Comparative study of the catalytic thermodynamic barriers for two homologous Mn- and Fe-non-heme oxidation catalysts

Article abstract of DOI:10.1016/j.jcat.2016.06.017

Two sets of homologous Mn- and Fe-catalysts, [Mn^IILCl₂], [Fe^IILCl₂] and [Mn^IIL(OAc)₂], [Fe^IIL(OAc)₂] have been synthesized. A detailed comparative study of their catalytic oxidative performance with H₂O₂, in tandem with EPR and Low-Temperature UV–vis spectroscopies has been carried out. The [Metal-L(OAc)₂] and [Metal-LCl₂] catalysts did not show any difference in their catalytic behavior i.e. there is no effect of the labile ligands on the studied catalysis. It is found that the Mn-catalysts consistently outcompeted the homologous Fe-catalysts i.e. TOFs (Mn) = 162 vs. TOFs (Fe) = 16. We found that the Fe-catalyst faces a significantly higher activation barrier than the Mn-catalyst i.e. E_a(Fe^IIL(OAC)₂) = 91KJ/mol ? E_a(Mn^IIL(OAC)₂) = 55 kJ/mole, while the free-energy difference, ΔG(Fe^IIL(OAC)₂) ～ ΔG(Mn^IIL(OAC)₂) ～ ?145 kJ/mole, did not make difference. Taken altogether the present data clarify that the main thermodynamic barrier, ultimately determining the overall catalytic performance, of these homologous Mn- and Fe-catalysts is the activation energy for the transient intermediates i.e. Mn^II to Mn^IV[Formula presented] for the Mn-catalysts and Fe^II to Fe^III[Formula presented] for the Fe-catalysts. A unified/consistent catalytic thermodymanic concept is discussed, that bears relevance to the catalytic behavior of many non-heme Mn- vs. Fe-oxidation catalysts.

Full text of DOI:10.1016/j.jcat.2016.06.017

Journal of Catalysis 341 (2016) 104–115
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Comparative study of the catalytic thermodynamic barriers for two
homologous Mn- and Fe-non-heme oxidation catalysts
M. Papastergiou , P. Stathi , E.R. Milaeva , Y. Deligiannakis , M. Louloudi a,
a
a
b
c,
⇑
⇑
a
Laboratory of Inorganic Chemistry, Department of Chemistry, University of Ioannina, 45110 Ioannina, Greece
Department of Medicinal Chemistry and Fine Organic Synthesis, Lomonosov Moscow State University 119991, Russia
Laboratory of Physical Chemistry of Materials & Environment, Department of Physics, University of Ioannina, 45110 Ioannina, Greece
b
c
a r t i c l e i n f o
a b s t r a c t
II
II
II
II
Article history:
2 2 2 2
Two sets of homologous Mn- and Fe-catalysts, [Mn LCl ], [Fe LCl ] and [Mn L(OAc) ], [Fe L(OAc) ] have
Received 12 May 2016
Revised 23 June 2016
Accepted 23 June 2016
been synthesized. A detailed comparative study of their catalytic oxidative performance with H
tandem with EPR and Low-Temperature UV–vis spectroscopies has been carried out. The [Metal-L
OAc) ] and [Metal-LCl ] catalysts did not show any difference in their catalytic behavior i.e. there is
2 2
O , in
(
2
2
no effect of the labile ligands on the studied catalysis. It is found that the Mn-catalysts consistently
outcompeted the homologous Fe-catalysts i.e. TOFs (Mn) = 162 vs. TOFs (Fe) = 16. We found that the
Keywords:
Catalytic epoxidation
Mn- and Fe-non-heme catalysts
H O activation
2 2
EPR
II
Fe-catalyst faces a significantly higher activation barrier than the Mn-catalyst i.e. E
a
(Fe L(OAC)
2
)
II
II
=
91KJ/mol ꢀ E
a
(Mn L(OAC)
2
) = 55 kJ/mole, while the free-energy difference,
D
G(Fe L(OAC)
2
) ꢁ
DG
II
(
Mn L(OAC)
2
) ꢁ ꢂ145 kJ/mole, did not make difference. Taken altogether the present data clarify that
the main thermodynamic barrier, ultimately determining the overall catalytic performance, of these
II
Thermodynamics
Low-T UV–vis
homologous Mn- and Fe-catalysts is the activation energy for the transient intermediates i.e. Mn to
IV
II
III
Mn @O for the Mn-catalysts and Fe to Fe AOOH for the Fe-catalysts. A unified/consistent catalytic
thermodymanic concept is discussed, that bears relevance to the catalytic behavior of many non-heme
Mn- vs. Fe-oxidation catalysts.
Ó 2016 Elsevier Inc. All rights reserved.
1
. Introduction
In fine chemicals technology, among the end-targets of applied
In general, the stability and selectivity of a homogeneous cata-
lyst are strongly related to its molecular structure. Consideration of
steric, electronic and conformational properties is necessary i.e. in
order to design appropriate ligands for metal complexes which will
serve as efficient catalysts. Among other properties, the ligands
must [i] be resilient to oxidation, and [ii] strongly electron donat-
ing, in order to achieve high oxidation states of the active metal [1].
In this context, polydentate ligands containing 2-pyridylmethyl or
2-pyridyl fragments attached to a tertiary nitrogen atom, contain
oxidation catalysis is to produce more efficient catalysts in tandem
with less expensive and more environmentally acceptable pro-
cesses [1]. Epoxidation reactions of olefins constitute an industri-
ally applicable process for production of epoxides i.e. widely
used as raw materials or intermediates for epoxy resin production,
paints, surfactants and medicines [2,3]. Environmental regulations
on industrial production of chemicals, force companies to diminish
environmental pollution [4–7]. Within this environmental/eco-
nomic context, the use of manganese and iron complexes as active
catalysts, associated with hydrogen peroxide as the primary oxi-
both
r-donor and p-acceptor binding sites. Thus, ligands of this
0
0
class (e.g. such as tris (2-pyridylmethyl) amine, tpa or N,N,N ,N -
tetrakis (2-pyridylmethyl) ethylenediamine, tpen) are capable of
stabilizing both high- and low-oxidation states in their metal
complexes [9–11]. Moreover, ligands containing 2,6-di-tert-butyl-
4-hydroxyphenol groups on the periphery of the ligand frame are
also interesting because they might provide antioxidant features
[12] and additional endurance to metal oxidation catalysts. 2,6-
di-tert-butyl-4-hydroxyphenol has been already used as sub-
stituent on the porphyrin ring offering remarkable performance
in oxidations catalyzed by metalloporphyrins [13–18]. Recently, a
new series of non-heme metal complexes with di-(2-picolyl)
amine ligand bearing this 2,6-di-tert-butyl-4-hydroxyphenol
2 2
dant, is highly desirable. H O is increasingly used as oxidant both
in industry and in academic research, since its only by-product is
water. In the same context, non-heme manganese and non-heme
iron catalysts present advantages e.g. convenient synthesis, low
production-cost, considerable stability, selectivity toward forma-
tion of epoxides [1,8].
⇑
Corresponding authors.
http://dx.doi.org/10.1016/j.jcat.2016.06.017
021-9517/Ó 2016 Elsevier Inc. All rights reserved.
0

M. Papastergiou et al. / Journal of Catalysis 341 (2016) 104–115
105
II
II
II
II
2 2 2 2
Fig. 1. Structures of [Mn L(OAc) ] (1), [Mn LCl ] (2), [Fe L(OAc) ] (3) and [Fe LCl ] (4) complexes.
moiety has been reported by one of us [19]. Herein, manganese and
iron complexes containing the [N-(3,5-Di-tert-butyl-4-hydroxyben
zyl)-N,N-di-(2-pyridylmethyl)]amine (L) ligand [19] (Fig. 1) have
been evaluated as homogeneous catalysts for alkene epoxidation
with hydrogen peroxide.
transient intermediates of the Fe-catalysts have higher E
the homologous Mn-catalysts, and as result Fe-catalysts have
lower catalytic performance than the Mn-catalysts.
a
than
2
. Materials and methods
Herein, we present a method where low-temperature UV–vis
spectroscopy was used in tandem with low-temperature Electron
Paramagnetic Resonance (EPR) to study the transient redox/cat-
2.1. Materials
II
alytic intermediates formed during the catalytic cycle of the [Mn -
II
All synthetic works were carried out under ambient air and at
LX
2
]
and [Fe LX
2
]
complexes. EPR is eminently suited for
III
II
room temperature (25 °C). All solvents were purchased from
Aldrich and used as received. The following chemical substrates
were obtained in their highest commercial purity, stored at 5 °C
and purified by passing through a column of basic alumina prior
to use (Aldrich): Cyclohexene, Cyclooctene, Limonene, 1-hexene,
cis-stilbene, styrene, trans-b-methylstyrene. Hydrogen peroxide
was 30% (w/w) aqueous solution. GC analysis was performed using
an 8000 Fisons chromatograph with a Flame Ionization Detector or
a Shimadzu GC-17A gas chromatograph coupled with a GC-MS-
quantitative monitoring of Fe and Mn states [20]. Transient oxi-
III
V
III
IV
dation states e.g. (Fe OAO), Fe , Mn , (Mn @O) can also be
detected by EPR [20–22]; however, under ambient temperature,
their population is usually rapidly evolving. Therefore they can
be trapped only by a systematic freeze-quench protocol that even-
tually allows optimal detection of these elusive species by EPR.
Previously, we have reported the detection of such a catalytic
IV
intermediate Fe state in a SiO
2
-[Fe-porphyrin] system [15]. In
IV
that system, the Fe -oxo center was shown to be magnetically
coupled with a porphyrin radical i.e. forming the highly-reactive
ferryl-species that determines the catalytic performance [15]. In
the case of Mn-catalysts, the evolution to higher oxidation states,
generated via oxidation of an initial Mn state has been reported
23] and EPR detection of Mn in Mn(salen) complexes has been
QP5000 mass spectrometer. Continuous-flow of
provided by a digitally-controlled syringe pump (SP101IZ WPI).
Solution potential E was measured in situ by a Metrohm platinum
H
2
O
2
was
h
III
redox electrode (type 6.0401.100).
IV
[
IV
also demonstrated [23]. In the aforementioned work [23] the Mn
state could not been trapped during a ‘standard’ catalytic run;
2.2. Catalytic procedure
IV
instead, a Mn species, was formed by reaction of the initial
In the reactions catalyzed by Mn-complexes 1 and 2, H O was
2
2
V
Mn -oxo intermediate with the solvent in the absence of a
slowly added, over a total time of 30 min, in an acetone/MeOH
(450/200 l) solvent mixture under ambient air, at room tempera-
ture (25 °C). Then, the alkene (1 mmol), ammonium acetate
(1 mmol) as cocatalytic additive, and manganese-catalyst (1 mol)
substrate.
l
Apart from the detection of the elusive transient intermediates,
their thermodynamic parameters i.e. activation energy, entropy,
enthalpy, Gibbs free energy, are important. Moreover, a direct
comparative study of the catalytic and thermodynamic parameters
of homologous non-heme Fe and Mn complexes should provide
profound insights into the catalytic mechanisms and energetics.
Recently, systematic differences have been documented in the cat-
alytic behavior of homologous Mn- vs. Fe-complexes [24–26], how-
ever a thermodynamic basis of these observations is still in need. In
l
were added and the resulting reaction solution was vigorously stir-
red in a glass round-bottom flask for 6 h under ambient conditions.
We underline that ammonium acetate is added as an additive that
is absolutely required in order to generate efficient Mn-catalytic
systems [27]. Acetophenone or bromobenzene was used as internal
GC standards. The molar ratio of [Mn-catalyst:H O :CH COONH :
2
2
3
4
substrate] was equal to [1:2000:1000:1000
volume of catalytic reaction was about 1 ml.
lmol]. The total
this context, herein we present the first comparative catalytic/ther-
II
modynamic study of the homologous complexes [Mn L(OAc)
2
] (1)
In the reactions catalyzed by Fe-complexes 3 and 4, H O2
2
II
II
II
vs. [Fe L(OAc)
2
] (3) and [Mn LCl
2
] (2) vs. [Fe LCl
2
] (4). A systematic
diluted in CH CN solvent (1:10 v/v) was slowly added (within a
3
freeze-quench EPR study was carried out to trap the transient
intermediates and then to monitor the full time-course of their
evolution. In parallel we have performed detailed kinetic studies
of the redox evolution of monitoring their UV–vis spectra at a
range of temperatures from +25 °C to ꢂ45 °C. Low-Temperature
UV–vis (LT-UV–vis) spectroscopy allows a continuous mapping of
the spectral evolution of the complexes. Then, a thermodynamic
analysis of the LT-UV–vis data allowed numerical estimates of
period of 5 min) to a CH CN solution containing the catalyst
3
(1
room temperature (25 °C). The molar ratio [Fe-catalyst:
H O :substrate] was equal to [1:500:1000 mol]. Acetophenone
lmol) and the substrate (1000 lmol) under N2 atmosphere at
2
2
l
or bromobenzene was used as internal GC standards. The total vol-
ume of catalytic reaction was approximately 1 ml.
The progress of the catalytic reactions was monitored by GC-MS
for 20 ll samples, periodically taken from the reaction mixture.
Quantitative analysis of the GC data was done by comparing the
integrals of the GC peaks vs. the internal standard, thus providing
the substrate conversion and product yield. Reactions were
à
à
à
the activation energies, E
a
, as well as
D
S ,
D
H ,
D
G
for the Fe-
and Mn-complexes. Based on the EPR and thermodynamic data,
we construct a reaction-path scheme which unravels that the

1
06
M. Papastergiou et al. / Journal of Catalysis 341 (2016) 104–115
completed within 6 h for Mn-catalysts and within 24 h for Fe-
catalysts. Error-bars indicated in the data were derived from at
least three replicates of each experiment.
oxidations providing significant yields (39–97%), and high selectiv-
ity for epoxide products (the mass balance is 99 ± 1%) in most of
the cases. In order to generate efficient catalytic systems, ammo-
Turnover Numbers and Turnover Frequencies Calculations
TONs were calculated using Eq. (1a)
nium acetate was added as additive. As reported many times pre-
II
viously [29–31], CH
3
COONH
4
plays a multiple role in Mn /H
2
O
2
and this can attributed to a dual acid-base role for CH
able to act as proton-donor and proton-acceptor [29–31]. More
3
COONH
4
moles of product
moles of catalyst
TONs ¼
ð1aÞ
ð1bÞ
precisely, according to the catalytic mechanism we have proposed
ꢂ
II
recently [i] CH
3
COO abstracts a proton from H
2
O
2
, promoting its
TOFs were calculated using Eq. (1b)
II
coordination to Mn and formation of Mn AOOH species and [ii]
TONs
t
+
II
subsequently, NH
4
, by acting as a proton donor to Mn AOOH,
TOFs ¼
IV
accelerates heterolytic OAO cleavage forming the active Mn @O
species which [iii] catalyzes alkene epoxidation [27]. Oxidation of
cyclooctene and hexene-1 catalyzed by 1 and 2 provided 100%
selectivity for cis-epoxide with 54.0–64.0% and 10.3–10.5% yield,
and 100% m.b. respectively (Fig. 2). The total cyclohexene oxidation
yield was 67.2% for 1 and 57.0% for 2. In particular, both 1 and 2
provided mainly epoxide with 61.4–53.3% yields, while the allylic
oxidation path resulted in small amounts of 2-cyclohexen-1-ol
2
.3. Low temperature UV–vis spectroscopy
Low-temperature UV–vis spectra were recorded in a Hitachi
spectrophotometer operating in the 190–900 nm wavelength
range, in 3 ml quartz cuvettes (1 cm optical path). The sample
was cooled in a Unisoku cryostat that was inserted inside the
UV–vis spectrophotometer beam-chamber. This system allows dig-
ital control of sample-temperature from +100 °C down to ꢂ100 °C.
Cooling of the sample was achieved by a controlled-flow of cold
-gas derived from heating of liquid-N
stabilization with an error of ±0.1 °C.
UV–vis sample preparation: For the low-temperature UV–vis
study of the catalytic reaction, the metal-complex was solubilized
(
3.6–2.1%) and 2-cyclohexen-1-one (2.2–1.7%). During the epoxi-
dation of styrene and cis-stilbene by 1 and 2, the corresponding
epoxide was the only product, cis-stilbene-epoxide in the case of
cis-stilbene, with yields 39.0–42.3% and 44.0–43.7% respectively.
The methyl-substituted derivative of styrene, trans-b-methyl styr-
ene, is more reactive than styrene giving total oxidation yields of
N
2
2
. This allows temperature
9
7.3% and 66.5% catalyzed by 1 and 2 respectively; in both cases,
the corresponding trans-epoxide was identified as single product.
The products detected from limonene oxidation were as follows:
3
to the suitable solvent i.e. CH CN for the Fe-complex or acetone/
MeOH mixture in the case of Mn-complex; the solution cuvette
was inserted to the cryostat, cooled to the desired temperature
and allowed to equilibrate for 10 min, under stirring. Then, the sub-
(
i) two epoxides (cis and trans) originated from epoxidation of
the electron-rich double bond in the 1,2-position and (ii) alcohols
derived from hydroxylation either at the 1-position or at the
2 2
strate plus co-catalyst, when Mn-catalysts were used, plus H O
6
8
-position. The total yield of the oxidation products was raised at
was added at this time-point (t = 0) and the collection of the
UV–vis spectra (240–900 nm) was started, with a [1 spectrum/
min] rate.
6.1% catalyzed by Mn-catalyst 1 and 60.6% by 2. Noticeably, the
II
II
[
2 2
Mn L(OAc) ] and [Mn LCl ] catalysts did not show any remark-
able difference in their catalytic behavior i.e. there is no detectable
effect of the labile ligands on the studied catalysis.
2
.4. EPR spectroscopy
II
Fe-catalysts: The catalytic evaluation of the Fe -complexes 3, 4
2 2
with H O , generally showed consistently much lower catalytic
Electron paramagnetic resonance (EPR) spectra were recorded
at liquid N temperature 77 K with a Bruker ER200D spectrometer
2
II
efficiency than the corresponding Mn -complexes 1, 2. Analogous
trend was reported for the catalytic performance of other homolo-
gous Fe-non-heme vs. Mn-oxidation catalysts [24]. Here, in oxida-
tion reactions of alkenes the Fe-catalysts provide varying yields
equipped with an Agilent 5310A frequency counter. The spectrom-
eter was running under home-made software based on Lab View.
EPR sample preparation: all EPR samples were prepared in 5 mm
internal diameter quartz tubes by Wilmad Glass. In the case of the
Mn-complexes, appropriate amounts of the catalytic reaction
components were dissolved in acetone/methanol mixture with
(
1.4–78.9%) (Fig. 3). We underline however the remarkable turn-
over numbers (see Table 1), achieved by the present Fe-non-
heme catalysts 3 and 4 approaching 394 TONs i.e. compared with
analogous non-heme Fe-catalysts in Fig. 4A.
The total cyclohexene oxidation yield was 26.4% with 132 TONs
for 3, and 19.5% with 97 TONs for 4. Allylic oxidation formed 2-
cyclohexen-1-ol (9.8–8.0%) and 2-cyclohexen-1-one (13.4–8.9%).
However cyclohexene epoxidation was also observed, with low
epoxide yields (3.1–2.7%). Oxidation of cyclooctene provides
a molar ratio [catalyst]: [H
1:2000:1000:1000 mol]. For the study of Fe-catalysts, CH
was used as solvent containing a molar ratio of [catalyst]:[H
cyclohexene] = [1:500:1000 mol].
Time-evolution of the EPR spectra: When all catalytic compo-
2
O
2
]:[CH
3
COONH
4
]:[cyclohexene] =
CN
]:
[
l
3
2 2
O
[
l
nents were added into the EPR tube, this was immediately i.e.
within 5 s, frozen at 77 K. The EPR spectrum of this sample is
referred as ‘‘t ”. To record the time-evolution of the EPR spectra,
0
the sample was thawed at room temperature and allowed to
evolve for a predetermined time interval, followed by rapid freez-
ing at 77 K, within 5 s. Numerical simulation of experimental EPR
spectra was performed with Easy Spin 4.5.1 [28] software.
1
00% selectivity for cis-cyclooctene epoxide with 19.5% and 16.5%
yield and 97 and 82 TONs, by 3 and 4 respectively. Hexene-1 that
is a rather hard oxidation substrate showed epoxide yields from
1
.4% to 3.0% by 3 and 4 respectively and 100% selectivity for the
cis-epoxide.
The products detected from limonene oxidation, were two
epoxides (cis and trans) originating from epoxidation of the
electron-rich double bond in the 1,2-position and alcohols derived
from hydroxylation in 1-, 2- and 6-position of the limonene ring.
Additionally, considerable amounts of the corresponding ketone
at 6-position were also formed (see details in Table 1). Oxidation
products from the more accessible, but less electron-rich 8, 9-
3
. Results
3
.1. Catalytic performance evaluation
Mn-catalysts: Table 1 lists the results of catalytic data obtained
double-bond were not observed. The total yield of the limonene
II
II
II
for the Mn - and Fe -complexes. Based on these, we observe that
oxidation products raised at 25.7% with 128 TONs for Fe L(OAc)
2
II
II
both non-heme Mn complexes 1 and 2 are efficient in alkene
2
and 29.3% with 146 TONs for Fe LCl .

M. Papastergiou et al. / Journal of Catalysis 341 (2016) 104–115
107
Table 1
Hydrocarbon oxidation catalyzed by Mn L(OAc)
II
II
II
II
2
, Mn LCl
2
, Fe L(OAc)
2
and Fe LCl
2
2 2
with H O .
a
a
LFe(OAc) b
LFeCl b
2
Substrate
Product
LMn(OAc)
Yield (%)a
2
LMnCl
2
2
TONc
TOFd
Yield (%)a
TONc
TOFd
Yield (%)b
TONc
TOFd
Yield (%)b
TONc
TOFd
Cyclohexene
cis-Epoxide
61.4
2.2
3.6
53.3
1.7
2.1
3.1
13.4
9.8
2.66
8.87
7.96
2
2
-Cyclohexenone
-Cyclohexenol
6
72
112
90
570
640
95
132
97
5
4
97
82
4
3
Cyclooctene
Limonene
cis-Epoxide
54.0
66.0
540
64.0
47.7
107
19.5
12.5
16.5
13.4
Epoxide
Limonene alcohol
Limonene ketone
20.1
12.9
7.2
9.5
6.4
6.0
8
61
143
17
606
105
101
17
128
7
5
146
15
6
Hexene-1
cis-Epoxide
10.3
44.0
103
10.5
43.7
1.4
0.3
3.0
0.6
cis-stilbene
cis-Epoxide
36.0
20.1
14.6
30.0
Benzaldehyde
440
390
973
73
437
423
665
73
280
235
390
11
9
223
296
394
9
Styrene
Epoxide
Benzaldehyde
39.0
97.3
42.3
66.5
5.4
41.7
15.01
44.3
65
70
12
Trans-b-methylstyrene
trans-Epoxide
Benzaldehyde
16.7
61.2
48.25
30.54
162
111
16
16
a
Conditions ratio of catalyst:H
2
O
2
:CH
3 4 3 3 3
COONH :substrate = 1:2000:1000:1000; Equivalent of catalyst = 1 lmol in 0,65 ml CH COOCH :CH OH (0,45:0,20); The reactions
were completed within 6 h.
b
Conditions ratio of catalyst:H
TON: turnover number, moles of epoxide formed per mole of catalyst.
TOF: turnover frequency which is calculated by the expression [epoxide]/[catalyst] ⁄ time (h ).
2
O
2
:substrate = 1:500:1000; equivalent of catalyst = 1
l
mol in 0.9 ml MeCN; the reactions were completed within 24 h.
c
d
1
Fig. 2. Bar chart representation of alkene epoxidations catalyzed by MnII complexes
and 2 in the presence of H . See Table 1 for further details.
1
2
O
2
Fig. 3. Bar chart representation of alkene epoxidations catalyzed by FeII complexes
and 4 in the presence of H . See Table 1 for further details.
3
2 2
O
Cis-stilbene was oxidized by 3 and 4 with total oxidation yields
of 56.1% (280 TONs) and 44.6% (223 TONs) respectively. The prod-
ucts were identified as cis-epoxide and benzaldehyde. Styrene oxi-
dation provided benzaldehyde as major product generated from
oxidative cleavage of the exo-cyclic double bond. However, epox-
ide has been also formed by direct oxidation of the same double
bond. Overall, styrene was oxidized by 3 and 4 with total oxidation
yields of 47.1–59.3% and 235, 296 TONs respectively. Trans-b-
methylstyrene oxidation by 3 provided benzaldehyde, as major
product (61.2%). Trans-epoxide has been also detected (16.7%).
On the contrary, oxidation of trans-b-methylstyrene by 4 provided
trans-epoxide as major product (48.2%) and, at the same time, con-
siderable amount of benzaldehyde has been also observed (30.5%).
Overall, trans-b-methyl-styrene was oxidized by 3 and 4 with total
oxidation yields of 77.9–78.7% and 390–394 TONs respectively.
The Fe catalysts 3 and 4 show no difference in their catalytic
behavior, thus -as in the case of the Mn-catalysts 1 and 2 -their
labile ligands i.e., chloride and acetate play no role in the catalysis
studied herein (present data, see Table 1).
For comparison, we notice that the present homogeneous Fe-
catalysts show remarkably higher turnover numbers (up to 390)
(see Table 1, Fig. 4A) compared with other non-heme Fe-catalysts
III
2
2
i.e. DPEIFe Cl [24] and FeCl (btaH) [32] studied previously by us
under the same catalytic conditions, which provided TONs close
to 46 (Fig. 4A); the enhanced catalytic activity of the present

1
08
M. Papastergiou et al. / Journal of Catalysis 341 (2016) 104–115
Fig. 5. Time dependence of cyclohexene epoxidation and solution redox potential
II
II
2 2
for the same reaction catalyzed by Mn L(OAc) and Mn LCl .
II
[
B] Fe-catalysts: The time-course profiles of the Fe L(OAc)
2
- and
-catalyzed oxidations of cyclooctene in conjunction with
the observed redox potential of solution E are presented in
Fig. 6. At the beginning of the reaction catalyzed by Fe L(OAc)
was +475 mV; after 12 h it dropped to +450 mV with a 8.8%
epoxidation yield and finally after a reaction time of 24 h, it
approached a value of E = +430 mV with a 19.5% yield. In the case
catalyzed epoxidation, at reaction time t = 0 h, E was
490 mV and then decreased to +460 mV (t = 12 h), providing a
product yield of 5.0% and at t = 24 h, E approached +441 mV with
6.5% epoxidation yield. Similar time course profiles of the homo-
II
Fe LCl
2
h
II
2
,
E
h
h
II
of Fe LCl
2
h
+
h
1
geneous catalysts have been observed for all substrates used
herein. The Turnover Frequencies (TOFs) listed in Table 1 have
been calculated on the basis that the Fe-depended oxidations were
completed within 24 h.
Thus, the data in Figs. 5 and 6 and the TONs, TOFs in Table 1
reveal an important difference between the Fe-catalysts 3 and 4
vs. the Mn-catalysts 1 and 2.
2 2
Fig. 4. (A) Total turnover numbers for alkene epoxidations with H O catalyzed by
II
II
III
Fe L(OAc)
FeCl (btaH)
2
OAc) and Fe LCl .
2
, Fe LCl
2
, compared with other non-heme Fe catalysts DPEIFe Cl and
II
II
3
2
and (B) comparison of TONs achieved by Mn (OAc) , MnLCl vs. Fe L
2 2
II
(
2
TONs; TOFs½Mn-systemsꢃ ꢀ TONs; TOFs½Fe-systemsꢃ
Fe-complexes 3 and 4 is assigned to the different coordination and
geometrical environments around the Fe-centers.
The Mn catalysts have TOFs in the range of 100 that are
about ꢄ 20-fold higher than the TOFs (ꢁ5) of the Fe-catalysts. Mn
catalysts cycle ꢄ 20-times more rapidly and consume much more
A comparison of the catalytic TONs of the present Fe-catalysts 3
and 4 vs. the homologous Mn-catalysts 1 and 2 (Fig. 4B), reveals
that Mn-catalysts show consistently remarkable higher turnover
numbers (TONs) than the homologous Fe-catalysts. As we show
hereafter, redox/thermodynamics of the higher oxidation states
determine this phenomenon.
2 2
H O equivalents than the homologous Fe-catalysts.
This striking difference has been observed previously for
homologous Mn, Fe non-heme catalysts [24–26] i.e. typically for
the same ligand, Mn-catalysts show higher TONs than the Fe-
catalysts, however a direct thermodynamic explanation has not
Catalytic Kinetics, TOFs: [A] Mn-catalyst: the time-course profiles
II
II
of the Mn L(OAc)
ene in conjunction with the observed solution potential of E
standard hydrogen electrode, SHE) are presented in Fig. 5. At the
2
- and Mn LCl
2
-catalyzed oxidations of cyclohex-
h
(vs.
II
beginning of the reaction catalyzed by Mn L(OAc)
425 mV; after 3 h it dropped to +305 mV with a 50% epoxidation
yield and finally after 6 h reaction time, it approached a steady
2 h
, E was
+
II
value of E
catalyzed epoxidation, at t = 0 h, E
decreased to +283 mV (t = 2 h), providing a product yield of
8.0%; at t = 6 h, E approached +270 mV with 57.4% epoxidation
h
= +296 mV with a yield of 67.2%. In the case of Mn LCl
2
-
h
was +337 mV and then
4
h
yield. Similar time course profiles of the homogeneous catalysts
have been observed for all substrates used herein demonstrating
that the oxidation reactions are completed within 6 h. Based on
this finding the Turnover Frequencies (TOFs) listed in Table 1 have
h
been calculated. Moreover, the very good correlation of the [E vs.%
yield] data in Fig. 5 shows that the catalytic reactions proceed in
tandem with consumption of oxidative equivalents, as we show
hereafter by EPR- via formation of the transient Mn @O states.
Fig. 6. Time dependence of cyclohexene epoxidation and solution redox potential
for the same reaction catalyzed by Fe L(OAc) and Fe LCl .
2 2
IV
II
II

M. Papastergiou et al. / Journal of Catalysis 341 (2016) 104–115
109
been attempted so far. To do this, in principle, we have to consider
Table 2
Spin Hamiltonian parameters used for simulation of the Mn (S = 5/2, I = 5/2) and
II
the rate-limiting steps that may be involved in such reactions: in
oxidative catalysis both Mn- and Fe-complexes usually advance
IV
II
Mn (S = 3/2, I = 5/2) EPR spectra for Mn L(OAC)
2
catalyst.
IV
IV
via high oxidation states i.e. Fe [33] and Mn [34,35]. The redox
D (G)
E/D
Aiso (G)
potential of the couples Fe -Fe3+/Fe is higher than the couple
2+
4+
II
Mn (S = 5/2, I = 5/2)
2
+
4+
II
Mn /Mn i.e. for this reason mild oxidants could suffice to acti-
vate Mn-complexes while stronger ones are necessary for Fe-
complex activation. Herein, quantitative insight into these trends
was achieved by in situ monitoring of the oxidative evolution of
the Fe- and Mn-catalysts, using low-temperature UV–vis and
freeze-quench EPR spectroscopies.
Mn L(OAC)
2
200
0.013 ± 0.005
107
ꢂ1
D (cm
)
E/D
Aiso (G)
IV
Mn (S = 3/2, I = 5/2)
IV
Mn L(OAC) -conformer-1 (65%)
2
4.0 ± 0.3
4.0 ± 0.3
0.21 ± 0.005
0.23 ± 0.005
110
110
IV
Mn L(OAC)
2
-conformer-2 (35%)
Numerical simulation of the MnIV (S = 3/2, I = 5/2) spectrum
Fig. 7A red line) revealed that consists of two Mn conformations
3
3
.2. Mechanistic spectroscopic studies
IV
(
.2.1. EPR spectroscopy
with slightly different E/D ratios E/D = 0.21 and E/D = 0.23. Aiso was
110 G for both species. These two sets of E/D parameters indicate
II
II
Mn L(OAC)
2
: Fig. 7 shows 77 K EPR spectra for Mn L(OAC)
2
in
II
IV
acetone/methanol. The spectra are typical for a mononuclear Mn
S = 5/2, I = 5/2) [20]. After addition of H and ammonium acet-
that there are two Mn (S = 3/2, I = 5/2) conformers differing
(
2
O
2
somehow in the ligand field symmetry. The best simulation was
achieved by assuming 65% of conformer-1 with E/D = 0.21 and
35% of conformer-2 with E/D = 0.23. In Table 2 the simulation
II
ate, the Mn signal intensity decreased gradually, and a new broad
signal with effective values in ꢁ5.5 and ꢁ3.0–3.5 was detected.
This new signal was maximized within 20 min, see inset Fig. 7C,
and then decayed again. The oxidative evolution of this signal indi-
cates that originates from high Mn oxidation states. Taking into
II
IV
parameters for Mn and Mn spectra are listed.
II
IV
The time evolution of the Mn and Mn species in Fig. 7C shows
that under the condition of the experiment described in Fig. 7, the
majority of the Mn centers are oxidized from the Mn to Mn state
within 20 min, and then the high oxidation states decline to EPR
IV
II
IV
account literature EPR data for Mn (salen) [21,22,36] this EPR sig-
nal can be assigned to a monomeric Mn (S = 3/2) spin system.
IV
II
Numerical calculations of EPR spectra for Mn (S = 5/2, I = 5/2)
silent states.
IV
II
: Fig. 8A (i) presents the EPR spectrum for FeIIL
and Mn
(S = 3/2, I = 5/2) were performed using the spin
Fe L(OAC)
2
Hamiltonian
2
(OAC) powder i.e. with no addition of solvent. The absence of sig-
III
nals shows the absence of Fe , indicating that all Fe centers are in
2
2
2
b
II
H ¼ g bB S þ D½Sz ꢂ ð35=12Þ þ ðE=DÞðSx ꢂ Sy Þꢃ þ SAI
ð2Þ
e
the ferrous state (Fe ) [20]. After addition of solvent and H O
(
2
2
III
Fig. 8A), a High Spin Fe (S = 5/2) signal was progressively formed,
where g
e
= 2.0023 and D and E are the axial and rhombic zero-
and maximized within ꢁ10 min, see plot of time evolution in
field splitting parameters respectively [20,37]. For D ꢀ h
v =
III
ꢂ1
Fig. 8C. In parallel, an EPR signal corresponding to Low spin Fe
0
.314 cm = 3400 G, the EPR spectrum is not sensitive to the
IV
(S = 1/2) was developed. This LS signal was maximized after
particular D-value [37]. Thus in the case of Mn a D value of
ꢂ1
II
20 min and then declined to zero within 30 min Fig. 8A and C.
4
.0 cm was used [38]. Numerical simulation of the Mn spectrum,
III
Numerical calculations of EPR spectra for Fe (S = 5/2) were
shows that the zero field splitting parameter D is 200 G, with
performed using the spin Hamiltonian
Aiso = 107 G and almost axial ratio E/D = 0.013 listed in Table 2.
II
IV
Fig. 7. EPR spectra for Mn L(OAC)
2
2
in methanol/acetone: (A) for reaction times t = 5, 10, 20, 30 min after addition of CH
3
COONH
4
and H
2
O
2
(Red line) theoretical Mn (S = 3/
II
II
IV
) EPR spectrum obtained using the spin-Hamiltonian parameters listed in Table 2, (B) same as in (A) zoomed in the Mn signals and (C) time evolution of Mn and Mn
species.

1
10
M. Papastergiou et al. / Journal of Catalysis 341 (2016) 104–115
3
+
3
+
HS-Fe (S=5/2)
3+
LS-Fe -OOH (S=1/2)
LS-Fe -OOH (S=1/2)
reaction time
t_35min
(B)
g= 2.19, 2.002, 1.91
t_30min
t_20min
t_15min
exp.
t_10min
t_5min
t_2min
t_30sec
theory
+
[Cyclohexane+HO ]
2
2
(A)
powder
0
1000
2000
3000
4000
5000
Magnetic Field (Gauss)
2
400
2800
3200
3600
4000
Magnetic Field (G)
H O Addition
2
2
(C)
1
0
0
0
0
0
.0
.8
.6
.4
.2
.0
3+
Fe (S=5/2)
3
+
Fe (S=1/2)-OOH
0
15
30
45
60
75
90
105
120
135
time (min)
II
and substrate, at reaction times t = 5–35 min, (B) a Low-Spin FeIII (S = 1/2) experimental and
Fig. 8. (A) EPR spectra for Fe L(OAC)
2
in CH CN:powder, after addition of H O
3 2 2
simulation (red line). Simulations were obtained using the spin-Hamiltonian parameters listed in Table 3 and (C) time evolution of the Fe species.
2
2
2
b
Detailed microwave-power dependence study of the low-spin
Fe3+ signal allowed the identification of the g-tensor components,
see Fig. S1 in Supporting Information.
The rhombicity (V/D) and tetragonality (D/k) parameters were
calculated from the g values [39]. The structural significance of
H ¼ g bB S þ D½Sz ꢂ ð35=12Þ þ ðE=DÞðSx ꢂ Sy Þꢃ
ð3Þ
e
where g
e
= 2.0023 and D and E are the axial and rhombic zero-field
splitting parameters respectively [20,37]. A D value of 3.0 cm
ꢂ1
ꢂ1
used, indicative of any D ꢀ h
v
= 0.314 cm that is typical of high
spin Fe-complexes [37]. The g-values detected in the experimental
spectrum are effective g-values resulting from the effect of D and
E on the energy levels.
these values can be parametrized by the so-called Peisach-
Blumberg diagram i.e. a plot of [V/
in Supporting Information. Fig. S2 shows that the [V/
D
] vs. [
D
/k] parameters, Fig. S2
] vs. [ /k]
D
D
In the case of the Low-Spin FeIII (S = 1/2) we used a simple Zee-
man Hamiltonian, with the g tensor having principal values the
parameters can vary among heme or non-heme FeAOOH systems,
indicating a high sensitivity to ligand-filed effects. On the other
experimental g
1
, g
2
, g
3
listed in Table 3.
III
hand the [V/
D
] vs. [D
/k] for our LS-LFe complex falls between
b
those of heme- and non-heme Fe-complexes. We consider that this
reflects the ‘‘semi-heme” structure of the ligand L. Overall, the g-
values, [V/D] vs. [D/k] parameters, the transient kinetics of the
H ¼ bBgS
ð4Þ
LS-EPR signal (together with the UV–vis spectra) corroborate its
assignment to a LS-Fe( AOOH state characteristic for Low-Spin
non-heme ferric complexes where the Fe atom is coordinated by
one hydroperoxo-unit [40–42]. From Fig. 8 we observe that [i]
III)
Table 3
Spin Hamiltonian parameters used for simulation of the Fe (S = 5/2) and FeIII (S = 1/
III
II
2
) EPR spectra for Fe L(OAC)
2
catalyst.
III
the development of Low-Spin LFe AOOH state follows the oxida-
ꢂ1
D (cm
)
E/D
3+
tive evolution of the High-Spin Fe after a time-lapse i.e. indicat-
III
High spin Fe (S = 5/2)
.0
ing that structural rearrangement events occur during this time –
3
0.30
III
not redox advancement- that result in the Low-Spin LFe AOOH
g
1
g
2
g
3
D
/k
V/D
configuration. Typically, formation of a Low-Spin sate is indicative
of a strong-ligand field around the iron center. [ii] then, after this
III
Low spin Fe (S = 1/2)
.19 2.005
2
1.91
7.6
0.7
III
Low-Spin LFe AOOH formation, its decline indicates formation
Error bars in g values ±0.05.
of higher Fe oxidation states [15,41,42], not detected in our EPR

M. Papastergiou et al. / Journal of Catalysis 341 (2016) 104–115
111
experiments probably due to very short lifetime, a typical case for
these highly reactive species [15,41,42]. We underline that, quan-
titation of the formed LS LFe AOOH centers indicates that only
dependence of k provides the energetic characteristic of the cat-
alytic system [45],
III
lnðkÞ ¼ ðꢂE
a
=RÞ ꢅ ð1=TÞ þ C
ð5Þ
ꢁ
30% of the total Fe is converted from the High-Spin state to the
III
LS LFe AOOH by H
2
O
2
. In contrast, H
2 2
O oxidizes almost 100% of
where E is the activation energy of the reaction in J/mole, T is the
a
II
IV
the Mn to Mn . Thus the present EPR data reveal a second funda-
mental difference between Fe L(OAC) and Mn L(OAC) :
2 2
temperature in K, R is the gas constant 8.314 J/K mol and C is a con-
stant to be defined by the fit to the experimental data. In a ln(k) vs.
II
II
(
1/T) plot, the Arrhenius process is reflected as a line with slope
½
% of oxidizable Mn in MnLðOACÞ2ꢃ
(ꢂE /R). Using equation [5], the best-linear fit, (solid line in
a
ꢀ
½% of oxidizable Fe in FeLðOACÞ2ꢃ
Fig. 10B), provides an estimate of the activation energy
a
E = 58.6 ± 3 kJ/mol, listed in Table 4. This activation barrier corre-
sponds to oxidation of Mn to Mn @O i.e. according to EPR data.
This phenomenon, together with the trends in TONs, TOFs, indi-
II
IV
cates the existence of different thermodynamic barriers that differ-
entiate the catalytic evolution of Mn L(OAC)
Further analysis of the kinetic rates, using the Eyring-Polani Eq.
II
II
2
vs. Fe L(OAC)
2
. Under
and
à
à
(
[
6) can provide the enthalpy
45]
D
H and entropy
D
S of the reaction
II
comparable solution potentials i.e. ꢁ430 mV in Mn L(OAC)
2
II
ꢁ
470 mV in Fe L(OAC)2, the same oxidant H
2 2
O is less efficient in
z
z
Fe advancement at higher oxidation states. This indicates that
k
T
D
R
H 1
kB
DS
II
ln ¼ ꢂ
þ ln h
þ
R
ð6Þ
2
the Fe L(OAC) catalyst will perform less catalytic turnovers, as
T
we show herein with the catalytic results, and thus will be less effi-
cient in catalysis, in agreement with the present catalytic data in
Fig. 4B and Table 1. To obtain numerical values of the thermody-
manic barriers, we have monitored the redox/catalytic kinetics of
where R is the gas constant = 8.314 J/K mol, T is the absolute
ꢂ34
temperature, h is the Planck constant = 6.626 ꢄ 10
J s, and k
B
J/K. According to
ꢂ23
the Boltzmann constant
k
B
= 1.380 ꢄ 10
II
II
Eq. (6), an Eyring-Polanyi plot, [ln(k/T) vs. 1/T], see Fig. 10C, provides
2 2
Mn L(OAC) vs. Fe L(OAC) in situ using low-temperature UV–vis
à
a line whose slope equals ꢂ
D
H /T and the y-intercept provides
B
S /R + ln(k /h). Accordingly, from the data in Fig. 8C we estimate
à
spectroscopy, as we analyze hereafter.
à
D
D
à
H = ꢂ55.0 ± 3.0 kJ/mole and
DS = +0.3 ± 0.05 kJ/K mole, listed in
3.2.2. Monitoring the redox/catalytic kinetics by low temperature UV–
Table 4.
II
vis spectroscopy
[Fe L(OAc)
2
]: An analogous thermodynamic analysis, was per-
II
II
[
2 2
Mn L(OAc) ]: The catalytic reaction for [Mn L(OAc) ] was
formed by low-temperature UV–vis experiments were performed
II
II
monitored in an [acetone: methanol] for a temperatures T = 0 °C
to ꢂ30 °C. Characteristic time-evolution of the spectra is displayed
in Fig. 9A (T = ꢂ30 °C) and 9B (T = ꢂ15 °C).
at different temperatures for [Fe L(OAc)
complex in CH CN, with no oxidant added, bands at 461 and
503 nm were observed at all temperatures, see Fig. 11A and B. After
addition of H a new band with a maximum at 420 nm was pro-
2
]. Starting with the Fe
3
Before addition of oxidant, this complex shows a characteristic
2 2
O
II
III
band at kmax = 420 nm characteristic of the Mn state [43,44]. After
gressively developed. This new band is characteristic for a Fe -
addition of H
2
O
2, the spectrum evolves, and a characteristic feature
AOOH species formation [40,46–48]. This is the Low-Spin HOO-
Fe (S = 1/2) intermediate detected by EPR, see Fig. 8A and B. More
IV
III
at kmax = 397 nm characteristic of a Mn @O species [43,44] is
emerging. The formation of this Mn @O species is in agreement
with our EPR data, Fig. 7.
IV
specifically Beller et al. [40] show the formation of Fe-OOH species
with absorbance peak at 465 nm. In reference [49] Que et al.
demonstrated that the configuration of the ‘‘OAO(H)” axial ligation
can –in certain cases- be distinguished by the UV–vis band-
à
à
a
Thermodynamic parameters E , DH , DS : The time-evolution of
(
IV)
the intensity of the Mn @O band at 397 nm (I397) is plotted in
Fig. 10(A) for T = 0, ꢂ15, ꢂ20, ꢂ30 °C respectively. At lower T, less
3
+
position in LS-Fe OAO(H) [49]. Chiraldi et al. reported that heme
pyridine iron complex upon oxidation with H forms a band at
IV
Mn centers achieve oxidation to the Mn state; thus at ꢂ30 °C, I397
2 2
O
IV
(III)
is ꢁ50% of that at 0 °C. The Mn @O formation kinetics in Fig. 10A
418 nm attributed to the Fe AOOH species. In a previous work
of our group [50], using a LFe complex were L = [3-{2—[2-(3-hy
III
shows a strong temperature dependence, revealing the existence of
an activation energy barrier. An Arrhenius plot, derived by plotting
the [initial kinetic rates (k)], i.e. estimated by the slope of the initial
part of the kinetic data, in Fig. 10A vs. [1/T] is displayed Fig. 10B.
According to the Arrhenius theory [45] the temperature
droxy-1,3-diphenyl-allylideneamino)-ethyl amino]}-1,3-diphenyl-
propen-1-ol] the formation of FeAOOH was studied by UV–vis. In
this study, after the addition of H
2
O
2
a band at 413 nm, attributed
to Fe( AOOH, progressively increased.
III)
II
Fig. 9. UV–vis spectra of redox/catalytic kinetics for [Mn L(OAc)
spectrum has been recorded within 60 s.
2
] in [acetone: methanol] (A) reaction temperature ꢂ30 °C and (B) reaction temperature ꢂ15 °C. Each

1
12
M. Papastergiou et al. / Journal of Catalysis 341 (2016) 104–115
T=0 °C
3
2
2
1
1
0
,0
,5
,0
,5
,0
,5
T=-15 °C
T=-20 °C
T=-30 °C
(A)
0
10 20
30 40 50 60 70 80
Time (min)
-
-
-
-
-
-
-
-
-
0,5
1,0
1,5
2,0
2,5
3,0
3,5
4,0
4,5
-6.0
-6.5
-7.0
H= -55 3.5 kJ/mol
S= 0.3 0.05kJ/mol
E =58.6 ± 3.0 kJ/mol
a
-7.5
k
-8.0
-8.5
-9.0
-9.5
(B)
(C)
-
10.0
3.6x10
-3
-3
-3
-3
-3
-3
-3
-3
3
,6x10
3,8x10
4,0x10
4,2x10
3.8x10
4.0x10
4.2x10
1/T (1/K)
1/T(1/K)
IV
II
Fig. 10. (A) Kinetics of the of Mn @O formation by [Mn L(OAc)
2
], monitored by the I397 at different temperatures, (B) Arrhenius plot and (C) Eyring-Polanyi plot of the
IV
formation of Mn
.
II
IV
Table 4
slower than the Mn to Mn kinetics, is agreement with the TOFs.
Under the same redox conditions, higher temperatures i.e. +5 to
ꢂ5 °C, had to be selected for the iron complex to achieve analogous
II
Thermodynamic parameters estimated for the reaction catalyzed by Mn L(OAc)
2
and
II
2
Fe L(OAc) .
E
a
(kJ/mol)
D
H (kJ/mol)
D
S (kJ/K mol)
D
G
(T-302K)
rates as for Mn complexes i.e. 0 °C to ꢂ30 °C, compare Figs. 10A vs.
(
kJ/mol)
II
1
2A. This reveals, that the iron catalyst Fe L(OAc)
2
has to overcome
II
II
Mn L(OAc)
2
+58.6 ± 3.0
+91.3 ± 3.0
ꢂ55.0 ± 3.0
ꢂ84.3 ± 3.0
+0.3 ± 0.05
+0.2 ± 0.05
ꢂ146 ± 6.0
ꢂ144 ± 6.0
a considerably higher thermodynamic barrier than the Mn L(OAc)
catalyst.
2
II
Fe L(OAc)
2
Quantitatively, the Arrhenius activation energy is estimated to
à
à
be
+
E
a
= 91.3 ± 3.0 kJ/mol, and
D
H = ꢂ84.3 ± 3.0 kJ/mole,
DS =
0.2 ± 0.05 kJ/K mole, listed in Table 4. At room temperature,
T = 302 K, the free energy change for this reaction is
The UV–vis spectrum shows also the transient character of the
à à
DG = DH –
Fe3+AOOH species. At T = +5 °C, see Fig. S3 in Supporting Informa-
à
TD
S = ꢂ84–60 = ꢂ144 ± 6 kJ/mole.
3+
tion the band at 420 nm, due to LFe AOOH, is maximized within
1
0 min, and then it decays rapidly within the next 10 min.
Fig. 12A, shows the kinetic evolution of the Fe AOOH band at
III
4
. Discussion
4
20 nm (I420) at different temperatures. The Arrhenius and
Eyring-Polanyi plots are presented in Fig. 12B and C respectively.
Notice that, the kinetic rate of Fe to Fe AOOH oxidation is much
The thermodynamic events: Overall, the present thermodymanic
data in Table 4 show that
II
III
II
Fig. 11. UV–vis spectra of redox/catalytic kinetics for [Fe L(OAc)
recorded within 60 s.
2
] in [CH
3
CN] (A) reaction temperature 0 °C and (B) reaction temperature ꢂ10 °C. Each spectrum has been

M. Papastergiou et al. / Journal of Catalysis 341 (2016) 104–115
113
2
2
2
2
1
1
1
.6
T=+5°C
T=0 °C
.4
.2
.0
.8
.6
.4
T=- 5 °C
(A)
0
10
20
30
40
50
60
70
80
time (min)
-
-
-
-
-
-
1.0
1.5
2.0
2.5
3.0
3.5
5,0
ÄH= -84.3± 3.5 kJ/mol
ÄS= 0.2± 0.05 kJ/mol
Ea=91.3KJ/mol
4
4
,5
,0
k
k
3,5
3
2
,0
,5
(B)
(
C)
2,0
-
3
-3
-3
-3
-3
3
.6x10
3.7x10
3.7x10
3.8x10
3.8x10
-
3
-3
-3
-3
-3
3
,6x10 3,7x10 3,7x10 3,8x10 3,8x10
/T (1/K)
1/T (1/K)
1
III
II
Fig. 12. (A) Kinetics of the of Fe AOOH formation by [Fe L(OAc)
formation of Fe AOOH.
2
], monitored by the I420 at different temperatures, (B) Arrhenius plot and (C) Eyring-Polanyi plot of the
III
II
IV
(
i) The oxidative advancements of LMn to LMn @O as well as
II
III
the LFe to LFe AOOH are strongly exothermic reactions. At
room temperature, T = 302 K, where typically our catalytic
à
reactions are performed, the entropic term is T
D
S = 302 -
ꢄ 0.3 ꢁ 91 kJ/mole; thus, the free energy change for this
à
à
à
reaction is
D
G =
for [Mn L(OAc) ], while
kJ/mole for [Fe L(OAc)
D
H – T
D
S = ꢂ55–91 = ꢂ146 ± 6 kJ/mole
II
à
à
2
D
Gv =
D
H – T
D
S = ꢂ144 ± 6
II
2
]. Thus we see that in thermodynamic
II
equilibrium terms, the overall evolution of [Mn L(OAc)
2
], has
]. This
difference is not sufficient to explain the huge differences
II
only slightly more negative
DG than the [Fe L(OAc)
2
II
II
observed in the TONs of the [Mn L(OAc)
2
] vs. the [Fe L
(
OAc) ] catalyst.
2
(
ii) Going to the kinetics, i.e. describing the events during the
out-of-equilibrium transient states, the activation energy
Scheme 1. Comparative thermodynamic reaction scheme for the oxidative evolu-
tion of Mn L(OAC)2 and Fe L(OAC)2 catalysts, according to the data of Table 4. The
II
II
II
barrier E
a
for the oxidative advancement of LMn to
IV
II
LMn @O is
E
a
= 58.6 ± 3.0 kJ/mol while
E
a
for LFe to
activation barrier is significantly higher for the Fe-complex than for the Mn-
à
III
complex, while the total free energy changes
D
G
are comparable.
LFe AOOH is E = 91.3 ± 3.0 kJ/mol. Thus the thermody-
a
namic data in Table 4 reveal that the oxidative advancement
of the Fe-catalyst faces a significantly higher activation
barrier than the Mn-catalyst
IV
II
and FeII to Fe AOOH for the Fe L
2
à à
III
II
Mn @O for the Mn L(OAC)
(OAC)
D
2
. On the other hand the equilibrium parameters
DG , DH ,
à
S , characteristic for the initial vs. final state, do not differ neither
E
aðLFeðOACÞ2Þ
ꢀ EaðLMnðOACÞ2Þ
in enthalpic nor entropic terms. This shows that molecular reorga-
nization events between initial vs. final states have the same
while the free-energy difference
II
II
energy cost in both Mn L(OAC)
2 2
and Fe L(OAC) catalysts. From
z
z
the redox-chemistry point of view i.e. using the fundamental equa-
D
G
ꢁ
DG
ðLFeðOACÞ2Þ
ðLMnðOACÞ2Þ
à
tion
D
G = ꢂn FE1/2, with E1/2 = redox potential of the equilibrium
à
à
is comparable in the Fe and Mn complexes. This thermodynamic
information can be visualized in Scheme 1.
Taken altogether the present data clarify that the main thermo-
dynamic barrier, ultimately determining the overall catalytic per-
formance, TONs/TOFs, of these homologous Mn and Fe catalysts
redox states, the comparable
D
G
(LFe(OAC)2)
ꢁ
DG
(LMn(OAC)2) implies
E1/2(LFe(OAC)2) ꢁ E1/2(LMn(OAC)2). Thus the redox potential between
the initial/final equilibrium states does not differentiate the Mn- vs
the Fe-complex. Instead the activation energy of the transient inter-
mediates is the parameters that makes the observed difference in
catalytic reaction rates.
II
is the activation energy for the transient intermediates i.e. Mn to

1
14
M. Papastergiou et al. / Journal of Catalysis 341 (2016) 104–115
H C CH3
3
C
CH3
N
HOO Fe
N
OH
N
CH3
C
H C CH3
H C CH3
3
3
C
H C
3
N
Mn O
N
HO
H C
N
IV
Fe =O
3
C
H C CH3
3
O
H C CH3
3
H C CH3
H C
3
3
C
CH3
C
N
Fe
N
N
X
X
X
N
OH
HO
H C
N
Mn
X
N
C CH3
3
C
H C
CH
3
3
H C CH3
+
3
+
H O
2
2
H O
2
2
II
II
Scheme 2. Catalytic reaction scheme for the oxidative evolution of Mn L(OAC)
2
and Fe L(OAC)
2
catalysts, according to our data presented herein.
II
A unified catalytic reaction mechanism: The thermodynamic reac-
tion scheme when incorporated to a pertinent catalytic cycle, see
and (ii) by enhancing heterolytic cleavage of Mn AOOH toward
the determinant Mn @O formation [27].
IV
Scheme 2, provides
follows.
a
unified physicochemical landscape as
For completeness, we reiterate that as shown many times by our
group [14–18] ammonium acetate has no effect on Fe-catalysts.
Ammonium acetate drives Mn-catalysis quickly and directly to
II
II
Addition of H
2
O
2
in Mn -catalysts produces Mn AOOH that
II
IV
IV
II
rapidly oxidizes Mn to Mn i.e. via a two-electron oxidation step
Mn @O i.e. Mn AOOH is not the rate-limiting intermediate but
II
IV
[
27]. In contrast, Fe is first oxidized to the intermediate low-spin
rather the Mn @O state. This is of great importance in oxidation
III
IV
IV
Fe AOOH state that further evolves to Fe @O [24]. Here it is
instructive to underline that, according to our EPR data, only the
Low-Spin Fe centers are catalytically active. As shown in Fig. 8B
the High-Spin-Fe complexes i.e. those Fe-complexes that fail to
catalysis, because the generation of the unique Mn @O oxidation
intermediate in conjunction with its low activation barrier is
responsible for the selective catalytic epoxidation of alkenes with
II
high TONs values observed herein and in analogous Mn -catalysts.
III
be converted to the Low-Spin Fe AOOH state do evolve also, how-
ever at much slower rates than the low-spin states. This superiority
of the Low-Spin vs. the High-Spin reaction path has been reported
previously for other non-heme [24] or heme [15,17] Fe-catalysts
also. Thus, in the case of the Fe-catalyst, practically the catalytic
activity is determined by the Low-Spin path, as shown herein.
Molecular catalytic events: Given that the catalytic profile of the
5
. Conclusion
Two sets of homologous Fe and Mn catalysts systematic [Fe L
II
II
II
II
(
2 2 2 2
Cl) ], [Mn L(Cl) ], [Fe L(OAc) ], [Mn L(OAc) ] have been synthe-
sized. A detailed comparative study of their catalytic oxidative per-
formance in tandem with EPR and Low-Temperature UV–vis has
been carried out. It is found that the Mn-catalysts consistently out-
competed the homologous Fe-catalysts i.e. TOFs (Mn) = 162 vs.
TOFs (Fe) = 16. The Fe-catalysts face a significantly higher activa-
tion barrier than the Mn-catalysts i.e. E
III
present iron catalysts is analogous to the non-heme-Fe (DPEIFe Cl)
explored by us recently [24], we consider that the catalytic oxida-
IV
tion mechanism should be similar one i.e. including not only Fe @O
Å
III
species but also OH formed by a homolytic cleavage of Fe -
II
a
(Fe L(OAC)
2
) = 91.3 kJ/-
IV
hydroperoxides [24]. Electrophilic attack of Fe @O species toward
II
mol ꢀ E
a 2
(Mn L(OAC) ) = 58.6 kJ/mole, while the free-energy
alkene could generate a radical [Fe-O-substrate] adduct which sub-
sequently, by intramolecular attack, may liberate epoxides as oxi-
dation products. Detection of allylic oxidation products indicates
à II à
II
difference
D
G (Fe L(OAC)
2
) ꢁ
D
G (Mn L(OAC)
2
) ꢁ ꢂ145 kJ/mole
did not make difference. A unified/consistent catalytic reaction
mechanism is discussed. We suggest that these mechanistic find-
ings bear immediate relevance to the catalytic performance of
other analogous non-heme Fe- vs. Mn-complexes. Taken altogether
the present data clarify that the main thermodynamic barrier, ulti-
mately determining the overall catalytic performance of these
homologous Mn- and Fe-catalysts is the activation energy for the
Å
IV
a first substrate attack by OH radicals followed by Fe @O associa-
II
2 2
tion. The Fe formed during catalysis is re-oxidized by H O [24].
On the other hand, the present Mn catalysts operate exclusively
via a 2-electron step forming Mn @O which is detected by both
II
IV
EPR and low temperature UV–vis. However, a key-parameter here
is ammonium acetate which assists -as additive-acting in two steps
II
IV
transient intermediates i.e. Mn to Mn @O for the Mn-catalysts
and Fe to Fe AOOH for the Fe-catalysts.
II
[
27] (i) by deprotonating H
2
O
2
thus promoting readily Mn AOOH
II
III

M. Papastergiou et al. / Journal of Catalysis 341 (2016) 104–115
115
[
21] W. Adam, C. Mock-Knoblauch, C.R. Saha-MÖller, M. Helderich, J. Am. Chem.
Soc. 122 (2000) 9685–9691.
Acknowledgments
[
22] D.P. Kessissoglou, X. Li, W.M. Butler, V.L. Pecoraro, Inorg. Chem. 26 (1987)
2487–2492.
This work was partially supported by the Russian Foundation
for Basic Research (grant no. 15-03-03057).
[23] K. Srinivasan, P. Michaud, J.K. Kochi, J. Am. Chem. Soc. 108 (1986) 2309–2320.
[
24] G. Bilis, P. Stathi, A. Mavrogiorgou, Y. Deligiannakis, M. Louloudi, Appl. Catal. A:
Gen. 470 (2014) 376–389.
Appendix A. Supplementary material
[25] S. Hong, K.D. Sutherlin, J. Park, E. Kwon, M.A. Siegler, E. Solomon, W. Nam,
Nature Commun. 5 (2014) (ar. no. 5440).
[
26] J. Chen, K.-B. Cho, Y.-M. Lee, Y.H. Kwon, W. Nam, Chem. Commun. 51 (2015)
3094–13097.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2016.06.017.
1
[27] A. Mavrogiorgou, M. Papastergiou, Y. Deligiannakis, M. Louloudi, J. Mol. Catal.
A 393 (2014) 8–17.
[
[
[
28] S. Stoll, A. Schweiger, J. Magn. Reson. 178 (2006) 42–55.
References
29] Ag. Stamatis, Ch. Vartzouma, M. Louloudi, Catal. Commun. 12 (2011) 475–479.
30] Ag. Stamatis, P. Doutsi, Ch. Vartzouma, K.C. Christoforidis, Y. Deligiannakis, M.
Louloudi, J. Mol. Catal. A 297 (2009) 44–53.
[
[
[
1] R.A. Sheldon, Top. Curr. Chem. 164 (1993) 21.
2] K.A. Jorgenson, Chem. Rev. 89 (1989) 431–458.
3] M. Islam, P. Mondal, S. Mokherjee, M. Mobarak, A.S. Roy, S. Mondal, S. Sarkar, J.
Chem. Technol. Biotechnol. 85 (2010) 460–470.
[31] G.B. Shul’pin, J. Mol. Catal. A 189 (2002) 39–66.
[32] N.C. Anastasiadis, G. Bilis, J.C. Plakatouras, C.P. Raptopoulou, V. Psycharis, C.
Beavers, S.J. Teat, M. Louloudi, S.P. Perlepes, Polyhedron 64 (2013) 189–202.
[33] W. Nam, Acc. Chem. Res. 48 (2015) 2415–2423.
[4] C.L. Hill, C.M. Prosser-McCartha, Coord. Chem. Rev. 143 (1995) 407–455.
[5] K. Sato, M. Aoki, R. Noyori, Science 281 (1998) 1646–1647.
[6] E.J. Larson, V.L. Pecoraro, in: V.L. Pecoraro (Ed.), Manganese Redox Enzymes,
Verlag-Chemie, New York, 1992, p. 1.
[34] G. Yin, Acc. Chem. Res. 46 (2013) 483–492.
[35] X. Wu, M.S. Seo, K.M. Davis, Y.-M. Lee, J. Chen, K.-B. Cho, H. Yoon, Y.J. Park, S.
Fukuzumi, Y.N. Pushkar, W. Nam, J. Am. Chem. Soc. 135 (2013) 6388–6391.
[36] M.J. Camenzind, F.J. Hollander, C.L. Hill, Inorg. Chem. 22 (1983) 3776–3784.
[37] B. Gaffney, H.J. Silverstone, in: L.J. Berliner, J. Reuden (Eds.), Biological
Magnetic Resonance, Plenum Press, New York, 1993, pp. 1–57.
[38] C. Duboc, Marie-Noelle Collomb, F. Neese, Appl. Magn. Reson. 37 (2010) 229–
245.
[39] C.P.S. Taylor, Bioch. Bioph. Acta 491 (1977) 137–149.
[40] F.G. Gelalcha, G. Anilkumar, M.K. Tse, A. Br Ückner, M. Beller, Chem. Eur. J. 14
(2008) 7687–7698.
[41] L. Duelund, H. Toftlund, Spectrochim. Acta A 26 (2000) 331–340.
[42] K. Chen, M. Costas, J. Kim, A.K. Tripton, L. Que, J. Am. Chem. Soc. 124 (2002)
3026–3035.
[43] R. Giovannetti, in: J. Uddin (Ed.), Macro to Nano Spectroscopy, InTech, 2012,
pp. 87–108.
[
[
[
7] B. Meunier, Biomimetic Oxidations Catalyzed by Transition Metal Complexes,
Imperial College Press, London, 2000.
8] A. Berkessel, M. Franenkron, T. Schwenkreis, A. Steinwetz, J. Mol. Catal A –
Chem. 117 (1997) 339.
9] J. Bjernemose, A. Hazell, C.J. McKenzie, M.F. Mahon, L.P. Nielsen, P.R. Raithby, O.
Simonsen, H. Toftlund, J.A. Wolny, Polyhedron 22 (2003) 875–885.
[
[
10] J.P. Wikstrom, A.Y. Nazarenko, W.M. Reiff, E.V. Rybak-Akimova, Inorg. Chim.
Acta 360 (2007) 3733–3740.
11] W.A. Chomitz, S.G. Manasian, A.D. Sutton, J. Arnold, Inorg. Chem. 46 (2007)
7
199–7209.
[
[
12] E.R. Milaeva, Curr. Top. Med. Chem. 11 (2011) 2703–2713.
13] E.R. Milaeva, O.A. Gerasimova, A.L. Maximov, E.A. Ivanova, E.A. Karachanov, N.
Hadjiliadis, M. Louloudi, Catal. Commun. 8 (2007) 2069–2073.
14] K.C. Christoforidis, M. Louloudi, E.R. Milaeva, Y. Sanakis, Y. Deligiannakis, Mol.
Phys. 105 (2007) 2185–2194.
[
[
[
[
[
[
[44] G.G. Martirosyan, A.S. Azizyan, T.S. Kurtikyan, P.C. Fond, Inorg. Chem. 45
(2006) 4079–4087.
15] K.C. Christoforidis, M. Louloudi, E.R. Milaeva, Y. Deligiannakis, J. Catal. 270
[45] K.J. Laidler, Chemical Kinetics, third ed., Harper & Row, New York, 1987.
[46] 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 Jr., Y. Zang, H. Kooijman, A.L. Spek,
L. Que Jr., B.L. Feringa, Inorg. Chem. 38 (1999) 1929–1936.
[47] R.A. Ghiladi, K.D. Karlin, Inorg. Chem. 41 (2002) 2400–2407.
[48] G.C. Silva, G.L. Parrilha, N.M.F. Carvahlo, V. Drago, C. Fernandes, A. Horn Jr., O.A.
C. Antunes, Catal. Today 133–135 (2008) 684–688.
[49] G. Roefles, V. Vrajmasu, K. Chen, R.Y.N. Ho, J. Rohde, C. Zondervan, R.M. La
Crois, E.P. Schudde, M. Lutz, A.L. Spek, R. Hage, B.L. Feringa, E. Munk, L. Que,
Inorg. Chem. 42 (2003) 2639–2653.
(
2010) 153–162.
16] K.C. Christoforidis, M. Louloudi, Y. Deligiannakis, Appl. Catal. B: Environ. 95
2010) 297–302.
17] K.C. Christoforidis, M. Louloudi, Y. Deligiannakis, Chem. Phys. Lett. 494 (2010)
89–294.
(
2
18] K.C. Christoforidis, E. Serestatidou, M. Louloudi, I.K. Konstantinou, E.R. Milaeva,
Y. Deligiannakis, Appl. Catal. B: Environ. 101 (2011) 417–424.
19] E.R. Milaeva, D.B. Shpakovsky, Y.A. Gracheva, S.I. Orlova, V.V. Maduar, B.N.
Tarasevich, N.N. Meleshonkova, L.G. Dubova, E.F. Shevtsova, Dalton Trans. 42
(
2013) 6817–6828.
[50] G. Billis, K.C. Christoforidis, Y. Deligiannakis, M. Louloudi, Catal. Today 157
(2010) 101–106.
[
20] J.R. Pilbrow, Transition Metal Electron Paramagnetic Resonance, Oxford
Science Publications, Clarendon Press, Oxford, England, 1991.