The catalytic function of SiO2-immobilized Mn(II)-complexes for alkene epoxidation with H2O2

Article abstract of DOI:10.1016/j.molcata.2009.11.021

Two symmetrical acetylacetone-based Schiff bases were immobilized on a silica surface by grafting and sol-gel procedure. The corresponding supported manganese complexes were prepared and evaluated as heterogeneous catalysts for alkene epoxidation with H₂O₂. These heterogeneous catalysts show remarkable effectiveness and selectivity towards epoxide formation in the presence of ammonium acetate. Moreover, the developed heterogeneous catalysts preserve the coordination and catalytic properties of the active-homogeneous manganese catalysts for alkene epoxidation vs. the competitive H₂O₂ dismutation. EPR spectroscopy shows that in heterogeneous manganese catalysts the Mn²⁺ centers are in a flexible, non-tight, coordination environment, as in the corresponding homogeneous manganese catalysts. However, after a first use of the heterogeneous catalysts, the Mn centers are detached from the ligand and are randomly dispersed on the SiO₂ surface. This is responsible for the loss of catalytic activity.

Full text of DOI:10.1016/j.molcata.2009.11.021

Journal of Molecular Catalysis A: Chemical 319 (2010) 58–65
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
The catalytic function of SiO2-immobilized Mn(II)-complexes for alkene
epoxidation with H2O2
Ag. Stamatisa, D. Giasafakia, K.C. Christoforidisb, Y. Deligiannakisb, M. Louloudia,∗
a Department of Chemistry, University of Ioannina, 45110 Ioannina, Greece
b University of Ioannina, Department of Environmental and Natural Resources Management, Laboratory of Physical Chemistry, Pyllinis 9, 30100 Agrinio, Greece
a r t i c l e i n f o
a b s t r a c t
Article history:
Two symmetrical acetylacetone-based Schiff bases were immobilized on a silica surface by grafting and
sol–gel procedure. The corresponding supported manganese complexes were prepared and evaluated as
heterogeneous catalysts for alkene epoxidation with H2O2. These heterogeneous catalysts show remark-
able effectiveness and selectivity towards epoxide formation in the presence of ammonium acetate.
Moreover, the developed heterogeneous catalysts preserve the coordination and catalytic properties of
the active-homogeneous manganese catalysts for alkene epoxidation vs. the competitive H2O2 dismu-
tation. EPR spectroscopy shows that in heterogeneous manganese catalysts the Mn2+ centers are in a
flexible, non-tight, coordination environment, as in the corresponding homogeneous manganese cat-
alysts. However, after a first use of the heterogeneous catalysts, the Mn centers are detached from the
ligand and are randomly dispersed on the SiO2 surface. This is responsible for the loss of catalytic activity.
Received 22 September 2009
Received in revised form
27 November 2009
Accepted 27 November 2009
Available online 21 December 2009
Keywords:
Catalytic epoxidation
Supported complexes
Manganese complexes
H2O2 activation
Modified silica
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
tions can explain why only a few such systems, based on supported
manganese catalysts and H2O2, have been reported for alkene
Epoxidation reactions that involve hydrogen peroxide, which
is still the ‘greenest’ terminal oxidant, is of great interest due to
the importance of epoxides in the manufacture of both bulk and
fine chemicals [1]. Despite success of homogeneous catalysts in the
epoxidations, there is a clear demand for heterogeneous catalysts
that provide advantages such as easy handling and product sep-
aration, catalyst recovery and less level of waste [2,3]. For these
reasons, the development of heterogeneous catalysts for epox-
idation reactions remains a very active field of research [4,5].
Preparation of heterogeneous catalysts, by covalent immobiliza-
tion of transition-metal complexes on silica surface represents one
of the most usable approaches [6–9]. Many of heterogeneous cata-
lysts contain metal ions and chemically modified silica gels with
active organic components [10]. Among the various transition-
metals for heterogeneous catalytic epoxidation, manganese stands
out as the most efficient, economical and environmental benign.
Various manganese complexes immobilized on an inorganic matrix
are known to be efficient and applicable as catalysts for the epoxida-
tion of a wide range of alkenes [11–15]. However, the development
of efficient manganese-systems that incorporate H2O2 has limita-
tions due to competitive H2O2 dismutation by both the Mn center
(catalase-type-activity) and the inorganic support. These limita-
epoxidation [16–20].
Thus, efficient, cost effective and robust manganese catalysts
remain at high demand in the field of hydrocarbon oxidation by
H2O2. Within this context, during the last years, our group has
developed imidazole based acetamide/Mn(II) systems [18,20] and
acetylacetone-based Schiff bases/Mn(II) systems [19,21] as homo-
geneous and heterogeneous catalysts for alkene epoxidation with
H2O2. More recently, we have presented homogeneous Mn(II)-
catalysts with new symmetrical acetylacetone-based Schiff bases
[22]. Their catalytic efficiency was shown to be switched-on by
ammonium acetate achieving remarkable effectiveness and selec-
tivity towards epoxides [22]. EPR spectroscopy provided evidence
that the catalytic centers are mononuclear Mn(II) complexes with
a flexible ligand-field environment [22].
Herein, the symmetrical acetylacetone-based Schiff bases are
immobilized on a silica surface by grafting and sol–gel procedure.
The obtained supported manganese complexes have been evalu-
ated as catalysts for alkene epoxidation with H2O2 and compared
with the corresponding homogeneous systems. EPR spectroscopy
has been also used to study the coordination environment of the
heterogeneous Mn-catalysts.
2. Experimental
∗
All substrates were purchased from Aldrich, in their highest
commercial purity, stored at 5 ◦C and purified by passage through
Corresponding author. Tel.: +30 26510 08418; fax: +30 26510 08786.
E-mail address: mlouloud@uoi.gr (M. Louloudi).
1381-1169/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2009.11.021

Ag. Stamatis et al. / Journal of Molecular Catalysis A: Chemical 319 (2010) 58–65
59
G-1MnIIacet: Metal loading: 0.28 mmol g−1. DRIFTS-IR (cm−1
,
selected peaks): 1704: ꢀ(C O); 1645: ꢀ(C N); 1548: ꢀsym(COO−);
1441: ꢀas(COO−); 1408: ꢀ(C–O). DRS (ꢁmax (nm)): 229, 310, 489.
a column of basic alumina with methanol as the eluting solvent,
prior to use. Hydrogen peroxide was 30% aqueous solution.
Elemental analyses (C, H, N) were obtained using a Perkin-Elmer
Series II 2400 elemental analyser. The manganese amount was
determined by flame atomic absorption spectroscopy on a Perkin-
Elmer AAS-700 spectrometer. Infrared spectra were recorded on a
Spectrum GX Perkin-Elmer FT-IR System. Continuous-wave (c.w.)
Electron Paramagnetic Resonance (EPR) spectra were recorded
with a Brucker ER200D spectrometer at liquid N2 temperature,
equipped with an Agilent 5310A frequency counter. The spec-
trometer was running under a home-made software based on
LabView. Diffuse reflectance UV–Vis spectra were recorded at
room temperature on a Shimadzu UV-2401PC with a BaSO4 coated
integration sphere. Thermogravimetric analyses were carried out
under atmosphere using Shimadzu DTG-60 analyser. GC analy-
sis was performed using an 8000 Fisons chromatograph with a
flame ionization detector and a Shimadzu GC-17A gas chromato-
graph coupled with a GCMS-QP5000 mass spectrometer. The N2
adsorption–desorption isotherms were measured at 77 K on a Sorp-
tomatic 1990, thermo Finnigan porosimeter. Specific surface areas
were determined with the Brunauer–Emmett–Teller (BET) method
using adsorption data points in the relative pressure (P/P0) range
of 0.05–0.30 and assuming a closely packed BET monolayer. H2O2
was added by a digitally controlled syringe pump type SP101IZ WPI
over 1 h under stirring. Solution potential Eh was measured by a
Metrohm platinum redox electrode (type 6.0401.100).
This material had an average surface area of ca. 286 m2 g−1
.
G-2MnIICl: Metal loading: 0.24 mmol g−1.DRIFTS-IR (cm−1
,
selected peaks): 1704: ꢀ(C O); 1645: ꢀ(C N); 1406: ꢀ(C–O); 1209:
ꢀ(CF3). DRS (ꢁmax (nm)): 227, 325, 476. This material had an average
surface area of ca. 198 m2 g−1
.
G-2MnIIacet: Metal loading: 0.25 mmol g−1.DRIFTS-IR (cm−1
,
selected peaks): 1706: ꢀ(C O); 1645: ꢀ(C N); 1528: ꢀsym(COO−);
1441: ꢀas(COO−);1402: ꢀ(C–O); 1209: ꢀ(CF3). DRS (ꢁmax (nm)):
227, 324, 487. This material had an average surface area of ca.
277 m2 g−1
.
2.4. Sol–gel procedure
By adaptation of grafting-methodology used, 1.2 mmol of 1
and 1.0 mmol of (3-chloropropyl)-trimethoxysilane were stirred
for 48 h in 10 ml of MeOH under N2 at 60 ◦C in order to gener-
ate OS-1 silicon precursor. After cooling, to this solution 2.4 mmol
of MnCl2·4H2O were added. The obtained solution was stirred
overnight at room temperature under N2 leading to metalated pre-
cursor OS-1MnIICl. The S-1MnIICl catalyst was prepared by sol–gel
method, through hydrolysis of TEOS and OS-1MnIICl precursor in
MeOH–H2O mixture in the presence of NaF as catalyst. For this, to
the precursor solution 4.5 ml TEOS, 3.6 ml H2O and 4.2 mg NaF were
added and mixed together for 10 min. The final molar ratio silicon
precursor:NaF:TEOS:H2O was equal to 1:0.1:20:200. The gelation
of the sol was observed in the next 3 h and the obtained gel was
kept for aging for 3 days at room temperature. The filtered solid
was extensively washed with water and methanol. The S-1MnIICl
material was further purified with MeOH using the soxhlet extrac-
tion method and dried under reduced pressure at 60 ◦C for 12 h. The
loading achieved is ca. 0.19 mmol g−1 determined by thermogravi-
metric and elemental analysis.
2.1. Ligand synthesis
Ligands 1 and 2 have been synthesized and characterized
recently [22].
2.2. Grafting procedure
In a typical synthesis, 2.0 mmol of 1 or 2 and 1.7 mmol of (3-
chloropropyl)-trimethoxysilane were stirred for 48 h in 50 ml of
MeOH under N2 at 60 ◦C. To this solution 3.0 g of SiO2 and 5 ml
of EtOH were added, and the slurred solution maintained at 60 ◦C
under N2 for 24 h. Filtering gave G-1 or G-2, respectively. The result-
ing materials were washed with MeOH and EtOH; they were further
purified with EtOH using the soxhlet extraction method and dried
under reduced pressure at 80 ◦C for 12 h. The loading achieved is
ca. 0.28 mmol g−1 for G-1 and 0.26 mmol g−1 for G-2 determined,
in both cases, by thermogravimetric and elemental analysis.
G-1: DRIFTS-IR (cm−1, selected peaks): 1652: ꢀ(C O); 1618:
ꢀ(C N); 1516: ꢀ(C C); 1463: ı(C–H); 1410: ꢀ(C–O). DRS (ꢁmax
(nm)): 232, 313, 491.
S-1MnIICl: Metal loading: 0.20 mmol g−1. DRIFTS-IR (cm−1
,
selected peaks): 1686: ꢀ(C O); 1658: ꢀ(C N); 1403: ꢀ(C–O). DRS
(ꢁmax (nm)): 244, 315, 480. This material had an average surface
area of ca. 465 m2 g−1
.
2.5. Representative catalytic conditions
The alkene (1 mmol), acetophenone or bromobenzene (inter-
nal standard, 1 mmol), catalyst (1 ␮mol) and CH3COONH4 additive
(1 mmol) in a acetone/MeOH (450 ␮l/400 ␮l) solvent mixture were
cooled to 0 ◦C. H2O2 (2 mmol) was added by a digitally controlled
syringe pump type SP101IZ WPI over 1 h under stirring. 10 min
later, the test tube was removed from the ice bath and allowed
to warm to room temperature 26 1 ◦C. The progress of the reac-
tion was monitored by GC–MS, for small samples taken from
the reaction mixture. GC analysis of the solution provided the
substrate conversion and product yield relative to the internal stan-
dard integration. To establish the identity of the epoxide product
unequivocally, the retention time and spectral data were compared
to those of an authentic sample. Blank experiments showed that
without Mn-catalyst or CH3COONH4, epoxidation reactions do not
take place.
G-2: DRIFTS-IR (cm−1, selected peaks): 1656: ꢀ(C O); 1619:
ꢀ(C N); 1511: ꢀ(C C); 1465: ı(C–H); 1410: ꢀ(C–O); 1212: ꢀ(CF3).
DRS (ꢁmax (nm)): 228, 320, 489.
2.3. Metalation procedure
Approximately
0.3 mmol
of
MnCl2·4H2O
or
Mn(CH3COO)2·4H2O and 0.5 g of G-1 or G-2 were stirred in
MeOH overnight before filtering. To remove any weakly coordi-
nated metal, the metalated materials, G-1MnIICl, G-1MnIIacet,
G-2MnIICl and G-2MnIIacet, were exhaustively washed with
MeOH, EtOH and Et2O and dried at 60 ◦C for 3 h. The amount of
manganese was determined by back-titration of the remaining
amount of metal ion into the solution and by atomic absorption
spectroscopy.
3. Results and discussion
3.1. Synthesis of the materials
G-1MnIICl: Metal loading: 0.26 mmol g−1. DRIFTS-IR (cm−1
,
selected peaks): 1706: ꢀ(C O); 1658: ꢀ(C N); 1406: ꢀ(C–O). DRS
(ꢁmax (nm)): 230, 315, 483. This material had an average surface
Immobilization of the ligands 1 and 2 requires their derivati-
zation which occurs via phenolic groups with (3-chloropropyl)-
trimethoxysilane (Scheme 1). The propyl-trimethoxysilane moiety
area of ca. 203 m2 g−1
.

60
Ag. Stamatis et al. / Journal of Molecular Catalysis A: Chemical 319 (2010) 58–65
Scheme 1. Silane-derivatization of the ligands 1 and 2.
of silicon precursors OS-1 and OS-2 allows covalent attachment of
1 and 2 on the silica surface.
on the support. Metalation of these materials with 2 equiv. of MnCl2
or Mn(CH3COO)2 per ligand molecule in methanol, gave G-1MnIICl,
G-1MnIIacet, G-2MnIICl and G-2MnIIacet, after removal of any
weakly associated metal with exhaustively washings. (ii) A mod-
ified sol–gel procedure involved an initial formation of metalated
precursor OS-1MnIICl followed by hydrolysis and co-condensation
in the presence of TEOS to yield S-1MnIICl. The catalytic epoxi-
dation reactivity of all prepared materials G-1MnIICl, G-1MnIIacet,
Two methods of covalent attachment were applied: grafting on
silica surface and sol–gel procedure providing materials G and S,
respectively (Scheme 2). (i) The grafting involved attachment of 1
and 2 to the support surface to form hybrid materials G-1 and G-2
with covalently attached ligands. Soxhlet extractions of G-1 and G-
2 ensure that only the covalently attached ligand molecules remain
Scheme 2. Schematic representation of the synthesis of manganese catalysts.

Ag. Stamatis et al. / Journal of Molecular Catalysis A: Chemical 319 (2010) 58–65
61
Fig. 2. Time dependence of cyclooctene epoxidation [black line] and solution redox
potential [red line] for the same reaction catalysed by G-1MnIIacet. (For interpre-
tation of the references to color in this figure caption, the reader is referred to the
web version of the article.)
3.2. Catalytic epoxidation with H2O2
To evaluate the catalytic behaviour of the prepared materials,
oxidations of several alkenes have been carried out at room temper-
ature. As in the case of homogeneous manganese catalysts [22], the
present heterogeneous catalyst required also ammonium acetate
as additive to produce efficient catalytic systems.
Fig. 1. EPR spectra of (i) homogeneous MnII1 system in acetone/MeOH (1/1, v/v),
(ii) unused heterogeneous G-1MnIICl catalyst in MeOH and (iii) recycled hetero-
geneous G-1MnIICl catalyst in MeOH. Experimental conditions: temperature 77 K,
modulation amplitude 5 G, microwave power 10 dB.
According to our data, the catalytic reactions were prac-
tically accomplished within 18 h. The time course profiles of
the G-1MnIIacet- and G-2MnIIacet-catalysed epoxidations of
cyclooctene, in conjunction with the observed redox potential of
solution Eh (vs. standard hydrogen electrode SHE) are given in
Figs. 2 and 3. At the beginning of the reaction catalysed by G-
1MnIIacet, Eh was +376 mV; after 2 h, it dropped at +243 mV with
a 37% epoxide yield and finally, after a reaction time of 18 h, it
approached a value of Eh = +175 mV with a 71% yield. In the case of
G-2MnIIacet-catalysed epoxidation, at reaction time t = 0, Eh was
+383 mV, at t = 2 h, then decreased at +246 mV providing a prod-
uct yield of 36% and at t = 18 h, Eh approached +178 mV with a
70% cyclooctene epoxide formation. Additions of new amounts of
oxidant slightly increased the Eh resulting in an additional 5% epox-
idation of the remaining substrate.
G-2MnIICl, G-2MnIIacet and S-1MnIICl are compared herein. When
the sol–gel procedure applied, starting with a covalent immobiliza-
tion of the ligand into the silica network, followed by metalation,
the materials obtained were inactive in catalytic epoxidations.
The DRIFTS-IR-spectra of the synthesized materials show sev-
eral bands corresponding to various structural units of the solids.
In the IR-spectra of G-1 and G-2, the bands at 1652 and 1656 cm−1
,
respectively are attributed to the ꢀ(C O) stretching vibrations;
accordingly the bands observed in both spectra, at ∼1618 cm−1
to the ꢀ(C N) stretching. Anchoring of the ligands 1 and 2 on
the silica surface can be conveniently followed also by DR UV
spectroscopy; in the corresponding spectra, the presence of the
␲ → ␲* transitions attributed to the immobilized ligands can be
seen. In the IR-spectra of the metalated materials G-1MnIICl, G-
1MnIIacet, G-2MnIICl, G-2MnIIacet, the C O and C N stretching
vibrations are clearly shifted when compared with the correspond-
ing non-metalated materials G-1 and G-2, indicating coordination
of manganese with the imine-nitrogen and the keto-oxygen of the
immobilized ligands 1 and 2. This is also supported by DR UV spec-
tra. The S-1MnIICl material presents the C O and C N stretching
vibrations at about 1686 and 1658 cm−1
.
Fig. 1(i) shows a 77K EPR spectrum for the heterogeneous G-
1MnIICl in [acetone/MeOH]. For comparison, an EPR spectrum for
the homogeneous MnII1 in [aceteone/MeOH] is also displayed,
Fig. 1(ii).
The EPR spectrum of G-1MnIICl (Fig. 1, spectrum (i)) corre-
sponds to mononuclear Mn2+ (S = 5/2, I = 5/2) centers [23]. Analysis
of the EPR spectrum based on the splittings of the semiforbidden
EPR transitions, using the method of Misra [24] or the inten-
sity method of Allen [25] as described previously [22] shows
that that the Zero-Field Splitting (ZFS) parameter for G-1MnIICl is
D = 0.40 0.5 GHz. The D-value for the homogenous MnII1 complex
is D = 0.38 0.5 GHz [22]. This shows that the ligands around the
Fig. 3. Time dependence of cyclooctene epoxidation [black line] and solution redox
potential [red line] for the same reaction catalysed by G-2MnIIacet. (For interpre-
tation of the references to color in this figure caption, the reader is referred to the
web version of the article.)
Mn2+ centers are slightly more strongly bound to the Mn2+
.

Table 1
a
Alkene epoxidations catalysed by supported manganese complexes in the presence of H2O2
.
b
II
II
II
b
II
II
Substrate
Products
MnCl
G-1Mn Cl
S-1Mn Cl
G-2Mn Cl
Mn(acet)
e
G-1Mn acet
G-2Mn acet
2
2
e
d
e
d
e
d
e
d
d
e
d
e
d
Yield
TON
TOF (%) Yield
TON
TOF (%) Yield
−1
TON
TOF (%) Yield
−1
TON
TOF (%) Yield
TON
TOF (%) Yield
TON
TOF (%) Yield
−1
TON
TOF (%)
−1
−1
c
c
c
−1
(h
c
−1
c
c
c
(h
)
(h
)
(h
)
)
(h
)
(h
)
(h
)
cis-Epoxide
13.1
0.4
0.2
–
40.5
2.5
1.1
49.0
4.1
2.2
29.6
2.8
1.5
12.4
0.4
0.2
56.4
3.0
2.3
53.0
4.0
2.8
Cyclohexene
2-Cyclohexenone
2-Cyclohexenol
cis-Diol
137
3.3
6
0.9
474
0.1
26.3
6.8
562
31.2
353
19.6
131
5
685
38.1
640
35.6
1.4
4.2
Cyclooctene
Hexene-1
cis-Epoxide
cis-Epoxide
18.0
4.6
180
46
8
2
58.6
9.5
586
95
32.6
5.3
51.0
14.0
510
140
28.3
7.8
49.7
9.3
497
93
27.6
5.2
16.8
4.1
168
41
7
2
70.9
10.0
709
100
39.4
5.6
70.1
10.0
701
100
38.9
5.6
Epoxide
Phenyl-acetaldehyde–
17.1
171
7
47.6
–
476
26.4
16.0
1.8
178
9.9
15.3
0.6
159
8.8
12.4
–
124
5
46.7
0.8
475
26.4
19.2
3.0
222
12.3
Styrene
trans-Methyl- trans-Epoxide
styrene
14.3
143
6
22.5
1.6
241
13.4
21.5
2.5
240
13.3
21.6
2.3
239
13.3
14.9
149
6
52.9
1.3
542
30.1
54.3
4.7
590
32.8
trans-Methyl-ketone–
–
f
Limonene
1,2-Epoxides
21.0
(11.8/9.2)
3.3
50.8
(27.9/22.9)
10.1
33.5
(18.5/15)
6.4
49
20.4
(11.7/8.7)
3.2
74.0
(41.5/32.5)
13.5
71.2
(38.9/32.3)
13.8
(26.8/22.2)
10.4
(cis-/trans-)
8,9-Epoxides
-Alcohol
-Ketone di-epoxide
0.1
243
–
10
0.7
0.8
624
34.7
1.7
1.0
0.9
435
128
24.2
7.1
1.1
0.8
613
305
34.1
16.9
0.1
237
194
10
8
0.9
1.1
895
265
49.7
14.7
2.0
1.1
881
320
48.9
17.8
cis-Epoxide
trans-Epoxide
-ketone
17.1
2.6
0.3
10.2
1.3
3.2
9.3
2.7
3.3
24.0
3.3
16.6
2.6
21.9
3.0
24.5
4.2
cis-
Stilbene
200
0.9
8
0.8
124
0.2
6.9
1.6
a
Conditions: ratio of catalyst:H2O2:CH3COONH4:substrate = 1:2000:1000:1000; equivalent of catalyst = 1 ␮mol in 0.85 ml CH3COCH3:CH3OH (0.45:0.40). Reactions were completed within 18 h.
MnCl2 and Mn(acet)2 have been evaluated as catalysts in homogeneous conditions. Reactions were completed within 24 h.
TOF: turnover frequency which is calculated by the expression [epoxide]/[catalyst] × time (h−1).
b
c
d
e
f
TON: total turnover number, moles of products formed per mole of catalyst.
Yields based on starting substrate and products formed. The mass balance is 98–100%.
Limonene 1,2-oxide was found as a mixture of cis- and trans-isomers and limonene 8,9-oxide as a mixture of two diastereoisomers.

Ag. Stamatis et al. / Journal of Molecular Catalysis A: Chemical 319 (2010) 58–65
63
Fig. 4. Bar chart representation of alkene epoxidations catalysed by MnCl2,
Mn(acet)2, G-1MnIICl, S-1MnIICl and G-1MnIIacet in the presence of H2O2. Reaction
Fig. 6. Distribution of oxidation products catalysed by G-1MnIIacet in the presence
of H2O2. Reaction performed in CH3COCH3/CH3OH (0.45:0.40 ml). See Table 1 for
further details.
performed in CH3COCH3/CH3OH (0.45:0.40 ml). See Table 1 for details.
According to the data in Table 1, epoxidation of a wide range of
olefins proceeds with high conversion and selectivity for the epox-
ide product (the mass balance is 98–100%) was achieved in most of
the cases (Table 1). For example, oxidation of cyclooctene provides
a 100% selectivity for cis-cyclooctene epoxide with 58.6%, 49.7% and
51.0% yield and 100% m.b. catalysed by G-1MnIICl, G-2MnIICl, and
S-1MnIICl, respectively (see Figs. 4 and 5). When the G-1MnIIacet,
and G-2MnIIacet catalysts were used, the cis-epoxide was also the
only product and the obtained yields were 70.9% and 70.1% (see
Figs. 4 and 5).
Hexene-1 is a rather hard oxidation substrate showing epoxide
yields from 9.3% to 14.0% and selectivity 100% for the cis-epoxide.
Cyclohexene achieves moderate epoxide yields, ranging from 29.6%
to 56.4%. In addition, small amounts of the corresponding diol,
probably as epoxide ring opening product, have been also detected
(0.9–6.8%). In some extent, allylic oxidation has also occurred with
the formation of 2-cyclohexene-1-ol and 2-cyclohexene-1-one. The
alcohol derivative determined to be 1.1–2.8% and the yield of the
corresponding ketone is 2.5–4.1%. Overall, cyclohexene is epoxi-
dized by the present catalytic systems with 82–87% selectivity (see
Fig. 6).
Styrene oxidation produces mainly epoxide with yields from
16.0% to 47.6%, and 90–98% selectivity; small amounts of
phenyl-acetaldehyde as side product have been detected. The
catalysts showed analogous ability towards the oxidation of trans-
methylstyrene resulted in 21.5–54.3% epoxide yield with >99%
retention of its configuration. Small amounts of the correspond-
ing methyl-ketone have been also observed (1.3–4.7%). The major
products detected from oxidation of limonene, were two epoxides
(cis- and trans-) derived from epoxidation of the electron-rich dou-
ble bond in 1,2-position and two diastereoisomers derived from
the more accessible but less electron-rich double bond in 8,9-
position. Additionally, small amounts of products formed by allylic
oxidation of the limonene ring have been observed. These were
identified as the corresponding derivatives of cyclohexene-1-ol and
cyclohexene-1-one. The total yield of epoxides (1,2- plus 8,9-) was
varied from 39.9% to 87.5%. It is noted that an additional 0.9% of di-
epoxide has been also formed. The corresponding cyclohexene-1-ol
was detected from 0.7% to 2.0% and the corresponding cyclohexene-
1-one from 0.8% to 1.1%. However, the epoxidation was clearly the
main reaction path resulting mainly in 1,2-epoxides (see Table 1
and Fig. 6). In the oxidation of cis-stilbene, the major product was
cis-epoxide with selectivity of 73–83% and yield ranging from 9.3%
to 24.5%. Moreover, considerable amounts of trans-stilbene epox-
ide (1.3–4.2%) and 1,2-diphenyl ethanone (0.8–3.3) have been also
detected.
When we inspect the influence of the R-substituents of the
ligands 1 and 2 on the catalytic activity, we notice that the het-
erogeneous catalysts based on either of these ligands are almost
equally efficient. This finding is in accordance with the catalytic
data observed by the corresponding homogeneous systems; EPR
data showed that the first coordination sphere of the Mn center
was not affected by R-substituents. Therefore the side chain modi-
fications play practically no-role in the catalytic activity [22].
When we evaluate the performance of the sol–gel material S-
1MnIICl and the corresponding grafting material G-1MnIICl for
alkene epoxidation based on the data of Table 1, no valid difference
is observed.
Fig. 5. Bar chart representation of alkene epoxidations catalysed by MnCl2,
Mn(acet)2, Mn(acet)2/LA G-2MnIICl and G-2MnIIacet in the presence of H2O2. Reac-
tion performed in CH3COCH3/CH3OH (0.45:0.40 ml). See Table 1 for details.
Competitive reactions show alkene reactivity to increase
with the electron density of the double bond when the

64
Ag. Stamatis et al. / Journal of Molecular Catalysis A: Chemical 319 (2010) 58–65
Table 2
Chemo- and region-selectivity in alkene epoxidation with H2O2 catalysed by G-1MnIICl, S-1MnIICl, G-2MnIICl, G-1MnIIacet, and G-2MnIIaceta.
Alkene 1
Alkene 2
Epoxide 1/epoxide 2b
G-1MnIICl
S-1MnIICl
G-2MnIICl
G-1MnIIacet
G-2MnIIacet
Styrene
Styrene
Cyclooctene
Cyclooctene
Cyclohexene
Cyclohexene
cis-1,2-Limon. epox.c
1,2-Limonene epoxided
trans-Methylstyrene
cis-Stilbene
Styrene
Cyclohexene
Hex-1-ene
1-Methylcyclohexene
trans-1,2-Limon. epox.c
8,9-Limonene epoxided
0.79
1.40
2.81
0.93
18.77
0.61
1.21
5.03
0.59
1.67
2.10
0.90
15.10
0.57
1.23
5.23
0.65
1.16
3.24
1.11
15.20
0.59
1.20
4.71
0.83
1.45
2.54
1.14
13.83
0.63
1.28
5.48
0.85
1.47
2.62
1.21
12.96
0.59
1.21
5.16
a
Conditions-ratio of catalyst:H2O2:CH3COONH4:substrate = 1:2000:1000:1000; equivalent of catalyst = 1 ␮mol in 0.85 ml CH3COCH3:CH3OH (0.45:0.40). Reactions were
usually complete within 18 h.
b
Yields based on starting substrate.
cis-1,2- vs. trans-1,2-Limonene epoxide.
1,2- vs. 8,9-Limonene epoxide.
c
d
more electron-rich olefin is oxidized preferentially, e.g. 1-
methylcyclohexene > cyclohexene > hex-1-ene (Table 2). These
data might be taken as evidence of the electrophilic nature of
oxygen transfer from manganese-oxo intermediate to the olefinic
double bond. However, additional effects of alkene shape are
obvious. For example, styrene epoxidation favours vs. cis-stilbene-
epoxidation and the electron-rich trisubstituted double bond of
limonene in 1,2-position gives approximately five times more
epoxides than the more accessible but less electron-rich double
bond in 8,9-position. The reactivity of other substrate types was
briefly investigated and the results are listed in Table 2.
Comparing the selectivity of the present heterogeneous cat-
alysts with the corresponding homogeneous systems [22], we
observe that the former prefer to epoxidize smaller substrate
molecules rather than bulky substrates. For example, the ratio
[styrene epoxide]/[cis-stilbene epoxide] varies from 1.16 to 1.67
for the heterogeneous catalysts (Table 2) vs. 0.73 to 0.99 for
the corresponding homogeneous catalysts [22]. This tendency is
enhanced by the sol–gel material S-1MnIICl supporting the sug-
gestion that the inorganic support prevents, in some extent, the
substrate approach to the catalyst active center due to steric hin-
drance. However, in general, the heterogeneous catalysts vs. the
homogeneous did not present dramatically different selectivity
with regard to the molecular dimension of the substrates. This
could be due to the elongate organic bridge which holds the active
catalyst in some distance from the silica support. Moreover, the
sol–gel material S-1MnIICl and the grafting material G-1MnIICl
present similar behaviour indicating that the S-1MnIICl-mediated
catalysis is accomplished mainly by the exposed surface active cen-
ters.
3.3. Comparison of chloride- vs. acetate-heterogeneous catalysts
The data of Table 1 and Figs. 4 and 5 show an enhanced
reactivity of the acetate-heterogeneous catalysts compared to
the chloride-heterogeneous catalysts, e.g. by 15–20%. This shows
that acetate has a beneficial influence on the heterogenized Mn-
catalysts studied herein. Strikingly, the same Mn-catalysts in the
homogeneous phase, showed no difference in their catalytic effi-
ciency between chloride and acetate [22]. This shows that in the
heterogenized catalysts, additional factors influence the catalytic
activity, e.g. compared with the homogeneous Mn-systems. These
factors can be (a) differences in the properties of the MnII-active
center. However our EPR and FT-IR data show only minor differ-
ences in the properties of the Mn(II)-complexes. (b) Significantly
higher surface areas measured for the acetate-heterogeneous cat-
alysts (285 m2 g−1 for the G-1MnIIacet, and 277 m2 g−1 for the
G-2MnIIacet) vs. the corresponding chloride-heterogeneous cat-
alysts (203 m2 g−1 for G-1MnIICl, and 198 m2 g−1 for G-2MnIICl).
Given that the catalysts present similar metal loading, we consider
that the increased surface area is correlated with enhanced catalytic
activity of the acetate-heterogeneous catalysts, in accordance with
previous observations [3,26].
3.5. Catalyst stability
Catalyst stability is of major concern in homogeneous and het-
erogeneous catalysis. In the present case, for the systems studied
herein two phenomena were screened in detail:
(a) Leaching on the Mn(II)-complexes from the SiO2 support: This was
tested by the ‘filtration method’ out according to Valkenberg
and Hölderich [3]. Accordingly, in typical catalytic experiments,
after 2 h, the solid catalysts were filtered. Into the filtrate, the
progress of the oxidation reaction was monitoring by GC. Our
data showed that, no evolution of the studied reactions was
observed in the filtrate. This ensures that no leaching of the
active supported Mn(II)-component occurs. Thus, the obtained
catalytic results derive exclusively from the heterogeneous cat-
alysts.
(b) Leaching of free Mn ions in solution: The catalytic performance of
the manganese salts MnCl2 and Mn(CH3COO)2 as homogeneous
catalysts was examined. For comparison the data are provided
in Table 1, and Figs. 4 and 5. As shown in Figs. 4 and 5 the
catalytic activity of the manganese salts in solution is much
inferior than those of the heterogeneous catalysts.
3.4. Comparison of hetero- vs. homogeneous catalysts
Comparing the present heterogeneous 1- and 2-based-
manganese catalysts with the corresponding homogeneous 1-
and 2-based-manganese catalysts [22], we observe that the het-
erogeneous systems lead to considerable epoxidation reactivity
with hydrogen peroxide. Moreover, the heterogeneous-acetate-
catalysts show enhanced catalytic activity compared with the
corresponding homogeneous-acetate-catalysts. That is, hetero-
genization process preserves the catalytic properties of the
active-manganese-centers for alkene epoxidation vs. the competi-
tive H2O2 dismutation and, in some cases increases their reactivity.
Overall, the control experiments provide strong evidence that
the observed catalytic results are derived by the heterogeneous
catalysts and no leaching of the active component or demetallation
occurs during the catalytic processes.

Ag. Stamatis et al. / Journal of Molecular Catalysis A: Chemical 319 (2010) 58–65
65
3.6. Catalyst recycling
4. Conclusions
It is known that reuse of heterogeneous catalysts as sup-
ported Schiff base complexes in epoxidation reactions is difficult
and remains challenging [27,28]. In our systems, to control the
behaviour of the heterogeneous catalysts after recycling, after a first
use the G-1MnIICl and S-1MnIICl catalysts were filtered, washed,
dried and reused under the same experimental conditions as in
Table 1. The data show that both G-1MnIICl and S-1MnIICl suffered
from significant loss of catalytic activity (>90%) after the first reuse.
In epoxidation reactions, as noted, upon recycling both activity and
selectivity deteriorate [27,28]. Often, upon epoxidation conditions
fracture of the ligand occurs, which limits recyclability. Herein we
provide further data which show that that ligand destruction is
occurring under epoxidation conditions.
The symmetrical acetylacetone-based Schiff bases 1 and 2 were
immobilized on a silica surface by two methods (a) grafting and (b)
sol–gel procedure. The corresponding supported manganese com-
plexes were prepared. The supported manganese complexes were
evaluated as heterogeneous catalysts for alkene epoxidation with
hydrogen peroxide and compared with the corresponding homo-
geneous systems published recently by our group [22].
The results obtained here demonstrate that the developed
heterogeneous catalysts preserve the coordination and catalytic
properties of the active-homogeneous manganese catalysts for
alkene epoxidation vs. the competitive H2O2 dismutation. More-
over, the data show remarkable effectiveness and selectivity
towards epoxide formation in the presence of ammonium acetate.
This indicates that the active catalytic features of the homogeneous
systems were successfully embedded onto the silica surface pro-
viding efficient heterogenized systems. EPR spectroscopy shows
that a flexible, not-tight, coordination environment is characteris-
ing the Mn-complexes in both homo- and hetero-phase. However,
the recycled catalysts suffer from significant loss of reactivity due to
ligand modification during catalytic experiments. EPR shows that
Mn ions are released and adsorbed randomly on the SiO2 surface.
3.7. Characterisation of used catalysts
DRIFTS-IR and DR UV spectra of the used catalysts showed
noticeable difference from the corresponding spectra of the unused
G-1MnIICl and S-1MnIICl indicating significant structural modifi-
cation of the organic ligands upon catalytic experiments. Moreover,
changes in the TGA-profile of the used catalysts were observed. The
EPR spectrum of the recycled G-1MnIICl shows a severe linebroad-
ening, see Fig. 1(iii). By measuring the EPR spectrum of centrifuged
catalyst, we have verified that the broad EPR spectrum is due
to MnII-ions on the SiO2. This line broadening originates from
non-specific spin–spin interactions between neighboring Mn spins
and/or distributed D-values [22,29]. Thus the EPR data show that
after the catalytic use of G-1MnIICl, the manganese centers are ran-
domly dispersed on the SiO2 surface. Only a small percentage <10%,
of mononuclear MnII center are resolved showing that most (90%)
of the MnII1 complexes have been destroyed.
Overall, the data demonstrate that during the catalytic pro-
cesses: (a) no leaching of the active-homogeneous catalysts occurs.
(b) After extended catalytic activity, i.e. hundred TONs, oxidative
destruction of the organic ligand occurs. (c) The Mn ions are ran-
domly dispersed on the SiO2 surface.
This shows that oxidative modification of the ligands is respon-
sible for loss of catalytic activity of the recycled catalyst, after
extended catalytic activity.
Acknowledgement
The authors thank the postgraduate program “Bioinorganic
Chemistry” at the Department of Chemistry, University of Ioannina.
References
[1] G. Grigoropoulou, J.H. Clark, J.A. Elings, Green Chem. 5 (2003) 1–7.
[2] W.F. Holderich, F. Kollmer, Pure Appl. Chem. 72 (2000) 1273–1287.
[3] M.H. Valkenberg, W.F. Holderich, Catal. Rev. 44 (2002) 321–374.
[4] D.E. De Vos, B.F. Sels, P.A. Jacobs, Adv. Synth. Catal. 345 (2003) 457–473.
[5] D.E. De Vos, M. Dams, B.F. Sels, P.A. Jacobs, Chem. Rev. 102 (2002) 3615–3640.
[6] D. Brunel, N. Bellocq, P. Sutra, A. Cauvel, M. Lasperas, P. Moreau, F.D. Renzo, A.
Galarneau, F. Fajula, Coord. Chem. Rev. 178–180 (1998) 1085–1108.
[7] J.S. Rafelt, J.H. Clark, Catal. Today 57 (2000) 33–44.
[8] J.H. Clark, D.J. Macquarrie, Chem. Commun. (1998) 853–860.
[9] D.E. De Vos, P.A. Jacobs, Catal. Today 57 (2000) 105–114.
[10] P.M. Price, J.H. Clark, D.J. Macquarrie, J. Chem. Soc., Dalton Trans. (2000)
101–110.
[11] P. Sutra, D. Brunel, Chem. Commun. (1996) 2485–2486.
[12] I.C. Chisem, J. Rafelt, M.T. Shieh, J. Chisem, J.H. Clark, R. Jachuck, D. Macquarrie,
C. Ramshav, K. Scott, Chem. Commun. (1998) 1949–1950.
[13] X.-G. Zhou, X.-Q. Yu, J.-S. Huang, S.-G. Li, L.-S. Li, C.-M. Che, Chem. Commun.
(1999) 1789–1790.
3.8. Mechanistic aspects
[14] P. Oliveira, A. Machado, A.M. Ramos, I.M. Fonseca, F.M. Braz Fernandes, A.M.
Botelho do Rego, J. Vital, Catal. Commun. 8 (2007) 1366–1372.
[15] T.J. Terry, T.D.P. Sttack, J. Am. Chem. Soc. 130 (2008) 4945–4953.
[16] Y.V.S. Rao, D.E. De Vos, T. Bein, P.A. Jacobs, Chem. Commun. (1997) 355–356.
[17] M.A. Martinez-Lorente, P. Battioni, W. Kleemiss, J.F. Bartoli, D. Mansuy, J. Mol.
Catal. A 113 (1996) 343–353.
[18] D. Zois, Ch. Vartzouma, Y. Deligiannakis, N. Hadjiliadis, L. Casella, E. Monzani,
M. Louloudi, J. Mol. Catal. A 261 (2007) 306–317.
[19] Ch. Vartzouma, E. Evaggellou, Y. Sanakis, N. Hadjiliadis, M. Louloudi, J. Mol.
Catal. A 263 (2007) 77–85.
[20] A. Serafimidou, A. Stamatis, M. Louloudi, Catal. Commun. 9 (2008) 35–39.
[21] M. Louloudi, Ch. Kolokytha, N. Hadjiliadis, J. Mol. Catal. A 180 (2002) 19.
[22] Ag. Stamatis, P. Doutsi, Ch. Vartzouma, K.C. Christoforidis, Y. Deligiannakis, M.
Louloudi, J. Mol. Catal. A 297 (2009) 44–53.
[23] G.H. Reed, G.D. Markham, in: L.J. Berliner, J. Reuben (Eds.), Biological Magnetic
Resonance, vol. 6, Plenum, New York, 1984, pp. 73–142.
The EPR spectra in Fig. 1(i) and (ii) show that the coordination
environment of the Mn2+ centers is comparable in the homoge-
neous and heterogeneous catalysts. The low D-values show that
the ligand field around the Mn2+ is weak in both cases. Impor-
tantly this shows that the heterogenization, e.g. in G-1MnIICl,
does not alter the coordination environment of Mn2+. A slightly
stronger ligand field, i.e. evidenced by the increase in the D-value
is the only effect of the heterogenization. Previously we have
provided evidence that the low D-value for MnII1 concurs with
good catalytic activity [22]. In contrast a severe increase in D-
value, indicative of a strong ligand-binding, has a severe inhibitory
effect on the catalytic efficiency, as observed for oxalate additive
[22].
[24] S. Misra, Physica B 203 (1994) 193–200.
[25] B.T. Allen, J. Chem. Phys. 43 (1965) 3820–3826.
This analysis shows that in the heterogeneous G-1MnIICl
catalyst the Mn2+ center retains a flexible coordination envi-
ronment as in the case of the homogenous MnII1 catalyst. This
explains the observed efficiency in catalytic activity as we show
herein.
[26] H. Chen, W.L. Dai, J.F. Deng, K. Fan, Catal. Lett. 81 (2002) 131–136.
[27] D.C. Sherrington, Catal. Today 57 (2000) 87–104.
[28] K.C. Gupta, A.K. Sutar, C.-C. Lin, Coord. Chem. Rev. 253 (2009) 1926–1946.
[29] A. Bencini, D. Gatteschi, Electron Paramagnetic Resonance of Exchange Coupled
Systems, Springer-Verlag, Berlin, 1990.