Oxygen atom transfer mediated by an iron(IV)/iron(II) macrocyclic complex containing pyridine and tertiary amine donors

Article abstract of DOI:10.1016/j.jinorgbio.2012.05.004

A new non-heme iron model complex containing a high-spin iron(II) complexed with N-methylated pyridine-containing macrocycle was synthesized and crystallographically characterized. The complex generates peroxo- and high-valent iron-oxo intermediates in reactions with tert-butylhydroperoxide and isopropyl 2-iodoxybenzoate, respectively, allowing to gain insight into the formation and reactivity of enzyme-like intermediates related to biological oxygen activation. The formation and reactivity of these intermediate species were investigated by the stopped-flow methodology. The mechanism of oxygen transfer to organic substrates involving reaction of oxoiron(IV) intermediate was elucidated on the basis of spectroscopic and kinetic data. Incorporation of a pyridine ring into the macrocycle increased the reactivity of the Fe ^IVO intermediates in comparison with polyamine tetraaza macrocyclic complexes: ferryl (Fe^IVO) species derived from 3 demonstrated electrophilic reactivity in transferring an oxygen atom to substituted triarylphosphines and to olefins (such as cyclooctene). However, iron(III) alkylperoxo intermediate was unreactive with cyclooctene.

Full text of DOI:10.1016/j.jinorgbio.2012.05.004

Oxygen Atom Transfer Mediated by an Iron(IV)/Fe(II) Macrocyclic Complex
Containing Pyridine and Tertiary Amine Donors
Wanhua Ye, Richard J. Staples, Elena V. Rybak-Akimova
PII:
DOI:
Reference:
S0162-0134(12)00157-2
doi: 10.1016/j.jinorgbio.2012.05.004
JIB 9080
To appear in:
Journal of Inorganic Biochemistry
Received date:
Revised date:
Accepted date:
6 November 2011
8 May 2012
10 May 2012
Please cite this article as: Wanhua Ye, Richard J. Staples, Elena V. Rybak-Akimova,
Oxygen Atom Transfer Mediated by an Iron(IV)/Fe(II) Macrocyclic Complex Contain-
ing Pyridine and Tertiary Amine Donors, Journal of Inorganic Biochemistry (2012), doi:
1
0.1016/j.jinorgbio.2012.05.004
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

ACCEPTED MANUSCRIPT
Oxygen Atom Transfer Mediated by an Iron(IV)/Fe(II) Macrocyclic Complex Containing
Pyridine and Tertiary Amine Donors
a
b
*a
Wanhua Ye , Richard J. Staples , Elena V. Rybak-Akimova
a
Department of Chemistry, Tufts University, Medford, MA 02155. FAX: 1-6176273443;
Email: erybakak@tufts.edu
b
Department of Chemistry and Chemical Biology, Michigan State University, East Lansing,
MI 48824
Corresponding author: Elena V. Rybak-Akimova. FAX: 1-6176273443; Email:
erybakak@tufts.edu
1

ACCEPTED MANUSCRIPT
Abstract
A new non-heme iron model complex containing a high-spin iron(II) complexed with N-
methylated pyridine-containing macrocycle was synthesized and crystallographically
characterized. The complex generates peroxo- and high-valent iron-oxo intermediates in
reactions with tert-butylhydroperoxide and isopropyl 2-iodoxybenzoate, respectively, allowing to
gain insight into the formation and reactivity of enzyme-like intermediates related to biological
oxygen activation. The formation and reactivity of these intermediate species was investigated
by the stopped-flow methodology. The mechanism of oxygen transfer to organic substrates
involving reaction of oxoiron(IV) intermediate was elucidated on the basis of spectroscopic and
kinetic data. Incorporation of a pyridine ring into the macrocycle increased the reactivity of the
IV
Fe =O intermediates in comparison with polyamine tetraaza macrocyclic complexes: ferryl
IV
(
Fe =O) species derived from 3 demonstrated electrophilic reactivity in transferring an oxygen
atom to substituted triarylphosphines and to olefins (such as cyclooctene). However, iron(III)
alkylperoxo intermediate was unreactive with cyclooctene.
Keywords
Non-heme iron, macrocyclic complex, X-ray crystal structure, oxidation, olefin epoxidation,
kinetics, reaction mechanisms
2

ACCEPTED MANUSCRIPT
1
. Introduction
Heme[1, 2] and non-heme[3, 4] iron-containing oxygenase enzymes have invited
interest in the enzyme-mediated oxygen atom transfer reaction mechanism. Physical
methods, structural biology, and mechanistic enzymology all contribute to the
understanding of the active iron center in these enzymes. Peroxoiron(III) species are
III
often cited as mechanistically important intermediates of iron enzymes. Fe –OOH
species has been characterized and proposed as an intermediate of dioxygen activation in
heme cytochrome P450 enzyme and non-heme bimolecular such as bleomycin (BLM) [5-
7
] and superoxide reductase (SOR) [8-13]. On the other hand, oxygen activation with
lipoxygenase (LO) proceeds through a well-characterized end-on-coordinated
alkylperoxoiron(III) species [14]. Several biomimetic models of the active site of these
enzymes have been prepared and mechanistically important intermediates of the catalytic
cycles have been spectroscopically characterized [3, 15, 16]. Synthetic model studies
have demonstrated that alkylperoxoiron(III) species [17-24] are more stable compared to
hydroperoxoiron(III) [25-39], yet the reactivity pathways of these peroxo complexes are
often similar.
Further investigation of the peroxoiron(III) intermediates have shown that O–O bond
cleavage can occur, yielding high-valent oxoiron(IV) species either via heterolysis or
-
IV
+
homolysis. Heterolytic two electron cleavage of the O–O bond leads to (P ˙)Fe =O (is
2
-
V
+
in resonance with (P )Fe =O ) in heme such as P450 enzymes[1], whereas one electron
IV
homolytic cleavage leads to formation of Fe =O species in non-heme systems [40, 41].
3

ACCEPTED MANUSCRIPT
Nam et al. have recently provided the first direct spectroscopic evidence that non-
heme alkylperoxoiron(III) species are not capable of oxygenating organic substrates such
as thioanisole, olefin, cyclooctanol, triphenylmethane; instead, the high-valent
oxoiron(IV) complex is the active oxidant [40, 42]. However, a high-spin iron(III)
hydroperoxo complex is reactive in oxidations of alkylaromatic compounds with weak
CH bonds [43]. The reactivity of iron(III) peroxo species and iron(IV)-oxo species with
organic substrates is currently under active investigation in a number of biological [44, 45]
and biomimetic [3, 15, 46-51] systems. Isolating and studying the oxoiron(IV) species
from peroxoiron(III) is often difficult, and distinguishing between the reactivity of
Fe(III)-peroxo vs. Fe(IV)-oxo intermediates in these systems is challenging. On the other
hand, such transient iron(IV)-oxo species can be generated by using oxygen atom donors
such as iodosylbenzene (PhIO) [52-57], m-chloroperoxybenzoic acid (m-CPBA) [58, 59],
peracetic acid (CH COOOH) [54, 60], potassium monopersulfate (KHSO ) [54], NaOX
3
5
(
X = Cl or Br) [61] or O [62, 63]. This has opened an opportunity to understand the
3
reactivity of the key intermediates in oxygen activation and oxygen atom transfer
reactions of iron enzymes and biomimetic complexes, thus clarifying the mechanisms of
these processes.
Olefin epoxidation is often studied as an example of a practically important oxygen
atom transfer process [64]. Biomimetic non-heme iron complexes that participate in
stoichiometric or catalytic epoxidations include aminopyridine complexes (e.g. tris-
picolylamine, TPA) and tetradentate azamacrocycles (e.g. [14]ane]N4 and [15]aneN4)
4

ACCEPTED MANUSCRIPT
[
40]. Interestingly, both classes of ligands support ferryl intermediates, but only the
IV
former complex, (TPA)Fe =O, directly transfers its oxygen atom to cyclooctene [60],
IV
while Fe =O intermediates supported by polyamine macrocycles do not react with
olefins directly, but facilitate epoxidations in the presence of excess PhIO [56]. Our
laboratory has utilized mononuclear non-heme iron(II) complexes of pyridine-containing
macrocycles to investigate their reactivity towards oxygen transfer reactions [65].
Complex 1 showed catalytic reactivity in olefin epoxidation using H O as oxidant,
2
2
whereas complex 2 was less reactive [65]. Moreover, we observed the ligand in 1 and 2 is
susceptible to oxidation, limiting the catalyst’s efficiency. We thus synthesized an iron(II)
complex with N-methylated pyridine-containing tetra-azamacrocycle L1 (3,7,11-
trimethyl-3,7,11,17-tetra-azabicyclo[11.3.1]heptadeca-1(17),13,15-triene) to prevent the
ligand oxidation (Fig. 1).
Herein we report the syntheses, structures, and properties of the iron(II) complexes
with ligand L1. tert-Butyl hydroperoxide (TBHP) and hydrogen peroxide (H O ), are
2
2
used as terminal oxidants in the stoichiometric oxidation of substrates (triaryl phosphines
or olefins) and in catalytic or stoichiometric olefin epoxidation studies. TBHP and
isopropyl 2-iodoxybenzoate were used to generate alkylperoxoiron(III) 3a and
oxoiron(IV) 3b species, respectively (Scheme 1). The formation and reactivity of these
intermediate species was investigated by the stopped-flow methodology. The mechanism
of oxygen transfer to organic substrates involving reaction of oxoiron(IV) intermediate
was elucidated on the basis of spectroscopic and kinetic data. Incorporation of a pyridine
IV
ring into the macrocycle increased the reactivity of the Fe =O intermediates in
5

ACCEPTED MANUSCRIPT
comparison with polyamine tetraaza macrocyclic complexes: ferryl(IV) species derived
from 3 transfer an oxygen atom to olefins (such as cyclooctene). However, iron(III)
alkylperoxo intermediate was unreactive with cyclooctene.
2
. Experimental
2
.1. General considerations
All reagents were obtained from commercial vendors and used without further
purification unless noted otherwise. 2,6-Pyridinedicarboxaldehyde [66] and isopropyl 2-
iodoxybenzoate [67] were prepared following a published procedure. The tetradentate
ligand L1 was synthesized according to previously published procedures [68, 69]. All
manipulations of air-sensitive materials were carried out in the MBraun glove box under
an atmosphere of additionally purified from oxygen and moisture ultra-high-purity argon.
18
18
16
2
18
18
H O (90% O-enriched, 2% solution in H O) and H O (98% O-enriched) were
2
2
2
purchased from Cambridge Isotope Laboratories Inc. Fe(OTf) ·2CH CN (OTf = CF SO ˉ)
2
3
3
3
was prepared from FeCl and CF SO SiMe in CH CN following an unpublished
2
3
3
3
3
procedure provided by Dr. Miquel Costas and Prof. Lawrence Que. m-CPBA was
recrystallized from dichloromethane prior to use. The olefin substrates were purified from
traces of alkylperoxides by passing through a short column of activated alumina and
stored in a glove box. All concentrations intended for the stopped-flow experiments were
before mixing, unless otherwise noted.
2
.2. Physical methods
6

ACCEPTED MANUSCRIPT
Room temperature UV-visible (UV–vis) spectra were recorded on a Jasco V-570
spectrophotometer. Low-temperature UV–vis spectra were recorded at –40 ˚C using AP
disposable dip probe UV (part No: FA-CTI01-AP2D) purchased from C Technologies,
Inc. ESI-MS was performed on a Thermo Finnigan LCQ quadruple field mass
spectrometer with electrospray ionization in the positive ion detection mode. NMR
spectra were acquired in CDCl at 300K on a Bruker DPX 300 MHz spectrometer;
3
chemical shifts in δ are reported versus tetramethylsilane. The EPR spectra were recorded
on a Bruker EMX EPR spectrometer (9.41 GHz) in CH CN at 120 K. GC–MS analyses
3
were performed using a Shimadzu QP5050A mass spectrometer connected to Shimadzu
®
GC-17A gas chromatograph equipped with Rtx -XLB column (length, 30 m; i.d.
0
.25 mm; film thickness, 0.25 μm). Helium at a flow rate of 45 cm/s was used as the
carrier gas. Injections were made in the splitless mode using an initial column
temperature of 50 ˚C. The temperature was raised at 15 ˚C/min until 320 ˚C. Full scan
was performed using ionization energy of 70 eV. Elemental analyses were performed by
Schwarzkopf Microanalytical Laboratory, Inc. (Woodside, NY).
II
2
.3. Synthesis of [Fe (L1)](OTf) (3)
2
Fe(OTf) ·2CH CN (440 mg, 1.01 mmol) dissolved in 2.0 mL of dry acetonitrile was
2
3
added to a solution of L1 (309 mg, 1.12 mmol) in 2.0 mL of CH CN with two drops of
3
dry methanol under an argon atmosphere. When mixing, the solution turned red. A pale
yellow solid was obtained after 3 days by crushing the mixture into diethyl ether. Light
yellow crystals suitable for X-ray diffraction analysis were obtained by dissolving the
solid in CH CN followed by adding solution dropwise into diethyl ether. Yield: 64 mg
3
7

ACCEPTED MANUSCRIPT
2+
(
43%). ESI(+)–MS (CH CN) m/z: [Fe(L1)(CN CN)] calcd for 186.6; Found, 186.6;
3
3
+
{
[Fe(L1)](OTf)} calcd for 481.1; Found, 481.3. Anal. Calcd for C H N FeO F S : C,
18 28 4 6 6 2
3
4.28; H, 4.48; N, 8.89; Fe, 8.87. Found: C, 34.52; H, 4.43; N, 8.96; Fe, 8.46.
II
Light green crystals of [Fe (L1)](OTf)Cl (4) were obtained according to procedure
prepared for complex 3 but using a mixture of Fe(OTf) ·2CH CN and FeCl (1:1) instead
2
3
2
+
of pure Fe(OTf) ·2CH CN. ESI(+)–MS (CH CN) m/z: {[Fe(L1)](Cl)} , calcd for 367.1;
2
3
3
Found, 367.4.
2
.4. X-ray crystallography
Samples were selected and mounted on a glass fiber in Paratone. The data collections
were carried out at a sample temperature of 173 K on a Bruker AXS platform three-circle
goniometer with a CCD detector. The first 100 frames were then re-collected at the end
of the data collection for monitoring decay. The data were processed utilizing the
program SMART [70] and SAINTPLUS [71] supplied by Bruker AXS. The structures
were solved by direct methods (SHELXTL v5.1, Bruker AXS) [72-74] in conjunction
with standard difference Fourier techniques. All non-hydrogen atoms were refined with
anisotropic displacement parameters. For complex 4, one chloride ion was found to be
disordered. Similarity restraints were applied using the SADI command. Structural
parameters for 3 and 4 are given in Table 1. The data were deposited in Cambridge
Crystallographic database: CCDC 818640 and 818641.
8

ACCEPTED MANUSCRIPT
2
.5. Cyclic voltammetry
Cyclic voltammetry experiments were carried out using a CH Instruments
electrochemical analysis system (Model 830) using a conventional three-electrode
configuration. A glassy carbon electrode (3 mm diameter) was used as working electrode
and platinum wires as the counter and reference electrodes. Measurements were
performed with 2 mM analyte in CH CN at ambient temperature under argon using 0.1 M
3
tetra-n-butylammonium hexafluorophosphate as the supporting electrolyte. The range of
-1
+
potentials was scanned at a scan rate of 100 mV s . The ferrocenium/ferrocene (Fc /Fc)
reference couple was used as an internal standard.
2
.6. Measurements of effective magnetic moments in solution
Measurements by Evans method were performed on a Bruker DPX-300 spectrometer
at 25 ˚C [75-77]. A solution of complex 1 (1 mM) in a 0.5 mL of CD CN (with 1%
3
tetramethylsilane, TMS) was prepared under an argon atmosphere. The effective
-1
magnetic moment ( ) was calculated from  = 8χ M T, where as the χ (cm g ) is
eff
eff
g
w
g
3
the corrected molar susceptibility derived from χ = 3δν /4πν CM + χ [78]. δν is the
g
o
w
o
shift in frequency (Hz) from the value found for the pure solvent, C is the concentration
-3
-1
of the complex (mol cm ), M is the molecular weight of the complex (g mol ), ν is the
w
o
operating frequency of the NMR spectrometer (Hz), and χ is the mass susceptibility of
o
-6
3
-1
the pure solvent (–0.682 × 10 cm g for acetonitrile). The 4π/3 shape factor is for a
cylindrical sample in a superconducting magnet. Then, the number of unpaired electrons
2
2
per molecule, n, can be easily derived from the magnetic moment using  = g S(S + 1),
9

ACCEPTED MANUSCRIPT
where S is the electronic spin and g the Landé factor. This can be further simplified into
n(n + 2) for g = 2 [79].
2
.7. Catalysis experiments
Solutions of the iron(II) complexes 3 and 4 were prepared in a glove box under an
argon atmosphere (0.8 μmol, 5% catalyst) in 1.0 mL of CH CN. Cyclooctene (25 μL,
3
0
.016 mmol) was added to the catalyst solution by a syringe. To this mixture was added
H O (30 wt.% in H O) or TBHP (70 wt.% in H O) (30 μL, 0.024 mmol) all at once. The
2
2
2
2
reaction was carried out at room temperature under argon atmosphere for 5 min and
1
8
analyzed by GC–MS. The ratio of catalyst/substrate/H O was 1/20/30. In the O-
2
2
1
2
8
18
labeling experiments, H O or H O (300 μL, 0.24 mmol) was added to the catalyst
2
2
solution following the method described as above.
2
.8. Stopped-flow kinetics
Stopped-flow spectrophotometry was performed on a TgK Scientific (formerly, Hi-
Tech Scientific, UK) SF-61DX2 cryogenic double-mixing stopped-flow system [80]
equipped with J&M TIDAS diode array spectrophotometer (MCS UVNIR 500-3, 200–
1
024 nm) [81]. Low temperatures were maintained through the use of a liquid-nitrogen-
cooled ethanol bath equipped with a cryo unit. All flow lines of the instrument were
extensively washed with oxygen-free, argon-saturated anhydrous CH CN before charging
3
the driving syringes with reactant solutions. CH CN solutions of the reagents were
3
prepared in a glove box filled with argon and placed in Hamilton gastight syringes.
TM
Kinetic Studio
[82] and J&M TIDAS software package were used to control the
1
0

ACCEPTED MANUSCRIPT
TM
instrument and collect data. The raw kinetic data were treated with Kinetic Studio
software from TgK Scientific, and with Spectfit/32 Global Analysis System software
from Spectrum Software Associates [83].
3
. Results and discussion
Iron(II) complexes with the pyridine-containing macrocycle L1 were prepared from
the appropriate iron(II) salt and the free ligand, and characterized by elemental analysis,
mass spectrometry, UV–vis spectrophotometry, and structural determination by single-
crystal X-ray diffraction. Crystallographic information was particularly revealing for
comparing the complexes of N-alkylated macrocycle with other pyridine-azamacrocyclic
(PyMAC) complexes prepared in our group [65, 84, 85].
3
.1. X-ray crystallographic analysis
X-ray crystallographic results demonstrated that 3 had an iron(II) ion in a distorted
octahedral environment and 4 adopted a square-pyramidal geometry (Fig. 2, 3). The
iron(II) center of 3 was essentially four-coordinate, with four nitrogen donors from the
‘folded’ macrocyclic ligand and two distant oxygen donors from two different weekly
bound triflate ions in a relative cis orientation completing the coordination sphere. The
crystal structure of 3 showed this pyridine-containing macrocycle was folded about the
N(2) to N(4) axis, and the apex was occupied by N(3) and located almost perpendicular
to the pseudo-planar atom set defined by the metal ion and three remaining nitrogens of
the macrocyclic amine. This conformation allowed easier N inversion at N(3), as
observed. The coordination sphere of iron in the crystal structure of 4 was also completed
1
1

ACCEPTED MANUSCRIPT
by a ‘folded’ structure, defined by the four nitrogen donors from the ligand and one
chloride anion in the apical site which occupied in a trans position to the pyridine N atom,
leaving one triflate as a counter ion. Selected bond lengths and angles were listed in
Table 2. For 3 and 4, the values of the N2–Fe1–N4 angles (153.12 (9)˚ and 152.58 (8)˚),
N1–Fe1–N3 angles (100.09 (9)˚ and 101.72 (8)˚) were observed, respectively. This
geometry has also been found in nickel [69, 86] and copper analogue of 3 and 4 [87].
The Fe–Otriflate bond lengths in 3 were 2.03(2) and 2.24(2) Å and the Fe–Cl bond
length in 4 was 2.29(9) Å. The Fe–N bond length in 3 and 4, which ranged from (2.07(2)
to 2.27(2) Å) and (2.07(2) to 2.20(2) Å), respectively, were in the range expected for
typical high-spin iron(II) complexes [88, 89], with the shortest bond length to the
pyridine nitrogens 2.07(2) Å in both cases. The Fe–N separation was similar to the
reported high-spin 2 (2.06–2.18 Å) [65].
For comparison, the Fe–N bond length (1.89–2.04 Å) for low-spin 2 has been reported
previously[65], which was shorter than the iron-nitrogen bond lengths in high-spin 3 and
4
. Low spin complex 1, bearing the ligand with several primary amine donors in the
macrocycle, adopted an octahedral geometry with coordinated CH CN in the axial
3
position, whose crystal-field stabilization energy favored low-spin octahedral complexes
6
for a d configuration. Nevertheless, tetradentate macrocycles of iron(II), nickel(II) and
copper(II) complexes [18, 90-93] were often coordinated to additional monodentate
ˉ
ˉ
ˉ
ˉ
ˉ
ˉ
2
anions (Cl , Br , I , N , NSC or NO ), displaying a strong structural preference for a
3
square-pyramidal geometry. Changes of the coordination geometry of the
1
2

ACCEPTED MANUSCRIPT
metallomacrocycles from the six-coordinate octahedral to a five-coordinate square-
pyramidal often resulted in changes from the low-spin to a high-spin configuration of the
central metal ion, evidenced by the appropriate changes in their magnetic and
spectroscopic properties [18, 94, 95].
3
3
.2. Properties of iron(II) complexes with L1
.2.1. Magnetic properties of iron(II) complexes in solution
Magnetic susceptibility measurements using Evans’ method was performed on 3 and 4
in CD CN. The effective magnetic moment values obtained ( = 5.2 B.M.) for
3
eff
6
complexes 3 and 4 were in the range expected for high-spin d electronic configuration of
the iron(II) centers in mononuclear complexes. The five-coordinate 2 had a typical high-
spin configuration (eff = 5.1 B.M.), whereas the low-spin 1 definitely showed no
splitting of TMS peak (Table 3) [65]. The N-methylation of secondary aminogroups
caused the iron(II) complex with ligand L1 to adopt high-spin configuration, while the
structurally similar complex 1 with the macrocycle containing non-alkylated secondary
aminogroups was low-spin. Meyerstein et al. demonstrated that transforming all the
nitrogen donor atoms into tertiary nitrogens elongated the Metal–N bonds.[95] This
structural effect has also been observed in 3, 4 and other reported iron(II) complexes with
N-methyl substituted TPA and cyclam-acetate [18, 94, 96]. These iron(II) complexes with
N-methylated ligand all favor the high-spin configuration of the metal center.
1
3

ACCEPTED MANUSCRIPT
3
.2.2. Cyclic voltammetry (CV)
Electrochemical characteristics of 3 and 4 derived from cyclic voltammetry are shown
in Fig. 4. The CV scan of 3 in CH CN showed a well-formed anodic peak E = 0.91 V
3
pa
III
II
assigned to Fe /Fe couple. During the cathodic reduction scan two responses were
observed: 0.715 V and 0.219 V. The observation of two cathodic processes implies the
change in the coordination sphere of iron after oxidation of the iron(II) center; formation
of the iron(III) oxo-bridged dimers is a likely possibility. The asymmetry in the wave’s
shape is also indicative of the deposition upon oxidation of conducting species at the
electrode. Other high-spin iron(II) complexes [18, 24] were also oxidized at high
potential, but the oxidations were irreversible in these cases. The chloride-coordinated
III
II
complex 4, however, displayed a quasi-reversible metal-centered redox process Fe /Fe
+
(
E = 0.41 V vs Fc /Fc, ΔE = 169 mV). The shift on oxidation potential of 4 compared
1
/2
p
to 3 is consistent with relative stabilization of the Fe(III) upon coordination of the
negatively charged chloride ligand. Coordination of chloride also protects the iron(III)
species from iorreversible hydrolysis and formation of oxo-bridged dimers or oligomers.
The previously reported high-spin 2 (E = 0.58–0.56 V with ΔE = 260 mV and ΔE =
1
/2
p1
p2
5
40 mV) also exhibited a quasi-reversible redox process. Furthermore, 2 differed from
low-spin 1, as the latter complex showed reversible one electron transfer process at 0.40
+
V vs Fc /Fc (non-coordinating counterions, perchlorate or triflate, were present in 1 and 2)
III
II
[
65]. Therefore, coordination sphere effects on Fe /Fe redox potential can be identified
by CV scan and reveal the relative reducing strength of iron(II) macrocyclic complexes in
a series of PyMACs.
1
4

ACCEPTED MANUSCRIPT
3
.2.3. Catalytic and stoichiometric olefin epoxidation
cis-Cyclooctene is commonly used as a suitable substrate for catalyst screening in
epoxidations with H O in the presence of non-heme iron catalysts [79]. Catalytic
2
2
epoxidation of cyclooctene with complexes 3 and 4 using H O and TBHP as oxidants
2
2
were examined to explore the effectiveness of the new complexes compared to that of
complex 1 and 2. Table 4 showed epoxide yields and turnover numbers (TON) after 5
min of reaction, carried out under argon atmosphere at room temperature with 1–4 as
catalysts.
Complex 3 did not efficiently catalyze olefin epoxidation, and acted as a
stoichiometric reagent instead, resulting in only 7% yield of epoxide (1 TON) using H O
2
2
as the oxidant. Other identified side products include diol (ca. 0.25% with respect to
1
8
cyclooctene) and allylic alcohol (ca. 0.25%). Incorporation of O into the epoxide and
1
8
one O into cis-diol were observed when the reaction was carried out in the presence of
18
16
18
H O . O-epoxide and trace of O of cis-diol were detected when the reaction was
2
2
1
8
carried out in the presence of H O and H O. These observations suggested that the
2
2
2
oxygen of epoxide and one oxygen atom of cis-diol were exclusively derived from H O .
2
2
One oxygen atom source of diol arose from H O. In a control experiment, diol was not
2
1
8
obtained from the epoxide and water (in the presence of H O or H O), indicating that
2
2
the incorporation of oxygen into diol only took place during the catalytic oxidation of the
olefin. Neither 2 nor 3 exhibited a reversible CV wave, which indicated that the small
turnover number might be related to poor stability of the oxidized forms of these
1
5

ACCEPTED MANUSCRIPT
complexes. In contrast, low-spin 1, which exhibited a classic redox wave in the cyclic
voltammogram, afforded about 6 TON of epoxide [65]. This suggested that the high-spin
iron(II) PyMAC complexes did not efficiently catalyze epoxidation, whereas the low-spin
ones were more reactive. It appeared that the weak-field ligands that form high-spin
iron(II) complexes did not sufficiently stabilize the oxidized forms of the catalysts (e.g.
Fe (III) species) with respect to partial or complete hydrolysis. Indeed, the loss of iron
was observed after several redox cycles with complex 3. Coordinatively saturated
complex 4 had no activity in olefin epoxidation using H O as the oxidant, presumably
2
2
because no vacant sites were available for initial coordination of the oxidant. Somewhat
higher activity of 3 compared to 4 underscored the importance of the vacant or labile site
at the iron centers in hydrogen peroxide activation, and highlights the critical role of the
coordination sphere of iron that controlled the catalytic activity of PyMAC complexes.
t-Butylhydroperoxide (TBHP) has also been examined as an oxidant in epoxidation
with 3, and it afforded the same product yield as the one obtained with H O as the
2
2
oxidant (7% epoxide with respect to cyclooctene). This led us to propose that similar
activity in both reactions may be attributed to the same intermediate responsible for
olefin epoxidation. In search of transient iron-based intermediates, reactions of complex 3
with oxidants were examined by in situ spectrophotometry at low temperature. In these
exploratory studies, low-temperature spectra were acquired with the dip probe connected
to a regular spectrophotometer; in detailed mechanistic studies, low-temperature stopped-
flow measurements allowed us to acquire accurate time-resolved spectra.
1
6

ACCEPTED MANUSCRIPT
3
.2.4. Reactions of complex 3 with oxidants
In order to further understand the details of stoichiometric or catalytic oxidations with
iron PyMAC’s, various oxidants have been examined in reactions with catalytically
active complex 3, generating putative intermediates in order to further understand the
details of the catalytic mechanisms. Complex 3 in CH CN gave a pale yellow solution.
3
The reaction of 3 with TBHP in CH CN at –40 ˚C produced a deep purple intermediate
3
which persisted for several hours at –40 ˚C, but quickly turned to yellow at room
temperature. The visible spectrum of this intermediate was consistent with an iron(III)-
alkylperoxo species [18, 20-22, 97-101]; somewhat lower observed extinction coefficient
may be explained by a less-than-quantitative yield of this intermediate. Treatment of 3
with H O , m-CPBA or isopropyl 2-iodoxybenzoate at –40 ˚C in CH CN generated a
2
2
3
transient brownish yellow solution, which also turned to yellow species upon warming up
to room temperature. Spectral changes included the growth of the weak near-IR
absorption band in the ca. 700–800 nm region, typical of the iron(IV)-oxo intermediates
[
21, 23, 97, 100, 102, 103]. Low-temperature UV–vis spectra for these transient species
are shown in Fig. 5. We then investigated the formation of these transient species and
their corresponding reactivity towards oxygen transfer reactions, using stopped-flow
methodology.
3
.3. Generation and reactivity of high-spin alkylperoxoiron(III) intermediates
In the stopped-flow experiment, treatment of 3 (2 mM) with 15-fold excess of TBHP at
–
20 ˚C in CH CN resulted in the rapid formation of a purple-colored transient species
3
-1
-1
which exhibited a broad absorption band with a maximum at 530 nm (ϵ  500 M cm )
1
7

ACCEPTED MANUSCRIPT
(Fig. 6) due to the charge transfer (LMCT) from the ligand to the high-spin iron(III)
center.[104] This assignment was consistent with reported spectroscopy of
alkylperoxoiron(III) complexes [18, 20-22, 97-100]. The reaction of TBHP with 3
III
t
2+
generating a deep purple species [Fe (L1)(OO Bu)] 3a was proposed and the overall
process was outlined in the eq (1):
Under pseudo-first-order conditions, excess TBHP (5–25 fold) was added to the
solution of 3 (2 mM) in CH CN at –20 ˚C. Pseudo-first-order fitting of the kinetic data
3
allowed us to determine kobs value for the reaction with TBHP (Fig. 7). The pseudo-first-
order rate constants were independent of concentrations of 3 and displayed linear
dependency on the concentration of TBHP. This led us to propose the rate law in eq (2)
-1 -1
and determine the second-order rate constants to be k = 0.40 M s at –20 ˚C.
2
Acquisition of kinetic data for the reaction of 3 (2 mM) with a 10-fold excess of
TBHP over a range of temperature (–40 to 0˚C) was performed for determining the
activation parameters of this process. An Eyring plot of the second-order rate constants
‡
was linear for the reaction and yielded the activation parameters of ΔH = 33.6 kJ/mol
‡
and ΔS = –107 J/mol K (Fig. 8). The moderately high activation enthalpy and large,
III
t
2+
negative activation entropy indicated the [Fe (L1)(OO Bu)] species 3a formed in an
associative process. Reaction (1) was a multi-step process that included a one-electron
1
8

ACCEPTED MANUSCRIPT
oxidation of Fe(II) into Fe(III) followed by the ligand substitution at the iron(III) center,
III t
eventually affording the (L)Fe - BuOOH species.[105, 106] While the activation
parameters for reaction (1) are consistent with the peroxide coordination to Fe(III), the
activation parameters for a multi-step process are composite values that likely reflect the
contributions from both the one-electron oxidation process and the subsequent ligand
binding process.
II
t
III
t
.
(
(
(
L)Fe + BuOOH →(L)Fe + BuOO
III
t
III
t
+
L)Fe + BuOOH → (L)Fe (OO Bu) + H
t
.
BuOO may abstract a hydrogen atom from the solvent, the ligand, or the substrate)
The purple intermediate 3a was also characterized by EPR and ESI–MS. An EPR
III
t
2+
spectrum of a frozen solutions of [Fe (L1)(OO Bu)] 3a in CH CN was obtained at 120
3
K (Fig. A.1). The intense purple intermediate exhibiting isotropic signals at g = 8.8 and
4
.3 indicated a high-spin iron(III) species (S = 5/2) with rhombic symmetry [18-23, 101];
the presence of additional high-spin iron(III) signals at g = 4.3 cannot be excluded. The
III
t
assignment to high-spin iron(III) was distinct from more common low-spin Fe –OO Bu
S = 1/2) with axial symmetry exhibiting g values of 2.28, 2.19 and 1.97 [17, 18, 24]. A
sharp signal at g = 2.0 possibly arose from organic radical. As reported, low-spin
(
II
2+
II
2+
[
Fe (TPA)(CH CN) ] and high-spin [Fe (6-Me -TPA)(CH CN) ] generated the
3 2 3 3 2
III
t
corresponding low-spin and high-spin Fe –OO Bu complexes, respectively [18].
Introduction of methyl substituents to the amine nitrogens of the ligand decreased the
ligand-field strength of the macrocycle and generated steric interactions to favor the high-
1
9

ACCEPTED MANUSCRIPT
III
t
spin state of the Fe –OO Bu complexes. Similar spin state changes upon alkylation of
macrocyclic nitrogen donors were reported for related iron(III) alkylperoxo complexes
[
96].
The ESI–MS spectrum was obtained by directly injecting the purple solution of
III
t
2+
[
Fe (L1)(OO Bu)] 3a in CH CN generated at –40 ˚C into mass spectrometer (Fig. A.2);
3
IV
2+
main peaks are listed in Table 5. Peaks m/z at 174.2, 497 correspond to [Fe (L1)(O)]
IV
+
and [Fe (L1)(O)(OTf)] , respectively. The oxoiron(IV) species was possibly formed
III
t
2+
from decay of [Fe (L1)(OO Bu)] 3a by O–O bond homolysis. This observation
III
t
demonstrated that the (L1)Fe O–O Bu intermediate underwent the homolytic cleavage of
the O–O bond under mass-spectrometric conditions, to yield a high-valent oxoiron(IV)
species 3b. In addition to the evidence of ESI–MS data, a low-intensity near-IR
absorption band at 800 nm may also indicate the formation of oxoiron(IV) [107, 108]
(Fig. 5). A similar homolytic O–O bond cleavage has been proposed for the
alkylperoxoiron(III) species with β-BPMCN and TPA ligands [21, 23], whereas the
III
t
homolytic LFe –OO Bu (iron-oxygen) bond cleavage was observed in other
alkylperoxoiron(III) species with 6-Me -TPA and Py(ProMe) ligands [22, 98].
3
2
III
t
2+
The purple [Fe (L1)(OO Bu)] species 3a did not react with cyclooctene at low
temperature. Therefore, other intermediate(s) must be responsible for olefin epoxidation
promoted by 3. The oxoiron(IV), that can result from a homolytic O–O bond cleavage in
III
t
LFe –OO Bu, was an obvious alternative for a potential reactive oxidant for olefin
epoxidation that should be considered. A direct way of generating oxoiron(IV) was
required for understanding the reactivity of this key intermediate.
2
0

ACCEPTED MANUSCRIPT
3
3
.4. Generation and reactivity of the oxoiron(IV) intermediates
IV
2+
.4.1. Formation of [Fe (L1)(O)] (3b) in the reaction of 3 with isopropyl 2-
iodoxybenzoate
In general, the formation of oxoiron(IV) complexes was accompanied by the
appearance of the low-energy d-d transition band in the near-IR region, characteristic for
the S = 1 oxoiron(IV) center [107, 108]; similar, but weaker near-IR absorption band was
IV
also observed for a high-spin (S = 2) Fe =O intermediate [109]. The formation of
oxoiron(IV) 3b from the iron(II) PyMAC complex 3 was indicated by an absorption band
at ca. 780 nm at low temperature (e.g., –40 ˚C) using H O , m-CPBA or isopropyl 2-
2
2
iodoxybenzoate as an oxidant (Fig. 5). The analysis of the UV–vis spectra of these
systems identified isopropyl 2-iodoxybenzoate as the most promising oxidant for clean
generation of the Fe(IV) species for subsequent reactivity studies. Indeed, the reaction of
1
with H O in CH CN exhibited a broad, but relatively weak absorption band at ca. 800
2 2 3
nm. With m-CPBA, the reaction showed a relatively strong absorption band at 780 nm
along with a less intense absorption at 550 nm, which can be tentatively assigned to an m-
CPBA adduct of iron(III) species; formation of two colored intermediates would
complicate the interpretation of kinetic data. Finally, the reaction of 3 with isopropyl 2-
iodoxybenzoate in CH CN monitored by a regular benchtop UV–vis spectrophotometer
3
equipped with an immersion probe (dip probe) generated a single species exhibiting an
-1
-1
absorption maximum at 780 nm (ϵ  50 M cm ). This rapid and clean reaction
(Scheme 1-(b)) was selected for detailed studies described below.
2
1

ACCEPTED MANUSCRIPT
Mass spectrometry is also consistent with the assignment of the yellow-green
intermediate generated in the reaction of 1 with isopropyl 2-iodoxybenzoate as the
iron(IV)-oxo species. This intermediate exhibited a peak m/z at 194.6 and 497.1 on ESI–
IV
2+
IV
+
MS, corresponding to [Fe (L1)(O)(CH CN)] and [Fe (L1)(O)(OTf)] , respectively
3
(Fig. 9). Additionally, a peak m/z at 497 slowly changed to a high relative abundance
III
+
peak at m/z = 498 which was assigned to [Fe (L1)(OH)(OTf)] , a decomposition product
of oxoiron(IV) species. This m/z 498 peak became predominant after 10 min, along with
the peak of a free ligand L1 (m/z at 277), suggesting the instability of the iron complex
(Fig. A.3). The resulting yellow acetonitrile solution precipitated a rust-colored solid
after one day at room temperature. ESI–MS showed the solution in CH CN mainly
3
contained the ligand, thus the precipitate was likely to be Fe O . A similar rust-colored
2
3
hydrolysis product was also obtained previously in a related system by adding TBHP to
II
8
8
high-spin [Fe (L py )(X
)] (L py = N,N’-bis(2-pyridylmethyl)-1,5-diazacyclooctane;
2
apical
2
X = triflate, benzoate or 4-methylbenzenethiolate) in CH Cl .[22] Relative instability of
2
2
3
b, combined with low yield of this species (as determined by low observed intensities of
the absorption bands in near-IR region) precluded a more detailed spectroscopic
characterization of this intermediate. A related intermediate obtained from 5 and
isopropyl 2-iodoxybenzoate also demonstrated UV–vis absorption peak at 705 nm (ε =
-1
-1
2
40 M cm ) and showed ferryl-derived peaks in the mass spectra analogous to the
ones observed for 3b. Moreover, the intermediate obtained from 5 and isopropyl 2-
iodoxybenzoate was characterized by Mössbauer spectroscopy, which unequivocally
showed an iron(IV) oxidation state of the metal center in this intermediate [84]. We
2
2

ACCEPTED MANUSCRIPT
propose that the intermediate 3b also contains a high-valent Fe(IV)-oxo species. This
formulation agrees with distinct UV-vis-NIR signatures of Fe(IV) species (low intensity
absorption band in the 700 – 800 nm region), and with mass spectroscopic results clearly
showing the peaks for (L)Fe=O intermediate, and is also consistent with the methods of
generating this intermediate from the Fe(II) precursor and several oxygen atom donors
(Figure 5) and its chemical reactivity discussed in detail below. In brief, a clean oxygen
1
8
atom transfer to substrates (phosphines or olefins) was observed; O label was
18
transferred to epoxide products in reactions with H O . Reactions of 3b with PPh yield
2
2
3
OPPh and an iron(II) complex 3. All these reactivity data are consistent with 3b being
3
being an iron(IV) species containing a transferable oxygen atom. In the absence of direct
spectroscopic data (e.g. Mössbauer spectra measured in applied magnetic field),
determining the spin state of iron(IV) in 3b is not possible at present.
Detailed studies of the reactivity of 3b were performed using an iodine(V) oxidant,
isopropyl 2-iodoxybenzoate, as an oxygen atom transfer reagent that cleanly reacted with
3
, forming 3b. Unlike other commonly used oxidants, such as PhIO or Me -PhIO (2,6-
2
dimethyliodosylbenzene) [110], isopropyl 2-iodoxybenzoate was soluble in CH CN (ca.
3
1
00 mM in CH CN at 25 ˚C). Due to the small value of the extinction coefficient of
3
oxoiron(IV) (3b), higher concentration of iron and oxidant solutions were needed to
obtain good time-resolved spectra for the stopped-flow measurements. Additionally,
most kinetic studies were performed at low temperature; thus, oxidants with good
solubility in CH CN were necessary to achieve the proper stoichiometry of reactants. The
3
reaction of excess isopropyl 2-iodoxybenzoate with 3 was monitored at –40 ˚C by
2
3

ACCEPTED MANUSCRIPT
stopped-flow spectroscopy and reveals an increase of the oxoiron(IV) species 3b at 780
nm (Fig. A.4).
A stopped-flow study of the reaction of 3 with isopropyl 2-iodobenzoate showed a
two-step process, with a very rapid (ca. 20 ms at – 40 ˚C) build-up of an absorption band
with λmax = 700 nm, which quickly converted into a much more stable band with λmax
=
7
80 nm (Fig. A.4). This shift of an absorption maximum of the iron(IV) species is likely
related to the coordination on the oxidant (isopropyl 2-iodobenzoate) or the product of its
partial reduction (an iodine(III) compound) to the initially formed Fe(IV) intermediate
(Scheme 2). When Fe(II) complex 3 was the limiting reactant (in reactions with 0.5 moles
or less of IOO oxidant per 1 mole of 3), both oxygen atoms of the iodine(V) oxidant were
eventually transferred to the iron center (the transfer of the first oxygen atom is very
rapid, while the second oxygen atom transfer is somewhat slower). Increasing the
concentration of the oxidant resulted in the expected increase in the values of kobs,
although limited series of the reactant concentrations could be explored due to low
extinction coefficient of 3b. Detailed kinetic analysis of the reaction between 3 and
isopropyl 2-iodoxybenzoate was not performed in this work; instead, the optimal
conditions of generating Fe(IV) intermediate were identified in single-mixing kinetic
experiments.
2
4

ACCEPTED MANUSCRIPT
3
.4.2. Reactions of Oxoiron(IV) 3b with PPh3
Using experimental conditions that allowed us to generate iron(IV)-oxo intermediate
3
b in single-mixing stopped-flow experiments described above, we investigated the
reactivity of this high-valent iron-oxo intermediate with substrates, such as PPh or
3
olefin (cyclooctene). PPh can be oxidized easily by isopropyl 2-iodoxybenzoate.
3
Therefore, less than 1 equivalent of isopropyl 2-iodoxybenzoate was used in the
formation of oxoiron(IV) species in order to avoid excess oxidant in subsequent
reactions with substrates.
Double-mixing stopped-flow experiments were performed to gain mechanistic insight
IV
2+
into the reactivity of [Fe (L1)(O)] 3b with substrates. In the first mixing step, solutions
of 3 (16 mM) and 0.3 equiv isopropyl 2-iodoxybenzoate in CH CN were combined at –
3
2
0 ˚C and incubated for 500 s to ensure maximum formation of oxoiron(IV). Then, a
solution of PPh in CH CN was added in the second mixing. Changes in absorbance at λ
3
3
=
780 nm are monitored. As shown in Fig. 10, when 3 equiv PPh (48 mM) was added to
3
the oxoiron(IV) solution, the reaction proceeds rapidly. It was completed within seconds
at –20 ˚C, yielding Ph PO quantitatively, which was confirmed by GC–MS. This implies
3
a fast oxygen atom transfer in the process. ESI–MS of this resulting solution shows a
IV
2+
predominant peak m/z at 481 indicating that the [Fe (L1)(O)] species 3b was converted
back to the starting complex 3 (Fig. A.5). These results, taken together, indicate a fast
IV
2+
oxygen atom transfer in the process. Moreover, generation of [Fe (L1)(O)] (3b) and its
decay upon addition of PPh can be repeated several times and monitored by regular
3
2
5

ACCEPTED MANUSCRIPT
spectrophotometry with an immersion probe (Fig. A.6), suggesting that the oxo transfer
could afford several catalytic cycles.
Acquisition of kinetic data for the reaction of 3b with 3 equiv of PPh over a
3
temperature range from –30 to 0 ˚C allowed us to determine activation parameters. An
Arrhenius plot was linear for the reaction (Fig. 11), and yielded rather high activation
‡
energy (Ea = 53 kJ/mol); this values exceeds the activation energy of cyclooctene
epoxidation discussed below. It can be concluded that the oxygen atom transfer from
oxoiron(IV) in 3b to PPh is a high-barrier process.
3
3
.4.3. Reactions of oxoiron(IV) 3b with cyclooctene
IV
2+
The pre-mixed solution containing [Fe (L1)(O)] 3b was generated in the stopped-
flow mixing cell as described above. The reaction of excess cyclooctene with 3b was
monitored at 20 ˚C under pseudo-first-order conditions by the stopped-flow
spectrophotometry and revealed a smooth decay of the oxoiron(IV) species. In a control
reaction, in which 3b was mixed with CH CN, little change in optical absorbance at 780
3
nm was observed (Fig. 12-a). The pseudo-first-order fitting of the kinetic data in the
reaction of 3b with cyclooctene allowed us to determine kobs values which increase
linearly with cyclooctene concentration, yielding the second-order rate constant k =
2
-1 -1
0
.032 M s at 20 ˚C (Fig. 12-b); the proposed mixed second-order rate law is shown in
eq (3). This rate law was consistent with a bimolecular reaction pathway depicted in
equation (4); epoxide products were quantitated by GC–MS in separate benchtop
experiments. A plot of the temperature dependence (25 to 35 ˚C; experimental difficulties
2
6

ACCEPTED MANUSCRIPT
did not allow us to perform the measurements over broader temperature range) for the
IV
2+
reaction of [Fe (L1)(O)] 3b with 60 equiv cyclooctene in CH CN fits the linear form
3
‡
‡
of the Eyring equation, giving the activation parameters of ΔH = 60 kJ/mol and ΔS = –
2
6
9 J/mol K (R = 0.9993) (Fig. 12-c). Thus, an oxygen atom transfer to cyclooctene from
oxoiron(IV) is an associative process characterized by a large negative activation entropy.
The ability of iron(IV) oxo intermediate 3b to transfer an oxygen atom to cyclooctene
under anaerobic conditions determined in kinetic experiments was mirrored by the
formation of cyclooctene epoxide in benchtop experiments. The organic oxidation
products were analyzed by gas chromatography – mass-spectrometry (GC–MS). For
example, mixing 1 mole of 3 with 0.3 moles of isopropyl 2-iodoxybenzoate at room
temperature, followed by the addition of 20 moles of cyclooctene afforded 0.3 moles of
epoxide (0.3 turnovers with respect to iron complex; 50% yield with respect to limiting
reactant of isopropyl 2-iodoxybenzoate). Low concentration of the oxidant and its rapid
(within seconds) consumption in the initial reaction with 3 ensured the absence of excess
isopropyl iodoxybenzoate in the process of cyclooctene eoxidation. It required about 15
min to reach the maximum yield of epoxide in this stoichiometric reaction of cyclooctene
and isopropyl 2-iodoxybenzoate in the presence of 3. This time scale was comparable to
the decay time of 3b in the presence of cyclooctene substrate. Additionally, no diol was
2
7

ACCEPTED MANUSCRIPT
produced, providing indirect evidence that the diol formation needed H O (and/or H O )
2
2
2
during the epoxidation/dihydroxylation reaction. These experiments demonstrate that an
iron(IV) oxo intermediate 3b is a competent oxo-transfer reactant in cyclooctene
epoxidation. This behavior is similar to the previously reported reactivity of ferryl
intermediates supported by aminopyridine ligands (such as TPA [60]), but differs from
the lack of reactivity of iron(IV) species bound to polyamine macrocycles [64]. It is
becoming increasingly clear that the reactivity of high-valent iron-oxo species can be
altered by the exact structure of the first coordination sphere of the metal [15, 40, 41,
1
11]. More electron donating tetraaza macrocycles stabilized ferryl species and rendered
them unreactive; more electron-withdrawing ligands bearing several pyridine rings
decreased the stability of high-valent iron-oxo intermediates and increased the reactivity
of these species with oxygen atom acceptors. The ligand L1 bears one pyridine rings and
three tertiary amine donors, and its ability to support reactive iron(IV)-oxo intermediates
falls between the reactivities of the ligands with several pyridine donors and those with
several amine donors.
The ability of complex 3 to support a high-valent iron(IV)-oxo intermediate that reacts
(albeit slowly) with cyclooctene did not lead to high catalytic activity of this complex in
IV
2+
epoxidations. Unfortunately, upon addition of cyclooctene to [(L1)Fe =O] (3b), this
iron(IV)-oxo intermediate did not convert back to the iron(II) precursor 3. Instead, 3b
decomposed into an as-of-yet uncharacterized iron(III) complex (that presumably results
from comproportionation of the iron(IV) species 3b and the iron(II) complex 3) and the
free ligand L1, that was detected by ESI–MS. This indicated that the low turnover
2
8

ACCEPTED MANUSCRIPT
numbers and efficiency of epoxidation with 3 was due to the iron dissociation from the
macrocycle. In contrast, as we have described above, several catalytic cycles can be
observed with stronger reducing agents, such as PPh , as can be estimated from the fast
3
observed rate of reaction between 3b and triarylphosphine substrate. Low catalytic
activity of the N-methylated iron macrocyclic complexes compared to their non-
methylated secondary amine precursors parallels the observations previously reported in
the literature. For example, Nam and Valentine [112] discovered catalytic activity of
iron(II) cyclam complex in epoxidation of olefins with hydrogen peroxide, while
analogous iron complex with tetramethylcyclam (TMC) was inactive. Interestingly,
tetramethylcyclam supported a relatively stable, crystallographically characterized
ferryl(IV) complex[52, 64]. Iron complex with tetramethylcyclam also supports a side-on
peroxo adduct that does not act as an active oxidant, and an end-on hydroperoxo species
that displays both nucleophilic and electrophilic reactivity with substrates, and generates
an iron(IV)-oxo species via hemolytic O-O bond cleavage [43]. It can be concluded that
the ability of the polydentate ligand to stabilize iron(IV)-oxo intermediates is insufficient
for designing highly reactive catalysts of oxygen atom transfer to olefin substrates; the
reactivity of the iron(IV)-oxo intermediates is the key. The presence of secondary
aminogroups in the immediate vicinity of the redox center appears to facilitate an oxygen
atom transfer to substrates;[84] similar second-sphere effects may increase the reactivity
of iron(IV) intermediates in non-heme iron enzymes [113, 114].
As we mentioned earlier, 4 did not promote olefin epoxidation with hydrogen peroxide
or TBHP. It is possible that the oxoiron(IV) intermediate cannot be formed because of the
2
9