Improved robustness of heterogeneous Fe-non-heme oxidation catalysts: A catalytic and EPR study

Article abstract of DOI:10.1016/j.apcata.2013.11.009

There is currently a rarity in production and in-depth catalytic study of heterogeneous non-heme Fe catalysts. Herein, two heterogeneous catalysts have been synthesized by covalent grafting of non-heme Fe-complexes, DPEIFe ^IIICl and HFEIFe^IIICl, on SiO₂. The catalytic performance of the obtained DPEIFe^III@SiO₂ and HFEIFe ^III@SiO₂ materials has been systematically studied for catalytic oxidation of cyclohexene. The catalytic data show that the present non-heme Fe catalysts are functional and can achieve higher activity compared to other non-heme Fe reported so far in the literature. Importantly, the heterogeneneous catalysts show a remarkable robustness and improved oxidative stability vs. the homogeneous ones. Studies by UV-vis and EPR reveal a common mechanistic pattern: CH₃CN interacts with the Fe-atom promoting the formation of a Low-Spin (S = 1/2) intermediate, in the presence of H ₂O₂, probably a Fe^III-OOH hydroperoxide. The role of radical intermediates was investigated in detail by spin-trapping techniques. Finally, taking into account the nature of oxidation products, a consistent catalytic mechanism, valid for both homogeneous and heterogeneous catalysts, is discussed.

Full text of DOI:10.1016/j.apcata.2013.11.009

Applied Catalysis A: General 470 (2014) 376–389
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Improved robustness of heterogeneous Fe-non-heme oxidation
catalysts: A catalytic and EPR study
G. Bilisa, P. Stathia, A. Mavrogiorgoua, Y. Deligiannakisb,∗, M. Louloudia,∗∗
a Laboratory of Inorganic Chemistry, Department of Chemistry, University of Ioannina, 45110 Ioannina, Greece
b Laboratory of Physical Chemistry, Department of Environmental and Natural Resources Management, University of Patras, Seferi 2, 30100 Agrinio, Greece
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 23 July 2013
Received in revised form 1 November 2013
Accepted 5 November 2013
Available online 15 November 2013
There is currently a rarity in production and in-depth catalytic study of heterogeneous non-heme Fe
catalysts. Herein, two heterogeneous catalysts have been synthesized by covalent grafting of non-
heme Fe-complexes, DPEIFeIIICl and HFEIFeIIICl, on SiO2. The catalytic performance of the obtained
DPEIFeIII@SiO2 and HFEIFeIII@SiO2 materials has been systematically studied for catalytic oxidation of
cyclohexene. The catalytic data show that the present non-heme Fe catalysts are functional and can
achieve higher activity compared to other non-heme Fe reported so far in the literature. Importantly,
the heterogeneneous catalysts show a remarkable robustness and improved oxidative stability vs. the
homogeneous ones. Studies by UV–vis and EPR reveal a common mechanistic pattern: CH3CN interacts
with the Fe-atom promoting the formation of a Low-Spin (S = 1/2) intermediate, in the presence of H2O2,
probably a FeIII-OOH hydroperoxide. The role of radical intermediates was investigated in detail by spin-
trapping techniques. Finally, taking into account the nature of oxidation products, a consistent catalytic
mechanism, valid for both homogeneous and heterogeneous catalysts, is discussed.
Keywords:
Nonheme iron
Oxidation catalysts
Biomimetic catalysis
Homogeneous
Heterogeneous
H2O2
Fe-OOH
Fe-spin states
TEMPO
© 2013 Elsevier B.V. All rights reserved.
DMPO
1. Introduction
and development of novel efficient catalysts. Achieving robust-
ness against ligand degradation under oxidizing conditions is also
The development of competitive technologies for hydrocarbon
catalytic oxidation remains an important challenge for synthetic
chemistry and technology. Moreover, developing catalysts able to
achieve high selectivity via environmentally friendly processes and
inexpensive oxidants, is at high demand with direct environmental
and financial benefits [1,2]. To satisfy these demands, we may be
inspired by natural enzymes e.g. since certain enzymes undergo
oxidative processes with astonishing prospects. For instance,
cytochromes P450 – a family of heme enzymes – are able to catalyze
quite efficiently the oxidation of various substrates including epox-
idation of olefins [3,4]. In addition, non-heme Fe enzymes – like
methane monoxygenase (MMO) – are capable for selective oxida-
tion of small hydrocarbon compounds via binding and activation
of dioxygen or hydro-peroxo units [5,6]. Detailed understanding
of the catalytic mechanism and more particularly the nature of
the active oxidizing species are of key-importance for the design
a standalone challenge that – although often overlooked – has to
be addressed [7]. Thus, within the “green chemistry/green tech-
nology” concept, immobilization of a catalyst on a solid matrix has
been proposed as a viable approach [8] aiming at lower toxicity
and economically benign catalysts e.g. thanks to catalyst recov-
ery and recycling, as well as stability/endurance and reusability
of the catalyst [9]. Therefore, currently, the majority of industrial
processes are carried out using heterogeneous catalysts. In this con-
text, the use of silica support has become of interest, because SiO2
exhibits features such as mechanical robustness [10,11], chemical
and thermodynamic stability [12], promoting the stability of the
associated homogeneous catalyst [13]. All these properties render
silica suitable for commercial catalytic applications [10].
From the mechanistic/catalytic point of view, the SiO2 matrix
can be viewed as acting in a manner analogous to the protein
matrix in a biological system [14] e.g. for example exerting steric,
hydrophobic, electronic effects [14]. Moreover our group [15,16]
and others [17,18] have demonstrated that hetrogenization by
covalent grafting on silica surface can improve decisively the sta-
bility and recyclability of Mn-based [17,18] and Cu-based [16] com-
plexes. On the other hand, despite the abundance of Fe in natural
systems, few synthetic non-heme Fe complexes allow conversion
∗
Corresponding author.
Corresponding author. Tel.: +30 2651 0 08418.
E-mail addresses: ideligia@cc.uoi.gr (Y. Deligiannakis), mlouloud@uoi.gr
∗∗
(M. Louloudi).
0926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.11.009

G. Bilis et al. / Applied Catalysis A: General 470 (2014) 376–389
377
of H2O2 toward selective hydrocarbon oxidation [19–22]. Limit-
ing factor is that the majority of the non-heme Fe complexes that
served as catalysts either suffer from oxidative damage of the ligand
during catalysis or show simple Fenton-type catalytic behavior i.e.,
characterized by low conversions and extensive H2O2 decomposi-
tion [23]. In the case of grafted Fe-complexes, the inorganic support
is expected to further favor H2O2 decomposition [24]. However, the
emergence of active heterogeneous non-heme Fe catalysts by H2O2
prompted us the last years to graft new Fe-biomimetic complexes
on SiO2 matrix and to investigate their potential as catalysts for
hydrocarbon oxidation using the clean oxidant H2O2 [25].
Overall, the aims of the present work were (a) to synthesize two
novel heterogeneous materials HFEIFeIII@SiO2, DPEIFeIII@SiO2
and to study their catalytic efficiency as oxidation catalysts. (b) To
compare the performance and stability of the heterogeneous vs.
the homogeneous systems. (c) To trap, identify and study the cat-
alytic transient reaction intermediates. (d) To develop a consistent
catalytic reaction mechanism.
2. Experimental
All solvents and reagents were of commercial grade, unless oth-
erwise stated, purchased from Merck or Aldrich. CH3CN and CH2Cl2
were distilled from CaH2 and acetone from MgSO4. Silica gel Grade
645 was purchased from Aldrich and was activated at 200 ◦C for
12 h before use. All substrates were purchased from Aldrich, in their
highest commercial purity, stored at 5 ◦C and purified by passage
through a basic alumina column prior to use. 30% aqueous solution
of fresh H2O2 was used. Elemental analyses (C, H, N) were obtained
using a Perkin Elmer Series II 2400 elemental analyzer. Thermo-
gravimetric analyses were carried out under air using a Shimadzu
DTG-60 analyser. Infrared spectra were recorded on a Spectrum GX
Perkin-Elmer FT-IR System. UV–vis spectra were recorded using a
UV/VIS/NIR JASCO Spectrophotometer and a Perkin-Elmer Lamda
35 with a diffuse reflectance setup. Fe quantitation was done
by Perkin-Elmer AAS-700 Flame Atomic Absorption spectroscopy
(FAAS). Mass spectra were measured on an Agilent 1100 Series
LC-MSD-Trap-SL spectrometer. GC analysis was performed using
an 8000 Fisons chromatograph with a flame ionization detector
or a Shimadzu GC-17A gas chromatograph coupled with a GCMS-
QP5000 mass spectrometer.
Moreover, as soon as H2O2 is involved in the catalytic reac-
tion, the nature of the reaction path is controversial. That is why
the mechanistic landscape for non-heme Fe synthetic catalysts
has started to be explored only lately. A transient hydroperoxo-
iron [FeIII-OOH] has been proposed to be the first key-species
for the oxidative catalytic cycle for active non-heme Fe synthetic
complexes [26]. The formation of such hydroperoxo-iron moi-
ety [FeIII-OOH] occurs (i) either via deprotonation of H2O2 – and
direct ligation to the metal center of the catalyst [27,28] or (ii)
as a stepwise product promoted by the reaction between FeIII
-
OH and H2O2 [29]. Which of the two paths will be dominant,
depends on the oxidation state of iron and its ligand environment.
Intriguingly, the spin-state of the FeIII atom e.g. High-Spin (S = 5/2)
or Low-Spin (S = 1/2), appears to be correlated with the catalytic
efficiency [26,30–34]. So far, these key-phenomena have been stud-
ied for homogeneous catalytic systems, while there is a rarity on
synthesis and catalytic study of heterogenized non-heme Fe cata-
lysts.
The cited literature works entail that the catalytic/mechanistic
landscape of the non-heme iron catalysts is multiparametric since it
may be influenced by (i) the reaction conditions: solvent, protona-
tion events, substrate type, (ii) the ligand scaffold, the coordination
environment, the spin-state of the metal and (iii) the oxidant type
and redox potential. In this context, a safe approach for understand-
ing and optimization of a given catalytic non-heme system requires
a combination of catalytic/analytical data with appropriate spectro-
scopic data recorded under comparable conditions [25,26,30–34].
In such an effort, key-parameters to be monitored include: (i) detec-
tion and quantitative determination of catalytic products, (ii) the
evolution of the redox and spin state of the non-heme Fe catalytic
center, (iii) the effect of the physicochemical environment e.g. sol-
vent, oxidant type and the inorganic matrix in the heterogeneous
catalysts. Such an integrated approach is still scarce in the litera-
ture.
2.1. Catalysts synthesis and characterization
Synthesis and characterization of Schiff base ligand,
1,1,1,5,5,5-hexafluoro-4-{2-[2-(4,4,4-trifluoro-3-hydroxy-
1-trifluoromethyl-but-2-enylidene-amino)-ethylamino]-
ethylimino}-pent-2-en-2-ol (HFEI) (Chart 1), was done by the
protocol developed previously by our group [35].
2.1.1. Preparation and characterization of HFEIFeIIICl
Reaction of HFEI (0.58 mmol) with FeCl3 (1.16 mmol) in an
acetonitrile/EtOH solution (v/v, 10/5 ml) resulted in the forma-
tion of the mononuclear complex HFEIFeIIICl. Anal. Calcd. for
C14H11ClF12FeN3O2 (%): C, 29.37; N, 7.34; H, 1.92. Found: C, 29.50;
N, 7.25; H, 2.01. IR (KBr, cm−1, selected peaks) 3148: ꢀ(NH); 1649:
ꢀ(C · · · O); 1619: ꢀ(C N); 1567, 1542: ꢀ(C C); 1448: ꢀ(C O); 1255,
1216, 1145: ꢀ(CF3). UV (CH3CN, ꢁmax (nm), ε (M−1 cm−1) 292
(13,470); 352 (5180); 471 (750); 502 (743). ESI-MS (m/z) 328.3
([M+2H.2CH3CN]2+); 573.3 ([MH]+).
In this context, the present work presents a combined cat-
alytic/analytical and mechanistic/spectroscopic study for two
heterogeneous non-heme Fe catalysts, derived by covalent
immobilization of two non-heme Fe complexes HFEIFeIIICl and
DPEIFeIIICl (Chart 1) on SiO2.
A well characterized catalyst
DPEIFeIIICl, whose synthesis and the basic catalytic evaluation in
CH3CN has been reported [25] was used. In addition a new homo-
geneous iron catalyst HFEIFeIIICl, associated with a macroacyclic
Schiff base HFEI (Chart 1) has been synthesized. The underlying
reason for choosing these two complexes was to examine the effect
of the terminal groups e.g. C6H5-groups in DPEI vs. CF3-groups in
HFEI, on the catalyst reactivity and stability.
2.1.2. Preparation and characterization of HFEI@SiO2
This synthetic procedure developed by our group [35]
is represented in Scheme 1: to
a stirred solution of 50 ml
toluene containing 1.0 mmol of HFEI, 1.0 mmol of (3-
glycidyloxypropyl)trimethoxysilane was added. The resulting
mixture was allowed to react at 80 ◦C for 24 h. To this solution 1.5 g
of SiO2 and 5 ml of EtOH were added, and the slurry solution was
stirred at 80 ◦C for 24 h. The so-functionalized HFEI@SiO2 material
was isolated by filtration and washed with MeOH and EtOH. It was
further purified with EtOH using the soxhlet extraction method.
The loading achieved is ca. 0.2 mmol g−1 determined by thermo-
gravimetric and elemental analysis. DRIFTS-IR (cm−1, selected
peaks): 1683: ꢀ(C · · · O);1634: ꢀ(C N); 1571, 1550: ꢀ(C C);1395:
ꢀ(C O); 1285: ꢀ(CF3). DRS (ꢁmax (nm)): 323.
The
synthesized
heterogeneous
HFEIFeIII@SiO2,
DPEIFeIII@SiO2 materials were evaluated for catalytic oxida-
tion of hydrocarbons using H2O2. The catalytic study was carried
out in various solvents by comparing performance of the hetero-
geneous vs. the homogenous complexes. Stability and reusability
of the heterogeneous HFEIFeIII@SiO2, DPEIFeIII@SiO2 materials
was studied in detail. In parallel, redox and spin-intermediates of
the Fe-centers have been trapped and detected using UV–vis and
EPR spectroscopies.

378
G. Bilis et al. / Applied Catalysis A: General 470 (2014) 376–389
Chart 1. Ligand and catalysts structures.
2.1.3. Preparation and characterization of HFEIFeIII@SiO2
To suspension of HFEI@SiO2 (0.3 g) in CH3CN, FeCl3
[26,36]) to the reaction mixture under inert (Ar) atmosphere,
at room temperature (26 ◦C). Acetophenone or bromobenzene
were used as internal standards. The optimum ratio [cata-
lyst:oxidant:substrate] was [1:20:1000] for substrates such as
cyclohexene, methyl-cyclohexene, cyclooctene, limonene and
cyclohexane and [1:50:1000] for styrene, trans-␤-methyl styrene
and cis-stylbene [37]. 1 ␮mol of homogeneous HFEIFeIIICl or
DPEIFeIIICl, and 5.0 mg of heterogeneous HFEIFeIII@SiO2 or
DPEIFeIII@SiO2 catalyst [corresponding to 1 ␮mol of Fe-complex]
were used. The total volume of the catalytic reaction was 1 ml.
It is essential to mention that, a large excess of substrate was
used to minimize over-oxidation of catalytic products and protect
the catalyst from potential oxidative degradation [38]. Products
were identified by GC–MS vs. commercially available compounds.
Blank experiments showed that without catalyst the oxidative
reactions did not take place. The conversion yields reported
herein is based on the amount of oxidant H2O2 converted to
oxygenated products. Reactions were complete within 4 h for
DPEIFeIIICl, 2 h for HFEIFeIIICl and 24 h for the heterogenized cata-
lysts. As we discuss herein, this considerably different completion
times among the two homogeneous catalysts reflects the ligand
a
(0.122 mmol) was added. The mixture was stirred for 24 h at room
temperature and the resulting yellow material HFEIFeIII@SiO2 was
filtered, washed with CH3CN, CH3COCH3, and Et2O and dried at
80 ◦C for 12 h. Metal loading, by flame atomic absorption spec-
troscopy: 0.2 mmol g−1. DRIFTS-IR (cm−1, selected peaks): 1655:
ꢀ(C · · · O); 1628: ꢀ(C N); 1570, 1552: ꢀ(C C); 1443: ꢀ(C O);
1278(br): ꢀ(CF3). DRS (ꢁmax (nm)): 243, 298, 334, 496.
2.1.4. Preparation and characterization of DPEIFeIIICl and
DPEIFeIII@SiO2
The preparation of DPEIFeIIICl was done as describe previously
[25]. The heterogeneous material has been prepared by covalent
grafting of DPEI on SiO2 as detailed previously [35].
2.2. Catalytic conditions
2.2.1. GC–MS
H2O2 (30%) diluted in CH3CN (1/10, v/v) was slowly added
(within 10 min to minimize Fe-catalyzed H2O2 disproportionation
H
N
SiO2
SiO2
F3C
N
N
CF3
OH
HO
OH
N
OH
O
O
F3C
CF3
N
N
HFEI
CF3
in CH3CN
F3C
F3C
F3C
SiO2, EtOH
80o C, 24h
CF3
FeCl3
N
O
N
N
O
toluene
Fe
+
OH
80o C, 24h
HO
Cl
O
MeO
CF3
F3C
CF3
Si
MeO
MeO
O
HFEIFeIII@SiO2
HFEI@SiO2
Scheme 1. Synthesis of the heterogenized catalyst HFEIFeIII @SiO2.

G. Bilis et al. / Applied Catalysis A: General 470 (2014) 376–389
379
structure differences i.e., C6H5-groups of DPEIFeIIICl vs. CF3-groups
of HFEIFeIIICl.
in 10 ml of methanol). After 30 s incubation at room temperature,
the 77 K EPR spectrum of the sample was recorded. Two types of
control samples were also examined. One (control-1) contained
all the reactants except Fe-catalyst, while control-2 contained all
reactants except oxidant.
2.2.2. Radical trapping during catalysis
In the experiments with radical traps, the appropriate spin-
trapping-compounds, e.g., CBrCl3 or Ph2NH (40 ␮mol) were added
to the catalytic reaction mixture.
2.3.4. Spin trapping with DMPO
DMPO (5,5-dimethyl-1-pyrroline N-oxide) was used to moni-
tor formation of OH radicals [40]. In 100 ␮l of reaction mixture
100 ␮mol of DMPO was added from a 12.5 M stock solution in
methanol. The EPR spectrum of the sample was recorded in capillar-
ies, at room temperature. Two types of control samples were also
examined. Control-1 contained all reactants except Fe-catalyst and
control-2 contained all reactants except oxidant. In both control
experiments no hydroxyl radicals were detected.
2.2.3. Catalyst reuse experiments
Catalyst reusability for the heterogeneous catalysts were tested
with cyclohexene as substrate, for [catalyst:H2O2:cyclohexene] =
[1:20:1000 ␮mol], under the catalytic conditions described above.
After completion of each oxidation reaction, the solid catalyst was
isolated from the reaction mixture by centrifugation, washed with
CH3CN and dried under vacuum (80 ◦C, 2 h). The mass of the recov-
ered solid was estimated by mass weighting. Subsequently the
recovered solid catalyst was reused for a new oxidation reaction
under the same catalytic conditions.
2.3.5. Simulation of the EPR spectra
Easy Spin 4.0.0 [41] software was used to simulate the experi-
mental EPR spectra.
2.2.4. UV–vis sample preparation
3. Results
UV–vis samples were prepared in quartz cuvette and cooled
rapidly at 0 ◦C in ice-bath. 102 ␮l of substrate were added in
2.876 ml of 0.33 mM homogeneous DPEIFeIIICl or HFEIFeIIICl solu-
tion. Then, 22 ␮l H2O2 (in acetonitrile 1/10, v/v) were added, and
UV–vis spectra were recorded at various reaction times. The molar
ratio [catalyst:H2O2:cyclohexene] was [1:20:1000] and the final
concentration of catalyst was 0.30 mM.
3.1. Catalytic evaluation
3.1.1. Effect of solvent
The most effective solvents for H2O2 activation are expected
to be polar, non-coordinating, non-basic and inert under oxidiz-
ing conditions [42]. Water-immiscible solvents have the advantage
of minimizing hydrolysis of epoxides, however, when aqueous
H2O2 is used as oxidant they result in mass-transfer limitations
[42]. On the other hand, halogenated solvents are undesirable
because they are toxic and suffer in safety and handling aspects
[42,43]. To monitor the effect of solvent, the catalytic activ-
ity of DPEIFeIIICl and HFEIFeIIICl were studied in five types of
solvent media: CH3CN, CH3CN/CH2Cl2 (v/v, 1:1), CH2Cl2, ace-
tone and tert-amyl alcohol. The catalytic data for cyclohexene
[catalyst:H2O2:substrate] = [1:20:1000] are presented in Fig. 1A
and B by DPEIFeIIICl and HFEIFeIIICl, respectively.
Fig. 1A and B shows that the highest cyclohexene oxidation
yield was achieved in CH3CN. Acetonitrile is a coordinating sol-
vent, which can act as a strong ligand of non-heme Fe complexes,
easily replacing anionic ligands [44]. It has also been proposed that
CH3CN is able to stabilize high- and low-spin intermediates of Fe
centers during the catalytic cycle [45,46] promoting, in this way,
the homolytic cleavage of O O bonds [45,47]. As we show in the
following by UV–vis and EPR data (see for example Fig. 8), in our
systems, addition of H2O2 [in CH3CN] promoted the formation of
low-spin (S = 1/2) DPEIFeIII-OOH and HFEIFeIII-OOH intermediates.
It has been reported that CH2Cl2 – as a halogenated solvent – can
be efficient solvent in catalysis with H2O2 in cases where CH2Cl2
can polarize solvated molecules, by weakening the O O bond and
making them more reactive [43]. In contrast, our data, Fig. 1A, show
that lower epoxide and allylic oxidation products were produced
by DPEIFeIIICl in CH2Cl2. Moreover, when HFEIFeIIICl was used as
catalyst for cyclohexene oxidation in CH2Cl2, Fig. 1B, zero oxida-
tion products are detected. As we show in the following e.g. in the
UV–vis and EPR study, different Fe-intermediates were identified in
CH2Cl2 vs. CH3CN. These observations imply a different mechanistic
landscape taking place in CH3CN vs. CH2Cl2.
2.3. EPR spectroscopy
Electron paramagnetic resonance (EPR) spectra were recorded
at liquid N2 (77 K) with a Bruker ER200D spectrometer as described
previously [16].
2.3.1. EPR sample preparation
All EPR samples were prepared in 5 mm internal diameter quartz
tubes by Willmad Glass. 5 ␮l of either 80 mM or 20 mM H2O2 aque-
ous stock solution were added in 95 ␮l of 0.2 mM DPEIFeIIICl or
HFEIFeIIICl solution in CH3CN, resulting in [catalyst:H2O2] = [1:20]
and [1:5], respectively. To study the effect of the substrate, 50 ␮l
of 2 M of substrate stock in CH3CN were added in 45 ␮l of 0.45 mM
DPEIFeIIICl or HFEIFeIIICl solution, incubated for 10 min, followed
by addition of 5 ␮l of 80 mM H2O2 aqueous stock solution, at
[catalyst:H2O2:substrate] = [1:20:1000].
2.3.2. EPR spectroscopic investigation of the catalyst
1 ␮M catalyst was added to EPR tube, the spectrum of the pow-
der was recorded, then 100 ␮l CH3CN was added and the spectrum
of the solution was recorded. Then 22 ␮l of H2O2 was added in the
tube and the spectrum of the reaction mixture was recorded at var-
ious reaction times after addition. In another set of measurements,
substrate (cyclohexene) was added in a similar reaction mixture
and the time-evolution of the EPR spectra were recorded. After
each addition, the sample was immediately frozen at 77 K within
5–10 s. The EPR spectrum of this sample is referred as “t0”. To record
the time-evolution of the EPR spectra, the sample was thawed at
room temperature and allowed to evolve for a predetermined time
interval, followed by rapid quenching at 77 K within 5 s.
2.3.3. Spin trapping with TEMPO
When acetone was used as solvent, the total yield of the oxida-
tion reaction was decreased dramatically, Fig. 1A and B. This could
be attributed to the strong polarity of acetone and its ability to
react with radical intermediates e.g. expected to be formed during
the catalytic reaction [33,45]. Oxidation of cyclohexene in tert-amyl
alcohol, catalyzed by DPEIFeIIICl and HFEIFeIIICl provided total oxi-
dant conversions of 79% and 86%, respectively. These values are
The formation of transient carbon based radicals was monitored
via their interaction with TEMPO (2,2,6,6-tetramethylpiperidine-
1-oxyl). This method was applied by Beller et al. [39] to study
radical intermediates generated by a homogeneous non-heme cat-
alyst [39]. In our experiments, 100 ␮l from the catalytic reaction
solution was mixed with 10 ␮l of TEMPO solution (1 mg TEMPO

380
G. Bilis et al. / Applied Catalysis A: General 470 (2014) 376–389
Fig. 1. H2O2 conversion for cyclohexene oxidation catalyzed by (A) DPEIFeIIICl and (B) HFEIFeIIICl in different solvents. Conditions: ratio of catalyst:H2O2:substrate = 1:20:1000
or 1:50:1000.
slightly lower vs. those when CH3CN used as solvent, 89% and 97%,
Fig. 1A and B.
Benzaldehyde which was detected as major product of trans-ˇ-
methyl-styrene, styrene and cis-stilbene oxidation, was produced
from substrate oxidative cleavage.
Cyclohexane catalytic oxidation resulted in the formation of
cyclohexanol and cyclohexanone with total yield 7.3–12.1% and
alcohol/ketone (A/K) ratio 1.60–1.75. It has been suggested that
a ratio A/K > 1 indicates that •OH radicals are formed by a metal-
based system which reacts directly with substrate to form the
corresponding alcohol [33,49].
Overall, the present data reveal significant solvent effects on the
catalytic performance of the studied non-heme Fe catalysts. Thus,
taking into account the superior performance in acetonitrile, all the
subsequent catalytic studies of the present work were performed in
CH3CN. For consistency, in all cases, olefin oxidation reactions have
been explored by using 30% H2O2 aqueous solution under inert (Ar)
atmosphere.
3.1.2. Mapping of catalytic products in CH3CN
3.1.3. Evidence for radical intermediates
Fig. 2 summarizes the catalytic performance and products for
various substrates in CH3CN. All catalytic data are fully detailed
in Table S1 of Supporting Information. Before we proceed to the
analysis of the results we notice that, according to Fig. 2, the homo-
geneous catalysts performed with faster kinetics e.g. accomplishing
catalysis within a couple of hours vs. 24 h required by the hetero-
geneous ones. This is in agreement with previous observations for
Mn-based and Cu-based heterogeneous catalysts on SiO2 [15,16].
More specifically, Fig. 2 shows that cis-cyclooctene as substrate
afforded a very clean reaction, producing cyclooctene oxide with
100% selectivity, with moderate yield however. Cyclooctene oxide
is known to be stable, since side-reactions of the epoxide group are
retarded by the steric hindrance of the ring, avoiding the epoxide
ring-opening [45].
In cyclohexene oxidation under inert (Ar) atmosphere, the
oxidant conversion was 88.5% and 96.9% by the homogeneous
DPEIFeIIICl and HFEIFeIIICl catalysts and 82.5% and 91.1% by
the heterogeneous DPEIFeIII@SiO2 and HFEIFeIII@SiO2 systems
respectively. Here, cyclohexene epoxide presents low-yield, while
ring-opening products have been not detected. However, for-
mation of 2-cyclohexenol and 2-cyclohexenone were strongly
favored (Fig. 2) probably via H-abstraction and stabilization of
long-living allylic radical species. The time-dependent oxida-
tion of cyclohexene catalyzed by HFEIFeIIICl or HFEIFeIII@SiO2 is
shown in Supporting Information, Figs. S1 and S2. The selectivity
2-cyclohexenol/2-cyclohexenone was close to 2:1 for the homoge-
neous catalysts and 1:2 for the heterogeneous systems.
To further investigate the type and role of radicals in the present
study, a carbon-radical trap CBrCl3 [50] and Ph2NH (an oxygen-
radical trap [51]) were used. Indeed these experiments provided
evidence for involvement of both C-based and O-based radicals.
Addition of 40 ␮mol of spin traps, CBrCl3 or Ph2NH to the reac-
tion mixture suppressed the formation of cyclohexene catalytic
oxidation products resulting in a drop of ca. 35–39% and 66–70%,
respectively (Table 1). These data indicate that for all the tested
DPEIFeIIICl, HFEIFeIIICl, DPEIFeIII@SiO2 and HFEIFeIII@SiO2 cata-
lysts a common radical path involving both C- and O-centered radicals
is implicated in the main catalytic mechanism. As we show in the
following by EPR spin-trapping experiments, the O-based radicals
could be identified as •OH and RO• radicals.
Overall the present data in Fig. 2, show that (i) the non-
heme-Fe catalysts in homogeneous and heterogeneous form,
are efficient for oxygen transfer in
a variety of substrates,
resulting in allylic oxidation products and epoxides. (ii) In the
majority of substrates tested herein, the main products were
allylic-oxidizing products, for both the homogeneous and the het-
erogeneous catalysts. This implies a similar mechanistic landscape
for both homogeneous and heterogeneous catalysts. (iii) Alkenes
with similar steric-character i.e., such as cyclohexene/1-methyl-
cyclohexene or styrene/trans-␤-methyl-styrene were oxidized
with quite different selectivity. This suggests that electronic effects
are also important in the catalytic path. (iv) Comparing the two
homogeneous-Fe-catalysts, it is observed that they provide com-
parable conversions; however, HFEIFeIIICl is more reactive since
the reactions studied were completed within t < 2 h, instead of 4 h
required for reactions catalyzed by DPEIFeIIICl. This shows that
the ligand peripheral groups play a key-role in the catalytic evo-
lution. This – often overlooked – issue has been highlighted by
Collins [7].
Both epoxidation and allylic-oxidation were more efficient for
methyl-substituted substrates i.e. 1-methyl-cyclohexene, trans-
␤-methyl-styrene, limonene, than for non-methyl-substituted
substrates, Fig. 2. This can be attributed to the electron donating
substituents on olefin double bond, which promote the oxi-
dation reactions by involvement of electrophilic reagents [48].

G. Bilis et al. / Applied Catalysis A: General 470 (2014) 376–389
381
Fig. 2. H2O2 conversion in oxidations catalyzed by DPEIFeIIICl (ꢀ), DPEIFeIII@SiO2 (᭹), HFEIFeIIICl (ꢀ) and HFEIFeIII@SiO2 (ꢁ). Conditions: ratio of
catalyst:H2O2:substrate = 1:20:1000 or 1:50:1000. Reactions were complete within 2 h for HFEIFeIIICl (ꢀ), within 4 h for DPEIFeIIICl (ꢀ), and within 24 h for the heterogenized
catalysts.
3.1.4. Comparison of heterogeneous vs. homogeneous catalysts
III
HFEI-Fe Cl
III
50
40
30
20
10
0
The observed efficiency of the present heterogeneous
DPEIFeIII@SiO2 and HFEIFeIII@SiO2 catalysts indicates that the
heterogenization process preserves the catalytic properties of the
active-iron-centers for hydrocarbon oxidation. Importantly, we
notice that the heterogenized DPEIFeIII@SiO2 and HFEIFeIII@SiO2
catalysts show only slightly lower catalytic activity compared
with the corresponding homogeneous DPEIFeIIICl and HFEIFeIIICl.
This is non trivial, since usually heterogensation often results
in lower catalytic activity [8]. A reason for the lower catalytic
activity can be anticipated by noticing that bulky substrates e.g.
such as cis-stilbene, accentuate this effect. This supports the idea
that SiO2 might prevent, to some extent, the substrate’s approach
to the catalyst active center due to steric hindrance. We should
notice, that the present systems are among the first experimental
evidence of heterogeneous Fe-non-heme catalysts showing com-
parable catalytic activity to the corresponding homogeneous ones.
To our knowledge, the first example of a functional heterogeneous
non-heme Fe catalyst was published only recently, in 2012, by
de Faria et al. [52], where it was reported that FeIII-picolinate
and -dipicolinate complexes immobilized on kaolinite can pro-
mote cyclooctene and cyclohexane oxidation [52]. However
this required excess of H2O2 [52]. The present heterogeneous
non-heme catalysts is the first example of heterogeneous active
Fe-non-heme catalysts that operate efficiently e.g. with an activity
comparable to that of the homogneous ones, under mild H2O2
conditions in ambient temperature. As we show in the following,
this is due to facile redox advancement of Fe under mild redox
potentials e.g. 345–395 mV.
HFEI-Fe
@SiO
2
III
DPEI-Fe Cl
III
DPEI-Fe
@SiO
2
0
20
40
60
80
100
120
μmol H O
2 2
Fig. 3. Cyclohexene oxidation upon progressive addition of H2O2 catalyzed by
HFEIFeIIICl ( ), DPEIFeIIICl ( ), DPEIFeIII@SiO2 ) and HFEIFeIII@SiO2
) catalysts.
(
(
According to Fig. 3, the first addition of 20 equiv. of H2O2,
resulted in 19.4 Turn Over Numbers (TON) by HFEIFeIIICl. Addi-
tion of a second dose of 20 equiv. of H2O2 provided 22.2 TONs.
Third addition of oxidant promoted slightly the reaction achiev-
ing 25.3TONs. This shows that the evolution of the catalytic reaction
catalyzed by the HFEIFeIIICl had practically stopped after the first addi-
tion of oxidant. Comparing the two homogeneous-Fe-catalysts, it is
concluded that HFEIFeIIICl was more reactive during the 1st cat-
alytic run, however, DPEIFeIIICl appeared to be more stable against
oxidative destruction.
3.1.5. Catalyst stability
Fig. 3 compares the evolution of catalytic reaction by the
homogeneous HFEIFeIIICl and the heterogeneous HFEIFeIII@SiO2
in CH3CN, with cyclohexene as substrate. Each point in Fig. 3
represents a sequential addition of 20 ␮mol H2O2, after comple-
tion of the reaction, i.e. every 2 h for homogeneous HFEIFeIIICl
and every 24 h for heterogeneous HFEIFeIII@SiO2. The initial molar
ratio [catalyst:H2O2:cyclohexene] was 1:20:1000. Our previous
data [25] for DPEIFeIIICl catalyst are also included for comparison
in Fig. 3.
In the oxidation reaction catalyzed by the HFEIFeIII@SiO2, the
first 20 equiv. of H2O2 resulted in 18.2 TONs (Fig. 3). Additional
H2O2 caused an almost linear increase of TONs, up to the third
addition where 34.4 TONs were achieved. Forth and fifth H2O2 addi-
tion provided 39.0 and 41.8 TONs, respectively. The corresponding
data for the heterogeneous DPEIFeIII@SiO2 catalyst were 16.5 TONs
(1st H2O2 addition); 26.0 TONs (2nd H2O2 addition; 48.0 5th H2O2
addition) [25].

382
G. Bilis et al. / Applied Catalysis A: General 470 (2014) 376–389
Fig. 4. Recyclability of DPEIFeIII@SiO2 (A) and HFEIFeIII@SiO2 (B) for cyclohexene oxidation with H2O2. Conditions: ratio of catalyst:H2O2:substrate = 1:20:1000.
1,75
1,50
1,25
1,00
0,75
0,50
0,25
Overall the data in Fig. 3, reveal two important trends (i) the
homogeneous catalysts, despite their faster kinetics, were inferior
in terms of oxidative stability compared with the heterogenized
ones. (ii) Among the two homogeneous catalysts, DPEIFeIIICl had
higher oxidative stability than HFEIFeIIICl.
3.1.6. Catalyst recyclability
The recyclability of the heterogeneous DPEIFeIII@SiO2 and
DPEIFeIII@SiO2 catalysts was examined by repeated recovering and
reusing of the solid catalyst for oxidative catalysis of cyclohex-
ene. A [catalyst:H2O2:substrate] = [1:20:1000] molar ratio was used
in CH3CN. After each catalytic run, the catalyst was recovered by
centrifugation washed with CH3CN 5–10 times and dried under
vacuum (80 ◦C, 2 h). Subsequently, the recovered solid catalyst was
reused for a new oxidation reaction under the same catalytic con-
ditions.
According to Fig. 4, DPEIFeIII@SiO2 presented excellent reusabil-
ity i.e. <18% loss of the catalytic activity and selectivity after 5
runs. No catalyst leaching or degradation has been detected by
Flame Atomic Absorption Spectrocopy and GC–MS. HFEIFeIII@SiO2
retained its high reactivity for two runs, and performed with 50% of
its maximum activity in the 3rd run. Detection of catalyst molecu-
lar fragments by GC–MS during the 4th run and changes recorded
in DRS spectrum of the recovered catalyst indicate degradation of
catalyst active site. Thus it appears that HFEIFeIII@SiO2 e.g. char-
acterized by a higher reactivity, suffers by faster degradation/lower
stability than the DPEIFeIII@SiO2.
400
500
600
700
800
wavelength (nm)
Fig. 5. DPEIFeIIICl titration with H2O2 followed by UV–vis in the absence of substrate
and in the presence of cyclohexene as substrate (inset figure).
underlying mechanisms we pursued a detailed spectroscopic study
of the reactions.
3.2. Mechanistic studies
3.2.1. UV–vis spectroscopy
Overall the present data on the stability of the catalysts and
efficiency can be summarized as follows
3.2.1.1. Titration of DPEIFeIIICl with H2O2. Slow addition of H2O2
(30% aqueous H2O2 solution: CH3CN, v/v = 1:50) into a catalyst
solution (1 ␮mol DPEIFeIIICl in 2.898 ml of CH3CN) at room tem-
perature, in the absence of substrate, caused a progressive increase
and blue-shift of the bands at 407 and 504 nm (attributed to Ligand
Metal Charge Transfer, LMCT), see Fig. 5. That is, by addition of a
first dose of H2O2 (4 ␮mol), an increase of intensity of these bands
was observed, which continued even after addition of 18 ␮mol of
H2O2 (upward arrows, in Fig. 5). Such behavior is attributed to a
[catalyststability] : DPEIFeIII@SiO2 > HFEIFeIII@SiO2
> DPEIFeIII > HFEIFeIII
(1)
[catalystefficiency] : HFEIFeIII > DPEIFeIII
> HFEIFeIII@SiO2 > DPEIFeIII@SiO2
rapid interaction of Fe-center with H2O2 forming probably a FeIII
-
(2)
OOH species [29,39,53]. Moreover, the observed spectral changes
are consistent with formation of a low-spin LFeIII-OOH intermedi-
ate [39,54,55], an observation directly proven by EPR, as discussed
below.
Upon addition of excess of H2O2 (200 ␮mol, non diluted), the
intensities of the bands at 407 and 504 nm were decreased and
were not detectable after 5 min they (down arrows in Fig. 5). This
is attributed to the catalyst oxidative destruction in the presence
of large excess of oxidant [26].
Subsequently, another set of titration has been monitored by
UV–vis, Fig. 5 (inset), in the presence of substrate, i.e. cyclohexene
(1000 ␮mol). By addition of a first amount of H2O2 (4 ␮mol), the
intensity of the bands at 407 and 504 nm were increased. However,
The present data reveal two main pieces of information (i) in
non-heme Fe catalysts, oxidative destruction of the catalyst i.e. not
only yield, is an important factor which may determine the eli-
gibility of a system as a promising catalyst. This is in agreement
with previous postulations by Collins and England et al. [7,23]. (ii)
Here we provide the first experimental evidence that the hetero-
geneous Fe-catalysts are far more stable than the corresponding
homogeneous ones. Although a full understanding of this benefi-
cial effect requires further research from the opposing treads (1)
and (2), we consider that SiO2 support protects the active cata-
lyst centers against oxidative destruction. To further peer into the

G. Bilis et al. / Applied Catalysis A: General 470 (2014) 376–389
383
Fig. 6. HFEIFeIIIC titration with H2O2 followed by UV–vis in the absence of substrate
and in the presence of cyclohexene as substrate (inset figure).
when a second amount of 4 ␮mol of H2O2 was added, these bands
were shifted to 397 and 472 nm, respectively; then, upon addition
of 10 ␮mol H2O2, they further deceased (Fig. 5 inset). These spectral
changes observed in the presence of substrate, can be attributed to
transformation of the Fe-OOH species [39,55], due to catalytic evo-
lution e.g. substrate molecules are oxidized by the Fe-OOH species
[39,55]. It is underlined that when analogous studies have been
done in CH2Cl2 or acetone – instead of CH3CN – no-spectroscopic
evidence has been recorded for a Fe-OOH intermediate (data not
shown). This corroborates the observation that the solvent type
e.g. CH3CN, is important for the promotion of the reactive catalytic
intermediates [46].
3.2.1.2. Titration of HFEIFeIIICl with H2O2. Upon, slow, addition of
dilute H2O2 (30% aqueous H2O2 solution: CH3CN, v/v = 1:50) into
a solution of HFEIFeIIICl catalyst (1 ␮mol HFEIFeIIICl in 2.898 ml
of CH3CN), in the absence of substrate, two bands were gener-
ated at 506 and 468 nm (Fig. 6, upwards black arrows) which are
characteristic for Fe-OOH intermediates [39,56]. We notice that in
Fig. 6 the development of these bands was much faster than for
the case of DPEIFeIIICl in Fig. 5. Further additions of H2O2 caused
rapid decrease of the intensity of these bands until they disap-
peared completely (downwards red arrows), Fig. 6. In the presence
of cyclohexene, a similar titration profile was observed followed by
a small shift of ∼8 nm of the bands (Fig. 6 inset).
The rapid evolution of these UV–vis bands i.e. more rapid in the
case of HFEIFeIIICl vs. DPEIFeIIICl, is in agreement with our cat-
alytic data which showed that HFEIFeIIICl is more reactive than
DPEIFeIIICl, e.g. as summarized in (2). This is clearly exemplified
by the data for cyclohexene oxidation showing a 97% total oxi-
dation yield within 2 h when catalyzed by HFEIFeIIICl, vs. 88% by
DPEIFeIIICl within double reaction time e.g. 4 h.
3.2.2. EPR spectroscopy
3.2.2.1. HFEIFeIIICl catalyst. Fig. 7(a) presents the powder EPR spec-
trum of the homogenous HFEIFeIIICl system e.g. with no addition
of solvent. The EPR spectrum in Fig. 7(a) is characterized by a weak
peak at g = 9.2 and a derivative-shaped signal with g values g1 = 4.6,
g2 = 4.2, g3 = 3.8, listed in Table 2. These g values are typical for
high-spin FeIII (S = 5/2) (herein named High SpinI, HSI) complexes
in rhombic environment [25,57].
3.2.2.2. Analysis of the EPR spectra: high-spin FeIII (S = 5/2). The EPR
spectrum of an S = 5/2 system can be described by the spin Hamil-
tonian:
H = geˇ BS + D[Sz2 − (35/12) + (E/D)(Sx2 − Sy2)]
ˆ
(3)

384
G. Bilis et al. / Applied Catalysis A: General 470 (2014) 376–389
I
H.S. (S=5/2)[LFe3+]
(I)
g = 4.6, 4.2, 3.8
High Spin
1,0
0,8
0,6
0,4
0,2
0,0
(a) Powder
Low Spin
H.S. (S=5/2)[LFe3+-CH3CN]
II
g = 6.22 5.66
I
H.S.
High
(II)
g=2.02, 1.96 1.86
Spin
(b) CH CN
3
L.S.(S=1/2)[LFe3+-CH CN+H O ]
exp.
sim.
3
2
2
0
20
40
60
80
(c) CH CN/H O
3
2
2
time after H2O2 addition (min)
0
1000 2000 3000 4000 5000 6000
Fig. 8. Time dependence of the EPR signal intensities for High Spin Fe3+(I)
Spin Fe3+(II) ), and Low Spin Fe3+(ꢂ) for the homogeneous HFEIFeIII in CH3CN in
the presence of H2O2 and cyclohexene. Lines connecting the points are guide-to-
the-eye. The vertical bars mark the relative error. Experimental conditions: 0.2 mM
HFEIFeIIICl in CH3CN in the presence of 4 mM H2O2.
(
), High
Magneꢀc Field (G)
(
Fig. 7. EPR spectra of HFEIFeIII complex (a) powder, (b) in CH3CN, (c) in CH3CN
plus 4 mM H2O2. Experimental conditions: Modulation amplitude 10G, microwave
frequency 9.51 GHz, microwave power 32mW, modulation frequency 100 KHz,
temperature 77 K. Solid lines: experimental spectra. Dotted, dashed lines: simulated
theoretical spectra. In (b) the contributions from the two types of High-Spin Fe3+
sub-spectra are shown by the faint-dotted line.
new peaks at g1 = 6.22, g2 = 5.66 appeared, Fig. 7(b). EPR signals of
this type have been detected for mononuclear FeIII centers in vari-
ous non-heme Fe-metalloproteins [60 and references therein]. For
example analogous signals have been reported for the non-heme
FeIII(S = 5/2) of Photosystem-II reaction centers [61]. Using the Spin
Hamiltonian (1) these signals can be simulated (herein named HSII,
dotted line in Fig. 7(b)), for S = 5/2 with E/D = 0.03, see Table 2.
Accordingly, both peaks at g1 = 6.22, g2 = 5.66 can be assigned to one
monocuclear high-spin FeIII(S = 5/2) center with quasiaxial ligand-
field symmetry. At prolonged incubation of HFEIFeIIICl in CH3CN
[t > 15 min] all the g 4 e.g. HS(I) has been converted to the HS(II)
form.
where D and E are the axial and rhombic zero-field splitting param-
eters, respectively [58]. The ratio E/D is a measure of the departure
of the electronic environment of the paramagnetic center from
axial symmetry. In the proper axis system, an axial symmetry
corresponds to E/D = 0 while a maximally rhombic symmetry corre-
sponds to E/D = 1/3 [58]. For geˇB ꢁ D e.g. D ꢂ 0.3 cm−1 at X-band,
the magnetic energy levels – in the absence of a magnetic field – are
non-degenerate i.e. there is a zero-field splitting, therefore for an
S = 5/2 they comprise of three Kramers doublets. In second-order
perturbation theory, the energy level separations are determined
from the value of E/D, which in turn can be determined directly
from the experimental g values. For E/D = 0 the relative energies
of the three doublets are 0, 2D, and 6D, while for E/D = 1/3 they
become equidistant, i.e. their relative energies are 0D, 3.75D, and
7.5D, respectively [59,60 and references therein]. A numerical sim-
ulation of the HSI spectrum calculated using the spin Hamiltonian
(1) is included in Fig. 7(a) (dotted line) indicating that in powder the
HSI (FeIII) complex is characterized by a rhombicity ratio E/D = 0.28,
see listed in Table 2.
Thus the present EPR data show that upon dilution in CH3CN,
the ligand-field symmetry of the HFEIFeIIICl complex was severely
modified from rhombic (E/D = 0.28) to
a
quasiaxial HFEIFeIII
(E/D = 0.03). This shows that CH3CN was coordinated on the
FeIII(S = 5/2) center, without modification of the spin-state of the
iron. Coordination of CH3CN on non-heme FeIII(S = 5/2) centers has
been reported for other analogous systems [32,62]. An X-ray crysta-
lographic structure reported by Que et al. [32] showed that CH3CN
can be coordinated via the N atom to analogous non-heme FeIII cen-
ters. In our case, upon CH3CN coordination, the iron did not change
spin state i.e. it remained in the high-spin state, thus in HFEIFeIII
CH3CN coordination was not-strong enough to cause a transition from
the high- to the low-spin state of the FeIII(S = 5/2) center. As we show
in the following, conversion to the low-spin state is achieved only
after addition of H2O2 and CH3CN.
3.2.2.3. Effect of the solvent CH3CN. When HFEIFeIIICl was dis-
solved in CH3CN, characteristic changes in the EPR spectrum were
detected. The signal at g 4 was diminished by almost 60% and two
Table 2
EPR parameters, effective g-values and ligand filed symmetry ratio E/D for the FeIII
species detected for the homogeneous HFEIFeIIICl catalyst.
3.2.2.4. Effect of H2O2. After addition of H2O2 in HFEIFeIII/CH3CN a
sharp EPR signal corresponding to a low-spin (S = 1/2) is developed
(see LS in Fig. 7(c)) at the expense of the High Spin FeIII signals.
This signal was maximized after 40 min. A detailed kinetics vs.
time for all Fe-EPR signals is presented in Fig. 8. The LS signal
can be simulated, see dotted line in for S = 1/2, using principal
g-values g1 = 2.02, g2 = 1.96, g3 = 1.86 listed in Table 2. Similar
EPR spectra of Low-Spin non-heme ferric complex with rhombic
character have been observed previously [63]. Pertinent to the
present work is the data of Beller et al. [39] where CH3CN had
Sample
g1
g2
g3
E/D
High-Spin FeIII(S = 5/2)
HSI [HFEIFeIII (S = 5/2)] powder
HSII [HFEIFeIII (S = 5/2)]-CH3CN
4.63
6.22
4.21
5.66
3.80
2.01
0.28
0.03
Sample
g1
g2
g3
ꢂ/ꢁ
V/ꢂ
Low-Spin FeIII(S = 1/2)
[HFEIFeIII (S = 1/2)]-CH3CN + H2O2
[DPEIFeIII (S = 1/2)]-CH3CN
2.02
2.052
1.96
2.005
1.86
1.80
22.2
6.2
0.6
1.3

G. Bilis et al. / Applied Catalysis A: General 470 (2014) 376–389
385
converted an initially High-Spin Fe complex to the Low-Spin state
with similar EPR characteristics. Similarly, our previous data for
DPEIFeIII complex [25] showed that CH3CN can convert the –
initially High-Spin – FeIII complex to the Low-Spin (S = 1/2) state
[25]. The LS-FeIII EPR spectrum reported previously [25,39] bears
strong similarities with the Low-Spin signal that we report here
for HFEIFeIII in CH3CN after addition of H2O2. Based on the g
values of the low-spin HFEIFeIII species the rhombicity (V/ꢂ)
parameter is estimated to be V/ꢂ = 22.2. The tetragonality (ꢂ/ꢁ)
parameter – which depends primarily on the detailed balance of
the charge of the iron atom [25] – is ꢂ/ꢁ = 0.6, listed in Table 2.
For comparison, in Table 2 we list the tetragonality (ꢂ/ꢁ) and
rhombicity (V/ꢃ) parameters for the DPEIFeIII complex studied
previously [25].
Overall the present EPR data reveal that
(a) in powder, the HFEIFeIII complex is in a HSI FeIII(S = 5/2) state
with rhombicity E/D = 0.28, then addition of CH3CN, as solvent,
modifies severely the ligand field of the high-spin HFEIFeIII cen-
ter, toward an axial E/D = 0.03. The iron remains in the High Spin
state. In contrast, in DPEIFeIII complex CH3CN convert the iron
to the Low-Spin state [25].
Fig. 9. Kinetics of the TEMPO EPR signal in catalytic mixture [CH3CN, H2O2 and
cyclohexene]: ( ) in the presence of homogeneous HFEIFeIII. ( ) heterogeneous
HFEIFeIII.SiO2. (♦) control data for TEMPO in the presence of HFEIFeIII only. EPR
conditions: Modulation amplitude 4G, microwave frequency 9.49 GHz, microwave
power 1.25 mW, modulation frequency 100 kHz, temperature 77 K.
(b) only when H2O2 interacts with the HFEIFeIII complex in CH3CN
the iron is converted to its Low-Spin state.
3.2.2.5. Time evolution of the HFEIFeIII spin sates. The time evolution
of the EPR signals for the HSI to HSII and LS was recorded in CH3CN
in the presence of H2O2. The time evolution of the intensity of the
signals in Fig. 8 shows that under the conditions of our experiments:
coordination of both CH3CN and H2O2 to the Low Spin HFEIFeIII
complex is not evident. Detailed computational studies by our
group are in progress to further probe this issue.
(a) the HSI state of the HFEIFeIII decays monotonically with a half-
life t1/2 = 10–15 min, toward the HSII and then to the LS sates.
(b) the HSII state forms as an intermediate, maximized within
10–15 min and decays toward the LS state.
3.2.3. Detection of radicals intermediates by EPR
3.2.3.1. Spin trapping with TEMPO during catalytic reaction. The role
of radical intermediates in the catalytic reaction was probed by
first following the method suggested by Beller et al. [39] e.g. trace
amount of TEMPO was added to the standard catalytic reaction mix-
ture and the EPR spectrum was recorded after freeze-quenching
at 77 K. Before addition of the oxidant e.g. in the presence of the
HFEIFeIII complex plus cyclohexene in CH3CN, the typical EPR
signal of frozen TEMPO was detected, see inset in Fig. 9. Com-
parable EPR spectra were recorded for TEMPO in the presence of
the homogenous or the heterogeneous HFEIFeIII@SiO2 catalyst, see
Fig. 9. In the absence of oxidant, the TEMPO EPR signal remained
unaltered – in line shape and intensity – for at least six hours (see
(♦) in Fig. 9).
(c) the LS state is maximized after 40 min, at the expense of the HSI
and the HSII state. Importantly, there is a broad time-window
e.g. from t = 20 to 40 min where all three-spin states coexist in
the same sample. This situation is depicted in the EPR spectrum
presented in Fig. 7b.
(d) after 60 min the LS state decays, with no appearance of any new
signal.
3.2.2.6. Structural implications. In the case of DPEIFeIII [25] as
well as in the complex reported by Beller et al. [39] a Low-Spin
FeIII(S = 1/2) state was formed rapidly within <5 min [25] in 100%
of the centers in the presence of CH3CN. In contrast, in the case
of HFEIFeIII – presented herein – the Low Spin FeIII(S = 1/2) state is
formed only in the presence of CH3CN and H2O2. Low-Spin state
formation requires coordination of strong ligands. Typical example
is CN− which can convert a High-Spin Iron to Low-Spin FeIII(S = 1/2)
[59,64].
Taking into account all this information, the present data for
HFEIFeIII can be interpreted as follows: (a) addition of CH3CN, as
solvent, modifies severely the ligand field of the HFEIFeIII center,
toward an axial E/D = 0.03. The Fe remains in the High Spin state.
This can be attributed to coordination of one CH3CN molecule,
which distorts the ligand field of the FeIII, however CH3CN ligation
is not strong enough to convert the HFEIFeIII to its Low Spin state.
In contrast, in the case of DPEIFeIII [25], as well as in the complexes
reported by Beller [39], or Que et al. [32,62] CH3CN was able to
convert these Fe-complexes to Low spin state. (b) When H2O2
interacts with the HFEIFeIII complex in CH3CN the iron is converted
to its low spin state. H2O2 in other solvents e.g. CH3COCH3, CH2Cl2
did not result in LS formation (data not shown). This shows that a
synergistic effect of H2O2 and CH3CN results in formation of Low
Spin of the HFEIFeIII complex. Whether this is due to concomitant
In the presence of the homogeneous HFEIFeIII catalyst, the
TEMPO signal decayed rapidly within 2 min after addition of H2O2,
see solid blue squares and zoomed-inset in Fig. 9. In the presence of
the heterogeneous HFEIFeIII@SiO2 catalyst, the kinetics of TEMPO
disappearance was much slower, see solid red circles in Fig. 9. We
underline that TEMPO disappearance occurred only under catalytic
conditions e.g. in the presence of [Fe-catalyst/H2O2/cycloehexene].
In the presence of HFEIFeIII catalyst with no oxidant added, the sig-
nal intensity did not change, see (♦) in Fig. 9. This demonstrates that
TEMPO disappearance is not related to an eventual side-reaction
with Fe-complex alone, or H2O2 but rather it reacts with a catalytic
intermediate.
The observed TEMPO signal decay can be attributed to the inter-
action of TEMPO with transient [C-based] radicals [39] formed
during the catalytic reaction, which could be trapped by the TEMPO
at very high rate. This finding is in agreement with the observed
inhibitory effect of the C-carbon radicals quencher (CBrCl3, see
Table 1) and provide strong evidence for production of C-based
radicals in the catalytic path.
According to the catalytic-kinetic data for the heterogeneous
HFEIFeIII@SiO2, the catalysis was completed within 15 h. Thus, in

386
G. Bilis et al. / Applied Catalysis A: General 470 (2014) 376–389
DMPO/ROH
(A)
100
(B)
DMPO/OH
80
*
t=40min
*
*
*
60
DMPO/OH
40
t=10min
20
0
DMPO/ROH
0
5
10
15
20
25
30
35
40
45
3340 3360 3380 3400 3420
Reacꢀon ꢀme (min)
magneꢀc Field (G)
Fig. 10. (A) EPR signals of the adduct formed in the presence of DMPO. (B) Kinetics of the adduct formation (ꢁ) DMPO/OH.
Table 3
herein, the EPR data reinforces the existence of a common catalytic
path in the cases of the homogeneous and the heterogeneous non-
heme Fe catalysts studied herein.
Hyperfine splittings for the formed adducts.
ˇ
ꢄ
Adduct
aN
aH
aH
DMPO/OH•
DMPO-RO•
11.8
12.1
11.8
11.4
–
1.25
4. Discussion
The mechanistic landscape of non-heme Fe catalysts is complex.
Thus, before discussing our data, we briefly review the current main
mechanistic views from literature.
the heterogeneous system, the production of the C-based radicals
is expected to occur at a much slower rate i.e. resulting in a much
slower decay of the TEMPO radical as observed in Fig. 9. Overall, we
conclude that the catalytic mechanism for both the homogeneous
HFEIFeIII and heterogeneous HFEIFeIII@SiO2 catalyst proceeds via
steps involving C-based radicals in solution.
Formation of hydroperoxo-iron FeIII-OOH intermediate has
been proposed as the key-species that promotes the oxidative
catalytic cycle of non-heme Fe complexes [26]. FeIII-OOH can be
formed (i) either via deprotonation of H2O2 and direct ligation to
the Fe atom [27,28] or (ii) as a stepwise product promoted by the
reaction between FeIII-OH and H2O2 [29].
3.2.3.2. Spin trapping with DMPO during catalytic reaction. The EPR
spin trapping of superoxide and hydroxyl radicals by DMPO (5,5-
dimethyl-1-pyrroline N-oxide) is a widely used tool to study the
production of superoxide and hydroxide radicals in chemical, bio-
chemical and biological systems [40]. In our case the study of the
catalytic reaction in the presence of DMPO reveals the existence of
two types of O-based radicals (see Fig. 10 and Table 3).
The first signal is developed at early reaction times, maxi-
mized within 10 min, then rapidly decays within the next 10 min.
Based in its characteristic 1:2:2:1 quadruplet this is assigned to
a DMPO–hydroxyl (DMPO/OH•) adduct [40,65] (Fig. 11). A second
radical adduct is formed at late reaction times, attains its maximum
after 20 min. Its maximum intensity accounts for <20% of the OH
radicals. This adduct – formed in excess of oxidant – does not fur-
ther evolve but remains unaltered. Based on its lineshape, taking
into account the possible catalytic products and the excess of oxi-
dant we assign this to a DMPO-RO• adduct attributable to a C6H8O•
radical generated by H-abstraction of the previously formed cyclo-
hexenol. The formation of HO• radicals is in agreement with the
inhibitory effect of Ph2NH – an oxygen radical trap [51] – on the
catalytic reaction. Thus together with the catalytic data presented
In certain non-heme systems, FeIII-OOH intermediates may be
sluggish oxidants, so a further activation step is required to elicit
hydrocarbon oxidation [66]. High-spin (S = 5/2) FeIII-OOH is less
efficient in hydrocarbon oxidation than low-spin (S = 1/2) FeIII
-
OOH [26]. The high-spin iron center is able to react directly with
substrate or can be activated – through isomerization – from a
FeIII-␩1-OOH form to the [FeIII-␩2-OO]+ species. Thus high-spin
iron is found to be capable of deformylating aldehydes via nucleo-
philic reactions but not capable of oxygenating substrates via
electrophillic reactions [66]. However, a high-spin FeIII-OOH com-
plex and its proton-assisted conversion to an FeIV O complex has
been reported [67] and the electrophilic reactivity of a high-spin
FeIII-OOH species for bromide oxidation in MeOH has been pub-
lished recently [68].
Low-spin FeIII centers were reported to result in weakening of
the O O bond and concomitant strengthening of the Fe O bond
[30], suggesting that the low-spin iron (III) promotes cleavage of
peroxide O O bond [30]. Then, the low spin FeIII-OOH intermediate
can produce more reactive high-valent oxo-species [31,32] via two
paths: (i) heterolyticaly forming the FeV O and OH, (ii) homolyti-
caly forming FeIV O and •OH [27]. FeV O is efficient in hydrocarbon
epoxidation and dihydroxylation of olefins [34,69]. FeIV O com-
plexes can form high-spin S = 2 [69,70] or S = 1 depending on the
ligand structure [70,71].
This mechanistic landscape favors the existence of a second
labile coordination site of the metal, that is required for the forma-
tion of an intramolecular hydrogen bond to promote the O O bond
heterolysis, forming finally the high-valent HO-FeV O [72–75].
Thus, epoxidation occurs by oxo-attack on the olefin, forming the
first C O bond and an intermediate carbon-based radical, which
then forms the second C O bond [76,77]. On the other hand, olefin
-
O
O
N
CH
CH
OH
3
N
H
CH
CH
3
3
HO
3
DMPO/OH
DMPO
Fig. 11. Reaction for the spin trapping of a transient HO• radical with DMPO.

G. Bilis et al. / Applied Catalysis A: General 470 (2014) 376–389
387
Scheme 2. Suggested catalytic reaction mechanism for the studied non-heme Fe catalyst. The states marked my a circle have been detected by EPR spectroscopy while
products marked with a triangle have been identified analytically by GC–MS.
which subsequently react with the new [FeIII-OH]+2 giving the cor-
responding alcohol [47,81], analogous of the Rieske Dioxygenases
[82]. Recent studies propose the formation of the FeIV O direct
from a FeII center without involvement of a FeIII intermediate
[72,81]. In any case, the structure of the substrate can play impor-
tant role in determination of the iron containing intermediates by
the course of the C O bond formation [83].
cis-dihydroxylation is initiated by hydroxo attack to form the first
O bond and an intermediate carbon-based radical, followed by
C
reaction with the oxo-group to form the second C O bond [78].
The experimental detection of a non-heme Fe(V) intermediate
has been documented [69,79]. The •OH radical in aryl substrates
reacts with the substrate to give a substrate radical which can
then be trapped by the FeIV O moiety to give alcohol [32,33,80]
or undergo a radical chain autoxidation reaction with O2 to gener-
ate equimolar amounts of alcohol and ketones [32,33]. Therefore,
FeIV O intermediates are efficient under inner conditions.
4.1. Mechanistic aspects
Another, mechanistic hypothesis is that the FeIV O itself
attacks the aliphatic C H bond in a range of hydrocarbon sub-
strates, causing the C H cleavage, forming the substrate radical
Our data show that for the catalytic activity of both DPEIFeIIICl
and HFEIFeIIICl catalysts, CH3CN as solvent is important via its
ligation to the iron center (Scheme 2). Recently de Visser et al.

388
G. Bilis et al. / Applied Catalysis A: General 470 (2014) 376–389
in Scheme 2. This is the first evidence for equivalence of homo-
geneous and heterogeneous non-heme Fe catalysts. Despite the
similarity of the first coordination sphere in both DPEIFeIIICl and
HFEIFeIIICl, important differences are observed in the spin con-
figuration, reactivity and stability. Thus the peripheral units of
the ligand play a determinant role in the stability and reac-
tivity of the catalysts. This significance of ligand stability in
catalysis by metal-complexes has been discussed by Collins [7].
Here, we provide evidence that heterogenisation improves sig-
nificantly the catalytic robustness and recyclability. However, the
pathology, sensitivity observed for the homogeneous catalysts
i.e., DPEIFeIII < HFEIFeIII is also observed for the heterogeneous
DPEIFeIII@SiO2 < HFEIFeIII@SiO2. This reveals that the local effects
i.e., the nature of the immediate ligand moiety, are determinant
in the catalytic stability despite the beneficial effect of the SiO2
matrix.
[84] presented theoretical evidence that CH3CN axial ligand is
orders of magnitude more reactive than chloride axial ligand for the
catalytic activity of a non-heme FeIV O species for aliphatic over
aromatic hydroxylation. However, the role of CH3CN may be more
subtle than previously thought. The present data exemplify that:
(a) CH3CN transforms the DPEIFeIIICl complex in the Low Spin
state as previously reported and for other Fe-non-heme systems
[32,39,62] resulting in a Low Spin DPEIFeIII-NCCH3 intermediate,
while (b) only modifies the ligand field symmetry of HFEIFeIIICl
complex which remains in the High Spin state providing a High
Spin HFEIFeIII-NCCH3 species. Given that catalytic reactions by both
DPEIFeIIICl and HFEIFeIIICl systems occur only in CH3CN, the coor-
dination ability of this solvent seems to play a crucial role for
the catalyst activity. We believe that the ability of the coordina-
tive CH3CN to transform the High Spin state of the FeIII catalyst to
the Low Spin one was welcomed, but obviously it is not the only
determinant for the catalyst activity, since the High Spin HFEIFeIII
-
-
NCCH3 catalyst is even more reactive than the Low Spin DPEIFeIII
5. Conclusions
NCCH3 one. This finding is in accordance with Britovsek’s sugges-
tion [23] that Fe-catalysts which exist as kinetically labile High Spin
complexes at room temperature in CH3CN deactivate more rapidly
than those which exist as kinetically inert Low spin complexes [23].
Nevertheless, the addition of H2O2 unifies the mechanistic path
of both catalysts, since the observed UV–Vis and EPR spectroscopic
evidence suggest the formation of a Low Spin state intermediate,
i.e., a FeIII-OOH species (Scheme 2).
The homogeneous catalytic systems DPEIFeIIICl and HFEIFeIIICl
have been compared with the heterogeneous ones e.g. where the
catalysts have been covalently immobilized in SiO2. The homo-
geneous catalysts DPEIFeIIICl and HFEIFeIIICl despite their faster
kinetics were inferior in terms of oxidative stability vs. the hetero-
genized ones, DPEIFeIII@SiO2 and HFEIFeIII@SiO2. Among the two
homogeneous catalysts HFEIFeIII is more rapid but less resistant to
At this point, the mechanistic landscape of non-heme ligands
oxidative stability than DPEIFeIII
.
favors a homolytic cleavage of hydroperoxides to form FeIV
O
In cyclohexene oxidation, the highest conversion was achieved
in CH3CN. UV–vis and EPR data show that, in DPEIFeIII complex,
CH3CN converts the iron to the Low Spin state. Addition of H2O2
promotes the formation of a low-spin intermediate which is sug-
gested to be an (S = 1/2) DPEIFeIII-OOH species. In the case of
HFEIFeIII, the CH3CN, as solvent, modifies severely the ligand field
of the HSII HFEIFeIII center, toward an axial symmetry. However,
the iron remains in the High Spin state. When H2O2 interacts with
the HFEIFeIII complex in CH3CN the iron is converted to its Low Spin
state forming HFEIFeIII-OOH. When HFEIFeIIICl was used as cata-
lyst in cyclohexene oxidation in CH2Cl2, no oxidation products have
been detected. UV–vis and EPR data show, no evidence for Low Spin
Fe-intermediates occurred in CH2Cl2 for neither DPEIFeIIICl nor
HFEIFeIIICl. The coordinative ability of CH3CN and the formation
of the Low Spin LFeIII-OOH intermediate seem to be determinant
for the catalytic activity of the present non-heme iron systems.
The higher reactivity and concomitant more rapid deactivation of
HFEIFeIIICl in CH3CN is attributed to the kinetically labile High Spin
HFEIFeIII-NCCH3 intermediate, whereas DPEIFeIIICl forms a rather
kinetically inert Low Spin DPEIFeIII-NCCH3 species.
Based on the detected intermediate Fe-states and transient C-
based and O-based radicals, a complex reaction-path landscape
is evidenced. It appears that interplay between solvent/spin state
of Fe/oxidative stability is key-features which frame the effi-
ciency/eligibility of a given non-heme Fe catalyst.
Herein taking into account the detected catalytic oxidation
products, a consistent catalytic mechanism is proposed. A simi-
lar mechanistic landscape is suggested for both homogeneous and
heterogeneous catalysts.
species and •OH [33].
Herein, the major oxidation products of alkenes are alcohols and
ketones, mainly derived by an allylic oxidation reaction. However,
in some cases, considerable amounts of epoxides have been also
formed. Thus, the catalytic oxidation does not follow a unique mech-
anistic path, and the participation of more than one active oxidative
species is required. This need can be covered by FeIV O and •OH,
since both of them are able to attack olefinic substrates.
Thus, we suggest that firstly •OH radicals attack the substrate
by proton abstraction forming a transient allylic radical (Scheme 2,
path a) which, during catalysis, is trapped by FeIV O providing
alcohols (Scheme 2, path i) and ketones (Scheme 2, path ii) as
allylic oxidation products. Alternatively, when substrate undergoes
electrophilic attack by FeIV O species (Scheme 2, path b), a radi-
cal [Fe O-substrate] adduct is generated which subsequently, by
intramolecular attack, liberates epoxides as oxidation products. The
FeII formed during catalysis is re-oxidized by H2O2.
When styrene, trans-ˇ-methyl-styrene and cis-stilbene are
used as substrates, their oxidation provided the corresponding
epoxides and, additionally, benzaldehyde. Benzaldehyde is pro-
posed to be formed by nucleophilic attack on the epoxide ring
providing oxiranyl radicals and subsequently resulting in the C
ring opening [85].
C
Finally, the strong radical mechanism character suggested in
Scheme 2 is substantiated by the radical trap and spin trap data.
We observe that cyclohexene oxidation by our systems is sup-
pressed in the presence of either a C-radical trap or an O-radical
trap e.g. resulting in a drop of catalytic efficiency by 35–39% and
66–70% respectively. A key finding of the present work is the trends
described in (1) and (2). The more reactive catalyst is the less sta-
ble may be. These two factors seem to correlate with the high-spin,
low-spin configuration. High-spin Fe-catalysts may be more reac-
tive; however they are deactivated easier than the low-spin ones.
Another important finding of the present work is that homoge-
neous and heterogeneous catalysts provide analogous distribution
of oxidation products presenting similar oxidation chemistry.
Based on this, we believe that homogeneous and heterogeneous
catalysts react through the same mechanistic pathways as depicted
Acknowledgment
This research has been co-financed by the European Union
(European Social Fund – ESF) and Greek national funds through
the Operational Program “Education and Lifelong Learning” of the
National Strategic Reference Framework (NSRF) – Research Fund-
ing Program: THALIS. Investing in knowledge society through the
European Social Fund.

G. Bilis et al. / Applied Catalysis A: General 470 (2014) 376–389
389
Appendix A. Supplementary data
[41] S. Stoll, A. Schweiger, J. Magn. Reson. 178 (2006) 42–55.
[42] G.J. ven Brink, B.C.M. Fernandez, M.C.A. van Vliet, I.W.C.E. Arends, R.A. Sheldon,
J. Chem. Soc. Perkin Trans. 1 (2001) 224–228.
[43] S. Ted Oyama, Mechanisms in Homogeneous and Heterogeneous Epoxidation
Catalysis, Elsevier, 2009.
[44] J.U. Rohde, L. Que, Angew. Chem. Int. Ed. 44 (2005) 2255–2258.
[45] M. Klopstra, G. Roelfs, R. Hage, R.M. Kellog, B.L. Feringa, Eur. J. Inorg. Chem.
(2004) 846–856.
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.apcata.
2013.11.009.
References
[46] C.V. Sastri, J.M. Park, T. Ohta, T.A. Jackson, A. Stubna, M.S. Seo, J. Lee, J.
Kim, T. Kitagawa, E. Münck, L. Que, W. Nam, J. Am. Chem. Soc. 127 (2005)
12494–12495.
[47] A.T. Fiedler, L. Que, Inorg. Chem. 48 (2009) 11038–11047.
[48] F.G. Gelalcha, B. Bitterlich, G. Anilkumar, M. Kin Tse, M. Beller, Angew. Chem.
Int. Ed. 46 (2007) 7293–7296.
[49] M. Costas, K. Chen, L. Que, Coord. Chem. Rev. 200-202 (2000) 517–544.
[50] L.M. Slaughter, J.P. Collman, T.A. Eberspacher, J.I. Brauman, Inorg. Chem. 43
(2004) 5198–5204.
[51] C. Di Nicola, Y. Yu Karabach, A.M. Kirillov, M. Monari, L. Pandolfo, C. Pettinari,
A.J.L. Pombeiro, Inorg. Chem. 46 (2007) 221–230.
[52] 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, Catal. Today 187 (2012) 135–149.
[53] G.C. Silva, G.L. Parrilha, N.M.F. Carvalho, V. Drago, C. Fernandes, A. Horn Jr.,
O.A.C. Antunes, Catal. Today 133–135 (2008) 684–688.
[54] M.R. Bukowski, S. Zhu, K.D. Koehntop, W.W. Brennessel, L. Que, J. Biol. Chem. 9
(2004) 39–48.
[55] M. Costas, M.P. Mehn, M.P. Jensen, L. Que, Chem. Rev. 104 (2004) 939–986.
[56] X. Shan, L. Que, J. Inorg. Biochem. 100 (2006) 421–433.
[57] J. Peisach, W.E. Blumberg, S. Ogawa, E.A. Rachmilewitz, R. Oltzik, J. Biol. Chem.
246 (1971) 3342–3343.
[58] J.R. Pilbrow, Transition Ion Electron Paramagnetic Resonance, Oxford Science
Publications, Clarendon Press, Oxford, England, 1991.
[1] B.S. Lane, K. Burgess, Chem. Rev. 103 (2003) 2457–2473.
[2] E.G. Kovaleva, J.D. Lipscomb, Nat. Chem. Biol. 4 (2008) 186–193.
[3] S.P. de Visser, D. Kumar, S. Shaik, J. Inorg. Biochem. 92 (2004) 1183–1193.
[4] P.R. Ortiz de Montellano, Cytochrome P-450: Structure, Mechanism, and Bio-
chemistry, second ed., Plenum Press, New York, 1995, pp. 3–40.
[5] E.P. Talsi, K.P. Bryliakov, Coord. Chem. Rev. 256 (2012) 1418–1434.
[6] L.G. Beauvais, S.J. Lippard, J. Am. Chem. Soc. 127 (2005) 7370–7378.
[7] T. Collins, Acc. Chem. Res. 27 (1994) 279–285.
[8] M.H. Valkenberg, W.F. Holderich, Catal. Rev. 44 (2002) 321–374.
[9] N.T.S. Phan, D.H. Brown, P. Styring, Green Chem. 6 (2004) 526–532.
[10] E.B. Mubofu, J.H. Clark, D.J. Macquarrie, Green Chem. 3 (2001) 23–25.
[11] O. Hemminger, A. Marteel, M.R. Mason, J.A. Davies, A.R. Tadd, M.A. Abraham,
Green Chem. 4 (2002) 507–512.
[12] J. Krammer, E. Nöllen, W. Buijs, W.L. Driessen, J. Reedijk, React. Funct. Polym.
57 (2003) 1–11.
[13] M.S.M. Moreira, P.R. Martins, R.B. Curi, O.R. Nascimento, Y. Iamamoto, J. Mol.
Catal. A: Chem. 233 (2005) 73–81.
[14] A.W. Van der Made, J.W.H. Smeets, R.J.M. Nolte, W. Drenth, J. Chem. Soc. Chem.
Commun. (1983) 1204–1206.
[15] D. Zois, Ch. Vartzouma, Y. Deligiannakis, N. Hadjiliadis, L. Casella, E. Monzani,
M. Louloudi, J. Mol. Catal. A: Chem. 261 (2007) 306–317.
[16] G. Grigoropoulou, K.C. Christoforidis, M. Louloudi, Y. Deligiannakis, Langmuir
23 (2007) 10407–10418.
[17] T.J. Terry, T.D.P. Stack, J. Am. Chem. Soc. 130 (2008) 4945–4953.
[18] R.I. Kureshy, I. Ahmad, N.H. Khan, S.H.R. Abdi, K. Pathak, R.V. Jasra, J. Catal. 238
(2006) 134–141.
[19] J. Kim, R.G. Harrison, C. Kim, L. Que Jr., J. Am. Chem. Soc. 118 (1996) 4373–4379.
[20] M.H. Lim, J.-U. Rohde, A. Stubna, M.R. Bukowski, M. Costas, R.Y.N. Ho, E. Munck,
W. Nam, L. Que Jr., Proc. Natl. Acad. Sci. U.S.A. 100 (2003) 3665–3670.
[21] G.J.P. Britovsek, J. England, A.J.P. White, Inorg. Chem. 44 (2005) 8125–8134.
[22] J. England, G.J.P. Britovsek, N. Rabadia, A.J.P. White, Inorg. Chem. 46 (2007)
3752–3767.
[23] J. England, C.R. Davies, M. Banaru, A.J.P. White, G.J.P. Britovsek, Adv. Synth. Catal.
3350 (2008) 883–897.
[24] C. Piovezan, K.A.D.F. de Castro, S.M. Drechsel, S. Nakagaki, Appl. Catal., A 293
(2005) 97–104.
[25] G. Bilis, K.C. Christoforidis, Y. Deligiannakis, M. Louloudi, Catal. Today 157
(2010) 101–106.
[26] K. Chen, M. Costas, J. Kim, A.K. Tipton, L. Que, J. Am. Chem. Soc. 124 (2002)
3026–3035.
[27] M. Fujita, L. Que, Adv. Synth. Catal. 346 (2004) 190–194.
[28] P.D. Oldenburg, L. Que, Catal. Today 117 (2006) 15–21.
[29] G. Roelfes, M. Lubben, K. Chen, R.Y.N. Ho, A. Meetsma, S. Genseberger, R.M.
Hermant, R. Hage, S.K. Mandal, V.G. Young, Y. Zang, H. Kooijman, A.L. Spek, L.
Que, B.L. Ferringa, Inorg. Chem. 38 (1999) 1929–1936.
[30] R.Y.N. Ho, G. Roelfes, B.L. Feringa, L. Que, J. Am. Chem. Soc. 121 (1999) 264–265.
[31] R. Mas-Balleste, M. Fujita, C. Hemmila, L. Que, J. Mol. Catal. A: Chem. 251 (2006)
49–53.
[59] Y. Deligiannakis, A. Boussac, H. Bottin, V. Perrier, O. Barzu, A.-M. Gilles, Bio-
chemistry 36 (1997) 9446–9452.
[60] B. Gaffney, H.J. Silverstone, in: L.J. Berliner, J. Reuben (Eds.), Biological Magnetic
Resonance, vol. 13, Plenum Press, New York, 1993, pp. 1–55.
[61] Y. Deligiannakis, V. Petrouleas, B.A. Diner, Biochim. Biophys. Acta 1188 (1994)
260–270.
[62] M. Costas, A.K. Tipton, K. Chen, D.-H. Jo, L. Que, J. Am. Chem. Soc. 123 (2001)
6722–6723.
[63] L. Duelund, H. Toftlund, Spectrochim. Acta A 56 (2000) 331–340.
[64] Y. Sanakis, V. Petrouleas, B.A. Diner, Biochemistry 33 (1994) 9922–9928.
[65] M. Davies, Biochim. Biophys. Acta 964 (1988) 28–35.
[66] M.J. Park, J. Lee, Y. Suh, J. Kim, W. Nam, J. Am. Chem. Soc. 128 (2006) 2630–
2634.
[67] F. Li, K.K. Meier, M.A. Cranswick, M. Chakrabarti, K.M. Van Heuvelen, E. Münck,
L. Que Jr., J. Am. Chem. Soc. 133 (2011) 7256–7259.
[68] A.K. Vardhaman, C.V. Sastri, D. Kumar, S.P. de Visser, Chem. Commun. 47 (2011)
11044–11046.
[69] O.Y. Lyakin, K.P. Bryliakov, E.P. Talsi, Inorg. Chem. 50 (2011) 5526–5538.
[70] J.U. Rolde, J.H. In, M.H. Lim, W.W. Brennessel, M.R. Bukowski, A. Stubna, E.
Münck, W. Nam, L. Que, Science 299 (2003) 1037–1039.
[71] L. Que, Acc. Chem. Res. 40 (2007) 493–500.
[72] F. Li, J. England, L. Que, J. Am. Chem. Soc. 132 (2010) 2134–2135.
[73] J. England, Y. Guo, E.R. Farquhar, V.G. Young, E. Münck, L. Que, J. Am. Chem. Soc.
132 (2010) 8635–8644.
[74] D.C. Lacy, R. Gupta, K.L. Stone, J. Greaves, J.W. Ziller, M.P. Hendrich, A.S. Borovik,
J. Am. Chem. Soc. 132 (2010) 12188–12190.
[75] Y. Feng, J. England, L. Que Jr., ACS Catal. 1 (2011) 1035–1042.
[76] R. Mas-Balleste, L. Que, J. Am. Chem. Soc. 129 (2007) 15964–15972.
[77] A. Bassan, M.R.A. Blomberg, P.E.M. Siegbahn, L. Que, Angew. Chem. Int. Ed. 44
(2005) 2939–2941.
[78] A. Bassan, M.R.A. Blomberg, P.E.M. Siegbahn, L. Que, J. Am. Chem. Soc. 124 (2002)
11056–11063.
[79] F.T. de Oliveira, A. Chanda, D. Banerjee, X.P. Shan, S. Mondal, L. Que, E.L. Bomi-
naar, E. Munk, T.J. Collins, Science 315 (2007) 835–838.
[80] A. Thibon, V. Jollet, C. Ribal, K. Senechal-David, L. Billon, A.B. Sorokin, F. Banse,
Chem. Eur. J. 18 (2012) 2715–2724.
[81] P. Comba, G. Rajaraman, H. Rohwer, Inorg. Chem. 46 (2007) 3826–3838.
[82] S. Chakrabaty, R.N. Austin, D. Deng, J.T. Groves, J. Lipscomb, J. Am. Chem. Soc.
129 (2007) 3514–3515.
[83] A. Company, L. Gómez, M. Güell, X. Ribas, J.M. Luis, L. Que, M. Costas, J. Am.
Chem. Soc. 129 (2007) 15766–15767.
[32] M.P. Jensen, S.J. Lange, M.P. Mehn, E.L. Que, L. Que, J. Am. Chem. Soc. 125 (2003)
2113–2128.
[33] G. Roelfes, M. Lubben, R. Hage, L. Que, B.L. Feringa, Chem. Eng. J. 6 (2000)
2152–2159.
[34] A. Company, Y. Feng, M. Güell, X. Ribas, J.M. Luis, L. Que, M. Costas, Chem. Eur.
J. 15 (2009) 3359–3362.
[35] C. Vartzouma, E. Evaggelou, Y. Sanakis, N. Hadjiliadis, M. Louloudi, J. Mol. Catal.
A: Chem. 263 (2007) 77–85.
[36] N.A. Vermeulen, M.S. Chen, M.C. White, Tetrahedron 65 (2009) 3078–3084.
[37] G. Bilis, M. Louloudi, Bioinorg. Chem. Appl. (2010), 861892.
[38] G.J.P. Britovsek, J. England, A.J.P. White, Dalton Trans. (2006) 1399–1408.
[39] F.G. Gelalcha, G. Anilkumar, M.K. Tse, A. Brücknerc, M. Beller, Chem. Eur. J. 14
(2008) 7687–7698.
[40] O.V. Zalomaeva, N. Truckhan, I. Ivanchicova, A.A. Panchenco, E. Roduner, E.P.
Talsi, A.B. Sorokin, V.A. Rogov, O.A. Kholdeeva, J. Mol. Catal. A: Chem. 277 (2007)
185–192.
[84] S.P. de Visser, R. Latifi, L. Tahsini, W. Nam, Chem.-Asian J 6 (2011) 493–504.
[85] P. Dowd, W. Zhang, Chem. Rev. 93 (1993) 2091–2115.