(Salen)nickel-catalysed epoxidations in the homogeneous and heterogeneous phase: The implications of oxygen on the efficiency and product selectivity

Article abstract of DOI:10.1002/ejic.200500314

The catalytic activity of several nickel(II) complexes with salen-type ligands and their zeolite-immobilised analogues is assessed in the epoxidation of irons-β-methylstyrene by NaOCl, in the absence and in the presence of oxygen. The complexes were immob

Full text of DOI:10.1002/ejic.200500314

FULL PAPER
(Salen)nickel-Catalysed Epoxidations in the Homogeneous and Heterogeneous
Phase: The Implications of Oxygen on the Efficiency and Product Selectivity
Rita Ferreira,[a][‡] Hermenegildo García,[b] Baltazar de Castro,[a] and Cristina Freire*[a]
Keywords: Alkenes / Epoxidation / Heterogeneous catalysis / N,O ligands / Nickel / Zeolites
The catalytic activity of several nickel(II) complexes with
salen-type ligands and their zeolite-immobilised analogues
is assessed in the epoxidation of trans-β-methylstyrene by
NaOCl, in the absence and in the presence of oxygen. The
complexes were immobilised in zeolites X and Y using the
“ship-in-a-bottle” procedure, and the resulting materials
were characterised by chemical analysis, XPS and IR, and
UV/Vis diffuse reflectance spectroscopy. The loading and the
distribution of the complexes in the materials were found to
be dependent on the ligand. All (salen)nickel complexes
proved to be active in the homogeneous and heterogeneous
epoxidation of trans-β-methylstyrene by NaOCl. The im-
mobilised complexes gave lower conversions and their cata-
lytic activity is dependent on the loading and distribution of
the complexes within the zeolites. In both homogeneous and
heterogeneous systems, the alkene conversion is influenced
by oxygen. Removing oxygen from the reaction media results
in lower alkene conversion but higher epoxide selectivity.
Based on these results, two competitive mechanisms are pro-
posed for the epoxidation reaction, one implying the forma-
tion of free radicals and other involving the binding of the
substrate to a (oxo)metal intermediate.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2005)
used.[11,12] Transition metal complexes can be immobilised
in several supports such as zeolites,[13] activated carbon[14]
and pillared clays.[15] Recently, we demonstrated that (salen)-
manganese complexes immobilised in carbon and pillared
clays can be used as heterogeneous catalysts in epoxidation
reactions without significant loss in their catalytic activity
upon re-use.[16–18]
The immobilisation and employment of (salen)nickel
complexes in heterogeneous catalysis has not been exten-
sively explored. Some examples include the use of (salen)-
Introduction
The use of transition metal complexes as catalysts for the
epoxidation of alkenes has been a subject of interest in the
past few decades due to the increasing use of epoxides in
agro- and pharmacological chemistry.[1] A large number of
publications concern the use of the (salen)Mn complexes
developed by Katsuki and Jacobsen as highly enantioselec-
tive catalysts in the epoxidation of alkenes.[2–4] In these
studies, oxidants such as tert-butyl hydroperoxide and iodo-
sylbenzene are often employed, although lately the use of
environmentally more friendly oxidants such as H2O2 and
NaOCl has been pursued.
Ni-type complexes in faujasites for the oxidation of phenol
[19]
with H2O2
and for the epoxidation of cyclohexene and
1-hexene by NaOCl.[20,21] A more recent study concerned
the use of [Ni(salen)] immobilised in several solid supports
as a catalyst for hydrogenation reactions.[22]
Although (salen)nickel() complexes have not been as
widely investigated as their manganese analogues, earlier
studies by Burrows and co-workers, and later by Kureshy,
This paper exploits the use of (salen)NiII complexes as
proved that nickel complexes with cyclam and salen ligands homogeneous catalysts and, after immobilisation in zeolites
are also active catalysts in the epoxidation of alkenes.[5–10]
X and Y, as heterogeneous catalysts in the epoxidation of
The immobilisation of transition metal complexes in so- trans-β-methylstyrene by NaOCl. Zeolites X and Y have a
lid supports has intrinsic practical advantages, namely the crystalline faujasite-type structure composed of small pores
heterogenised catalysts are easy to handle and can be re- (diameter 7 Å), which access internal cavities with a dia-
meter of 13 Å. This topology originates two special proper-
ties: (i) zeolites can act as molecular sieves, and (ii) they can
[a] REQUIMTE/Departamento de Química,
be hosts by accommodating and restricting molecules inside
the cavities. When considering the potential use of zeolites
as supports for catalysts, the above characteristics can result
in very important features such as reactant/product size-
and-shape selectivity and the possibility to host metal com-
plexes that have proved to be good catalysts in solution.
Nevertheless, it is noteworthy that the catalyst performance
Rua do Campo Alegre, 4169-007 Porto, Portugal
Fax: +351-22-608-2959
E-mail: acfreire@fc.up.pt
[b] Instituto de Tecnología Química, UPV-CSIC,
Av. de los Naranjos s/n, 46022 Valencia, Spain
[‡] Present address: ESTG-Instituto Politécnico de Viana do
Castelo,
Av. do Atlântico, 4900-348 Viana do Castelo, Portugal
E-mail: rferreira@estg.ipvc.pt
4272
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/ejic.200500314
Eur. J. Inorg. Chem. 2005, 4272–4279

(Salen)nickel-Catalysed Epoxidations
FULL PAPER
can be affected by the immobilisation in a solid material Table 1. Nickel content (% weight) obtained by chemical analysis
(CA) and Ni/Si ratios obtained by chemical analysis (CA) and XPS
with its own physical and chemical properties.
for the immobilised (salen)nickel()complexes.
Four different (salen)NiII complexes (Scheme 1) were en-
Catalyst
wt.-% Ni
(CA)
Ni/Si
capsulated in both zeolites X and Y using the “ship-in-a-
bottle” procedure[23] accomplished by the “flexible ligand”
CA
XPS
approach. This two-step synthesis consists of (i) the intro- [Ni(salen)]@X
1.0
1.9
1.0
1.8
1.0
1.3
0.4
1.5
0.040
0.053
0.033
0.050
0.019
0.055
0.011
0.031
0.176
0.142
0.196
0.161
0.101
0.074
0.117
0.074
duction of NiII cations by ion exchange and (ii) solvent-free
[Ni(salhd)]@X
[Ni(α,αЈ-Me2salen)]@X
diffusion of the ligands, followed by complexation with the
nickel ions. The four (salen)nickel() complexes meet the
requirements needed for this method of immobilisation be-
cause their ligands have a suitable size to diffuse through
the pores and reach the cavities containing the metal ions.
[Ni(3-MeOsalen)]@X
[Ni(salen)]@Y
[Ni(salhd)]@Y
[Ni(α,αЈ-Me2salen)]@ Y
[Ni(3-MeOsalen)]@Y
The results in Table 1 show that the amount of complex
and its distribution within the solids differs with the size of
the complex and with the zeolite type. Loadings between
0.4 and 1.8 wt.-% Ni were obtained, but generally zeolite X
presents a slightly higher complex content than zeolite Y.
This is explained by the higher number of exchangeable cat-
ions present in the parent NaX (Na/Si = 0.7) when com-
pared with NaY (Na/Si = 0.4), which results in a higher
number of NiII cations in the cavities and, consequently, a
higher complex loading.
In both zeolites X and Y, the complexes [Ni(3-MeO-
salen)] and [Ni(salhd)] have a higher content inside the zeo-
lites. Also, the distribution of the complexes depends on the
ligand type. All the complexes show a higher Ni/Si ratio by
XPS than by chemical analysis, pointing to a preferential
location of the complexes in the outer layers. This difference
is more significant for [Ni(salen)] and [Ni(α,αЈ-Me2salen)],
suggesting that these complexes occupy more the surface
layers while [Ni(3-MeOsalen)] and [Ni(salhd)] are more uni-
formly distributed.
Scheme 1. Structure of (salen)nickel() complexes.
The catalytic activity of the materials was then tested in
the epoxidation of trans-β-methylstyrene by NaOCl and
compared with the results obtained for the complexes in the
homogeneous phase. In order to understand the effect of
molecular oxygen, a possible co-oxidant, all the reactions
were performed in the presence and in the absence of oxy-
gen. These experiments have enabled us to obtain insights
into the mechanism of the alkene epoxidation by nickel
complexes.
The preferential location in the outer layers could be due
to diffusion of the complexes out of the inner cavities, prob-
ably during the extraction step.
The IR spectra of the solids were recorded and compared
with the IR spectra of the corresponding free complexes.
Figures 1a and 1b show the IR spectra of [Ni(salen)]@X
and [Ni(salen)]@Y, respectively, in the region 2000–
1000 cm–1. For comparison, the spectra of the parent mate-
rials, NiX, NiY, and of [Ni(salen)] are also included. All the
materials present intense bands in this region due to zeolite
lattice vibrations (1200–1000 cm–1) and a large band
centred at 1620 cm–1 attributed to adsorbed water.[24]
The characteristic bands of the complex are observed in
the region between 1600 and 1200 cm–1, where the zeolites
do not absorb, and represent the ν(C=N), ν(C=C), ν(C–C)
and ν(C–O) vibrations. The frequencies of these vibrations
are practically coincident with those of the complexes, thus
Results and Discussion
Materials Characterisation
The ship-in-a-bottle synthesis resulted in an efficient providing evidence for their presence inside the zeolites
loading of the nickel complexes inside the zeolites. The without significant interaction with the matrixes.
loading of the complexes and their distribution within the
The encapsulation of the complexes was also confirmed
material was deduced from the nickel content determined by UV/Vis diffuse reflectance. Figures 2a and b display typ-
by bulk chemical analysis and from XPS. Table 1 presents ical UV/Vis diffuse reflectance spectra of [Ni(salen)]@X
the nickel content of the materials (wt.-% Ni) obtained by and [Ni(salen)]@Y, respectively. The spectra of the corre-
chemical bulk analysis and the Ni/Si ratio obtained by sponding original materials, NiX and NiY, and of [Ni-
chemical analysis and by XPS.
(salen)] are included for comparison.
Eur. J. Inorg. Chem. 2005, 4272–4279
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4273

R. Ferreira, H. García, B. de Castro, C. Freire
FULL PAPER
strate/catalyst ratio = 0.025:1, in dichloromethane) with
NaOCl as oxidant. The experiments were run in the pres-
ence and in the absence of oxygen. Figure 3 shows the al-
kene conversion as a function of the reaction time for the
epoxidation of trans-β-methylstyrene catalysed by [Ni-
(salen)], [Ni(salhd)], [Ni(α,αЈ-Me2salen)] and [Ni(3-MeO-
salen)] in the presence of oxygen. It is clear from this figure
that all complexes are active catalysts in the reaction. The
conversion follows a sigmoid curve with an induction
period of about 1 h, and reaches maximum conversion after
10 h. The presence of an induction period suggests that the
oxidant efficiency is limited.[25]
Figure 1. a) IR spectra of [Ni(salen)]@X (solid line), [Ni(salen)]
free complex (dotted line) and the parent NiX (dashed line), and
b) IR spectra of [Ni(salen)]@Y (solid line), [Ni(salen)] free complex
(dotted line) and the parent NiY (dashed line).
Figure 3. Conversion vs. time plot of the epoxidation of trans-β-
methylstyrene with NaOCl catalysed by [Ni(salen)] (∆), [Ni(salhd)]
(᭺), [Ni(α,αЈ-Me2salen)] (᭡) and [Ni(3-MeOsalen)] (×) in the pres-
ence of oxygen.
Figure 4 shows the alkene conversion as a function of
the reaction time obtained in the epoxidation of trans-β-
methylstyrene catalysed by [Ni(salen)], [Ni(salhd)], [Ni(α,αЈ-
Me2salen)] and [Ni(3-MeOsalen)], in the absence of oxygen.
All the complexes are also active in the absence of oxygen
and the evolution with time pursues a similar pattern as
observed in the presence of oxygen: a sigmoidal curve with
an induction period, although the reaction is slower and
the time needed to reach the maximum conversion is about
20 h. The induction period is higher in this case, and de-
pendent on the complex. Table 2 summarises the results ob-
tained after 24 h for the reactions catalysed by the nickel()
complexes in the homogeneous phase under aerobic or an-
aerobic conditions.
Figure 2. a) UV/Vis solid-state diffuse reflectance spectra of [Ni-
(salen)]@X (solid line), [Ni(salen)] (dotted line) and NiX (dashed
line), and b) UV/Vis solid-state diffuse reflectance spectra of [Ni-
(salen)]@Y (solid line), [Ni(salen)] (dotted line) and NiY (dashed
line).
All the materials exhibit bands in the range λ = 300–
700 nm. The bands with λmax Ͻ 450 nm were attributed to
charge-transfer bands. Characteristic d-d bands appear at
510–560 nm in the spectra of the free complexes as well as
in the spectra of the zeolite-immobilised complexes. How-
ever, the materials show slight shifts of the transition bands
in relation to the free complexes, thus suggesting some de-
gree of distortion around the metal centre, probably in-
duced by the immobilisation.
Figure 4. Conversion vs. time plot of the epoxidation of trans-β-
methylstyrene with NaOCl catalysed by [Ni(salen)] (∆), [Ni(salhd)]
(᭺), [Ni(α,αЈ-Me2salen)] (᭡) and [Ni(3-MeOsalen)] (×) in the ab-
sence of oxygen.
Homogeneous Catalysis
The catalytic activity of the four (salen)nickel() com-
Alkene conversions of between 61 and 100% were ob-
plexes was first studied in the homogeneous phase (sub- tained in the presence of oxygen and between 47 and 65%
4274
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org Eur. J. Inorg. Chem. 2005, 4272–4279

(Salen)nickel-Catalysed Epoxidations
FULL PAPER
Table 2. Epoxidation reactions with the free complexes in the presence and absence of oxygen.[a]
[b]
Catalyst
O2
Conv. [%][c]
Selectivity [%] (Yield [%])
Epoxide[d]
Benzaldehyde
Others
[Ni(salen)]
[Ni(salhd)]
[Ni(α,αЈ-Me2salen)]
[Ni(3-MeOsalen)]
[Ni(salen)]
[Ni(salhd)]
[Ni(α,αЈ-Me2salen)]
[Ni(3-MeOsalen)]
present
present
present
present
absent
absent
absent
absent
100
99
84
61
62
65
48
47
54 (54)
54 (54)
58 (49)
52 (32)
60 (37)
65 (42)
69 (33)
47 (22)
9 (1)
11 (11)
11 (9)
12 (7)
8 (5)
15 (10)
9 (4)
37 (4)
35 (3)
31 (26)
36 (22)
32 (20)
20 (13)
22 (11)
44 (21)
9 (4)
[a] Solvent: CH2Cl2; alkene/catalyst ratio = 1.0:0.025; results after 24 h. [b] Oxygen is present as in normal atmosphere (1 atm). For the
conditions in the absence of oxygen, see Exp. Sect. [c] Conversion (as disappearance of alkene) determined by GC using n-decane as
internal standard. [d] Obtained as 100% trans-epoxide.
in its absence; epoxide was the major product but benzalde- dation reaction is not completely understood, it is known
hyde and other products were also obtained.[26]
that the reaction selectivity depends on how the oxygen
In the presence of oxygen, the reactions catalysed by [Ni- atom is transferred to the alkene. A rebound mechanism
(salen)] and [Ni(salhd)] show the highest conversion, reach- involving the formation of a (oxo)metal intermediate and
ing almost 100%. The introduction of two methyl groups alkyl radicals has been proposed for Cytochrome P450, a
at the α-imine carbon atom in [Ni(α,αЈ-Me2salen)] decreases (porphyrin)iron compound.[27] Recently, it has been proved
the conversion but does not affect the epoxide selectivity. that these short-lived radicals are mainly free in solution
The effect of an electron-donor group in the third position rather than being bound to the (oxo)metal species.[28]
of the aldehyde fragment is more significant: the methoxy
group in [Ni(3-MeOsalen)] decreases the conversion signifi-
cantly (61%), although the epoxide selectivity is not affec-
ted. No significant changes were observed when the ali-
phatic imine bridge was substituted by a cyclohexyl bridge,
as in [Ni(salhd)]. In general, the selectivity of the products
does not seem to be dependent on the ligand: [Ni(α,αЈ-Me2-
salen)] shows the highest value for the epoxide selectivity
(58%) whereas [Ni(3-MeOsalen)] has the lowest value
(52%).
For the reactions in the absence of oxygen, the alkene
conversion is lower when compared with reactions under
aerobic conditions (Table 2), but the complexes [Ni(salen)]
and [Ni(salhd)] still present the highest values (62 and 65%,
respectively) followed by [Ni(α,αЈ-Me2salen)] (48%) and fi-
nally [Ni(3-MeOsalen)] (47%). On the other hand, the ep-
oxide selectivity increases for all complexes except for the
[Ni(3-MeOsalen)]-catalysed reaction. As observed for the
reactions in the presence of oxygen, the highest epoxide se-
lectivity was obtained for [Ni(α,αЈ-Me2salen)], whereas
[Ni(3-MeOsalen)] shows the lowest value. The dependency
of the epoxide selectivity on the ligand is noteworthy, and
is more significant when oxygen is absent.
The involvement of an (oxo)metal species of high oxi-
dation state (Mn+=O; n = 3, 4 or 5) has also been proved
for (salen)manganese- and -chromium-catalysed epoxida-
tions.[29] In the case of the (salen)nickel-catalysed epoxida-
tions, the formation of an (oxo)metal intermediate is consis-
tent with the dependency of the alkene conversion with the
ligand type, although, as suggested by Burrows and co-
workers, another intermediate with a radical character can
also be formed, namely NiIII–O·. This species can instigate
the formation of free radicals that compete in the oxidation
of the substrate and, therefore, result in reactions with low
epoxide selectivity (Scheme 2, mechanism A).[6,9] Alterna-
tively, the alkene can bind to the (oxo)metal intermediate
and different paths can then follow (Scheme 2, mechanism
B).[30] In the case of epoxidations catalysed by (cyclam)
nickel complexes, two main pathways have been suggested
for the subsequent steps: (1) a concerted mechanism in
which the substrate–(oxo)metal intermediate suffers a het-
erolytic cleavage leading to the desired epoxide, and (2) a
stepwise pathway implying the formation of a carbocation
or a carbon radical, which forms the product by homolytic
cleavage. The heterolytic cleavage is known to afford high
yields of epoxide, while the other preferably performs allylic
oxidations;[31] both are summarised in Scheme 2.
Overall, the alkene conversion is higher and the reaction
faster in the presence of oxygen, although the epoxide selec-
tivity is higher in the absence of oxygen.
On the basis of the catalytic results obtained in the ab-
sence and presence of oxygen, the formation of free radicals
(mechanism A) cannot be excluded. Free radicals can cap-
ture oxygen and consequently enhance the reaction rate and
alkene conversion;[32,33] therefore, a free-radical mechanism
Reaction Mechanism
Salen-type complexes have been used in epoxidation re- would explain the higher rates and alkene conversions ob-
actions owing to their special coordination environment, served for the reactions run under oxygen. On the other
which, like porphyrins, provides a square-planar geometry hand, the dissimilarity in product selectivity obtained with
around the metal atom that allows reactants and/or oxidant different complexes in the absence of oxygen points to the
to access the active centre. Although the oxidation mecha- involvement of
a
substrate–(oxo)metal intermediate
nism through which the metal complexes act in an oxi- (mechanism B), which, through a concerted or a stepwise
Eur. J. Inorg. Chem. 2005, 4272–4279
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4275

R. Ferreira, H. García, B. de Castro, C. Freire
FULL PAPER
Table 3 summarises the results obtained in the hetero-
geneous phase using the nickel complexes immobilised in
zeolites X and Y, in the presence of oxygen.
Table 3. Results obtained for the heterogeneous catalysis in the
presence of oxygen.[a]
Catalyst
Conv. [%][b]
Selectivity [%] (Yield [%])
Epoxide[c] Benzaldehyde Others
[Ni(salen)]@X
[Ni(salhd)]@X
[Ni(α,αЈ-Me2salen)]@X
[Ni(3-MeOsalen)]@X
[Ni(salen)]@Y
[Ni(salhd)]@Y
[Ni(α,αЈ-Me2salen)]@Y
[Ni(3-MeOsalen)]@Y
71
27
88
19
30
44
62
19
40 (28)
28 (8)
53 (47)
22 (4)
34 (10)
48 (21)
45 (28)
25 (5)
23 (16)
27 (7)
15 (13)
36 (7)
37 (26)
45 (12)
32 (28)
42 (8)
50 (15)
28 (12)
33 (20)
15 (3)
16 (5)
23 (10)
22 (14)
60 (11)
[a] Solvent: CH2Cl2; alkene/catalyst ratio = 1.0:0.025; results after
48 h. [b] Conversion (as disappearance of alkene) determined by
GC using n-decane as internal standard. [c] Obtained as 100%
trans-epoxide.
Scheme 2. Proposed mechanism for the trans-β-methylstyrene
epoxidation by NaOCl with (salen)nickel() complexes.
The reaction time increased significantly and thus the re-
sults in Table 3 refer to values obtained after 48 h. The re-
sults show that all the immobilised complexes are active in
the epoxidation of the alkene, but the catalytic activity is
lower than in the homogeneous reactions. The reactions ca-
talysed by zeolites with [Ni(α,αЈ-Me2salen)], which reached
conversions close to those obtained in the homogeneous
phase, are excluded from this conclusion. In fact, the maxi-
mum conversion (88%) among the heterogeneous catalysts
was observed for [Ni(α,αЈ-Me2salen)]@X, together with the
highest epoxide selectivity.
Lower conversions and longer reactions times are usually
observed on going from solution to immobilised catalysts
due to the diffusion constraints imposed by the pore net-
work.[18] Although the solid support can constitute a pro-
tective environment around the metal atom, the physical re-
strictions imposed by the zeolite can make the access of
reactants to the metal centre more difficult and, as a result,
the activity decreases.
In this case, a correlation between the catalytic activity
of the complexes and their distribution within the solids
confirms the effect of diffusion constraints: the complexes
localised in the outer layers of the zeolites (see above),
which therefore can be more easily accessed by the reac-
tants, give higher catalytic activities.
Surprisingly, the epoxide selectivity was lower in all the
heterogeneous reactions than in the homogeneous phase,
resulting in higher benzaldehyde selectivity. The chemical
microenvironment imposed by the zeolites could be the
mechanism, allows a good proximity of the substrate to the
ligand. Also, as the carbon radical formed in the stepwise
pathway in mechanism B is expected to be very sensitive to
the presence of O2, resulting in highly oxidised products
(such as benzaldehyde), this mechanism could explain the
lower epoxide selectivity observed in the reactions run in
the presence of oxygen.
The substantially lower epoxide selectivity obtained with
the complex that bears electron-donor substituents, [Ni(3-
MeOsalen)], in the absence of oxygen is noteworthy. Studies
on the epoxidation of alkenes by (salen)chromium com-
plexes have shown that complexes bearing methoxy groups
in the ortho-position to the hydroxy group of the salicyl-
aldehyde moiety are easily destroyed by the oxidant.[34]
In this context, our results support the occurrence of two
competitive mechanisms (Scheme 2): a mechanism where
the (oxo)metal species originates free radicals that, in the
presence of oxygen, cause higher and faster alkene conver-
sions but indistinguishable epoxide selectivity (mechanism
A), and a mechanism involving binding of the substrate to
the (oxo)metal intermediate that, through a concerted or a
stepwise pathway, results in lower conversions, but higher
epoxide selectivity (mechanism B).
Mechanism A is favoured in the presence of oxygen,
while mechanism B is favoured in its absence.
Heterogeneous Catalysis
To study the catalytic activity of the immobilised nick- cause of this result. Epoxide ring opening in acidic media
el() complexes, the epoxidation of trans-β-methylstyrene forms benzaldehyde, thus the acid character of the zeolites
by NaOCl was carried out using the prepared materials as contributes to the lowering of the epoxide selectivity. Again,
heterogeneous catalysts. The reactions were run under the the zeolites that bear complexes closer to the surface, such
same conditions as described for the homogeneous phase, as [Ni(α,αЈ-Me2salen)]@X and [Ni(α,αЈ-Me2salen)]@Y, are
using an amount of zeolite corresponding to the same sub- less affected by the zeolite characteristics and present nearly
strate/catalyst ratio (1:0.025) in dichloromethane (see Exp. the same epoxide selectivity as in the homogeneous phase.
Sect.). As in the homogeneous phase, the reactions were
The effect of the zeolites’ acidity in the reaction selectiv-
performed in the presence and in the absence of oxygen. ity can be noted by comparing the epoxide selectivity ob-
4276
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2005, 4272–4279

(Salen)nickel-Catalysed Epoxidations
FULL PAPER
tained for a complex immobilised in zeolite X with the same Re-Use of Heterogeneous Catalysts
complex immobilised in zeolite Y (Table 3). Apart from the
The possibility to re-use [Ni(3-MeOsalen)]@X and [Ni(3-
MeOsalen)]@Y was evaluated in the subsequent way: after
the first run, the catalysts were filtered from the solution,
washed with copious amounts of dichloromethane, dried
and re-used according to the general procedure described
for the heterogeneous reactions. The results of the re-utilis-
ation and the relation with the first use are illustrated in
Figure 5.
For [Ni(3-MeOsalen)]@X, the conversion is similar to
that obtained in the first use (19%), but the epoxide selec-
tivity shows a significant increase from 22% to 53%. In the
case of [Ni(3-MeOsalen)]@Y, an increase in both conver-
sion (from 19 to 30%) and epoxide selectivity (from 25 to
46%) was observed.
other factors already discussed in this paper, the complexes
in zeolite X have a higher epoxide selectivity than the com-
plexes in Y, a zeolite with more acidic sites.
The activity of the immobilised catalysts was also evalu-
ated in the absence of oxygen and the results are collected
in Table 4. As observed for the homogeneous phase, the ab-
sence of oxygen results in lower conversions and higher ep-
oxide selectivity, except for [Ni(3-MeOsalen)], which regis-
ters an increase in the conversion when compared with the
reaction run under oxygen. Once more, as observed for the
homogeneous system, this is the complex with the lowest
epoxide selectivity, due to its fast decomposition in the reac-
tion media.
The immobilised catalysts show similar conversions in
their re-use compared to their first utilisation: the conver-
sion is equal or even higher than in the first use and there
is an increase in the epoxide selectivity. These positive re-
sults can be attributed to a change in the overall chemical
environment around the complex. In the first use the mate-
rials are in contact with NaOCl and a chemical modifica-
tion from acidic to more basic environment might have oc-
curred. This effect is more significantly reflected in the reac-
tions catalysed by the zeolite with higher initial acidity (zeo-
lite Y). Also, a possible migration of the complex to the
outer surface layers of the zeolites cannot be excluded; this
decreases the diffusion constraints and therefore favours an
increase in conversion.
Table 4. Results obtained for the heterogeneous catalysis in the ab-
sence of oxygen.[a]
Catalyst
Conv. [%][b]
Selectivity [%] (Yield [%])
Epoxide[c] Benzaldehyde Others
[Ni(salen)]@X
[Ni(salhd)]@X
[Ni(α,αЈ-Me2salen)]@X
[Ni(3-MeOsalen)]@X
[Ni(salen)]@Y
[Ni(salhd)]@Y
[Ni(α,αЈ-Me2salen)]@Y
[Ni(3-MeOsalen)]@Y
62
13
62
39
48
43
64
32
60 (37)
41 (5)
13 (8)
7 (1)
28 (17)
52 (7)
59 (37)
46 (18)
60 (29)
59 (25)
64 (41)
38 (12)
12 (7)
14 (5)
18 (9)
14 (6)
16(10)
31 (10)
29 (18)
40 (16)
22 (11)
27 (12)
20 (13)
31 (10)
[a] Solvent: CH2Cl2; alkene/catalyst ratio = 1.0:0.025; results after
48 h. [b] Conversion (as disappearance of alkene) determined by
GC using n-decane as internal standard. [c] Obtained as 100%
trans-epoxide.
However, as [Ni(3-MeOsalen)] suffers decomposition in
both homogeneous and heterogeneous media, additional
experiments should be performed to reach safer conclusions
about the catalysts’ re-utilisation.
The catalytic activity in the heterogeneous phase follows
the same pattern as the homogeneous-phase reactions,
hence we propose that the conclusions concerning the reac-
tion mechanism drawn for the homogeneous system can be
extended to the heterogeneous counterpart. In this context,
the two mechanisms presented above are also competing
in the heterogeneous reaction: a mechanism involving the
Conclusions
All the (salen)nickel() complexes are active catalysts in
formation of free radicals and a mechanism occurring the epoxidation of trans-β-methylstyrene by NaOCl in
through a substrate–(oxo)metal intermediate. homogeneous media, both in the presence and in the ab-
Figure 5. Conversion and selectivity obtained in the first use and in the re-use of [Ni(3-MeOsalen)]@X and of [Ni(3-MeOsalen)]@Y in
the epoxidation of trans-β-methylstyrene by NaOCl.
Eur. J. Inorg. Chem. 2005, 4272–4279
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4277

R. Ferreira, H. García, B. de Castro, C. Freire
FULL PAPER
AES. The samples were previously treated by fusion with LiBO2
and the mixture dissolved in nitric acid. Surface chemical analysis
of the materials was performed by X-ray photoelectron spec-
troscopy at the Centro de Materiais da Universidade do Porto (Por-
tugal) in a VG Scientific ESCALAB 200A spectrometer, using Mg-
Kα radiation (1253.6 eV). Electronic spectra were recorded with a
Shimadzu UV/3101PC spectrophotometer equipped with a BaSO4
sphere and using the same salt as reference. FTIR spectra were
recorded from KBr pellets of the materials or complexes with a
JASCO FT/IR-460 Plus spectrometer. GC analysis was carried out
sence of oxygen. Conversions of up to 100% and epoxide
selectivity of around 54% are obtained in the presence of
oxygen. However, in the absence of oxygen, the reaction
is slower, with lower conversions but with higher epoxide
selectivity.
These results lead to a proposal of two simultaneous
mechanisms for this reaction: a mechanism implying the
formation of free radicals, which is favoured in the presence
of oxygen, in competition with a mechanism where the sub-
strate binds to an (oxo)metal intermediate. The first would with a Varian 3400 CX equipped with a fused silica Chrompack
capillary column (30 m×0.53 mm) and a TCD detector. The col-
umn temperature was programmed from 60 °C (3 min) to 170 °C,
at a rate of 5 °Cmin–1, and the injection temperature was 200 °C.
lead to higher alkene conversions but lower and indistin-
guishable epoxide selectivity, while the other would result
in higher epoxide selectivity dependent on the ligand, but
lower conversions.
The immobilisation of the four nickel complexes in zeo-
lites X and Y results in materials with different loading and
distribution of the complexes. Complexes [Ni(salhd)] and
[Ni(3-MeOsalen)] have higher loadings than [Ni(α,αЈ-Me2-
salen)] and [Ni(salen)]. The former are also more uniformly
distributed than the latter, which are preferentially localised
in the surface layers of the zeolites.
General Procedure for the Synthesis of Nickel Complexes: The li-
gands H2salen, H2(salhd), H2(α,αЈ-Me2salen) and H2(3-MeOsalen)
and the respective nickel() complexes N,NЈ-bis(salicylidene)-1,2-
diamine, [Ni(salen)] (1), N,NЈ-bis(salicylidene)-1,2-cyclohexene-
diamine, [Ni(salhd)] (2), N,NЈ-bis(α-methylsalicylidene)-1,2-di-
amine [Ni(α,αЈ-Me2salen)] (3) and N,NЈ-bis(3-methoxysalicyl-
idene)-1,2-diamine, [Ni(3-MeOsalen)] (4) (Scheme 1) were prepared
and fully characterised in refs.[35,36] according to standard pro-
cedures.[37]
All the materials are active catalysts in the epoxidation
of trans-β-methylstyrene by NaOCl, both in the presence of
oxygen and in its absence, but the conversions are lower
than in the homogeneous phase. The zeolites bearing the
metal complexes closer to the surface give higher conver-
sions than those having higher loadings, but a more uni-
form distribution of the complexes. These results can be
explained by the physical restrictions imposed by the zeolite
pore network, which makes the access of the reactants to
the metal centre more difficult, resulting in lower activities.
Furthermore, the zeolite chemical environment, namely its
acidic character, influences the selectivity of the reaction.
Nevertheless, the zeolites can constitute an important pro-
tective environment around the ligand and an opportunity
to improve the catalyst life, i.e., by catalyst re-utilisation.
General Procedure for the Preparation of the Immobilised Catalysts:
The preparation of the immobilised (salen)nickel complexes fol-
lowed the “ship-in-a-bottle” procedure[23] using the “flexible li-
gand” method. In a typical preparation, a suspension of calcinated
NaY or NaX (10 g) in an aqueous solution containing Ni(NO3)2
(6 mmol) was heated at 70–80 °C for 24 h. The solid was then fil-
tered and dried at 120 °C under vacuum, and the ion-exchanged
zeolite (2 g; NiY or NiX) was mixed with the corresponding ligand
(metal/ligand molar ratio = 1:2) and heated to 130–150 °C for 2 h.
A change in the colour of the material from pale green (characteris-
tic of hexahydrated NiII) to orange-brown [typical of (salen)NiII
complexes] confirmed complex formation. The solids were then
purified by solid-liquid Soxhlet extraction with appropriate sol-
vents (dichloromethane or ethanol) to remove unreacted ligands
and surface-bound complexes. The materials were denoted as [Ni-
(salen)]@X, [Ni(salhd)]@X, [Ni(α,αЈ-Me2salen)]@X [Ni(3-MeO-
salen)]@X, [Ni(salen)]@Y, [Ni(salhd)]@Y, [Ni(α,αЈ-Me2salen)]@Y
and [Ni(3-MeOsalen)]@Y.
Experimental Section
General Procedure for the Epoxidation Reactions: In a typical ex-
periment, a solution of the oxidant (5 mL) buffered to pH = 11
(with a solution of Na2HPO4) was stirred at room temperature in
a solution of dichloromethane (5 mL) enriched with trans-β-meth-
ylstyrene (2 mmol), the nickel catalyst (0.05 mmol) (substrate/cata-
lyst = 1:0.025) and benzyltributylammonium bromide (0.06 mmol),
used as phase-transfer agent. A similar procedure was employed
for the heterogeneous reactions using an amount of material corre-
sponding to a substrate/nickel complex ratio of 1:0.025. In this case
the catalysts were first activated at 120 °C for 2 h just before use
and no phase-transfer agent was employed. The reactions were run
at room temperature with continuous magnetic stirring and were
followed by GC using n-decane as internal standard. For the reac-
tions carried out in the absence of oxygen, the solvent was first
deoxygenated with argon for some minutes in a Schlenk tube, and
the reaction was run under argon.
Reagents and Solvents: The solvents used in the preparation of the
complexes were of reagent grade and were used as received. Dichlo-
romethane used in the epoxidation reactions was HPLC grade. An
aqueous solution of NaOCl (available chlorine 10–13%, Aldrich)
was used as oxidant. Salicylaldehyde, 2-hydroxy-3-methoxybenz-
aldehyde, 2Ј-hydroxyacetophenone and the diamines 1,2-cyclohex-
anediamine (as cis/trans mixture) and ethylenediamine were pur-
chased from Aldrich. The substrate trans-β-methylstyrene was also
purchased from Aldrich and n-decane was obtained from Fluka.
Nickel acetate, nickel nitrate pentahydrate and sodium hydrogen
phosphate were obtained from Merck. All reagents were used with
no further purification, except the diamines, which were distilled
prior to use. The GC standard epoxide was prepared by epoxid-
ation of trans-β-methylstyrene with m-chloroperbenzoic acid (m-
CPBA) at 4 °C and stored at that temperature. The phase-transfer
catalyst benzyltributylammonium bromide was obtained from
Fluka. The zeolites in their sodium form, NaY and NaX, were
kindly provided by Grace GmbH (Germany) and were calcinated
at 600 °C under oxygen before use.
Acknowledgments
Instrumentation: Silicon, aluminium, sodium and nickel contents
This work was partially funded by Fundação para a Ciência e a
were determined by Kingston Analytical Services (UK) by ICP-
Tecnologia (FCT) and FEDER, through the project ref. POCTI/
4278
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2005, 4272–4279

(Salen)nickel-Catalysed Epoxidations
FULL PAPER
[19] M. R. Maurya, S. J. J. Titinchi, S. Chand, J. Mol. Catal. A:
Chem. 2003, 201, 119–130.
[20] D. Chatterjee, A. Mitra, J. Mol. Catal. A: Chem. 1999, 144,
363–367.
CTM/45449/2002. R. F. acknowledges FCT for a PhD fellowship.
We are indebted to Dr. M. J. Sabater (Valencia) for helpful dis-
cussions.
[21] D. Chatterjee, S. Mukherjee, A. Mitra, J. Mol. Cat. A: Chem.
2000, 154, 5–8.
[22] V. Ayala, A. Corma, M. Iglesias, J. A. Rincón, F. Sanchez, J.
Catal. 2004, 224, 170–177.
[23] A. Corma, H. García, Eur. J. Inorg. Chem. 2004, 1143–1164.
[24] V. Crupi, D. Majolino, V. Venuti, J. Phys. Condens. Matter
2004, 16, S5297–S5316.
[25] V. Ramaswamy, M. Sivarama, K. Krishnan, A. V. Ramaswamy,
J. Mol. Catal. A 2002, 181, 81–89.
[1] L. Canali, D. C. Sherrington, Chem. Soc. Rev. 1999, 28, 85–93.
[2] E. N. Jacobsen, A. Pfaltz, H. Yamamoto, Comprehensive
Asymmetric Catalysis, Springer, Berlin, 1999, vol. III.
[3] H. Nishikori, C. Ohta, T. Katsuki, Synlett 2000, 1557–1560.
[4]
[5]
[6]
[7]
[8]
M. J. Sabater, A. Corma, A. Domenech, V. Fornés, H. García,
Chem. Commun. 1997, 1285–1286.
J. F. Kinneary, T. R. Wagner, C. J. Burrows, Tetrahedron Lett.
1988, 29, 877–880.
J. F. Kinneary, J. S. Albert, C. J. Burrows, J. Am. Chem. Soc.
1988, 110, 6124–6129.
R. I. Kureshy, N. H. Khan, S. H. R. Abdi, P. Iyer, A. K. Bhatt,
J. Mol. Catal. A: Chem. 1998, 130, 41–50.
R. I. Kureshy, N. H. Khan, S. H. Abdi, S. t. Patel, P. Iyer, E.
Suresh, P. Dastidar, J. Mol. Catal. A: Chem. 2000, 160, 217–
227.
[26] Mainly chlorinated and allyclic hydroxylation products (iden-
tified by GC-MS).
[27] J. T. Groves, G. A. McClusky, J. Am. Chem. Soc. 1976, 98, 859–
861.
[28] S. Shaik, S. Cohen, S. P. de Visser, P. K. Sharma, D. Kumar, S.
Kozuch, F. Ogliaro, D. Danovich, Eur. J. Inorg. Chem. 2004,
207–226.
[29] C. T. Dalton, K. M. Ryan, V. M. Wall, C. Bousquet, D. G.
Gilheany, Top. Catal. 1998, 5, 75–91.
[9]
H. Yoon, C. J. Burrows, J. Am. Chem. Soc. 1988, 110, 4087–
4089.
[30] A. M. Daly, M. F. Renehan, D. G. Gilheany, Org. Lett. 2001,
[10] H. Yoon, T. R. Wagner, K. J. O’Connor, C. J. Burrows, J. Am.
Chem. Soc. 1990, 112, 4568–4570.
[11] R. A. Sheldon, R. S. Downing, Appl. Catal. A 1999, 189, 163–
183.
[12] J. M. Thomas, T. Maschmeyer, B. F. G. Johnson, D. S.
Shepard, J. Mol. Catal. A: Chem. 1999, 141, 139–144.
[13] R. Ferreira, M. Silva, C. Freire, B. de Castro, J. L. Figueiredo,
Microporous Mesoporous Mater. 2000, 38, 391–401.
[14] A. R. Silva, M. Martins, M. M. A. Freitas, J. L. Figueiredo, C.
Freire, B. de Castro, Eur. J. Inorg. Chem. 2004, 2027–2035.
[15] R. Ferreira, C. Freire, B. de Castro, A. P. de Carvalho, J. Pires,
M. Brotas de Carvalho, Eur. J. Inorg. Chem. 2002, 3032–3038.
[16] I. Kuzniarska-Biernacka, A. R. Silva, R. Ferreira, A. P. de Car-
3, 663.
[31] W. Adam, K. J. Roschmann, C. R. Saha-Möller, D. Seebach, J.
Am. Chem. Soc. 2002, 124, 5068–5073.
[32] For this reason, in certain oxidations using peroxides, such as
the Fenton reaction, H2O2 is often seen as an initiator rather
than an oxidant because it contributes only to initiate the for-
mation of radicals from O2. It is therefore interesting to explore
the catalytic activity of the complexes in the absence of oxygen.
[33] W. R. Sanderson, Pure Appl. Chem. 2000, 72, 1289–1304.
[34] K. M. Ryan, C. Bousquet, D. G. Gilheany, Tetrahedron Lett.
1999, 40, 3613–3616.
[35] C. Freire, B. de Castro, J. Chem. Soc., Dalton Trans. 1998,
1491–1497.
valho, J. Pires, M. Brotas de Carvalho, C. Freire, B. de Castro, [36] C. Freire, B. de Castro, Polyhedron 1998, 17, 4227–4235.
New J. Chem. 2004, 28, 853–858.
[37] R. H. Holm, G. W. Everett Jr., A. Chakravorty, in Progress in
Inorganic Chemistry (Ed.: F. A. Cotton), John Wiley and Sons,
New York, 1966, vol. 7, p. 183.
[17] A. R. Silva, J. L. Figueiredo, C. Freire, B. de Castro, Micro-
porous Mesoporous Mater. 2004, 68, 83–89.
[18] A. R. Silva, C. Freire, B. de Castro, Carbon 2004, 42, 3003–
3042.
Received: April 15, 2005
Published Online: September 21, 2005
Eur. J. Inorg. Chem. 2005, 4272–4279
www.eurjic.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4279