Grafted Ta-calixarenes: Tunable, selective catalysts for direct olefin epoxidation with aqueous hydrogen peroxide

Article abstract of DOI:10.1016/j.jcat.2010.07.010

Supported Ta(V) oxide catalysts were prepared by grafting calixarene-Ta(V) and TaX₅ complexes on SiO₂ at coverages less than 0.25 Ta nm^-2. Thermogravimetric and elemental analyses and UV-visible spectroscopy indicate that catalysts consist of isolated, 1:1 Ta:ligand surface sites. Catalysts obtained by this one-pot procedure were studied in cyclohexene and cyclooctene epoxidation with H₂O₂. In sharp contrast with bare oxides, calixarene-containing catalysts had initial cyclohexene direct epoxidation turnover rates of 3.9 ± 0.1 × 10^-2 s ^-1 unaffected by surface density, demonstrating single-site character. Calixarene-containing catalysts were up to 3× more active than the corresponding TaCl₅-based catalysts at high surface densities and were also up to 95% selective to direct (non-radical) cyclohexene epoxidation versus <65% selectivity for TaCl₅-based catalysts or only ～20% selectivity for grafted calixarene-Ti(IV) catalysts. Capping silanols with octanol reduced epoxide hydrolysis from >50% to <30%. These catalysts demonstrate the utility of ligand-protected, supported Ta catalysts for epoxidation with H₂O₂ and demonstrate that a surface ligand on an otherwise traditionally prepared supported oxide can selectively direct oxidation down mechanistically distinct pathways.

Full text of DOI:10.1016/j.jcat.2010.07.010

Journal of Catalysis 275 (2010) 191–201
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Grafted Ta–calixarenes: Tunable, selective catalysts for direct olefin
epoxidation with aqueous hydrogen peroxide
Natalia Morlanés, Justin M. Notestein *
Chemical and Biological Engineering Department. Northwestern University, 2145 Sheridan Road, Technological Institute E136, Evanston, IL 60208, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 18 March 2010
Revised 7 July 2010
Accepted 10 July 2010
Available online 21 August 2010
Supported Ta(V) oxide catalysts were prepared by grafting calixarene–Ta(V) and TaX complexes on
5
ꢀ2
SiO
2
at coverages less than 0.25 Ta nm . Thermogravimetric and elemental analyses and UV–visible
spectroscopy indicate that catalysts consist of isolated, 1:1 Ta:ligand surface sites. Catalysts obtained
by this one-pot procedure were studied in cyclohexene and cyclooctene epoxidation with H . In sharp
contrast with bare oxides, calixarene-containing catalysts had initial cyclohexene direct epoxidation
2 2
O
ꢀ
2 ꢀ1
turnover rates of 3.9 ± 0.1 ꢁ 10
Calixarene-containing catalysts were up to 3ꢁ more active than the corresponding TaCl
at high surface densities and were also up to 95% selective to direct (non-radical) cyclohexene epoxida-
tion versus <65% selectivity for TaCl
-based catalysts or only ꢂ20% selectivity for grafted calixarene–
Ti(IV) catalysts. Capping silanols with octanol reduced epoxide hydrolysis from >50% to <30%. These
2 2
catalysts demonstrate the utility of ligand-protected, supported Ta catalysts for epoxidation with H O
s
unaffected by surface density, demonstrating single-site character.
Keywords:
Calixarene
Tantalum
5
-based catalysts
Heterogeneous catalyst
Supported catalyst
Supported oxide
Single-site catalyst
Olefin epoxidation
Hydrogen peroxide
Green chemistry
Allylic hydroperoxide
5
and demonstrate that a surface ligand on an otherwise traditionally prepared supported oxide can selec-
tively direct oxidation down mechanistically distinct pathways.
Ó 2010 Elsevier Inc. All rights reserved.
1
. Introduction
Tantalum complexes are known to mediate a number of inter-
butyl hydroperoxide and cumene hydroperoxide are often neces-
sary to achieve high activity [10]. Titania–silica supported and
mixed oxide materials are highly studied and active catalysts for
olefin and allylic alcohol epoxidation, but comparable group five
Nb and Ta catalysts have been shown in a limited number of stud-
ies to give high rate, selectivity and stability, especially when using
esting catalytic transformations including olefin oxidation [1] and
polymerization [2], arene hydrogenation [3], metathesis of alkanes
[
4], imines and olefins [5], or activation of C–N bonds as in hydro-
denitrogenation [6] and hydroamination [7]. However, a limited
range of ancillary ligands have been employed to modulate the
reactivity of the tantalum center. Moreover, most Ta catalysts are
employed as homogenous catalysts, while the development of
heterogeneous catalysts lags behind, comprising approximately
2 2
aqueous H O as the oxidant [11,12]. In particular, group five cat-
ions have been observed to be more tolerant to the presence of
water than the analogous Ti(IV) systems, as a result of the ability
to coordinate an additional anionic ligand when equal numbers
of surface Si–O–M linkages are assumed [11].
0
.05% of supported oxide catalysts [8]. In this respect, immobiliza-
For supported oxides in general, active site structures can, in
principle, be controlled by grafting well-defined transition metal
complexes onto a solid support. If retained during catalysis, the or-
ganic ligand cooperates with the rigid inorganic oxide surface to
define the active site, rather than it arising solely as a consequence
of the surface density of the metal oxide and the relative strength
of metal–metal and metal–support interactions [13]. Bulky, multi-
dentate ligands are ideal for this purpose, since they create stable
surface species, and their bulk ensures metal atom site isolation
during catalytic turnover. Here, calixarene–metal complexes simi-
lar to those reported previously for Ti or V active species are ex-
tended to Ta to take advantage of the diverse chemistry of this
cation [14–16]. It is shown here that the calixarene ligands lead
to well-defined behavior, with epoxidation rates that are not a
tion of homogeneous complexes may be a viable strategy route to
increase the diversity and utility of Ta heterogeneous catalysts.
Epoxidation is one of the most studied reactions in organic syn-
thesis, for both commodity and fine chemicals. In particular, ‘‘green
oxidants” such as H
only by-product (H O) is environmentally benign and the process
economics are not dependant on selling alcohol co-product [9].
However, a number of supported transition metal oxide catalysts
employed in this reaction are poisoned by coordination of water
to the active site, so non-aqueous organic oxidants such as tert-
2 2
O have become highly desirable, since the
2
*
Corresponding author. Fax: +1 847 491 3728.
E-mail address: j-notestein@northwestern.edu (J.M. Notestein).
0
021-9517/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2010.07.010

1
92
N. Morlanés, J.M. Notestein / Journal of Catalysis 275 (2010) 191–201
function of surface density, and are higher than those of the bare
Ta–SiO
Since catalytic epoxidation with aqueous H
complicated reaction network that includes direct epoxidation
with H , radical oxidation of the olefin to yield allylic hydro-
peroxides that in turn act as stoichiometric oxidants, epoxide
hydrolysis, and unproductive decomposition, reaction
became pink, in contrast to
a
yellow, previously reported
ꢃ
2
.
Cp Ta(CH
3
)
3
OSiPh
3
[25]. The color changes were not accompanied
2
O
2
can result in a
by dramatic changes in diffuse reflectance spectra, as discussed
below. Solid Ta-containing catalysts were stored inside the glove
box. For comparison, an analogous grafted calixarene–Ti(IV) catalyst
0.2-dmCal4Ti(OSiꢄ) was prepared as described previously [14].
2 2
O
2 2
H O
pathway control is critical. It is also shown here that grafted
calixarene–Ta is substantially more selective to direct epoxidation
2.3. Heat treatment
than either the bare Ta–SiO
2
or an analogous grafted calixarene–Ti.
Some catalysts were heated at 550 °C for 6 h in air to remove
Finally, it is shown that the diversity of commercially available
calixarenes enables tuning of the reactivity and selectivity of the
organic ligands. The resulting catalysts, denoted, e.g. 0.2-
Cal4Ta(OSiꢄ)
2
-c are assumed to have identical Ta content per g
Ta center in cyclohexene epoxidation with H
2
O
2
, particularly to
SiO as the starting materials.
2
minimize the relative rate of unselective radical reactions.
2.4. Analytical characterization
2
. Materials and methods
The mass loss of all grafted materials was determined by thermo-
gravimetry (TGA; TA Instruments Q500) in a flow of dry synthetic
2.1. General remarks
ꢀ1
air at a heating rate of 8 °C min
from room temperature to
8
00 °C. Mass losses were corrected for hydroxyl condensation,
2
All syntheses were carried out under N using standard Schlenk
which was determined from TGA of unmodified SiO previously
treated identically to the catalyst supports. Ligand contents were
estimated by assuming that mass losses reflect combustion of a
2
lines techniques or under Ar in a controlled atmosphere glove box.
Cyclohexene, cyclooctene, and solvents were obtained at the high-
est purity available, dried as recommended [17], degasified and
ꢀ1
molecular weight of 612 g mol site in material Cal4Ta(OSiꢄ)
2
stored inside the glove box. Aqueous H
2 2
O solution 30 wt.% (Al-
and were adjusted accordingly for other grafted species.
drich) was used as received. Fresh tantalum pentachloride (TaCl
5
)
Ta content in the catalyst was measured using inductively cou-
anhydrous 99.9%, pentamethylcyclopentadienyltantalum tetra-
ꢀ1
pled plasma techniques. Tantalum standards (from 1000 lg mL
ꢃ
chloride (Cp TaCl
4
) 98%, and pentakis(dimethylamino)tantalum
Specpure AAS standard solution) were prepared for the calibration
of the instrument, and linear calibration curves were obtained in
the range 1–50 Ta ppm. To achieve quantitative removal of Ta for
the support, it was necessary to use 0.7 mL piranha reagent
(
Ta(dma)
Benzylmagnesium chloride 1.0 M in diethyl ether (Aldrich) was
used as received to synthesize pentabenzyltantalum (Ta(Bz)
according to published procedures, to yield a brown complex
18]. p-tert-Butylcalix[4]arene (Cal4) 98% (Aldrich), p-tert-butyl-
calix[6]arene (Cal6) 98% (Aldrich), and p-tert-butylthiacalix[4]arene
SCal4) 98% (TCI) were used as received. 1,3-Dimethoxy-tert-butyl-
5
) 99% (all from Strem Chemicals) were used as received.
5
)
(
H
2
SO
in an ultrasound bath, and diluted to 14 mL. [CAUTION: make
and handle piranha reagent with extreme care.] Residual H
4 2 2
:H O = 3:1) per 10 mg catalyst, which was stirred for 1 h
[
2 2
O
(
was decomposed before the ICP measurement. Multiple trials were
averaged to obtain accurate results.
calix[4]arene (dmCal4) was prepared according to a previously pub-
lished procedure [19].
UV–visible spectra were measured at ambient conditions using
a UV–visible spectrophotometer (UV-3600 Shimadzu). Absorption
spectra of Ta complexes were collected in toluene solution and dif-
fuse reflectance spectra of Ta-containing catalysts were collected
using a Harrick Praying Mantis diffuse reflectance accessory. Sam-
ples for diffuse reflectance were exposed to ambient air. Pressed
poly-(tetrafluoroethylene) powders were used as perfect reflector
standard in calculating Kubelka–Munk pseudoabsorbances. Optical
edge energies of the catalysts were calculated by assuming that the
lowest energy transitions (calixarene–Ta LMCT) were allowed in-
direct transitions [26]. Solid-state C CP/MAS NMR was performed
at the Northwestern IMSERC facility using a solids 400 Varian spec-
trometer at 400 MHz, a spinning rate of 7 kHz and 5680 scans. Dif-
fuse Reflectance Infrared Fourier Transform spectroscopy (DRIFTS)
was performed in the Keck-II facility of NUANCE Center at North-
western using a Nexus 870 FTIR spectrometer with diffuse reflec-
tance accessory. Spectra were normalized with respect to the
2
.2. Catalyst synthesis
Tantalum calixarene complexes were synthesized by adding
ꢃ
0
.1 mmol TaCl
5
or Cp TaCl
4
to a stirred suspension of the calix-
arene (0.11 mmol) in anhydrous, degassed toluene (25 mL) at room
temperature inside the glove box. Sealed flasks were removed from
the glove box, and the mixtures were heated to reflux for 14–36 h
under N
20], dmCal4TaCl
Cal4TaCp [24]. In all cases, yellow or orange complexes were
formed. Specific synthesis details are given in the Supplementary
material. Other Ta compounds were dissolved in anhydrous, de-
2
following previously described methods to yield Cal4TaCl
[21], SCal4TaCl [22], Cal6TaCl [23], and
1
3
[
2
2
ꢃ
gassed toluene at the same concentration. SiO
2
(Selecto Scientific,
00 m g surface area, average pore diameter 6.0 nm, particle
sizes 32–63 m, treated under dynamic vacuum at 300 °C for
4 h) was transferred to the flasks containing the previously syn-
thesized precursors. Ratios of 0.02, 0.05, 0.1, 0.15, 0.2, and
.0 mmol Ta per g SiO were used in the grafting step by altering
the amount of SiO added. Nomenclature used for the catalysts
is, e.g. for Cal4TaCl precursor, n-Cal4Ta(OSiꢄ) , where n is the syn-
thesis ratio. Suspensions were refluxed under N with the excep-
tion of Ta(Bz) which was stirred at room temperature in the
2
ꢀ1
5
l
ꢀ
1
2
2
SiO support. Sixty-four scans were averaged with 8 cm
resolution.
1
2
2
2.5. Catalysis
2
2
Reaction flasks were prepared inside the glove box, where the
catalysts were stored to prevent contact with atmosphere. A 6-
mL screw-cap flask was loaded with 30 mg of the catalyst, 4 mL
acetonitrile, 200
4 mmol of substrate (400
The reaction flask was sealed and moved outside the glove box just
immediately prior to the injection of aqueous H (200 L) to
start the reaction. Initial cyclohexene and H concentrations
5
glove box. The color of the Ta complexes was gradually transferred
from the solutions to the solid. After 24 h, the solid was filtered and
lL
dichlorobenzene (internal standard), and
washed with 300 mL of hot anhydrous toluene under N
2
and dried
l
L cyclohexene or 250 L cyclooctene).
l
under dynamic vacuum at 25 °C overnight. When dry, all solids
acquired the color of the soluble complex, with the exception of
2
O
2
l
ꢃ
Cal6Ta(OSiꢄ)
2
, which became green and Cp TaO(OSiꢄ)
2
, which
2 2
O

N. Morlanés, J.M. Notestein / Journal of Catalysis 275 (2010) 191–201
193
were 0.82 M and 0.41 M, respectively, giving a 50% maximum con-
version of cyclohexene. Analogous cyclooctene experiments were
one-pot procedures in controlled atmosphere conditions to pre-
vent premature hydrolysis of the Ta chloride precursors. The syn-
theses of all calixarene–Ta(V) complexes are known to the
literature to be very high yielding, and thus no effort was made
to purify the intermediates, making this catalyst synthesis method
general and scalable.
performed with 20 mg catalyst, 2 mL acetonitrile, 100
benzene, 250 L cyclooctene, and 130 L aqueous H
flasks were stirred and heated at 65 °C. Aliquots (10
l
L dichloro-
. Reaction
l
l
2
O
2
lL) were re-
moved periodically using a syringe, filtered to remove the catalyst,
and analyzed using a Shimadzu 2010 GC system equipped with a
flame ionization detector using a TR-1 capillary column.
Although understanding the atomic connectivity of supported
metal oxides remains an area of intense research, a structure for
2
SiO -grafted calixarene–Ta(V) is proposed based on known molec-
ular species. In particular, the reaction between Cal4TaCl and ex-
cess sodium phenoxide coordinates two monomeric phenolates
3
. Results
[
27]. By analogy, it is proposed that the surface Ta coordinates
3
.1. Materials synthesis and characterization
ꢀ
two SiO species on the surface, and three of the calixarene pheno-
lates. Depending on the environment and ligand, a calixarene phe-
nol or ether oxygen, or H
5
The synthesis of grafted calixarene–Ta(V) and TaX complexes
O or oxidized products, will complete the
is depicted in Schemes 1 and 2. All catalysts were prepared using
2
ꢃ
Scheme 1. Calixarene–Ta(V) syntheses and grafting on SiO
2
. (a) Cal4, (b) Cal4TaCl, (c) dmCal4, (d) dmCal4TaCl
2
, (e) Cal6, (f) Cal6TaCl
2
, (g) SCal4, (h) SCal4TaCl, (i) Cp TaCl
4
, (j)
ꢃ
Cpꢃ
Cal4TaCp , (k) generic calixarene–Ta species on surface, Cal4Ta(OSiꢄ)
2
from b, dmCal4Ta(OSiꢄ)
2
from d, Cal6Ta(OSiꢄ)
2
from f, SCal4Ta(OSiꢄ)
2
from h and Cal4Ta (OSiꢄ)
2
ꢃ
ꢃ
from j. Pentamethylcyclopentadienyl Ta grafting directly produces (m) Cp TaCl4ꢀx(OSiꢄ)
x
, which forms (n) Cp TaO(OSiꢄ)
2
upon air exposure. Regardless of precursor, (l)
TaO(OSiꢄ)
3
are oxo-Ta species formed after heat treatment. The structure of generic supported calixarene–Ta(V) (k) is by analogy to known molecular species [27]. The
structure of heat-treated catalyst (l) is as proposed for traditionally synthesized, supported Ta oxides [28,29].

1
94
N. Morlanés, J.M. Notestein / Journal of Catalysis 275 (2010) 191–201
Table 1
p
O
l
Tantalum (ICP) and ligand (TGA) contents for catalysts synthesized with 0.02–
1.0 mmol Ta per g SiO and various precursors. Ligand loadings are accurate to
0.02 mmol/g. Ta loadings are reproducible to 0.2 wt%.
X
X
X
X
X
X
SiO2
TaCl5
Ta(bz)5
Ta(dma)5
2
Ta
O
Ta
or
Ta
toluene,
O
X
O
O
air
O
O
reflux, 24h
Catalyst
Ta in
mmol
Ta ICP
analysis
(%)
Ta:ligand
synthesis
ligand/
(
%)
g
catalyst
Scheme 2. Grafting TaX
5
on SiO
2
directly forms (p) TaX
4
(OSiꢄ) or TaX
3
(OSiꢄ)
2
,
0
0
0
0
0
0
0
0
1
0
.2-TaCl5ꢀx(OSiꢄ)
x
3.5
3.1
3.3
0.4
0.9
1.7
2.4
3.1
15.4
3.1
2.9
3.1
3.3
3.0
–
–
–
3.4
2.5
0.5
0.5
0.8
1.8
2.6
2.9
3.3
3.3
3.3
3.3
3.5
2.4
–
–
–
which forms the surface oxo–Ta complex (l) upon standing in moist air. The
structure of heat-treated catalyst (l) is as proposed for other supported Ta oxides at
low coverage [28,29].
.2-(dma)5ꢀxTa(OSiꢄ)
x
.2-(Bz)5ꢀxTa(OSiꢄ)
x
.02-Cal4Ta(OSiꢄ)
.05-Cal4Ta(OSiꢄ)
.10-Cal4Ta(OSiꢄ)
.15-Cal4Ta(OSiꢄ)
2
2
2
2
0.05
0.06
0.09
0.16
0.19
0.19
0.19
0.16
0.18
0.19
0.13
n/a
n/a
1.1
0.9
0.9
0.9
1.0
1.1
1.0
1.0
1.0
coordination sphere. Before exposure to any moisture, grafting
TaX complexes likely results in a mixture of Ta species connected
5
to the surface by one or two bonds [1,4].
Calixarene and Cp contents were determined from the sharp
mass loss near 350 °C in thermogravimetric analysis in O
.2-Cal4Ta(OSiꢄ)
.0-Cal4Ta(OSiꢄ)
2
2
.2-dmCal4Ta(OSiꢄ)
2
ꢃ
0.2-Cal6Ta(OSiꢄ)
2
(Fig. 1
0.2-SCal4Ta(OSiꢄ)
2
2
ꢃ
0
0
.2-Cp TaO(OSiꢄ)
2
and Supplementary material, Fig. S1). The organic content was esti-
mated from mass losses assuming complete loss of the calixarene
Cpꢃ
.2-Cal4Ta (OSiꢄ)
2
ꢃ
or Cp fragment shown in Scheme 1, and gain of two oxygen atoms
ꢀ
1
to form TaO(OSiꢄ)
3
.
This corresponds to 612 g mol
for
ꢀ1
ꢀ1
Cal4Ta(OSiꢄ)
2
, 641 g mol for dmCal4Ta(OSiꢄ)
2
, 936 g mol for
detectable calixarenes are attached to the surface under these
conditions.
ꢀ
1
ꢀ1
Cal6Ta(OSiꢄ)
2
, 684 g mol
for SCal4Ta(OSiꢄ)
2
, 135 g mol
for
ꢃ
ꢀ1
Cpꢃ
Cp TaO(OSiꢄ)
2
, and 779 g mol
for Cal4Ta (OSiꢄ)
2
. In this
experiment, TaCl5ꢀx(OSiꢄ)
x
,
(dma)5ꢀxTa(OSiꢄ)
x
and (Bz)5ꢀx
-
3.2. Materials spectroscopic characterization
Ta(OSiꢄ)
x
show negligible mass loss compared to SiO due to the
2
instability of these ligands under air atmosphere. Hydrolysis is ex-
pected to occur after only brief exposure of these supported com-
pounds to air. Table 1 compares ligand content determined by TGA
and Ta content determined by ICP (Supplementary material, Tables
S1 and S2) for the calixarene-containing catalysts. The table shows
Ta:ligand ratios of approximately 1:1 for all samples except in the
low-loaded catalysts, where small mass losses do not allow suffi-
ciently precise measurement of the ligand content. This equi-
molecular ratio is consistent with intact grafting of species that
deposit as monomers.
Solution and diffuse reflectance UV–visible (DRUV–vis) spec-
troscopies were used to characterize the precursors and solid cat-
alysts, respectively. For d metals, the energy of the ligand to metal
charge transfer (LMCT) band is dependent on the local metal co-
ordination environment. In general, a red-shift in the LMCT band
occurs upon increasing the coordination number of the metal
center or agglomeration via bridging oxygen ligands [11,28,29].
One advantage of the grafted calixarene–Ta(V) complexes is that
the molecular precursor contains most of the Ta–O connectivity
of the final catalyst and is thus a spectroscopic model for the active
0
ꢀ4
Near quantitative grafting of Ta and calixarene is observed up to
site prepared after grafting. In 10 M toluene solution, the starting
calixarene ligands have an absorption maximum at 270–290 nm
(295–305 nm for SCal4) arising from
no absorption features above 350 nm. calixarene–Ta(V) complexes
ꢀ2
0
.2 mmol per g SiO
interactions between Ta, surface, and calixarene. At larger amounts
of Ta complex in the synthesis solution (1 mmol Ta per g SiO ), Ta
contents in the final catalyst do not pass the maximum geometrical
2
(0.25 groups nm ) which indicates strong
ꢃ
p–p
transitions and have
2
ꢃ
have the same
p
–
p
transitions as well as a broad phenol-to-Ta
ꢀ2
surface density of calixarenes (at ꢂ0.25 groups nm ), consistent
LMCT centered near 350 nm (Fig. 2A). This band is very weak for
SCal4TaCl. Related transitions are also weak for Cp TaCl and
ꢃ
with previous observations of Ti-containing calixarenes [14] and
4
ꢃ
consistent with a Ta–calixarene interaction that is stronger than
Cal4TaCp (Fig. 2B). TaCl5 (OSiꢄ) , (dma) Ta(OSiꢄ) , and
ꢀx
x
5ꢀx
x
Cpꢃ
the Ta–surface interaction in refluxing toluene. Cal4Ta (OSiꢄ)
2
(Bz)5 Ta(OSiꢄ) show negligible absorbance above 300 nm.
ꢀx
x
grafts at a lower surface density than other species, consistent with
DRUV–vis spectra of the calixarene-containing catalysts (Fig. 3)
ꢃ
hindered loss of the large Cp ligand. In the absence of Ta, no
show three features: charge transfer of the bare oxide at ꢂ230 nm,
Fig. 1. TGA for calixarene–Ta(V) grafted catalysts with 0.2 mmol Ta per g SiO
2
. Mass loss has been scaled to 150 °C to account for solvent or weakly adsorbed fragments of
ꢃ
monomeric precursors. From top to bottom: SiO
2 2
, (h) 0.2-SCal4Ta(OSiꢄ) , (b) 0.2-Cal4Ta(OSiꢄ)
2
, (d) 0.2-dmCal4Ta(OSiꢄ)
2
, (f) 0.2-Cal6Ta(OSiꢄ)
2
2
, (n) 0.2-Cp TaO(OSiꢄ) , and
Cpꢃ
(
j) 0.2-Cal4Ta (OSiꢄ)
2
. Letters corresponds to the catalyst precursors in Scheme 1.

N. Morlanés, J.M. Notestein / Journal of Catalysis 275 (2010) 191–201
195
ꢀ
4
Fig. 2. UV–visible absorption spectra for synthesized tantalum compounds in toluene 10 M. (d) 0.2-dmCal4TaCl
2
, (f) 0.2-Cal6TaCl
2
, (b) 0.2-Cal4TaCl, (h) SCal4TaCl, (i)
ꢃ
ꢃ
4
Cp TaCl , and (j) 0.2-Cal4TaCp . Letters corresponds to the catalysts precursors in Scheme 1.
Fig. 3. DRUV–vis spectra for calixarene–Ta(V) complexes grafted on silica surface. (f) 0.2-Cal6Ta(OSiꢄ)
2
, (d) 0.2-dmCal4Ta(OSiꢄ)
2
, (h) SCal4Ta(OSiꢄ)
2
, (b) 0.2-Cal4Ta(OSiꢄ)
2
,
ꢃ
Cpꢃ
ꢃ
(
l) 0.2-Cal4Ta(OSiꢄ)
2
-c, (n) 0.2-Cp TaO(OSiꢄ)
2
, (j) 0.2-Cal4Ta (OSiꢄ)
2
and (l) 0.2-Cp TaO(OSiꢄ)
2
-c. Letters correspond to catalysts precursors in Scheme 1.
ꢃ
the
5
p
–p
of the ligand at ꢂ290 nm, and LMCT that extend to 400-
500 °C have identified numerous surface species: crystalline
Ta with peak maxima at 265–270 nm, non-crystalline tantala
at 260 nm, and isolated terminal tantalum-oxo structures at
220–235 nm [28,29]. The edge energies of Ta films were re-
ported to be 4.18 eV (297 nm) [30]. Thus, the peak maxima at
ꢂ230 nm for heat-treated catalysts here indicates that these spe-
cies are isolated terminal tantalum-oxo structures with minimal
00 nm, which are more intense than for the precursor structures.
These latter features are used diagnostically for calixarene-to-Ta
connectivity. The as-synthesized forms of TaCl5ꢀx(OSiꢄ)
dma)5ꢀxTa(OSiꢄ) and (Bz)5ꢀxTa(OSiꢄ) are insufficiently air-sta-
2 5
O
x
,
2 5
O
(
x
x
ble to provide meaningful diffuse reflectance UV–visible spectra
in this experiment. The position of the calixarene–Ta LMCT edge
is extrapolated from the linear region to the intercept (Table 2).
The LMCT edge energies of the chloride-containing precursors are
lower than those of the resulting catalysts, consistent with loss
of halide upon grafting and confirming that the Ta precursor has
been covalently grafted to the surface. No significant or systematic
changes in LMCT edge energies were observed when comparing
two different Ta loadings, providing first evidence that these cata-
lysts consist of a single type of active site unaffected by surface
density. Neither were large nor systematic differences observed
in the edge energy for different calixarene precursors used, indicat-
ing electronically similar sites in all catalysts. Heating the catalysts
in dry air at 550 °C for 6 h removed all absorption features above
2 5
crystalline or non-crystalline Ta O . This is consistent with an as-
synthesized structure of an isolated Ta atom protected by a calix-
arene ligand. Table 2 also shows the position of the absorption
maximum and the edge energy for all catalysts after heat treat-
ment. Additional DRUV–vis spectra are included in the Supplemen-
tary material (Figs. S2–S4).
DRIFTS of 0.2-Cal4Ta(OSiꢄ)
2
, 0.2-Cal4Ta(OSiꢄ)
2
-c and calixa-
rene covalently grafted to SiO
2
in the absence of Ta [31] (Supple-
mentary material, Fig. S5) show evidence of the intact calixarene
ligand as synthesized and its complete removal by heat treatment.
1
3
The solid state C CP/MAS NMR of 0.2-Cal4Ta(OSiꢄ)
2
(Supplemen-
tary material, Fig. S6) also confirms that the calixarene ligand was
grafted intact. As observed previously for Ti-based catalysts [14], a
resonance at ꢂ158 ppm for the ispo carbon indicates calixarene–Ta
coordination on the surface.
2
90 nm, confirming that those electronic transitions arise from
the calixarene or Cp ligands or from the ligand-Ta connectivity.
Previous assignments of supported Ta catalysts treated above
ꢃ

1
96
N. Morlanés, J.M. Notestein / Journal of Catalysis 275 (2010) 191–201
Table 2
Absorption edge energies for the lowest energy LMCT band for Ta–calixarenes in
toluene solution, after grafting on SiO with different Ta:SiO ratios and calixarene–
complexes, and after heat treatment for 0.2 mmol Ta g catalysts.
which is catalyzed by both Brønsted and Lewis acids. These path-
ways are illustrated in Scheme 3. A key metric therefore is the
pathway selectivity to epoxide via radical or direct routes, which
2
2
ꢀ
1
Ta(V) and TaX
Precursor
5
1 1 2
is the ratio of apparent rate constants S = k /(k + k ). Selectivity
Edge (eV) Grafted on silica, edge (eV)
Treated at
550 °C
to direct epoxidation or non-radical processes is defined in Eq.
ꢀ
4
in 10
toluene
M
(
1) from the ratios of concentrations of various products.
0
.05 mmol Ta
0.2 mmol Ta Edge
Max
solution
per g SiO
2
per g SiO
2
(eV)
(nm)
epoxide þ diol ꢀ cyclohexenol
TaCl
Ta(dma)
Ta(Bz)
Cal4TaCl
dmCal4TaCl
Cal6TaCl
5
–
–
–
3.1
3.0
3.1
3.6
2.9
3.0
–
–
–
3.4
3.3
3.3
3.3
3.1
3.1
–
–
–
3.4
3.4
3.2
3.6
3.0
3.0
4.7
4.9
4.8
4.8
4.9
5.0
4.7
4.8
4.8
240
228
230
225
229
222
231
228
232
Direct Epoxide Selectivity ¼
ð1Þ
all C6 products
5
5
In parallel, these catalysts can also decompose H
2
O
2
without
generating product, accounting for lower than maximum conver-
sion of cyclohexene (maximum conversion is ꢂ50% under these
conditions).
2
2
SCal4TaCl
ꢃ
Cp TaCl
4
First, dmCal4Ta(OSiꢄ)
compared under identical conditions in cyclohexene epoxidation
with H (Table 3, entries 8 vs. 9). Ti–SiO is a ubiquitous catalyst,
2
and dmCal4Ti(OSiꢄ) catalysts were
ꢃ
Cal4TaCp
2
O
2
2
and dmCal4Ti(OSiꢄ) has been shown previously to be an extre-
mely active and selective catalyst in cyclohexene epoxidation with
organic hydroperoxides [14]. Here, the Ti-containing catalyst is
slightly more productive than the Ta-based catalyst with respect
to total cyclohexene conversion. However, the initial direct epoxi-
dation turnover rates are higher for Ta due to a much higher direct
epoxide selectivity (90% for Ta vs. 21% for Ti). The Ta catalyst also
shows much lower hydrolysis activity. This substantial improve-
ment in epoxide yield when compared to a workhorse catalyst
motivates a more extensive investigation into these materials.
3
.3. Catalytic behavior
The catalytic activity of these materials was probed with cyclo-
hexene and cyclooctene oxidation with aqueous H in aceto-
nitrile. For all supported calixarene–Ta catalysts tested here, the
selectivity to cyclooctene epoxide is higher than 98%. Allylic oxida-
tion and epoxide hydrolysis to the trans-diol are both negligible, as
is typical for cyclooctene in similar supported d metal oxide cata-
lysts [32]. Oxidation rates are similar for all calixarene–Ta(V) cata-
5
lysts but higher than for TaX -type precursors. Yields are given in
the supporting data (Table S3) and a full kinetic analysis will be
published elsewhere.
Table 3 shows initial rates and the 3-h conversions, turnover
numbers, and product distributions at 3 h for cyclohexene oxida-
2 2
tion with aqueous H O in acetonitrile for 0.2 mmol g Ta and Ti
catalysts. Full data for other metal loadings are included in the
Supplementary material (Tables S4 and S5). In contrast to cyclo-
octene, cyclohexene oxidation typically yields
products under similar conditions with other supported d metal
oxide catalysts [33], making this substrate more useful for
detecting patterns of reactivity among different catalysts.
2 2
O
0
Cal4Ta(OSiꢄ)
Ta loadings of 0.02, 0.05, 0.15, and 0.2 mmol per g SiO
catalysts have similar initial direct epoxidation
2
and TaCl5ꢀx(OSiꢄ)
x
catalysts were compared for
2
. All
Cal4Ta(OSiꢄ)
2
ꢀ1 ꢀ1
rates (r
i
= (molepoxide + moldiol ꢀ molcyclohexenol) molTa
s ) up to
the maximum surface density of Ta possible. Low Ta loadings also
give nearly identical turnovers to products as a function of time
over 3 h. (Fig. 4 and Table 4; complete product distributions and
kinetic data in the Supplementary material Table S4 and Fig. S6).
ꢀ1
a number of
2
The maximally loaded catalyst 0.2-Cal4Ta(OSiꢄ) maintains the
0
same initial turnover rate but uniquely shows a drop off in
reactivity, the mechanistic origin of which is under investigation.
Table 4 shows that selectivity to direct epoxidation through the
non-radical route increases with calixarene–Ta(V) surface density.
Given that direct epoxidation rates are constant with surface
density, this is equivalent to decreasing radical oxidation with
increasing surface density. To check the role of lower-selectivity
background reactions that may be competitive at the lower total
conversions of the low surface density catalysts, an experiment
Following reaction networks postulated for Ti–SiO
2
oxidation
[
33], allylic oxidation produces cyclohexenyl hydroperoxides, and
this unstable molecule can oxidize cyclohexene to generate cyclo-
hexane epoxide and cyclohexanol, or it can decompose to cyclo-
hexanone. Alternately, cyclohexane epoxide can be formed
directly via a direct pathway catalyzed by Lewis acids. Epoxide
generated from either pathway can hydrolyze to the trans-diol,
2
was run with 250 mg of 0.02-Cal4Ta(OSiꢄ) to give a Ta content
Table 3
2
Cyclohexene epoxidation with Ta and Ti-grafted catalysts with 0.2 mmol Ta per g SiO .
Catalyst
Conv (%)a
TONb
Direct epoxide initial rate
Selectivity
Direct epoxide selectivityc
ꢀ
1
ꢀ1
2
mol molTa
s
ꢁ 10
Epoxide
Diol
Anone
Enol
Enone
0
0
0
0
0
0
0
0
0
0
0
0
.2-TaCl5ꢀx(OSiꢄ)
x
6.7
8.9
2.2
8.7
9.9
39
51
15
53
55
56
54
58
64
83
95
34
1.7
1.2
1.2
3.6
3.9
3.1
2.5
3.8
2.0
3.1
5.4
1.1
39
45
8
59
41
12
15
28
6
33
38
67
30
52
85
78
66
40
59
82
17
9
5
1
1
2
1
0
1
29
1
9
8
11
6
3
2
3
3
24
5
3
11
5
12
4
3
1
4
3
1
7
63
75
63
83
89
95
89
90
21
91
87
53
.2-(dma)5ꢀxTa(OSiꢄ)
x
.2-(Bz)5ꢀxTa(OSiꢄ)
x
.2-Cal4Ta(OSiꢄ)
.2-Cal4Ta(OSiꢄ)
.2-Cal6Ta(OSiꢄ)
2
2
2
-cap
9.9
.2-SCal4Ta(OSiꢄ)
2
10.2
10.1
12.3
16.8
13.4
6.1
.2-dmCal4Ta(OSiꢄ)
2
.2-dmCal4Ti(OSiꢄ)
ꢃ
.2-Cp TaO(OSiꢄ)
2
28
9
34
Cpꢃ
.2-Cal4Ta (OSiꢄ)
2
1
2
6
2
.2-Cal4Ta(OSiꢄ)
2
-c
45
a
b
c
Reaction conditions: catalyst = 30 mg, Vtotal = 4.6 ml, [cyclohexene] = 0.82 M, [H
TON: mmol oxidation products/mmol Ta.
Direct epoxide selectivity = 100 ꢁ (epoxide + diol ꢀ cyclohexenol)/(epoxide + diol + cyclohexenol + cyclohexenone + cyclohexanone).
2
O
2
] = 0.41 M, T = 65 °C, t
R
= 180 min.

N. Morlanés, J.M. Notestein / Journal of Catalysis 275 (2010) 191–201
197
2 2
Scheme 3. Proposed mechanism for cyclohexene oxidation with H O over Ta grafted catalysts.
Fig. 4. Turnover numbers vs. reaction time for Cal4Ta(OSiꢄ)
Cal4Ta(OSiꢄ) , s 0.20-Cal4Ta(OSiꢄ) . (B) ꢀ 0.05-TaCl5ꢀx(OSiꢄ)
cyclohexene] = 0.82 M, [H ] = 0.41 M, T = 65 °C.
2
and TaCl5ꢀx(OSiꢄ)
x
catalysts with different Ta content. (A) ꢀ 0.02-Cal4Ta(OSiꢄ)
, N 0.15-TaCl5ꢀx(OSiꢄ) , s 0.20-TaCl5ꢀx(OSiꢄ)
2
, h 0.05-Cal4Ta(OSiꢄ)
2
, N 0.15-
2
2
x
, h 0.1-TaCl5ꢀx(OSiꢄ)
x
x
x
. Reaction conditions: Cat. = 30 mg,
[
2 2
O
in the reactor equivalent to that for the 0.15-Cal4Ta(OSiꢄ)
2
trial.
rates and conversions under these conditions. There is some varia-
tion of initial direct epoxidation rates within the series, but all
This increased cyclohexene conversion to 10%, but maintained di-
rect selectivity constant at 76%. For all reactions, direct epoxide
selectivities held constant from 10 min to 3 h, regardless of conver-
sion, indicating that the reaction conditions themselves were not
responsible for the observed behavior. Thus, the presence of calix-
arene at increasing surface coverage is shown to suppress radical
reaction pathways.
ꢃ
calixarene- and Cp -containing catalysts give ꢂ90% selectivity to
5
direct epoxidation, whereas materials derived from TaX -type pre-
cursors give <75% direct epoxidation selectivity. Furthermore, heat
treatment of the calixarene-containing catalysts, e.g. to produce
0.2-Cal4Ta(OSiꢄ)
2
-c, dramatically reduces initial rates, total con-
versions, and selectivity to direct epoxidation. This result demon-
strates that, rather than potentially limited by steric congestion
at the metal, rates are enhanced in the presence of the organic li-
gand. Perhaps more importantly, the presence of the ligand during
turnover decreases radical reactivity and enhances non-radical
reactivity, with a total increase in the direct selectivity.
5
In contrast, materials prepared with different TaCl contents on
SiO showed initial direct epoxidation rates and 3-h turnovers that
2
decreased markedly with increasing Ta loading (Fig. 4 and Table 4).
The rates and selectivities to direct epoxidation are also substan-
tially lower than those of Cal4Ta(OSiꢄ)
2
, and direct epoxidation
selectivities are similar to the blank experiment.
2 2
The decomposition of H O over Ta catalysts was studied by
Different calixarenes and Ta precursors were next compared at
iodometric titration [34] to further understand the reactivity of
this system. In the control experiment with neither catalyst nor
Ta contents of 0.2 and 0.05 mmol per g SiO
tent (Table 3), calixarene- and Cp -containing catalysts exhibit 2-3
times higher initial epoxidation rates than catalysts derived from
2
. At the higher Ta con-
ꢃ
SiO
concentration after 6 h at 65 °C. With only SiO
within 93%. With 0.2-Cal4Ta(OSiꢄ) , H remained within 89%.
2
, the concentration of H
2
O
2
remained within 98% of the initial
2
, H remained
2 2
O
TaX
5
5
precursors. Ta(Bz) -derived catalyst gives particularly low
2
2 2
O

1
98
N. Morlanés, J.M. Notestein / Journal of Catalysis 275 (2010) 191–201
Table 4
Cyclohexene epoxidation results as a function of Ta content.
Loading (mmol g 1
ꢀ
)
TaCl5ꢀx(OSiꢄ)
Cal4Ta(OSiꢄ)
x
2
Conv. (%)a
TONb
Direct epoxide
initial ratec
Direct epoxide
selectivityd
Conv. (%)
a
TONb
Direct epoxide
initial rate
Direct epoxide
selectivity
c
d
Blank
0.8
0.8
–
–
–
–
54
50
0.8
0.8
3.8
10.0
54
50
73
76
0.00
0.02
0.02
0.05
0.10
0.15
0.20
163
73
3.6
2.9
e
4.1
7.5
8.4
6.7
129
95
55
3.0
2.7
2.5
1.7
55
63
57
63
7.3
12.3
9.9
165
143
55
4.0
4.0
3.9
80
90
89
39
a
b
c
2 2 R
Reaction conditions: catalyst = 30 mg, Vtotal = 4.6 ml, [cyclohexene] = 0.82 M, [H O ] = 0.41 M, T = 65 °C, t = 180 min.
TON: mmol oxidation products/mmol Ta.
ꢀ1 ꢀ1
2
Rate as mol molTa
s
ꢁ 10 .
d
e
Direct epoxide selectivity = 100 ꢁ (epoxide + diol ꢀ cyclohexenol)/(epoxide + diol + cyclohexenol + cyclohexenone + cyclohexanone).
2 2 R
Reaction conditions: Catalyst = 250 mg, Vtotal = 4.6 ml, [cyclohexene] = 0.82 M, [H O ] = 0.41 M, T = 65 °C, t = 180 min.
Finally, over 0.2-Cal4Ta(OSiꢄ)
2
-c, the H
2
O
2
concentration fell to
4 vs. 5). These results demonstrate that while bulk hydrophobicity
effects can control the rates of hydrolysis by minimizing the role of
exposed SiOH, these effects are independent of those of the calix-
arene-containing precursors that directly alter the active site
chemistry.
7
0% of the initial value, clearly showing that coordination of the
calixarene ligand inhibits radical peroxide decomposition relative
to the bare oxide catalyst.
Results obtained in cyclohexene epoxidation with low surface
coverage catalysts are given in the Supplementary material
(
Table S5). The lower surface coverages decrease the selectivity
3.4. Catalytically relevant species and stability
to direct epoxidation, but also decrease the extent of subsequent
epoxide hydrolysis. While overall conversions at completion are
lower at these lower surface densities, 3-h turnover numbers to
The catalytic behavior of dmCal4Ta(OSiꢄ) was compared to
2
that of the dmCal4TaCl homogeneous complex in solution.
product are higher. By comparing these data with those in Table 3,
Fig. 5A shows that turnover numbers for the solid catalyst are sig-
nificantly higher than for the same compound in solution. Indeed,
the homogeneous precursor gives comparable initial rates, but
deactivates completely within 10 min. Further control experiments
confirmed that the active catalyst site resides on the solid surface
and that it is stable against leaching during the reaction. Fig. 5B
shows that catalyst removal by hot filtration from the reactor after
60 min immediately halts cyclohexene conversion. In another
Cpꢃ
Cal4Ta(OSiꢄ)
2
and Cal4Ta (OSiꢄ)
2
are both seen to have initial
direct epoxidation rates that are largely insensitive to surface
density.
Catalysts 0.20-Cal4Ta(OSiꢄ)
treated with excess octanol (5 mmol per g catalyst) in refluxing tol-
uene for 24 h, then washed and dried to study the effect of SiO
2
and 0.05-Cal4Ta(OSiꢄ)
2
were
2
surface end-capping on reaction rates and selectivities [35,36].
After the surface modification, the catalyst weight increase in
TGA corresponded to 0.32 and 0.53 mmol octanol per g catalyst
experiment, 0.2-Cal4Ta(OSiꢄ) recovered after 12 h under stan-
2
dard reaction conditions showed only a nominal 0.2 wt% absolute
(
ꢂ9% and ꢂ14% of the hydroxyl groups on surface) for 0.2-
change in Ta content, a small increase in mass loss by TGA, and
no change to the DRUV–vis edge energy (Supplementary material,
Table S1 and Fig. S8), indicating the absence of leaching or ligand
loss.
To demonstrate that the decreasing rates with time in the batch
reactors are due to reactant depletion, and not catalyst deactivation
or poisoning by a product at typical reactor concentrations, an
Cal4Ta(OSiꢄ)
2
and 0.05-Cal4Ta(OSiꢄ)
2
, respectively (Supplemen-
tary material, Fig. S7). Accounting for the overall weight increase,
the capped catalysts show only minimal decreases in conversion,
turnovers, initial rates, and selectivity to direct epoxidation after
capping, while the extent of hydrolysis, defined as moldiol
(
/
moldiol + molepoxide), decreases substantially from ꢂ55% to ꢂ30%
(Table 3, entry 4 vs. 5 and Supplementary material, Table S5, entry
additional charge of H O (200 lL) was added after 3 h to standard
2
2
Fig. 5. Cyclohexene epoxidation turnovers or conversion with 0.2-Cal4Ta(OSiꢄ)
2
solid catalyst (ꢀ) and (A) dmCal4TaCl
2
in solution (s), (B) with filtration of the catalysts after
2
or 0.007 mmol 0.2-dmCal4TaCl ,
one hour reaction (s), or (C) with H
2
O
2
addition (200
l
L) after 3 h reaction (s). Reaction conditions: 30 mg 0.2-Cal4Ta(OSiꢄ)
2
[
2 2
cyclohexene] = 0.82 M, [H O ] = 0.41 M and T = 65 °C.

N. Morlanés, J.M. Notestein / Journal of Catalysis 275 (2010) 191–201
199
reaction conditions. Fig. 5C shows that after the second injection of
, product formation rates are similar to those at the start of the
reaction.
Together, these results indicate that calixarene–Ta(V) catalysts
epoxide hydrolysis. The catalysts studied in the current work also
exhibit a range of extents of epoxide hydrolysis. A parametric plot
for all catalysts (Fig. 6) shows that catalysts that promote direct
oxidation generally have higher extents of hydrolysis, defined as
2 2
H O
are stable with respect to ligand loss and inhibition/poisoning
and do not leach into solution either as active or inactive homo-
geneous species. Active species are permanently immobilized
under reaction conditions reported here, and immobilization of
both the calixarene ligand and the Ta atom are required to observe
optimal catalyst behavior.
moldiol/(moldiol + molepoxide) at 3 h. For example, Cal6Ta(OSiꢄ)
2
has the highest selectivity to direct epoxidation and highest diol
selectivity at 3 h. This trend is consistent with the same Lewis acid
site leading to direct epoxidation and to hydrolysis, but also to the
simple presence of increased concentrations of epoxide over more
Lewis acidic catalysts, which will increase hydrolysis regardless of
mechanism. The latter is supported by increasing extents of hydro-
lysis with time (conversion). As shown here, modifying the surface
by end-capping of hydroxyl groups reduces the extent of hydro-
lysis, but has a weaker and slightly detrimental effect on selectivity
to direct epoxidation, indicating that the radical and non-radical
routes to epoxide are both controlled by the same Ta site, which
is influenced primarily by directly attached ligands.
4
. Discussion
The proposed reaction pathway for cyclohexene epoxidation
with H O on these catalysts was introduced in Scheme 3 [33].
2 2
All epoxidation steps, direct by Lewis acids or an indirect radical
route via allylic hydroperoxides, are metal catalyzed. Under the
conditions studied here, allylic radicals are expected to be exclu-
sively generated at the metal center and to have very short life-
times in solution. H O decomposition occurs both in solution
2 2
and at the metal center. Epoxide hydrolysis is catalyzed by Lewis
and Brønsted acids, and this may occur at the Lewis acidic Ta site
The changes in reactivity for the different calixarene ligands are
likely due to a combination of different steric bulk, different num-
bers of potentially coordinating groups (phenols, phenyl ethers,
and thioethers), and different local H-bonding in the vicinity of
the Ta cation. The ability to control the coordination environment
at the grafted metal without perturbing the isolated nature of the
active site is a unique feature of calixarene ligands for creating
grafted oxide catalysts [40]. The different direct epoxidation rates
and selectivities for different calixarenes, and the increasing direct
epoxidation selectivities with surface density are not accompanied
by correlated changes in the energy of relevant LMCT bands in the
DRUV–vis spectra. Thus, the different calixarenes, which are all
contain tert-butylphenol subunits, do not differently alter the
2
or the Brønsted acidic SiO surface. Thus, control of the metal envi-
ronment is the key to designing better catalysts for this reaction.
Similar to that which was proposed for calixarene–Ti(IV) cata-
lysts, the combination of the bulky multidentate calixarene ligand
and the oxide surface improves catalyst reactivity, in large part by
preventing the formation of larger, less active oxide clusters, espe-
cially as Ta-surface density increases and under conditions of turn-
over when monodentate ligands would be displaced by ligand
exchange [14]. However, the differences between calixarene-con-
taining and bare Ta–SiO catalysts in epoxidation with H O are
2 2 2
much more pronounced than observed for Ti-containing catalysts
and epoxidation with organic hydroperoxides.
Here, the higher rates of direct epoxidation for the grafted calix-
arene–Ta when compared to the bare oxide may reflect the role of
the ligand in increasing the local hydrophobicity, thus decreasing
0
Lewis acidity for these d cations. This is in agreement with previ-
ously reported deperoxidation rates on SiO
complexes, which showed no effect of the number of bonds to
SiO or the identity of the alkoxo ligands [41]. Elsewhere, calix-
2
-supported alkoxo–Ta
2
arene para-substitution had little effect on edge energies of modi-
fied photoluminescent materials [42].
The reactivity differences observed here are proposed to be
strongly influenced by the different H-bonding environments in
these different materials. Although no role of para-substitution of
grafted calix[4]arene–Ti complexes was seen for the epoxidation
rate of cyclohexene with organic hydroperoxides [43], changing
the macrocycle structure as done here may be a more dramatic
2
the binding of H O, which is a likely inhibitor. Likewise, the ligand
bulk may help prevent coordination of co-product alcohols or
especially the trans-cyclohexanediol, which would rigidly coordi-
nate and block reaction sites and is more electron-donating than
the calixarene phenoxide ligands that can partially delocalize oxy-
gen lone pairs into the ring [37]. The multidentate calixarene li-
gand may also distort the coordination environment away from
the lowest energy configuration that would be adopted by the bare
oxide, thus increasing reactivity. Spectroscopic support for an im-
proved understanding of the role of the ligand in tuning the active
site is under way.
change to the local environment at the coordinated metal. Control
+
over H transfer is known to be critical in H
2
O
2
activation down
homolytic or heterolytic pathways [38]. Each of the calixarenes
will have a different intrinsic pKa, and different active site surface
densities will alter the average pKa of the remaining surface
The greater selectivity toward direct epoxidation and decreased
100
2 2
H O decomposition relative to the ligand-free case are consistent
with the ligand-bearing Ta being more Lewis acidic than the pure
oxide under reaction conditions. In other studies, increasing Lewis
acidity of metal centers enhances the rate of heterolytic relative to
8
0
0
0
0
homolytic cleavage of the O–O bond of coordinated H
enhancing the rate of olefin epoxidation relative to H
sition [38]. In sealed reactors such as in these experiments, homo-
lytic decomposition of H generates O that has been proposed
2
O
2
, as well
6
4
2
2 2
O decompo-
2
O
2
2
to be responsible for allylic oxidation of cyclohexene [33].
Others have modified supported Ti and Ta oxides with ligands
to control Lewis acidity; however, this has primarily resulted in
changes in total rates and extents of hydrolysis. Fraile et al. [39]
noted lower rates of hydrolysis for ligands that decreased the
LUMO energy, creating less Lewis acidic sites, and Ruddy et. al.
0
2
0
40
60
80
100
selectivity to direct epoxidation
[
11] created more Lewis acidic and more hydrophobic sites by
silanizing supported Ta oxides, which were then less prone to
Fig. 6. Relationship between hydrolysis and non-radical epoxidation processes.

2
00
N. Morlanés, J.M. Notestein / Journal of Catalysis 275 (2010) 191–201
gands, support modifications, and reactions are currently being ex-
plored for an improved understanding and expanded scope of
grafted Ta catalysts, which will be reported in due course.
O
O
o
O
O
R
Ta
Acknowledgments
O
-
O
N.M. acknowledges the Ministry of Science and Innovation in
Spain for a postdoctoral fellowship. J.M.N. acknowledges the
donors of the American Chemical Society Petroleum Research Fund
and Northwestern University for support of this research. ICP-OES
+
H O
2 2
-
H O
2
1
3
and C NMR were carried out at the Northwestern IMSERC facility.
The DRIFTS work was performed in the Keck-II facility of NUANCE
Center at Northwestern University, supported by NSF-NSEC, NSF-
MRSEC, Keck Foundation, the State of Illinois, and Northwestern
University.
O
O
O
o
O
O
O
o
O
O
O
H
H
R
R
H
H O
Ta
Ta
Appendix A. Supplementary material
O
O
O
O
Supplementary materials associated with this article can be
found, in the online version, at doi:10.1016/j.jcat.2010.07.010.
References
Scheme 4. Proposed cyclohexene epoxidation cycle on Ta–calixarene catalysts. An
equivalent mechanism can be proposed for breaking Ta–O–surface bonds for H
coordination.
2 2
O
[
1] D. Meunier, A. Piechaczyk, A. de Mallmann, J.M. Basset, Angew. Chem. Int. Ed.
8 (1999) 3540.
3
[
[
2] Y. Matsuo, K. Mashima, Angew. Chem. Int. Ed. 40 (2001) 960.
3] D.R. Mulford, J.R. Clark, S.W. Schweiger, P.E. Fanwick, I.P. Rothwell,
Organometallics 18 (1999) 4448.
silanols. At this time, it cannot be deduced which groups directly
coordinating to Ta (phenol or silanol) are formally protonated in
catalytically relevant steps when coordinating and activating
[
4] E. Le Roux, M. Chabanas, A. Baudouin, A. de Mallmann, C. Coperet, E.A.
Quadrelli, J. Thivolle-Cazat, J.M. Basset, W. Lukens, A. Lesage, L. Emsley, G.J.
Sunley, J. Am. Chem. Soc. 126 (2004) 13391.
H
2
O
2
, or whether outer-sphere effects of the general H-bonding
[5] M.C. Burland, T.W. Pontz, T.Y. Meyer, Organometallics 21 (2002) 1933.
[
[
[
6] K.J. Weller, P.A. Fox, S.D. Gray, D.E. Wigley, Polyhedron 16 (1997) 3139.
7] L.L. Anderson, J. Arnold, R.G. Bergman, Org. Lett. 6 (2004) 2519.
8] B.M. Weckhuysen, D.E. Keller, Catal. Today 78 (2003) 25.
environment override these local effects. Scheme 4 assumes that
one of the bonds to the calixarene is broken, reflecting this
proposal, but a surface bond could also be cleaved in catalytically
relevant steps by analogy to a previous result with grafted
calixarene–Ti in non-aqueous systems [43]. The mechanism
remains under investigation.
[9] J.M. Thomas, R. Raja, D.W. Lewis, Angew. Chem. Int. Ed. 44 (2005) 6456.
10] R.A. Sheldon, J. Mol. Catal. 7 (1980) 107.
11] (a) R.L. Brutchey, C.G. Lugmair, L.O. Schebaum, T.D. Tilley, J. Catal. 229 (2005)
[
[
72;
(b) D.A. Ruddy, T.D. Tilley, J. Am. Chem. Soc. 130 (2008) 11088.
12] (a) I. Nowak, M. Ziolek, Chem. Rev. 99 (1999) 3603;
[
(
b) D. Meunier, A. de Mallmann, J.M. Basset, Top. Catal. 23 (2003) 183;
5
. Conclusions
(c) C. Coperet, M. Chabanas, R.P. Saint-Arroman, J.M. Basset, Angew. Chem. Int.
Ed. 42 (2003) 156.
[
[
[
13] J.M. Notestein, A. Katz, Chem. Eur. J. 12 (2006) 3954.
14] J.M. Notestein, E. Iglesia, A. Katz, J. Am. Chem. Soc. 126 (2004) 16478.
15] N. de Silva, S.J. Hwang, K.A. Durkin, A. Katz, Chem. Mater. 21 (2009) 1852.
Epoxidation catalysts were obtained by grafting Ta compounds
on SiO
thesizes catalysts that are more active and selective for direct
non-radical) epoxidation than are monodentate TaX precursors.
2
. Using calixarene–Ta(V) complexes as the Ta source syn-
[16] (a) E. Hoppe, C. Limberg, B. Ziemer, Inorg. Chem. 45 (2006) 8308;
b) E. Hoppe, C. Limberg, B. Ziemer, C. Mugge, J. Mol. Catal. A 251 (2006) 34.
[
(
5
17] L.F.W. Armarego, C.L.L. Chai, Purification of Laboratory Chemicals, fifth ed.
Elsevier.
Combined TGA, ICP, and diffuse reflectance UV–visible spectro-
scopy show that isolated, 1:1 Ta:calixarene complexes are formed
on the surface and are stable under reaction conditions. The elec-
tronic structure and the reactivity per site were shown not to be
a function of surface density, consistent with a single type of iso-
lated catalyst site being present on the surface. Finally, the sites
combining calixarene ligand, Ta, and surface are more active in
epoxidation than either the ligand-free or the homogeneous cases.
It must be noted that while some of the same characteristics
[18] S. Groysman, I. Goldberg, M. Kol, Organometallics 22 (2003) 3793.
[
[
[
[
19] A. Casnati, A. Arduini, E. Ghidini, A. Pochini, R. Ungaro, Tetrahedron 47 (1991)
221.
2
20] B. Castellano, E. Solari, C. Floriani, N. Re, A. Chiesi-Villa, C. Rizzoli, Chem. Eur. J.
5 (1999) 722.
21] B. Castellano, A. Zanotti-Gerosa, E. Solari, C. Floriani, Organometallics 15
(
1996) 4894.
22] C. Floriani, R. Floriani-Moro, Adv. Organometall. Chem. 47 (2001) 167.
[23] C. Redshaw, M. Rowan, D.M. Homden, M.R.J. Elsegood, T. Yamato, C. Perez-
Casas, Chem. Eur. J. 13 (2007) 10129.
[
[
24] J.A. Acho, L.H. Doerrer, S.J. Lippard, Inorg. Chem. 34 (1995) 2542.
25] A. Conde, R. Fandos, A. Otero, A. Rodríguez, P. Terreros, Eur. J. Inorg. Chem. 19
(2008) 3062.
(
high non-radical selectivity, moderately high epoxidation rates)
ꢃ
are exhibited by Cp TaCl
4
as a precursor, this precursor is approx-
[
26] (a) O. Banakh, P.A. Steinmann, L. Dumitrescu-Buforn, Thin Solid Films 513
imately 4ꢁ the cost to synthesize Cal4TaCl. Moreover, the syn-
thetic diversity of the calixarene ligands opens up many more
possibilities for modification of catalytic species. Finally, and as
noted by others, Ta may be an ideal metal for Lewis acid catalysts
because it can coordinate five anionic ligands, allowing for multi-
dentate connection each to ligand and to surface for stability in
coordinating solvents, while still leaving an exchange site for
hydroperoxide [1]. This is experimentally bore out, as the grafted
calixarene–Ta complex was substantially more selective than the
equivalent grafted calixarene–Ti complex. Other calixarene li-
(
(
2006) 136;
b) J. Tauc, A. Menth, J. Non-Cryst. Solids 8-10 (1972) 569.
[27] A. Zanotti-Gerosa, E. Solari, L. Giannini, C. Floriani, A. Chiesi-Villa, C. Rizzoli,
Chem. Commun. (1997) 183.
[
28] (a) E.L. Lee, I.E. Wachs, J. Phys. Chem. C 111 (2007) 14410;
b) Y.S. Chen, J.L.G. Fierro, T. Tanaka, I.E. Wachs, J. Phys. Chem. B 107 (2003)
243.
(
5
[29] M. Baltes, A. Kytokivi, B.M. Weckhuysen, R.A. Schoonheydt, P. Van Der Voort,
E.F. Vansant, J. Phys. Chem. B 105 (2001) 6211.
[
[
30] X. He, J. Wu, L. Zhao, J. Meng, X. Gao, X. Li, Solid State Commun. 147 (2008) 90.
31] J.M. Notestein, A. Katz, E. Iglesia, Langmuir 22 (2006) 4004–4014.
[32] A. Batler, M.J. Clague, G.E. Meister, Chem. Rev. 94 (1994) 625.

N. Morlanés, J.M. Notestein / Journal of Catalysis 275 (2010) 191–201
201
[
[
33] (a) J.M. Fraile, J.I. Garcia, J.A. Mayoral, E. Vispe, J. Catal. 233 (2005) 90;
b) J.M. Fraile, J.I. Garcia, J.A. Mayoral, E. Vispe, Appl. Cat. A: Gen. 245 (2003)
63.
34] G.H. Jeffery, J. Bassett, J. Mendham, R.D. Denney, Vogel’s Textbook of
Quantitative Chemical Analysis, fifth ed., Wiley, NY, 1989.
35] J.M. Fraile, J.I. Garcia, J.A. Mayoral, E. Vispe, J. Catal. 189 (2000) 40.
36] D.A. Ruddy, T.D. Tilley, Chem. Commun. (2007) 3350.
37] T.W. Coffindaffer, B.D. Steffy, I.P. Rothwell, K. Folting, J.C. Huffman, W.E. Streib,
J. Am. Chem. Soc. 111 (1989) 4742–4749.
[39] J.M. Fraile, J.I. Garcia, J.A. Mayoral, L. Salvatella, E. Vispe, D.R. Brown, G. Fuller, J.
Phys. Chem. B 107 (2003) 519.
[40] J.M. Notestein, L.R. Andrini, V.I. Kalchenko, F.G. Requejo, A. Katz, E. Iglesia, J.
Am. Chem. Soc. 129 (2007) 1122–1131.
[41] R. Petroff Saint-Arroman, B. Didillon, A. Mallmann, J.M. Basset, F. Lefebvre,
Appl. Catal. A: Gen. 337 (2008) 78.
[42] J. Notestein, E. Iglesia, A. Katz, Chem. Mater. 19 (2007) 4998.
[43] J.M. Notestein, A. Solovyov, L.R. Andrini, F.G. Requejo, A. Katz, E. Iglesia, J. Am.
Chem. Soc. 129 (2007) 15585.
(
3
[
[
[
[
38] (a) N.A. Stephenson, A.T. Bell, J. Mol. Catal. A Chem. 275 (2007) 54;
(
b) N.A. Stephenson, A.T. Bell, J. Mol. Catal. A Chem. 272 (2007) 108.