Chiral Co(II) metal-organic framework in the heterogeneous catalytic oxidation of alkenes under aerobic and anaerobic conditions

Article abstract of DOI:10.1021/cs401003d

The chiral Co(II) MOF [Co(l-RR)(H₂O)·H₂O] _∞ [1; l-RR = (R,R)-thiazolidine-2,4-dicarboxylate] has been exploited in the catalytic oxidation of different alkenes (cyclohexene, (Z)-cyclooctene, 1-octene) using either tert-butyl hydroperoxide ( ^tBuOOH) or molecular oxygen (O₂) as oxidants. Different chemoselectivities are observed, both substrate- and oxidant-dependent. A moderate enantioselectivity is also obtained in the case of prochiral precursors, revealing the chiral induction ability of the optically pure metal environment. The interaction of O₂ with the exposed metal sites in 1 (after material preactivation and consequent removal of the coordinated aquo ligand) has been studied through TPD-MS analysis combined with DFT calculations, with the aim of probing effective oxygen uptake by the heterogeneous catalyst and unraveling the nature of the active species in the catalytic oxidation process under aerobic conditions. Theoretical results indicate the presence of an η¹-superoxo species at the cobalt center, with concomitant Co(II) ? Co(III) oxidation. Finally, the experimental estimation of the O₂ adsorption enthalpy is found to be in good agreement with the calculated binding energy.

Full text of DOI:10.1021/cs401003d

Research Article
pubs.acs.org/acscatalysis
Chiral Co(II) Metal−Organic Framework in the Heterogeneous
Catalytic Oxidation of Alkenes under Aerobic and Anaerobic
Conditions
Giulia Tuci,† Giuliano Giambastiani,†,‡ Stephanie Kwon,§ Peter C. Stair,⊥ Randall Q. Snurr,§
and Andrea Rossin*,†,‡
†Consiglio Nazionale delle Ricerche, Istituto di Chimica dei Composti Organometallici (ICCOM-CNR), Via Madonna del Piano 10,
50019 Sesto Fiorentino (Firenze), Italy
‡Consorzio Interuniversitario per la Scienza e Tecnologia dei Materiali (INSTM), Via G. Giusti 9, 50121 Firenze, Italy
⊥
§Department of Chemical and Biological Engineering, Department of Chemistry, Northwestern University, 2145 Sheridan Road,
Evanston, Illinois 60208, United States
S
* Supporting Information
ABSTRACT: The chiral Co(II) MOF [Co(L-RR)(H2O)·H2O]∞
[1; L-RR = (R,R)-thiazolidine-2,4-dicarboxylate] has been exploited
in the catalytic oxidation of different alkenes (cyclohexene, (Z)-
cyclooctene, 1-octene) using either tert-butyl hydroperoxide
(tBuOOH) or molecular oxygen (O2) as oxidants. Different
chemoselectivities are observed, both substrate- and oxidant-
dependent. A moderate enantioselectivity is also obtained in the
case of prochiral precursors, revealing the chiral induction ability of
the optically pure metal environment. The interaction of O2 with the
exposed metal sites in 1 (after material preactivation and consequent
removal of the coordinated aquo ligand) has been studied through
TPD-MS analysis combined with DFT calculations, with the aim of
probing effective oxygen uptake by the heterogeneous catalyst and
unraveling the nature of the active species in the catalytic oxidation process under aerobic conditions. Theoretical results indicate
the presence of an η1-superoxo species at the cobalt center, with concomitant Co(II) ↔ Co(III) oxidation. Finally, the
experimental estimation of the O2 adsorption enthalpy is found to be in good agreement with the calculated binding energy.
KEYWORDS: metal−organic frameworks, cobalt (II), alkene oxidation, heterogeneous catalysis, DFT calculations, dioxygen activation
from the metal coordination environment of the “unactivated”
material (mostly H2O, when operating under hydrothermal
synthetic conditions; pure water or aqueous binary mixtures are
the solvents of choice par excellence for MOF synthesis). The
preactivation of the material is expected to create open metal
sites potentially available for promoting catalysis while keeping
the framework crystalline texture intact. In addition, the
heterogeneous nature of a MOF can be very useful to separate
the catalyst from the products of interest; recover it after simple
filtration procedures; and finally, regenerate it for successive
catalytic runs.
One industrially important reaction is the selective oxidation
of hydrocarbons to higher-value chemical feedstocks.7 Tradi-
tional catalytic procedures make use of polluting oxidizing
agents (such as MnO2 or CrO3) that are environmentally
unfriendly and difficult to recycle. Thus, in recent years, the
INTRODUCTION
■
The extremely high degree of freedom in the structural design
of metal−organic frameworks (MOFs), achieved through a
virtually infinite combination of inorganic secondary building
units (SBUs) and organic linkers, opens new perspectives in the
practical application of this class of materials. Thus, in addition
to the initial exploitation of MOFs for gas storage purposes,1 in
the recent literature, more diversified applications have started
to appear, spanning from separation and purification of gas
mixtures2 to luminescence,3 ion sensing,4 drug delivery,5 and
heterogeneous catalysis.6 The employment of MOFs as
catalysts for selected chemical transformations relies on the
presence of catalytically active sites within their 3D networks.
Transition metals, in particular, are good candidates for the
activation of small molecules, which can coordinate to a vacant
site within the metal’s coordination sphere (the basic principle
of homogeneous catalysis performed by transition metal
organometallics).
Received: October 31, 2013
Revised: February 17, 2014
Published: February 17, 2014
In some MOFs, empty sites on the inorganic SBU can be
generated through the simple removal of small neutral ligands
© 2014 American Chemical Society
1032
dx.doi.org/10.1021/cs401003d | ACS Catal. 2014, 4, 1032−1039

ACS Catalysis
Research Article
search for greener processes that could work with more benign
oxidants has grown considerably. Among them, organic
peroxides (R−OOH; R = alkyl chain), hydrogen peroxide
(H2O2), and dioxygen (O2) are the most popular.
under both anaerobic [tert-butyl hydroperoxide (tBuOOH) as
oxidant] and aerobic (O2 as oxidant) conditions.
EXPERIMENTAL SECTION
■
Several examples of MOF-catalyzed hydrocarbon oxidation
reactions using peroxides or dioxygen as oxidizing agents have
appeared in the literature, exploiting metals coming from the
first transition 3d series, such as chromium,8 manganese,9
iron,10 and copper,11 combined with assorted organic spacers.
Cobalt, in particular, has been shown to be well-performing in
olefin oxidation by organic peroxides or dioxygen with
satisfactory activity and (chemo)selectivity under mild reaction
conditions.12 The main reaction products strongly depend on
the nature of the oxidant and the presence of auxiliary radical
initiators, such as N-hydroxyphtalimide (NHPI). If the
heterogeneous catalyst of choice is optically active (because
either it contains a chiral organic spacer or it bears optically
active catalytic single sites embedded into the crystal lattice),
enantioselective catalysis is also achievable.13 Some of us have
recently prepared and characterized a chiral Co(II) MOF based
on an optically pure thiazolidine dicarboxylate ligand: [Co(L-
RR)(H2O)·H2O]∞ [1; L-RR = (R,R)-thiazolidine-2,4-dicarbox-
ylate, Figure 1].14 The Co(II) coordination sphere contains one
General Considerations. All catalytic tests have been
performed either under inert atmosphere in flame-dried flasks
using standard Schlenk-type techniques or under O2 atmos-
phere at variable pressures using a stainless steel autoclave (15
mL internal volume) equipped with a magnetic stirrer, a Teflon
inset, a pressure controller, and a liquid/gas inlet. 1-Octene,
cyclohexene, (Z)-cyclooctene and o-DCB (ortho-dichloroben-
zene) were purified according to literature procedures15 and
stored over 4 Å molecular sieves under nitrogen. Unless
otherwise stated, all other chemicals were purchased from
commercial suppliers and were used as received without further
purification. GC/MS analyses were performed on a Shimadzu
QP2010S apparatus equipped with a flame ionization detector
and a Supelco SPB-1 fused-silica capillary column (30 m length,
0.25 mm i.d., 0.25 μm film thickness). Chiral GC analyses were
performed on a Shimadzu 17A apparatus equipped with a flame
ionization detector and a Lipodex-E column (50 m length, 0.25
mm i.d., 0.25 μm film thickness).
General Procedure for the Catalytic Alkene Oxidation
by 1act Using tBuOOH As Oxidant. A mixture of the selected
t
alkene (16 mmol), BuOOH (8 mmol), and either o-DCB
(used as internal standard for cyclohexene oxidation trials, 2
mmol) or 2,6-dimethylphenol (used as internal standard for 1-
octene and (Z)-cyclooctene oxidation trials, 2 mmol) are added
in one portion to the activated catalyst 1act (0.095 mmol)
weighed into a two-necked 25 mL flask.16 The resulting
suspension is stirred at 70 °C for 24 h. Afterward, the mixture is
allowed to cool to room temperature, filtered over a Celite pad,
and analyzed through gas chromatography. Catalyst recovery
and recycling is operated under inert atmosphere by means of a
careful removal of the supernatant and subsequent washing of
the solid catalyst with dry and degassed n-pentane (5 mL). The
solid residue is then left to dry under high vacuum for 1 h to
remove all volatiles before resubmitting it for further oxidation
cycles. Finally, hot filtration experiments were performed to
confirm the heterogeneous nature of the catalyst, thus
excluding any cobalt leaching from the activated MOF.
Catalytic Cyclohexene Oxidation by 1act under
Aerobic Conditions (O2 As Oxidant). Compound 1act
(0.095 mmol) is weighed under inert atmosphere in a Teflon
sample holder and placed in a 15 mL volume stainless steel
autoclave equipped with a magnetic stirrer. The reactor is
purged with three nitrogen−vacuum cycles and then filled with
a solution of cyclohexene (16 mmol) and o-DCB (used as
internal standard, 2 mmol). The reactor is then pressurized
with oxygen at the desired pressure (1, 3, or 5 bar) and kept at
70 °C for 24 h. Oxygen is continuously fed to keep the reactor
pressure constant throughout the catalytic test. After that time,
the system is allowed to cool to room temperature, filtered over
a Celite pad, and analyzed through gas chromatography.
Catalyst recovering and recycling is operated following a
procedure similar to that described above for catalysis using
tBuOOH as oxidant.
Figure 1. Cobalt coordination sphere (a) and 3D network (b) of 1.
TPD-MS Measurements. The TPD-MS analysis is
performed through the Hiden Analytical CATLAB instrument
(http://www.hidenanalytical.com) on sample 1act pretreated
with O2 or ethylene at p = 5 bar in a closed vessel at 70 °C for 2
h. Thermal desorption on the pressurized samples is carried out
under an argon flow (flow rate = 20 mL/min) in the 20−300
water molecule that can be removed upon thermal treatment of
the as-synthesized material at 190 °C for 24 h. The activated
form [Co(L-RR)]∞ (1act) maintains the same lattice structure as
1. In the present work, 1act has been exploited as a
heterogeneous catalyst for the oxidation of selected alkenes
1033
dx.doi.org/10.1021/cs401003d | ACS Catal. 2014, 4, 1032−1039

ACS Catalysis
Research Article
°C temperature range (sample heating rate = 5 °C/min). The
quadrupole mass detector limit is 2 × 10−14 Torr.
Table 1. Cyclohexene Oxidation Catalyzed by 1act Using
a
tBuOOH As Oxidant
Computational Details. Periodic DFT calculations were
performed using the gradient-corrected Perdew−Burke−
Ernzerhof (PBE) functional17 as implemented in the VASP
software package.18 The long-range dispersion interactions
were included using the PBE-D2 method based on Grimme’s
parameters.19 The core electrons were described by the
projector augmented wave (PAW) method.20 In all of the
VASP calculations, spins were polarized. For the cell and
geometry relaxations, the systems were fully relaxed until the
total energy was converged to 10−4 eV. The cell parameters
were first calculated by optimizing the cell volume and atomic
positions. The Co-MOFs with and without water on Co metal
sites were optimized using unit cells with 80 atoms and 68
atoms, respectively. The energy cutoff was set at 800 eV, and a
Monkhorst−Pack grid of (2 × 2 × 2) was used.21 For O2
adsorption calculations, the lattice constants were fixed at the
calculated value with Co open metal sites, and only atomic
positions were allowed to relax using the energy cutoff of 400
eV. A 4 × 4 × 8 k-point mesh was used. Atomic charges were
obtained based on Bader charge analysis.22 We also tested and
optimized all geometries through a GGA(PBE)+U approach.23
The self-interaction error in the 3d orbitals of Co was
corrected, and 3.5 eV was used as the Ueff parameter.24
Calculated adsorption energies were close to the GGA
calculated values and are listed in the Supporting Information
(SI). The oxygen adsorption energy of 1act was calculated on
the basis of the equation ΔEads = E(MOF − O2) − [E(MOF) +
E(O2)]. For loadings higher than one molecule of O2/cell, the
average O2 adsorption energies were calculated as follows:
ΔEavg = 1/n × {E(MOF − nO2) − [E(MOF) + nE(O2)]}. The
differential O2 adsorption energies were calculated through the
simplified equation: ΔEdiff = E(MOF + nO2) − {E[MOF + (n
− 1)O2] + E(O2)}.
b
,
c
c
c
c
c
entry
1act (mmol)
conv %
a %
b %
c %
d %
1
2
0
1.6
58
d
23
19
0.095
18.6
100
e
f
3
19.0
99
100
93
<1
<1
g
4
0.095
0.095
0.095
19.0
18.4
16.7
h
5
5
5
1
1
1
2
i
6
92
a
t
Oxidation conditions, BuOOH (8 mmol); reaction temperature, 70
°C; reaction time, 24 h. Calculated from the GC trace using o-DCB
as internal standard. Average values calculated over three independent
runs. ee, 24%. Hot filtration test; reaction of the filtered supernatant
deriving from entry 2 after additional 2 h at 70 °C. Cobalt leaching
measured by GF-AAS analysis ≈ 0.002%. Catalyst recycling from
entry 2 (2nd run). Catalyst recycling from entry 4 (3rd run).
b
c
d
e
f
g
h
i
Catalyst recycling from entry 5 (4th run).
lower than those measured for MFU-1 (27% after 22 h),12f
[Co3(BTC)2(HCOO)4(DMF)]·H2O (84% after 24 h),12a or
MFU-3 (62% after 12 h)12g under similar experimental
conditions; nonetheless, it is slightly higher than that of
MFU-2 (16% after 22 h).12e As for chemoselectivity, when
compared with the previous examples, 1act shows a higher
selectivity toward a 100% conversion to the tert-butyl-2-
cyclohexenyl-1-peroxide product, a. Indeed, only the 83% of
product a was obtained with MFU-3 as catalyst, whereas 66%
was the maximum amount of a that can be obtained from
MFU-1 or MFU-2 under similar reaction conditions. The
compound [Co3(BTC)2(HCOO)4(DMF)]·H2O gave a com-
pletely different selectivity toward the ketone, b (95% after 24
h). The peroxide (a) was the main oxidation product deriving
from the cobalt-mediated allylic substitution on cyclohexene
RESULTS AND DISCUSSION
Catalytic Alkene Oxidation by 1act Using BuOOH As
■
t
25
t
Oxidant. The activated chiral MOF 1act has been preliminarily
scrutinized as a heterogeneous catalyst for the oxidation of
selected alkenes (cyclohexene, (Z)-cyclooctene, 1-octene)
when BuOOH was used as oxidant. Thus, the proposed
mechanism of cyclohexene conversion into a in the presence of
tBuOOH (SI Scheme S2) is of the same kind as that reported
for MFU-1.12f In line with this mechanistic hypothesis, 1 equiv
t
using BuOOH as oxidant. For the sake of comparison,
t
substrates and experimental conditions used for the oxidation
process have been chosen according to those reported in
similar literature precedents.12e−g A blank reaction test carried
out in the absence of 1act has also been used to elucidate the
catalyst role in terms of process activity and selectivity. Because
the material has only very small pores (pore limiting diameter
of 3.96 Å, as calculated from the crystal structure),14 it can be
assumed that the catalytic transformations take place at the
crystallites’ outer surface only, without diffusion into the MOF
3D lattice.
of the byproduct BuOH (in a 1:1 stoichiometric ratio with
respect to a) was also observed on the GC traces of the reaction
solutions. With the aim of excluding any possible contami-
nation by homogeneous metal active sites responsible for the
observed catalyst activity and selectivity (cobalt leaching), a hot
filtration test and an ICP-MS analysis of the metal content in
the supernatant were performed. Accordingly, the oxidation
(Table 1, entry 2) was allowed to proceed for 24 h before being
split into two fractions, one containing the suspended catalyst
and the other filtered to remove any solid precipitate. The
whole procedure was performed at 70 °C under inert
atmosphere to prevent any oxygen contamination of both
catalyst and solution. The solution was finally allowed to react
for an additional 24 h at 70 °C before being analyzed via GC/
MS (Table 1, entry 3).
As can be inferred from the analysis of reaction products
obtained from the resubmission of the filtered supernatant
(entry 3), only a negligible increase in the substrate conversion
was observed, and new oxidation products [namely, traces of 2-
cyclohexen-1-one (b) and 2-cyclohexen-1-ol (c)] started to
appear. The slight conversion increase (∼0.4%; Table 1, entries
The liquid-phase cyclohexene oxidation by 1act/tBuOOH is a
relatively fast process with fairly good catalyst activity, together
with remarkable selectivity toward the tert-butyl-2-cyclohex-
enyl-1-peroxide product (a, Table 1). Notably, cyclohexene
oxidation under the same experimental conditions but in the
absence of the catalyst is almost negligible, with ∼1.6% of
substrate conversion after 24 h at 70 °C and a moderate catalyst
selectivity (Table 1, entries 1 vs 2).
Using 0.095 mmol of 1act in the presence of a large excess of
tBuOOH, the highest substrate conversion achieved after 24 h
at 70 °C was 19.0%. The conversion obtained in this case is
1034
dx.doi.org/10.1021/cs401003d | ACS Catal. 2014, 4, 1032−1039

ACS Catalysis
Research Article
2 vs 3) is reasonably ascribed to the action of the excess of
tBuOOH, whereas the role of homogeneous CoII species can be
definitively ruled out from the GF-AAS analysis of the
cyclohexene solution at the end of the catalytic cycle (Table
1, entry 3). Indeed, a content of ≈0.002% of “leached” cobalt
ions was detected in solution, thus labeling the process as truly
heterogeneous in nature. In addition, product a from entry 2
showed a moderate enantiomeric excess (ee 24%) determined
via chiral GC analysis of the reaction mixture;26 this issue is in
line with the chiral nature of the employed Co-MOF catalyst.
To the best of our knowledge, this is the first example of a
heterogeneous alkene oxidation performed with an optically
pure Co-MOF catalyst. On the other hand, a racemic mixture
(50:50) of the two (R,S) isomers of a was obtained in the
absence of the catalyst (Table 1, entry 1).
Catalyst recovery and recycling was straightforwardly
performed through complete solvent removal at the end of
the first catalytic cycle. The solid residue was then rinsed with
freshly distilled and degassed n-pentane (1 × 5 mL), filtered,
and dried at room temperature under vacuum before being
treated with an additional amount of cyclohexene and
tBuOOH. This procedure was repeated under inert atmosphere
at the end of each catalytic cycle. Compound 1act maintained its
catalytic activity almost unchanged during three successive
oxidation cycles (Table 1, entries 2 vs 4−6), whereas a slight
reduction of the catalyst selectivity became appreciable from
the second catalyst recycling (Table 1, entries 2 and 4 vs 5).
The contamination from secondary oxidation products (b−d)
in the recycling runs is ascribed to the successive additions of a
large excess of the oxidant, and the slightly reduced catalyst
activity during the recycling is tentatively ascribed to its
progressive deactivation caused by either adventitious moisture
or catalyst loss.
observed on the catalyst [Co3(BTC)2(HCOO)4(DMF)]·H2O
(64%)12a under similar reaction conditions. Most importantly,
the catalytic process led to mixtures of oxidation products in
which the epoxide (a) and the tert-butyl-2-cyclooctenyl-1-
peroxide (c) constitute ∼90% of the mixture. Only a minor
component (<10%) ascribed to the cyclooct-2-enone (b) was
detected from the GC trace.
As for the process selectivity, the two main oxidation
products (a and c) were obtained in nearly equal amounts.
Although the epoxidation of (Z)-cyclooctene is expected to be
the most favored process (allylic functionalization is more
difficult than for other cycloalkenes, because of almost
orthogonal allylic C−H bonds),27 allylic oxidation still
represents one important oxidation path under the exper-
imental conditions used. Only a small fraction of 2-cycloocten-
1-one (up to 8%) was detected in the mixtures. Compound 1act
is also less selective than [Co3(BTC)2(HCOO)4(DMF)]·H2O,
for which a selectivity of 78% toward the epoxide (a) was
recorded.12a
Catalyst recycling was also performed on the cyclooctene
oxidation trials, following the same experimental approach.
Again, 1act maintained its catalytic activity and selectivity
unchanged for (at least) two successive oxidation cycles, (Table
2, entries 2 vs 3 and 4).
For the sake of completeness, 1-octene was finally selected as
a model linear olefin for the catalytic oxidation tests. The
experimental conditions used were the same as those applied to
cyclic olefins, and catalytic outcomes are summarized in Table
3. 1-Octene oxidation proceeded with overall substrate
Table 3. 1-Octene Oxidation Catalyzed by 1act using
a
tBuOOH As Oxidant
The oxidation protocol was applied to another model cyclic
olefin, (Z)-cyclooctene. Catalytic results are listed in Table 2.
b
,
c
c
c
c
entry
1act (mmol)
conv %
1.1
a %
b %
c %
Table 2. (Z)-Cyclooctene Oxidation Catalyzed by 1act Using
a
1
2
0
50
e
50
52
53
57
tBuOOH As Oxidant
d
0.095
0.095
0.095
18.8
17.3
17.9
37
11
12
8
f
3
35
35
g
4
a
t
Oxidation conditions, BuOOH (8 mmol); reaction temperature, 70
b
°C; reaction time, 24 h. Calculated from the GC trace using 2,6-
dimethylphenol as internal standard. Average values calculated over
three independent runs. Cobalt leaching measured by GF-AAS
analysis ≈ 0.003%. ee 18%. Catalyst recycling from entry 2 (2nd
run). Catalyst recycling from entry 3 (3rd run).
c
b,
c
c
c
c
entry
1act (mmol)
conv %
a %
b %
c %
d
1
2
0
2.1
27.6
25.2
26.1
49
41
45
49
5
8
4
5
46
51
43
46
e
f
0.095
0.095
0.095
g
d
3
e
4
a
t
Oxidation conditions, BuOOH (8 mmol); reaction temperature, 70
conversions up to 19% after 24 h at 70 °C, with the production
of mixtures made of three different oxidation products: oct-1-
en-3-ol (a), (E/Z)-oct-2-en-1-ol (b) and the 2-hexyloxirane (c),
the first two components representing about 90% of the
mixture. Similarly to cyclohexene oxidation, the epoxidation
product represents the minor component in the mixture. As can
be argued from the product distribution, the allylic oxidation
mechanism, proceeding via αH-abstraction and generation of
an allylic radical, seems to be the main operating oxidation path
in the process. Accordingly, the allylic substitution generated
the two allylic alcohols with a branched to linear ratio of ∼0.7.
Substrate conversion and selectivity were almost completely
preserved in the two successive oxidation cycles in which 1act
was recovered and reused according to the operative conditions
used for the previously discussed oxidations on cyclic olefins.
b
°C; reaction time, 24 h. Calculated from the GC trace using 2,6-
dimethylphenol as internal standard. Average values calculated over
c
d
three independent runs. Catalyst recycling from entry 2 (2nd run).
e
Catalyst recycling from entry 3 (3rd run).
According to the procedure outlined above, a blank reaction
test carried out in the absence of 1act again was used to establish
the role of the catalyst in terms of activity and selectivity.
As reported in Table 2, cyclooctene oxidation generally
proceeds with higher substrate conversions (up to 28% after 24
h at 70 °C) but with a reduced selectivity compared with
cyclohexene. A slightly higher percentage of the oxidation
products was also observed in the blank test. The total
conversion obtained in this reaction was lower than that
1035
dx.doi.org/10.1021/cs401003d | ACS Catal. 2014, 4, 1032−1039

ACS Catalysis
Research Article
The action of leached homogeneous CoII species was
definitively ruled out from the GF-AAS analysis of the crude
reaction mixture at the end of the catalytic cycle (Table 3, entry
2). A content of ≈0.003% of “leached” cobalt ions was detected
in the solution, thus confirming a truly heterogeneous process.
In addition, a modest enantiomeric excess (ee 18%) was
measured via chiral GC analysis on product a only; such a result
confirms the (even moderate) induction ability of the
enantiomerically pure cobalt coordination environment.
Catalytic Cyclohexene Oxidation by 1act Using
Molecular O2 As Oxidant. To make a step forward in the
oxidation of cyclic olefins, the activated chiral MOF 1act has also
been used under aerobic conditions, using molecular oxygen as
oxidant. At odds with previously reported data on the aerobic
oxidation of cyclic olefins,12d,f,28 1act is capable of catalyzing
cyclohexene oxidation efficiently without the employment of
cocatalysts acting as electron-transfer mediators. As a matter of
fact, cyclohexene conversions up to 37% were obtained after 24
h at 70 °C under constant O2 pressure (Table 4, entry 3),
chemical selectivity is completely different, the main oxidation
product being 2-cyclohexenyl-1-peroxide (not observed in our
catalytic trials).8,12f In the absence of radical initiators, the two
main products are the ketone (a) and the alcohol (b) in 78%
and 90% total selectivity for 1act and [Co2(DOBDC)(H2O)2]·
8H2O,12d respectively.
With the assumption that CoIII-superoxo species are
responsible for the observed catalytic activity (vide infra and
SI) and in the absence of electron-transfer cocatalysts, the
process selectivity confirms the coexistence of two oxidation
paths: epoxidation and allylic oxidation.27 Although epoxide d
and diol c are supposed to be generated from an epoxidation
mechanism followed by an overoxidation of the epoxide to diol
and regeneration of the pristine CoII species (SI Scheme S1, eqs
2 and 5), a mechanism initiated by a substrate αH-abstraction
with generation of an allylic radical can be reasonably invoked
to justify the production of the allylic alcohol b (SI Scheme S1,
eq 3). The major mixture component (a) is ultimately ascribed
to an overoxidation path of the intermediate allylic alcohol b
(SI Scheme S1, eq 4). The absence of exogenous sacrificial
reductants (such as DMF) capable of promoting the
regeneration of the pristine CoII species28 is apparently
balanced by a substrate overoxidation (SI Scheme S1, eqs 3
and 5). Nevertheless, additional data and a more detailed
mechanistic study (which are beyond the scope of the present
work) are needed to provide unambiguous evidence for all the
postulated intermediates.
Table 4. Cyclohexene Oxidation Catalyzed by 1act Using
a
Molecular O2 As Oxidant
1act
p(O2)
(bar)
conv
b,c
Oxygen is continuously fed to the reactor to keep the desired
pressure constant throughout the catalytic test. Although no
appreciable variation of the product distribution is observed
when the oxidation process is performed at increasing O2
pressures, substrate conversions follow a more precise trend,
reaching a maximum value for pO2 = 2 bar. Thus, higher O2
pressures (>2 bar) translates into a reduced substrate
conversion. Such a trend may be ascribed to the “saturation”
of all the exposed catalytic sites at pO2 around 2 bar. On a
speculative ground, the observed reaction trend could also be
ascribed to the generation of binuclear Co-peroxo species
forming at higher O2 pressures between exposed metal ions at
spatially close crystallite edges (SI Scheme S1, eq 1).28a,29
Catalyst recycling in successive oxidation tests has also been
accomplished for catalysis performed under aerobic conditions.
Thus, the catalyst is recovered at the end of the first oxidation
cycle by cooling and depressurizing the reactor, decanting the
solid catalyst, and removing the supernatant solution via a
Teflon cannula under an O2 flow. The solid catalyst is dried
under a stream of O2 before being rinsed with an additional
amount of cyclohexene and repressurized at the desired O2
pressure. As Table 4 shows, 1act maintains its catalytic activity
and selectivity almost unvaried for two successive oxidation
cycles, the small conversion decrease being attributed to either
a partial catalyst removal during the recovery procedure or
catalyst poisoning by adventitious moisture (Table 4, entries 3
vs 6 and 7).
c
c
c
c
c
entry (mmol)
%
a % b % c % d % e %
1
2
3
4
0
2
1
2
3
5
2
2
d
0.095
0.095
0.095
0.095
0.095
0.095
27.7
36.7
33.9
20.3
31.2
32.6
54
49
43
59
51
54
29
29
17
22
29
26
7
11
23
9
1
9
11
17
10
13
9
5
e
6
7
f
7
9
2
a
Oxidation conditions: O2 (1−5 bar); reaction temperature, 70 °C;
reaction time, 24h. Calculated from the GC trace using o-DCB as
internal standard. Average values calculated over three independent
runs. Cobalt leaching measured by GF-AAS analysis ≈0.005%.
Catalyst recycling from entry 3 (2nd run). Catalyst recycling from
entry 6 (3rd run).
b
c
d
e
f
whereas no substrate conversion was shown when the reaction
was carried out in the absence of the catalyst (Table 4, entry 1).
The achieved conversion is the highest reported to date for this
reaction catalyzed by Co(II) MOFs. MFU-1 showed a 35%
conversion after 24 h at T = 35 °C in the presence of the radical
initiator N-hydroxyphtalimide (NHPI),12f whereas almost 33%
after 20 h was obtained with [Co2(DOBDC)(H2O)2]·8H2O
(DOBDC = 2,5-dihydroxyterephthalate) under similar reaction
conditions to ours.12a Higher conversions were recorded only
for Fe(II)-containing MOFs, such as Fe-MIL-101 (up to 44%
after 16 h), even if in this case a small amount of tBuOOH was
added to the reaction mixture to initiate the process.
The aerobic oxidation protocol exhibited only moderate
selectivity: indeed, five oxidation products have been detected,
whose identity has been unambiguously assigned, together with
small amounts of other unidentified oxidation byproducts
(Table 4, e). The complete mass balance and substrate
conversion were obtained from GC analysis of the crude
reaction mixtures using o-DCB as internal standard. In the
Temperature-Programmed Desorption of O2 and
Ethylene on 1act. To assess the effective interaction of
dioxygen with the heterogeneous catalyst and to verify if olefin
coordination could also take place during the process, TPD-MS
experiments were carried out on 1act pretreated under O2 or
ethylene (as gaseous olefin model compound) under the same
experimental conditions used for the catalytic trials. For
dioxygen, the ion current vs temperature graphs for m/z =
32 a.m.u. (Figure 2) clearly indicate desorption of O2, with a
t
presence of radical initiators such as NHPI or BuOOH, the
1036
dx.doi.org/10.1021/cs401003d | ACS Catal. 2014, 4, 1032−1039

ACS Catalysis
Research Article
Figure 2. TPD-MS plot for O2 desorption from 1act.
maximum centered around 37 °C. The estimated ΔHdes at the
curve maximum (inferred through the TPD Plotter integrated
software30) equals 85.4 kJ/mol, in good agreement with the
value coming from the computational results for the uptake of
one O2 molecule per unit cell (vide infra). The same
experiments performed with ethylene (m/z = 26 and 27
a.m.u., SI Figure S1) revealed that no significant interaction
between 1act and the CC bond occurs, the two TPD profiles
being identical for the blank sample and that pretreated with
ethylene. This may be due to the fact that the kinetic diameter
of ethylene (4.5 Å) is slightly bigger than the MOF
crystallographic pore size (3.96 Å, see above), thus preventing
any chemical interaction between the olefin and the metal sites.
DFT Analysis of O2 Uptake by 1act. To get insight about
how 1act can activate O2, we modeled the system through
periodic DFT calculations. In the calculations, the interactions
between O2 and the cobalt sites in 1act have been investigated.
As noted above, the catalysis probably occurs on the outer
surfaces of the MOF crystallites; however, given our lack of
information about these sites, the interior open metal sites are a
useful model for a preliminary understanding of O2 activation
in this system. In addition, the TPD measurements likely probe
both inner and outer metal sites, since the kinetic diameter of
O2 (3.46 Å)31 is slightly smaller than the pore limiting diameter
of the MOF (3.96 Å). Consequently, it is reasonable to think
that a partial O2 diffusion through the particles outmost crystal
layers occurs.
A cell size and geometry optimization was initially performed
to obtain optimal lattice constants and structures of 1 and 1act.
The experimental X-ray structure was used as an initial guess.
The lattice parameters calculated for 1 were very close to the
experimental data, as shown in SI Table S1. During the
geometry optimization, both 1 and 1act had 12 unpaired
electrons within a cell (three on each Co atom), which implies
that Co2+ sites are in a high spin state. In this state, the 3D
lattice can have different magnetic configurations. As shown in
Figure 3, three different magnetic configurations were
calculated. In the ferromagnetic configuration, all metal ions
have the same spin states (Figure 3a), but in the other two
possible antiferromagnetic configurations, neighboring metal
ions have opposite spins (Figure 3b,c). Our results show that
the antiferromagnetic configuration depicted in Figure 3b has
the lowest total energy (ground state electronic energy), lower
by 0.97 kJ/mol per Co atom than the ferromagnetic
configuration (Figure 3a) and by 0.57 kJ/mol per Co atom
Figure 3. Magnetic configurations of the cobalt centers in 1act: (a)
ferromagnetic; (b, c) antiferromagnetic. The blue and red surfaces
represent spin density surfaces with values of 0.3 and −0.3,
respectively. The (2 × 2 × 1) extended cells are shown in the figure
for better visualization.
than the other antiferromagnetic configuration (Figure 3c).
This type of antiferromagnetic nature has also been shown to
be the lowest energy configuration for other cobalt-containing
MOFs, such as CPO-27-Co.32
For O2 adsorption, the optimized geometry of 1act was used.
The cell parameters were kept constant (at the previously
calculated value) throughout the calculation; only the atomic
positions were allowed to relax. First, one oxygen molecule was
introduced into the system, and the overall structure was fully
optimized. Figure 4 shows the optimized geometry of O2
adsorption on one Co site. The O−O bond length was
elongated after adsorption from 1.235 Å (as found in free
dioxygen) to 1.285 Å, showing a typical value of the superoxo-
Co(III) fragment found in other literature examples (1.28 Å).33
A Bader charge analysis was performed to investigate the charge
distribution. Upon adsorption, the O2 ligand gains charge from
Co electronic back-donation; in addition, the coordinated
oxygen atom bears a higher negative charge compared with the
1037
dx.doi.org/10.1021/cs401003d | ACS Catal. 2014, 4, 1032−1039

ACS Catalysis
Research Article
Figure 4. DFT-optimized geometry of one O2 molecule/cell adsorbed
by 1act. (a) The coordination environment of the cobalt center after O2
adsorption. The internuclear O−O and Co−O distances (Å) are
reported. Atomic charges transferred to O atoms are reported in
parentheses (see main text for details.) (b) The unpaired spin density
surfaces in 1act after O2 adsorption. The blue and red isosurfaces
represent 0.3 and −0.3, respectively.
Figure 5. DFT optimized geometry of O2 adsorption on 1act at higher
O2 loadings. The geometry around the four Co centers is shown. (a)
Two O2 per unit cell adsorbed on opposite Co sites, (b) two O2 per
unit cell adsorbed on adjacent Co sites, (c) three O2 per unit cell, and
(d) four O2 per unit cell.
distal oxygen atom, as shown in Figure 4a. The distribution of
unpaired spin was also examined. After O2 adsorption, the
unpaired spin is localized mainly on the O2 fragment, but not
on the cobalt atom attached to it, which implies that, in the
Co−O2 adduct, the metal oxidation state is +3 (Co3+ = 3d6),
with all paired electrons. The crystal field splitting energy (Δ)
of Co3+ is much larger than that of Co2+, thus favoring a low
spin state on Co3+. According to crystal field theory, the
splitting is also affected by the ligand strength. The superoxo
both anaerobic and aerobic conditions. Different activities and
selectivities have been observed, together with a moderate
chiral induction as measured on selected chiral oxidation
products. The results using O2 as the oxidant are consistent
with a mechanism that proceeds through the generation of an
η1-superoxo species with no evidence for any significant
interaction of the olefin with the metal center and its chiral
environment. This is probably why only a moderate ee is
observed for the analyzed chiral oxidation products. Among the
various olefins considered, cyclohexene is the most efficiently
oxidized substrate. During its anaerobic oxidation, a 100%
selectivity toward the chiral tert-butyl-2-cyclohexenyl-1-per-
oxide product has been achieved, albeit with a moderate
conversion (at t = 24 h). The highest conversion (37% after 24
h) reported to date for a Co-MOF catalyzed cyclohexene
oxidation (but a lower selectivity) is achieved under aerobic
conditions, leading to the α,β-unsaturated ketone 2-cyclo-
hexenone as the main product. As a complement to the
experimental work, a DFT computational analysis of O2
interaction with the exposed metal sites in 1act was performed.
The optimized structures, the calculated atomic charges, and
the spin state are consistent with the η1-superoxo nature of
bound O2. The calculated adsorption energy of one O2
molecule per unit cell is in good agreement with the value
obtained experimentally from TPD-MS. Following these
results, new thiazole-based MOFs containing exposed metal
sites are currently being prepared in our laboratories for the
study of new catalytic processes.
−
(O2 ) ligand is only slightly weaker than NH3, again favoring a
low spin configuration for the Co3+−O2 fragment.34
−
The PBE-calculated O2 adsorption energy on 1act is −72 kJ/
mol, and the (PBE+U)-calculated value is −75 kJ/mol. Higher
loadings of O2 were further analyzed by introducing more O2
molecules into the computational model. For the adsorption of
two oxygen molecules in the unit cell, two possible adsorption
geometries were considered: the O2 molecules on two adjacent
Co sites and on two opposite Co sites. The latter was found to
be slightly more stable (ΔΔE = 1 kJ/mol) at the computational
level used (Table 5). For ΔEdiff related to three O2 ligands/cell,
a
Table 5. O2 Binding Energies on 1act Calculated Using
b
Periodic GGA Methods
binding energy (kJ/mol)
coverage per unit cell
av
differential
1 O2
−72
−71
−70
−67
−56
2 O2 (opposite)
2 O2 (adjacent)
3 O2
−70
−68
−58
−25
4 O2
a
b
In kJ/mol. The corresponding values obtained with a periodic GGA
+U approach are listed in Table S2.
the E[MOF + 2O2] reference value was that of the geometry in
which the two O2 ligands were placed on opposite Co sites.
The adsorption energy decreases for higher loadings, especially
for a fully saturated loading (Figure 5 and Table 5). This can be
attributed to a repulsive electrostatic effect.
ASSOCIATED CONTENT
■
S
* Supporting Information
Complementary DFT tables and Cartesian coordinates of the
optimized geometries, hypothesized reaction schemes, TPD-
MS results for ethylene. Additional experimental data (GC
traces) are available from the authors on request. This material
is available free of charge via the Internet at http://pubs.acs.org.
CONCLUSIONS
■
The chiral Co(II) MOF [Co(L-RR)(H2O)·H2O]∞ has been
exploited for the catalytic oxidation of selected alkenes under
1038
dx.doi.org/10.1021/cs401003d | ACS Catal. 2014, 4, 1032−1039

ACS Catalysis
Research Article
Mollmer, J.; Sauer, J.; Volkmer, D. Chem.−Eur. J 2011, 17, 8671−
̈
AUTHOR INFORMATION
Corresponding Author
*E-mail: a.rossin@iccom.cnr.it.
■
8695. (f) Tonigold, M.; Lu, Y.; Bredenkotter, B.; Rieger, B.;
̈
Bahnmuller, S.; Hitzbleck, J.; Langstein, G.; Volkmer, D. Angew.
̈
Chem., Int. Ed. 2009, 48, 7546−7550. (g) Lu, Y.; Tonigold, M.;
Notes
Bredenkotter, B.; Volkmer, D.; Hitzbleck, J.; Langstein, G. Z. Anorg.
̈
Allg. Chem. 2008, 634, 2411−2417.
The authors declare no competing financial interest.
(13) Wang, C.; Liu, D.; Lin, W. J. Am. Chem. Soc. 2013, 135, 13222−
13234.
ACKNOWLEDGMENTS
■
(14) Rossin, A.; Di Credico, B.; Giambastiani, G.; Peruzzini, M.;
Pescitelli, G.; Reginato, G.; Borfecchia, E.; Gianolio, D.; Lamberti, C.;
Bordiga, S. J. Mater. Chem. 2012, 22, 10335−10344.
(15) Perrin, D. D.; Armarego, W. L. F.; Perrin, D. R. Purification of
Laboratory Chemicals, 2nd ed.; Pergamon: Oxford, New York, 1980;
Vol. 1.
Thanks are given to the project FIRENZE HYDROLAB 2
supported by ECRF and to the U.S. Department of Energy
(Grant DE-FG-02-03ER15457). Dr. Vladimiro Dal Santo
(ISTM-CNR, Milano) is acknowledged for help with the GF-
AAS measurements. A.R. thanks Dr. Tiziano Montini
(16) In the case of 2,6-dimethylphenol, the standard was added after
the catalysis to avoid the inhibition of free radicals formation (and
related conversion decrease) caused by the same phenol.
(17) Perdew, J.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77,
3865−3868.
̀
(Universita di Trieste and ICCOM-CNR, Trieste Unit) for
fruitful discussions.
REFERENCES
■
(1) (a) Zhao, D.; Timmons, D. J.; Yuan, D.; Zhou, H.-C. Acc. Chem.
Res. 2011, 44, 123−133. (b) Chen, B.; Xiang, S.; Qian, G. Acc. Chem.
Res. 2010, 43, 1115−1124. (c) Li, J.-R.; Kuppler, R. J.; Zhou, H.-C.
Chem. Soc. Rev. 2009, 38, 1477−1504. (d) Britt, D.;
Tranchemontagne, D.; Yaghi, O. M. Proc. Natl. Acad. Sci. U.S.A.
2008, 105, 11623−11627. (e) Mueller, U.; Schubert, M.; Teich, F.;
Puetter, H.; Scherie-Arndt, K.; Pastre,
626−636.
(2) (a) Li, J.-R.; Sculley, J.; Zhou, H.-C. Chem. Rev. 2012, 112, 869−
932. (b) Li, J.-R.; Kuppler, R. J.; Zhou, H.-C. Chem. Soc. Rev. 2009, 38,
1477−1504.
(18) (a) Kresse, G.; Furthmuller, J. Comput. Mater. Sci. 1996, 6, 15−
̈
50. (b) Kresse, G.; Furthmuller, J. Phys. Rev. B 1996, 54, 11169−
̈
11186.
(19) Grimme, S. J. Comput. Chem. 2006, 27, 1787−1799.
(20) Kresse, G.; Joubert, D. Phys. Rev. B 1999, 59, 11−19.
(21) Monkhorst, H. J.; Pack, J. D. Phys. Rev. B 1976, 13, 5188−5192.
(22) Sanville, E.; Kenny, S. D.; Smith, R.; Henkelman, G. J. Comput.
Chem. 2007, 28, 899−908.
(23) Dudarev, S. L.; Botton, G. A.; Savrasov, S. Y.; Humphreys, C. J.;
Sutton, A. P. Phys. Rev. B 1998, 57, 1505−1509.
(24) Wang, L.; Maxisch, T.; Ceder, G. Phys. Rev. B 2006, 73, 195107.
(25) Koola, J. D.; Kochi, J. K. J. Org. Chem. 1987, 52, 4545−4553.
(26) The absolute structure (R or S) of the main enantiomer was not
determined.
́
J. J. Mater. Chem. 2006, 16,
(3) (a) Kreno, L. E.; Leong, K.; Farha, O. K.; Allendorf, M.; Van
Duyne, R. P.; Hupp, J. T. Chem. Rev. 2012, 112, 1105−1125.
(b) Shekhah, O.; Liu, J.; Fischer, R. A.; Woll, C. Chem. Soc. Rev. 2011,
40, 1081−1106.
(4) (a) Li, Y.-W.; Li, Y.-R.; Wang, L.-F.; Zhou, B.-Y.; Chen, Q.; Bu,
X.-H. J. Mater. Chem. A 2013, 1, 495−499. (b) Jayaramulu, K.;
Narayanan, R. P.; George, S. J.; Maji, T. K. Inorg. Chem. 2012, 51,
10089−10091. (c) Luo, F.; Batten, S. R. Dalton Trans. 2010, 39,
4485−4488.
(5) (a) Huxford, R. C.; Della Rocca, J.; Lin, W. Curr. Opin. Chem.
Biol. 2010, 14, 262−268. (b) Keskin, S.; Kizilel, S. Ind. Eng. Chem. Res.
2011, 50, 1799−1812.
(6) (a) Dhakshinamoorthy, A.; Garcia, E. Chem. Soc. Rev. 2012, 41,
5262−5284. (b) Yoon, M.; Srirambalaji, R.; Kim, K. Chem. Rev. 2012,
112, 1196−1231. (c) Lillerud, K. P.; Olsbye, U.; Tilset, M. Top. Catal.
2010, 53, 859−868.
(7) Caron, S.; Dugger, R. W.; Ruggeri, S. G.; Ragan, J. A.; Ripin, D.
H. B. Chem. Rev. 2006, 106, 2943−2989.
(8) Skobelev, I. Y.; Sorokin, A. B.; Kovalenko, K. A.; Fedin, V. P.;
Kholdeeva, O. A. J. Catal. 2013, 298, 61−69.
(27) Neuenschwander, U.; Hermans, I. J. Org. Chem. 2011, 76,
10236−10240.
(28) (a) Quek, X.-Y.; Tang, Q.; Hu, S.; Yang, Y. Appl. Catal., A 2009,
361, 130−136. (b) Zhang, X.; Zeng, C.; Zhang, L.; Xu, N. Kinet. Catal.
2009, 50, 199−204. (c) Lu, X. H.; Xia, Q. H.; Fang, S. Y.; Xie, B.; Qi,
B.; Tang, Z. R. Catal. Lett. 2009, 131, 517−525. (d) Salavati-Niasari,
M.; Abdolmohammadi, S.; Oftadeh, M. J. Coord. Chem. 2008, 61,
2837−2851. (e) Jiang, J.; Li, R.; Wang, H.; Zheng, Y.; Chen, H.; Ma, J.
Catal. Lett. 2008, 120, 221−228. (f) Jinka, K. M.; Sebastian, J.; Jasra, R.
V. J. Mol. Catal. A: Chem. 2007, 274, 33−41. (g) Patil, M. V.; Yadav,
M. K.; Jasra, R. V. J. Mol. Catal. A: Chem. 2007, 277, 72−80.
(h) Kantam, M. L.; Rao, B. P. C.; Reddy, R. S.; Sekhar, N. S.; Sreedhar,
B.; Choudary, B. M. J. Mol. Catal. A: Chem. 2007, 272, 1−5. (i) Tang,
Q. H.; Zhang, Q. H.; Wu, H. L.; Wang, Y. J. Catal. 2005, 230, 384−
397. (j) Tang, Q. H.; Wang, Y.; Liang, J.; Wang, P.; Zhang, Q. H.;
Wan, H. L. Chem. Commun. 2004, 440−441.
(29) Holleman, A. F.; Wiberg, E. Lehrbuch der Anorganischen Chemie,
101st ed.; De Gruyter: Berlin, New York, 1995.
(30) Waugh, K. C. Appl. Catal. 1988, 43, 315.
(31) Breck, D. W. Zeolite Molecular Sieves; John Wiley & Sons: New
York, 1974, p 636.
(32) Yu, D.; Yazaydin, A. O.; Lane, J. R.; Dietzel, P. D. C.; Snurr, R.
(9) (a) Song, F.; Wang, C.; Falkowsky, J. M.; Ma, L.; Lin, W. J. Am.
Chem. Soc. 2010, 132, 15390−15398. (b) Cho, S.; Ma, B.; Nguyen, S.
T.; Hupp, J. T.; Albrecht-Schmitt, T. E. Chem. Commun. 2006, 2563−
2565.
(10) (a) Balu, A. M.; Lin, C.; Liu, H.; Li, Y.; Vargas, C.; Luque, R.
Appl. Catal., A 2013, 455, 261−266. (b) Dhakshinamoorthy, A.;
Alvaro, M.; Garcia, H. J. Catal. 2012, 289, 259−265.
Q. Chem. Sci. 2013, 4, 3544−3556.
(33) (a) Bloch, E. D.; Murray, L. J.; Queen, W. L.; Chavan, S.;
Maximoff, S. N.; Bigi, J. P.; Krishna, R.; Peterson, V. K.; Grandjean, F.;
Long, G. J.; Smit, B.; Bordiga, S.; Brown, C. M.; Long, J. R. J. Am.
Chem. Soc. 2011, 133, 14814−14822. (b) Greenwood, N. N.;
Earnshaw, A. Chemistry of the Elements, 2nd ed.; Butterworth
Heinemann: Burlington, MA, 2002, pp 604−616,
́ ́
(11) (a) Luz, I.; Leon, A.; Boronat, M.; Llabres i Xamena, F. X.;
Corma, A. Catal. Sci. Technol. 2013, 3, 371−379. (b) Brown, K.;
Zolezzi, S.; Aguirre, P.; Venegas-Yazigi, D.; Paredes-Garcia, V.; Baggio,
R.; Novak, M. A.; Spodine, E. Dalton Trans. 2009, 1422−1427.
(c) Wu, Y.; Qiu, L.; Wang, W.; Li, Z.; Xu, T.; Wu, Z.; Jiang, X.
Transition Met. Chem. 2009, 34, 263−268.
(34) Miskowski, V. M.; Robbins, J. L.; Treitel, I. M.; Gray, H. B.
Inorg. Chem. 1975, 14, 2318−2321.
(12) (a) Hamidipour, L.; Farzaneh, F. React. Kinet. Mech. Catal. 2013,
109, 67−75. (b) Zhang, J.; Biradar, A. V.; Pramanik, S.; Emge, T. J.;
Asefa, T.; Li, J. Chem. Commun. 2012, 48, 6541−6543. (c) Beier, M. J.;
Kleist, W.; Wharmby, M. T.; Kissner, R.; Kimmerle, B.; Wright, P. A.;
Grunwaldt, J.-D.; Baiker, A. Chem.Eur. J. 2012, 18, 887−898. (d) Fu,
Y.; Sun, D.; Qin, M.; Huang, R.; Li, Z. RSC Adv. 2012, 2, 3309−3314.
(e) Tonigold, M.; Lu, Y.; Mavrandonakis, A.; Puls, A.; Staudt, R.;
1039
dx.doi.org/10.1021/cs401003d | ACS Catal. 2014, 4, 1032−1039