene
◦
with O2 and H2O2, provided that rather mild reaction conditions (T < 50 C, [H2O2] < 0.2 M) are used
with the latter oxidant.
© 2008 Elsevier Inc. All rights reserved.
Molecular oxygen
Hydrogen peroxide
1. Introduction
as inorganic nature, thermodynamic stability to oxidation, hy-
drostability, tunability of redox and acid properties, and so on
[6–18]. Transition metal monosubstituted POMs (M-POMs) are use-
ful single-site catalysts where active metal M is isolated in and
strongly bound to an inorganic metal-oxide matrix and thus is pre-
vented from oligomerization and leaching [19,20]. Much work is
directed to immobilization of POMs on various solid supports via
dative [21], covalent [22], or electrostatic [23–28] binding.
Recently, Férey and co-workers have demonstrated a successful
immobilization of a Keggin heteropolytungstate within the cages
of the nanoporous metal-organic framework MIL-101, which has a
rigid zeotype crystal structure, large pores, and outstanding sur-
face area and also is resistant to air, water, common solvents, and
thermal treatment [29]. To date, no examples of catalysis over
POM/MIL-101 composite materials have been presented.
The design of active, selective, and recyclable heterogeneous
catalysts is a challenging goal of oxidation catalysis. Most known
solid catalysts, including redox molecular sieves, encapsulated,
grafted, and tethered redox metal complexes, suffers from leach-
ing of the active metal while using them in liquid-phase selective
oxidation processes and, in effect, behave as homogeneous rather
than heterogeneous catalysts [1–3]. A rare example of a highly effi-
cient and true heterogeneous catalyst is the microporous titanium
silicalite TS-1 developed by Enichem, which is currently used in
three industrial processes for selective oxidation of small organic
substrates, specifically phenol, cyclohexanone, and propene [4,5].
Many research efforts are directed toward synthesis of mesoporous
M,Si-catalysts, but most of them undergo irreversible deactivation
in the liquid phase due to hydrolysis, oligomerization, and subse-
quent leaching of the active metal.
Here we report on the preparation of novel heterogeneous
catalysts, based on MIL-101 and catalytically active Co- and Ti-
5−
monosubstituted Keggin-type heteropolyanions, [PW11 CoO39
]
Transition-metal oxoanion nanosized clusters or polyoxometa-
lates (POMs) have received increasing attention as oxidation cat-
alysts because of their numerous advantageous properties, such
5−
and [PW11 TiO40
]
, incorporated within the coordination polymer
nanocages. We describe their catalytic performance in liquid-phase
oxidation of three representative alkenes—cyclohexene, α-pinene,
and caryophyllene—using green oxidants—O2 and H2O2.
α-Pinene is the main component of gum turpentine, whereas
caryophyllene is a constituent of some essential oils (e.g., clove oil).
Corresponding author. Fax: +7 383 330 95 73.
E-mail address: khold@catalysis.nsk.su (O.A. Kholdeeva).
*
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.05.014

316
N.V. Maksimchuk et al. / Journal of Catalysis 257 (2008) 315–323
Both are low-cost and readily available renewable raw materials
for the production of a wide variety of fine chemicals, such as fra-
grances, flavors, drugs, and agrochemicals [30–32]. Caryophyllene
oxide has received FDA approval as a food and cosmetic stabilizer
[33]. It is also used as a flavoring substance, along with the α-
pinene allylic oxidation products verbenol and verbenone [31].
method [35]. The catalysts also were characterized by XRD and FT-
IR spectroscopy.
2.3. Catalytic oxidation
Catalytic experiments were carried out in thermostatted glass
◦
vessels under vigorous stirring at 50 C for α-pinene oxidation
2. Experimental
with molecular oxygen and caryophyllene oxidation with H2O2,
◦
◦
at 30 C for α-pinene oxidation with H2O2, and at 70 C for cy-
clohexene oxidation with H2O2. Typically, the reactions of alkene
oxidation with H2O2 were initiated by adding H2O2 (0.12 mmol
for α-pinene, 0.10 mmol for caryophyllene, and 0.40 mmol for cy-
clohexene oxidation) to a mixture containing an organic substrate
2.1. Materials
α-Pinene, containing 98% of α-pinene and 2% of β-pinene, was
obtained by vacuum rectification of gum turpentine. Hydrogen per-
oxide was used as 30% solution in water; its precise concentration
was determined iodometrically before use. All other reactants were
obtained commercially and used without further purification.
Acid sodium salt NaH3[PW11 Ti(OH)O39] (NaH4PW11 TiO40, Ti-
POM) was synthesized as described previously [34]. Tetrabuty-
lammonium (TBA) salt [Bu4N]4H[PW11 Co(H2O)O39] (Co-POM) was
prepared by metathesis of Na5PW11 CoO39 with TBABr in water
at pH 2.7 as described previously [26]. The presence of one acid
proton in Co-POM was confirmed by potentiometric titration with
methanolic TBAOH (Aldrich).
(0.1 mmol of α-pinene or caryophyllene, or 0.2 mmol of cyclohex-
−4
ene), 14 mg of the catalyst (6 × 10
mmol of Ti-POM), an internal
standard (biphenyl), and 1 mL of acetonitrile.
In α-pinene oxidation with molecular oxygen, 0.1 mmol of α-
pinene was added to a preliminary blown with dioxygen mixture,
−4
which contained 14 mg of the catalyst (4×10 mmol of Co-POM),
internal standard (biphenyl), and 1 mL of acetonitrile. Aliquots of
the reaction mixture were withdrawn periodically during the reac-
tion course by a syringe through a septum. Each experiment was
reproduced at least 2 or 3 times. The reaction products were iden-
tified by GC-MS and quantified by GC. After the reactions, catalysts
were filtered off, washed with acetonitrile, dried in air at room
temperature overnight, and then reused.
2.2. Catalyst preparation and characterization
2.2.1. MIL-101 synthesis
The coordination polymer MIL-101 was prepared as described
by Férey et al. [29]. In a typical synthesis, a mixture of 1.2 g
(3 mmol) of Cr(NO3)3·9H2O, 500 mg of terephthalic acid (H2bdc,
3 mmol), and 0.6 mL of 5 M HF (3 mmol) in 15 mL H2O was
2.4. Instrumentation
GC analyses were performed using a Tsvet-500 gas chromato-
graph equipped with a flame ionization detector and a Supelco
MDN-5S quartz capillary column (30 m × 0.25 mm). GC-MS anal-
yses were carried out using an Agilent 6890 gas chromatograph
(with an HP-5ms 30 m × 0.25 mm quartz capillary column)
equipped with an Agilent MSD 5973 quadrupole mass-selective
detector. FT-IR spectra were recorded using KBr pellets contain-
◦
heated at 220 C for 8 h in a Teflon-lined stainless steel bomb. The
resulting green solid was passed through a coarse glass filter to
remove the unreacted colorless crystals of H2bdc and then filtered
on the dense paper filter. Then the green raw product was washed
◦
◦
in hot DMF (100 C, 8 h, 2 times) and in hot EtOH (80 C, 8 h, 2
◦
times), filtered off, and dried overnight in an oven at 75 C.
ing 0.3 wt% of sample on a BOMEM-MB-102 spectrometer in the
−1
250–4000 cm
range. Nitrogen adsorption at 77 K was stud-
2.2.2. Immobilization of M-POMs on MIL-101
ied using an ASAP-2020 instrument within
a partial pressure
Adsorption measurements were carried out in a glass reactor
range of 10−6–1.0. Before measurements, the samples were de-
◦
at 25 C. The adsorbent (MIL-101, 40 mg) was placed into the reac-
◦
gassed at 90 C for 48 h. UV–vis spectra were recorded using a
tor, and a solution of Ti-POM in acetonitrile (1 mL, 1–10 mmol/L)
was added. Samples of the solution were obtained by a syringe
(100 μL) after 24 h and diluted with water (50 mL). Then the
Ti-POM concentration in solution was determined by UV–vis (λ =
250 nm).
Specord M40 spectrophotometer (l = 10 mm, accuracy of mea-
surements 10%). Particle size was measured using a Shimadzu
SALD-2101 laser diffraction particle size analyzer. XRD measure-
ments were performed on a high-precision X-ray diffractometer
mounted on beamline 2 of the VEPP-3 storage ring at the Siberian
Synchrotron Radiation Center (SSRC). The radiation wavelength was
λ = 0.15393 nm. The high natural collimation of synchrotron radi-
ation beam, flat perfect crystal analyzer, and parallel Soller slit on
the diffracted beam provided very high instrumental resolution of
◦
Desorption measurements were carried out at 25 C as follows.
MeCN (0.5 mL) was added to 40 mg of a MIL-101 sample con-
taining Ti-POM (with the amount of Ti-POM determined from the
adsorption measurements). Samples of the solution (100 μL) were
obtained by a syringe after 24 h and diluted with water (50 mL);
then the Ti-POM concentration in the solution was determined
by UV–vis. The amount of POM irreversibly adsorbed on MIL-101
was determined from the point at which the desorption curve
cut the axis (Y ). A similar procedure was used for the adsorp-
tion/desorption measurements for Co-POM.
◦
◦
the diffractometer in a small-angle region of 2θ = 0.5 –10 and
higher.
3. Results and discussion
3.1. Catalysts preparation and characterization
Supported M-POM/MIL-101 samples were prepared by dissolv-
ing 11 mg of M-POM (NaH4PW11 TiO40 or TBA4HPW11 CoO39) in
MeCN (1.5 mL), adding 100 mg of MIL-101, stirring for 3 h, storing
overnight at room temperature, filtering off, washing with MeCN,
and then drying in air until the weight remained constant. The
complete disappearance of M-POM in the MeCN solution was con-
firmed by UV–vis. The M-POM loading also was evaluated by the
elemental analysis.
MIL-101 was prepared as described by Férey et al. [29]. The
XRD data confirm the MIL-101 structure reported in [29]. The XRD
pattern calculated for MIL-101 based on the structural informa-
tion (Fig. 1A) and the experimental XRD pattern (Fig. 1B) were
very close. The size of the MIL-101 particles was in the range of
5–10 μm; such small particle sizes suggests no significant diffusion
limitations inside the MIL-101 crystals.
The textural characteristics of the supported catalysts were de-
termined from nitrogen adsorption isotherms. The specific sur-
face area and pore volume were measured by the comparative
Co-POM/MIL-101 and Ti-POM/MIL-101 composites (shown sche-
matically in Fig. 2) were prepared by M-POM adsorption from

N.V. Maksimchuk et al. / Journal of Catalysis 257 (2008) 315–323
317
Table 1
Physico-chemical properties of the catalysts studied
M/W/Cra (wt%) M/W/Crb (wt%) Ac (m2/g) V p (cm3/g)
d
Sample
MIL-101
–
–
2220
25e, 80f
1.13
0e,f
10% Ti-POM/MIL-101 0.17/7.4/12.0
7% Co-POM/MIL-101 0.14/5.2/12.4
0.13/4.1/12.1
0.14/5.1/14.4g
1930
0.96
0.14/4.15/15.1
0.14/4.2/14.8h
2050
1610i
1.03
0.81i
a
Evaluated from the adsorption data.
Based on the elemental analysis data.
Specific surface area.
Mesopore volume.
After three treatments with 0.4 M H2O2 at 70 C for 6 h.
After four treatments with 0.1 M H2O2 at 70 C for 6 h.
After five treatments with 0.2 M H2O2 at 50 C for 4 h.
After three catalytic cycles of α-pinene oxidation with molecular oxygen, the
b
c
d
e
f
◦
◦
g
h
◦
Fig. 1. Calculated XRD pattern for MIL-101 based on the structural informa-
tion (A) and experimental XRD patterns for MIL-101 (B), Ti-POM/MIL-101 (C),
Co-POM/MIL-101 (D), and Co-POM/MIL-101 after five cycles of α-pinene oxidation
with molecular oxygen (E).
reaction conditions as in Table 2.
After five catalytic cycles of α-pinene oxidation with molecular oxygen, the re-
action conditions as in Table 2.
i
1 M solution of Bu4NClO4 in MeCN, demonstrating the electro-
static character of binding between POM anion and the positively
charged surface of the MIL-101 nanocage,
ꢀ
ꢁ
5 Cr3O(bdc)3(H2O)2F + POM5−
→ Cr3O(bdc)3(H2O)2 5POM + 5F .
ꢀ
ꢁ
−
Some redistribution of XRD reflexes intensities was observed for
the Co-POM/MIL-101 and Ti-POM/MIL-101 composites compared
with the initial MIL-101 support (Fig. 1). A similar redistribution
was reported for POM-modified MIL-101 by Férey [29] and at-
tributed to interaction of POM anion with MIL-101 surface; how-
ever, both the positions and width of the X-ray reflections re-
mained unchanged compared with the initial MIL-101 sample.
Thus, the composite materials obtained did not lose their crys-
talline structure during impregnation with POM. This was also
confirmed by the FT-IR study (vide infra)
Table 1 gives the elemental analysis data for the supported
POM/MIL-101 samples along with the M/W/Cr values calculated
for a specific POM loading acquired from the adsorption measure-
ments. Keeping in mind that elemental analysis of POMs in the
presence of MIL is not an easy task, we may conclude that the
amount of POM in the composite materials obtained by the two
methods were quite similar.
Fig. 2. Schematic representation of M-POM encapsulated within MIL-101 matrix.
◦
acetonitrile solutions at 25 C. The adsorption/desorption pro-
cesses were monitored by UV–vis (see Section
2 for details).
The adsorption and desorption curves for NaH4PW11 TiO40 and
[Bu4N]4HPW11 CoO39 are shown in Figs. 3a and 3b, respectively.
An irreversible character of the sorption of about 10 wt% of
PW11 TiO5− or 7 wt% of PW11 CoO5− (∼1 POM anion per MIL-
Table
1 also presents the textural properties (i.e., specific
surface areas and mesopore volumes) acquired from the low-
temperature N2 adsorption measurements. Fig. 4a shows N2 ad-
sorption isotherms for both MIL-101 and supported POMs in
semilogarithmic coordinates. The isotherms are characterized by
40
39
101 nanocage) indicated a strong interaction between POM and
the support. The immobilized POM could be reextracted using a
◦
Fig. 3. Adsorption/desorption isotherms for (a) NaH4PW11 TiO40 and (b) [Bu4N]4HPW11 CoO39 on MIL-101 (MeCN, 25 C). Adsorption and desorption curves are marked by (2)
and (a), respectively.

318
N.V. Maksimchuk et al. / Journal of Catalysis 257 (2008) 315–323
Fig. 4. (a) N2 adsorption isotherms and (b) adsorption isotherms in αS coordinates for MIL-101 (A), Co-POM/MIL-101 (B), and Ti-POM/MIL-101 (C).
(a)
Fig. 5. FT-IR spectra of MIL-101 (A), Ti-POM/MIL-101 (B), and Co-POM/MIL-101 (C).
four pressure ranges within which the steplike behavior is ob-
served: 4 × 10−6–8 × 10−6, 5 × 10−4–10−3, 10−2–2 × 10−1, and
7 × 10−1–1.0. This steplike behavior in adsorption most likely is
related to the presence of different kinds of interactions at the
MIL-101 surface, condensation near cage entrances, and in the
cages themselves. At present, we do not have an appropriate model
for quantitative evaluation of pore volume–size distributions in
MIL-type materials; the development of such a model is currently
in progress.
The surface areas were evaluated using the comparative plot
method (Fig. 4b) [35]. Despite the existence of several types of
pores, the comparative plots demonstrated a linear behavior in the
range of almost complete monolayer formation (0.1 < αS < 0.4).
The corresponding interpolation lines passed through the origin,
indicating the essential nonmicroporous nature of the materials
studied. The observed affinity of the reference and study materi-
als provides strong evidence of the correctness of the surface area
measurements. Of note, determining surface area by means of BET
or Langmuir interpolation is not reliable, because some points on
adsorption isotherms (which comprise both layered adsorption and
condensation) can be included in the interpolation. The second lin-
ear range was observed for αS near 2.0. The corresponding slope
gave the surface area outside the structural mesopores, and the in-
tercept on the ordinate axis gave the volume of mesopores. The
measured surface areas and volume of pores are shown in Table 1.
The volume of mesopores for initial MIL-101 was close to the the-
oretical approximation of 1.37 cm3/g. The decreased pore volume
on POM deposition corresponds to the mass increase due to the
introduction of POM (Table 1).
(b)
Fig. 6. FT-IR spectra of (a) Co-POM and (b) Ti-POM: M-POM (A), M-POM/MIL-101 (B),
4 catalytic cycles of α-pinene oxidation (C). (B) and
(C)—after subtraction of the spectrum of MIL-101.
and M-POM/MIL-101 after
sented in Fig. 5. These spectra look quite similar because of the
rather low POM loading; the most intense peaks belong to the
MIL-101 matrix. Strong bands in the region of 1800–1300 cm
correspond to νas(COO), νs(COO), and ν(C–C) vibrations, implying
the presence of dicarboxylate linker in the MIL framework. Rather
weak and narrow bands at 1017 and 749 cm can be attributed to
δ(C–H) and γ (C–H) vibrations of aromatic rings, respectively. Weak
bands in the spectral region of 700–400 cm
to in-plane and out-of-plane bending modes of COO-groups [36].
Fig. 6 shows the IR spectra of bulk M-POMs and of the MIL-
supported M-POMs after subtraction of the spectrum of MIL-101.
The IR spectra of the M-POM/MIL composites exhibit the principal
stretching modes of the Keggin POM unit: 1077 and 1073 νas(PO4);
−1
−1
−1
are most likely due
−1
950 and 969 νas(W=O); and 883, 823, and 890, and 803 cm
νas(W–O–W), for Co-POM and Ti-POM, respectively. This definitely
indicates that the heteropolyanion structure was maintained after
immobilization.
FT-IR studies were performed to check whether the M-POM and
MIL structures were preserved in the composite materials. FT-IR
spectra of MIL-101, Ti-POM/MIL-101, and Co-POM/MIL-101 are pre-

N.V. Maksimchuk et al. / Journal of Catalysis 257 (2008) 315–323
319
◦
Fig. 7. Hydrogen peroxide decomposition: (A) blank experiment at 50 C, (B) over
MIL-101 at 30 C, and (C) over MIL-101 at 50 C.
Fig. 8. Effect of α-pinene conversion on verbenol/verbenone selectivity in oxidation
with (A) H2O2 over Ti-POM/MIL-101 and (B) O2 over Co-POM/MIL-101.
◦
◦
3.2. Catalytic oxidations
cline on immobilization of POMs. In both catalytic systems studied,
α-pinene was converted mainly to the allylic oxidation products
verbenol and verbenone, along with small amounts of campholenic
aldehyde.
The catalytic performance of the new M-POM/MIL-101 compos-
ites was assessed in α-pinene oxidation with H2O2 and O2, as well
as in cyclohexene and caryophyllene oxidation with H2O2 in ace-
tonitrile (see Scheme 1).
The α-pinene allylic oxidation to verbenol/verbenone with both
molecular oxygen and H2O2 showed a strong negative effect of the
reaction conversion on its selectivity due to the high reactivity of
the products (Fig. 8), which is typical for allylic oxidation [37–39].
The selectivity to verbenol/verbenone reached a maximum (81–
84%) at a rather low substrate conversion (15–25%) and then de-
clined due to overoxidation processes [37–40]. Interestingly, the
MIL-supported POMs were more selective in α-pinene oxidation
than the silica-supported POMs [38,39], likely due to the MIL-101’s
zeotype porous structure and smaller pore diameter, which may
have protected allylic oxidation products from further oxidative
polymerization.
Oxidation of cyclohexene with H2O2 over Ti-POM/MIL-101 also
gave predominantly the allylic oxidation products cyclohexene-
2-ol and cyclohexene-2-one, pointing to a homolytic oxidation
mechanism [41]. The yield of trans-cyclohexane-1,2-diol was lower
than that of the homogeneous Ti-POM. In turn, caryophyllene ox-
idation with H2O2 over Ti-POM/MIL produced caryophyllene 4,5-
Table 2 presents the results of the alkene oxidations over Ti-
and Co-POM/MIL-101 catalysts, along with blank experiments and
experiments performed in the presence of the corresponding ho-
mogeneous M-POMs. The catalytic study revealed that both Ti- and
Co-POM incorporated into MIL-101 matrix catalyzed alkene oxi-
dation with H2O2 and O2, respectively, whereas the matrix alone
was quite inert in most cases (Table 2). The MIL matrix exhibited
some activity only in caryophyllene oxidation with H2O2, in con-
trast with α-pinene oxidation with H2O2, where the matrix was
almost inert. This may be due to the higher reaction temperature
used for the caryophyllene oxidation. Indeed, no H2O2 decomposi-
tion over MIL-101 occurred at 30 C, whereas some activity of the
matrix was observed at 50 C (Fig. 7).
It is noteworthy that the activity of the M-POM/MIL catalysts
expressed in TOF values determined from the initial rates of sub-
strate consumption was close to the activity of the corresponding
homogeneous M-POMs; this indicates that the activity did not de-
◦
◦
Scheme 1. Selective alkene oxidations over M-POM/MIL-101.

320
N.V. Maksimchuk et al. / Journal of Catalysis 257 (2008) 315–323
Table 2
Alkene oxidation with H2O2 and O2 in the presence of M-POM/MIL-101 catalysts
−1
Catalyst
Oxidant
Time (h)
Substrate
TOFa (h
)
Selectivityb (%)
conversion (%)
c
d
–
1 atm O2
1 atm O2
1 atm O2
1 atm O2
0.12 M H2O2
0.12 M H2O2
0.12 M H2O2
0.12 M H2O2
1
1
2
2
5
5
5
5
15
45
14
45
16
42
18
40
–
80
–
80
–
40
–
33
36
14
29
7
25
8
32
27
24
7
27
7
18
7
30
Co-POMe
MIL-101f
Co-POM/MIL-101
d
–
Ti-POMe
MIL-101f
Ti-POM/MIL-101
40
g
d
–
0.1 M H2O2
0.1 M H2O2
0.1 M H2O2
0.1 M H2O2
0.2 M H2O2
4
4
4
4
4
13
60
40
71
88
–
92
–
82
130
23
52
41
80
Ti-POMe
MIL-101f
Ti-POM/MIL-101
Ti-POM/MIL-101
100
h
Ti-POMe
0.4 M H2O2
0.4 M H2O2
0.4 M H2O2
6
6
6
7
8
39
4
–
21
7
60
32
32
10
46
MIL-101f
Ti-POM/MIL-101
a
TOF = (moles of substrate consumed)/[(moles of M-POM) × time]; determined from the initial rates.
b
c
d
e
f
GC yield based on substrate consumed (high boiling overoxidation products were also formed along with the main detectable products).
−4
−4
◦
◦
0.1 M α-pinene, 14 mg catalyst (4 × 10
mmol of Co-POM or 6 × 10
mmol Ti-POM), 1 mL MeCN, 50 C (for oxidation with O2) or 30 C (for oxidation with H2O2).
No catalyst was present.
−4
−4
Homogeneous catalyst (4 × 10
mmol of [Bu4N]4HPW11 CoO39 or 6 × 10
−4
mmol of NaH4PW11 TiO40).
MIL-101 matrix without M-POM.
g
h
◦
0.1 M caryophyllene, 14 mg catalyst (6 × 10
mmol Ti-POM), 1 mL MeCN, 50 C.
−4
◦
0.2 M cyclohexene, 0.4 M H2O2, 20 mg catalyst (9 × 10
mmol Ti-POM), 1.5 mL MeCN, 70 C.
Fig. 9. Oxidation of α-pinene (a) with H2O2 over Ti-POM/MIL-101 and (b) with O2 over Co-POM/MIL-101. (c) Oxidation of caryophyllene with H2O2 over Ti-POM/MIL-101.
monoepoxide with good to excellent selectivity. Importantly, the
selectivity of caryophyllene epoxidation was significantly higher for
MIL-supported Ti-POM than for homogeneous Ti-POM (Table 2).
This may be due to deprotonation of the Ti-POM anion on immo-
bilization within the MIL-101 cages via anion exchange (vide supra),
which protected the acid-sensitive epoxide from the ring opening.
The substrate conversion and epoxide selectivity increased with in-
creasing H2O2 concentration (Table 2).
3.3. Catalyst stability and recycling
The essential properties of solid catalysts are the stability of
the active component with respect to leaching and other transfor-
mations, and the stability of the porous structure. The nature of
the catalysis is also a crucial factor in liquid-phase processes. The
experiments with fast catalyst filtration performed at the reaction
temperature showed no further α-pinene and caryophyllene con-

N.V. Maksimchuk et al. / Journal of Catalysis 257 (2008) 315–323
321
Fig. 10. (a) Co-POM/MIL-101 recycling in α-pinene oxidation with molecular oxygen and (b) FT-IR spectra of Co-POM/MIL-101: (A) initial and (B) after 5 catalytic cycles of
α-pinene oxidation with O2.
Fig. 11. (a) Ti-POM/MIL-101 recycling in cyclohexene oxidation with H2O2 and (b) FT-IR spectra of Ti-POM/MIL-101: (A) initial and (B) after 4 catalytic cycles of cyclohexene
oxidation with H2O2.
version in the filtrate after the catalysts were removed (Fig. 9). This
indicates that the oxidation process is truly heterogeneous and oc-
curs on the catalyst surface rather than in solution. Even if very
small amounts of titanium or cobalt species are leached from the
solid matrix during the oxidation process, the observed catalytic
activity is not due to these species. Furthermore, elemental anal-
ysis data (Table 1) showed no evident POM leaching under the
conditions used in this study.
An important issue that must be addressed while studying
liquid-phase oxidations over a solid catalyst is a possibility of cat-
alyst recycling. We found that the Co-POM/MIL-101 catalyst main-
tained its activity and selectivity for at least 5 catalytic cycles
of the α-pinene oxidation with O2 (Fig. 10a). The XRD pattern
(Fig. 1D) of this catalyst demonstrated no significant changes in
the structure. The N2 adsorption measurements (Table 1) revealed
a 20% decrease in the specific surface area and pore volume, which
most likely can be explained by a mass increase of the cata-
lyst with hardly removable reaction byproducts. The FT-IR study
indicated preservation of both the POM and MIL-101 structures
(Fig. 10b).
In contrast, when H2O2 was used as the oxidant, the reac-
tion conditions had a significant impact on catalyst stability. Thus,
the catalytic performance of Ti-POM/MIL-101 in cyclohexene ox-
idation changed significantly from cycle to cycle. After the third
cycle, the catalytic activity sharply increased, and the ratio of prod-
ucts changed significantly, indicating that the reaction mechanism
had become more heterolytic (Fig. 11a). This most likely resulted
from oxidative degradation of the MIL-101 structure accompanied
by Ti-POM leaching. Indeed, the FT-IR-spectra demonstrated de-
struction of the MIL-101 structure after four catalytic cycles of
cyclohexene oxidation. This was manifested by decreasing intensity
of the bands attributed to vibration of benzene ring at 1017 and
−1
749 cm
and carboxylic groups of terephthalic acid in the spec-
−1
tral region of 1800–1300 cm
(Fig. 11b). Moreover, new bands
corresponding to stretching C=O vibrations appeared at 1709 and
−1
1686 cm
along with new stretching OH vibrations at 3300–
2700 cm−1, most likely indicating protonation of COO-groups [36].
In sharp contrast, the Ti-POM/MIL-101 catalyst did not lose its
activity/selectivity during several catalytic cycles of both α-pinene
and caryophyllene oxidation (Fig. 12). This may be explained by
taking into account the milder reaction conditions used in the oxi-
dation of these two substrates compared with the conditions for
cyclohexene oxidation. Indeed, the FT-IR findings confirmed the
preservation of both the POM and MIL-101 structures after cata-
lyst recycling in α-pinene and caryophyllene oxidation with H2O2
(Fig. 13). In effect, after four catalytic cycles, the principal stretch-
ing modes of both the Keggin POM unit and terephthalic acid
remained unchanged.
To gain insight into the factors crucial for the stability of the
composite POM/MIL-101 materials, we studied the resistance of
the MIL-101 matrix to aqueous H2O2. Treatment of MIL-101 with
H2O2 was followed by FT-IR, small-angle XRD, and N2 adsorp-
tion. All of the methods collectively demonstrated the matrix was
unstable at rather high oxidant concentrations (0.4 M; Table 1,
Fig. 14a). After three consecutive treatments of MIL-101 with 0.1 M
◦
H2O2 at 70 C, the FT-IR spectrum remained practically unchanged
(Fig. 14b), but no XRD was observed. This may indicate preser-
vation of the short-range order, but accompanied by loss of the

322
N.V. Maksimchuk et al. / Journal of Catalysis 257 (2008) 315–323
◦
◦
Fig. 12. Ti-POM/MIL-101 recycling in (a) α-pinene oxidation with 0.12 M H2O2 at 30 C and (b) caryophyllene oxidation with 0.1 M H2O2 at 50 C.
(a)
(b)
Fig. 13. FT-IR spectra of Ti-POM/MIL-101: (A) initial, (B)–(F) after 1–5 catalytic cycles of (a) α-pinene and (b) caryophyllene oxidation with H2O2.
(a)
(b)
◦
◦
Fig. 14. FT-IR spectra of MIL-101: (A) initial, (B)–(E) after 1–4 treatments with (a) 0.4 M H2O2 at 70 C and (b) 0.1 M H2O2 in MeCN at 70 C for 6 h.
supramolecular structure. After both treatments with H2O2, the
catalyst lost its porous structure, as demonstrated by the textural
data in Table 1.
mesoporous zeotype coordination polymer MIL-101. The compos-
ite POM/MIL-101 materials demonstrate fairly good catalytic ac-
tivity and selectivity in α-pinene allylic oxidation (81–84% ver-
benol/verbenone selectivity at 15–25% substrate conversion) and
caryophyllene epoxidation (100% selectivity at 88% conversion)
with green oxidants H2O2 (Ti-POM) and O2 (Co-POM). The Co-
POM/MIL-101 catalyst can be used repeatedly with no loss of ac-
tivity and selectivity in oxidation with molecular oxygen. Reaction
4. Conclusion
Our findings demonstrate that Ti- and Co-monosubstituted Keg-
gin POMs can be attached electrostatically to the surface of the

N.V. Maksimchuk et al. / Journal of Catalysis 257 (2008) 315–323
323
conditions have a significant impact on the stability of the Ti-
POM/MIL-101 catalyst when H2O2 is used as the oxidant. Under
rather mild conditions (<50 C, [H2O2] < 0.2 M), the Ti-POM/MIL-
101 composite is rather stable, behaves as a true heterogeneous
catalyst, does not undergo POM leaching, and can be recycled with
no loss of catalytic properties. At higher H2O2 concentrations and
higher temperatures, the MIL-101 matrix is destroyed, and the
POM/MIL-101 catalyst undergoes deactivation.
[16] M.T. Pope, in: A.G. Wedd (Ed.), Comprehensive Coordination Chemistry II,
vol. 4, Elsevier, New York, 2004, p. 635.
[17] R. Neumann, A.M. Khenkin, Chem. Commun. (2006) 2529.
[18] C.L. Hill, in: A.G. Wedd (Ed.), Comprehensive Coordination Chemistry II, vol. 4,
Elsevier, New York, 2004, p. 679.
[19] J.M. Thomas, R. Raja, Top. Catal. 40 (2006) 3.
[20] O.A. Kholdeeva, Top. Catal. 40 (2006) 229.
[21] B.J.S. Johnson, A. Stein, Inorg. Chem. 40 (2001) 801.
[22] R.J. Errington, S.S. Petkar, B.R. Horrocks, A. Houlton, L.H. Lie, S.N. Patole, Angew.
Chem. Int. Ed. 44 (2005) 1254.
◦
[23] R. Neumann, H. Miller, J. Chem. Soc. Chem. Commun. (1995) 2277.
[24] W. Kaleta, K. Nowinska, Chem. Commun. (2001) 535.
[25] N.M. Okun, T.M. Anderson, C.L. Hill, J. Am. Chem. Soc. 125 (2003) 3194.
[26] O.A. Kholdeeva, M.P. Vanina, M.N. Timofeeva, R.I. Maksimovskaya, T.A. Tru-
bitsina, M.S. Melgunov, E.B. Burgina, J. Mrowiec-Bialon, A.B. Jarzebski, C.L. Hill,
J. Catal. 226 (2004) 363.
[27] M.V. Vasylyev, R. Neumann, J. Am. Chem. Soc. 126 (2004) 884.
[28] J. Kasai, Y. Nakagawa, S. Uchida, K. Yamaguchi, N. Mizuno, Chem. Eur. J. 12
(2006) 4176.
Acknowledgments
The authors thank V. Utkin for identifying the oxidation prod-
ucts by GC/MS and E.V. Kulko for providing the particle size mea-
surements. The research was partially supported by the Russian
Foundation for Basic Research (RFBR-CNRS grant 05-03-34760).
[29] G. Férey, C. Mellot-Draznieks, C. Serre, F. Millange, J. Dutour, S. Surble, I. Mar-
giolaki, Science 309 (2005) 2040.
[30] W.F. Erman, Chemistry of the Monoterpenes: An Encyclopaedic Handbook,
Dekker, New York, 1985.
References
[1] R.A. Sheldon, M. Wallau, I.W.C.E. Arends, U. Schuchardt, Acc. Chem. Res. 31
(1998) 485.
[31] K. Bauer, D. Garbe, H. Surburg, Common Fragrance and Flavor Materials,
Wiley–VCH, New York, 1997.
[2] I.W.C.E. Arends, R.A. Sheldon, Appl. Catal. A Gen. 212 (2001) 175.
[3] J.S. Rafelt, J.H. Clark, Catal. Today 57 (2000) 33.
[32] J.D. Connoly, R.A. Hill, Dictionary of Terpenoids. Vol I: Mono- and Sesquiter-
penoids, Chapman & Hall, London, 1991.
[4] B. Notari, Adv. Catal. 41 (1996) 253.
[5] G. Centi, S. Perathoner, in: I.T. Horvath (Ed.), Encyclopedia of Catalysis, vol. 6,
Wiley–Interscience, NJ, 2003, p. 239.
[6] M.T. Pope, Heteropoly and Isopoly Oxometalates, Springler-Verlag, Berlin,
1983.
[33] Rules and Regulations, Food Drug Adm. 95 (38) (1973) 12913.
[34] O.A. Kholdeeva, T.A. Trubitsina, M.N. Timofeeva, G.M. Maksimov, R.I. Maksi-
movskaya, V.A. Rogov, J. Mol. Catal. A Chem. 323 (2005) 173.
[35] V.B. Fenelonov, V.N. Romannikov, A.Y. Derevyankin, Micropor. Mesopor. Mater.
28 (1999) 57.
[7] J.B. Moffat, Metal-Oxygen Clusters: The Surface and Catalytic Properties of
Heteropoly Oxometalates, Kluwer/Plenum, New York, 2001.
[8] C.L. Hill, C.M. Prosser-McCartha, Coord. Chem. Rev. 143 (1995) 407.
[9] N. Mizuno, M. Misino, Chem. Rev. 98 (1998) 199.
[36] D.W. Brown, A.J. Floyd, M. Sainsbury, Organic Spectroscopy, Willey, New York,
1988.
[37] E.F. Murphy, T. Mallat, A. Baiker, Catal. Today 57 (2000) 115.
[38] N.V. Maksimchuk, M.S. Melgunov, J. Mrowiec-Białon´ , A.B. Jarze¸ bski, O.A.
Kholdeeva, J. Catal. 235 (2005) 175.
[10] R. Neumann, Prog. Inorg. Chem. 47 (1998) 317.
[11] J.E. Lyons, P.E. Ellis, V.A. Durante, Stud. Surf. Sci. Catal. 67 (1991) 99.
[12] N. Mizuno, M. Tateishi, T. Hirose, M. Iwamoto, Chem. Lett. (1993) 2137.
[13] N. Mizuno, T. Hirose, M. Iwamoto, J. Mol. Catal. 88 (1994) L125.
[14] H. Weiner, R.G. Finke, J. Am. Chem. Soc. 121 (1999) 9831.
[15] R. Neumann, M. Fahan, J. Chem. Soc. Chem. Commun. (1995) 2277.
[39] N.V. Maksimchuk, M.S. Melgunov, Yu.A. Chesalov, J. Mrowiec-Białon´ , A.B.
Jarze¸ bski, O.A. Kholdeeva, J. Catal. 246 (2007) 241.
[40] P. McMorn, G. Roberts, G.J. Hutchings, Catal. Lett. 67 (2000) 203.
[41] R.A. Sheldon, J.K. Kochi, Metal-Catalyzed Oxidations of Organic Compounds,
Academic Press, New York, 1981.