Molybdenum versus tungsten for the epoxidation of cyclooctene catalyzed by [Cp*2M2O5]

Article abstract of DOI:10.1002/ejic.201300260

The catalytic activity of [Cp*₂M₂O₅] (M = Mo, W; Cp* = pentamethylcyclopentadienyl) for the homogeneous epoxidation of a solution of cyclooctene in MeCN/toluene follows the order Mo >> W when using tert-butyl hydroperoxide (TBHP)/decane as oxidant, in contrast to the inverse order (W >> Mo) when using aqueous H ₂O₂ as oxidant. The catalytic activity for the Mo system strongly depends on the solvent used to deliver the oxidant (TBHP/decane >> TBHP/H₂O). The low activity of the W system is also decreased when using TBHP/water in place of TBHP/decane. For both metals, H ₂O₂/H₂O is a better oxidant than TBHP/H ₂O. However, whereas the Mo-based catalyst is much more active for the TBHP/decane epoxidation in spite of the lower TBHP oxidizing power (TBHP/decane > H₂O₂/H₂O > TBHP/H ₂O), the W-based system is much more active for the H ₂O₂/H₂O epoxidation in spite of the negative effect of water (H₂O₂/H₂O > TBHP/decane > TBHP/H₂O). The kinetic profile of the TBHP/decane epoxidation process is affected by product inhibition. Initial rate measurements show that the rate law is first order with respect to substrate and has saturation behavior with respect to the oxidant. Copyright

Full text of DOI:10.1002/ejic.201300260

FULL PAPER
DOI: 10.1002/ejic.201300260
Molybdenum versus Tungsten for the Epoxidation of Cyclooctene Catalyzed by
[
Cp* M O ]
2
2
5
Pelin Sözen-Akta s¸ ,[a,b,c] Eric Manoury,[a,b] Funda Demirhan, and Rinaldo Poli*[a,b,d]
[c]
Keywords: Homogeneous catalysis / Epoxidation / Molybdenum / Tungsten / Cycloalkenes
The catalytic activity of [Cp*
2
M
2
O
5
] (M = Mo, W; Cp* =
whereas the Mo-based catalyst is much more active for the
TBHP/decane epoxidation in spite of the lower TBHP oxidiz-
pentamethylcyclopentadienyl) for the homogeneous epoxid-
ation of a solution of cyclooctene in MeCN/toluene follows
the order Mo ϾϾ W when using tert-butyl hydroperoxide
ing power (TBHP/decane Ͼ H
W-based system is much more active for the H
idation in spite of the negative effect of water (H
TBHP/decane Ͼ TBHP/H O). The kinetic profile of the
2
O
2
/H
2
O Ͼ TBHP/H
2
O), the
O epox-
/H O Ͼ
2
O
2
/H
2
(
TBHP)/decane as oxidant, in contrast to the inverse order (W
ϾϾ Mo) when using aqueous H as oxidant. The catalytic
activity for the Mo system strongly depends on the solvent
used to deliver the oxidant (TBHP/decane ϾϾ TBHP/H O).
2
O
2
2
2
O
2
2
TBHP/decane epoxidation process is affected by product in-
hibition. Initial rate measurements show that the rate law is
first order with respect to substrate and has saturation behav-
ior with respect to the oxidant.
2
The low activity of the W system is also decreased when
using TBHP/water in place of TBHP/decane. For both metals,
2 2 2 2
H O /H O is a better oxidant than TBHP/H O. However,
Introduction
vestigations were then reported by the groups of Romão,
[
1c,6]
Kühn, Gonçalves, and Abrantes,
the notable features
Epoxidation catalysis continues to attract a lot of atten-
tion in the academic[1] and industrial research laboratories
in view of the versatility of the epoxide products as reaction
intermediates in the fine-chemical and polymer industries.
New emphasis on this process comes from the conversion
of new substrates obtained from biomass[ and from the
desire for greener catalytic protocols for the transformation
of traditional fossil resources.[
being the very high initial activities achieved for certain sys-
tems in ionic-liquid media, and the possibility to anchor the
catalyst on a solid support with a certain degree of recycl-
ability. The major problems of these systems are their deac-
tivation, which seems to be caused at least in part by prod-
[2]
3]
uct inhibition, and their lower efficiency in the presence of
[6g]
water or when H O is used as oxidant.
In many cases,
2
2
4]
however, the shortcomings of H O could be simply related
2
2
In recent years, the catalytic performance of half-sand-
wich Group 6 (mostly Mo) complexes has been of particu-
lar interest in a few laboratories, including ours. The cata-
to mass-transport problems in the biphasic systems.
We first observed and reported that the dinuclear com-
pounds [Cp* M O ] (M = Mo, W) are excellent catalysts
2
2
5
lytic activity of [Cp*MoO Cl] (Cp* = pentamethylcyclopen-
2
for the oxidation of thiophene derivatives to the corre-
tadienyl) in olefin epoxidation by tert-butyl hydroperoxide
sponding sulfoxides and sulfones by aqueous H O2 in
2
5]
(
TBHP) was first described by Bergman et al.[ Many in-
MeCN. Not only was this the first report of the use of half-
VI
sandwich W systems in catalytic oxidation, but the study
also unexpectedly revealed that the W system is approxi-
mately 100 times more active than the Mo system.[ The
study further revealed, through kinetic investigations on the
W system, that the catalyst does not degrade during the
process, and that the reaction rate is first order with respect
to the substrate and zero order with respect to H O . There-
7]
[
a] CNRS, LCC (Laboratoire de Chimie de Coordination),
2
05 route de Narbonne, BP 44099, 31077 Toulouse Cedex 4,
France
Fax: +33-5-61553003
E-mail: rinaldo.poli@lcc-toulouse.fr
Homepage: http://www.lcc-toulouse.fr/equipe_g/pages/poli/
index.html
2
2
fore, half-sandwich Mo and W systems are highly capable
of working with H O /H O as oxidant, provided the system
2
2
2
[b] Université de Toulouse, UPS, INPT,
is homogeneous, that is, that substrate, oxidant, and catalyst
31077 Toulouse Cedex 4, France
[
c] Celal Bayar University, Faculty of Sciences & Liberal Arts, are in the same phase. However, additional experiments run
Department of Chemistry,
with the more active W catalyst have revealed that water
has a negative effect on the reaction rate when added in
large amounts to the catalytic mixture.
45030 Muradiye-Manisa, Turkey
[d] Institut Universitaire de France,
1
03, bd. Saint-Michel, 75005 Paris, France
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201300260.
2728
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2013, 2728–2735

2 2 5
Epoxidation of Cyclooctene Catalyzed by [Cp* M O ]
Encouraged by these results, the same catalytic system Results and Discussion
(
[Cp* M O ] in MeCN) was investigated for the epoxid-
2
2
5
ation of cyclooctene by aqueous H O . The only difference Solvent Effect
2
2
relative to the thiophene-oxidation study was the need to
add toluene as an inert cosolvent (1:3 relative to MeCN) to
maintain the solution homogeneity (cyclooctene is not fully
miscible in MeCN at the concentration needed for the ex-
periments). The results paralleled those of the oxidation of
the thiophene derivatives: the oxidation was efficient, selec-
tive (no products other than cyclooctene oxide were de-
tected), and much faster for the W system, a factor of ap-
proximately 50 being measured in this case.[ Once again,
A preliminary investigation of catalytic activity for the
cyclooctene epoxidation by TBHP/decane with H NMR
spectroscopic monitoring was carried out at room tempera-
ture with 1% precatalyst (or 2% metal) to screen solvents.
The investigation included chloroform, acetone, acetoni-
trile, and methanol, and the results are shown in Table 1.
1
Table 1. Cyclooctene oxide yield for the cyclooctene epoxidation
8]
2 2 5
by TBHP catalyzed by 1% [Cp* M O ] (M = Mo or W) at room
the kinetics investigation indicated a first-order dependence temperature in different solvents.[
a]
on the oxidized substrate (cyclooctene in this case) and
Solvent
M
t
Cyclooctene oxide yield [%]
zero-order dependence in H O . Further investigations con-
2
2
CDCl
CDCl
3
3
Mo
W
Mo
W
Mo
W
Mo
W
20.5 h
14 d
7 d
7d
5 d
6 d
8 d
2 d
84.5
0.02
–
–
47.1
–
firmed the expected first-order dependence on the catalyst,
as well as the stability of the catalytic system against deacti-
vation. Like in the thiophene-oxidation study, the addition (CD ) CO
(CD
3
)
2
CO
3
2
MeOD
MeOD
of large amounts of water revealed a negative effect on the
catalytic activity. Finally, a study at variable temperature in
the 25–55 °C range for the W system allowed the derivation
CD
CD
3
CN
CN
89.0
–
3
‡
–1
‡
of the activation parameters [ΔH = 10.2(6) kcalmol ; ΔS
–32(2) e.u.] for the catalytic cycle. This study also in-
[
5
a] Conditions: Cyclooctene (0.1 mL, 0.78 mmol), TBHP (0.28 mL;
.5 m in decane, 1.56 mmol), [Cp* ] (4.2 mg for Mo; 5.6 mg
=
2 2 5
M O
cluded a comparative computational investigation of the
catalytic cycle for the Mo and W systems on the basis of the
hypothesis that the reaction follows a pathway that involves
–3
for W; 7.8ϫ10 mmol), and deuterated solvent (0.3 mL).
The immediate striking observation is the much greater
VI
VI
oxidant activation by protonation of an M=O function to efficiency of the dinuclear Mo compound than the W
α
β
yield an {M(OH)(O O H)} intermediate, followed by trans- analogue, which is in the inverse order relative to the use of
α
fer of the O atom to the substrate. This pathway had al- aqueous H O epoxidation in MeCN/toluene (see above).
2
2
[
9]
ready been previously calculated for the Mo system, with Chloroform is the best solvent for this reaction and gives a
only small changes relative to a proposition made by Thiel very high yield in the shortest time (84.5% in 20.5 h), but
[
10]
et al. and inspired by the original hypothesis formulated only for the Mo compound. The W analogue proves very
by Sharpless et al.[ The calculations indeed showed that inefficient under the same conditions with only trace
the barrier of the catalytic cycle is slightly smaller for the amounts of product measured after 14 days. The result for
W system. The explanation for the better catalytic activity [Cp* Mo O ] in this solvent reproduces that reported in a
11]
2
2
5
[
6g]
of the W system was attributed to the greater oxophilicity previous contribution, including the observed formation
and Lewis acidity of the W center, which interacts more of a white precipitate, the nature of which is unknown.
β
α
strongly with the O atom in the transition state of the O - Thus, although this reaction is initially very fast and homo-
transfer process.
geneous, it rapidly turns into a heterogeneous system. The
The above-mentioned comparative calculations[8] were yield and conversion are in good correspondence, and the
run only for H O as the oxidant, whereas those previously mass balance is excellent (see details in Figure S1 of the
2
2
published for the Mo system were also run for MeOOH as a Supporting Information). The catalytic mixtures in all other
computational model for TBHP. The profiles are essentially investigated solvents, on the other hand, remained homo-
identical, although the barrier for the MeOOH system is geneous.
marginally higher. The validity of these calculations and of
There was no conversion at all in (CD ) CO for both
3 2
the mechanistic hypothesis would lead us to predict that catalysts. A negative effect of solvent coordination through
[
Cp* W O ] should also be
a
better catalyst than the carbonyl oxygen atom, which blocks the Lewis acidic
2
2
5
[
Cp* Mo O ] when using TBHP as oxidant. We have there- sites of the catalyst, does not seem a reasonable rationaliza-
2
2
5
fore investigated the relative activity of the two compounds tion for this lack of catalytic activity, because the same phe-
for the epoxidation with TBHP/decane and found, contrary nomenon should then also occur in MeCN and MeOH. On
to the above prediction, that the Mo system is a much bet- the other hand, the reaction takes place in these solvents,
ter catalyst under these conditions. Along this investigation, although more slowly than in chloroform, but once again
we have observed dramatic effects on the rate of the cata- much more rapidly for the Mo-containing catalyst (no sig-
lytic process depending on the nature of the metal (Mo ver- nificant conversion was measured for the W-based catalyst
sus W), of the oxidant (TBHP versus H O ), and of the under these conditions). Plots of yield and conversion ver-
2
2
solvent that delivers the oxidant to the substrate solution sus time for the Mo-based catalyst in these two solvents,
(
decane versus H O). We report here the results of this in- obtained from 1H NMR spectroscopic monitoring, are
2
vestigation.
given in the Supporting Information (Figures S2 and S3).
Eur. J. Inorg. Chem. 2013, 2728–2735
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2729

P. Sözen-Akta s¸ , E. Manoury, F. Demirhan, R. Poli
FULL PAPER
Amongst these two solvents, CD CN affords the most ef- and one for the oxidant, depending on the position of the
3
ficient catalytic system, with nearly complete conversion af- catalyst–oxidant association equilibrium (saturation kinet-
[
12]
ter eight days and very high selectivity, with cyclooctene ics). We believe that the kinetics are affected by an inter-
epoxide being the only observed product.
fering phenomenon that makes the olefin consumption look
like a second-order decay. One possibility is a dampening
effect of the catalytic activity by the produced tert-butanol,
namely, a product-inhibition phenomenon. Indeed, it has
been previously highlighted that certain Mo-catalyzed
epoxidations are product-inhibited, with a slowing of the
Kinetics Study
All subsequent catalytic experiments were carried out in
a 3:1 CH CN/toluene combination to keep the reaction
3
[13]
catalytic activity as the epoxide product accumulates.
mixture homogeneous at high substrate concentrations. The
catalytic runs were carried out at 55 °C, and the reactions
were followed by gas chromatography with detection and
quantification of the cyclooctene and cyclooctene oxide
peaks. The conditions used are identical to those of the pre-
[8]
vious study with the H O oxidant. A first run was carried
2
2
out under pseudo-first-order conditions (TBHP/cyclooc-
tene = 8:1), which allowed in principle the determination of
the reaction order in cyclooctene. The GC monitoring
showed that cyclooctene oxide was the terminal product
with no additional visible peaks of byproducts. There was Scheme 1. Kinetic scheme and rate law for the Mo-catalyzed olefin
epoxidation.[
12]
a good mass balance (Ͼ90%), with the amount of produced
cyclooctene oxide each time being nearly the same as the
To avoid the complication of product inhibition, we car-
amount of consumed cyclooctene. Under these conditions,
ried out an investigation of initial rates: (a) on the one
essentially all the cyclooctene was completely converted
hand, with variable concentration of cyclooctene at con-
into the oxide product in approximately three hours (see
stant TBHP concentration to determine the reaction order
Figure 1).
in cyclooctene; and (b) on the other hand, with variable
concentrations of TBHP at constant cyclooctene concentra-
tion to determine the reaction order in TBHP. All experi-
ments were carried out with an excess amount of TBHP
and the initial rates were measured on the basis of the cy-
clooctene consumption (see details in Figures S6–S11 of the
Supporting Information), during a time interval in which
the TBHP concentration can be assumed to remain approx-
imately constant, even for the experiments run with a low
[TBHP]/[cyclooctene] ratio. The initial rate values are re-
ported in Table 2 and are represented graphically in Fig-
ure 2. Once again, the reaction selectivity was excellent,
with a good mass balance, and product inhibition was ob-
served from the first-order kinetic analyses. From part (a)
of the study, it is clear that the data are in better agreement,
as expected, with a first-order dependence of the reaction
Figure 1. GC monitoring of the epoxidation of cyclooctene by rate on the olefin concentration and not with a second-or-
TBHP (5.5 m in decane) in CD
Cp* Mo ]. Conditions: catalyst (6.4 mg, 0.0119 mmol), cyclo-
octene (0.16 mL, 1.19 mmol), TBHP (1.72 mL; 5.5 m in decane,
.50 mmol), and CH CN/toluene (6.05 mL, 3:1) at 55 °C.
3
CN/toluene catalyzed by 1% der dependence, as suggested by the kinetic analysis of the
[
2
2 5
O
entire oxidation process. From part (b), the rates show a
dependence on TBHP between zero and one, which sug-
9
3
Kinetic analysis shows that the cyclooctene decay does
Table 2. Initial rates of cyclooctene oxidation by TBHP catalyzed
not fit a first-order rate law as expected (Figure S4 in the
Supporting Information). It fits a second-order decay law
better (Figure S5 in the Supporting Information). However,
such rate-law dependence is unexpected and unreasonable,
because it would indicate that two molecules of cyclooctene
must be involved at the level of the rate-determining transi-
tion state. The generally accepted kinetic scheme for this
process and the corresponding rate law are as shown in
Scheme 1, which predicts that the reaction should always
2 2 5
by 1% [Cp* Mo O ] at different cyclooctene and TBHP concentra-
tions.
[a]
[
TBHP]
0
= 1.5 m;
[cyclooctene]
[Cp* Mo ] = 2.5ϫ10
Run [TBHP] [m] v0 [ms ]
0
= 0.25 m;
–3
–3
[Cp*
2
Mo ] = 2.0ϫ10
2
O
5
m
2
2 5
O
m
–1 [b]
–1 [b]
Run [cyclooctene] [m] v [ms ]
0
0
0
–4
–4
–5
–5
–5
1
2
3
0.199
0.412
0.769
4.61ϫ10
10.6ϫ10
4
5
6
0.496
0.911
1.563
9.39ϫ10
12.6ϫ10
15.0ϫ10
14.7ϫ10–4
3
[a] Solvent = 3:1 CH CN/toluene, T = 55 °C. [b] Rates measured
be first order in alkene and with an order between zero from the initial slope (5 min) of the [epoxide] versus time plot.
2
730 Eur. J. Inorg. Chem. 2013, 2728–2735
www.eurjic.org
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

2 2 5
Epoxidation of Cyclooctene Catalyzed by [Cp* M O ]
gests a balance in the substrate–oxidant adduct formation adduct with MeOOH (used as a model of tBuOOH),
equilibrium and is in agreement with saturation behavior [Cp*MoO(OH)(OOMe)Cl], lies at higher energy than the
(
Scheme 1; K [TBHP] not varying much from 1). A fit of corresponding adduct with H O , [Cp*MoO(OH)(OOH)-
1
2
2
[
9]
these data to the rate equation of Scheme 1 for the indepen- Cl]. Therefore, the adduct formation equilibrium is ex-
dent determination of k and K is prevented by the limited pected to lie less extensively on the left-hand side for TBHP
2
1
number of points and by the approximate values of the (Figure 3, case A) and lead to the observed behavior in Fig-
measured initial rates. ure 2, b, whereas H O is able to further displace the adduct
2
2
formation equilibrium (Figure 3, case B) and give an ob-
served zero-order dependence on H O . This difference can
2
2
be easily understood in terms of steric bulk, with the bigger
tBu group destabilizing the catalyst–oxidant adduct.
Another interesting question is why there is product inhi-
bition when using TBHP, whereas the cyclooctene concen-
[8]
tration follows clean first-order decay when using H O .
2
2
In the TBHP system, product inhibition can be related to
coordination of the tert-butanol, the reaction byproduct
generated from TBHP, to a catalyst coordination site,
thereby making it less available to operate the catalytic cy-
cle, or to formation of a catalyst–tert-butanol adduct that
has greater stability than the catalyst resting state. In each
case, an energetically more stable off-loop species would
form after the beginning of the reaction, gradually increas-
ing the energy span of the catalytic cycle. When the oxidant
is aqueous H O , however, the byproduct of the reaction
2
2
is water, which is already present in the medium from the
beginning of the reaction. Therefore, a retardation effect is
not expected.
The temperature dependence of the catalytic reaction was
investigated in the temperature range 35–65 °C under stan-
dard conditions, namely, with 1% catalyst and with a
[TBHP]/[cyclooctene] ratio of 10 to insure pseudo-first-or-
der conditions. The kinetic profiles are given in Figure S12
in the Supporting Information. The cyclooctene decay was
again affected by retardation at high conversions. Hence,
the observed pseudo-first-order rate constants were ob-
Figure 2. Dependence of the initial rate of cyclooctene oxidation tained from the first-order decay analysis restricted to the
by TBHP catalyzed by 1% [Cp*
constant [TBHP] and [Cp* Mo
cyclooctene] and [Cp* Mo ].
2
Mo
2 5
O ] on (a) [cyclooctene] at
initial phase of the reaction (Ͻ40 min; see Figure S13 in the
Supporting Information). The observed rate constants are
collected in Table 3.
2
2 5
O ], and (b) [TBHP] at constant
[
2
2 5
O
Figure 3 shows a qualitative energy profile that fits the
The subsequent Eyring analysis of the data is not
above-determined trends. Previously reported DFT calcula- straightforward because of the saturation behavior. The ob-
tions run on the [Cp*MoO Cl] catalyst first used by tained kobsd./[cat.] values reported in Table 3 correspond to
2
[
5]
Bergman et al. (which can also be considered to be a the expression {k K [TBHP]/(1 + K [TBHP])}. According
2
1
1
model of the [Cp* Mo O ] catalyst) have shown that the to transition-state theory, the rate constant k of a bimolecu-
2
2
5
Figure 3. Energy profiles of two limiting rate laws for the kinetic scheme shown in Scheme 1.
Eur. J. Inorg. Chem. 2013, 2728–2735
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2731

P. Sözen-Akta s¸ , E. Manoury, F. Demirhan, R. Poli
FULL PAPER
Table 3. Rate constants for the [Cp*
of cyclooctene by TBHP at different temperatures.
2
Mo
2
O
5
]-catalyzed epoxidation as the enthalpic barrier from the pre-equilibrium situation
[a]
to the rate-determining transition state, the large positive
‡
°
–
1
obsd./[cat.] [s–1 m–2]
value of ΔS* (ΔS with some contribution from ΔS ) con-
T [°C]
k
obsd. [s ]
k
2
1
‡
1.17ϫ10–4
trasts with the large negative ΔS value that was previously
determined for the [Cp* W O ]-catalyzed cyclooctene epox-
35
45
55
65
0.468
1.43
2.46
4.88
–4
3.57ϫ10
2
2
5
6.16ϫ10–
4
idation by H O (–32 calmol K ). The latter is consis-
–1 –1 [8]
2
2
1.22ϫ10–3
Mo
0.048 m; [TBHP] = 0.48 m; solvent: MeCN/toluene (2:1).
tent with the associative nature of the transition state rela-
(Figure 3,
] = 4.8ϫ10–4 m; [cyclooctene] tive to the catalyst–oxidant adduct and H
O
[
=
a] Reaction conditions: [Cp*
2
2
O
5
2
2
case B). Therefore, under the reasonable hypothesis that
‡
ΔS for the oxidation by TBHP is also negative, the ob-
2
lar collision can be expressed as shown in Equation (1). For served positive value of ΔS* must originate from a large
the reaction analyzed here, this would be appropriate for positive contribution of the ΔS° term. This could result, for
the second step of the process in Scheme 1 (k ), the reaction instance, from the presence of several interacting solvent
2
between the catalyst–oxidant adduct and cyclooctene. How- molecules (e.g., by hydrogen bonding) in the initial reagent
ever, the measured kobsd./[cat.] also incorporates the tem- mixture (cyclooctene + TBHP + catalyst) that need to dis-
perature dependence of the term K [TBHP]/(1
+
sociate to allow activation of the oxidant molecule.
1
K [TBHP]), which depends on the equilibrium of the first
1
step. This term could be eliminated under saturation condi-
tions (K [TBHP] ϾϾ 1), in which kobsd./[cat.] simplifies to
1
Effect of Water
k . Under the opposite limiting conditions (K [TBHP] ϽϽ
2
1
1
), kobsd./[cat.] would simplify to (k K [TBHP]), and a
The observed opposite relative order of catalytic activity,
2
1
modified Eyring expression could be written for the re- under the same experimental conditions (MeCN/toluene
sulting (k K ) values as Equation (2), in which the lnK
solvent mixture), for the epoxidation of cyclooctene by
term has been developed according to van’t Hoff’s expres- TBHP/decane in the present study (Mo ϾϾ W) and by
2
1
1
[
8]
sion of the equilibrium constant temperature dependence. H O /H O in the previous study (Mo ϽϽ W) is an inter-
2
2
2
Under these conditions, the measured thermodynamic pa- esting puzzle, which motivated us to run additional experi-
rameters would correspond to the difference between the ments to arrive at a mechanistic interpretation. Before con-
rate-determining transition state and the (cyclooctene + cat- templating a possible specific mechanistic difference in-
alyst + TBHP) mixture. In an intermediate situation, as is duced by the nature of R in the oxidant ROOH (R = H or
°
experimentally observed, the thermodynamic ΔH and tBu), it was necessary to examine the possible effect of the
1
°
ΔS values of the pre-equilibrium give an intermediate con- solvent that delivers ROOH to the olefin solution. H O2
1
2
tribution to the measured activation parameters from the was delivered as an aqueous solution, whereas TBHP was
Eyring analysis. Unfortunately, under our experimentally delivered as the commercially available solution in decane.
accessible conditions of solvent and reagent concentrations, It is therefore possible that the presence of water negatively
it was not possible to access either limiting situation of the affects the catalytic activity of the Mo system to a greater
rate law. The Eyring plot (shown in Figure 4) yields a good extent relative to the W system.
linear correlation, and the values determined from the slope
Since TBHP is also commercially available as an aqueous
(
ΔH*) and intercept (ΔS*) are 15.1 kcalmol–1 and solution, we carried out a comparative study of the cyclooc-
–
1
–1
37.4 calmol K , respectively.
tene epoxidation by TBHP/H O under analogous experi-
2
mental conditions used for the previous studies. The result
of this investigation (see Figure 5) clearly shows that water
has a dramatic negative effect on the catalytic activity. The
‡
‡
‡
ln[(kh)/(k
B
T)] = –ΔG /RT = –ΔH /RT + ΔS /R
(1)
= –ΔG‡
/RT – ΔG°
ln[(k
2
K
1
‡
h)/(k
B
T)] = ln[(k
)/RT + (ΔS
2
h)/(k
B
T)] + lnK
1
2
1
/RT
(2)
+ ΔH°
‡
+ ΔS°
selectivity in epoxide remains high for the TBHP/H O
2
=
–(ΔH
2
1
2
1
)/R
epoxidation (excellent mass balance). Therefore, the lower
epoxide yield must be the result of a reduced catalyst ac-
tivity in the presence of water. The activity for the oxidation
by TBHP/H O seems even lower (ca. 12% conversion after
2
3
0 h; see Figure 5) than that of the oxidation by H O /H O
2 2 2
(ca. 60% conversion after 25 h under the same conditions
of temperature, reagents, catalyst concentrations, and sol-
[
4,7]
vent).
This comparison clearly shows that H O is the
2 2
better oxidant. In fact, there is even more water in the sys-
tem for the faster oxidation with H O (30% in H O) than
2
2
2
in the slower oxidation with TBHP (70% in water). Use
of TBHP/H O also slightly reduces the catalytic activity of
2
Figure 4. Eyring plot of the kinetic data in Table 3.
[
Cp* W O ] relative to TBHP/decane, for which the activity
2 2 5
Whereas the ΔH* value (essentially equivalent to ΔH 2‡ is already quite low (see Figure S14 in the Supporting Infor-
°
with some contribution from ΔH ) can be easily interpreted mation). Therefore, the much lower activity of the W-based
1
2732
www.eurjic.org
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2013, 2728–2735

2 2 5
Epoxidation of Cyclooctene Catalyzed by [Cp* M O ]
catalyst in the epoxidation by TBHP/decane versus H O / dation (Figure S14 in the Supporting Information), and
2
2
[
8]
H O is attributable to the nature of the oxidant (TBHP ϽϽ earlier for the H O oxidation, that water has a negative
2
2
2
H O ), as the nature of the solvent is expected to give a effect on the W-catalyzed process, just like for the Mo-cata-
2
2
positive contribution (decane Ͼ water), which is, however, lyzed process. However, whereas the negative effect of re-
insufficient for this catalyst to inverse the effect. The nega- placing decane with water overrules the positive effect of
tive effect of water was already pointed out for the replacing TBHP with H O when using the Mo-based cata-
2
2
[8]
[
Cp* W O ] catalyst in the epoxidation by H O .
lyst, the opposite is true for the W-based catalyst. This op-
posite behavior might simply be related to a different in-
hibiting effect of water for the two catalysts (much stronger
for Mo than for W), or to a different rate-enhancement fac-
tor upon going from TBHP to H O (much stronger for W
2
2
5
2
2
2
2
than for Mo) or to the combination of both effects. It is
also possible, however, that competing reaction mechanisms
also affect the system. The possibility of a new catalytic
epoxidation pathway, which is specific for H O , has been
2
2
[
9,14]
presented in recent computational studies.
This path-
way still requires a first step of oxidant activation to form
α
β
the same activated {M(OH)(O O H)} intermediate, but
β
then involves transfer of the O atom to the incoming olefin
Figure 5. Cyclooctene epoxide yield versus time for the substrate, concomitant with a 1,2 shift of the proton from
β
α
[10]
[
Cp*
and TBHP/water (diamonds) in 3:1 MeCN/toluene at T = 55 °C.
Concentrations for the TBHP/decane run: [Cp* Mo
.83ϫ10–3 m; [cyclooctene] = 0.734 m; [TBHP] = 1.457 m. Concen-
2 2 5
Mo O ]-catalyzed epoxidation with TBHP/decane (squares) O to O , instead of the commonly proposed transfer of
α
the O atom. This new mechanism clearly cannot take place
2
2 5
O ] =
α
β
for the corresponding {M(OH)(O O tBu)} intermediate.
The calculations have shown that this mechanism has a
much lower activation barrier for the W catalyst.
1
–
3
2 2 5
trations for the TBHP/water run: [Cp* Mo O ] = 1.99ϫ10 m;
[cyclooctene] = 0.798 m; [TBHP] = 4.13 m.
Activity Comparisons
The [Cp* M O ]-catalyzed cyclooctene epoxidation
Conclusion
2
2
5
shows complex behavior under identical conditions except
In the present investigation of the [Cp* M O ]-catalyzed
2
2
5
for the nature of R in the ROOH oxidant and the small olefin epoxidation using the model cyclooctene substrate,
amount of solvent (decane versus water) that delivers we have reported new experimental results that reveal that
ROOH to the substrate solution in MeCN/toluene. The dramatic effects are induced by the nature of the oxidant
present study shows that the Mo system is a much more (H O versus TBHP), of the solvent that delivers the oxi-
2
2
efficient catalyst than the W system when using TBHP/dec- dant (decane versus H O), and of the metal in the precata-
2
ane, whereas previous studies have shown that the opposite lyst (Mo versus W). Under identical experimental condi-
[8]
is true when using H O /H O. Under comparable condi- tions, the process is much faster for Mo than for W when
2
2
2
tions (solvent, temperature, concentrations, oxidant/sub- using TBHP/decane as oxidant. Use of TBHP/water re-
strate ratio), the rate in the presence of the Mo catalyst duces the activity dramatically. However, as previously re-
[
8]
follows the order TBHP/decane ϾϾ H O /H O. On the ba- ported, the W system is much more active than the Mo
2
2
2
sis of the result of the last experiment (see above), this trend analogue when using H O /H O, whereas water still nega-
2
2
2
can be easily rationalized as a consequence of the presence tively affects the catalytic activity. The negative effect of
of water in the oxidant solution, since a dramatic loss of water is easily rationalized as an inhibiting effect and is
activity is observed upon going from TBHP/decane to qualitatively common to both metals, although quantitative
TBHP/water. This trend can be rationalized quite simply on differences might exist. A comparison of activity shows that
the basis of the catalyst inhibition by water, similar to the H O is a better oxidant than TBHP, and this result is also
2
2
product inhibition observed in the TBHP reaction. As qualitatively common to both metal catalysts. However,
pointed out above, the H O epoxidation does not show whereas the oxidant change is more important for the W-
2
2
product inhibition because the byproduct water is already based catalyst (H O /H O ϾϾ TBHP/decane Ͼ TBHP/
2
2
2
present in the system from the very beginning of the reac- H O), the nature of the solvent that delivers the oxidant is
2
tion, but this presence is reflected dramatically in the reac- more important for the Mo-based catalyst (TBHP/decane
tion rate.
ϾϾ H O /H O Ͼ TBHP/H O). The intervention of a new
2 2 2 2
The opposite trend, however, is observed for the W sys- mechanism, which is possible only for H O and which
2
2
tem. Under comparable conditions (solvent, temperature, seems much more favored for the W-based catalyst, possibly
concentrations, oxidant/substrate ratio), the rate in the pres- contributes to yield the observed effect. The rate depen-
ence of the W catalyst follows the order H O /H O ϾϾ dence on the oxidant (intermediate saturation behavior for
2
2
2
TBHP/decane. It has been shown here for the TBHP oxi- TBHP) is consistent with the generally accepted mechan-
Eur. J. Inorg. Chem. 2013, 2728–2735
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2733

P. Sözen-Akta s¸ , E. Manoury, F. Demirhan, R. Poli
FULL PAPER
2
483–2491; c) C. Freund, M. Abrantes, F. E. Kühn, J. Or-
istic proposals that require oxidant activation in the first
step to yield a catalyst–oxidant adduct as intermediate and
with equilibrium for this activation step (Figure 3).
ganomet. Chem. 2006, 691, 3718–3729; d) A. J. Burke, Coord.
Chem. Rev. 2008, 252, 170–175; e) S. Shylesh, M. J. Jia, W. R.
Thiel, Eur. J. Inorg. Chem. 2010, 4395–4410.
[2] a) J. G. Serafin, A. C. Liu, S. R. Seyedmonir, J. Mol. Catal. A
1
998, 131, 157–168; b) N. End, K. U. Schoning, Immobilized
Catalysts 2004, 242, 241–271; c) H. U. Blaser, B. Pugin, F. Spin-
dler, J. Mol. Catal. A 2005, 231, 1–20.
Experimental Section
General: All preparations and manipulations were carried out with
Schlenk techniques under an argon atmosphere. Solvents were
dried by standard procedures and distilled under argon prior to
use. H NMR spectra were recorded with a Bruker AM 250 op-
erating at 250 MHz. Chemical shifts are expressed in ppm down-
field from Me Si. Coupling constants are given in Hertz. The gas-
4
chromatographic analyses were completed with an Agilent 7890A
instrument equipped with an SPB-5 capillary column
[3] a) K. A. D. Swift, Top. Catal. 2004, 27, 143–155; b) C. B. Woiti-
ski, Y. N. Kozlov, D. Mandelli, G. V. Nizova, U. Schuchardt,
G. B. Shul’pin, J. Mol. Catal. A 2004, 222, 103–119; c) Y. Kon,
Y. Ono, T. Matsumoto, K. Sato, Synlett 2009, 1095–1098; d)
P. P. Jiang, M. Chen, Y. M. Dong, Y. Lu, X. Ye, W. J. Zhang,
J. Am. Oil Chem. Soc. 2010, 87, 83–91; e) C. A. Gamelas, P.
Neves, A. C. Gomes, A. A. Valente, C. C. Romão, I. S. Gonç-
alves, M. Pillinger, Catal. Lett. 2012, 142, 1218–1224.
1
[
4] a) B. Lane, K. Burgess, Chem. Rev. 2003, 103, 2457–2473; b)
L. J. Cao, M. Yang, G. Wang, Y. Wei, D. B. Sun, J. Polym. Sci.,
Part A 2010, 48, 558–562; c) S. Bawaked, N. F. Dummer, D.
Bethell, D. W. Knight, G. J. Hutchings, Green Chem. 2011, 13,
127–134; d) M. Herbert, F. Montilla, A. Galindo, J. Mol. Ca-
tal. A 2011, 338, 111–120; e) J. Pisk, D. Agustin, V. Vrdoljak,
R. Poli, Adv. Synth. Catal. 2011, 353, 2910–2914; f) E. H.
de Faria, G. P. Ricci, L. Marcal, E. J. Nassar, M. A. Vicente,
R. Trujillano, A. Gil, S. A. Korili, K. J. Ciuffi, P. S. Calefi, Ca-
tal. Today 2012, 187, 135–149; g) M. A. Katkar, S. N. Rao,
H. D. Juneja, Rsc Adv. 2012, 2, 8071–8078; h) D. J. Xuereb, J.
Dzierzak, R. Raja, Catal. Today 2012, 198, 19–34; i) J. Morlot,
N. Uyttebroeck, D. Agustin, R. Poli, ChemCatChem 2013, 5,
(
30 mϫ0.25 mmϫ0.25 μm) and with a flame-ionization detector.
Compounds [Cp* ] (M = Mo, W) were prepared as described
Cyclooctene (Fluka), dodecane (Aldrich), and
tBuOOH solutions in decane (5.5 m) or in water (70%, ca. 19 m)
Aldrich) were used as received.
2
15]
2 5
M O
[
in the literature.
(
Catalytic Epoxidation Experiments Monitored by 1H NMR Spec-
troscopy: The catalytic reactions were carried out in an NMR spec-
troscopy tube at room temperature. For each experiment, the NMR
spectroscopy tube was charged with cyclooctene (0.086 g,
0
(
7
.78 mmol) and the deuterated solvent [CDCl
CD CO] (0.3 mL). Subsequently, the catalyst ([Cp*
.8 μmol) was added, and finally the catalytic reaction was started
3
, MeOD, CH
3
CN,
3
)
2
2
M
2
O
5
],
6
01–611.
5] M. B. Trost, R. G. Bergman, Organometallics 1991, 10, 1172–
178.
[
[
by the addition of TBHP (283 μL of a 5.5 m solution in decane,
.56 mmol) and monitored by NMR spectroscopy. After 15 min,
1
1
6] a) M. Abrantes, A. Santos, J. Mink, F. Kühn, C. Romão, Orga-
nometallics 2003, 22, 2112–2118; b) A. Sakthivel, J. Zhao, M.
Hanzlik, F. E. Kuhn, Dalton Trans. 2004, 3338–3341; c) M.
Abrantes, S. Gago, A. A. Valente, M. Pillinger, I. S. Gonçalves,
T. M. Santos, J. Rocha, C. C. Romão, Eur. J. Inorg. Chem.
formation of a precipitate in considerable quantities was observed
in all solvents except in acetonitrile. All attempts to isolate the pre-
cipitate (which is soluble only in DMSO and acetonitrile) were un-
successful. The percentage yield of cyclooctene oxide was calcu-
lated from the NMR spectroscopic intensity.
2004, 4914–4920; d) J. Zhao, A. M. Santos, E. Herdtweck, F. E.
Kühn, J. Mol. Catal. A 2004, 222, 265–271; e) F. E. Kühn, J.
Zhao, M. Abrantes, W. Sun, C. A. M. Afonso, L. C. Branco,
I. S. Gonçalves, M. Pillinger, C. C. Romão, Tetrahedron Lett.
Catalytic Epoxidation Experiments Monitored by Gas Chromato-
graphy: The epoxidations were carried out under an argon atmo-
sphere in Schlenk tubes equipped with a magnetic stirrer and im-
mersed in a thermostatic bath set at the desired temperature. In a
typical experiment (which corresponded to run 4 in Table 2), the
tube was charged with cyclooctene (132.7 mg, 1.204 mmol), the in-
ternal standard dodecane (134.9 mg, 0.792 mmol), the
2
005, 46, 47–52; f) A. Sakthivel, J. Zhao, M. Hanzlik, A. S. T.
Chiang, W. A. Herrmann, F. E. Kuhn, Adv. Synth. Catal. 2005,
47, 473–483; g) A. M. Martins, C. C. Romão, M. Abrantes,
3
M. C. Azevedo, J. Cui, A. R. Dias, M. T. Duarte, M. A. Lemos,
T. Lourenço, R. Poli, Organometallics 2005, 24, 2582–2589; h)
J. Zhao, A. Sakthivel, A. M. Santos, F. E. Kuhn, Inorg. Chim.
Acta 2005, 358, 4201–4207; i) M. Abrantes, A. Sakthivel, C. C.
Romão, F. E. Kühn, J. Organomet. Chem. 2006, 691, 3137–
[
Cp*
solution in decane, 0.43 mL, 2.36 mmol) in acetonitrile/toluene
4 mL, 3:1). Aliquots of the reaction mixture (0.2 mL) were with-
2 2 5
Mo O ] catalyst (6.5 mg, 12.0 μmol), and TBHP (as a 5.5 m
(
3145; j) J. Zhao, E. Herdtweck, F. E. Kühn, J. Organomet.
drawn at predetermined times, quenched by the addition of manga-
nese dioxide, diluted with diethyl ether (2 mL), and filtered through
a Pasteur pipette filled with silica to eliminate the residual MnO .
2
The silica was washed by diethyl ether (3 mL). The resulting or-
ganic phase was analyzed by gas chromatography.
Chem. 2006, 691, 2199–2206; k) J. Zhao, K. R. Jain, E.
Herdtweck, F. E. Kühn, Dalton Trans. 2007, 5567–5571; l)
K. R. Jain, F. E. Kühn, Dalton Trans. 2008, 2221–2227; m)
A. M. Al-Ajlouni, D. Veljanovski, A. Capapé, J. Zhao, E.
Herdtweck, M. J. Calhorda, F. E. Kühn, Organometallics 2009,
2
8, 639–645; n) M. Abrantes, F. A. A. Paz, A. A. Valente,
Supporting Information (see footnote on the first page of this arti-
cle): Plots of all kinetics runs.
C. C. L. Pereira, S. Gago, A. E. Rodrigues, J. Klinowski, M.
Pillinger, I. S. Gonçalves, J. Organomet. Chem. 2009, 694,
1826–1833; o) A. Capapé, A. Raith, F. E. Kühn, Adv. Synth.
Catal. 2009, 351, 66–70; p) P. J. Costa, M. J. Calhorda, F. E.
Kühn, Organometallics 2010, 29, 303–311; q) D. Betz, A. Raith,
M. Cokoja, F. E. Kühn, ChemSusChem 2010, 3, 559–562; r) M.
Abrantes, P. Neves, M. M. Antunes, S. Gago, F. A. A. Paz,
A. E. Rodrigues, M. Pillinger, I. S. Gonçalves, C. M. Silva,
A. A. Valente, J. Mol. Catal. A 2010, 320, 19–26; s) P. Neves,
C. C. L. Pereira, F. A. A. Paz, S. Gago, M. Pillinger, C. M.
Silva, A. A. Valente, C. C. Romão, I. S. Gonçalves, J. Or-
ganomet. Chem. 2010, 695, 2311–2319; t) M. Abrantes, S. M.
Bruno, C. Tome, M. Pillinger, I. S. Gonçalves, A. A. Valente,
Catal. Commun. 2011, 15, 64–67; u) N. Grover, F. E. Kuhn,
Acknowledgments
The authors are grateful to the Centre National de la Recherche
Scientifique (CNRS), to the Institut Universitaire de France (IUF),
and to Celal Bayar University (project FBE 2010-081) for financial
support of this work.
[
1] a) T. Katsuki, Adv. Synth. Catal. 2002, 344, 131–147; b) F. E.
Kühn, A. M. Santos, W. A. Herrmann, Dalton Trans. 2005,
2734
www.eurjic.org
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2013, 2728–2735

2 2 5
Epoxidation of Cyclooctene Catalyzed by [Cp* M O ]
Curr. Org. Chem. 2012, 16, 16–32; v) S. A. Hauser, M. Cokoja,
M. Drees, F. E. Kuhn, J. Mol. Catal. A 2012, 363, 237–244.
7] M. Ciclosi, C. Dinoi, L. Gonsalvi, M. Peruzzini, E. Manoury,
R. Poli, Organometallics 2008, 27, 2281–2286.
[11] K. B. Sharpless, J. M. Townsend, D. R. Williams, J. Am. Chem.
Soc. 1972, 94, 295–296.
[
[
[
[
[12] A. O. Chong, K. B. Sharpless, J. Org. Chem. 1977, 42, 1587–
1590.
8] C. Dinoi, M. Ciclosi, E. Manoury, L. Maron, L. Perrin, R.
Poli, Chem. Eur. J. 2010, 16, 9572–9584.
[13] G. Lewandowski, React. Kinet. Catal. Lett. 2002, 77, 325–334.
[14] C. Dinoi, R. Poli, L. Perrin, L. Maron, Dalton Trans. 2012, 41,
1131–1133.
[15] C. Dinoi, G. Taban, P. Sözen, F. Demirhan, J.-C. Daran, R.
Poli, J. Organomet. Chem. 2007, 692, 3743–3749.
Received: February 22, 2013
9] A. Comas-Vives, A. Lledós, R. Poli, Chem. Eur. J. 2010, 16,
2147–2158.
10] a) W. R. Thiel, T. Priermeier, Angew. Chem. 1995, 107, 1870;
Angew. Chem. Int. Ed. Engl. 1995, 34, 1737–1738; b) W. R.
Thiel, J. Mol. Catal. A 1997, 117, 449–454.
Published Online: April 11, 2013
Eur. J. Inorg. Chem. 2013, 2728–2735
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2735